Detroit Sexual Assault Kit Action Research Projectort

The author(s) shown below used Federal funds provided by the U.S.

Department of Justice and prepared the following final report:

Document Title:

The Detroit Sexual Assault Kit (SAK) Action

Research Project (ARP), Final Report
Author(s):
Rebecca Campbell, Ph.D., Giannina FehlerCabral, Ph.D., Steven J. Pierce, Ph.D., Dhruv B.
Sharma, Ph.D., Deborah Bybee, Ph.D., Jessica
Shaw, Ph.D., Sheena Horsford, Ph.D., Hannah
Feeney, B.A.
Document No.:
248680
Date Received:
March 2015
Award Number:
2011-DN-BX-0001
This report has not been published by the U.S. Department of Justice.
To provide better customer service, NCJRS has made this federally
funded grant report available electronically.
Opinions or points of view expressed are those

of the author(s) and do not necessarily reflect
the official position or policies of the U.S.
Department of Justice.
THE DETROIT
SEXUAL ASSAULT KIT (SAK)
ACTION RESEARCH PROJECT (ARP)
2011-DN-BX-0001
FINAL REPORT
February 25, 2015
Principal Investigator
Rebecca Campbell, Ph.D.

Department of Psychology
Michigan State University
Co-Investigator
Giannina Fehler-Cabral, Ph.D.

Harder+Company Community Research
Los Angeles, CA
Statisticians
Steven J. Pierce, Ph.D.

Dhruv B. Sharma, Ph.D.
Center for Statistical Training & Consulting
Deborah Bybee, Ph.D.

Research Associates
Jessica Shaw, Ph.D.

Sheena Horsford, Ph.D.
Hannah Feeney, B.A.
ii
ACKNOWLEDGEMENTS
ThisfinalreportforNIJAward2011DNBX0001waswrittenbytheresearch/evaluationteamat
MichiganStateUniversity,basedonourresearchfindingsfromtheDetroitSexualAssaultKit(SAK)
ActionResearchProject(ARP).Theviewsrepresentedinthisreportarethoseoftheauthorsanddonot
reflecttheofficialpositionsofanyparticipatingorganizationsortheNationalInstituteofJustice.
AswillbeexplainedinChapter1ofthisreport,thenamesoftheorganizationsandindividualswho
participatedinthisactionresearchprojectwillnotbereleased.Therefore,theseAcknowledgements
cannotspecificallynamethemanypeopleandgroupswhoareduethanksfortheirdilligenceand
commitmenttothisproject.Withthoseparametersinmind,thePrincipalInvestigatorofthe
research/evaluationteamwishestoacknowledgesthefollowing:
TotheMembersoftheDetroitSexualAssaultKitActionResearchProject,Iamhumbledbyyour
dedicationtothiscityandallitscitizens,especiallythosemostvulnerable.Iappreciatethetrustyou
placedinmeandinthisproject.Ithasbeenanhonorworkingwithallofyou.
ToourColleaguesattheNationalInstituteofJustice,Iappreciateyourgudianceandcounsel
throughoutthisprojectandyourcommitmenttoactionresearchasavehicleforimprovingthecriminal
justicesystemresponsetosexualviolence.
TomyStatisticanColleagues,Iambeyondgratefulthatyouagreedtotakeonthisprojectandallits
complexities.Youwentwayaboveandbeyondtoensurethatthisprojectwouldbehelpfulto
practitioners,policymakers,andsexualassaultsurvivors.
TomyCoInvestigatorandResearchAssociates,Iamsofortunatethateachofyoudecidedtojumpinto
thiswithme.IcannotpossiblylisteverythingIamthankfulfor,butIdoknowthatwithoutallofyou,
thiswouldnothavebeenpossible.
This document is a research report submitted to the U.S. Department of Justice. This report has not
been published by the Department. Opinions or points of view expressed are those of the author(s)
and do not necessarily reflect the official position or policies of the U.S. Department of Justice.
iii
EXECUTIVE SUMMARY
Detroit,MichiganisoneofagrowingnumberofU.S.citiesthathavelargenumbersofuntested
sexualassaultkits(SAKs).InAugust2009,representativesfromthelocalpolice,statepolice,andthe
prosecutorsofficetouredaremotepropertystoragefacilitytodiscusshowtobestmanagethevolume
ofevidenceinpolicecustody.Duringthattour,anassistantprosecutornoticedalargenumberof
storageboxesonshelvingunits,andwhenaskedwhattheywere,policepersonnelindicatedthatthey
wererapekits.Whenpressedfordetailsaboutthekits,policeofficialswerenotabletoverifyhow
manySAKswereinpolicepropertyandhowmanyofthoseSAKshadbeentested.
Todeveloplongtermstrategiesforresolvingthisproblem,amultidisciplinaryactionresearch
projectwascreated,TheDetroitSexualAssaultKit(SAK)ActionResearchProject(ARP),whichbrought
togetherresearchersandpractitionersfromlawenforcement,prosecution,forensicsciences,forensic
nursing,andvictimadvocacytoaddressfourprimarygoals:
1) ToassessthescopeoftheproblembyconductingacompletecensusofallSAKsinpoliceproperty;
2) ToidentifytheunderlyingfactorsthatcontributedtowhyDetroithadsomanyunsubmittedSAKs;
3) TodevelopaplanfortestingSAKsandtoevaluatetheefficacyofthatplan;
4) Tocreateavictimnotificationprotocolandevaluatetheefficacyofthatprotocol.
Thefirstgoalofthisprojectwastoassessthescopeoftheproblembyconductingacensusofall
SAKsinpoliceproperty(currenttoNovember1,2009).Thecensustook15weekstocompleteand
revealedthattherewere11,303SAKsinpolicecustody.Postcensusreviewofpropertyrecords
indicatedthat84SAKsneededtoberemovedfromthecount(typicallybecausetheSAKdidnotcontain
sexualassaultmedicalforensicevidence;theboxhadbeenusedtostoreothertypesofcrimescene
evidence),therebyrevisingthecensuscountto11,219.Someofthesekits(2,512)hadlaboratoryID
iv
numbers,indicatingthattheyhadbeensubmittedfortesting,butitwasunclearhowmanyhadinfact
beentestedforDNA.ThevastmajorityoftheSAKsinpoliceproperty(8,717)hadnotbeensubmitted
forforensictesting.TheDetroitSAKARPcreatedastepbystepsummaryofthecensusprocedures
usedinthisprojecttoguideotherjurisdictionsonhowtoconductacensusofSAKsinpoliceproperty.
ThesecondgoalofthisactionresearchprojectwastostudytheunderlyingreasonswhyDetroit
hadsomanyunsubmittedSAKs.Theresearch/evaluationteaminterviewedcurrentandformer
employeesinallorganizations,andexaminedpubliclyavailabledocumentsandinternalorganizational
recordstoassesstheresourcesavailableforservingrapevictims(ingeneral)andtestingSAKs
(specifically).TheresultsofthishistoricalcontextualanalysisindicatedthatallorganizationsinDetroit
thatserverapevictimshavestruggledfordecadeswithchronicunderstaffingandresourcedepletion
relativetootherU.S.citieswithsimilarpopulations,racial/ethniccompositions,and/orcrimerates.
Policepersonnelacknowledgedthatbudgetandstaffingcutscompromisedinvestigationquality
suchthatcuttingcornersbecamenormative.Ananalysisof1,268sexualassaultpolicereports
associatedwithSAKsthathadnotbeensubmittedfortestingrevealedthatmostcaseswereclosedafter
minimalinvestigationaleffort.Inboththestakeholderinterviewsandintheactualpolicereports,law
enforcementpersonnelexpressednegative,victimblamingbeliefsaboutsexualassaultvictims.Rape
survivorswereoftenassumedtobeprostitutesandthereforewhathadhappenedtothemwas
consideredtobetheirownfault.Adolescentswereassumedtobelying,tryingtoavoidgettinginto
troublebyconcoctingafalsestoryaboutbeingraped.Policesaidthatthosewhohadbeenassaultedby
friendsandacquaintanceshadgotwhattheygotbecausetheyhadchosentoassociatewiththe
perpetrator.Caseaftercasewaslabeledadealgonebadorotherwisedismissedasnotreallya
rape,andtheseattitudesdirectlyaffectedlawenforcementpersonnelsdecisionsregardingwhetherto
submitarapekitforforensictesting.Withoutconsistentsupervisionandtrainingtochallengethese
practices,unsubmittedSAKscontinuedtoaccumulate.Thisresearchidentifiedindividuallevel,
organizationallevel,andsystemicfactorsthatmayplacecommunitiesatriskfordevelopingthis
problem(i.e.,stockpilesofuntestedrapekits),whichcanhelpotherjurisdictionstakestockoftheir
pastandpresentpracticesregardingsexualassaultinvestigationsandSAKtesting.
Thethirdgoalofthisprojectwastodevelopandevaluateaplanfortestingtheseunsubmitted
SAKs.Atthebeginningofthisproject,DetroitdidnothavesufficientfundingtotestallSAKsinpolice
property.However,practitionersdisagreedastowhetherallSAKsshouldbetested,eveniffundswere
available.IsitusefultotestaSAKiftheassailantisalreadyknown(nonstranger)?Doesitmakesense
totestaSAKifthecaseisbeyondthestatuteoflimitations?Thesesamequestionscameupinour
researchinterviewswithstateandnationalstakeholdersfromthelawenforcement,prosecution,
forensicscience,medical/nursing,andvictimadvocacy,sotheDetroitSAKTestingplanwasdesignedto
gatherdataaboutthesefundamentalquestions.PoolingfundsfromtheDetroitSAKARPbudget,the
statepolicedepartmentsNIJDNABacklogReductionGrants,andtheresourcesofauniversitybased
forensiclaboratory(whichwasseparatelyfundedbyNIJ),theprojectwasabletotest1,600SAKs(1,595
actuallytested).KitswererandomlysampledandplacedintofourTestingGroups,eachonedesignedto
addressspecificresearchquestionsregardingtheutilityofSAKtestingunderdifferentcase
circumstances.ThisdesignallowedustoexaminetheutilityofSAKtestingforstrangerperpetrated
sexualassaults(TestingGroup1),nonstrangerperpetratedsexualassaults(TestingGroup2),andsexual
assaultcasesthatwerepresumedtobebeyondthestatuteoflimitations(SOL)(TestingGroup3).For
TestingGroup4,SAKswererandomlyassignedtotwodifferentDNAtestingmethodstoexamine
whetheranemergingtestingmethod,selectivedegradation,couldofferfaster,lessexpensivetesting
options,withoutsacrificingaccuracy(relativetotraditionalDNAtestingmethods).AllTestingGroups
werecomparedwithrespecttotheirratesofCODIS(CombinedDNAIndexSystem)entries(theSAK
containedaDNAeligibleprofileforCODIS),CODIShits(aDNAmatchtoaprofileinCODIS),andserial
sexualassaulthits(aDNAmatchacrosstwoormoreSAKs).
vi
Inthisproject,1,595SAKsweretested,whichyielded785CODISeligibleprofiles(49%ofthe
SAKstested),455CODIShits(28.5%oftheSAKstested;58%oftheprofilesentered),and127serial
sexualassaults(8%oftheSAKstested;28%oftheCODIShits).Aseriesofstatisticalmodelswere
evaluatedthatcomparedtheprobabilitiesofCODISentryrates,CODIShitrates,andserialsexualassault
hitrates,accountingforthesamplingdifferencesbetweenthefourTestingGroups.Usingcontinuation
ratiomodels,theresultsfromtheconditionalandunconditionalprobabilityrateanalysesshowedno
significantdifferenceinCODIShitratesasafunctionofeithervictimoffenderrelationshiporSOLstatus.
Inotherwords,SAKsassociatedwithcasesthatwerestrangerperpetratedhadstatisticallyequivalent
CODIShitratesascasesperpetratedbynonstrangers;similarly,ratesdidnotsignificantlydifferby
statuteoflimitationsstatus.SomestakeholdersinDetroit(aswellasthoseatthestateandnational
level)advocatedforprioritizingSAKsfortestingbyvictimoffenderrelationship(toprioritizestranger
perpetratedcrimes)and/orskippingoverSAKsassociatedwithcasesthatarepresumedtobebeyond
thestatuteoflimitations;however,theseresultsdonotsupportsuchaplanbecausetheratesofCODIS
hitsdonotsignificantlydifferasafunctionofthesevariables.Theseresultsindicatethatthereismerit
intestingbothstrangerandnonstrangerSAKs,andpresumedSOLexpiredandnonexpiredSAKs,in
termsofexpectedyieldsforCODISentries,CODIShits,andidentificationofserialsexualassaults.
InTestingGroup4,twodifferentmethodsofDNAtestingwerecompared:traditionalvs.
selectivedegradation.TherewasnosignificantdifferencebetweenthetwogroupsinCODSentryrates,
indicatingthattheselectivedegradationmethodhadnodecrementinperformancerelativeto
customarymethods.Comparisonsofmaterialscostswerealsoequivalentacrossthetwogroups,butthe
selectivedegradationmethodsaved1.10hoursofstafftimeperSAK.Thesesavings,whenaggregated
acrossalargecollectionofSAKs,maysubstantiallyreducepersonnelcosts.Theseresultsmerit
replicationpriortobroadbasedimplementation.
vii
ThefourthgoaloftheDetroitSAKARPwastodevelopandevaluateavictimnotification
protocol.TheDetroitcollaborativehadatwodayplanningretreattodevelopavictimcentered,
traumainformednotificationprotocol(withanaccompanyingstepbystepguideandsampleFAQ
documentsforotherjurisdictions).Theprotocolstipulatedthatamultidisciplinaryteamwouldreview
casesthathadCODIShitsanddiscussifandhowtonotifysurvivors,giventhespecificcircumstancesof
eachcase.Thenotificationswouldproceedinatwostageprocess,wherebythegoalsofthefirstcontact
weretoexplaintothevictimthather/hisrapekithadnotbeentestedatthetimes/hemadethepolice
report,butnowithadbeenfoundandtested;offeranapologytothesurvivorthather/hisSAKhadnot
beentested;andofferafollowupmeetingtodiscusstheissuesinmoredetail.Atthatsecond,follow
upmeeting,aninvestigatorandcommunitybasedadvocatewouldprovidemoredetailedinformation,
discussoptions,andconnectthesurvivortocommunityservices.
Intheevaluationofthisprotocol,41caseswereselectedfornotificationbythemultidisciplinary
reviewteam,andtheinvestigatorswereabletofind31survivors(2caseswereclosedoutbecausethe
investigatorshadexhaustedallpossibleleadstryingtofindthevictims;8caseswerestillpendingatthe
timetheevaluationdatacollectionperiodclosed)(95%findrate).Theaveragelengthoftimebetween
whentheassaultoccurredandthetimeofnotificationwasnineyears.Mostsurvivors(65%)couldbe
foundwithrelativelylowinvestigationaleffort:databasessearches(e.g.,LEINLawEnforcement
InformationNetwork),plus04phonecalls,and01inpersonvisitsto1address.Survivorswhowere
hardertolocatewantedtoparticipateintheprosecutionoftheircasesatacomparableratetothose
whowereeasiertofind,suggestingthatvictimslocateabilityshouldnotbeaselectioncriterionfor
eitherSAKtestingorvictimnotification.
Thefirstcontactwiththesurvivorswasmadebyinvestigatorsaffiliatedwiththeprosecutors
office(notthefocalpolicedepartment)andtypicallyoccurredatthevictimshomes.Somesurvivors
hadstrongnegativereactions(16%)(e.g.,anger,refusaltotalktoinvestigators),morehadstrong
viii
positivereactions(29%)(e.g.,happiness,relief),andmost(55%)didnotexhibitstrongemotional
reactionstheywereopentohearingwhattheinvestigatorshadtosay,butwerereservedand
cautious.Mostsurvivors(64%)wantedafollowupmeetingwiththeinvestigatorsandanadvocateto
discussoptionsinmoredetail,andintheend,most(57%)alsodecidedthattheywantedtoparticipate
intheinvestigationandprosecutionprocess.Thisrateofreengagementwashigherthanexpected
giventhepervasivevictimblamingtreatmentmanysurvivorshadexperiencedfromlawenforcement
personnelatthetimetheyhadfiledthepolicereport.
Victimswerelesslikelytoreactpositivelyandtoreengagethelongertheperiodoftime
betweentheassaultandthenotification(beyondnineyears),whichhighlightstheimportanceoftimely
testingofSAKsandinvestigationofreportedsexualassaults.Survivorswhowere1624yearsoldatthe
timeoftheassaultweresomewhatmorelikelytohavehadnegativereactionstothenotificationand
weresomewhatlesslikelytowanttohavecontinuedcontactwiththecriminaljusticesystem.Given
thatpriorresearchhasfoundthatvictimsinthisagegroupareathighriskforvictimblamingtreatment,
thesegirls/youngwomenmayhavehaddifficultencountersyearsago,andassuch,mayhavebeen
disinclinedtoreengage.Inthisevaluation,onlyasmallnumberofnotificationswereconductedwith
victimsofnonstrangerrape,butpreliminaryfindingssuggestedthattheywerenotaslikelyasvictimsof
strangerrapetocontinuecontactwiththecriminaljusticesystempostnotification.
Theresultsofthisprojectwereinfluentialincreatinganumberofsignificantchangesinpolicy
andpractice,including,butnotlimitedto:apolicychangeinthelocalpolicedepartmenttosubmitall
SAKsforforensictesting;trainingforpoliceandotherpractitionersonvictimcentered,trauma
informedservicesandoffenderfocusedinvestigations;securing$4millionfromthestateAttorney
GeneralsOfficetotestasmanyremainingDetroitSAKsaspossible;andthepassageofnewstatewide
legislationrequiringalllawenforcementagenciesinthestateofMichigantosubmitSAKsfortesting(if
releasedfortestingbytherapevictim)(theSexualAssaultKitEvidenceSubmissionAct(PA227)).
ix
TABLE OF CONTENTS
CHAPTER 1: Introduction
The Detroit Sexual Assault Kit (SAK) Action Research Project (ARP)
UnsubmittedSexualAssaultKits(SAKs):AGrowingNationalProblem

AnActionResearchApproachtotheProblemofUnsubmittedSAKs
OneCitysShockingDiscovery:TheProblemofUnsubmittedSAKsinDetroit
AnOverviewoftheDetroitSexualAssaultKit(SAK)ActionResearchProject(ARP)
ProjectGoals
11
13
13
TheCollaborativeTeam
TheCollaborativePartnershipProcess
Research/EvaluationComponent
14
16
16
ASpecialNoteAboutIdentity,Confidentiality,andPrivacy
AboutThisReport
27
33
CHAPTER 2: The Scope of the Problem

How Many Unsubmitted Sexual Assault Kits (SAKs) In Detroit
TheDiscoveryoftheKitsandInitiatingaCensus

TheProcessofConductingtheCensusinDetroit
TheResultsoftheCensus:HowManyUnsubmittedSAKsinDetroit
Summary&Conclusions:TheScopeoftheProblem
36
39
47
56
58
CHAPTER 3: Underlying Reasons

Why So Many Unsubmitted Sexual Assault Kits (SAKs) in Detroit
EcologicalSystemsTheory:UnderstandingInterdependentOrganizations
HistoricalContext:OrganizationalResources&Practices19802009
Overview
HistoryofDNATestingandCODISAccess19802009
PoliceDepartmentResources&Policies19802009
CrimeLaboratoryResources19802009
ProsecutorsOfficeResources19802009
MedicalSystemResources19802009
VictimAdvocacyResources19802009
TheImpactandLegacyofChronicResourceDepletion
UnderlyingFactors:FrontLinePractices&InterOrganizationalCommunications
Overview
PoliceDepartmentFrontLineDecisionMaking&Practices
ThePoliceandtheCrimeLab:IntraOrganizationalPractices&Communication
ThePolice,theCrimeLabandtheProsecutorsOffice:InterOrganizational
Communication
ThePoliceandtheMedicalSystem:MinimalInterOrganizationalCommunication
ThePoliceandVictimAdvocacyOrganizations:Intra&InterOrganizational
Communication
Summary&Conclusions:WhySoManyUnsubmittedSAKsinDetroit
60
62
68
68
73
79
84
87
91
95
98
101
101
104
118
123
128
132
135
xi
CHAPTER 4: Testing Kits

Developing & Evaluating a Sexual Assault Kit (SAK) Testing Plan
138
141
TheProcessofDevelopingaSAKTestingPlan
ImplementingtheDetroitSAKTestingPlan
149
Overview
149
DefiningtheSamplingCriteriaforSAKSelection
150
SelectingSAKsfortheTestingGroups
152
ChallengesImplementingtheTestingPlanandCoordinatingPostTestingActivities 155
EvaluatingtheDetroitSAKTestingPlanDescriptiveFindingsAbouttheCases/SAKs
TestedinthisProject
160
EvaluatingtheDetroitSAKTestingPlanDescriptiveFindingsontheForensic
TestingOutcomes
168
BackgroundContext:AnOverviewofDNATestingandCODIS
168
DescriptiveResults:CODISHits&SerialSexualAssaultsintheOverallSample
173
DescriptiveResults:CODISHits&SerialSexualAssaults,byTestingGroup
176
EvaluatingtheDetroitSAKTestingPlanInferentialFindingsontheEffectofVictimOffender
RelationshipandStatuteofLimitationsStatusonForensicTestingOutcomes
187
Overview
187
VictimOffenderRelationshipEffectonForensicTestingOutcomes:
Stranger&NonStrangerSexualAssaults
194
StatuteofLimitationsEffectonForensicTestingOutcomes:PresumedSOLExpired
andUnexpiredSAKs
206
xii
EvaluatingtheDetroitSAKTestingPlanInferentialFindingsontheEffectof
DNATestingMethodonForensicTestingOutcomes
Overview
EffectofDNATestingMethodonDNATestingRatesandCODISEntryRates
ComparisonsofTestingCostsandPersonalEffortBetweenTestingMethods
Summary&Conclusions:EmpiricalFindingsRegardingtheUtilityofSAKTesting
CHAPTER 5: Notifying Victims

Developing & Evaluating a Victim Notification Protocol
213
213
214
226
228
232
233
TheProcessofDevelopingtheVictimNotificationProtocol
TheDetroitSAKActionResearchProjectPilotVictimNotificationProtocol
245
GuidingPrinciples
245
TheDetroitPilotVictimNotificationProtocol
246
ChallengesImplementingthePilotVictimNotificationProtocol
250
EvaluatingthePilotVictimNotificationProtocol
256
DevelopingtheEvaluationDesign
256
TheDecisiontoNotify:SelectingCasesforVictimNotification
258
LocatingVictims:StrategiesandSuccessRates
259
TheFirstNotificationContact:VictimsEmotionalReactionstotheInvestigators
264
TheSecondNotificationContact:VictimsDecisionsRegardingFurtherInvolvement
withtheCriminalJusticeSystem
270
InvestigatorsandAdvocatesReflectionsontheVictimNotifications
275
Summary&Conclusions:AVictimCentered,TraumaInformedApproach
toVictimNotification
278
xiii
CHAPTER 6: Discussion
Summary of Findings, Implications, and Community Changes
MajorFindingsoftheDetroitSAKActionResearchProject
Goal1:DetermineHowManyUnsubmittedSAKsinDetroitThrougha
CompleteCensus
Goal2:ExamineWhyDetroitHadSoManyUnsubmittedSAKsinPoliceProperty
Goal3:DevelopandEvaluateaSAKTestingPlan
Goal4:DevelopandEvaluateaSAKVictimNotificationProtocol
CreatingaMultidisciplinaryTeamtoAddresstheProblemofUnsubmittedSAKs:
OverallProjectLessonsLearnedandImplications
WhatsDifferentNow:TheOutcomesoftheDetroitSAKActionResearchProject
EvaluatingtheSuccessofanActionResearchProject
EvidenceofProcessUse:ChangesinStakeholdersKnowledge,Attitudes,and
BeliefsaboutResearchandEvaluation
EvidenceofConceptualUse:ChangesinStakeholdersAttitudesandBeliefsabout
SexualAssaultandSAKTesting
EvidenceofInstrumentalUse:ChangesinPolicy&Practice
Summary&Conclusions
APPENDIX A:
References
281
281
281
292
302
315
327
333
333
336
340
343
350
351
xiv
APPENDIX B: Methodology
Design, Sampling, Procedures, & Analysis
377
GuidingEvaluationTheory:DevelopmentalEvaluationTheory
GuidingResearchDesign:SequentialExploratoryMixedMethodsDesign
DataCollectedintheDetroitSAKActionResearchProject
EthnographicObservations
IndividualInterviews
ArchivalRecords
FocusGroups
QualitativeDataAnalysis
DataPreparation&TriangulationAssessments
DataCoding&AnalysisProcedures
Validity(Trustworthiness)Assessments
QuantitativeDataAnalysesModelingSAKSubmissionRatesOverTime
DataAnalyticMethods
DetailedStatisticalResults
QuantitativeDataAnalysesModelingCODISHitRates
DataAnalyticMethods
DetailedStatisticalResults
382
385
388
388
390
395
399
400
400
402
405
411
411
415
426
426
440
xv
APPENDIX C: Victim Notification Resources
449
AppendixC1:SamplePlanningToolforDevelopingaVictimNotificationProtocol
AppendixC2:SampleFrequentlyAskedQuestions(FAQ)ResourcePacket
AppendixC3:SampleVictim/SurvivorCommunityResourceBrochure
APPENDIX D: Data Collection Instruments
AppendixD1:InterviewProtocolforDetroitStakeholders(FirstInterview)
AppendixD2:InterviewProtocolforDetroitStakeholders(SecondInterview)
AppendixD3:InterviewProtocolforNationalStakeholdersfromCriminalJustice/
ForensicScienceandViolenceAgainstWomenSocialServiceOrganizations
AppendixD4:QuestionsAskedofPublicOfficialsinComparableCities
AppendixD5:PoliceReportCodingSheetsforVictim,Assailant,andCaseCharacteristics
AppendixD6:ForensicTestingOutcomesCodingSheets
AppendixD7:LawEnforcementTrackingSheetsforVictimNotification
AppendixD8:CommunityBasedAdvocatesTrackingSheetsforVictimNotification
AppendixD9:FocusGroupProtocolforDetroitStakeholders
450
451
454
456
458
460
461
462
463
464
xvi
LIST OF TABLES
Table1.1OverviewoftheDetroitSAKARPDataCollectionMethods
Table1.2OverviewoftheQuantityofDataCollectedintheDetroitSAKARP
23
26

Table2.1SAKsinPoliceProperty:UnderstandingTestingStatusandAdjudicationStatus
37

Table3.1ComparableCities:Population,Race/Ethnicity,andUCRCrimeRate
Table3.2DistributionofSAKsOverTime,ByDNA/CODISEra
Table3.3MultiLevelLogisticRegressionResults,PredictingSAKSubmissionby
DNA/CODISEra
Table3.4ComparableCities:Mayoral&PoliceDepartmentLeadership
Table3.5ComparableCities:NumberofDNAScientists
Table3.6ComparableCities:ProsecutorLeadershipTurnover
Table3.7ComparableCities:SpecializedProsecutionPrograms
Table3.8ComparableCities:SpecializedSANE/SAFEServices
Table3.9MultiLevelLogisticRegressionResults,PredictingSAKSubmissionby
ImplementationofSANEProgram
Table3.10ComparableCities:VictimAdvocacyServices
72
76
78
80
85
87
90
92
94
95
xvii

Table4.1SummaryofSamplingCriteria(asImplemented)fortheSAK
TestingGroups(N=1,595)
154
Table4.2TheDetroitSAKs:Victim,Assailant,andAssaultCharacteristics
164
Table4.3RatesEstimatedfromContinuationRatioModelExaminingEffectofVictimOffender
RelationshiponTestingOutcomes
196
Table4.4LogisticRegressionPredictingCODISHitsWithinTestingGroup1
(StrangerRapes)
201
(NonStrangerRapes)
204
Table4.6RatesEstimatedfromContinuationRatioModelExaminingEffectofStatuteof
LimitationsonTestingOutcomes(TestingGroup3[PresumedSOLExpired]andTestingGroup4
[DNAMethod/UnexpiredSOL])
208
(PresumedSOLExpired)
212
Table4.8RatesEstimatedfromContinuationRatioModelComparingDNATestingMethods
(TestingGroup4[DNAMethod])
219
Table4.9RatesEstimatedfromContinuationRatioModelComparingDNATestingMethodsby
SpermAbsent/Present(TestingGroup4,SpermPresent/Absent)
224
Table4.10CostandPersonnelEffortComparisons(TestingGroup4[DNAMethod])
227

Table5.1VictimsDecisions,byEmotionalReactionsatFirstContact
273
xviii
CHAPTER 6:
Table6.1SAKCODISEntry&CODISHitRates:ResultsfromDetroit,LosAngeles,
andNewOrleans.
303
TableB.1Validity(Trustworthiness)Assessments,UsingLincoln&Gubas(1985)Standards
TableB.2ModelingSAKSubmissionRatesOverTime,DescriptiveStatistics(N=10,817)
TableB.3BaselineModelsforSubmissionofSAKstoCrimeLab
TableB.4ModelsExaminingtheEffectsofDNAEra
TableB.5ModelsExaminingtheEffectsofDOJBacklogReductionGrantFunding
TableB.6ModelsExaminingtheEffectsofthePolicePolicyChangeRegarding
SAKSubmission
TableB.7ModelsExaminingtheEffectsofthePoliceSexCrimeUnitStaffingCuts
TableB.8ModelsExaminingtheEffectsoftheEstablishmentoftheSANEProgram
TableB.9ModelsExaminingtheEffectsofDNAErasandDOJBacklogReduction
GrantFunding
TableB.10ModelsExaminingtheEffectsofDNAErasandPolicePolicyChange
RegardingSAKSubmissions
TableB.11ModelsExaminingtheEffectsofDNAErasandEstablishmentofthe
SANEProgram
TableB.12SamplesUsedintheQuantitativeAnalysesExaminingForensicTestingOutcomes
TableB.13ContinuationRatioModelforVictimOffenderRelationshipEffect
TableB.14ContinuationRatioModelforStatuteofLimitations(SOL)Effect
406
412
414
418
419
420
421
422
423
424
425
427
441
442
xix
TableB.15LogisticRegressionPredictingCODISHitRatesAmongTestingGroup1
(StrangerRape)
(NonStrangerRape)
TableB.18ContinuationRatioModelforDNATestingMethodEffect(ModeratorOmitted)
TableB.19ContinuationRatioModelforDNATestingMethodEffect(ModeratorIncluded)
443
444
445
446
447
xx
LIST OF FIGURES
Figure1.1TheActionResearchProcess(McEwen,2003)
Figure1.2TheDetroitSAKARPCollaborativePartners
15

Figure2.1TheStepbyStepProcessofPlanningandConductingtheDetroitSAKARPCensus 49
Figure2.2SummaryoftheDetroitSAKARPCensusResults
57

Figure3.1SAKRelatedInterOrganizationalSystemicMap
Figure3.2SummaryofHistoricalContextAcrossAllDetroitOrganizations
Figure3.3PoliceProcessingofSexualAssaultCasesAssociatedwithUnsubmittedSAKs
Figure3.4PoliceAttitudesinSexualAssaultCasesAssociatedwithUnsubmittedSAKs
Figure3.5IntraOrganizationalCommunicationBetweenthePolice&CrimeLab
Figure3.6InterOrganizationalCommunicationBetweenthePolice,CrimeLab,
&ProsecutorsOffice
Figure3.7InterOrganizationalCommunicationBetweenthePolice&MedicalSystem
Figure3.8Intra&InterOrganizationalCommunicationBetweenthePolice&
VictimAdvocacy
64
99
108
110
122
126
131
134
xxi

Figure4.1TheStepbyStepProcessofCreatingtheDetroitSAKARPTestingPlan
Figure4.2AnOverviewoftheDetroitSAKActionResearchProjectTestingPlan
Figure4.3Challenges&SolutionsEncounteredImplementingtheDetroit
SAKARPTestingPlan
Figure4.4StagesofSAKDNATesting,CODISEntry,andCODISResults
Figure4.5AnOverviewoftheStructureofCODIS
Figure4.6TypesofCODISHitsandIdentificationofSerialSexualAssaultsinCODIS
Figure4.7NumberofCODISHitsintheOverallSample(N=1,595)
Figure4.8NumberofSerialSexualAssaultsintheOverallSample(N=1,595)
Figure4.9NumberofCODISHits,TestingGroup1(Stranger)(n=445)
Figure4.10NumberofCODISHits,TestingGroup2(NonStranger)(n=449)
Figure4.11NumberofCODISHits,TestingGroup3(PresumedSOLExpired)(n=351)
Figure4.12NumberofCODISHits,TestingGroup4(DNATestingMethod)(n=350)
Figure4.13NumberofSerialSexualAssaults,TestingGroup1(Stranger)(n=445)
Figure4.14NumberofSerialSexualAssaults,TestingGroup2(NonStranger)(n=449)
Figure4.15NumberofSerialSexualAssaults,TestingGroup3(PresumedSOLExpired)
(n=351)
Figure4.16NumberofSerialSexualAssaults,TestingGroup4(DNATestingMethod)
(n=350)
Figure4.17TheDetroitSAKTestingGroupsandPopulationsRepresented
Figure4.18TheEffectofVictimOffenderRelationshiponCODISEntry,CODISHits,and
SerialSexualAssaultRates(TestingGroup1[Stranger]andTestingGroup2[NonStranger])
143
149
156
168
170
172
173
175
177
177
180
180
182
183
185
186
189
196
xxii
Figure4.19UnconditionalCODISEntry,CODISHit,andSerialSexualAssaultRates,
byVictimOffenderRelationship(TestingGroup1[Stranger]andTestingGroup2
[NonStranger])
Figure4.20TheEffectofStatuteofLimitationsonCODISEntry,CODISResults,and
SerialSexualAssaultRates(TestingGroup3[PresumedSOLExpired]andTestingGroup4
[DNAMethod/SOLUnexpired])
Figure4.21UnconditionalCODISEntry,CODISHit,andSerialSexualAssaultRates,
byStatuteofLimitations(TestingGroup3[PresumedSOLExpired]andTestingGroup4
[DNAMethod/SOLUnexpired])
Figure4.22TheEffectofTestingMethodonCODISEntry,CODISHits,andSerialSexual
Assaults(TestingGroup4[DNAMethod])
Figure4.23UnconditionalDNATestingandCODISEntryRates,byDNATestingMethod
Figure4.24SimpleMainEffectsofTestingMethodonDNATestingandCODISEntryRates
bySpermAbsent/Present(TestingGroup4[DNAMethod])
Figure4.25UnconditionalSimpleMainEffectsofTestingMethodonDNATestingand
CODISEntryRatesbySpermAbsent/Present(TestingGroup4[DNAMethod])
200
207
210
218
220
221
225

Figure5.1TheStepbyStepProcessofCreatingtheDetroitVictimNotificationProtocol
Figure5.2TheDetroitSAKARPPilotVictimNotificationProtocol
Figure5.3Challenges&SolutionsEncounteredImplementingtheDetroitSAKARP
VictimNotificationProtocol
Figure5.4NumberofCasesSelectedforNotificationinwhichtheVictims
WereUltimatelyLocated
Figure5.5EffortRequiredtoLocateVictimsforNotification
Figure5.6VictimsEmotionalReactionstotheFirstContactNotifications
236
247
251
260
263
268
xxiii
Figure5.7VictimsDecisionsRegardingContinuedParticipationintheCriminal
JusticeSystem
274
CHAPTER 6:
Figure6.1LessonsLearned:ConductingaSAKCensus
Figure6.2LessonsLearned:DevelopingaSAKTestingPlan
Figure6.3LessonsLearned:ConductingVictimNotifications
Figure6.4LessonsLearned:ConductinganActionResearchProject
Figure6.5EvaluatingtheOutcomesofActionResearch:AnEvaluationInfluenceModel
Figure6.6MovingForward:ChangingSAKPolicyandPractice
285
309
320
328
334
347
FigureB.1SequentialExploratoryMixedMethodsDesign
385
xxiv
LIST OF EXHIBITS
Exhibit2.1ExcerptsFromFirstLetterfromProsecutortoPoliceChief
41
Exhibit2.2ExcerptsFromSecondLetterfromProsecutortoPolice
41
Exhibit2.3ExcerptsFromMediaReportsRegardingtheScopeoftheProblem
43
Exhibit2.4SummaryStatementfromthePoliceInternalAffairs
45
Exhibit2.5ReasonsWhySAKsWereNotSubmitted
46
112

Exhibit3.1SexualAssaultPoliceReportofSuspectedProstitutionDealGoneBad
Exhibit3.2SexualAssaultPoliceReportofanAdolescentVictimsCoveringUpBadBehavior 114
Exhibit3.3SexualAssaultPoliceReportofaKnownOffenderAssailant
117
Sexualviolenceisapervasivesocialproblem:nationalepidemiologicaldataindicatethat18%
25%ofwomenaresexuallyassaultedintheiradultlifetimes(Blacketal.,2011;Fisher,Cullen,&Turner,
2000;Kilpatricketal.,2007;Tjaden&Thoennes,2006).1Whenvictimsturntotheircommunitiesfor
assistanceaftertheassault,mostareadvisedtohaveamedicalforensicexam(MFE)(Campbell,2008;
Martin,2005).Thepurposeofthisexamistoprovidehealthcaretovictims(IAFN,2009;Ledray,
Burgess,&Giardino,2011;Lynch2006),whichincludes:caringforinjuriessustainedintheassault,
offeringemergencycontraceptiontopreventpregnancy(ifapplicable),andadministeringprophylaxis
forsexuallytransmittedinfectionsthatmighthavebeencontractedintheassault(Departmentof
Justice,2013).Inadditiontothesehealthcarecomponents,themedicalforensicexamcanincludethe
collectionofasexualassaultkit(SAK)topreservethephysicalevidencefromsurvivorsbodiestoaidin
theprosecutionofthecrime(Campbell,Patterson,&Lichty,2005;DuMont&White,2007;Fry,2007;
Ledray,1999;Martin,2005).TheprocessofcollectingaSAKistimeconsuming(usuallyfourhourslong)
andhighlyinvasiveforvictims,asitincludes:pluckingheadandpubichairs;swabbingthevagina,anus,
mouth,and/orbreaststocollectsemen,blood,orsaliva;andobtainingfingernailscrapingsintheevent
theassailantwasscratchedduringtheattack.
AfteraSAKhasbeencollectedbyahealthcareprofessional,itistakenintocustodybylaw
enforcementpersonnel.PolicearethenresponsibleforsubmittingtheSAKtoaforensiclaboratoryfor
testing,whichincludesscreeningthesamplesinthekitforwhethertheycontainbiologicalevidence
Toclarifythemeaningofkeytermsusedinthisreport,sexualviolencereferstoabroadrangeofsexuallyviolatingevents,
includingrape(anonconsensualactoforal,vaginal,and/oranalpenetrationcommittedbytheuseofforce,threatofforce,
orwhenanindividualisunabletoprovideconsent)andsexualassault(abroaderrangeofnonconsensualcontactandnon
contactsexualoffenses,uptoandincludingrape).Weusethetermsthetermsvictimandsurvivorinterchangeablytoreflect
thatsexualassaultisaviolentcrimethattakestremendousstrengthandcouragetosurvive(seeCampbell&Townsend,2011).
(i.e.,bodilyfluids),andifso,thenanalyzingthesamplesforDNA(seeButler,2005,2010,2012for
reviews).2TheresultingDNAprofilecanbeuploadedtoCODIS(CombinedDNAIndexSystem),the
nationalforensicDNAdatabase,whichconsistsofreferenceDNAprofilesfromarrestees/convicted
offendersandfromsamplesobtainedatcrimescenes(Butler,2005;Jobling&Gill,2004;Stevens,2001).
ThesampleintheSAKiscomparedtothosereferencesamplesandifthereisamatch(termedahit)
thenlawenforcementpersonnelhaveapromisinginvestigativeleadastotheidentityoftheoffender
and/ortoapatternofrepeatoffending.
Thismultistepprocessofcollectingandanalyzingsexualassaultmedicalforensicevidenceis
longandarduous,particularlysoforvictims,whooftencharacterizetheexperienceashighlytraumatic,
onethatleavesthemfeelingviolated,blamed,depressed,andreluctanttoseekfurtherhelp(Campbell,
2005,2009;Campbell&Raja,2005).Butfordecades,rapesurvivorshaveenduredtheexamand
evidencecollectionkitbecausetheyweretoldtheyhadto(Martin,2005),becausetheybelievedthere
wasnootherwaylawenforcementwouldtaketheircaseseriously(Parnis&DuMont,2006),and/or
becausetheywantedtokeepotherwomensafe(Patterson&Campbell,2010).Theyassumedthatthe
evidenceinthesexualassaultkitwouldbetestedforDNAandtheresultswouldbeutilizedbythe
criminaljusticesystemfortheinvestigationandprosecutionoftheassault(Tofte,2013).However,a
growingnumberofmediareportsandsocialsciencestudiessuggestthatformanyrapesurvivorsthisis
not,infact,whathappens.Inbothmajorurbancitiesandsmallerjurisdictions,policefrequentlydonot
submitSAKsforforensictesting,andinstead,thekitistypicallyplacedinpoliceproperty,unexamined
anduntested(Strom&Hickman,2010).
Somepolicedepartmentshavetheirowninternalforensiclabtoanalyzeevidence(i.e.,thelabisaunitwithinthepolice
department).Inotherjurisdictions,theforensiclabmayberegionalorstatewide,wherebymultiplelawenforcementagencies
submitkitstoasinglelabortoamultisitelaboratorysystem.Typically,theselabsarealsoaffiliatedwithlawenforcement
agencies(e.g.,thestatepolice).In2009,aNationalAcademyofSciencesPanel(StrengtheningForensicScienceintheUnited
States)recommendedthatforensiclaboratoriesbecomeindependentfromlawenforcement,butthesereformshavebeen
slowtoimplementandmostforensiclabsareaffiliatedwiththecriminaljusticesystem(Cowan&Koppl,2010).
Unsubmitted Sexual Assault Kits (SAKs): A Growing National Problem

In1999,multiplemediaoutletsreportedthatNewYorkCityhadover16,000SAKsstockpiled
collectedfromvictims,yetnevertestedbyacrimelab(seeTheWeinerReport,2002).Publicoutcry
grewasitbecameclearerthatseveralothermajorcitiesalsohadstaggeringnumbersofuntestedSAKs.3
HumanRightsWatch(2009)reportedthattheLosAngelesPoliceDepartment,SheriffsDepartment,and
47otherindependentpolicedepartmentsinthecountyhad12,669unprocessedSAKsinpolicestorage
facilities.InIllinois,only1,474of7,494sexualassaultkitsbookedintoevidencefrom19952009had
beenconfirmedastested(HumanRightsWatch,2010).Sincethen,largenumbersofuntestedSAKs
havebeendocumentedinmajorurbanareassuchasHouston(~4,000),SanAntonio(~11,000),
Cleveland(~4,000),aswellassmallercitiessuchasAmarillo,Texas(~1,000)andDavenport,Iowa(~600).
Emergingsocialsciencedatasuggestthesemediareportsmayindeedbecredible.InaNIJ
fundednationalsurveyof1,692lawenforcementagencies,Lovrichandcolleagues(2004)estimatedthat
therewere169,000rapecasesdatingbackto1982thatcontaineduntestedbiologicalevidence.Amore
recentNIJfundedstudybyStromandHickman(2010)surveyed2,250lawenforcementagenciesand
estimatedthat18%(27,595)ofallunsolvedrapecasessince2003containedunsubmittedforensic
evidence.Smallerscaleregionalstudiesareyieldingsimilarresults.InareviewofSAKscollectedfrom
adultvictimsinaMidwesternsexualassaultnurseexaminer(SANE)program,PattersonandCampbell
(2012)foundthat41%werenotsubmittedforanalysis,andShawandCampbell(2013)documented
that41%ofkitscollectedfromadolescentvictimswerealsonotsubmittedtothecrimelab.4
ConsistentwithNIJrecommendationsregardingappropriatenomenclature(Nelson,2010andRitter,2011),wedistinguish
betweenbackloggedSAKs(i.e.,thosethathavebeensubmittedtoacrimelaboratoryfortesting,butstillawaittesting)and
unsubmittedSAKs(i.e.,thosethathaveNOTbeensubmittedfortesting).Thefocusofthisreport/projectisunsubmittedSAKs.
4
Samplingcriteriaexcludedcasesinwhichtherewasnopolicereport;inotherwords,amongadultandadolescentvictimswho
hadSAKscollectedANDreportedtothepolice,41%ofthetimelawenforcementdidnotsubmittheirkitsfortesting.
Bothnationalandregionalstudiessuggestthatlawenforcementpersonneldonotconsistently
submitSAKsforforensictesting,whichraisesthequestion:whyaresomeSAKssubmittedfortesting,
butothersarenot?Todate,onlyahandfulofstudieshaveexaminedthisissue,andtheirfindings
suggestthatpoliceoftendoubttheevidentiaryvalueofthekit.InStromandHickmans(2010)national
surveyoflawenforcementagencies,44%oftherespondentsindicatedthattheydidnotsubmit
evidenceifasuspecthadnotbeenidentified,24%ifthesuspecthadalreadybeenadjudicated,19%if
thecasehadbeendismissed,17%iftheydidnotthinktheevidencewasuseful,15%iftestingwasnot
requestedbytheprosecutor,and12%ifthesuspecthadbeenidentifiedbutnotcharged.InPatterson
andCampbells(2012)study,policewerelesslikelytosubmitSAKsifthevictimhadbathedpostassault,
mostlikelybecauselawenforcementbelieveddoingsohadcompromisedtheevidence.
Whereaspolicemaydownplaytheevidentiaryutilityofthekit,itappearsthattheydoattuneto
perceivedvictimcredibilityandassaultseriousnesswhendecidingwhethertosubmitSAKs.Shawand
Campbell(2013)foundthatvictimsaged1315yearsweremorelikelytohavetheirSAKssubmittedthan
oldervictims,aged1617yearsold,whichisconsistentwithotherresearchindicatingthatpolicefind
childrenmorecrediblethanadolescentsandthattheyinvestmoreeffortincasesinvolving
children/youngeradolescentsthanthoseolderadolescentsandadults(Campbelletal.,2012;Cross,
Walsh,Simone,&Jones,2003).Amongadultvictims,PattersonandCampbell(2012)foundthatSAK
submissionwasmorelikelyincasesinwhichthevictimwasinjured,theperpetratorusedforce,andthe
assaultitselfinvolvedmultiplesexualpenetrations.Takentogether,theresultsofthesestudiessuggest
thatwhenvictimsfitstereotypicallyrootedbeliefsaboutwhatconstitutesgoodvictims,realvictims,
and/orrealcrimes(Caringella,2008;Lonsway&Archambault,2012;Spohn&Holleran,2001;Spohn&
Tellis,2012),theirkitsaremorelikelytobesubmittedfortesting.
IfpolicedonotsubmitSAKsbecausetheydoubttheevidentiaryvalueofthekitsand/ordoubt
thecredibilityofthevictims,thenmedicalforensicevidencehaslittlechanceofimpactingcriminal
proceedings.Indeed,Petersonandcolleagueshavefoundthattheevidenceinrapekitsisoftennot
usedtoinformarrestdecisions,andinfact,itisrarelyevenpresentedincourt(Johnson,Peterson,
Sommer,&Baskin,2012;Peterson,Johnson,etal.,2012;Peterson,Hickman,Strom,&Johnson,2013).
Similarly,DuMontandWhite(2007)reviewedstudiesfromtheUnitedStates,Canada,andseveral
Scandinaviancountriesontheimpactofmedicalforensicevidence(e.g.,injuriesdetected,documented
presenceofsperm)onsexualassaultcaseoutcomes,andfoundthatsuchevidenceisusuallynot
influentialtolegaloutcomes.5
However,somerecentstudiessuggestthattheimpactofforensicevidenceoncaseprocessing
maybebettercharacterizedasanindirecteffect,ratherthanadirectone.Forinstance,Peterson,
Hickmanetal.(2013)suggestedthatthemereexistenceofforensicevidencemayhelpbuildmomentum
foracase,particularlyifitcorroborateskeyelementsofthecrime,whichinturncanhaveapositive
effectoncaseprogression.Campbell,Bybeeetal.(2012)testedamediationalmodeloftheimpactof
sexualassaultmedicalforensicexamsonpolicereferraldecisionsandfoundthatSANEcollected
exams/kitswereassociatedwithincreasedlawenforcementeffort(e.g.,collectingothertypesof
evidence,interviewingsuspectsandwitnesses),whichinturnpredictedhigherratesofcasereferralto
prosecutors.However,ifthereislagbetweentheassaultandevidencecollection(e.g.,morethan24
hours),suchdelaysappeartohaveadirectnegativeeffectoncaseoutcomes(Campbell,Pattersonetal.,
2009),whichcouldbebecausethepassageoftimedecreasesthelikelihoodoffindingevidence
(Johnson,Peterson,etal.2012),and/orbecauseanyhesitancybyvictimstoreportthecrimemayhurt
theirperceivedcredibilitywithpoliceandprosecutors(Frohmann,1997;Kerstetter,1990).
ItisimportanttonotethattheDuMontandWhite(2007)reviewdoesnotaddresstheissueofwhetherDNAtestingandDNA
evidencespecificallyisinfluentialtocaseoutcomes.
StromandHickman(2010)notedthatwhenSAKsarenottested,justice[is]denied(p.382)
becausethereisnoopportunityforthatevidencetoinformcriminalproceedings,eithertoaidinthe
prosecutionofaperpetratorortoexoneratesomeonefalselyaccused.Similarly,HumanRightsWatch
(2009)arguedthatthefailuretotestkitssendsamessagetosurvivorsthatwhathappenedtothemis
notofsocietalconcern,andtoassailants,itsendsthemessagethattheywillnotbeheldaccountablefor
theircrimes.ThelackofconsistentSAKtestingalsomakesitdifficultforcriminaljusticesystem
personneltoidentifyserialsexualoffenders.Forstrangerperpetratedserialcrimes,testingcouldreveal
thepotentialidentityoftheoffenderandDNAmatchesacrossmultiplecrimes(termedcasetocase
associations).SAKtestingcanalsohelpidentifyserialoffendingamongnonstrangersexualassaults.
Forexample,ifassailantidentitywasknowninCaseA,policemaynottestthekit;similarly,ifidentity
wasknowninaseparatecrime,CaseB,again,policemaynottestthekit.However,ifAandBwereboth
tested,itispossibletheywouldmatch,indicatingapatternofserialsexualoffenses.Priorresearch
consistentlyfindsthatmostrapistsareserialrapists(bothstrangerandnonstrangerperpetrated)
(Abbey&McAuslan,2004;Abbey,Wegner,Pierce,&JacquesTiura,2012;Lisak&Miller,2002;
McWhorteretal.,2009;Swarthoutetal.,2011),soitisimportanttoconsiderhowmanyrepeat
offendersarenotbeingidentifiedbecauseSAKsarenotbeingtested.
Whentestinghasnotoccurredconsistentlyandajurisdictionhasaccumulatedalargenumber
ofuntestedSAKs,whatthenshouldacommunitydoabouttheproblem?Giventhelackofresearchon
thistopic,mostcitiesstrugglingwiththisissuehavehadtodeveloplocallevelsolutionswithoutthe
benefitofevidencebasedstrategiestoguidetheirefforts.Forexample,NewYorkCitydecidedtousea
forkliftapproach(Bashford,2013),wherebyall16,000SAKswereoutsourcedtoprivatelaboratory
vendorsfortesting.Onthereturnside,prosecutorshadthechallengingtaskofsiftingthrough
thousandsoflabreportstodecidewhichcasestopursueforprosecution.LosAngelesalsodecidedto
testallkits,whichagainprovidedanearlyoverwhelmingamountofdataforlawenforcementand
prosecutorsonthereturnside.ThesehighprofilecasestudiesinNewYorkCityandLosAngelesraisea
numberofimportantissuesthatmeritscientificstudy:Ifacommunityhasalargenumberofuntested
SAKsinpoliceproperty,howcanit(relatively)quicklygetanaccuratecountofthekitsinordertogauge
theextentoftheproblem?ShouldallSAKsbetestedwhataretheadvantagesanddisadvantagesof
testallvs.testsomestrategies?IfonlysomeSAKsaretobetested,isitpossibletodevelop
empiricallybasedselectioncriteriathataddresstheneedsandconcernsofdiverseconstituents(e.g.,
police,prosecutors,victimadvocates,andsurvivors)?Forthesurvivorsthemselves,whenandhow
shouldtheybenotifiedaboutwhatwasorwasnotdonewiththeirkits?
An Action Research Approach to the Problem of Unsubmitted SAKs

Inthepastfifteenyearstherehasbeenamarkedshiftinhowresearchersapproachthestudyof
complexcriminaljusticeproblems.Historically,socialscientistshavebeenononesideoftheproblem
theoutsideworkingseparatelyandindependentlyfromthepractitionersandthevictimsontheinside
(seeGaines,Worrall,&Southerland,2003;EklandOlson&Martin,1988;Hudzik&Cordner,1983for
reviews).Intheircallforanewcriminaljustice,Klofas,Hipple,andMcGarrell(2010)arguedfora
paradigmshift,oneinwhichresearchersareengagedwithlocalcommunities,workingcollaboratively
withmultidisciplinarystakeholdergroupstoidentifyevidencebasedsolutions.Oftentermedanaction
researchapproach,theoverarchinggoalistointegratescienceandsocialreformsuchthatthepursuit
ofchangeoccurssimultaneouslywithresearchandevaluationinacyclicalpatternofcriticalreflection.
Teamactivitiesarecontinuallyrevisedandimprovedasnewinformationbecomesavailable(Kemmis&
McTaggart,2005;McEwen,2003).Figure1.1(nextpage)providesaconceptualoverviewoftheaction
researchprocess.
FIGURE1.1TheActionResearchProcess(McEwen,2003)
ApplyFindings
toPractice
Identifythe
Problem
Conduct
Reconnaissance
Develop
Research
Steps
Conduct
Research
Steps
Communicate
Findings
Analyze
Research
Results
PLANNINGPHASE IMPLEMENTATIONPHASE
Klofas,Hipple,andMcGarrells(2010)reviewofsuccessfulcriminaljusticeactionresearch
projectshighlightsthreecoredefiningfeaturesofthisapproachtosocialproblemsolving.First,action
researchprojectsaregroupprojectsamultidisciplinaryworkinggroup/steeringcommitteeisformed
thatincludeslocalleaders,frontlinepractitioners,andresearchers.Theresearchersarefull,
contributingmembersoftheworkinggroup,notoutsidepassiveobservers,but(usually)notthe
directiveleaderseither;typically,apractitionerisappointedastheprojectcoordinatortoprovide
oversightandleadership.Practitionerprojectcoordinatorsareusuallybetterpositionedtogarner
supportandbuyinfromthelocalcommunityandtofacilitatelongtermsustainabilityoftheinitiative
(seeMinkler,2012andMinkler&Wallerstein,2008forreviews).Forexample,intheNIJfundedProject
SafeNeighborhoodsinitiativetoreducegunviolence,eachofthe93nationaltaskforceswas
coordinatedbytheirlocal/closestU.S.AttorneysOffice(McGarrell,2010a;McGarrelletal.,2009).
Similarly,NIJsStrategicApproachestoCommunitySafetyInitiativealsoemployedpractitionerleaders
(again,usuallypersonnelfromU.S.AttorneysOffices),andinterestingly,theyfoundthatthecitiesin
that10siteprojectthathadthebenefitofafulltimecoordinatorweremoresuccessfulinachieving
theirgoals(Roehl,Rosenbaum,etal.,2006;Rosenbaum&Roehl,2010).
Second,actionresearchprojectsincludeadistinct,datadrivenplanningphase.Theultimate
goalistocreateanempiricallyinformedintervention,andtothatend,theresearcherscollectdata
aboutexistingconditionstotakestockandgaugethenatureandextentoftheproblembefore
launchingintoanewinitiative.Forexample,intheoneofthefirstNIJfundedactionresearchprojects,
theBostonGunProjectsOperationCeasefire,thestudyteamgatheredextensivearchivalrecordsto
identifythesourcesofgunsandguntrafficking,tocreatemapsofgunandknifeslayingsinBoston,and
todevelopsocialnetworkmodelsthatillustratedgangactivity(Kennedy,2012;Kennedy,Braga,Piehl,&
Waring,2001).Thesedatawerethensharedwiththeworkinggroupsothatpractitionerscouldreflect
onthisinformationanduseittoinformnextsteps.Alltoooftenpractitionershavetomakedecisions
basedonimpressionsandanecdotes;theactionresearchparadigmprovidesempiricaldatathatreflects
theirlocalconditionssothattheyhaveanaccuratepicturetoguideinterventiondevelopment.
Third,thiscommitmenttodatainformeddecisionmakingcontinuesastheinterventionis
developedandevaluated.Researchersareactiveparticipantsincreatingtheprogram,bringingtheir
knowledgeaboutempiricalbestpracticestothetable.Astheprogramisimplemented,theresearchers
(typically)collectbothprocessandoutcomedataregardingitseffectiveness,sharinginterimfindings
withthemultidisciplinaryworkinggroupsothatmidcourseadjustmentscanbemade(ifnecessary).In
traditionalresearchprojects,resultsaresharedafterthecompletionoftheproject,andalthoughan
afterthefactanalysisofwhatworkedandwhatdidnotworkiscertainlyhelpful,itisoftenmore
usefultoknowwhatsnotworkingwhenthereisstillopportunitytocreatechange.Forinstance,in
ProjectSafeNeighborhoods,thelocalresearchers/evaluatorsprovidedcontinuousfeedbacktotheir
taskforcesforongoingrevisionsandimprovementoftheirgunviolencereductionprograms
(McGarrell,2010b).
10
Severalprojectsthathavefollowedthesecoreprinciplesofgroupcollaboration,datadriven
planning,andongoingevaluationhaveachievedsuccessfuloutcomes.TheBostonGunProjectsawa
63%reductioninyouthhomicidespermonthanda25%decreaseingunassaultspermonth(Kennedyet
al.,2001).IntheStrategicApproachestoCommunitySafetyInitiative,fivecitiesreporteddramatic
decreasesintheirtargetcrimes(e.g.,homicide,youthviolence,firearmsviolence)(Rosenbaum&Roehl,
2010).ThejurisdictionsparticipatinginProjectSafeNeighborhoodsexperiencedan8%declineinviolent
crime(relativetocomparablecitiesnotpartoftheinitiative)andfederalprosecutionsincreased60%
(McGarrell,2010a).Thesehighprofileprojectshighlightthepromiseoftheactionresearchmodel,
thoughitisimportanttonotethatsuchprojectsareoftencharacterizedashighstress,highstakes
endeavors(seeKennedy,2012),andthatresearcherpractitionerpartnershipsprovidenoguarantee
thataninitiativewillbesuccessfulinmeetingitsaims(seeMinkler,2012andMinkler&Wallerstein,
2008).Nevertheless,inalloftheseprojects,complex,seeminglyintractablecrimeproblemswere
systematicallywhittleddownandfundamentallychangedinthecontextofsustainedmultidisciplinary
collaborationsbetweenresearchersandpractitioners.
Giventhesesuccesses,anactionresearchapproachmaybeapromisingstrategyforaddressing
thegrowingnationalproblemofunsubmittedSAKs.Thoughlawenforcementpersonneltypicallyhave
theresponsibilityofsubmittingaSAKforforensictesting,theyworkwithinamultidisciplinarynetwork
ofsexualassaultresponders,includingprosecutors,forensicscientists,victimadvocates,and
medical/nursingproviders.Eachofthosedisciplineshasuniquerolesandresponsibilitiesregarding
postassaultservicesforrapevictims.Bringingthesedifferentprofessionstogethertoexaminecritically
thepurposeandutilityofSAKtestingmaybeafruitfulstrategyforunderstandingwhysomanykitsare
notbeingsubmittedfortestingandhowthisproblemcanberemedied.Tothatend,NIJreleaseda
solicitationinOctober,2010tofundthreetofivejurisdictionswithsubstantialnumbersofuntestedkits
11
toconductamultiyearactionresearchproject(StrategicApproachestoSexualAssaultKitEvidence:An
ActionResearchProject).Thesolicitationhighlightedtwoprimarygoals:
Thepurposeofthisstudyistwofold.First,NIJisinterestedinlearningabouttheunderlying
factorsthatcontributetothisunsubmittedSAKevidence.Second,basedonwhatislearned,
promisingstrategieswillbedevelopedandimplementedtoreduceandeliminatetheuntested
kitsinthatjurisdiction.Throughmonitoringandassessment,allstrategieswillbeevaluatedfor
theireffectivenessandsustainability.(p.4.)
Thissolicitationstipulatedthattheformalapplicantmustbeastateorlocalunitofgovernment(e.g.,
policeagency,crimelab,prosecutorsoffice)andthateachsiteshoulddemonstrateateamapproach
thatincludesaneffectivepartnershipbetweenthepractitioneragenciesandaresearchorganizationor
partner(pp.56).Consistentwithcoreprinciplesoftheactionresearchparadigm(Klofasetal.,2010;
McEwen,2003),thegrantwouldprovidefundingforasixmonth,datadrivenplanningphase,andthen
siteswouldreapplyforsecondphaseofworktoimplementtheirplans.
One Citys Shocking Discovery: The Problem of Unsubmitted SAKs in Detroit

WhenNIJreleasedtheSAKActionResearchProjectSolicitationinOctober,2010,thecityof
Detroitwasstilltryingtomakesenseofwhatmanytermedashockingdiscovery.OnAugust17,2009,
representativesfromlocalpolice,statepolice,andtheprosecutorsofficetouredaremotepolice
propertystoragefacilitytodiscusswhattodoaboutthevolumeofevidenceinpolicecustodyandhow
itshouldbebestmanaged.TheissueofforensicevidencehadbeenacriticalconcerninDetroitever
sincethepolicedepartmentcrimelabhadbeenclosedonSeptember25,2008duetoahigherrorrate
inballisticstestingandbroadersystemicconcernsregardingitsprocessingofcrimesceneevidence
(Baker,2009).Duringthetour,anassistantprosecutornoticeddozensofstorageboxesandaskedwhat
12
theycontained:rapekits,approximately10,000rapekits.Theassistantprosecutorpressedfordetails
Howmanykits?Havetheybeentested?Policepersonneldidnothaveimmediateanswers.
Afterthetour,theassistantprosecutorinformedtheElectedProsecutor,whothenmade
repeatedeffortstogettheanswerstothosefundamentalquestions.Phonecallsandmeetingsbetween
theChiefofPoliceandtheProsecutor,andthenmoremeetings,loopinginmorelocalstakeholders,did
notproducetheanswers.FormalmemosfromtheProsecutortotheChiefrequestingtheimmediate
productionofalistofthekitsinquestion(andtheirtestingstatus)andrequestinganindependentaudit
ofthecontentsofthepropertystorageroomwentunanswered.TheProsecutorcontinuedtopressthis
issue,soastategovernmentviolenceagainstwomenagencyorganizedamultidisciplinaryteam,
includingprosecutors,lawenforcement,medicalprofessionals,andcommunityadvocatesnoneof
whomwereemployedbythelocalpoliceorprosecutorsofficetobeginanindependent,indepth
reviewoftheproblem.
WithfederalfundingfromtheOfficeofViolenceAgainstWomen(OVW),The400Projectbegan
inApril,2010:400SAKswererandomlyselectedfrompoliceproperty,allwereoutsourcedfortesting,
andthenprojectstaffbeganthelaboriousprocessoftrackingdowntherecordsassociatedwitheach
casetogetthestorybehindeachkit.BylateFallof2010,preliminaryfindingssuggestedthatindeed,
mostSAKsinpolicepropertyprobablyhadnotbeentested.Basedontheirsampleof400,theyhad
ascertainedthatsomekitshadbeentestedandthenrefiledinproperty,butthevastmajoritydidnot
appeartohavebeentested.SomeSAKswereassociatedwithcasesthathadalreadybeenadjudicated
(someofwhichhadbeenadjudicatedwithouttestingthekit);someSAKswereassociatedwithcases
thatwerebeyondthestatuteoflimitations;andastillundeterminednumberofSAKswereassociated
withcasesthatcouldstillbeprosecuted.ThoughThe400Projectfinalreportwas(atthattime)many
monthsaway,itwasalreadyclearthatDetroithadalong,longroadahead.
13
WhentheNIJSolicitationforactionresearchprojectsonunsubmittedSAKswasreleased
whichwas,byhappenstance,aboutthesametimeDetroitpractitionersweretryingtoabsorbtheearly
findingsfromThe400Projectstakeholdersreadilyagreedthiswasanexcellentopportunityforthe
communityandtheprosecutorsofficewouldbethelocalunitofgovernmenttoapply.TheDirectorof
thestategovernmentagencyleadingThe400ProjectconnectedtheElectedProsecutortoasexual
assaultresearcheratMichiganStateUniversity,whohadconductedseveralNIJfundedstudiesonthe
criminaljusticeresponsetosexualassault.Asmallworkinggroupofresearchers,representativesfrom
theprosecutorsoffice,andthestategovernmentviolenceagainstwomenagencycametogetherto
preparethegrantapplication.OnApril8,2011theDepartmentofJustice,OfficeofJusticePrograms
issuedapressreleaseannouncingthatWayneCounty,MichiganandtheCityofHouston,Texas,had
beenawardedgrantsunderthesolicitation.6
An Overview of the Detroit Sexual Assault Kit (SAK) Action Research Project (ARP)
Project Goals
TheDetroitSAKARPbeganApril1,2011andprojectactivitiesendedSeptember30,2013(2.5
years/30months).ConsistentwithMcEwens(2003)actionresearchapproach,thefirstsixmonthswas
adatadrivenplanningphase;theremaining24monthsfocusedoncompletingreconnaissancetasks
initiatedduringtheplanningphaseandthenimplementingtheresponseplan.Specifically,theDetroit
SAKARPhadfourprimarygoals:
1) ToobtainanaccuratecountofthenumberofSAKsinpolicepropertythoughacompletecensus
ofeverySAKinpoliceproperty(uptoNovember1,2009);
SeesectionASpecialNoteAboutIdentity,Confidentiality,andPrivacy(laterinthischapter)formoredetailsregardinghow
thisreportwillhandleidentifying/potentiallyidentifyinginformationassociatedwiththisprojectanditsparticipants.
14
2) ToidentifytheunderlyingfactorsthatcontributedtowhyDetroithadsomanyunsubmitted
SAKsinpoliceproperty;
3) TodevelopanempiricallybasedplanfortestingSAKsandtoevaluatetheefficacyofthatplan;
4) Tocreateavictimnotificationprotocol(i.e.,whenandhowvictimswouldbenotifiedregarding
thestatus/findingsoftheirkits)andtoevaluatetheefficacyofthatprotocol.
The Collaborative Team

Figure1.2(nextpage)depictstheeightdisciplinesrepresentedwithintheDetroitSAKARP
(spanningnineorganizations).Representingtheprosecutorialperspective,assistantprosecuting
attorneysfromthelocalprosecutorsoffice(oneofwhomwastheprojectcoordinatorforthemajority
oftheproject)andonememberofastatelevelprosecutingattorneysassociationwereteammembers.
Localpolicewererepresented,includingfrontlinedetectivesfromthesexcrimesunit,theunits
supervisor,andhighercommandstaff(uptothelevelofDeputyChief).Thelocalpolicedepartmenthad
aninternalvictimadvocacyprogram(oftentermedasystembasedadvocacyprogram),staffedby
MSWlevelsocialworkers.Thesesystemsbasedadvocateswerenotinitiallyinvolvedintheproject,but
oncepolicecommandstaffbroughtthisoversighttotheteamsattention,theywereincludedaswell.
Duringtheplanningphase,itbecameabundantlyclearthattheteamneededtheassistanceof
informationtechnologyspecialists;thelocalpolicedepartmenthadaCityofDetroitITstaffmemberon
calltothem,andsothatindividualwasaskedtojointhecollaborative.Fortheforensicsciences
perspective,membersoftheforensicsdivisionofthestatepolicewereteammembers;asnoted
previously,thelocalpolicedepartmentscrimelabclosedin2008andafterthat,forensicevidencefor
Detroitwashandledbythestateforensicsciencelabs.Itwasnotclearatthebeginningoftheproject
15
FIGURE1.2TheDetroitSAKARPCollaborativePartners
StateLevel
ProsecutorsAssoc.
Prosecutors
LawEnforcement
ForensicScientists
StateGovernment
SystemBased
VAWAgency
CommunityBased
VictimAdvocates
Victim Advocates
NationalVAW
Foundation
InformationTechnology
Medical/Nursing
Specialist
Researchers
NOTE:TeammembersfromLawEnforcement,SystemBasedAdvocacy,andInformationTechnology
Specialistswerefromthesameorganization(localpolice,tealoutline).TwoCommunityBasedAdvocacy
organizationsparticipatedinthisproject:onewasacombineddomesticviolence/sexualassaultprogram,the
otherwasacombinedSexualAssaultNurseExaminer(SANE)/sexualassaultprogram(purpleoutline).
thatthelocalpolicedepartmentstillhadaforensicsciencecoordinator,andoncethatwasknown,that
individualwasalsoaskedtoparticipateinthecollaborative.Detroithastwocommunitybased
advocacyprogramsthatprovidesexualassaultservicesandbothofwhichwerepartnersintheproject:
onewasacombineddomesticviolence/sexualassaultsocialserviceagency;theotherwasacombined
SexualAssaultNurseExaminer(SANE)program/sexualassaultadvocacy(membersoftheSANEwing
ofthatorganizationrepresentedthemedical/nursingperspectiveinthecollaborative).Thestate
governmentviolenceagainstwomenagencythatoversawThe400Projectwasalsoinvolved,aswasa
nationalnonprofitviolenceagainstwomenfoundationwithalongstandinginterestintheissueof
unsubmittedsexualassaultkits.Theresearchteamconsistedofoneprincipalinvestigatorandoneco
investigator(bothPh.D.level),threeM.A.levelresearchassociates,andthreePh.D.levelstatisticians.
16
Eachofthenineorganizationsinthecollaborativetypicallyhad1to4representativesontheteam;thus,
atanyonepointintime,thecollaborativehadbetween9to36members(mostmeetingsseebelow
wereattendedby9to18members).
The Collaborative Partnership Process

TheDetroitcollaborativemetonsiteattheprosecutorsofficetwiceamonth(fortwohours
eachmeeting)for30months;additionalphone/inpersonmeetingswereheldasneededtoattendto
urgentissues.Inaddition,theteamhadanextendedtwodayretreattodevelopthevictimnotification
protocol,andaseparateVictimNotificationReviewTeam(NRT)wasthenformed(seeChapter5:
Developing&EvaluatingaVictimNotificationProtocol).Thebimonthlyteammeetingswerefacilitated
bytheprojectcoordinator(arepresentativeoftheprosecutorsoffice);formalmeetingminuteswere
taken,whichwerereviewed/approvedateachsubsequentmeeting.Thoughthefocusofmeetings
variedthroughouttheproject,ingeneral,eachdisciplinedidareportoutregardingitsactivitiessince
thelastmeetingandthentherewasusuallytimeallocatedtodiscusscurrentproblemsandchallenges.
Thedecisionmakingprocesswasparticipatoryinthesensethatalldisciplineshadinput(orhadthe
opportunitytoprovideinput)intotheissuesathand,butfinalauthorityrestedwiththeseniormost
officialwithineachparticipatingorganization(e.g.,ElectedProsecutor,ChiefofPolice).
Research/Evaluation Component
Theactionresearchparadigmstipulatesamarkedlydifferentroleforresearchersthanwhatis
typicalintraditionalsocialscienceresearchprojects.Asnotedpreviously,theresearcherswerefull
membersoftheDetroitcollaborativeteam,taskedwithcollectingformativedatatoguidetheplanning
process,providingsubstantiveinputregardingthedevelopmentoftheresponseplan,andevaluatingits
efficacy.Thoughtherearelongstandingdebatesintheevaluationliteratureregardingwhetheritis
17
appropriateforresearcherstobeinvolvedwiththedevelopmentandevaluationofaninitiative(see
Alkin,2004;Cousins&Chouinard,2012;Patton,2011;Scriven,1997forreviews),theactionresearch
paradigmsquarelylocatestheresearchersroleonthesidethatitisnotonlyacceptabletohavedual
tasks,butnecessarytohelpguideempiricallybaseddecisionmaking(seeKlofas,Hipple,&McGarrell,
2010forareview).Forthreeofthefourmainprojectgoals,suchablendedrolewasrelatively
straightforwardtoenvisionandenact:theresearcherswouldbeinvolvedintheplanningandexecuting
datacollectionfromtheSAKcensusandanalyzingthedata;wouldhelpcreateaSAKtestingplanand
evaluateitsefficacy;andwouldparticipateinthedevelopmentofthevictimnotificationprotocolsand
evaluatetheirimpact.
However,theNIJSAKactionresearchprojectsolicitationalsostated,NIJisinterestedin
learningabouttheunderlyingfactorsthatcontributetothisunsubmittedSAKevidence(p.4).
UnderstandinghowandwhyDetroithassomanyunsubmittedSAKsrequiresanindepthlookwithin
eachorganizationatitsleadership,staffing,resources,anddecisionmaking,aswellasanexamination
oftherelationshipsamongorganizationsovertime.Essentially,thisisastudyofhowandwhykey
individualsandorganizationsdidnotdowhatwasexpectedofthembyvictims,byotherorganizations
inthecommunity,bysocietyatlarge.Insuchsituationsmeaning,thosethatareripeforcontroversy
anddiscordthescholarlyliteratureislessdividedonthenatureoftheresearchersrole,as
collaborativeprocessescansometimesunderminethequalityandcredibilityofthework(seeScriven,
1997).AsChelimsky(1997)noted,[theevaluatorsjob]istomakeobjectiveinformationavailable,
especiallyinahostilepoliticalclimate(p.57).Echoingtheimportanceofindependenceandobjectivity,
Stake(1997)emphasizedhowrigorousmethodologyisallthemorecriticalinpoliticallycharged
environmentsbecausesciencehastraditionalmechanismsforvalidation,forexposingideologyand
purgingmisrepresentation(p.474).Notwithstandingtheacademicdebateastowhetherscienceisin
factneutralandvaluefree(seeEagly&Riger,2014andHesseBiber,2007forreviews),independent
18
datacollection,analysis,andverificationcanhelpstakeholdersstepbackandlookatchallenging
informationinanewandoftenlessemotionallychargedway,whichcanfacilitateproblemsolvingand
systemicchange(Patton,2011).Therefore,forthisparticularprojectgoal(identifyingunderlying
factors),theresearchersworkedindependentlyfromthelargergroup,but,consistentwiththeaction
researchmodel,interimfindingsweresharedwiththecollaborativepartnershipforfeedbackinhope
thatgroupdiscussionwouldsuggestnewavenuesforfurtherstudy,andmoreimportantly,bolster
effortstochangetheseunderlyingfactorstopreventthereoccurrenceofthisproblem.
Thefourgoalsofthisprojectspandiversesubstantivetopics(e.g.,SAKtesting,victim
notification)andmethodologicalaims(e.g.,basicresearchandevaluation),whichnecessitatedtheuse
ofmultiplemethodsofdatacollection.Belowisabriefsummaryofeachofthefourdatacollection
methodsusedinthisprojectethnographicobservations,interviews,archivalrecords,andfocus
groupshighlightingwhyeachtechniquewasselectedandhowitwasimplementedinthisaction
researchproject.Table1.1(followingpages)summarizeshowthesefourmethodswereusedtoassess
eachofthefourmainprojectgoals;Table1.2(followingpages)summarizesthequantityofdata
collected(organizedbymethod).AppendixB:ProjectMethodologyprovidescompletedetailsregarding
theprojectssampling,measures,datacollectionprocedures,analytictechniques,and
verification/authenticationprocesses.
EthnographicObservations.Ethnographicmethodsweretheprimarydatacollectiontechnique
inthisactionresearchproject.Realtimeobservationsofeventsastheyarehappeningoffer
unparalleledcapacityforcapturingtimesensitiveevents(Atkinsonetal.,2001;Fetterman,2010;
Wolcott,2005).Moreover,ethnographyisparticularlyusefulinsituationswherethe
researcher/evaluatorwillbedeeplyinvolvedinthecontext,andwhenaholistic,longtermapproachis
necessarytounderstandthephenomenonofinterest(Langhout,2003).Ethnographiesarealsowell
suitedforcapturingprocessesovertime,especiallywhensignificantevents,transitions,andconflictsare
19
expected(Smith,2005).Interestingly,ethnographicmethodshavenotbeenamainstayincriminal
justiceactionresearchprojects(seeKennedyetal.,2001;McGarrelletal.,2009;Roehletal.,2006);
however,formalizingobservationsofgroupprocessintotheethnographicfieldnotesallowsfor
documentingkeyquestions,dilemmas,decisions,andlessonslearned.Giventhatthereisno
roadmapforcommunitiesstrugglingwithlargenumbersofuntestedSAKs,ethnographicmethodsare
wellsuitedforcapturingeachstepalongthewayanditsassociatedchallengesandsolutions.
Inthisproject,theresearchteammemberswereparticipantobserversinallSAKcollaborative
meetings:oneresearchteammemberwasdesignatedobserveronlyandthatindividualtranscribed
thediscussionsatthemeetingsastheywasoccurring,andone(ormore)researchersengagedinthe
discussionsandalsotooknotes.AllmembersoftheSAKcollaborativewerebriefedindividuallyandasa
groupregardingIRBproceduresforethnographicobservationssothattheyunderstoodthattheir
remarkswouldbewrittendown.InaccordwiththemethodsoutlinedbyEmerson,Fretz,andShaw
(1995),fieldnoteswerewrittenwithin72hoursofanobservation;thesenotesincludeddirectquotes,
timelinesandsequencesofevents,keydecisionsmadebythegroup,andrichdescriptionsofthe
interactions;aseparatetranscriptofeachmeetingwasalsopreserved.
IndividualInterviews.Interviewmethodsarecommonlyusedinactionresearchprojectsto
obtainindividualteammembersthoughts,experiences,andopinions(Greenwood&Levin,2006;
Stringer,2013).Interviewsarealsoanintegralpartofethnographicprojectstocreateprivatespacesfor
individualstodiscusskeyissueswiththeresearchers(Adler&Adler,2002;Heyl,2001;Wolcott,2005).
Typically,suchinterviewsarequalitativeinnature:openendedquestionswithinasemistructured
interviewguidethatchangesinresponsetotheinformationprovidedbytheparticipants(Patton,2002;
Rubin&Rubin,2011).Longitudinalinterviewingisvaluablewhenthereisreasontoexpectthat
individualsbeliefsandperceptionswillchangeovertimeasaresultofsustainedinteractionswithother
stakeholdergroupsintheirenvironment(Britten,2007;Saldana,2003;Seal,Eldrige,&Kacanek,2007).
20
Inthisproject,individualinterviews(bothcrosssectionalandlongitudinal)wereusefulforallofthese
reasons,plustheyhelpedfocusthecollaborativeteammeetingsondevelopingresponsestrategies,
ratherthantakingtheentiregroupstimeforgatheringresearchrelatedinformation.
Inthisproject,theresearchteamconductedbothformalandinformalindividualinterviews.
Formalinterviewsfollowedstandardinterviewingprocedures:separatewrittenrequestsweremade
tomultipleindividualswithineachparticipatingorganization(lawenforcement,prosecution,forensic
sciences,medical/nursing,systemsadvocacy,andcommunitybasedadvocacy),askingthemto
participateinaconfidentialoneononeinterviewthatwouldbeaudiorecordedandtranscribed.
Formalinterviewswerealsoconductedwithnationalstakeholdersfromcriminaljustice/forensicscience
andviolenceagainstwomenorganizationstogainabroader(lessDetroitcentric)perspectiveonSAK
testing,sexualassaultinvestigations,andvictimnotification(seeTables1.1and1.2).
Inethnographicresearch,itisalsotypicalthatresearchershavecontactwithindividualsoutside
offormalmeetings/settings,andtheseinteractionsareoftenopportunitiestoaskquestionsabout
currentevents.Theseinformalinterviewsprovideanotherwayofdocumentingeventsthroughouta
project.UnderIRBconsentforethnographicobservation,theresearchershadongoinginformal
interviewswithrepresentativesfromeachparticipatingDetroitareaorganization(seeTables1.1and
1.2).Theseconversationswereoftenamixoffactualprojectupdateswithdisclosuresthatwereprivate
(e.g.,ventingfrustrations,personalreflections);therefore,consistentwithstandardpracticein
ethnographicresearch(seeFetterman,2010;Wolcott,2005),onlyfactualinformationwasrecorded,
unlesstheresearcherspecificallyaskedpermissiontomakenoteoftheothercontent.
ArchivalRecords.TheunsubmittedSAKsinDetroitpolicepropertydatedbacktothe1980s,so
studyingaproblemthreedecadesinthemakingposesuniquemethodologicalchallenges.Forexample,
manyofthekeypersonnelineachorganizationwhocouldspeaktopoliciesandprocedureshavelong
sinceretiredormovedontootherpositions.Currentpersonnelmayormaynotbeawareofkey
21
historicalshiftswithintheirorganizationsovertime.Insuchsituations,researchersoftenturnto
archivalrecordstopiecetogethereventsofyearsago,thoughitisnotuncommonthatsuchdocuments
areincompleteanddisorganized(Corti,2007;Hill,1993;Singleton&Straits,2010).Recordsmaybe
publicallyavailablethroughFreedomofInformationAct(FOIA)requests,orinternal/privateand
thereforeaccessibleonlyatthediscretionofkeystakeholders.Thus,archivalresearchrequires
extensiverelationshipbuildingsothatorganizationswillinvesttheefforttofindkeydocumentsand
disclosethem,whichmaybedifficulttonegotiateifthereisariskthattheinformationcouldbe
incriminatingorotherwisedisparaging.Inotherwords,thechallengeinthisprojectwasnotjusttrying
todocumentthirtyyearsofhistory,buttodocumentthirtyyearsinwhichsome(andperhapsmany)of
thoseyearsmightnotbeviewedfavorably,byboththoseinsidetheDetroitcommunityaswellas
outsiderslookingin.
Theactionresearchparadigmwashelpfulinaddressingtheseissues.Themultidisciplinaryteam
approachprovidedastructureforstakeholderstoairdifferencesinaconstructivesettingandtoset
expectationsforfulldisclosure,transparency,andchange.Thisapproachalsoallowedpractitionersto
workcloselywiththeresearchersandtodiscusshowtherequestsforarchivalrecordsfitintothebig
pictureoftheproject.Assuch,theresearchteamsrequestsforpublicrecordsandinternalrecords
regardingleadership,staffing,resources,anddecisionmakingregardingSAKtestingfortheyears1989
2009werefulfilled.Similarly,requestsweregrantedtoreviewpolicereportsassociatedwithSAKsthat
werenotsubmittedfortesting,whichenabledtheresearcherstoexplorewhethertherecommon
featuresaboutthevictims,assaults,and/orinvestigationalpracticesinthesecases.However,itis
importanttonotethatDetroitareaorganizationshavenotbeenabletodevelopandmaintain
comprehensive,wellorganizedinformationsystems,sotherearenumerousgapsinthearchivalrecord
(e.g.,staffingrecordsovertimearespotty,policereportscouldnotbefoundforallSAKs).Furthermore,
inallformsofarchivalresearch,itisdifficultifnotimpossibletogaugethecompletenessofthedata
22
becausetheremayberecordsresearchersarenotevenawareofandhencecannotrequest.Therefore,
inthisproject,archivalrecordswereusedasasupplementaldatasourcetotriangulatefindings
generatedthroughothermethods(e.g.,interviewdata,ethnographicobservations)(seeAppendixB:
ProjectMethodologyforfulldetailsregardingdatatriangulation).
FocusGroups.Focusgroupscanbeausefulsupplementtoindividualinterviewsinthatthey
provideopportunitiesforparticipantstoshareperspectivesandreacttoideaspresentedbyothers,
whichisoftenhelpfulforgeneratingnewideas,clarifyingissues,andrevealingdifferencesofopinion
(Krueger&Casey,2008;Morgan&Krueger,1997;Stewart,Shamdasani,&Rook,2006).Ingroupbased
ethnographicresearch,tosomeextenteverymeetingisafocusgroupofsorts,butthereisadifference
betweenobservingandrecordingwhatagroupisnaturallydoingandaskingthegrouptoengageina
specificdiscussionexplicitlyforresearchpurposes.Thus,atruefocusgroupfollowsstandard
proceduresofthemethodology,including,butnotlimitedto,aformalizedprotocolforthediscussion,
scriptedquestions,establishedrulesforthediscussion,aformalmoderator,andfacilitatedguidance
throughout(seeKrueger&Casey,2008).Inthisproject,focusgroupswereusedsparingly(threetotal),
andalloccurredinthefinalmonthsoftheactionresearchprojectasatechniqueforgatheringlessons
learnedabouteachmaincomponent(census,testing,victimnotification,andoverallprojectissues).
23
TABLE1.1OverviewoftheDetroitSAKARPDataCollectionMethods
DATACOLLECTIONMETHODS
Goal1:
ArchivalRecords
FocusGroups
DocumenttheDiscoveryoftheUnsubmittedSAKsandEventsThatOccurredPriortotheBeginningofCensus
ConductaCensusof
Observegroupdiscussions
SAKsinPoliceProperty aboutthediscoveryofthe
unsubmittedSAKsandthe
interorganizational
communicationsthereafter
InterviewDetroitstakeholders
aboutthediscoveryofthekitsand
theeventsthereafter
Reviewinterorganizationalmemos
re:thediscoveryofthekits&intra
organizationalinvestigations/audits
Reviewmediareportsaboutthe
discoveryofthekits
(NA)
DocumenttheProcessoftheSAKCensus
Observetheprocessof
conductingthecensusfor
keyquestions,issues,and
decisions
aboutthesuccesses,challenges,and
lessonslearnedfromconductinga
census
Reviewrecordsinthepolice
propertydatabase
forensiclaboratorytesting
spreadsheet
Discusssuccesses,
challenges,and
lessonslearned
re:conductinga
census
Goal2:
DocumentHistoricalContextinDetroitSexualAssaultOrganizations
Identifythe
UnderlyingFactors
Re:WhyDetroitHas
UnsubmittedSAKs
aboutpolicies,practices,and
resourcesavailableineach
organizationovertime
InterviewDetroitstakeholdersfrom
eachdisciplinere:dailyoperations/
servicesprovided;staffinglevels;
andreportingstructure,trainingand
supervision
Interviewpublicofficialsinfour
comparablecitiesre:services&
staffinglevels
Reviewrecordsre:leadership,
staffing,&resourcesinDetroit
organizations(19892009)
staffing,&resourcesfromsexual
assaultorganizationsinfour
comparablecities(Philadelphia,
Dallas,Baltimore,&NewOrleans)
(NA)
24
TABLE1.1(continued)
Goal2(continued)
ArchivalRecords
FocusGroups
ExamineFrontLineServices&DecisionMakinginSexualAssaultCases
aboutstandardoperating
procedures&decisionmaking
insexualassaultcases
eachdisciplinere:decisionmaking
processesinsexualassaultcases
Goal3:
DocumentProcessofDevelopingTestingPlan
DevelopSAKTesting
PlanandEvaluate
Efficacy
Observetheprocessof
developingaSAKtestingplan
re:keyquestions,issues,and
decisions
Reviewcriminalsexualassault
policereports(19892009)re:
investigationalpracticesand
decisionmaking
(NA)
InterviewDetroitstakeholdersand
(NA)
stakeholdersinnationalorganizations
concernedwithcriminaljustice,
forensicsciences,andviolenceagainst
womenregardingthepurpose&utility
ofSAKtesting
Discusssuccesses,
challenges,and
lessonslearned
re:developinga
testingplan
(NA)
(NA)
EvaluateTestingPlan
Observegroupdiscussionsre:
testingresultsandthe
implicationsofthefindings
Reviewpolicefilesassociated
withSAKstestedinthisproject
forvictim,assailant,andcase
characteristics
Documentnumberandtypeof
CODIShitsassociatedwithSAKs
testedinthisproject
25
TABLE1.1(continued)
Goal4:
DocumentProcessofDevelopingVictimNotificationProtocols
DevelopVictim
NotificationProtocols
andEvaluateEfficacy
Observetheprocessof
developingvictimnotification
protocolsre:keyquestions,
issues,anddecisions
InterviewDetroit
stakeholdersand
stakeholdersinnational
organizationsconcernedwith
criminaljustice,forensic
sciences,andviolenceagainst
womenregardinghowand
whenvictimsshouldbe
notifiedabouttestingresults
ArchivalRecords
(NA)
FocusGroups
Discusssuccesses,
challenges,andlessons
learnedre:victim
notificationandthe
utilityofaformalized
VictimNotification
ReviewTeam
EvaluateVictimNotificationProtocols
Documenttheprocessesand
decisionsoftheVictim
NotificationReviewTeam
(NA)
Trackinvestigatorseffortstolocate (NA)
victimsfornotificationand
perceptionsregardingthe
notifications
Documentcommunitybased
advocates(deidentified)
notifications
26
TABLE1.2OverviewoftheQuantityofDataCollectedintheDetroitSAKARP
METHODS
DATAQUANTITY
Ethnographic
Observations
N=81observations,~186hoursofobservation
n=53collaborativeteammeetings,~106hoursofobservation
n=18impromptumeetings,~32hoursofobservation
n=6shadowingobservationsofstakeholdersconductingtheirjobs,~18hoursofobservation
n=1planningretreat,~12hoursofobservation
n=3VictimNotificationReviewteammeetings,~18hoursofobservation
Individual
Interviews
N=42formalinterviewswithDetroitstakeholders(16onetime/crosssectionalinterviews;
26longitudinalinterviews10peopleinterviewedtwotimes,2peopleinterviewthreetimes)
N=187informalinterviewswithDetroitstakeholders(30people,numberofinterviewsvaried)
N=5interviewswithnationalcriminaljustice/forensicsciencestakeholders
N=5interviewswithnationalviolenceagainstwomenorganizationstakeholders
N=35interviewswithpublicofficialsincomparablescities
ArchivalRecords
N=2databasesreviewedre:thenumberofunsubmittedSAKs:policepropertydatabase
(~11,000entries)andpoliceforensicsciencestestingspreadsheet(~2,500entries)
N=5intraandinterorganizationalrecords(andN=6mediareports)re:discoveryofthe
unsubmittedSAKsinAugust,2009
N=93(publicallyavailableandinternal)fromDetroitorganizationsre:leadership,staffing,
resources,servicesprovided,andpolicies&proceduresovertime
N=33records(publicallyavailable)fromorganizationsincomparablecitiesre:leadership,
staffing,andresources
N=1,268policereportsreviewedre:investigationalpracticesanddecisionmakinginsexual
assaultcasesandcodedforvictim,assailant,&casecharacteristics
N=1,595SAKDNAtestingresultsreports
N=31investigatorrecordsandN=18communitybasedadvocaterecordsre:victim
notificationsconductedinthisproject
FocusGroups
N=3focusgroupsre:successes,challenges,andlessonslearned
27
A Special Note About Identity, Confidentiality, and Privacy

Inresearchonsensitivetopics,extracareiswarrantedregardingboththeconductofthe
research/evaluationitselfaswellasthereportingofitsfindings(forreviews,see:Deckeretal.,2011;
DicksonSwift,James,&Liamputtong,2008;Liamputtong,2007;Lee,2000;Miller,Forte,Wilson,&
Greene,2006;Renzetti&Lee,1992).Thedatacollectionandanalysisproceduresusedinthisproject
conformedtotheAmericanPsychologicalAssociationsguidelinesfortheethicaltreatmentofhuman
subjectsinresearch,andwerereviewedandapprovedbytheMichiganStateUniversityInstitutional
ReviewBoard(IRB)(theinstitutionoftheleadresearcher/evaluator).Todefinetheparameters
regardingidentity,confidentiality,andprivacywhenreportingtheresultsofthisproject,multipleexpert
sourceswereconsulted:MichiganStateUniversitysIRBandOfficeoftheGeneralCounsel;the
NationalInstituteofJustice;theethicalstandardsoftheAmericanPsychologicalAssociationandthe
AmericanEvaluationAssociation;nationallyrecognizedevaluationethicistsandlegalethicists;the
academicliteraturesonethicsinactionresearch,ethnography,andresearchwithvulnerable
populations;andtheindividualsandorganizationswhoparticipatedinthisactionresearchproject.
Belowissummaryofhowthisreportwillattendtothesethreeethicalelements.
Identity.Insocialscienceresearch,identityoridentificationreferstowhetherthenames
and/orotheridentifiableinformationaboutthespecificsite,city,setting(s),organization(s),and/or
individualswhoparticipateinaprojectwillbeprotectedorreleased(BelmontReport,1979;Bernard,
2011;Sieber,2004;Singleton&Straits,2010).Historically,manycriminaljusticeactionresearch
projectshaverevealedthenamesofparticipatingcities,organizations,andspecificindividualswhowere
involvedintheinitiative(e.g.,Kennedyetal.,2001;McGarrelletal.,2009;Roehletal.,2006).However,
inethnographicresearch,thereisawellestablishednormthatidentitiesmustnotberevealed,orthat
onlymacrolevelinformation(e.g.,thenameofthecityorgeographicregioninwhichtheworkwas
conducted)canbeshared(Fetterman,2010;Murphy&Dingwall,2001;Wolcott,2005).
28
Inthisproject,theidentityofthecityinwhichthisworkwasconductedwasreleasedbythe
DepartmentofJustice,OfficeofJusticePrograms,viaapressreleasenamingthetwositesthatwere
awardedgrantsundertheSexualAssaultKitActionResearchProjectSolicitation.Assuch,thisreport
hasthusfarandwillcontinuetomakereferencetothecityofDetroitasthecommunityinwhichthis
workisconducted.Theidentitiesofthespecificorganizationsandindividualswhowereinvolvedinthis
actionresearchprojectwillnotberevealed(regardlessofwhethertheyarepublic/electedfigures),per
therequirements,recommendations,and/orpreferencesoftheindividuals,groups,anddoctrines
consultedregardingthesematters(seelistabove).ThespecificorganizationsinDetroitthatserve
victimsofsexualassaultare,ofcourse,publicrecord,butakeyprincipleinethnographicresearchisthat
theparticularcity/community/organizationanditsspecifichistoryisoftennotthepointoftheresearch;
thepointistheresultingfindingsregardinghuman,social,and/ororganizationalbehaviorandtheir
generalizabilitytoothercontexts(Fetterman,2010;Johnson,1990;Wolcott,2005).Becausethe
problemofuntestedSAKsisagrowingnationalproblem,theissueisnotsomuchwhathappenedinX
PoliceDepartment,YProsecutorsOffice,ZForensicLaboratory(andsoon),butwhatcanlearnedabout
howtobringmultidisciplinaryorganizationstogethertoresolvelargequantitiesofpreviously
unsubmittedSAKsandtopreventthereoccurrenceoftheproblem.
Confidentiality.Insocialscienceresearch,confidentialityreferstoprotectingtheidentityof
theparticipant/datasource,whichtypicallyinvolvesremovingnamesandotheridentifyinginformation
fromthedataandfromanydistributionofthedataand/orfindings(oftentermeddeidentifyingthe
data)(BelmontReport,1979;Bernard,2011;Sieber,2004;Singleton&Straits,2010).Inactionresearch
projects,researchersneedtoconsidercarefullywhatcouldbeidentifyinginformationbecausethose
outsideaprojectorcommunitycouldreadadeidentifiedinterviewexcerptandhavenoideawho
providedtheinformation,butthoseinsidecouldlookatthesamedeidentifiedexcerptand
29
recognizeitssource(bythecontent,wordchoice,speechmannerisms,etc.).Therefore,protecting
confidentialitycanbecomequitecomplexincloseknitgroupprojects.
Toaddressthesechallenges,confidentialityguidelineswerecreatedforeachofthefourmain
typesofdatacollectedinthisproject.Theethnographicobservationsyieldedrichdescriptivedata
regardingtheissues,dilemmas,anddecisionsmadethroughouttheprojectallofwhichareknownto
theteammembersbecausetheywerepartofthatprocess.Assuch,thiscomponentoftheprojectis
necessarilyknowntotheinsiders;therefore,thekeyconfidentialitytaskissummarizingthefindingsin
waysthatdonotrevealconfidentialinformationtooutsiders.Consequently,thepresentationof
thesefindingsemphasizestheendresultsofthegroupprocess(i.e.,whatdidthegroupultimately
decideateachjunctureandwhy),notingkeydifferencesofopinionatanorganizationallevelofanalysis
(e.g.,representativesfromthelocalpolicedepartmentexpressedconcernsabout...ratherthanthe
localpolicedepartmentDeputyChiefexpressedconcernsabout...).7
Theethnographicobservationsalsoyieldeddirectionquotationsfromteammembersasdid
theindividualinterviews(formalandinformal)andthefocusgroups.Statementsmadeinteam
meetingsand/orthefocusgroupswereknowntoothercollaborativepartners,butcommentsmadein
individualinterviewswerenot.Thoughitcouldbepossibletodistinguishquotestakenfromgroup
settingsvs.individualsettingsinthepresentationofthefindings,itseemedmoresensibletotreatall
directquotedataasinformationthatneededtobeconfidentialtobothinsidersandoutsiders.To
protecttheconfidentialityofthesedata,mostofthequotesselectedforinclusioninthisreport
reflectedthesentimentsexpressedbymanyindividuals(i.e.,thequotecouldhavecomefromoneof
Therewereinstancesinwhichdocumentingtheworkofthegroupnecessitatedhighlightingspecificdecisionsmadeby
specificindividuals,asthosedecisionswerekeyturningpointsfortheproject(seeChapter2:HowManyUnsubmittedSAKsin
Detroit,Figure2.1TheStepbyStepProcessofPlanningandConductingtheDetroitSAKCensusasanexample).Theseevents
werealreadywellknowntotheinsidersandrevealingthetitle/roleofthoseindividualsdoesindeedrevealkeyinformation
tooutsidersaswell.However,thecollaborativepartnersconcurredwiththeresearchersthatthislevelofdetailwas
appropriateintheseinstancesinordertoreflectaccuratelywhathappenedinthisproject.
30
many,asmultiplepeoplestatedthatsameidea).Tooutsidersitisoftenhelpfultohavesomedegree
ofattributionforaquote(forcontext),andbyandlarge,itwaspossibletoincludetextsuchas,a
memberofthelocalpolicedepartmentnotedthat...oraforensicscientistsstatedthat...without
compromisingconfidentiality.Quotesattributedasamemberofthecollaborativenotedthat...
reflectsentimentsexpressedbymembersfromalldisciplines/organizations,sothatspecifying
organizationalaffiliationwasnotnecessaryasthisparticularideawassharedthroughouttheentire
collaborative.However,thereareoftenuniquequotesideasnotexpressedbymanyindividualsthat
meritinclusioninaresearchreportpreciselybecausetheyexpressadivergentpointofview.Inthese
instances,attributionmustbenonspecific(e.g.,anindividualsaid...)inordertoprotect
confidentiality.Allquotespresentedinthisreportwerereviewedcarefullytoassesspotential
identifabilitytothoseinsidetheproject,andasnecessary,materialwaslightlyeditedtoremove
distinctivespeechmannerisms,turnsofphrase,etc.8
Protectingtheconfidentialityofarchivalrecordsistypicallystraightforward.Usually,
researchersareworkingwithpublicallyavailablerecords,andthenanyidentifyinginformationwithin
thosedocumentsmustberedacted(e.g.,victimsnameandidentifyingcharacteristicsmustberedacted
fromcopiesofthepolicereportandfromanyexcerptreleasedofthatreport).Forthisactionresearch
project,excerptsfrompublicallyavailablerecordsfollowedtheseconventions.However,protecting
confidentialityismorecomplexwithinternalorganizationalrecords.Revealingexcerptsfromsuch
recordsmustfollowsimilarredactingrules,butanyexcerptrevealsthattheresearcherhastherecordin
thefirstplaceandforinsidersthatmaybeadefactoidentificationofthepersonwhoprovidedthe
documents(whichmaynotbesomethingthatindividualwishestobeknown).Inthisreport,excerpts
Withinthequalitativeresearchparadigm,therearedifferencesofopinionastowhetherquotedmaterialshouldeverbe
edited(seeSandelowski,1994).Giventhatthisisnotaphenomenologicalorhermeneuticalanalysis(seeAppendixB:Project
Methodology),whichrequireverbatimtext,confidentialityconcernswereprivilegedandsomequoteswerelightlyedited.
31
frominternaldocuments(thosewouldnotbetypicallyreleasedthroughFOIA)werereviewedbythe
individual(s)whoprovidedthemandtheirpermissionforinclusioninthisreportwasobtained.
Privacy.Insocialscienceresearch,privacyreferstopersonshavingcontrolovertheextent,
timing,andcircumstancesofsharinginformationaboutoneselfwithothers(BelmontReport,1979;
Bernard,2011;Sieber,2004;Singleton&Straits,2010).TheBelmontReports(1979)principleofrespect
forpersonsstipulatesthatresearchparticipantshavearighttoprivacy,whichiscodifiedinfederallaw
(HHSandFDARegulations(45CFR46.111(a)(7)and21CFR56.111(a)(7)).Privateinformationmustbe
keptconfidential,soresearchersmustdeidentifydataandtakemeasurestoensurethatthe
presentationofthefindingsdoesnotincludeidentifiableorpotentiallyidentifiableinformation(see
sectionabove).However,insomecircumstancesitisworthaskingwhethercertaindatashouldbe
revealedatallevenindeidentifiedformbecausetheinformationisprivate.Inethnographic
researchthisiskeyconcernbecauseoverthecourseofmanymonths,oryears,ofcontactwiththe
participants,itisquitelikelythattheresearcherswillsee/hearaboutexperiencesthataredeeply
personaltotheindividualsand/orthecommunitymoregenerally(Fetterman,2010;Murphy&Dingwall,
2001;Wolcott,2005).Assuch,inethnographicresearchthereisalong,valuedtraditionofrespecting
participantsprivacy;asWolcott(2005)noted,Nofieldworkereverhaslicensetotellall(p.141).Butif
someinformationwillbewithheld,thentherewillbequestionsbythoseinsideandoutsidethe
projectaboutwhatisbeingwithheldandwhy.Wolcott(2005)arguedthatethnographersmustgive
theirreadersasenseofwhatisexcludedbecausesuchtransparencybolstersthecredibilityofthework
anditstrustworthinessamonginsidersandoutsiders.
Whendecidingwhatshouldandshouldnotberevealed,Wolcott(2005)notedthat
fieldworkersshouldalwayshaveinmindtheboundariesoftheirinquiries...staywithinthelimitsof
theresearchfocus(pp.143143).Eventsthatoccuroverthecourseoftheprojectthataredirectly
relevanttotheresearchaimsmustbereported,andthosethatinvolvehighlysensitiveinformation
32
shouldbereportedwithcarefulattentiontoconfidentiality.However,materialthatisnotdirectly
relevantshouldnotbeincluded,thoughWolcott(2005)maintainedthatfortransparency,
ethnographersshouldmakementionofkeyexclusions.Inthisactionresearchproject,themost
commonofftopicissuewasthetremendousstressandstrainofthisworkontheteammembers
themselves.Thisprojectwasnotintendedtobeastudyofvicarioustraumatherewasnoindicationin
anyoftheresearchmaterialsorconsentformsthatindicatedtotheparticipantsthatthiswouldbea
subjectofinquiry.However,participantsshared,bothprivatelywiththeresearchers,and,toalesser
extent,publicallywitheachotheratteammeetings,thatsortingthroughalltheseuntestedSAKstooka
terribleemotional,physical,andspiritualtollonthem.Anextendedanalysisofthisissueisnotwithin
thescopeofthisproject,anddoingsowouldbeaviolationoftheparticipantsprivacy.However,
becausethiswasasalientissueforteammembers,therearelessonslearnedregardingthe
importanceofsupportingstaffmembersinthisverydifficultwork,recommendationsthattheteam
reviewedandendorsedforinclusioninthisreport(seeChapter2:HowManyUnsubmittedSAKsin
Detroit,Figure2.1TheStepbyStepProcessofPlanningandConductingtheDetroitSAKCensusasan
example).However,therewillnotbeanysustaineddiscussionorexamplesofthistopicinthisreport.
Therecouldalsobematerialthatisclearlywithinthescopeoftheprojectthatperhapsshould
notbereleasedoutofrespectforindividualand/orcommunityprivacy.Wolcott(2005)arguedthat
ethnographershavearesponsibilitytoupholdthetrustthecommunityhasplacedinthem:Ultimately,
however,questionsofdiscretionareuptoeachindividualfieldworker.Onlythatindividualisawareof
alltheprofessionalandpersonaldimensionstobefactoredin(Wolcott,2005,pp.230).Again,the
extentpossible,ethnographersshouldlettheiraudienceknowingeneraltermswhatontopic
matterstheyhavedecidedtowithhold.Inthisreport,thedepictionsofgroupprocessnotethatthere
wereargumentsandtensionsamongstakeholdersandthecontentofkeydebatesthatinfluenced
laterdecisionsaredescribedbutspecificcomments,particularlythoseclearlymadeintheheatofthe
33
moment,neednotbeincluded.Itissufficienttosaythattensionsranhighsometimesandpaying
attentiontogroupfacilitationisimportant(seeChapter5:Developing&EvaluatingaVictimNotification
Protocol,Figure5.1TheStepbyStepProcessofCreatingtheDetroitVictimNotificationProtocolasan
example).Thisreportincludesinformationthatisnotflattering,isoftendamning,andisnearlyalways
heartbreaking.Noteverydetailcouldbesharedresearchisalwaysaselectiveprocessbutthe
detailsthatarenotshareddonotchangethesubstanceofthefindingspresented.
About This Report

ThisreportsummarizesthefindingsoftheDetroitSexualAssaultKit(SAK)ActionResearch
Project(ARP),9whichisasocialscienceresearchstudyontheproblemofunsubmittedSAKs.Todate,
muchofwhathasbeenwrittenonthistopicstemsfrominvestigativereportingprojects,suchasthose
conductedbyHumanRightsWatch(e.g.,HumanRightsWatch,2009,2010),whicharemarkedly
differentinmethods,tone,andstyle.Forexample,theHumanRightsWatchprojectshavehadan
explicitaimofdemandingpublicaccountabilityandadvocatingforpolicyreform,whichisconsistent
withtheirorganizationsmissionandpurpose.Bycontrast,socialscienceresearchstudiesincluding
thoserootedinanactionresearchparadigmhavedifferentaims,namelyunderstandinghowandwhy
problemsoccur,documentingtheprocessoftryingtochangethem,andevaluatingtheeffectivenessof
thosestrategies.Assuch,thisreportisnecessarilydifferentfromthoseotherworks,andhopefullyadds
newinformationandnewperspectivesforthepublicdiscourseontheproblemofunsubmittedSAKs.
Thisreportwaswrittenbytheresearch/evaluationteam,basedondatacollectedoverthirty
monthsandanalyzedinarigorousprocessofcrosschecking,triangulation,andauthentication(see
AppendixB:ProjectMethodologyfordetails).Thefindingsforeachmajorgoalweresharedwiththe
ThroughoutthisreportwewillrefertothisprojectastheDetroitSAKARP,ormoresimply,thecollaborative,the
collaborativeteam,collaborativepartnership,ortheteaminreferencetothemultidisciplinary,multiorganizationalgroup
thatworkedtogetherfor30monthstodevelopandimplementanactionplanfortheuntestedSAKsinDetroit.
34
collaborativepartnershipthroughouttheproject(informalpowerpointpresentations,insmallwithin
organizationdiscussiongroups,andwithindividualteammembers)andadraftofthisreportwas
providedtoallparticipatingorganizationsforcommentpriortoitssubmissiontotheNationalInstitute
ofJustice(andsubsequentreleasethroughtheNationalCriminalJusticeReferenceService[NCJRS]).
Theresearch/evaluationteamreviewedthestakeholdersfeedbackonthedraftreport,recheckedeach
issueinquestionagainstthedata,solicitednewdocumentation/datatoresolvediscrepancies,andif
warranted,revisedthereportaccordingly;dissentingfeedbackthatcouldnotberesolvedthroughthis
processisnotedthroughoutthereport.Itisimportanttoemphasizethatalthoughallmembersofthe
collaborativehadanopportunitytoreviewandcommentonthefindings,thisreportdoesnotreflect
theofficialpositionsofanyparticipatingorganization.
Theremainingchaptersinthisreportareorganizedbyeachofthefourmajorgoals,followedby
adiscussionofthefindingsandadetailedmethodologicalappendix:10
Chapter2:HowManyUnsubmittedSAKsinDetroitdescribeshowthecollaborativeconducted
acensusofallrapekitsinpoliceproperty(GOAL1).
Chapter3:WhySoManyUnsubmittedSAKsinDetroitexaminestheunderlyingfactorsthat
contributedtowhyDetroithadsomanyunsubmittedSAKsinpoliceproperty(GOAL2).
Chapter4:Developing&EvaluatingaSAKTestingPlandescribeshowthecollaborative
developedandevaluatedanempiricallybasedplanfortestingSAKs(GOAL3).
10
Giventhenumberoftopics(andoveralllength)ofthisreport,wehaveusedcolorcodingthroughouttohelpguidethereader
througheachmaincomponentoftheproject.DarkredwillbeusedtodenoteGOAL1(Census);darkgrayforGOAL2
(UnderlyingFactors);darkgreenforGOAL3(Testing);purpleforGOAL4(VictimNotification).Otherchaptersinthisreport
(Introduction,Discussion,Appendices)willbeinblue.
35
Chapter5:Developing&EvaluatingaVictimNotificationProtocolexplainshowthe
collaborativedevelopedandevaluatedavictimcentered,traumainformedvictimnotification
protocol(GOAL4).
Chapter6:SummaryofFindings,Implications,andCommunityChangessummarizesthe
projectfindingsandimplicationsforpolicyandpractice.
AppendixB:ProjectMethodologyprovidestechnicalmethodologicalandanalyticdetailsfor
eachcomponentoftheproject.
36

OneofthefirsttaskstobetackledwhenacommunityhaslargenumbersofunsubmittedSAKsis
determiningthescopeofproblem.AsnotedinChapter1:Introduction,thefocusofthisreport/project
iskitsthatwereneversubmittedbylawenforcementtoaforensiclaboratoryfortesting;11therefore,at
issueishowmanySAKsareinpolicepropertyfacilities.Inthisera,many(butcertainlynotall)law
enforcementagencieshavecomputerizedrecordsregardingwhathasbeenenteredintoproperty,soit
isquitelikelythatajurisdictionwillbeabletoproducealistofallSAKsinpolicestoragebutthislist
mayormaynotanswerthequestionofhowmanyunsubmittedSAKsareinthatjurisdiction.
Why?SomeSAKsinpolicepropertymayhavealreadybeentested,andwerethenrefiled;
othersmayhavebeensubmittedfortesting,andwerereturnedtopropertyiftestingwasstillpending
(e.g.,becausereferencesampleswereneededtoruleoutaconsensualpartner);andstillothersmay
haveneverbeensubmittedfortesting.Therefore,oneissuetobeparsedoutisthetestingstatusofthe
SAKsinpoliceproperty.InTable2.1(nextpage),therowsdepictthreepossiblescenariosfortesting
status:neversubmittedfortesting;submittedfortesting,buttestingwaspartial/incomplete;12
submittedfortestingandtestingiscomplete.
AnotherissuetoattendtoistheadjudicationstatusofthecasesassociatedwiththeSAKsin
policeproperty.InTable2.1,thecolumnsdepictthreepossibleoutcomes:thecasehasnotbeen
adjudicatedandthecrimeisstillwithinthestatuteoflimitations(SOL);thecasehasnotbeen
adjudicatedandislikelySOLexpired;thecasehasbeenadjudicated.Forexample,someSAKsinpolice
11
AsopposedtobacklogSAKs,whichweresubmittedtoaforensiclaboratorybuthavenotyetbeentested(seeNelson,
2010;Ritter,2011).
12
Examplesofpartial/incompletetestinginclude:testingthatisincomplete/pendingwhileawaitingreferencesamples;
testingthatwasconductedinthepreDNAeraandthereforeisincompletevisvisDNAtesting.
37
propertyhavebeentestedandtheircaseswereadjudicated,butthekitsareinpropertybecausethey
arebeingretainedintheeventofanappealorarerequiredtoberetainedpursuanttostatelaw.13
SomeSAKshaveneverbeentested,butthecaseshavebeenadjudicatedandthekitsareinproperty
(mostlikely)becausetheagencyhaspoliciesnottodestroyevidence.OtherSAKshaveneverbeen
testedandmaystillbeeligibleforprosecution.Inotherwords,theSAKsinpolicepropertyarelikelya
jumbleofdifferentsituationsandscenarios,andsortingtheseoutiscriticalfordeterminingthescopeof
theproblemandidentifyingthefocusofaresponseproject.
TABLE2.1SAKsinPoliceProperty:UnderstandingTestingStatusandAdjudicationStatus
ADJUDICATIONSTATUS
NONADJUDICATED
StillWithinSOL
TESTING
STATUS
ADJUDICATED
SOLExpired
Neversubmittedfortesting
KEYFOCUS
KEYFOCUS*
Submittedfortesting,
buttestingwaspartial
POSSIBLESECONDARYFOCUS
Defensecounselmayseek
reviewoftheseSAKs/cases
Defensecounselmayseek
reviewoftheseSAKs/cases
(e.g.,DNAtestingnotperformed,
testingpausedwhilewaitingfor
referencesample)
Submittedfortesting,
testingcomplete
SAKswererefiledinpoliceproperty
*TheDetroitSAKARPincludedunsubmittedpresumedSOLexpiredSAKsinitsprimaryfocus,thoughotherjurisdictionsmay
choosenottoprioritizethesecasesinthesamemanner.
13
Forexample,inMichigan,MCL770.16(12)requiresthattheinvestigatinglawenforcementagencypreserveanybiological
materialidentifiedduringtheinvestigationofacrimeorcrimesforwhichanypersonmayfileapetitionforDNAtestingunder
thisstatute.
38
Testingstatusandadjudicationstatusmayormaynotbeinformationfieldsinapoliceproperty
computersystem.Ifthatinformationistrackedwithinthepropertysystem,thenassessingtheexact
statusofeachkitisasimpledatabasequery;ifnot,thenitisnecessarytolinkpolicepropertyrecordsto
forensictestingandcourtrecords.Again,dependingonthequalityoftheinformationsystemsina
particularjurisdiction,thiscouldbefairlystraightforwardcomputertask,oritcouldbecomeanarduous
manualsearchforinformation.
DetroitisoneofmanyresourcestrappedcommunitiesintheUnitedStatesthathasnotbeen
abletoinvestinstateoftheartcriminaljusticeinformationsystems.Thepropertycomputersystem
generatedalistoftheSAKsinpolicepropertybutkeystakeholdersquestionedthecompletenessand
accuracyofthatinformation(seeTheDiscoveryoftheKitsandInitiatingaCensusbelow).Linkingthe
propertyrecordstoforensictestingrecordsseemedimpossiblebecausethepolicecrimelabdidnot
haveacentralizedevidencesubmissionandtestingdatabase,anditwasseveralmonthsintotheproject
beforethecollaborativelearnedthatacrimelabstaffmemberhadinfactbeentrackingSAKtestingin
astandaloneExcelspreadsheet.Linkingpropertyrecordstoadjudicationrecordswasnot
straightforwardbecausethecomputerizedcourtrecordsdidnotinterfacewiththepoliceproperty
database.Furthermore,adjudicationstatuscouldnotalwaysbedeterminedfromthecomputerized
records,andsoitwasnecessarytoperformmanualsearchesofhandwrittenpolicelogbooks.Inshort,
countingthenumberSAKsinpolicepropertyanddiscerningtheirtestingstatuswasacomplicated,
painstaking,15weektask,consumingapproximately2,365hoursofstafftime.
ThepurposeofthischapteristodescribehowtheDetroitSAKARPcollaborativedetermined
howmanyunsubmittedSAKswereinpolicepropertyandtopresenttheresultsofthatmonthslong
processoflocating,sorting,andconnectingrecords.First,bywayofbackground,theeventssurrounding
thediscoveryofthekitswillbedescribedbecausewhathappenedinthattourofpolicepropertyand
themonthsthereafterhadadirecteffectonkeydecisionsregardingthetaskofcountingtheSAKsin
39
policeproperty.ThiscontextisessentialforunderstandingwhytheDetroitcollaborativeconductedthe
censusinthewaythatitdid.Second,theprocessofconductingtheSAKcensuswillbedescribed,
highlightingthekeyquestions,issues,anddecisionsmadebythecollaborative.Thissectionmaybe
particularlyhelpfultootherjurisdictionsthatneedtoplanandexecuteacensus.Finally,theresultsof
thecensuswillbepresented,highlightinghowmanySAKswerefoundinDetroit.
The Discovery of the Kits and Initiating a Census

HowdoesacommunityknowwhetherithaslargenumbersofunsubmittedSAKs?Giventhe
increasingnationalattentiontothisproblem,somejurisdictionsarenowproactivelycheckingthe
contentsofpolicepropertyfacilities,butformanyofthefirsthighprofilecities,thediscoverywas
accidentalandunexpected.Forinstance,inNewYorkCity,areorganizationofpolicepropertybrought
theproblemtolight:rapekitsthathadbeenpreviouslydispersedthroughoutstoragehadbeenlocated
andgroupedtogether,whichrevealedthattheyhadaverylargestockpile(Bashford,2013;Tofte,2013).
TheNewYorkCitypoliceinformedtheirprosecutors,forensicscientists,andMayorsOfficeaboutthe
discovery,andthentheyworkedtogethertodevelopanactionplanfortestingallkits.
DetroitsdiscoverywassimilarinsomewaystoNewYorks:in2002,policeinitiatedamassive
reorganizationofpropertyevidence,whichincludedpullingrapekitsthathadbeeninstoredinbins
alongsideotherevidence(e.g.,ballisticsevidence,crimesceneevidence),puttingthemtogetherin
bankerstyleboxes,andmovingthemtoanoffsitestoragefacility.However,unlikeNewYorkCitys
discovery,thepolicedidnotreachouttocityofficials,theprosecutorsoffice,orforensicscientists
aboutthematterbecausesomelawenforcementpersonnelinDetroitdidnotperceivethistobea
problemandothersthoughttheproblemdidnotmeritalarmandimmediatereaction(seebelowfor
40
furtherdiscussion).14Theboxescontinuedtoaccumulateinpropertyforapproximatelysevenmore
years,untilAugust17,2009whenrepresentativesfromlocalpolice,statepolice,andtheprosecutors
officetouredaremotepolicestoragefacilitytodiscusswhattodoaboutthestaggeringvolumeof
evidenceinpolicecustodyandhowitshouldbebestmanaged.Duringthetour,anassistantprosecutor
noticedthoseboxes:
Werewalkingthrough,IseetheselikesteelshelvingunitswithboxesandIsay,Whatare
those?Theysaid,Thosearerapekits.Isaid,Rapekits!Whatarealltheserapekitsdoing
here?Iestimated...10,000ormore...And[Iasked],Aretheytestedoruntested?Andthe
officers[said],Idontknow.(emphasesinoriginal).
Afterthetour,theassistantprosecutorinformedtheElectedProsecutor,whomaderepeated
effortstoascertainthestatusofthesekits.Exhibit2.1(nextpage)andExhibit2.2(nextpage)are
excerptsfromletterssentbytheProsecutortothethenChiefofPolice.Thefirstletterrequeststhatan
independentbodyconductafullauditofthecontentsofthepolicestoragefacility;thesecondletter
requestsanitemizedlistoftheSAKsinquestionandtheirtestingstatus.Itdoesnotappearthatthe
Chiefprovidedawrittenreplytoeitherletter(i.e.,theresearchteamaskedthepoliceforsuchrecords
andnonewereprovided,andtheprosecutorsofficehasnorecordofawrittenreplytoeitherletter).
14
Thereareotherkeyhistorical/contextualdifferencesbetweenNewYorksandDetroitsSAKdiscoveriesthatmayalso
explain,inpart,whypolicechosetodisclose(ornotdisclose)theexistenceofthousandsofuntestedSAKs.InNewYorkCity,
thediscoveryoccurredinthelate1990s,whenthecityscoffershadsufficientfundstotestallthekitsandthethenMayorwas
promotingtoughoncrimeinitiatives(Tofte,2013).Bycontrast,Detroitwasstrugglingfinanciallyandin2008,thepolice
departmentcrimelabcameunderscrutinyfortheaccuracyofitswork.
41
EXHIBIT2.1ExcerptsFromFirstLetterfromProsecutortoPoliceChief
...
...
...
EXHIBIT2.2ExcerptsFromSecondLetterfromProsecutortoPolice
42
Formonths,andindeedyears,aftertheAugust,2009discoveryofthekits,thepolice
departmentwasnotforthcomingwithdetailedinformationaboutthesekits.Itdoesnotappearthatthe
departmentreleasedalistofkitsuntiltheOVWfundedThe400ProjectbeganinApril,2010(andthelist
wasnotgiventotheDetroitSAKARPProjectCoordinatoruntilfourmonthsafterthestartofthis
project).Inthemonthssincethediscoveryofthekits,therewerenumerousmeetingsandmedia
reportsinwhichpolicedepartmentrepresentativesquestionedwhatpreciselanguageshouldbeusedto
describethestatusofthekits.Forinstance,policeofficialsobjectedtoworddiscoveryasitcouldimply
thatthekitshadbeenlostandwerethenfoundorthattheyhadbeendeliberatelyhidden.Police
executivesemphasizedthatthekitswereneverlost,wereneverhidden,andhadalwaysbeenproperly
accountedfor.ThesedebatesaboutsemanticswerefrustratingtomanyotherDetroitareaandstate
levelstakeholders,asamemberofthecollaborativenoted:15
Doesitmatterwhatwecallit?Alltheirarguingdidntsitwellwithme,andIknowitdidntsit
wellwithothers.Imean,lookatitthousandsandthousandsofkitsarejustsittingthere,
ignoredforyears,andsoyourresponseistoarguewhatexactwordweregoingtouseto
describethisincredibletravestyofjustice?Ok,fine,ifwewontusetheworddiscovery
anymore,will[thepolice]admittheresaproblemhere?
Policeofficialsalsosteadfastlydeniedthattherewere11,000untestedSAKsinpoliceproperty,andthe
numbersdebateplayedoutinthepress,asseeninExhibit2.3(nextpage).
15
AsexplainedinChapter1:Introduction,quotesattributedtoamemberofthecollaborativereflectsentimentswidely
sharedthroughoutthecollaborative(i.e.,acrossdisciplinesandorganizations).
43
EXHIBIT2.3ExcerptsFromMediaReportsRegardingtheScopeoftheProblem
CBS News.com, March 4th, 2010

In Detroit, the Wayne County Prosecutor Kym Worthy recently said she is worried about 10,000
untested rape kits in the Detroit evidence storage facility. John Roach, spokesman for the
Detroit Police Department says there are really about 7,000 kits in storage with an estimated
5,800 kits that are untested. He attributes this to cases where there is a known alleged
assailant, no charges being pressed, a guilty plea or insufficient evidence to proceed.
Again,otherstakeholderswerefrustratedbythepolicedepartmentsfocusonthenumbers,particularly
becausedocumentationthatcouldhavesubstantiatedtheirclaimswasnotforthcoming;asamember
ofthecollaborativenoted:
Ifitsnot11,000thenprovidesomeproofabouthowmanyitisthen...ontheonehand,I
understandwhytheyrearguingaboutthenumbersbecauseifitsnot11,000thenweshouldnt
besayingits11,000.Buttheyvegottocomeupwithsomedocumentationtoshowthat,and
theyhavent.
Forotherstakeholders,thenumbersdebatewasfrustratingbecauseitdetractedfromwhattheyfelt
wastherealissue:thousandsofunsubmittedSAKssatinpolicepropertyandthepolicewerenot
acknowledgingthatthiswasaseriousproblem.Asamemberofthecollaborativenoted:
Itslikeanumbersgametothem(thepolice),arguingwhetherthismanythousandorthat
manythousand.Really,soifits7,500insteadof10,000,itsokay?Likehaving7,500untested
kitsisacceptable?Idontthinktheyseethisasaproblem.
Inthemonthsafterthediscoveryofthekits,andextendingtothebeginningoftheDetroitSAK
ARP(twoyearslater),policeofficialsdidnotmakeanypublicstatementsthatconveyedalarmorserious
concernregardingthekitsinproperty.Bycontrast,asseeninExhibit2.3(above)policeofficials
44
maintainedthattherewerejustifiablereasonsforwhySAKshadnotbeentested,buttheydidnot
elaborateatthetimeastowhatthosereasonsmightbe.
Aninternalreportthatsurfacedinlate2012providedsomeinsightastohowpoliceofficials
viewedthesituationandwhypolicerepresentativesmighthaverespondedastheydidafterthe
discoveryofthekits.Afterthepropertytourin2009,thepolicedepartmentlaunchedanInternalAffairs
investigationofthekits,culminatinginafivepagereport.16AsshowninExhibit2.4(nextpage),the
InternalAffairsinvestigationinvolvedselecting36SAKsfrompoliceproperty(onebatchof10SAKsand
asecondbatchof26SAKs)andthenpullingthecorrespondingpolicereportstoseehowtherespective
officersincharge(OIC)characterizedeachcaseandwhethertheydocumentedareasonastowhythe
kitwasnotsubmitted.17TheInternalAffairsreportconcludedthattherewerejustifiablereasonswhyall
36SAKshadnotbeensubmittedfortesting.
TheInternalAffairsreportlistscasebycasewhyeachSAKwasnotsubmitted(seeExhibit
2.5,followingpage).Inthirtyfiveofthe36cases,thereportmaintainsthatpolicefollowedproper
procedures.18In71%ofthecases,thereasonlistedfornotsubmittingthekitwaseitherastatement
aboutthevictimsbehaviororanoveralljudgmentofthevictimscredibility.Victimbehaviorsthatwere
commonlycitedincludedrefusingtocooperatewithprosecutionornottakingstepsthatwouldbe
16
ThatreportwasnotsharedwiththeprosecutorsofficeoranyotherDetroitareaorstatelevelorganizationswhowere
vestedinthisissue,norwasitdisclosedtotheDetroitSAKARP;thereportonlybecameknowntothecollaborativeinlate2012,
whenitwasreleasedtothemediainresponsetoaFOIArequestfromanationalnewsorganizationdoingafeaturestoryonthe
DetroitSAKs.
17
TheInternalAffairsreportstatesthatthe36SAKswererandomlyselectedfrompoliceproperty,butitisstatistically
improbablethatselectionwasrandombecause33ofthe36SAKswerecollectedbetweentheyears20052008(and22were
from2008alone).The400ProjectdocumentedthatSAKsdatedbacktothe1980s,whichwasconfirmedinthecomplete
censusconductedinthisproject(seeTheResultsoftheCensuslaterinthischapter).
18
Onecasefrom2007wasdenotedDefendanttoLocate,meaningthatthecasewasstalledpendingidentificationofthe
defendant;in2007,thedepartmentpolicywastosubmitallToLocatecases/SAKsfortesting,soeventhoughthesummary
statementintheInternalAffairsreportstipulatesthattherewerejustifiablereasonsinallcases,this2007casewasnot
consistentwithdepartmentpolicy.Theextenttowhichtheother35caseswerehandledappropriatelyissubjectto
interpretation(seeaboveandChapter3:WhySoManyUnsubmittedSAKsinDetroit).
45
expectedforvictimsseekingprosecution(e.g.,notshowingupforscheduledappointments).Forsome
cases,thereasonlistedwasageneralstatementthatthevictimwasnotcredible,whileforothers,the
reportstatedthatthevictimhadbeenproventobelying.
EXHIBIT2.4SummaryStatementfromthePoliceInternalAffairs
...
...
46
EXHIBIT2.5ReasonsWhySAKsWereNotSubmitted
AsnotedinChapter1:Introduction,priorresearchhasfoundthatpoliceinactioninsexual
assaultcasesisoftenattributedtosomefaultofthevictim(Caringella,2008;Lonsway&Archambault,
2012;Spohn&Tellis,2012).Furthermore,theseeminglyneutrallabelcomplainantrefusedto
prosecutemaynotreflectwhatactuallyhappenedgiventhatmultiplestudieshavefoundthatvictims
frequentlywithdrawfromtheinvestigationduetoinsensitive,victimblamingtreatment(HumanRights
Watch,2013;Kelley&Campbell,2013;Patterson,2011a,2011b).ThiscontentofthisInternalAffairs
reportiscertainlysuggestiveofthesepatterns.Policemaynothaveseenreasonforalarmorprovided
detailedinformationaboutthekitsbecausetheyfelttheyhadlookedintothematterandconcluded
thattestingwasnotwarrantedinthesecases.19Thethoroughnessandvalidityoftheinternal
investigationiscertainlysubjecttodebate,butitdoesshedsomelightastohowpoliceofficialswere
viewingtheissueatthetime.
19
ItseemslikelythatthiskindofinternalauditingiswhattheElectedProsecutorsoughttopreventwhencallingforan
independentauditoftheSAKsinpoliceproperty(seeExhibit2.1,above).
47
Thisinternalaffairsreport,aswellasanyotherdetailsabouttheSAKs,wasnotsharedwith
otherstakeholders,andsoformonths,theprosecutorsoffice,statepolice,andadvocacyorganizations
hadverylittleinformationaboutthestatusofthesekits.Intheabsenceofinformation,itisperhapsnot
surprisingthatbythetimetheDetroitSAKARPbegan,therewasastrongsentimentamongmany
stakeholdersthatthepolicewerenotbeingtransparentandforthcoming.Furthermore,thesustained
debateaboutthenumberofkitsinpropertyraisedconcernabouttheaccuracyofanyinformation
suppliedbythepolice,giventhatitseemedtootherDetroitstakeholdersthatthepolicewere
minimizingtheextentoftheproblem.Therefore,whentheARPbeganandthefirsttaskwastoassess
thescopeoftheproblem,theProsecutordecidedthattheonlycertainwaytoknowhowmanySAKs
wereinpropertywastohaulthemout,onebyone,andcountthem.
The Process of Conducting the Census in Detroit
Asnotedpreviously,forsomejurisdictions,determiningthenumberofunsubmittedSAKsin
policepropertycouldbeaccomplishedbyqueryingcomputerdatabases;however,ifcomputerized
recordsdonotexistand/orthereareconcernsabouttheaccuracyandcompletenessofthoserecords,
thenamanualcensusmaybenecessary.InDetroit,therewerestrongconcernsamongstakeholders
aboutthequantityandqualityofinformationprovidedbythepolice,soamanualcountwasseenasthe
bestwaytodeterminethetruescopeoftheproblem.However,suchanendeavoristimeconsuming,
resourceintensive,andfraughtwithnumerouslogisticalchallenges,including(butnotlimitedto):
AccessingtheSAKsPolicepropertyfacilitiesarelimitedaccessandonlycertainpersonnel
mayenterinordertomaintainthesecurityandintegrityofthecontents.Sendinginteamsof
staffandvolunteerstocountkitswasnotpossible;rather,thekitswouldneedtobebrought
outandcountedonsite,underthesupervisionofpropertyofficers;
48
MarshallingthepersonpowerforamanualcensusIndividualsselectedtoparticipateinthe
censusmustbetrustworthytoconducttheworkaccuratelyandrespecttheintegrityofthetask.
Furthermore,eachorganizationparticipatinginthecensuswouldneedtofreeupstaff
memberstimetoconductand/orsupervisethework.
DeterminingwhatinformationtocollectduringthecensusDefiningthescopeandpurposeof
thecensuswascriticalisthegoalmerelytocounthowmanySAKsareinproperty?Or,isthe
goaltotrytoextractasmuchinformationaspossibleabouteachSAK(e.g.,victimname,dateof
offense,testingstatusifknown,adjudicationstatusifknown)whileintheprocessof
countingthekits?SAKscannotbeopenedduringthecensus(theycanonlybeopenedatthe
testingfacility),soitwouldbenecessarytodeterminewhatinformationwasavailableabout
eachcasebasedonthedocumentationontheoutsideofthebox;
RecordingandtrackinginformationIfamanualcensusisbeingconducted,thenitislikely
thatcomputerizedrecordsdonotexistorthereisconcernaboutthecompletenessand
accuracyoftherecords.Assuch,thecensusprovidesanopportunitytocreatenewdatasystems
forcollecting,tracking,andsharinginformation.
Figure2.1TheStepByStepProcessoftheDetroitSAKCensus(followingpages)describeshow
Detroittackledtheseissuesandmanyotherissuesintheprocessofconductingitscensus.Figure2.1
listseachissuethathadtoberesolved,asummaryofthediscussionanddebateabouthowbestto
addresseachissue,anddecisionsmadebythecollaborativeteam(andwhytheydecidedwhattheydid).
Thecensuswasindeedalong,painstakingprocess,butitproducedhighquality,credibledatathat
answeredthekeyquestionregardinghowmanyunsubmittedSAKswereinpoliceproperty;these
findingswillbepresentedinthefollowingsection(TheResultsoftheCensus).
49
FIGURE2.1TheStepbyStepProcessofPlanningandConductingtheDetroitSAKARPCensus
ISSUE1
ISCONDUCTING
ACENSUS
NECESSARY?
DISCUSSION.Afterthe
discoveryofthekitsin
August,2009,the
Prosecutormademultiple
requestsforacomplete
countofallSAKsinpolice
property.Althoughthe
policedepartmenthada
computerizedproperty
trackingsystem,key
stakeholderswere
concernedastowhether
theinformationcontained
thereinwascomplete.
DECISION.Atthetimethisproject
beganinApril,2011,itwasstillnotclear
howmanySAKswereinpoliceproperty
andhowmanyofthosekitshadnever
beensubmittedfortesting.Therefore,
theProsecutordecidedthatacensusof
allSAKsinpolicepropertywasnecessary
inordertoobtainanaccuratecountand
understandthescopeoftheproblem.
DISCUSSION.The
prosecutorsofficestaff
assumedresponsibilityfor
planningandconducting
thecensus,butthepolice
expressedconcernsabout
theprosecutorsofficestaff
leadingthisworkbecause
itseemedlikelythatthe
censuswouldinvolvea
manualcountofSAKs(see
Issue3below);they
objectedtononpolice
personnelenteringtheir
securepropertyfacilities.
DECISION.Theprosecutorsofficestaff
expressedconcernsastowhetherdata
providedbythepolicewouldbe
accurateduetothefactthatthepolice
hadnotbeencompletelyforthcoming
regardingthenumberofSAKsin
propertyatthetimethekitswere
discoveredinAugust2009.Assuch,the
Prosecutordecidedthatthistaskwould
bedonebyprosecutorsoffice
personnel,undercloseobservationby
policepropertypersonnel.
ISSUE2
WHOSHOULDBE
INCHARGEOF
CONDUCTING
THECENSUS?
50
ISSUE3
CANTHECENSUS
BECONDUCTED
ELECTRONICALLY
ORWILLTHE
COUNTHAVE
TOBEDONE
MANUALLY?
ISSUE4
IFTHECENSUS
HASTOBEDONE
MANUALLY,
WHOCANHAVE
ACCESSTOTHE
KITSANDHOW
WILLACCESSBE
MANAGED?
DISCUSSION.Theproject
coordinatorreviewedwhat
informationwasavailable
inthepoliceproperty
databasetoascertain
whetheritwouldbe
sufficientforthepurposes
ofthecensus.The
databasetrackedlimited
information(e.g.,evidence
tagnumber,casenumber,
nameoforiginalofficerin
chargeofthecase,seize
date,victimname).
DECISION.Atthetimethatthecensus
wasbeingplanned,itwasnotyetknown
thattheformerpolicecrimelabhadan
Excelspreadsheetdocumentingwhich
SAKshadbeensubmittedfortesting
(whichcouldhavehelpednarrowdown
whichkitshad/hadnotbeensubmitted
fortesting).Therefore,basedonthe
informationavailable,itappearedthat
thecensuscouldnotbeconducted
electronicallyandthattheSAKswould
needtobemanuallycounted.
DISCUSSION4A.Amanual
countoftheSAKswould
requiretheassistanceof
manyindividuals,ifthat
taskwastobecompleted
inareasonableperiodof
time;therefore,itwas
importanttoconsiderwho
couldbeallowedtohave
accesstotheSAKs.
DECISION4A.Theprojectcoordinator
developedastaffingplanwherebythe
censuswouldbeconductedbyapoolof
prosecutorsofficestaff(assistant
prosecutingattorneys)andvolunteers
(lawstudents).
DISCUSSION4B.The
policevoicedconcerns
abouthavingnonpolice
personnelentertheir
securepropertyfacilityto
counttheSAKsbecause
eachSAKcontained
potentialevidencefora
criminaltrialandbecause
thepropertyfacility
containedotherevidence
relatedtoothercases.
DECISION4B.Becausenonpolice
personnelcouldnotenterthepolice
propertyroomtocounttheSAKs,the
SAKswerebroughtoutinbatchesand
countedinasecureroomwithinpolice
headquartersbytheprosecutorsoffice
personnel,underconstantsupervisionof
policepropertypersonnel.Thisdecision
satisfiedsecurityconcerns,butitwas
logisticallyburdensomeandtook
considerablepolicetime.
51
ISSUE5
AREALLTHESAKs
TOBECOUNTEDIN
ONEPHYSICAL
LOCATIONORARE
THEYSTOREDIN
MULTIPLEPLACES?
DISCUSSION.SomeSAKs
werestoredinanoffsite
propertystoragefacility;
someSAKswereinthe
mainpropertyroomin
policeheadquarters.
DECISION.AllSAKsweremovedtothe
mainpolicedepartmentpropertystorage
atpoliceheadquarterspriortobeginning
thecensus.
ISSUE6
HOWWILLTHE
TEAMBEABLE
TODISTINGUISH
OLDSAKs(THE
FOCUSOFTHE
CENSUS)FROM
CURRENTSAKs?
DISCUSSION6A.The
expressedconcernthat
becausetheSAKswere
dispersedthroughoutthe
propertyroomandstored
amongstothertypesof
evidence,itmadeitmore
difficulttoascertainthe
extentoftheproblem.
Also,thecutoffdatefor
theprojectscensushad
beensetatNovember1,
2009,soSAKsbefore&
afterthisdateneededto
bedistinguished.
DECISION6A.Thepoliceagreedto
restructuretheirstoragefacilityto
groupSAKsandtoseparatethepre
2009SAKs(thoseinthescopeofthis
project)frompost2009SAKs.Pre2009
SAKsincludedinthecensuswere
labeledandstoredseparately.
DISCUSSION6B.Asthe
policepersonnelbegan
pullingpolicerecordsto
assistinthecensus(see
Issue8below),the
notedthatthesefilesmay
needtobeaccessedagain.
DECISION6B.Thepolicesexcrimesunit
storedtherecordsassociatedwiththe
casesinthecensusseparatelytofacilitate
easyaccesstothosefiles.
52
ISSUE7
HOWSHOULD
SAKsBE
TRACKED
DURINGTHE
CENSUSTO
ENSURETHEY
ARENT
COUNTED
TWICE?
ISSUE8
HOWMUCH
INFORMATION
ABOUTEACHSAK/
CASESHOULDBE
EXTRACTED
DURINGTHE
CENSUS?
DISCUSSION.Thephysical
logisticsforthecensus
wereburdensomeand
manystaffandvolunteers
fromtheprosecutors
officeworkedonthe
census(seeIssue4above),
sotherewasconcernthat
inthiscomplexprocess,
someSAKscouldbe
accidentallycountedtwice
(orsomecouldbemissed).
DISCUSSION8A.The
outsideoftheSAKbox
containedlimited
informationaboutthe
kit/case;moredetailed
informationwasavailable
onthepaperworkinside
thekit,buttheSAKscould
notbeopenedduringthe
census.
DECISION.Theprojectcoordinator
createdanewlabelingsystemsuchthat
SAKsthathadalreadybeencountedin
thecensusweretaggedwithacolor
codedstickertopreventduplicate
counting.Astheteamwasableto
ascertainthetestingsubmissionstatus
ofeachkit(seeIssue8below),another
colorcodingschemewasintroducedto
distinguishkitssubmittedfortestingvs.
unsubmittedkits.
DECISION8A.Theprojectcoordinator
createdaformtorecordkeyinformation
fromtheoutsideoftheSAKs.The
prosecutorsofficestaffcompletingthe
censuswouldcopyinformationfromthe
SAKstothisform(whichwouldthenbe
enteredintocomputerdatabaseata
latertime).
Issue8continuedonthenextpage.
53
ISSUE8
HOWMUCH
INFORMATION
ABOUTEACHSAK/
CASESHOULDBE
EXTRACTED
DURINGTHE
CENSUS?
DISCUSSION8B.The
informationontheoutside
oftheSAKwasnotenough
todeterminewhethereach
SAKhadbeensubmittedto
thecrimelabfortesting
and/orifthecorresponding
casehadbeenadjudicated.
Theprosecutorswantedto
knowhowmanycaseshad
notbeenpreviously
adjudicatedinorderto
planhowmanynewcases
theymightbeopening.
DECISION8B.Theprosecutorsoffice
recommendedareviewofthepolice
recordscorrespondingtoeachSAK,as
theserecordscouldholdmore
informationaboutacase(e.g.,ifthe
casehadbeensenttotheprosecutors
officeforwarranting).Policerecords
mayalsohavedocumentationfromthe
medicalproviderand/orcrimelab
indicatingiftheSAKhadbeensubmitted
fortesting.(NOTE:Itwasstillnotyet
knownthattheformerpolicecrimelab
hadanExcelspreadsheetdocumenting
whichSAKshadbeensubmitted.)
DISCUSSION8C.Pullingthe
policefilesforeachSAK
wouldhelptoascertain
testingstatusand
adjudicationstatus,but
doingsowouldrequire
extensivetimebypolice
personnel,whichthey
expressedconcernsabout.
Furthermore,theydidnot
wantpolicerecordstaken
offsitefrompolice
headquarters/unitoffices.
DECISION8C.Thepoliceretrievedthe
requestedfilesasquicklyastheycould,
giventheirstaffingresources.The
policeallowedonlytheprosecutors(as
opposedtoprosecutorsandtheir
volunteers)toreviewthefiles.Thefiles
hadtoremainonsiteatpolice
headquarters/unitoffices.
54
ISSUE8
HOWMUCH
INFORMATION
ABOUTEACH
SAK/CASE
SHOULDBE
EXTRACTED
DURINGTHE
CENSUS?
DISCUSSION8D.Asthe
policereportswerepulled
andreviewed,additional
detailsaboutthe
SAKs/casesbecameclearer.
Theprosecutorsoffice
stafffeltthatitwas
importanttostart
capturingthisinformation
asitwouldlikelybeneeded
forlaterdecisions
regardingtesting,
investigation,and
prosecution.
DECISION8D.Prosecutorsofficestaff
createdtwonewformstodocument
additionalinformationabouteach
SAK/case(oneformforthereviewof
complaintbooks/warrantbooklog;one
formforthecasefilereview).Asthe
prosecutorsreviewedpolicerecords,
informationwastransferredtothese
forms(whichwouldthenbeentered
intocomputerdatabaseatalatertime).
DISCUSSION8E.Asthe
policereportswerepulled
andreviewed,itbecame
clearthattheserecordsdid
notconsistentlydocument
whetheraSAKhadbeen
submittedfortesting,what
testingwasperformed,and
theresultsofthattesting.
DECISION8E.Theprosecutorsoffice
requestedthecrimelabreportsto
ascertaintestingresults.Afterthisitwas
reportedthattheformerpolicecrimelab
hadanExcelspreadsheettrackingwhich
SAKshadbeensubmittedfortesting.
Comparingthelistofkitsinpolice
propertywiththislistallowedtheteam
todeterminewhichkitshadbeen
submittedtothelab.
55
ISSUE9
HOWSHOULD
INFORMATION
COLLECTED
DURINGTHE
CENSUSBE
TRACKEDAND
STORED?
ISSUE10
ASSAKsARE
BEINGCOUNTED,
SHOULDTHEYBE
PRIORITIZEDFOR
TESTINGATTHE
SAMETIME?
DISCUSSION.Therewasno
single,unifieddatabase
thattrackedinformation
oneachSAK.Itwasnoted
thatthereneededtobea
waytosynthesizethe
informationgatheredfrom
thevarietyofsourcesfor
theSAKsthatwerethe
focusofthecensusand
thatlongtermplansfor
betterdatasystems
neededtobecreated.
DISCUSSION.Asthe
prosecutorsbegan
reviewingpolicefiles,they
becameincreasingly
concernedabouttesting
SAKsandmovingcases
forwardforprosecution.
Theywereparticularly
concernedaboutcasesthat
wereclosetoreaching
statuteoflimitation(SOL).
Inaddition,astheywere
examiningpolicefiles,they
sawmanycasesthatthey
feltshouldhavebeen
forwardedforwarranting
atthetimeoftheoriginal
reportandwereeagerto
begininvestigationand
prosecutionofthesecases.
DECISION.Theprojectcoordinator
createdanExcelspreadsheettotrack
theinformationcollectedinthecensus.
Later,anassistantprosecutorpulled
thatinformationintoanAccess
database,asthatsoftwareofferedmore
optionsfordatastorageandretrieval.
AfterthepoliceITstaffmemberbecame
involvedinproject,theAccessdatabase
wasexpandedtoaccommodatenew
informationbeingcollectedforthe
testingplanandvictimnotification.
DECISION.Theprosecutorsoffice
createdaprioritizationsortingplanin
whichcasesnearingtheirstatuteof
limitationandcasesinvolvedstranger
perpetratedcrimes,weredesignatedto
besenttothelabfirst.Thecollaborative
wasconcernedaboutthisplanonthe
groundsthatithadbeencreated
withoutteaminput.
Theprosecutorsofficeagreedtohalt
theirinitialprioritizationplanandto
workwiththefullcollaborativetoco
developatestingplan.
56
The Results of the Census: How Many Unsubmitted SAKs in Detroit

ThemanualcensuscountedallSAKsinpoliceproperty(uptoNovember1,2009),20which
totaled11,303SAKs21(seeFigure2.2nextpage).Thekitsspannednearlythirtyyears,from1980(the
dateoftheoldestkitfound)to2009(thecutoffforthecensus).WhilecountingtheSAKs,staffnoticed
thatsomeevidencetagsontheoutsideoftheboxeshadnotesindicatingthatthecontentsinsidewere
notsexualassaultrelatedevidence.Forexample,sometagsidentifiedtheevidenceintheboxas
clothes,knife,bottle,etc.,anditwasnotclearwhethertheitemswerecrimesceneevidence
relatedtoasexualassaultcase,orwhetherpolicepersonnelhadusedaSAKboxasacontainertostore
evidence.TheSAKscouldnotbeopenedduringthecensus,sostaffmadeanoteoftheseinstancesand
countedthemanyway.Afterthecensuswascomplete,theProjectCoordinatorandITconsultantfor
thepolicedepartmentworkedtogethertodeterminehowmanySAKswerenotinfactsexualassault
medicalforensicexamevidenceandtoresolveotherassortedissuesthatcroppedupinthecensusthat
couldaffecttheoverallcount(e.g.,possibleduplicateIDnumbers,transposednumbers,etc.).Basedon
theirreviewoftheevidencetagsandsupplementalinformationinthepolicepropertydatabase,84SAKs
weresubtractedfromthecount,leaving11,219SAKsinthecensus.
TheDetroitSAKARPcollaborativethenattemptedtodeterminethetestingandadjudication
statusofthese11,219SAKs(seeTable2.1,above).Asnotedpreviously,althoughthepolicecrimelab
didnothaveacentralizeddatabasethatcouldprovidetestingstatusforeachkit,aforensicscientistin
theunithadcreatedanExcelspreadsheetofkitsthathadbeensubmittedfortesting.Basedonthose
20
The400SAKsthatwererandomlysampledfortheOVWfundedThe400Projectwereincludedinthiscensusinorderto
obtainacomprehensive,completecountofallSAKsinpoliceproperty,currenttoNovember1,2009.
21
Monthslater,whenoneoftheSAKsthatwasselectedtobetestedaspartoftheDetroitSAKARPwasopenedatthelab,it
containedbiologicalsamplesfromtwodifferentvictims;thereafter,theDetroitcollaborativerevisedthetotalcensuscountto
11,304(+1giventhatthesamplesforanothervictimhadbeendiscovered).
57
records,itwasdeterminedthatofthese11,219SAKs,2,512hadalaboratoryIDnumberassignedto
them,22indicatingthattheyhadbeensubmittedfortesting,thoughitwasnotclearhowmanykitshad
infactbeentestedforDNA.23Atotalof8,707SAKshadneverbeensubmittedfortesting.
FIGURE2.2SummaryoftheDetroitSAKARPCensusResults
11,303SAKsCountedintheCensus
Minus84SAKs
11,219SAKsTestingStatusToBeDetermined
2,512
SAKs
Submitted
forTesting
8,707SAKs
Unsubmitted/Untested
22
AsnotedintheIntroduction,membersoftheDetroitSAKARPreviewedadraftofthisfinalreportbeforeitwassubmitted
anddissentingfeedbackfromstakeholderswouldbenotedthroughoutthereport.Duringthisreview,representativesfromthe
policedepartmentprovideddocumentationtoindicatethat2,915SAKshadlaboratoryIDnumbers(not2,512).Whenthe
researchteamcomparedthematerialscollectedatthetimeofthecensusandthoseprovidedlater,wenoticedthatthe
dates/timeframeusedforthecensusdidnotmatchthedatesonthesupplementaldocuments,whichlikelyexplainsthe
differencesinthenumbers.Inthisreport,wepresentthe2,512numberasitisthefigurewecouldindependentlyverifyas
correspondingtotheexactdates/timeframeforthecensus.
23
ThroughoutthedurationoftheDetroitSAKARP,teammembersfrommultipleorganizationssoughttoclarifyhowmanyof
theSAKsthathadbeensubmittedtothepolicedepartmentcrimelaboratoryhadinfactbeentestedforDNA.Atthetimethis
finalreportwasprepared,thisissuehadnotbeenfullyresolved,butbasedonthedatathatwereavailable,itwasclearthatnot
allofthesubmittedkitshadbeentestedforDNA(thoughtheexactnumberwasstillunderreview).Assuch,itisnotaccurate
toassumeorconcludethatallSAKssubmittedtothepolicecrimelabfortestinghadindeedbeentestedforDNA.
58
Astotheadjudicationstatusofthesekits,thelaborrequiredtoobtainthatinformationfor
11,000+kitswaswellbeyondthetimeandstaffingavailableforthisproject.Computerizedcourt
recordsdidnotinterfacewiththepolicepropertydatabase,andadjudicationstatuscouldnotalwaysbe
determinedfromthecomputerizedrecords,soprojectstaffwouldneedtoconductmanualsearchesof
paperrecordstoobtaintheinformationneeded.GiventhelackofintegrateddatasystemsinDetroit,
thelaborexpendituretoscreeneachandeverykitforadjudicationwasprohibitive,andthereforewe
werenotabletodeterminetheadjudicationstatusforallSAKsinpoliceproperty.Thecollaborative
decidedinsteadtotrytoscreenasmallersubsetofSAKsforadjudicationstatus,specificallythosethat
wouldbetestedinthecontextofthisproject(seeChapter4:DevelopingandEvaluatingaSAKTesting
Plan).InThe400Project,all400randomlysampledSAKswerescreenedforadjudicationstatus,and
thosefindingssuggestthatmost(~87%)ofthe11,000+SAKswerelikelyassociatedwithnon
adjudicatedcases.
Summary & Conclusions: The Scope of the Problem

Amanualcensuswasconductedduetoconcernsaboutthepolicedepartmentstransparency
andengagementintheyearsaftertheAugust,2009discoveryofthousandsofSAKsinpoliceproperty.
Theinitialestimatesmadeatthetimeofthediscovery(~10,000to11,000SAKs)wereremarkably
accurate,asthecensusresultsrevealedthattherewere11,219SAKswereinpoliceproperty(asof
November1,2009).Inthisproject,weattemptedtodeterminethetestingstatusoftheseSAKs,and
basedontherecordsavailable,2,512SAKshadlaboratoryIDnumbers,indicatingthattheyhadbeen
submittedtothepolicedepartmentcrimelab,and8,717hadneverbeensubmittedfortesting.Notall
oftheSAKssubmittedtothecrimelabweretestedforDNA,althoughitwasnotpossibleinthetimeline
ofthisprojecttodeterminehowmanywerenot.
59
Intheprocessofconductingthecensus,itwasclearthatthenumberofunsubmittedSAKswas
notconstantovertimesomeyearshadmoreunsubmittedkitsthanothers.Staffmemberswho
countedthekitsnotedthatsomeyearsseemedtogoonforever(i.e.,therewerealargenumberof
kitstocountthatyear)andotheryearsweremuchquickertocount(i.e.,therewerefewerkitsthat
year).Why?Topursuethisissue,webeganourworkonthesecondgoalofthisproject(toidentifythe
underlyingfactorsthatcontributedtowhyDetroithadsomanyunsubmittedSAKs)withadetailed
historicalanalysisoftheorganizationalresourcesandpoliciesregardingSAKsforthecrimelab,police
department,prosecutorsoffice,medicalsystem,andvictimadvocacyalloftheorganizationsthat
havedirectorindirectinfluenceonSAKprocessing.Inthenextchapterofthisreport,Chapter3:WhySo
ManyUnsubmittedSAKsinDetroit,wewilldescribeourfindingsfromthishistoricalanalysisandthen
returntothecensusresultstoexplorewhythenumberofunsubmittedSAKsfluctuatedovertime.
60

Why So Many Unsubmitted Sexual Assault Kits (SAKs) In Detroit
ThesecondgoaloftheDetroitSAKARPprojectwastoidentifytheunderlyingreasonswhythere
weresomanyunsubmittedSAKsinpoliceproperty.24Howdoesapolicedepartmentaccumulate
11,000+kitsoverthirtyyears,mostofthemnevertested?AsnotedinChapter1:Introduction,this
projectgoalanditsassociatedquestionsareuniquefromtheotherprojectaimsinthattheyrequire
takingastepbacktounderstandwhatwentwrongandwhy.Giventhatfocus,theresearchersworked
independentlyfromthelargercollaborativeteamtoprovide,totheextentpossible,anoutside,
independentexaminationoftheseissues.25
Thepurposeofthischapteristopresentourfindingstothisfundamentalquestionofhowand
whyDetroithassomanyunsubmittedSAKs.26Specifically,thischapterhasthreemainsections.First,to
setthestageforthisresearch,wewilldescribethetheoreticalmodelthatguidedthisinquiry:ecological
systemstheory.Thistheoreticalframeworkiswellestablishedingenderbasedviolenceresearch(see
Whiteetal.,2011forreviews)anditiswellsuitedforthecurrentstudyasitprovidesamultisystemic
frameworkforunderstandinghowinterdependentorganizationsfunctionovertime.
24
TheNIJSolicitationspecificallymentionedthisasaprioritytopic:NIJisinterestedinlearningabouttheunderlyingfactors
thatcontributetothisunsubmittedSAKevidence(p.4).
25
Thisanalysisisindependenttotheextentpossiblegiventhatresearchteamwasamemberofthecollaborative(consistent
withtheactionresearchparadigm.Inanefforttoseparatethisworkasmuchaspossiblefromtheotherprojectgoals,the
researchteamdidnotpresentthespecificquestions,methods,andresultstothefullcollaborativeuntilafterawellwarranted
setoffindingshadbeenestablished(seeErickson,1986),andthosefindingswerereadyformemberchecking(consistentwith
standardpracticesinqualitativeandmixedmethodsresearch,seeAppendixB:ProjectMethodology).
26
ThroughoutthisChapter,thepronounsour/wereferonlytotheresearchteam,anddoesnotincludeorimplythe
involvement,views,orofficialpositionsofanyoftheorganizationswithintheDetroitSAKARP.
61
Second,toplaceourfindingsincontext,weconductedathirtyyearhistoricalanalysisofallkey
organizations(police,policecrimelab,prosecution,medical,systemsbasedvictimadvocacy,and
communitybasedvictimadvocacy)tounderstandtheirpoliciesandpracticesaswellasresourcesthey
hadavailabletotestSAKs/respondtosexualassaultvictimsfrom1980(thedateoftheoldestkitfound
inthecensus)to2009(theendofthescopeofthisproject).27Usingamultistage,sequential
exploratorymixedmethodsdesign(Creswell,2010;Creswell&Clark,2011;Creswelletal.,2003),we
collectedmultipletypesofqualitativedata(ethnographicobservations,stakeholderinterviews,and
archivalrecords)todocumentkeyturningpointsandchangesoverthosethirtyyearsinSAKsubmission
practices(specifically)andSAservices(generally)(seeAppendixB:ProjectMethodology).Todetermine
whetherthosehistoricalchangesaffectedSAKsubmissionratesovertime,wereturnedtothecensus
dataandusingquantitativemultilevellongitudinalmodeling,weexaminedwhetherthesefactors(e.g.,
staffingcuts,policychanges)weresignificantlyassociatedwiththeprobabilityofSAKsubmissions.
Perhapsnottoosurprising,theresultsofthequalitative/quantitativehistoricalanalysisraisedas
manynewquestionsasitanswered.ThenumberofunsubmittedSAKsinpolicepropertyincreasedand
decreasedovertime,andsomeofthesepatternscouldbeexplainedbykeyhistoricalevents,but
overall,thekeyfindingfromthisanalysiswasthatthevastmajorityofSAKseachyearwerenot
submittedfortesting.Therefore,inthethirdsectionofthischapter,wewillpresentfindingsfroman
additionalstageofqualitativedatacollectionthatsoughttoidentifythefrontline,ontheground
practicesinthepolicedepartmentandotherorganizationsinthissystemicnetworkandtoexaminehow
thosepracticesaffectedSAKsubmissions.28Bringingtogetherdatafromethnographicobservations,
stakeholderinterviews,andsexualassaultpolicereportsassociatedwithunsubmittedSAKs,we
27
Itisimportanttoemphasizethatthisanalysisreflectshistoricalpracticesintheorganizationsstudiednottheircurrent
resources,leadership,policies,andpracticesregardingsexualassaultinvestigationsandrapekittesting.
28
Again,thisanalysisfocusedonpastpractices(upto2009)inthefocalorganizations,nottheircurrentapproachestosexual
assaultinvestigationsandrapekittesting.
62
examinedthedecisionmakingprocessesandinstitutionalnormsofpolicepersonnel,astheywerethe
entityresponsibleforsubmittingSAKsfortesting.Wealsoexploredhowinteractionswithother
organizationsinthisinterdependentsysteminfluencedpolicedecisionmakingregardingSAK
submissions.Takentogether,thesequalitativeandquantitativedatahelpshedsomelightonthehistory
ofDetroitsresponsetosexualassaultandwhysomanyunsubmittedSAKsendedupinpoliceproperty.
Ecological Systems Theory: Understanding Interdependent Organizations

OurresearchonhowandwhyDetroithassomanyunsubmittedSAKswasinformedby
ecologicalsystemstheory(Bronfenbrenner,1979,1986,1995;Kelly,1966,1968,1971;Trickett,1984,
1996,2011;Trickett,Kelly,&Vincent,1985).Thistheoreticalmodelpositsthathumanbehaviorand
socialphenomenaareshapedbymutuallyinfluencingrelationshipsamongindividualsandthesettings
inwhichtheyliveandwork(Bronfenbrenner,1979;Kelly,1970).Inotherwords,individualbehaviors
and/orthecollectivebehaviorsofindividualswithinorganizationsdonotoccurinisolation,butare
activelyshapedbyinteractionswithothersandareresponsivetofeedback(bothpositiveandnegative),
whichaffectssubsequentbehaviors(Kelly,Ryan,Altman,&Stelzner,2000).Furthermore,settinglevel
factors,suchasleadership,resources,andnormsofanorganizationorcommunityalsodictatebehavior
(Schensul&Trickett,2009;Trickettetal.,1972).AcoretenetofthistheoryisKellys(1968)Principleof
Interdependence,whichstatesthatcomponentswithinasocialsystemfunctioninrelationtoeachother
andchangesinonecomponentofasystemwillproducechangesinanother.Becausetheremaybe
expectedandunexpectedchanges(bothpositiveandnegative),researchersmustfocusnotonlyonthe
targetpopulationorsetting,butalsoonextendedpersons/settingswhohavedirectorsecondary
contactwiththosetargets.Therefore,ecologicalsystemsresearchmustmapthelocalcontextidentify
whichorganizationsworktogether,determinewhatrolesandresponsibilitiestheyhavetoeachother
andtooutsideparties,andexaminehowforcesexternaltothesystemmayalsoaffectitsfunctioning.
63
Ecologicalsystemstheoryhasbeenwidelyusedinresearchongenderbasedviolence,including
effortstomaptheunderlyingetiologyofvictimizationandperpetration(Grauerholz,2000;Heise,1998;
White&Kowalski,1998),developpreventiveinterventions(CentersforDiseaseControlandPrevention,
2004;WorldHealthOrganization/Jewkes,Sen,&GarciaMoreno,2002andKrug,Mercy,Dahlberg,&
Zwi,2002),assessposttraumasequelae(Campbell,Dworkin,&Cabral,2009;Harvey,1996;Koss&
Harvey,1991;Neville&Heppner,1999),andevaluatevictimspostassaulthelpseekingexperiences
(Campbell,1998;Campbell,Patterson,&FehlerCabral,2010;Campbelletal.,1999;Campbelletal.,
2001).Applyingthismodeltothecurrentcontext(unsubmittedSAKs)focusestheresearchonthe
systemoforganizationsthatworktogethertotestSAKs,investigatereportedcrimes,prosecutecriminal
sexualoffenses,andprovidesupportandadvocacytosurvivors.Figure3.1(nextpage)depictsthe
systemsmapinDetroitoftheorganizationsandlinkagesbetweenorganizationsthatwerethefocusof
thisresearch.
InDetroit,asinmanyotherjurisdictions,thepolice(usuallythesexcrimesunit/sexcrimes
investigators)areresponsibleforsubmittingaSAKtothecrimelabforanalysis;29inturn,thecrimelab
analyzestheevidenceandreportsthefindingsbacktothepolice.30InDetroit,thepolicesexcrimes
unitandthecrimelabwerewithinthesameorganization(i.e.,thepolicedepartment),31andboth
entitieswouldbeexpectedtobeinfluencedbydepartmentwideissuesandproblems(e.g.,budget
29
Prosecutorscanalsorequesttohavearapekittested;however,inmanyjurisdictionsthroughouttheUnitedStates,thevast
majorityofsexualassaultsreportedtothepolicearenotreferredtoprosecutors(Campbell,Bybee,Shaw,Townsend,&Karim,
2014),whichmeansthatprosecutorsdonotknowthatarapekitevenexistsandneedstobetested.
30
Insomejurisdictions,testingresultsarealsoreportedsimultaneouslytotheprosecutorsoffice.Historically,thiswasnotthe
practiceinDetroit,buthassincebeenimplementedafterthediscoveryoftheunsubmittedSAKsinpoliceproperty(seeChapter
6:SummaryofFindings,Implications,andCommunityChanges).
31
InDetroit,thepolicehadtheirowncrimelabuntilSeptember25,2008;afterthat,forensicevidencewassenttoand
processedbythestatepoliceforensiclab,whichisamultisite,statewidelaboratorysystem.Althoughthecurrenttrendin
forensicsciencesistoestablishindependentcrimelabs(i.e.,independentfromlawenforcementagencies)(Cown&Koppl,
2011;Edwards&Gotsonis,2009),thismodeloflabswithinpolicedepartmentswascommoninthe1990sand2000s.
64
FIGURE3.1SAKRelatedInterOrganizationalSystemicMap
cuts),thoughthesexcrimesunitandthelaboratoryhaddifferentchainsofcommand.Thesetwounits
arealsointerdependentsuchthatpolicesSAKsubmissionspoliciesarelikelydeterminedinpartbythe
labsSAKtestingpolicies.Inotherwords,thepolicecansubmitonlywhatthelabwillaccepttotest,so
understandingthatdynamicwillbeessential.Thepoliceandcrimelabalsointeractwiththe
prosecutorsoffice,ascasesareforwardedforpossibleprosecution(e.g.,warrantrequests).Thesethree
units(police,crimelab,andprosecution)areinterdependentsuchthatallmustworktogetherifacase
isgoingtobeadjudicated,sounderstandingthepatternsofcommunicationamongthesethreepartsof
thecriminaljusticesystemiscritical.32
32
Ecologicalsystemstheoryemphasizestheimportanceofunderstandingtheinfluenceofexternalforcestothefunctioningof
asystem.Inthecriminaljusticesystem,theworkofthelawenforcement,forensicsciences,andprosecutionisaffectedbythe
65
Inthatsamevein,theworkoflegalorganizations(police,crimelab,prosecutorsoffice)is
dependentuponthemedicalsystemtoconductthemedicalforensicexamandcollecttheSAKs.Over
thepasttwentyyears,therehavebeenradicalchangesinhowthemedicalsystemrespondstothe
needsofsexualassaultvictimsandthetaskofforensicevidencecollection.Manycommunities
throughouttheUnitedStateshaveimplementedSexualAssaultNurseExaminer(SANE)/SexualAssault
ForensicExaminer(SAFE)programs,wherebyspeciallytrainednursesnowprovidetheseservices(see
DepartmentofJustice,2013;Campbell,Patterson,&Lichty,2005;Ledray,1999forreviews).Itis
importanttoexaminehowmedicalforensicexamshavebeenconductedinDetroit(i.e.,whetherby
speciallytrainedhealthcareprovidersorstandardhospitalemergencydepartmentpersonnel),and
howthemedicalcommunityhascommunicatedwithmembersofthelegalsystemovertime.
Throughouttheentiremedicalandlegalprocess,sexualassaultvictimsoftenneedsupportand
advocacytonavigatethesesystemsandaccesstheservicestheyneed.Therearetwomaintypesof
victimadvocacyprograms:systemsbasedprograms(e.g.,policevictimadvocacyprograms)and
communitybasedprograms(e.g.,rapecrisiscenters)(DepartmentofJustice,2006;Cole&Logan,2008).
Ingeneral,systemsbasedadvocacyprogramsareorganizationallyembeddedwithinthelegalsystem
(e.g.,aunitwithinthepolicedepartment)andfinanciallytiedtotheirhostorganization(e.g.,staffing
andsupervisionistypicallyprovidedbythepolicedepartment).Typically,theseprogramscannotoffer
victimsconfidentialcommunication,meaningthatanythingsurvivorsdisclosetosystemsbased
advocatescanbesharedwithpolicepersonnel.Bycontrast,communitybasedadvocatesare
organizationallybasedinnonprofitagenciesandtheycanassureconfidentiality,suchthatany
communicationwithsurvivorsorthefactthatthereevenwascommunicationcannotbedisclosedto
anotherparty,withoutwrittenconsentofthesurvivor(DepartmentofJustice,2006;Cole&Logan,
normsandexpectationsofthejudgesbeforewhomcaseswillbetried.Inthisproject,wemadeconsiderableeffortstoengage
membersoftheDetroitjudicialcommunityinthiscomponentoftheproject,butalldeclinedtoparticipateinstakeholder
interviews.Assuch,weacknowledgethatoursystemicanalysisislimitedandincomplete.
66
2008;Martin,2005).Furthermore,communitybasedadvocatesmayinterveneonbehalfofsurvivorsto
challengeinstitutionalpracticesiftheneedsandwishesofthevictimsarenotbeingrespected
(Campbell,Baker,&Mazurek,1998;Cole&Logan,2008;Maier,2008;Martin,2005).InDetroit,the
policedepartmenthashadalongstandingsystemsbasedvictimadvocacyprogram,buttherewas
limitedcommunitybasedadvocacyavailable,sounderstandinghowthesetwotypesofadvocacy
programsassistedvictimswillbeimportanttocapture.
Inmostcases,thisentiresystemoflegal,medical,andadvocacyserviceswillnotbepressedinto
dutyifsurvivorsdonotfirstcomeforwardtoseekamedicalforensicexamand/orreporttheassaultto
thepolice.Priorresearchsuggeststhatvictimsreasonsforseekinghelparecomplex(Campbelletal.,
2009;ClayWarner,&McMahonHoward,2009;Dumont,White,&McGregor,2009;Patterson,
Greeson,&Campbell,2009;Paul,Zinzow,McCauley,Kilpatrick,&Resnick,2013;Resnick,Acierno,
Holmes,Dammeyer,&Kilpatrick,2000).Someareconcernedabouttheirhealth(e.g.,pregnancyand
sexuallytransmittedinfections)(Campbelletal.,2009;Pauletal.,2013)andaretoldtheymustahave
medicalforensicexamtoobtainthatcare,33whichinvokesalegalcomponentthatvictimsmayormay
nothavewanted(Martin,2005;Young,Bracken,Goddard,&Matheson,1992).Otherspurposelyreport
tothepoliceinordertoprotecttheirownsafety,totryprotectotherwomen,andtotrytohold
perpetratorsaccountablefortheiractions(Johnson,1985;NationalCenterforVictimsofCrime,2008).
Forsomesurvivors,thedecisiontocontactthelegalandmedicalsystemwasnottheirsandwasinstead
madebysomeoneelse(e.g.,adolescentvictimswhoarebroughtforcarebytheirparents/guardians;
survivorswhowereunconsciousorseriouslyinjuredandcarewassoughtontheirbehalfbyothers)
(Campbell,Greeson,Bybee,Kennedy,&Patterson,2010;ClayWarner&McMahonHoward,2009).
33
ThefederalViolenceAgainstWomenActof2005(whichwentintoeffectin2009)changedthispracticesuchthatvictimscan
nowobtainamedicalforensicexamwithoutfilingpolicereport/legalinvolvement.However,atthetimethattheseSAKswere
collected,victimswerelikelytoldthathadtomakeapolicereportinordertoreceivepostassaulthealthcare.
67
Thereasonsandpathwaysbywhichvictimsenterintothelegalandmedicalsystem
undoubtedlyaffecttheirexpectationsandexperiencesandinturn,affecthowsystempersonnel
respondtosurvivors(Campbell,Greeson,FehlerCabral,&Kennedy,2014).Understandingthese
interdependenciesiscritical,butpracticallydifficultinthecontextoftheproblemofunsubmittedSAKs.
AsnotedpreviouslyinChapter2:HowManyUnsubmittedSAKsinDetroit,theDetroitSAKsdateback
manyyearsuptothirtyyearsandrecontactingvictimsraisescomplexlegal,psychological,and
ethicalissues.Therefore,theresearchteamdecided,inconjunctionwithourInstitutionalReviewBoard
(IRB),thatwewouldnotattempttorecontactvictimsforthisspecificresearchtask/projectgoal(i.e.,
researchingtheunderlyingreasonswhyDetroithassomanyunsubmittedSAKs).Victimsmayneedto
berecontactedbylegalsystempersonnelaftertheirkitsaretested(seeChapter5:Developing&
EvaluatingaVictimNotificationProtocol),andwedidnotwantanypriorcontactwiththeresearchteam
tointerferewiththosenotifications.34Therefore,thesystemicanalysisdescribedinthischapteris
notablymissinganessentialpartofthatthesystemthesurvivors.Weacceptthislimitationgiven
possibleunintendednegativeconsequencesonfutureprosecutions.35
Insummary,anecologicalsystemstheoryperspectivefocusesonunderstandingthedynamics
withinandbetweenorganizationsthatworktogetherasinterdependentpartsofalargersystem.To
exploreourfocalquestionhowandwhydoesDetroithavesomanyunsubmittedSAKsweexamined
34
Iftheresearchteamhadcontactwithvictimspriortolegalnotificationabouttestingresults,itispossiblethatthe
researcherscouldbecomeapartytothecase(i.e.,awitness)andalthoughcommunicationswiththeresearchteamwouldbe
federallyprotectedandcouldnotbedisclosed,itwouldcreateadditionalcomplicationsinalreadycomplicatedcases,which
wasafocalconcernoftheElectedProsecutor.Thepossibilityofinterviewingonlythevictimswhosecaseswouldnotbere
openedraisedbothpracticalandscientificconcerns.Fromapracticalpointofview,itwasimpossibletoknowwhichcases
mightbenotifiedandwhichmightnot(ormightnotbeintheimmediatefuture,butcouldbelater)becausethecollaborative
wasstillworkingthroughitsvictimnotificationprotocols.Fromascientificpointofview,interviewingonlysurvivorswhose
caseswouldnotbereopenedisproblematicasitisclearlyanonrepresentative,incompletesample.Assuch,thescopeofour
researchrelatedtothisprojectgoalwaslimitedtosystemstakeholdersandsystemarchivalrecords.
35
Weappreciatethatnotallvictimsmaychoosetoparticipateinprosecution(orthatallcasescouldorwouldbeprosecuted).
However,keepingoptionsopenforsurvivorsisakeytenetofavictimcenteredapproach(seeSexualViolenceJusticeInstitute,
2008),andwefeltthattheresearchmethodsinthisactionresearchprojectneededtobeconsistentwiththoseaims.
68
thedecisionmakingprocessesandinstitutionalnormsofpolicepersonnel,astheywereultimately
responsibleforsubmittingaSAKfortesting.Giventhatthepoliceareonepartofabroadersystemof
serviceproviders,wealsoexaminedhowinteractionswiththecrimelab,prosecutorsoffice,medical
system,andvictimadvocacyorganizationsinfluencedpolicedecisionmakingregardingSAKsubmissions.
Historical Context: Organizational Resources & Practices 1980 - 2009

Overview
Whentheresearchteambeganinformal(andthenformal)interviewswithrepresentativesfrom
Detroitareaorganizationsaboutwhythereweresomanyunsubmittedsexualassaultkits,stakeholders
acrossallorganizationsemphasizedthattheproblemmustbeplacedinitsproperhistoricalcontext.As
onememberofthecollaborativenoted:
Weretalkingaboutkitsfromalongtimeago,letsnottakeour2011expectationsofDNA,
CODIS,andCSIandallthat,andapplytodaysstandardstobackthen.WedidnthaveDNA
testingforalongtime,didnthaveCODIS,weneedtojudgewhathappenedbasedonwhatwas
possible,atwhattime...wevegottoputwhathappenedincontext.
Stakeholdersalsoemphasizedthattheproblemneededtobeconsideredwithinthebroadercontextof
Detroitshistoryasacitythathasstruggledfordecadeswithchronicresourcedepletion:
ThisisDetroit,notNewYorkCity,notLosAngeles.ThisisDetroitandtheresnocityintheU.S.
thatslikeDetroit...YouhavetokeepinmindwhatwaspossibleinDetroit.Whattheyhadin
othercities,youknow,money,personnel,technology,well,wedidnt.Everythingcameonline
hereyearsafteritdidinotherplaces.Thingsothercitiesweredoing,wecouldntdo.Notthat
wedidntwantto,wedidnthavetheresources(emphasesinoriginal).
69
Giventheseimportantcaveats,webeganbyexaminingtheorganizationalresourcesofthe
policedepartmentandthentheotherorganizationsinthesystem(crimelab,prosecutorsoffice,
medical/SANE,andadvocacy)forSAKtestingandvictimservicesfrom1980(thedateofoldestkitfound
inthecensus)to2009(theendofprojectsscope).IntheAppendixB:ProjectMethodology,wedescribe
thedatacollectionandanalysisproceduresforthisworkindetail,butbriefly,wedrewuponthreedata
sources:ethnographicobservations,stakeholderinterviews,andarchivalrecords.Inthecollaborative
teammeetings,stakeholdersoftendiscussedresourceconstraintsandchallenges,whichwere
documentedinourethnographicfieldnotes.Buildingonthosedata,weconductedindividualinterviews
withstakeholdersacrossallorganizations,interviewingbothcurrentandformeremployeesinthese
organizations,andinterviewingindividualsatalllevels/roleswithineachorganization.Weasked
stakeholdersaboutdailyoperations/servicesprovided;staffinglevelsovertime;andreportingstructure,
trainingandsupervision.36Becausememorygapsaretobeexpectedinretrospectivedatacollection
(Bradburn,Sudman,&Wansink,2004;Sudman,Bradburn,&Schwarz,1996),wealsoconductedan
extensivereviewofarchivalrecordswithineachorganization.Theinformationobtainedthroughthese
threemethods(observations,interviews,andarchivalrecords)wascrosscheckedandtriangulatedprior
toanalysistoensurethattheresultswereaccurateandcredible(seeAppendixB:ProjectMethodology
fordetailsontriangulationassessmentsandtrustworthinessofthedata).
Aswewerecollectingthesedataaboutorganizationalresources,weoftenwondered:iswhat
wereseeinginDetroittypical?AretheresourcelevelsinDetroitsimilartootherurbanareaswith
similarlyhighcrimerates?Toaddressthesequestions,wealsocollecteddescriptivedataincomparable
36
Wealsoaskedaboutinterorganizationalcommunication,SAKtestingpoliciesandpractices,anddecisionmakinginsexual
assaultinvestigationsandSAKsubmissions;thedataresultingfromthosequestionswerethefocusofthesubsequentsectionin
thischapter(UnderlyingFactors:FrontLinePracticesandInterOrganizationalCommunication).
70
citiestoprovidesomecontextastowhetherthefindingsinDetroitweresimilarelsewhere.37However,
asonememberofthecollaborativenotedabove,theresnocityintheU.S.thatslikeDetroit,and
indeed,itwasachallengetofindcitiesintheU.S.thathavesimilarpopulations,racial/ethnicminority
distributions,andcrimeratesasDetroit.Basedondatafromthe2000Censusandthe2000FBIUniform
CrimeReport(asthiswastheerainwhichSAKswereaccumulatingrapidly),fourcitieswereselectedas
comparisonsforthishistoricalanalysis,thoughtherewasnoonecitythatwassimilaralongallthree
factors(overallpopulation,racial/ethniccomposition,crimerate)(seeTable3.1,followingpages):38
Philadelphia,PA:ThoughPhiladelphiaspopulation(1.5million)exceedsDetroits(951,270),
theircrimeratesaresimilar:UCRmodifiedcrimeindextotalsare100,581and97,776,
respectively;adjustedper100,000people,therateinPhiladelphiaislowerthough(6,751vs.
9,848).AsubstantialproportionofPhiladelphiasresidentsareAfricanAmerican(45%),though
nottothesameextentasDetroit(82%).
Dallas,TX:DallasisalsolargerthanDetroit(1,188,580vs.951,270)anddoesnothavea
comparableracial/ethniccomposition(25.9%ofDallasresidentsareAfricanAmericanvs.82%in
Detroit),butlikePhiladelphia,ithasacrimeratesimilartoDetroits:UCRmodifiedcrimeindex
totalsare106,460and97,776,respectively;adjustedper100,000people,theoverallcrimerate
inDallasremainssimilartoDetroits(9,382vs.9,848).
37
Thescopeofthisactionresearchproject,asstipulatedintheRFP,didnotincludecrossjurisdictioncomparisons,butwe
collectedasmuchcomparabledatainothercitiesaswasfeasible,thoughwenotethatwedonothaveperfectlyparalleldata
acrossallorganizations,acrossallcities.SomedatawerequitechallengingtotrackdowninDetroitanditwasnotpracticalfor
othercitiestodevotetime/energytocollectparalleldata,thoughweweresuccessfulinsecuringsomecomparativedataforall
organizations,acrossallcities.
38
WedidnotselectHouston,TXasacomparisoncityforthisanalysis,eventhoughitwastheothersitefundedintheNIJSAK
ActionResearchProjectsolicitation,becauseitisquitedifferentfromDetroitwithrespecttoourkeycomparativefactors
(population,racial/ethniccomposition,crimerate,andresources).WereferthereadertotheFinalReportfromtheHouston
siteformoreinformationaboutthelocalcontextandresourcesofthatcity.
71
Baltimore,MD:BaltimoreissmallerthanDetroit(651,154vs.951,270),butitsracial/ethnic
compositionismoresimilarinthatbothcitiesarepredominatelyAfricanAmerican(64%and
82%respectively).ThecrimerateinBaltimoresimilartoDetroits(whenadjustedfor
populationsize):UCRmodifiedcrimeindextotalsare65,886and97,776,respectively;adjusted
per100,000people,theratesare10,168and9,848.
NewOrleans,LA:NewOrleansisalsopredominatelyAfricanAmerican(67%).Itis
approximatelyhalfthesizeofDetroit(484,674vs.951,270,respectively)anditscrimerateis
alsosubstantiallylower:UCRmodifiedcrimeindextotalsare34,208and97,776,respectively;
adjustedper100,000people,therateinNewOrleansbecomessomewhatmorecomparableto
Detroits(7,216and10,055).
OnceweunderstoodthehistoricalcontextofeachofthemainDetroitsexualassault
organizationsandhowtheycomparedtothoseincomparablecitieswereturnedtothecensusdata
toexaminewhetherthekeyturningpoints/changesweidentified(throughthequalitativedata)were
associatedwithSAKsubmissionratesovertime.Usingmultilevellongitudinalquantitativemodeling,
weexaminedwhetherkeyhistoricaleventsidentifiedinthisanalysiswereassociatedwiththeobserved
ratesofSAKsubmissionsovertime.
72

TABLE3.1ComparableCities:Population,Race/Ethnicity,andUCRCrimeRate
2000CensusEthnicMinorityPopulation*
City
2000Census
Population*
White
Black/African
American
Asian
Other
Race
Hispanic/
Latino
2000UCRCrimeStatistics**
Crime
Index
Total
Modified
Crime
Index
Total***
Forcible
Rape
2000UCRCrimeStatistics
(per100,000)**
Crime
Index
Total
Modified
Crime
Index
Total***
Forcible
Rape
116,599 775,772
951,270 (12.3%) (81.6%)
9,268
(1.0%)
24,199
(2.5%)
47,167
(5.0%)
95,761
97,776
811
9,848
10,055
83
Philadelphia
683,267 655,824
1,517,550 (45.0%) (43.2%)
67,654
(4.5%)
72,429
(4.8%)
128,928
(8.5%)
98,000
100,581
1,021
6,751
6,929
70
Dallas
604,209 307,957
1,188,580 (50.8%) (25.9%)
32,118
(2.7%)
204,833 422,587
(17.2%) (32.6%)
105,050
106,460
633
9,382
9,506
56
Baltimore
205,982 418,951
651,154 (31.6%) (64.3%)
9,985
(1.5%)
4,363
(0.7%)
11,061
(1.7%)
65,886
66,397
366
10,168
10,247
56
NewOrleans
135,956 325,947
484,674 (28.1%) (67.3%)
10,972
(2.3%)
4,498
(0.9%)
14,826
(3.1%)
34,001
34,208
227
7,216
7,260
48
Detroit
*USCensusBureau(2000)
**FBIUniformCrimeReport(UCR)(2000)
***TheModifiedCrimeIndextotalisthesumoftheCrimeIndexoffensesincludingarson.
73
History of DNA Testing and CODIS Access 1980 2009

ContextualFindings.Arapekitcontainsmultiplesamplesofbiologicalevidencecollectedfromthe
victimsbody,mosttypicallyoral,vaginal,andanalswabs.39Overthepastthirtyyears,therehavebeen
revolutionarychangesinhowthisevidencecanbetestedandusedbythecriminaljusticesystem.Itis
beyondthescopeofthisreporttoreviewthetechnicalevolutionofDNAtesting(seeButler,2005,2010,
2012forreviews),butforsummarypurposes,itmaybehelpfultodenotekeyhistoricaldevelopments:
PreDNA:PriortothedevelopmentofDNAtesting,thebiologicalsamplesinrapekitswere
testedusingdiscriminatingproteinmarkers,suchasABObloodtyping;however,suchmethods
havelowdiscriminatorypowerandproteinscandegradequickly,sotheutilityofthisinformation
inaninvestigativecontextwaslimited.
DNATesting,CODISNotYetDeveloped:DNAtestingisamultistageprocessthatbeginswitha
serologyscreeningofthesamplesinthekittodetermineiftheycontainbiologicalevidence(e.g.,
semen,saliva,blood).40Ifthesamplesinthekitdocontainbodilyfluids,thenthenextstepsinthe
testingprocessinvolveextractingtheDNAfromthesamples,quantifyingoftheamountofDNA
extracted,separatingtheDNA,andfinally,analyzingandinterpretingtheresults.Overtheyears,
differentmethodshavebeendevelopedfortheextraction,amplification,separation,andanalysis
39
Hairsamples(headhairandpubichair)arealsousuallycollectedinthemedicalforensicexam,butarerarelyanalyzedby
forensiclaboratories(Peterson,Sommers,Baskin,&Johnson,2010).Olderkits(i.e.,thosecollectedinthe1980sand1990s)may
alsoincludeabloodsamplefromthevictim.Newerkits(i.e.,thosecollectedinthe2000s)maycontainadditionalswabstaken
fromotherbodyparts(e.g.,breasts,neck)thatweretouched/harmedintheassault.Despitethesevariations(overtimeand
jurisdictionalcollectionpolicies),theoral,vaginal,andanalswabhavebeenandcontinuetobethemostprobativeevidencein
theSAKandthereforearetheprimaryfocusintesting.
40
Inthepastdecade,newermethodshavebeendevelopedthatallowsforensicscientiststoskiptraditionalserologyscreeningin
favorofafasterscreeningmethodsthatdetermineifthereisanymaleDNAinthesamples(Yscreeningmethods);ifso,thekit
willproceedtothenextstagesintheprocessforDNAtesting.SeeChapter4:Developing&EvaluatingaSAKTestingPlanfor
extendeddiscussionofthestagesofDNAtesting.
74
steps.41However,aDNAprofile(howeverextracted)haslimitedutilityinandofitself.AsButler
(2010)noted,aDNAprofilebyitselfisfairlyuselessbecauseithasnocontext.DNAanalysisalways
requiresthatacomparisonbemadebetweentwosamples(p.9).Therefore,aDNAprofile
extractedfromarapekithadthepotentialtobehelpfultoinvestigators,butwithoutreference
samplesforcomparison,thelikelihoodofidentifyinganoffenderwaslow.
DNATestingwithCODIS:Anationaldatabase,CODIS(CombinedDNAIndexSystem),was
authorizedbythefederalDNAIdentificationActof1994andprovidedtwoindexingsystems:the
offenderindex,containingtheDNAprofilesofconvictedoffenders,andtheforensicindexsystem,
containingDNAprofilescollectedfromcrimescenes.WiththeadventofCODIS,aDNAprofile
extractedfromarapekitcouldbeloadedintothedatabase(provideditmetstate/federal
requirementsforupload)andthenthedatabasecouldbesearchedtodetermineiftherewasa
matchbetweenthenewprofileandanexistingsample(eitherintheoffenderindexorforensic
index).AhitmeantthattherapekitsamplematchedaDNAprofilealreadyinCODIS,thereby
identifyingthepossibleoffender.CODISlaunchedin1998andlaboratorieshadtocompleteFBI
auditingrequirementsand/orseekaccreditationfromindependentforensicscienceorganizations
toaccessCODIS(i.e.,toloadprofilesandcompareprofilestoexistingreferencesamples).In2004,
federalregulationsrequiredthatlaboratorieshaveaccreditationfromspecifiedcredentialing
organizations(tobecompletedbyOctober,2006).
41
IntheRFLP(RestrictionFragmentLengthPolymorphism)technique,thefirstmethodtogainwidespreadadoptioninforensic
settings,aDNAsampleisbrokenintopiecesbyrestrictionenzymes,resultinginrestrictionfragmentsthatareseparatedaccording
totheirlengthtoevaluatedifferentpatternsacrosspersons.Intheearlytomid1990s,anewamplificationtechnique,PCR
(PolymeraseChainReaction)methods,offeredfasteranalysiswithsmallersamples;atthattimeforensicscientistsalsobeganusing
STR(ShortTandemRepeat)methods,inwhichshorter,specificunitsofDNAwerecopiedandexamined.
75
WiththisnationalcontextregardingDNAtestingandCODISestablished,wethenexamined
Detroitspracticesforrapekittestingovertheyears.From1980to1993,SAKssubmittedtothepolice
departmentcrimelabwereanalyzedwithABObloodtypingmethods(aswasnationalpracticeatthetime).
From1994to1997,thecrimelabwasdoinglimitedscaleDNAtesting,butgiventhatCODISdidnotyet
exist,DNAtestingwasnotroutine.WhenCODIScameonlinein1998,thepolicedepartmentcrimelabdid
nothaveaccessforfouryears(19982001),thoughtheycouldappealdirectlytotheFBItosearchDNA
samplesincasesofsuspectedserialrapists.From2002to2005,thepolicedepartmentcrimelabhad
securedaccesstoCODISbypassingFBIauditingrequirements,thoughtheywereintheprocessofseeking
accreditation(per2004federalrequirements).Stakeholdersnotedthattheaccreditationprocessrequired
substantialadministrativetime,whichreducedthenumberofhoursthatwereavailableforbenchwork
(i.e.,testingkits).Thelabwasaccreditedin2006,whichremainedineffectuntilitwasclosedin2008.
QuantitativeModeling:DNA/CODISHistory.TheavailabilityofDNAtestingandCODISarecritical
contextualfactorsthatmayhaveaffectedratesofSAKsubmissionsovertime.Specifically,inthepre
DNA/preCODISera,fewerkitswouldbeexpectedtobesubmitted,giventhelimitedutilityoftestingat
thattime.AfterCODIS,SAKsubmissionswouldbeexpectedtoincrease,givenhowhelpfulDNAtesting
couldbetotheinvestigationalprocess.Toexploretheseideas,wereturnedtothecensusdatatoexamine
howmanySAKsweresubmittedfortestingineachofthesehistoricaleras.Forthesestatisticalmodels,we
workedwithasubsampleof10,817SAKs(fromthe11,219censuscount),removingcasesfromThe400
Project,42asthedatacollectionmethodsinthatproject(whichcamebeforetheDetroitSAKARP)didnot
capturedatafieldsinthewayweneededfortheseparticularanalyses.Thecasesthatwerenotincludedin
theseanalysesrepresentedapproximately4%ofthetotalcensus(andhadbeenrandomlyselectedinthe
firstplace),soitisunlikelythattheirexclusionwouldmarkedlyaffectthefindingsandconclusions.
42
Twoadditionalcaseswereremovedduetomissingdata(i.e.,402casestotalwereremovedfortheseanalyses).
76
Table3.2(below)showsthattheratesofSAKsubmissionwerelowestinthepreDNAandpre
CODISeras(14%and13%,respectively).AfterthedevelopmentofCODIS,submissionrateswerehigher,
exceptfortheperiodoftimeinwhichthepolicedepartmentcrimelaboratorywasseekingaccreditation
(20022005).However,oncetheaccreditationprocesswascomplete,thesubmissionsratesincreased,
whereby35%oftheSAKsinpolicepropertycollectedfrom2006to2009hadbeensubmittedfortesting.
However,acrossalleras,mostSAKswerenotsubmittedfortesting(rangeof65%to86%).
TABLE3.2DistributionofSAKsOverTime,ByDNA/CODISEra
DNAEra
PreDNA
NumberofKits
inPolice
Property
PercentageofKits
SubmittedtotheCrime
LabforTesting
PercentageofKits
Unsubmittedtothe
CrimeLabforTesting
2,261
14%
86%
2,751
13%
87%
2,026
33%
67%
2,070
24%
76%
1,709
35%
65%
(19801993)
DNATesting,CODISNotYetDeveloped
(19941997)
DNATesting,CODISExists,CrimeLabDidNot
HaveAccesstoCODIS
(19982001)
DNATesting,CODISExists,CrimeLabHad
AccesstoCODIS,butSeekingAccreditation
(20022005)
DNATesting,CODISExists,CrimeLab
Accredited
(20062009)
Giventhatthesedescriptiveanalysessuggestedthatsubmissionratesvariedovertime,weused
mixedeffectslogisticregression(withrandomeffectsofkitsnestedwithinyears)(seeAppendixB:Project
Methodologyfortechnicaldetailsregardingtheseanalyses)tomodeltheprobabilityofSAKsubmissionsas
afunctionofwhetherkitwascollectedduringthe:
77
PreDNAEra(19801993)
DNATesting,CODISNotYetDevelopedEra(19941997);
DNATestingWithCODISButTheCrimeLabDidNotHaveAccessToCODISEra(19982001)
DNATestingWithCODISButTheCrimeLabWasSeekingAccreditationEra(20022005)
DNATestingWithCODISAccreditationCompleteEra(20062009).
AsshowninTable3.3(nextpage),DNA/CODISerawasasignificantpredictorofSAKsubmissions.
SAKsfromthepreDNAera(19801994)andtheCODISNotYetDevelopedera(19941997)were
significantlylesslikelythankitsfromothererastobesubmitted.Inotherwords,whenDNAtestingand
CODISwerenotdeveloped,SAKsubmissionswere,asexpected,significantlylower.Alsoasexpected,kits
collectedintheeraofDNATestingwithCODISwithlabaccreditation(20062009)hadsignificantlyhigher
ratesofsubmission.However,intheyearsinwhichthelabwasseekingaccreditation(20022005),SAK
submissionsdippedsignificantlyinfact,submissionswerehigherwhenthelabdidnotyethaveCODIS
access(19982001).Thereasonswhysubmissionswerelowerduringtheaccreditationprocessmeritsmore
exploration,andwepursuedthisissuewithadditionaldatacollection,whichwillbepresentedin
subsequentsectionsofthischapter(seeThePoliceandtheCrimeLab:IntraOrganizationalPractices&
Communication).Overall,thesequantitativeresultssupportthequalitativedataregardingtheimportance
ofplacingthisprobleminhistoricalcontextratesofsubmissiondidinfactvaryasafunctionofwhatwas
availableregardingDNAtechnologyandaccesstoCODIS.WhenDNAtestingresourceswerefullyavailable
toDetroitstakeholders,theratesofsubmissionweresignificantlyhigher.
78
TABLE3.3MultiLevelLogisticRegressionResults,PredictingSAKSubmissionbyDNA/CODISEra
Block
Log
Odds
Intercept
1.157
DNAErasinwhichSAKwascollected
Odds
Ratio
0.314
Confidence
Interval
0.242 0.409 <.001
PreDNAEra(19801993)
0.764
0.466
0.325 0.667
<.001
DNAbutNoCODIS(19941997)
0.751
0.472
0.324 0.686
<.001
0.440
1.553
1.0742.246
0.021
1.0362.412
0.034
DNA&CODISbutCrimeLabDoes Not
HaveCODISAccess(19982001)
DNA&CODISbutCrimeLabSeeking
CODISAccreditation(20022005)
DNA&CODISandCrimeLabAccredited
(20062009)
Log
Likelihood
#
Parameters
15351.13
15344.86
LRchi
square
12.54
ReferenceCategory
0.437
1.548
N=10,817kits,nestedwithin28years;analysiswasconductedusingmixedeffectslogisticregression,withrandomeffectsofkitsnestedwithinyears
0.028
79
Police Department Resources & Policies 1980-2009

ContextualFindings.Perhapsthemostimportantresourceforanypolicedepartmentisits
personnelleadership,staffing,andotherprofessionalsupports(e.g.,training).Asinmanycities,the
DetroitPoliceChiefisappointedbythecitysMayorandservesatthepleasureoftheMayor.Throughout
thethirtyyearsinthisanalysis,therewerefivedifferentMayors,whichisconsistentwiththeratesof
mayoralturnoverinthefourcomparablecitiesexamined(Dallassomewhathigheratseven)(seeTable3.4,
nextpage).InDetroit,thosefiveMayorsappointed10differentPoliceChiefs,who,onaverage,served2.70
yearsinoffice,whichissubstantiallylowerthannationalaverages.Theaveragetenureforametropolice
chiefinthe1990s(whichwaswhenmanyoftheDetroitSAKswereaccumulating)was3.50to4.50years
(Peak&Glensor,1996),and4.93yearsinjurisdictionswithover500,000residents(PoliceExecutive
ResearchForum,1997;Rainquet&Dodge,2001).However,Detroitsturnoverisnotdissimilartoother
urbancitieswithsimilarcrimeratesand/orracialcompositions,thoughitisatthehighendofthe
comparison(seeTable3.4).Comparablecitieshad710PoliceChiefsinthatsameperiodoftime,with
averagetenuresrangingfrom2.60years(Baltimore,MD)to4.00years(Dallas,TX).
TakingacloserlookatPoliceChiefturnoverfromthe1990sto2009asthesewerethekeyyears
inwhichunsubmittedSAKswereaccumulatingweseeaslightlydifferentpatternemerge.InPhiladelphia
andDallastwocitiesthat,generallyspeaking,arebetterresourcedthanDetroittherewerefourPolice
Chiefs,comparedtonineinDetroit.InBaltimoreandNewOrleanscitiesmoresimilartoDetroitinracial
compositionandsimilarlylowerresourcedoverallthereweresevenandsixchiefs(respectively),again,
comparedtonineinDetroit.TheaveragetenureofaDetroitPoliceChieffrom1991to2009was2.06
years,whichisthelowestamongthefourcomparablecities,particularlysorelativetoPhiladelphiaand
Dallas(4.25and4.20,respectively).
80
TABLE3.4ComparableCities:Mayoral&PoliceDepartmentLeadership
City
MayoralTurnover
19802009
Detroit
5Mayors
7.40YearsAvg.
Tenure
Philadelphia
5Mayors
6.00YearsAvg.
Tenure
Dallas
7Mayors
3.80YearsAvg.
Tenure
Baltimore
5Mayors
7.80YearsAvg.
Tenure
NewOrleans
4Mayors
7.75YearsAvg.
Tenure
PoliceChiefTurnover PoliceChiefTurnover
19802009
Since1990*
10Chiefs
2.70YearsAvg.
Tenure
9Chiefs
3.20YearsAvg.
Tenure
7Chiefs
4.00YearsAvg.
Tenure
10Chiefs
2.60YearsAvg.
Tenure
8Chiefs
3.50YearsAvg.
Tenure
9Chiefs
2.06YearsAvg.
Tenure
4Chiefs
4.25YearsAvg.
Tenure
4Chiefs
4.20YearsAvg.
Tenure
7Chiefs
2.60YearsAvg.
Tenure
6Chiefs
2.90YearsAvg.
Tenure
*Computationsforthenumberofchiefs/yearsoftenuresince1990variedslightlytoaccommodatenaturalvariationsinturnovers
on/about1990ineachcity.InDetroit,thedatesusedwere19912000;Philadelphia,19922009;Dallas,19912009;Baltimore,
19892009;NewOrleans,19912009.
InterviewswithpolicestakeholdersinDetroitsuggestthattheinstabilityinleadershiphasalways
beenproblematic;asonepolicedepartmentmembernoted,Theonlythingconstantaroundhereis
change,andanothercommented,wevehadsomanyreorganizationsthatyoujustcantkeepup.43
Fortheworkofthesexcrimeunitspecifically,theleadershiptransitionsweredifficultbecauseeachnew
Chiefusuallymeanttherewouldbeanewunitsupervisoraswell;inotherwords,theleadershipchangesat
thetoptrickleddowntotheindividualtaskedwithsupervisingthesexcrimesunit.44Whenaskedwhat
43
Inthe30months(2.5years)ofthisActionResearchProject,therewerethree(additional)newPoliceChiefs;thesethreearenot
countedinTable3.4becausethescopeofthehistoricalanalysisendedin2009(however,theinstabilitycontinuedthereafter).
44
EvenwithinaparticularChiefstenure,itwasnotuncommonthattherewouldbemultiplechangesinthesupervisionofthesex
crimeunit.Forexample,from2005to2009,therewerethreeChiefsandfivedifferentsupervisingLieutenantsoftheunit.
81
thosechangesdidforcontinuityofpractice,oneintervieweepantomimedpullingadraweroutofdesk,
dumpingoutallthecontents,thenreplacingthedrawer:
Iveseenabunchofchiefscomeinand..[whenthey]camein,theytookeverydraweroutofthe
cabinet,dumpeditout,puteverythingbackin,butjustindifferentdrawers,sotheymoved
everybody,theyswitcheditallup.
Anothermemberofthepolicedepartmentdescribedtheleadershiptransitionsassuch:Westartover,
everysingletime.WithrespecttohowthisinstabilityinleadershipmayhaveaffectedtheissueofSAKs
submissionspecifically,stakeholdersnotedthattheconstantchangemadeithardtostartandsustainnew
initiativesfortheunit:somethinggood,positivefortheunit[getsstarted],andthentheyregone.The
constantturnoveralsomadeitdifficulttoidentifyproblemsandlowperforminginvestigators:
Peoplecanhidebadworkforonlysolong...andthatsabouthowlonganysupervisorwas
around,soyouassoontheydfigureitout,theyremovedon.
Aseniorpoliceofficialsummedupthenegativeimpactoftheseperpetualtransitionsinleadership:Those
peoplethathavecontrolofthedaytodayoperation,whenyoulosethosepeopleyoulosealot.
Inadditiontothechallengeofconstantsupervisoryturnover,thenumberoffrontlineofficersand
detectivesinthesexcrimeunitsdwindledovertime.IntheyearsofthepreDNA/preCODISera,thepolice
departmenthad,onaverage,2030sexcrimeunitinvestigatorsandstaff(e.g.,adesignatedproperty
officer),andthatstaffinglevelcontinueduntilearly2000s.In2002,thestaffinglevelsinsexcrimeswere
cutapproximately50%(downto12investigators/staff).In2008,thesexcrimesunithadanother50%cutin
staffing,downto68investigators/staff.Thesestaffingcutsfundamentallychangedthewaytheunit
operates,includingasonestakeholderexplainedbelow,nolongerhaving24hourservice:
82
[Afterthebudgetcuts]weoperatewithtwoshiftsandweusedtobe24houroperational...we
workfrom8a.m.to4p.m.fordaysand4p.m.to12a.m.forafternoons...from12a.m.to8a.m.
itsaclosedshiftandweoperateonarecallbasis...[yougeta]callinthemiddleofthenightifwe
hadasituationgoingon...and[thenyou]callamembertogetthemuptorespondbacktothe
scene...thatreallyaffectsourabilitytoproperlyinvestigatesomeofthesesexualassaults.
Bothfrontlinepolicestaffandcommandstaffwerewellawarethatthesestaffingcutsweregoingtohave
alongtermnegativeimpactonsexualassaultinvestigations.Asoneseniorpoliceofficialcommented:
Isawthiscoming(meaningproblemswithSAKsandsexualassaultinvestigations)...whenwe
startedlosingmanpower,Icouldseethatcornersweregoingtobecutbecausetherewasnoway
youcouldjustkeepupwiththedemandofcasestheyweregettingandyouwantresults.
Whenaskedabouttrainingforsexcrimesunitstaff(e.g.,trainingonDNAanalysisanditsutilityto
sexualassaultcases,trainingoneffectiveinvestigativetechniquesforsexualassaultcases,trainingon
workingwithspecialpopulations),policepersonnelnotedthatprofessionalinserviceswereinfrequentand
alotoftheireducationcamefromwhateverthelabfolkssaidtouswhenwedroppedoffkits.When
askedabouttrainingandsupportforburnoutandvicarioustraumafromconstantexposuretotraumaand
violence,policepersonnelinvariablyshruggedandstatedthatwassimplynotpartofthedepartments
culture.Asonestakeholdernoted,ThisisDetroit.Everyonehastoseeawfulthingseveryday.
TurningtothepolicedepartmentspoliciesregardingSAKsubmissionsoverthisthirtyyear
historicalanalysis,itisunclearwhetherinfacttherewerewrittenpoliciestheresearchteamwasnotable
toverifytheexistenceofsuchdocumentation,thoughpoliceofficialsstatedthattherewerewritten
guidelines.Basedondatafromstakeholderinterviews,itappearsthatfromthe1980sto1999,thepractice
83
wastosubmitonlythosekitsassociatedwithknownsuspectSAKs.45Inotherwords,ifpolicehada
possiblesuspectidentified(e.g.,knownoffenderassaults,suchasacquaintancerapes,intimatepartner
rapes;orsufficientinformationaboutunknownoffenderssuchthattheycouldbeidentifiedbypolice),
thenSAKtestingmighthavebeenconducted;withouttheexistenceofCODISoraccesstoCODIS,aprofile
ofanunknownoffenderhadlimitedinvestigationalutility.In2000,itappearsthatthepolicedepartment
changedtheirpractice,mostlikelyduetotheemergenceofCODIS(thoughthelaboratorydidnotyethave
accesstoCODIS),andfromthatpointforwardbothknownsuspectandnosuspectkitswere
consideredfortesting.Again,thecensusdataclearlyshowthatnotallSAKsweresubmittedfortesting,
butthesedatahelpcontextualizewhysomecaseswerenotsubmitted:dependingonthevictimoffender
relationshipandtheyearinwhichtheSAKwascollected,departmentpracticesweresuchthatparticular
kindsofcaseswouldbeunlikelytobesubmittedfortesting.
QuantitativeModeling:PoliceResourcesandPoliciesOverTime.Oneofthemoststrikingfindings
fromthequalitativedatawasthehighrateofturnoverinpoliceleadership;however,itwassofrequent
(approximatelyeverytwoyears)thatattemptingtomodelthesechangeswithinthemultilevellogistic
regressionanalyseswasnotpracticalbecauseleadershiptransitionswereessentiallyaconstant.However,
therewereotherkeychangesinresourcesandpolicyovertheyearsthatmighthaveaffectedSAK
submissionsandcouldbequantitativelyevaluated.Thesexcrimesunitfacedtwo50%cutsintheirstaffing
levels(onein2002,theotherin2008).Followingsuchstaffingcuts,itispossiblethatratesofSAK
submissionwoulddropsignificantly,asthesexcrimeunitstruggledtorespondtoreportedcases.
However,thishypothesiswasnotsupportedinthedata.AftercontrollingforDNA/CODISera(whichprior
analysesdemonstratedwasasignificantfactoraffectingSAKsubmissionrates),thesechangesinstaffing
levelswerenotsignificantlyassociatedwithSAKsubmissionrates:OR=.895(95%CI=.4351.839),p=
45
ItisclearfromthecensusdatathatindeednotallknownsuspectSAKswereinfactsubmittedfortesting;thepointhereisthat
onlyknownsuspectcaseswereconsideredforsubmission.
84
.753.Thus,eventhoughthequalitativedatahighlightedhowdifficultthesecutsweretothefunctioningof
thesexcrimesunit,theywerenotdirectlyrelatedtoSAKsubmissionrates.Inotherwords,theseresults
suggestthattheproblemofunsubmittedSAKsisnotsimplyapersonpowerissue,andthatotherissues
withinapolicedepartmentmustbeconsidered(seePoliceDepartmentFrontLineDecisionMaking&
Practicesforadditionalfindingsonthisissue).
WealsoexaminedwhetherthepolicedepartmentspoliciesregardingSAKtestinghada
demonstrableaffectonsubmissions.In2000,policeofficialsstatedthattheirpolicychangedsothatboth
knownoffenderandunknownoffenderSAKscouldbesubmittedfortesting,whichwouldbeexpectedto
resultinasignificantincreaseinthenumberofSAKsubmitted(giventhatessentiallyallSAKswerenow
eligible).However,aftercontrollingforDNA/CODISera,thischangeinpracticewasnotassociatedwitha
significantincreaseinSAKsubmissions:OR=1.105(95%CI=.5122.382),p=.792.Asnotedpreviously,it
isunclearwhetherthispolicywascodifiedinwriting,sothesenonsignificantresultsarenotentirely
surprising.Indeed,theyunderscoretheimportanceofdelvingdeeperintopolicepracticesregardingSAK
submissionasthesemacrolevelchangesinresourcesandpolicydidnotappeartohaveadirecteffecton
submissionrates.
Crime Laboratory Resources 1980-2009

ContextualFindings.AkeyresourceforDNAtestingis,ofcourse,personnelandfundingavailable
toconductsuchtesting.Withrespecttopersonnel,laboratoriesmusthaveaminimumoftwoscientists
(forqualityassurancepurposes),asstipulatedbytheFBIandprofessionalaccreditingorganizations(see
Butler,2005,2010,2012).InDetroit,thelabtypicallyhadtwotothreeDNAscientists:onepositionwas
primarilyadministrative,dedicatedtotheworkofsecuringaccreditationforthelab,so,onaverage,only
twoscientistswereavailableforactualtesting.Forcityof900,000+residents,withahighUCRcrimerate,
thesestaffinglevelscertainlysoundlow,andindeed,basedondataincomparablecities,itisobjectively
85
low.Table3.5(below)presentsthestaffinglevelsinthecomparablecities:threecities(Philadelphia,
Dallas,andBaltimore)hadtwotothreetimesasmanyDNAscientists(NewOrleanshadsimilarstaffing
levelsasDetroit,butapopulationapproximatelyhalfthatofDetroit).Putanotherway,Dallashadthree
laboratories;DetroithadthreeDNAscientists.Wealsointerviewednationallevelstakeholdersinthe
forensicscienceprofessiontoassesswhatthoseinthefieldconsidertobetypicalandreasonable
staffinglevelsforlaboratoriesservinghighcrimeurbancities.Verylargecities(populationsof2million
andup)canhave40+DNAscientists,whilesmallerurbanareas(500,000to1million)varytremendously,
ofteninthe610range.Detroitsstaffinglevelsof23scientistsisconsiderablybelowthistypicalrangeof
610.Takentogether,thesedatasuggestthattheDetroitcrimelaboratorywassorelyunderresourced
relativetoitsserviceareapopulationandcrimerate.Crimelabofficialswereawarethattherewasmore
demandthancapacity,andinthemid2000stheywereabletoleveragesomeadditionalresources.Inlate
2006,thepolicecrimelabwasabletohirefiveadditionalscientists,butafterthelengthyperiodoftraining
newstaff,theyneverrealizedthefruitsoftheirlaborbecausethecrimelabclosedin2008.
TABLE3.5ComparableCities:NumberofDNAScientists
AverageNumberofDNAScientistsPerYear
19902009
City
Detroit
23
Philadelphia
Dallas*
68
Baltimore
NewOrleans
*TherewerethreecrimelabsthatprocessedcasesforDallas(andothernearbycities);numbersreportedhereare
rangesacrossthethreelabs.
86
WithrespecttoothernecessaryresourcesforDNAtesting,thepolicecrimelabdidnothavefiscal
supportfromcity,county,orstatesourcestooutsourcetestingtoprivatelaboratories.However,inlate
2004,thepolicecrimelabreceivedafederalDepartmentofJusticeDNABacklogReductionGrant,which
allowedthemtostartoutsourcingsometestingin2005.Whenaskedabouttrainingandsupportofcrime
labpersonnel,stakeholdersnotedthattheyregularlyparticipatedinscientifictrainingonDNAanalysis,as
wasnecessarytopassFBIauditsandobtainaccreditation;trainingspecifictoSAKtestingwasnever
provided,norwasanytrainingonburnoutorvicarioustraumafromworkingonthesekindsofcasesover
prolongedperiodsoftime.
QuantitativeModeling:CrimeLabResourcesOverTime.Thecrimelabdidnothavemany
changesinitsresourcesovertime,astheystruggledwithchronicunderstaffing/underfundingthroughout
thisthirtyyearanalysis.However,onekeychangethatmighthaveaffectedSAKsubmissionsrateswasthe
receiptoffederalDOJDNABacklogReductionfunds,whichallowedforoutsourcingofSAKsstartingin2005
(i.e.,ratesofsubmissionmighthaveincreased).AftercontrollingforDNA/CODISera,thereceiptoffederal
fundshadnosignificanteffectonSAKsubmissionrates:OR=.731(95%CI=.3721.439),p=.350.This
findingwassurprising,giventhatstakeholdersacrossmultipleorganizationsemphasizedtheimportanceof
theDOJDNABacklogReductionfunds;theremaybestatisticalreasonsforthiseffect(seefootnote).46
Thesegrantfundsmayhavekepttherateofsubmissionsrelativelyconstant,preventingadropinrates
(hence,anonsignificanteffect).Giventhattheresearchteamwasunabletodetermineexactlyhowthose
fundswereused(i.e.,howmanykits,fromwhichyearsweresubmittedfortesting),itisalsopossiblethat
theeffectofthesemoniesishardtoidentifyiftheywereusedtotestkitsinatestsomehere,somehere,
somehereacrosstimedistributionpattern.Itisalsopossiblethatthefundswereusedtotestbiological
evidencefromother,nonSAKcases(ifso,therewouldnotbeanimpactonSAKsubmissionrates).
46
DOJfundsarecollinearwithDNA/CODISera,sotheadditionaleffectofthisvariablewillbedifficulttodistinguishfromtheeffect
ofitsrespectiveera.
87
Prosecutors Office Resources 1980-2009

ContextualFindings.Asinmanyotherjurisdictions,theProsecutorisanelectedpositioninWayne
County,andoverthisthirtyyearhistoricalreview,therewerefourelectedprosecutors,servingonaverage
eightyearsinoffice.Comparablecities(Philadelphia,Dallas,Baltimore,andNewOrleans)hadasimilar
numberofprosecutorsoverthatsameperiodoftime(threetofive)(seeTable3.6,below).Theaverage
tenureinWayneCountywassomewhatshorterthantheothercommunities,butonlybecausetherewas
oneProsecutorintheearly2000swhohadanatypicallyshorterterm(threeyears);historically,most
ProsecutorsinWayneCountyserve10+yearsinoffice(asinthecomparablecities).
TABLE3.6ComparableCities:ProsecutorLeadershipTurnover
ElectedProsecutor
Turnover19802009
City
Detroit
Philadelphia
Dallas
Baltimore
NewOrleans
ElectedProsecutor
TurnoverSince1990
4Prosecutors
3Prosecutors
8.00YearsAvg.Tenure
8.33YearsAvg.Tenure
3Prosecutors
1Prosecutor
10.70YearsAvg.Tenure
4Prosecutors
3Prosecutors
7.30YearsAvg.Tenure
3Prosecutors
2Prosecutors
9.30YearsAvg.Tenure
5Prosecutors
5Prosecutors*
7.70YearsAvg.Tenure
7.70YearsAvg.Tenure
**NewOrleanss19902009turnoverratesarethesameastheiroverallratesbecauseoneprosecutorservedfrom1973to2003.
88
Detroitareastakeholdersinvariablycharacterizedtheprosecutorsofficeasmorestablethanany
othercomponentofthelocalcriminaljusticesystem;asonememberofthecollaborativenoted:
[Theprosecutors]hadtoweatherallthosechangesinpolicedepartmentChief...thataffected
themtoo,alwayshavingtostartoverwithnewleadershipoverthere(thepolicedepartment)...
eventhoughthey(theprosecutors)didntchange,theywereforcedtodealwiththeinstabilityin
othergroups,whichnodoubtmadeitharderforthemtoprosecutecases.
Speakingtothisissueofhowthefrequentturnoverinpoliceadministrationaffectedtheprosecutors
office,oneprosecutornoted:
Youneverknewwhogottransferredwhere,whowasgoingtopickupthephone,whotocalltoget
stuffmovingiftheyddroppedtheball...andwhethertheyweregoingdowhattheyneededtodo
intime...weneverknew.
Eventhoughtheprosecutorsofficehadmorestabilityinleadershipthantheircounterpartsinthe
policedepartment,theyalsostruggledwithinsufficientstaffingtohandlecases.From2000to2009,the
prosecutorsofficehadapproximately92trialattorneysperyear,andinthatperiodoftime,theyissued,on
average,17,907felonycases(including,butnotlimitedto,sexualassaultcases).Inthisdecade,therewere
substantialfluctuationsintheirstaffinglevels(overall,trendingdownward),dippingtoalowof82trial
attorneysduetobudgetcutsfromCountygovernment.47Itisimportanttonotethattheprosecutorsoffice
servestheentirecounty,andthatthreeothercitiesinWayneCountywithpopulationsnear100,000also
havesubstantialcrimerates,perthe2000FBIUCRdata(Dearborn,Livonia,andWestland).However,
prosecutorsalsopointedoutthatregardlessoftheirstaffinglevels,theyhadanobligationtopublicsafety:
47
Weattemptedtocollectcomparabledataonprosecutionstaffinginthefourcomparisoncities,butgiventhatsuchrecordsare
oftendifficult/timeconsumingtoaccess,wewereabletoobtaindatafromonlyonecommunity:DallasCounty,TX,servingthecity
ofDallas.From20002009,theDallasCountyprosecutorsofficehadanaverageof99trialattorneys,comparedto92forWayne
County.However,thestaffinglevelsinDallasCountytrendedupwards(endingat108trialattorneysin2009),whereasWayne
Countytrendeddownwards(87trialattorneysin2009,whichtendedfurtherdownwardto<80atthetimeoftheARP).
89
Weveneverhadenoughprosecutors,especiallygiventhecrimeratesinDetroitandothercities
aroundhere...but,wehaveobligationtopublicsafety,sowehavetodealwithit.
Oneofthechallengesstemmingfromchronicunderstaffingwasthattheprosecutorsofficewas
onlyrecentlyabletoformaspecializedsexualassaultunit(2009asanunfundedvolunteerprogram;2011
asafundedprogram).Bycontrast,threeofthecomparablecitiesweexaminedhadtheresourcestoform
designatedsexualassaultunitsmuchsooner(Philadelphia,Dallas,andBaltimore;NewOrleanshasyetto
doso)(seeTable3.7,nextpage).Thoughtheexactstructureandoperationsofsuchunitsvary,theyare
typicallystaffedbyprosecutorswithexpertise/interestinworkingexclusively(ornearlyexclusively)on
sexualassaultcases(seeBeichner&Spohn,2005;Hemmens,Brody,&Spohn,2012;Lord&Rassel,2008).
Someprogramsfollowverticalprosecutionwherebyoneprosecutorworksacasefromstarttofinish,
ratherthanthemoretypicalhorizontalmethodsuchthatdifferentprosecutorshandleeachstepthrough
theprocess(e.g.,oneprosecutorreviewsthewarrantrequest,anotherconductsthepreliminary
examination,whileyetanotherconductsthetrialproceedings).
ThehorizontalmethodwasthenorminDetroit,until2009whentheprosecutorsofficeformeda
volunteersexualassaultunitwithverticalprosecution.Thoughtheydidnothavedesignatedfundingfor
theunituntil2011,threeAPAsvolunteeredtoberegularlyassignedtothesekindsofcases(inadditionto
theirregularduties).Asoneprosecutordescribed:
Wereallyneededit(aspecialunit)andthe(Elected)Prosecutorwasreallyopentotheidea...the
victimsneedspecialattention,theydeserveitandtheyneedit;andsecondly,inorderto
successfullyprosecutethecases,youhavetoreallyknowwhatyouredoingwiththesewhat
issuestolookfor,howtopresentittoajury...weneededexperiencedattorneys,experiencedwith
thesekindsofcases.
90
TABLE3.7ComparableCities:SpecializedProsecutionPrograms
City
YearinWhichSpecializedSexualAssaultProsecutionUnit
WasFounded
2009(unfundedvolunteerprogram)
Detroit
2011(fundedprogram)
Philadelphia
1977,thenreorganizedin1991
Dallas
~1993(atleasttwentyyearsago)
Baltimore
1978,thenreorganizedin2012
NewOrleans
(NONE)
Thoughhistoricallytheprosecutorsofficehadlimitedresourcesforinhousetraining,themost
recentElectedProsecutoractivelyencouragedAPAstotakeadvantageofeventshostedbythestate
prosecutingattorneysassociationandstatelevelviolenceagainstwomenagencies.AsoneAPAnoted,
Wealwaysknewwecould,andshould,seekouteveryopportunitytolearnmoreanditwouldbe
supportedby[theElectedProsecutor].
QuantitativeModeling:ProsecutorsOfficeResourcesOverTime.Stakeholdersacrossmultiple
organizationsnotedthatitwouldbeunlikelythatchangesintheElectedProsecutorwouldhavehada
directeffectonSAKsubmissionrates,unlessaspartofestablishinganewadministration,theElected
Prosecutorestablishedinteragencypoliciesregardingkittesting.Aswillbediscussedlaterinthisreport
(Chapter6:SummaryofFindings,Implications,andCommunityChanges),suchpoliciesnowexistinDetroit,
butfrom1980to2009,theproblemofunsubmittedSAKswasnotyetknown.Assuch,statisticalmodeling
ofprosecutortransitiondidnotseemfruitful(andindeed,therewerenoeffectsattheunivariatelevel).
Theonlyothersignificantpolicychangethatoccurredwithinthishistoricalanalysisthatmighthave
affectedSAKsubmissionsrateswastheestablishmentoftheverticalprosecutionsexualassaultunit;
however,thisunitbeganin2009(onavolunteerbasis),whichwasattheendofthetimeframestudied,
andthereforeitsimpactcouldnotbeevaluatedquantitatively.
91
Medical System Resources 1980-2009

ContextualFindings.InDetroit,victimsofsexualassaulttypicallyreceivedtheirmedicalcareand
forensicexamsatoneofthreelarge,hospitalemergencydepartments(EDs).48AllthreeEDshandledhigh
volumesofcrimerelatedinjuries,andasoneEDstaffmembernoted:withallthegunshots,beatings,
drugoverdoses,sexualassaultvictimswerelowpriority...wehadtotakerealemergenciesfirst.
Althoughspecializedsexualassaultnurseexaminer(SANE)/sexualassaultforensicexaminer(SAFE)
programswerefoundedinmanyjurisdictionsinthe1990sandearly2000s(seeCampbell,Patterson,&
Lichty,2005andDepartmentofJustice,2013forreviews),DetroitdidnothaveaSANEprogramuntil2006,
twelveyearsafterBaltimoreestablishedthefirstofitstwoprograms,andsixyearsafterNewOrleans
foundedtheirs.However,PhiladelphiaandDallasestablishedtheirSANEprogramsevenlater(2011and
2012,respectively)(seeTable3.8,nextpage).ThestaffinglevelsintheDetroitSANEprogram(typically56
providers)wereconsistentwiththoseinNewOrleans,butagain,NewOrleansishalfthesizeofDetroit,
with~40%fewerforciblerapes(per100,000).Bycontrast,Baltimorehasapproximatelyeighttimesas
manysexualassaultforensicexaminers(acrossitstwoprograms)asDetroit,buthasapproximately
300,000fewerresidentsand~33%fewerforciblerapes(per100,000).
EvenaftertheSANEprograminDetroitwasfoundedin2006,therewerestillseveraltransitional
yearsduringwhichmanyvictimshadexamsperformedbynonSANEpersonnel(20062009)becausethe
programdidnotyethavethecapacitytotreatallvictimsinthecity.Therefore,itisreasonabletoinferthat
thevastmajorityoftheunsubmittedSAKsinthecensus(19802009)werecollectedbynonSANEmedical
personnel.WhenaskedaboutthequalityofcarevictimsreceivedinthepreSANEyears,onemedical
providerindicatedthatitwasgenerallyquitepoor:
48
Childandadolescentvictimsweretypicallytreatedatthecitysonechildrenshospital.
92
[Intheemergencydepartment]theresalackofequipmentandalackoftrainingandalackof
privacy,andnorealsupportiveserviceseither.Norealsocialwork...[sothere]wasntalways
[someone]thereforrapevictims.
Intermsoftheforensicadequacyofthekits(i.e.,whethertheywerecollectedcorrectlyandhadgood
forensicutility),moststakeholdersweinterviewedstatedthattheysimplydidnotknow.Indeed,most
notedthattheexams/kitswereknownfortheirunknownquality,sotospeak;asonememberofthe
collaborativeexplained:
Weretheexamsanygood?Whoknows?Werethekitscollectedcorrectly?Whoknows?...its
safetosaytheywereknownforbeingkindofacrapshoot...most(exams/kits)wereprobably
rushed,youknow,swabswab,moveon.
WithrespecttothetraininghospitalEDpersonnelreceivedonperformingmedicalforensicexams,SAK
evidencecollectiontechniques,workingwithsurvivors,andburnout/vicarioustrauma,onestakeholder
simplysaid,None.Therewasnotraining.Theinstructionsheetinthekit,thatwasthetraining.
TABLE3.8ComparableCities:SpecializedSANE/SAFEServices
YearinWhichSAFE/SANE
ProgramEstablished
AverageNumberof
SAFE/SANEMedicalProviders
(FromBeginningofProgramto2009)
Detroit
2006
56Providers
Philadelphia
2011
(ProgramEstablishedafter2009)
Dallas
2012
(ProgramEstablishedafter2009)
Program1:1994
Program1:2730providers
Program2:2004
Program2:1015providers
2000
56Providers
City
Baltimore
NewOrleans
93
QuantitativeModeling:MedicalSystemResourcesOverTime.ThekeychangeinDetroitwith
respecttomedicalservicesforsexualassaultvictimswastheemergenceoftheSANEprogramin2006.
PriorresearchhasestablishedthattheimplementationofSANEprogramshaspositivedirecteffects(e.g.,
betterqualityforensicevidencecollection)andindirecteffects(e.g.,betterpostassaultcareandcrisis
intervention,whichcanincreasevictimsengagement)onsexualassaultprosecutions(Campbelletal.,
2009;Campbelletal.,2010).Assuch,weexaminedwhethertheemergenceoftheSANEprogramwas
associatedwithincreasedSAKsubmissions.AftercontrollingforDNA/CODISera,thiseffectwasstatistically
significantsuchthatmorekitsweresubmittedfortestingafterthisprogramwaslaunched(seeTable3.9,
nextpage).Thus,consistentwithpriorresearch,theimplementationofaSANEprogramcanhavea
positive,directeffectonhowthecriminaljusticesystemprocessessexualassaultcases,andinthis
instance,itsimpactwasonsubmissionofSAKsforforensictesting.Insubsequentanalyses(seeThePolice
andtheMedicalSystem:InterOrganizationalCommunication),wewillexaminewhytheSANEprogram
mayhavebeensoinfluentialtopolicedecisionmakingregardingkitsubmission.
94
TABLE3.9MultiLevelLogisticRegressionResults,PredictingSAKSubmissionbyImplementationofSANEProgram
Block
Log
Odds
Intercept
1.157
DNAErasinwhichSAKwascollected
0.314
Confidence
Interval
0.242 0.409 <.001
PreDNAEra(19801993)
0.764
0.466
0.325 0.667
<.001
DNAbutNoCODIS(19941997)
0.751
0.472
0.324 0.686
<.001
0.440
1.553
1.0742.246
0.021
DNA&CODISbutCrimeLabDoes Not
HaveCODISAccess(19982001)
DNA&CODISbutCrimeLabSeeking
CODISAccreditation(20022005)
Odds
Ratio
#
Parameters
LRchi
square
15351.13
15344.86
12.54
0.028
15342.74
4.24
0.039
ReferenceCategory
DNA&CODISandCrimeLabAccredited
(20062009)
0.437
1.548
1.0362.412
0.034
YearSANEwasestablished(2006)
0.596
1.814
0.992 3.316
0.053
Chisquare
SD
RandomInterceptvariance
Log
Likelihood
0.231
df
22
119.564
<.001
N=10,817kits,nestedwithin28years;analysiswasconductedusingmixedeffectslogisticregression,withrandomeffectsofkitsnestedwithinyears
95
Victim Advocacy Resources 1980-2009

ContextualFindings.Asnotedpreviously,therearetwomaintypesofvictimadvocacy
programs:systemsbasedprograms(e.g.,policedepartmentvictimadvocacyprograms)andcommunity
based/nonprofitprograms(e.g.,rapecrisiscenters).InDetroit,thepolicedepartmenthashadalong
standingsystemsbasedsexualassaultvictimadvocacyprogram.Foundedin1977,theprogramhas
typicallyhad1014MSWlevelstaffmembers,whoprovideonsiteassistanceinhospitalemergency
departmentsandfollowupcounselingservices(ifrequestedbysurvivors).AsshowninTable3.10
(below),apolicedepartmentbasedadvocacyprogramissomewhatunusual,asonlyoneofthefour
comparablecities(Dallas)hadthisresource.
TABLE3.10ComparableCities:VictimAdvocacyServices
SystemsBased(Police)
AdvocacyServices
City
CommunityBased(NonProfit)
AdvocacyServices
Yes1014MSWlevel
advocates
Limited1AdvocatePosition
No
Yes18Advocates
Yes2advocates
Yes3Advocates/Counselors
Baltimore
No
YesVolunteerAdvocatesOnly
NewOrleans
No
Yes4Advocates
Detroit
Philadelphia
Dallas
TheDetroitsystemsbasedvictimadvocacyprogramhasthesamechainofcommandasthesex
crimesunit(i.e.,theybothreporttothesameseniorcommandstaff).From20022009,thepolice
departmentvictimadvocacyprogramprovidedservicesto10,648adultsexualassaultvictims.49Notall
49
Recordspriorto2002werenotavailable.Theprogramservesbothvictimsandtheirsignificantothers(e.g.,family/friends
whomayaccompanyvictimstothehospital),sosomeofthose10,648individualsservedwerenotvictims,butsignificant
othersofvictims.Givenhowtheprogramtracksservices,itwasnotpossibletoseparatenumberofvictimsvs.numberof
significantothers;however,programstaffindicatedthatthevastmajorityofthoseservedwerevictims(notsignificantothers).
96
ofthesevictimshadSAKscollectedorfiledpolicereports,butthesedatadohelpshedsomelightonthe
numberofvictimsinDetroitseekingtreatmentinhospitalemergencydepartmentsforsexualassaults,
andshowsthatthepolicevictimadvocacyprogramhadcontactwithsubstantialnumberofvictimsover
thoseeightyears.AsrecipientsofVOCA(VictimsofCrimeAct)funding,thepolicevictimadvocacy
organizationparticipatedinongoingtrainingsandseminarsonvictimservices,andtried,totheextent
possible,toprovidesupporttotheircolleaguesinthesexcrimesunit,butstakeholdersnotedthat
discussionsofburnoutandvicarioustraumawerenotcommonplaceinthedepartment.
CommunitybasedadvocacyintheDetroitareahasasubstantiallyshorterhistory.Nonprofit
rapecrisiscenterswereformedthroughouttheUnitedStatesinthe1970sand1980s(Martin,2005),
butDetroitdidnothavesuchanorganization.Thisisatypical,giventhatallfourcomparablecitieshave
suchagencies(seeTable3.10).50InefforttoexpandcommunitybasedadvocacyinDetroit,in2000,one
sexualassaultdesignatedadvocatepositionwasestablishedwithinadomesticviolenceagency.51By
contrast,communitybasedadvocacyprogramsincomparablecitiestypicallyhadthreetoeighteen
sexualassaultadvocatepositions.AstowhyDetroithadsolittlebywayofcommunitybasedadvocacy,
onestakeholdernoted:
50
Rapecrisiscenterstypicallyoffer:24hourhotlines,informationandreferral,victimadvocacyservices,assistancewithcrime
victimcompensation,andcrisiscounseling(amongotherservices,perDepartmentofJustice,OfficeofVictimsofCrime).Tothis
day,thereisnooneagencyinDetroitthatoffersallofthesecoreservices:hotlinesareoperatedbythedomesticviolence
agencies;advocacyandcounselingareprovidedbythesystemsbasedpoliceprogram(butservicesarenotconfidential)and,
onaverylimitedscale,bythedomesticviolenceagencies(withconfidentiality).In2010,thesexualassaultforensicexam
programreceivedfundingtoexpandtheirscopetoincludecommunitybasedadvocacyservicessothatvictimshadmore
optionsforconfidentialadvocacyandcounseling.
51
Archivalrecordsregardingthenumberofvictimsservedbythatonecommunitybasedadvocatewerenotavailable,butitis
reasonabletoassumethatthenumberwaslikelyquitelow.
97
Itwasassumedthatthe[policeadvocacyprogram]wastakingcareofit...theywerefunded
todohospitaladvocacy...theyhadalotoffundingforthat,alotofstaff...andtheywere
there[atthehospital,withvictims]butweretheyreallythere?Advocatingforvictims,pushing
backagainstthepolice,whobytheway,werethereemployers.Idontknow...givenhow
many[unsubmitted]kitswehavenow,itdoesntseemlikely.(emphasisinoriginal)
Eventhoughtherearecriticaldifferencesbetweensystemsbasedandcommunitybasedadvocacy,it
waseasytoturnablindeyeandnotpayattention,asonestakeholdernoted,becausetherewas
limitedstatelevelfundingavailableforcommunitybasedadvocacyandDetroithadalongstanding,
reasonablywellfundedsystemsbasedvictimadvocacyprogram.Inhindsight,manystakeholders
questionedwhethervictimswereadequatelyservedinalloftheirneedsgiventhatthesystembased
advocatescouldnotprovideconfidentialityandthescopeoftheirservices(inpractice)appearstohave
beenfocusedonprovidingpostassaultsupporttovictimsinhospitalemergencydepartments,
informationandreferral,and(ifrequested),followupcounseling.
QuantitativeModeling:VictimAdvocacyResourcesOverTime.Systemsbasedadvocacywas
availablethroughoutthethirtyyearsinthishistoricalanalysis,andwedidnotdocumentanysubstantial
changes/reductionsinstaffingpersonnelwithinthisprogram.Communitybasedadvocacybeganin
2000,albeitwithonlyonestaffposition,anditseemsunlikelythatoneadvocatecouldhavehada
significantimpactonSAKsubmissionrates(indeed,attheunivariatelevel,therewasnosignificant
association).Assuch,multilevelstatisticalmodelingoftheimpactofadvocacyresourcesovertimewas
notwarranted.
98
The Impact and Legacy of Chronic Resource Depletion
TheresultsofthishistoricalcontextualanalysisindicatethatallorganizationsinDetroits
systemicnetworkstruggledwithchronicunderstaffingandresourcedepletionovertheyears(seeFigure
3.2,nextpage).Thepolicedepartmenthadmajororganizationalchangesintheirleadership
approximatelyeverytwoyears,aturnoverratethatexceedsnationalnormsandisatypicalamongother
urbancitieswithcomparablecrimerates.Theconstantchangesinleadershipmadeitdifficulttosustain
positiveinitiativesandtoidentifyandcorrectproblematicpoliciesandlowperformanceamongstaff.
Maintaininghighperformancewasalsodifficultgiventhatthesexcrimesunitsustainedtwo50%cutsin
theirstaffinglevelsovertheyearsexaminedinthisanalysis.Thepolicecrimelabtypicallyhadonly23
DNAscientistsonstaff,anumbersubstantiallylowerthaninotherurbancitieswithcomparablecrime
rates.Theprosecutorsofficehadfarmorestabilityinleadership,butgivencountylevelbudgetcuts,
thenumberoftrialattorneysdeclinedsignificantlyovertime.SexualAssaultNurseExaminer(SANE)
programsarewidelyconsideredtobebestpractice(seeDepartmentofJustice,2013),butDetroitdid
nothavesuchaprogramuntil2006,andtheoverwhelmingmajorityofSAKsinthecensuswere
collectedbymedicalpersonnelwhohadnotraininginforensicevidencecollection.Thepolice
departmenthadalongstandingvictimadvocacyprogram,butcommunitybasedadvocacyandits
promiseofconfidentialserviceswaslargelynotavailable.
Takentogether,theseyearsofchronicresourcedepletioncreatedwhatsomestakeholders
referredtoasThePerfectStorm:
ItslikeThePerfectStorm...theconditionswereripeforthistohappen...honestly,looking
back,Ithinktherealquestionisnt,howdidithappen,buthowcoulditnothavehappened?Not
asanexcuse...asrealityhowcoulditnothavehappened?
99
FIGURE3.2SummaryofHistoricalContextAcrossAllDetroitOrganizations
UNDERSTAFFING&BUDGETCUTS
MULTIPLELEADERSHIPTRANSITIONS
50%CutinPoliceSexCrimesUnitStaffing
TwoCrimeLabDNAScientists
OneCommunityBasedAdvocate
10PoliceChiefsOver30Years
MajorLeadershipChanges~2Years
CHRONICRESOURCE
DEPLETION
INNOVATIONSSLOWTOIMPLEMENT
MINIMALSUPPORTFORPROVIDERS
2006:DNACODISAccreditation
2009:SpecializedProsecution(Volunteer)
2006:SANEProgram
2000:OneCommunityBasedAdvocate
LimitedProfessionalTraining
LimitedVicariousTraumaSupport
HighBurnout
Thepolicedepartment,crimelab,prosecutorsoffice,medicalsystem,andvictimservice
agenciesfunctionedunderchronicresourcedepletion,whichraisesthequestion,howdoesworking
undersuchconditionsdayinanddayoutaffectpeopleandorganizations?Drawingfromresearch
acrossmultipledisciplines(e.g.,psychology,sociology,economics),theliteraturesuggeststhatsuch
conditionscreatewhatisoftenreferredtoasacultureofscarcity:apervasive,widelyheldmindset
thatthereisfartoomuchtodoandnotenoughresourcesavailable,soeverythingsimplycannotget
done(Fried,1982;Kramer,1990;Lipsky,2009;Mullainathan&Shafir,2013;Rouxetal.,2012;Walsh,
1961).Researchonchronicscarcityhasfoundthatsomepeoplemeetsuchchallengeswithtremendous
resiliencyanddetermination,andcontinuetobededicatedandworkhardwithhumilityandpatience
(seeBonanno,2004,2005forreviews).However,mostdonotrespondinthatmannerandevenif
theycouldforawhile,theyareunlikelytosustainsuchcommitmentovertimeandinallcircumstances
100
undersuchharshconditions.Moretypically,chronicscarcitytendstobringoutnegativebehaviorsin
peopleandorganizations.
Specifically,theliteraturesuggeststhattherearethreemainnegativeoutcomesstemmingfrom
chronicresourcedepletion.First,individualsworkinginunderresourcedorganizationstendtotake
shortcutstomanageandlimittheirworkloadsifthenormativebeliefisthatitisnotpossibletoget
everythingdone,thenitbecomesnecessarytotakemeasurestomakeitmanageable(Edney,1982;
Lipsky,2009;Mullainathan&Shafir,2013;Rouxetal.,2012;Tetlock,2000).Workersselecttasksthat
areperceivedasemergencies,highprofile,and/orimportant(basedonsettingnorms),andoffload
workthatisperceivedaslowpriority,hard,and/orcomplicated(duetotimeconstraints)(Lipsky,2009;
Tetlock,2000).Second,chronicscarcitytendstodecreaseindividualsempathyforothers(Fried,1982;
Kramer,1990;Lipsky,2009;Rouxetal.,2012).Forthoseworkinginthehelpingprofessions,itisnot
uncommonthatstaffwilldehumanizetheverypeopletheyaresupposedtohelp,disregardingtheir
sufferinganddistancingthemselves(Lipsky,2009).Inaddition,itbecomesincreasinglydifficultforstaff
tohaveconcernandempathyfortheircolleagues,bothwithintheirownorganizationandforthose
workinginotherorganizations(Lipsky,2009).Finally,chronicscarcitytendstopromotebunkersand
silos:individualsbunkerdownwithintheirownorganizationandfunctioninindependentsilos,
cuttingoffcommunicationandcollaborationwithothergroups(Fried,1982;Kramer,1990;Lipsky,2009;
Mullainathan&Shafir,2013;Rouxetal.,2012;Walsh,1961).
Theliteratureonchronicscarcitysuggeststhatlongtermresourcedepletionisinandofitself
detrimental,butmorethanthat,itcausesindividualsandorganizationstochangetheirbehaviorover
time,changesthatareoftenharmful,particularlytothoseinneedofhelpfromtheseovertaxedgroups.
Therefore,inthecontextofDetroitsproblemwithlargenumbersofunsubmittedSAKs,resource
depletionisonlyonepartoftheexplanation;theotherpartrequiresdelvingdeeperintohow
professionalswereapproachingtheirworkonsexualassaultcases,howtheyweretreatingvictims,and
101
howtheyweretreatingeachother.Highlightingthispointthatresourcedepletionisonlypartofthe
story,onememberofthecollaborativenoted:
Thisisntthewholestory...sure,forthirtyyears,andevenlongerthanthat,thiscityhasnt
had[whatit]needstocareforitscitizens.Butalotofcaringdidhappen.Thisdidnt.Thiscrime
(sexualassault)wasntcaredabout...thesecitizens(sexualassaultvictims),specifically,werent
caredabout...therewerentenoughresourcestogoaround...whatlittletherewas,itdidnt
gohere(tohelprapevictimsandtestSAKs)...Therewasachoice,lotsofchoices,choicesevery
daynottohelpavictim.(emphasisinoriginal).
Theresultsofthecensusindicatethatthousandsoftimes,therewasadecisionnottosubmitaSAKfor
testing.Why?Toanswerthisquestion,weneededtoexaminethefrontline,daytodaypracticesin
theseorganizationstounderstandhowandwhysomekitswereshelvedinpoliceproperty.
Underlying Factors: Front-Line Practices & Inter-Organizational Communications

Overview
WhenweaskedstakeholderswhytheythoughtthereweresomanyunsubmittedSAKsin
Detroit,nearlyallmentionedthatgenderwasundoubtedlyakeyfactor.Asonestakeholdersaid,I
thinkthat'sprobablythe#1reason[whykitsarentsubmitted],itaffectsmostlywomen...ifmenwere
gettingraped,Ithinkthatitwouldntbelikethat.Similarly,anothermemberofthecollaborativesaid,
Itsnotthatcomplicatedtofigureout...thisisacrimethataffectswomen,andinthiscity,thatmeans
Blackwomen,poorBlackwomen...theresagoodchunkoftheexplanationrightthere.Sexualassault
isacrimethatdisproportionatelyaffectswomen,andtherefore,becauseoftheirdevaluedstatusin
society,andthemorehighlyvaluedstatusoftheir(usually)maleperpetrators,thiscrimeisunlikelyto
receivewidespreadsocietalattention(Bergoffen,2005;Funk,2006;Rush,2010).
102
InDetroit,acitythatispredominatelyAfricanAmerican(82%inthe2000Census,and82.7%in
the2010Census)withathirdofitsinhabitantslivingunderthepovertylevel(21.7%offamiliesand
26.1%ofindividualsin1999;anestimated32.3+/1.9%offamiliesandanestimated37.6+/1.8%of
individualsin2009),theintersectingoppressionsofgender,race,andsocialclassplacemanywomenat
riskforabuse(seeDavies,Francis,&Greer,2007).Manystakeholdersalsonotedthatthesefactors
undoubtedlyaffectedwhichSAKsweresubmittedfortestingandwhichoneswerenot,asthesethree
quotesillustrate:
Manyofthemarepoor...manyofthemarelivingnotonlyalegacyofracismbutactive
racism,activemisogyny...andtheyhavemultipleproblems[intheirlives]...andthecriminal
justicesystem[isjustgoingto]exacerbate[that.]
AhousewifefromGrossePointeShores[who]getsraped...thatrapekitsgonnago(for
testing)andthehouseisgoingtobedusted(forprints)versusalowermiddleclassblack
womantheyretreateddifferently.Ithinkthatthere'sthismentalitythatsomeofthesewomen
mayhavecausedittohappen.
Ifitsnotawhitegirl,whitewoman,oranaffluentblackperson[whocanmake]somenoise,
they(thepolice)willfeelverycomfortable...ignoringit.
Onememberofthecollaborativeexpressedthesentimentsofmany(butcertainlynotall)team
memberswhens/henoted,you(theresearchers)probablycantproveitinthescienceway,butwe
knowthishaseverythingtodowith[thefactthattheseare]Blackwomen,mostofthempoorBlack
women.52Indeed,whetherthisnotioncouldbeempiricallyprovenisdebatable(giventhe
homogeneityofthesample);instead,itmaybemorehelpfultotreatsexism,racism,andclassismas
sensitizingconcepts,whichare,byBlumers(1954)classicdefinition,conceptsthatgivetheusera
52
Inreviewingadraftofthisreport,policeofficialsexpressedstrongdisagreementwiththissection(i.e.,thematerialregarding
theroleofgender,race,andclassinSAKsubmissions),highlightingthatmanymembersofthedepartmentareAfrican
Americanandmanyarefemale,andthatamongthehigherleadershipranksofthedepartment,womenandethnicminorities
havebeenwellrepresentedthroughouttheyears.
103
generalsenseofreferenceandguidanceinapproachingempiricalinstances(p.7).Inotherwords,
sensitizingconceptsprovidealensthroughwhichthedatacanbeviewedtounderstandtheobserved
processesandoutcomes.Therefore,theintersectingoppressionsinthelivesofthesesurvivors,andhow
thoseinequalitiesmightmanifestintheirinteractionswithsocialsystempersonnel,wasanintegralpart
ofourresearchontheunderlyingreasonswhyDetroithassomanyunsubmittedSAKs.
Specifically,thiscomponentofourresearchhadtwoprimaryaims.First,wewantedto
understandhowchronicscarcityaffectedthefrontlinepracticesofpolice,andhow,asonestakeholder
notedabove,therewerechoiceseverydaynottohelpavictim.53However,thepolicearealsopartof
aninterdependentsystemofmultipleorganizationsthatrespondtosexualassaultvictims.Therefore,
oursecondaimwastoexploreinvestigatorsinteractionswithotheragenciesinthesystemicnetwork
andhowthoseinteractionsmayhavehadanindirectinfluenceonpolicedecisionmaking.Inother
words,wewantedtounderstandwhathadbeenhappeningonthefrontlineswithinthepolice
departmentandbetweenthepoliceandotherserviceproviders.54
TounderstandpastpolicepracticesinsexualassaultinvestigationsandSAKsubmissions,we
drewuponthreedatasources:ethnographicobservationsofcollaborativeteammeetings;stakeholder
interviewswithpolicepersonnel,includingcurrentemployeesandformeremployees,spanningalllevels
oftheorganization(patrol,sexcrimesunitinvestigators,supervisors,andseniorcommandstaff);and
archivalrecords(i.e.,1,268sexualassaultpolicereports)(seeAppendixB:ProjectMethodologyfor
detailsregardingdatacollection,dataanalysis,andtriangulation/datacredibility).55
53
Again,thisanalysisfocusedonpastpractices(upto2009)inthepolicedepartment,nottheircurrentapproachestosexual
assaultinvestigationsandrapekittesting(whichwasoutsidethescopeofthecurrentstudy).
54
Theseinterorganizationalanalyseswerealsofocusedonpastpractices(upto2009).
55
Itisimportanttoemphasizethatthisanalysisofpastpracticeswouldnothavebeenpossiblewithoutthesupportofthe
currentleadership(ofthesexcrimesunitspecificallyandthedepartmentmoregenerally).Allowingtheresearchteamaccess
tothese1,268policefilesreflectsthecurrentleadershipscommitmenttotransparencyandchange,eventhoughthe
informationthereinwouldlikelynotbeflatteringtotheorganization.
104
Toexaminehowpolicepracticesmighthavebeeninfluencedbyinteractionswithother
organizationsinthesystemicnetwork,wefocusedonthedatainthestakeholderinterviews.Ourgoal
wastounderstandwhatmessageswerebeingcommunicatedfromoneorganizationtoanother,and
howwerethosemessagesbeingheardandactedupon.Wewerenottryingtodocumentwhathad
happenedinspecificcases,butrathercommonthemesininterorganizationalcommunicationsacross
manycasesandmanyyearsofworkingtogether.Datacollectionwasaniterativeprocessofidentifying
themesincrossorganizationalcommunicationandthencheckingwithrepresentativesfromeach
organizationregardinghowtheyinterpretedthosemessages(seeAppendixB:ProjectMethodology).
Police Department Front-Line Decision Making & Practices

ReframingtheProblem:TheCasevs.TheKit.Inourinterviewswithpolicestakeholders,we
askedspecificquestionsaboutSAKsubmissionguidelinesanddecisions(e.g.,Canyoutellmeabout
whyaSAKwouldorwouldnotbesubmitted,Canyoutellmeabouthowyoudecidewhethertosubmit
aSAKfortesting).Inansweringthosequestions,policepersonnelexplainedthatkitsubmissionwas
onlyonepartofalargerdecisionmakingprocess,asonepoliceofficialexplained:
Thekitsnottheissue...itstheinvestigation,figuringoutwhatwecoulddoto...if[theres]
somethingwecoulddo,thenwewould.(Q:wouldthatincludesubmittingthekitfortesting?)
Depends.Ifitwasastranger,wherewewereatwithCODIS,allthat.So,sometimes,yes,
sometimesno...you(referringtotheresearchinterviewer)keepaskingaboutthekit.Wrong
question.Whatmatteredwasthecaseandwhetheritwasrealandwhetherwecoulddo
anythingaboutit.
OtherinterviewswithpolicepersonnelconfirmedthatthedecisionaboutSAKtestingwasnotadiscrete,
separatedecision;whetherakitwouldbetesteddependedonprecedingdecisionsandevaluations
regardingtheoverallmeritofthecase,asthisseriesofthreequotesillustrates:
105
Itsnotlikethedecisionaboutthekitwasseparateoranything...wehadtojudgethecaseas
awholeanddecidewhattodoaboutit...whetherithadmerit.
Everyonesallup[inarms]aboutthekits...fromourpointofview,itstheinvestigation...a
kitwouldntbetestediftherewasnocasetomake.
Ifareportisfalseorwecantestablishelementsofthecrimeoritsunfoundedorthe
complainantrefusestoprosecute,thenno,werenottakingthekittothelab.Whywouldwe?
...thatswhytheressomanykits[inproperty].
Inlightofthisinformation,ourfocusshiftedtounderstandingthemoregeneralprocessofhow
policeapproachsexualassaultinvestigations.Casesassociatedwithunsubmittedkitswerenot,asa
group,thoroughlyinvestigated.Inmanyinstances,itwouldbedifficulttoclaimthattheywere
investigatedatall.Forinstance,whenwewerereviewingthecasefilesassociatedwithoneofthefour
testinggroups(TestingGroup4;seeChapter4:Developing&EvaluatingaSAKTestingPlan),84%ofthe
reportsfrom20022005were12pageslong,consistingonlyoftheinitialpolicereportmadebythe
victim.56Inmanyrespects,theuntestedkitswereatangiblesignaboutthedispositionsofthesecases
thecasehadbeenshelved,figuratively;thekithadbeenshelved,literally.Therefore,wewantedto
understandwhythishappenedsofrequently,orasonepoliceofficialputit:Thekits[thatwerent]
testedwerecasesthatwecouldntorwouldntdoanythingabout.57Whywouldthepolicedecidea
caseissomethingthattheycouldntorwouldntactupon?
56
WealsosawsomefilesassociatedwithunsubmittedSAKsthatwereverythoroughlyinvestigated,filesmanyinchesthickof
documentation;however,thosetypesofcase/filesweretheminorityofthosewereviewed.
57
ThisanalysisfocusesoncasesinwhichtheSAKswerenottested,whichlikelydoesnotreflectpoliceinvestigationspractices
incasesinwhichtheSAKsweresubmittedfortesting.
106
CouldntDoAnythingAboutIt:TheImpactofChronicResourceDepletion.Policeofficials
spokeatlengthabouthowdifficultithasbeentosustainqualitypoliceworkunderdecadesofresources
cuts.Budget/personnelcutswereparticularlyhardhitting,aspolicedidnothaveenoughinvestigators
tohandleallthecasesthatwerereported.Asonememberofthepolicedepartmentexplained:
Itsoundslikeanexcuse,[and]theresnoexcuse,[but]whenyouhavefewerinvestigatorsand
thesamecaseload,youhavelesstimetoinvestigateeachcase[when]youdonthaveenough
peopletoinvestigateit,youknow,thatcancomebacktoreallyhauntyou.
Otherpolicestakeholdershighlightedhowbudgetcutsnegativelyaffectbothhighandlowperforming
investigators.Fortopnotchemployees,itbecomesincreasinglydifficulttogive100%performanceto
allcaseswhentheywerestretchedsothin;forothers,slackperformanceisdifficulttocorrectwithout
adequatesupervisionandopportunitiesfortraining,bothofwhichwerehistoricallylackinginDetroit.A
policesupervisordescribedhowthisnegativedynamicemergesandhowdifficultitistocorrect:
Letssayyouhaveareallyremarkableteamofpeopleandtheyredoingagreatjobandthen
[with]budgetcutstheythrowpeopleinthatmightnotbesogreattheygo,okayImadeone
phonecallcouldntgetaholdofthevictim,thecaseisclosed...youdonthaveanychecksand
balancesonthat...thebosssaying,whatdidyou[do]beforeyouclosedsomething.BecauseI
thinkonsomeoftheseolderones[kits]yousee...thelackofeffort...Theydidnthavetime.
Highcaseloadsandchronicunderstaffingledtochronicburnout,whichalsonegativelyaffected
investigationalquality,asonepolicestakeholdernoted:
Idothinkitsaresourceissue;Idothinkthatofficersaresooverwhelmedwiththenextcase
walkingthroughthedoorthattheytakeshortcuts,andthattheydontdoeverythingthey
shoulddo.Ithinktheburnoutisveryhigh;Ithinkpeopleareinpositionslongerthanthey
shouldbe.
107
Howdopoliceapproachsexualassaultinvestigationswhentheyareunderstaffed,exhausted,
andburnedout?Figure3.3(nextpage)depictstheprocessthatemergedfromourstakeholder
interviewsandreviewof1,268sexualassaultpolicereports.Lawenforcementofficialswerecandidthat
theinvestigationsformanyofthesexualassaultcasesassociatedwithunsubmittedSAKswerenot
thorough.Onestakeholdersummeduphowmanyofthesecaseswerehandled:Okay,Imadeacouple
phonecalls,Ithrewmycardinthedoorandthatstheendofthat.Closeit.Speakingtothisissueofthe
thoroughnessofinvestigation,onepoliceofficialsaid:
Whenyouhavefewerinvestigatorsandthesamecaseload...therealityofitthatyoumaynot
crossallyourtsanddotallyouris.wheredoyoucut?Canvassing,lookingforadditional
witnesses...personalcontacts...Thosearethings,thatswhereyoullseecornerscutorwe
shouldbereachingoutmoretalkingtopeople,wejustdont,wedonthaveachanceto.
Anotherstrategyfordecidinghowtoallocatelimitedtimeforunlimitedcaseswastowaitforthe
victimtoinitiatefollowupcontactwiththepolice.Asonepolicesupervisordescribed:
IgetcallsfromvictimssayingmydetectivehasntcalledmeandIlllookinthenotesanditsays,
Imadeaphonecall,Icalledat3onTuesdayafternoon,therewasnoanswer,Ileftamessage..
.[theinvestigator]maynotputintheeffortthatwewouldliketoseehimputin...hemayjust
say,makeaphonecallandthatsit.Thenifshewantstocomplainorshewantstoprosecute,
Imsureshellbecontactingme.
Inotherwords,policewouldshifttheburdentothevictimtocontactthemandpursuethecase,oras
oneinvestigatorsaid,shehastoproveshewantsthis...thenIlltakealook.Whenquestionedabout
thispractice,policereadilyacknowledgedthatitwasnotfairtovictims,butthatitsoneofmany
strategiesinvestigatorsusetomanageworkloadifthevictimisengaged,thenthatcasemightbeone
inwhichtheycouldinvesttimeandeffortifnot,thennot.
108
FIGURE3.3PoliceProcessingofSexualAssaultCasesAssociatedwithUnsubmittedSAKs
AsshowninFigure3.3,resourceconstraintsledtocuttingcornersanddecreasedefforton
cases;whenvictimsdidnotrespondtoinvestigators,admittedly,minimaleffortstocontactthemand
developtheinvestigation,policeoftenlabeledvictimsasnotcooperativeorrefusedtoprosecute.
Asonepoliceofficialexplained:
Itcomesdowntotheindividualofficerwhethertheythinksomeonescredibleornot...you
mighthaveanofficerwhomaynotbeparticularlymotivatedorambitiousandwhomightsimply
markafilethatthecomplainantrefusestocallbackorrefusestoappear,andthatmaynotbe
accurate...Wefindthatalotoftimeswhenanofficerhassortofdeemedthatacomplainantis
uncooperative,thatsnotreallythecase.Sotheresjusttoomuchdiscretioninthatprocess.
Butonceavictimwaslabeledasnotcooperative,moreoftenthannot,thecasewasnotpursuedfurther
(including,notsubmittingtheSAK).Thetermnotpursuedisuseddeliberatelyherebecausethefinal
casedispositionsdocumentedinthefilesvariedwidely.Somewereofficiallyclosed(e.g.,unableto
109
establishtheelementsofthecrime,unfounded),butothersremainedopen(nearlyadecadelater)
(e.g.,tolocate[meaning,offenderstilltobelocated],pendingvictimcontact).Whatwascommon
acrossthecaseswithunsubmittedSAKswasthatmoreoftenthannot,therewasminimal
documentationandminimaleffortinvestedinthecase.
WouldntDoAnythingAboutIt:TheImpactofVictimBlamingAttitudes.Thoughresource
depletionundoubtedlyhadanegativeeffectonthesexcrimesunit,insufficientstaffingwasnottheonly
reasonwhysomanykitswerenotsubmitted.Indeed,thequantitativemodelingpresentedintheprior
sectionofthischapterfoundnostatisticalassociationbetweenstaffingcutsandSAKsubmissionrates,
whichsuggeststhattherewereotherfactorsatplay.Policenotedthattherewerecasestheywouldnt
doanythingabout,andbasedonthedatafromthestakeholderinterviewsandthesexualassaultpolice
reports,itappearsthatthiswasduetonegativebeliefsandstereotypesaboutvictims,whichadversely
affectedthequalityoftheinvestigation,andtherefore,SAKsubmission.Asnotedpreviously,all
stakeholdersnotedthatnegativeattitudesrootedinsexism,racism,andclassismwerecontributing
factorstothisproblem;however,whenfocusingspecificallyonthepolice,thestakeholderinterviews
andpolicereportshighlightedthreespecificattitudesthatappeartohavenegativelyimpactedcase
investigations(andultimatelySAKsubmissions)(seeFigure3.4,nextpage).
First,policepersonneloftenassumedthatvictimsreportingsexualassaultswereengagedin
prostitution/sexwork.Inboththestakeholderinterviewsandinthepolicereports,therewerefrequent
referencestodealsgonebad:
[sometimesitwas]adealgonebad,shegotherselfcaught.(Q:gotherselfcaught,whatdoes
thatmean?)Shewasprostitutingandsheagreedtothemoneyandhedidntpayher...shesays
itsrape.Itsnot,itsadealgonebad.
110
FIGURE3.4PoliceAttitudesinSexualAssaultCasesAssociatedwithUnsubmittedSAKs
Supportingtheirpositionthatthereportwasadealgonebad,policeoftenhighlightedthelocationand
timeoftheincident,whichofcourseshouldbenotedinapolicereport,buttheneighborhood/areain
whichtheassaultoccurredwasoftenheavilyemphasizedinthereport.Whenaskedaboutthispractice,
oneinvestigatorexplained:
Wehavetonotetimeandplacethatsbasicpolicework.(Q:Iappreciatethat,butwhyisit
mentionedrepeatedlyinthenarrative?)Becauseitsrelevant.IfshesonJohnRatthattimeof
day,wellwhatelseisshedoing?...Itskindalikecodeamongus(theinvestigators),makea
pointaboutwhatneighborhoodithappenedin,andwell,enoughsaid.
Inthestakeholderinterviews,weaskedpolicehowandwhytheysuspectedavictimmighthavebeen
involvedinprostitution,whichoneinvestigatordescribedas:
Itsnotonething,usually,neighborhood,street,circumstanceoftheassault,likeifsheaccepted
aridewithsomeone...howshelooked...cantputyourfingeronitexactly,butyoudothis
longenough,youcantell.
111
Whenaskedifwomeninvolvedinsexworkcouldindeedbevictimsofrape,somepoliceofficialsagreed
(e.g.,ofcoursetheycanberapedsomerapistspreyonthemcausepeoplewontbelievethem),but
again,theoverridingconcernwasthatifitwasadealgonebad,thentheywouldntwanttoinvest
limitedresourcesinvestigatingthecase.Whenaskedifitwaspossiblethatinvestigatorswouldlabela
caseinthatwayinordertomanagetheirworkloads,onepoliceofficialstated:Idliketothinkthat
didnthappen,butityeah,Imsureitdid.
ReturningtoFigure3.3,ifpolicebelievedthatvictimsmayhavebeeninvolvedinprostitution,
theyreadilyacknowledgedthattheytreatedthemdifferently,oftenactingindeliberatelyintimidating
waystotestthem,asonofficerputit,toseeiftheyweretellingthetruth,andsometimestonudge
themoutofthesystemanddiscouragethemfromcontinuedpursuitoftheirreport:
Imjaded,Icomeoff...asjaded,lousydemeanor...shesaprostitutebutshesstilla
womanandwasvictimized,shejustsays,youknowwhat,enoughofthis,Imoutofhere,Illjust
keepongoingaboutmybusiness.
Similarly,anotherpoliceofficialstated,
Iguessoneofthewaysyoucoulddiscourageisbeingthejadedpoliceprofessional....Just
comingoffwrong...nowinmymindtheyreprostitutes,theyrethisorthat...mylineof
questioningor...demeanor[showsthat].YoulookatmeandgowhatthehellamIdoing
here?[Its]discouragingthem.
Afterbeingtreatedinsuchaway,victimsmightwithdrawfromprocessormightnotcooperatewiththe
police,therebybolsteringinvestigatorsassessmentsthatthecaseshouldnotbepursuedfurther.
Bringingtheseideastogether,Exhibit3.1presentsapolicereportinwhichthevictimisassumed
tobeaprostituteandthatbeliefappearstohavedictatedthecourseoftheinvestigation,tothepointof
discountingpossibleevidenceofacrime.Inthiscase,a22yearoldwomanwasattackedoutsidea
112
liquorstorebyaknownacquaintance,whopulledherintoavacanthouse,threatenedherwithagun,
andthensexuallyassaultedher.Afterwards,sherantoanearbyfirestationandthestafftookhertothe
hospital.Belowaretheofficersnotesfromtheinterviewconductedwiththevictimatthehospital:
EXHIBIT3.1SexualAssaultPoliceReportofSuspectedProstitutionDealGoneBad
Intimidating
Behavior
Deal
Gone
Bad
Burdenon
VictimtoMake
Contact
Thisexamplehighlightshowthedealgonebadassumptioncanaffectaninvestigation.The
victimsconcernsaboutdiscussingtheassaultarereframedasduplicityonherpart;thoughwedonot
knowhowsheinterpretedtheinvestigatorscommentsaboutfilingfelonychargesagainsther,itseems
likelythatsuchactionswereattheveryleastintimidating,andatworst,threatening.Thereportalso
showstheletthevictimcontactusstrategydiscussedbefore,wherebypoliceappeartobemanaging
113
theirworkloadsbyexpectingsurvivorstopursuetheirowncases.Thereportendswithamentionof
physicalinjuriestothevictim,butthesefactsdonotappeartobeinfluentialtotheinvestigator.58
Thesecondcommonstereotypethatemergedinthestakeholderinterviewsandpolicereports
wasthatthecredibilityofadolescentvictimswasquestionableandthattheymadeclaimsofrapeto
coverupforbadbehavior(e.g.,beingoutlate,beingsomewhere/withsomeonethattheywerent
supposedto,experimentingwithalcoholand/ordrugs).Forcontext,itisimportanttonotethata
substantialportionofthevictimswhosekitswerenottestedwereadolescentvictims(e.g.,ofthe1,595
kitstestedinthecontextofthisproject,43%ofthevictimswereunder18).Inthestakeholder
interviews,policedescribedtheirdoubtsaboutyoungervictims,asthisseriesoftwoquotesillustrates:
Sometimeswiththeschoolgirls...anexperiencedinvestigatorknowsmorethanan
inexperiencedone.Youkindoflookandgo,yeahitdidnthappenlike[she]saidithappened.
Theyounggirls,theysaysomethingwentdownsotheywontgetintroublewiththeirmammas
causetheywereoutlateorwitholdermenordoingsomethingtheyshouldntbedoing.
ReturningtoFigure3.3,policenotedthattheywouldlayitonthickwithadolescentvictims,
describingthecriminaljusticesystemprocessindeliberatelyscaryterms,warningthemaboutnegative
consequencesforthemiftheywerefoundtobelying.Oneinvestigatorstated,
Forthegirls,yeah,Idtrytowarnthemwhatsitlike...soiftheyretellingthetruth,theyknow
whattheyreupagainst...iftheyrelying,theyllknowhowmuchtroubletheyllgetin.
58
AsnotedinAppendixB:ProjectMethodology,casestudyexampleswerecarefullyselectedtoreflecttypicalreport
characteristics,circumstances,language,experiences,etc.amongthe1,268policereportswereviewedinthisproject.
114
Inthiscaseexample(Exhibit3.2,below),a14yearoldgirlreportedthatshehadbeenabducted
bytwomen,takentoanabandonedburnedoutbuildingandraped.Atthebeginningofthecasereport
(toppartofExhibit3.2),theofficerstateshis/herbeliefthatthereportisfalseastatementthat
precededanydetailsabouttheassaultitself.Thelastsentenceofthisreporthighlightsthatthevictim
wassupposedtobehomeat7:00pm,butdidntreturnuntil8:30pm.Therelevanceofthisstatementis
unclear,thoughitimpliesthattheaccountmayhavebeenfabricatedtocoverupforbeinglate.The
investigatorssupplementalnotes(bottompartoftheExhibit3.2)elaborateonthesethemes,
highlightinghowtheofficerdidnotbelievethevictimsaccountoftheassault.
EXHIBIT3.2SexualAssaultPoliceReportofanAdolescentVictimCoveringUpBadBehavior
VictimAssumedtobeCovering
UpforBadBehavior
VictimAssumedtobeCovering
UpforBadBehavior
115
Thiscasewasclosed(UTEEC=unabletoestablishelementsofthecrime),anditdoesnotappearfrom
thedocumentationinthefilewhichconsistsofonlythesetwodocumentsthattherewasany
attempttoreviewthesceneofthecrime,canvassforwitnesses,orsearchfortheassailants.Whenthis
caseexamplewaspresentedtopolicestakeholders(aspartofthemembercheckingprocess),they
expressedalarmatthetone,content,andlanguageofthereport,buttheyacknowledgedthatdisbelief
ofadolescentvictimswasapervasiveproblem.
Thethirdcommonstereotypewastodisbelievevictimswhoknewtheirassailants:police
doubtedvictimscredibilityiftheykneworwereevenminimallyacquaintedwiththeassailant.Inthe
stakeholderinterviews,policeexpressedfrustrationaboutthesekindsofcasesbecausetheaccused
perpetratorsoftenclaimthattheincidentwasconsensual,whichlawenforcementfeltwasdifficultto
proveordisprove:[its]impossibletoprovelackofconsent.Whenaskedaboutwhyitwas
impossible,oneinvestigatorclarifiedthattheycanestablishtheelementsofthecrime,includinglack
ofconsent,butthatitisoftentimeconsumingtodosoandtimetoinvestinsuchcasesisoftenlimited.
Policepersonnelalsoquestionedwhetherreportsbetweentwoknownpartieswerefalsely
madebecausethevictimlaterregrettedwhatsheddone.Asanotherofficerexplained:
Idonthavetimetodealwith...wakeupandregret.Youdidwhatyoudid.Thatsthat.Its
notacrimeanddonttakeupourtimewithit.
Investigatorsemphasizedthatgivenhowpressedtheyweretokeepupwiththeircaseloads,theywere
particularlyfrustratedandskepticalofanycasethatseemed,tothem,likearevengereport,meaning,
thevictimwasmakinganaccusationofrapetogetbackatafriendorpartnerbecausetheywere
upsetorbecausetheywereregrettingwhathadhappened.Whenaskedabouthowcommonitwas
thatknownassociates,friends,and/orpartnersrapetheirpartners,policeacknowledgedthatitdoes
happen,but,intheirbelief,notthatoften:Trulyrape?Sometimes.Butnotmostofthetime.
116
Intermsofhowpolicerespondedtovictimswhoknewtheirassailants(seeFigure3.3),
stakeholdersacknowledgedthattheysometimestreatedvictimsofnonstrangerassaultbrusquely,as
oneinvestigatordescribedhis/herapproachtoknownoffendercasesassuch:
Idontbelievethem,sureIletthemknowthat...ifthisboyfriendgirlfriendstuff,thenthats
notmybusinessandItellthemthat...iftheysaynevermind,ok,complainantrefusedto
prosecute.Closeitandmoveontothenextone.
Otherpolicestakeholdersnotedthatvictimsoughttoexpectwhattheygetiftheyinvitesomeone
overoragreetogosomewherewiththem.Asoneinvestigatorstated,itmightnotberight,butits
whathappens,yougooverthere,whatdoyouthinksgonnahappen?
InthecaseexampleinExhibit3.3(nextpage),theinitialrespondingofficersreportisthree
detailed,handwrittenpages,sequencinghowa14yearoldgirlvictimwasthreatenedwithbeing
beatenupandkilled,orallyandvaginallypenetratedbyforce,andsubjectedtoattemptedanal
penetrationbyaknownacquaintanceinhis20s(firstexcerpt,handwritten);thesedetailswere
condensedtothesecondexcerptshowninExhibit3.3:compl(complainant)invitedknownperp
(perpetrator)overtowatchTV;perpforcedsexualintercourseandforcedcompltoperformfellatio.
Thevictimsbehaviorisnowtheleadingelementincasefile.Theinitialcasedispositionwastolocate,
meaningthattheassailant(whoisknowntothevictim),neededtobefoundbypoliceandinterviewed.
Thecasewasthenhandedofftoasexcrimesunitinvestigator(thirdsection,handwrittenin
cursive),whoquestionsthevictimsstatement(statementandtimeframehassomeholesinit).The
victimprovidedadditionaldetailsabouttheperpetrator,namelythathehadbeenarrestedbefore,but
theinvestigatorcharacterizesthevictimashostilewhenaskedformoreinformation.Giventhatpolice
personnelacknowledgedthattheywillpurposelyaskquestionsinintimidatingandchallengingwaysin
ordertodiscouragevictimscontinuedengagement,itseemsreasonabletowonderwhathadtranspired
117
inthisexchange.Intheend,thefinalcasedispositionchangedtounabletoestablishelementsofthe
crime.Thereisnodocumentationinthefiletosuggestthatofficerseversearchedfor,contacted,
and/orinterviewedthesuspect.
EXHIBIT3.3SexualAssaultPoliceReportofaKnownOffenderAssailant
EstablishingVictimInvited
PerpetratorIntoHerHome
QuestioningVictims
Truthfulness
VictimLabeledasNon
CooperativeorBecomingNon
CooperativeinResponsetoLaw
EnforcementTreatment
118
Throughoutthisanalysisofpoliceattitudestowardssexualassaultvictims,wehavehighlighted
casestudyreportstoillustratehowthesebeliefsappeartobeenactedinpractice;however,itis
importanttoacknowledgethatpolicereportsdonottellthefullstoryofaninvestigationeitherfrom
theinvestigatorspointofvieworfromthevictims.Yet,whatwasexpressed,clearlyandfrequently,in
thereportswereviewedwasawidespreaddisbeliefofvictims,particularlythosewhomighthavebeen
involvedinsexwork,thosewhowereadolescents,andthosewhoknewtheiroffenders.
The Police and the Crime Lab: Intra-Organizational Practices & Communication
Thecrimelabistheorganizationmostproximaltothepolicesexcrimesunitinthissystemic
network,asbothwereunitswithinthepolicedepartment.Investigatorsandcrimelabscientists
interactedfrequentlyandbothpartiesnotedthattheywereinregular,casebycasecommunication
aboutSAKs.Basedonstakeholderinterviewswithcrimelabpersonnelandpolice,wedocumented
threethemesintheircommunicationsregardingSAKsubmission.First,bothentitiesconfirmedthatlab
personnelroutinelytoldpolicethattheywouldnotacceptakitfortestingunlessitwascompleteand
properlysubmitted.Crimelabpersonnelexpressedconsiderablefrustrationregardinginvestigators
lackofcompliancewithwhattheyconsideredtobebasicprinciples,suchasproperlysealingevidence:
Theybringstuffin,itsnotproperlysealed,andyousay,Icanttakethislikethis...Ourpolicy
iswhenitcomesthroughthedoor,itmustbeproperlysealed...[ourpolicies]maybecontrary
towhatyourpoliciesare,but,weretheonesthataregoingtotaketheheatifthisiswrong.
However,therealstickingpointseemedtobelabpersonnelsrequirementthatthekitwascomplete:
bycomplete,theymeantthatiftheyweregoingtobetestingforDNA,thentheywouldneeda
119
referencesample,sopolicewouldneedtoobtainandsubmitabuccalswabfromtheperpetrator
and/oranyrecentconsensualsexualpartner(s).59Asoneforensicscientistexplained:
Icouldntgetthrough[tothem]Ihavetohaveareferencesample...Imnottestingitunless
Ihaveit...you(thepolice)needtogogetthereferencesamplebeforeIcandomyjob.And
theyjustwouldntdoit.
Inpractice,itappearsthatcrimelabpersonneldidsometimesacceptkitswithoutreferencesamples
andbegantesting,buthowoftenthatoccurredwasnotsomethingwewereabletodetermine.Whatis
clearisthatsexcrimeunitinvestigatorsweretoldrepeatedlythatSAKsubmissionrequirednotonlythe
kititself,butalsosupplementalsamplestobeobtainedbythepolicethemselves.
Asecondcommonthemeinthecommunicationbetweenpoliceandlabpersonnelfocusedon
thevolumeoftestingthatthelaboratorycouldreasonablyhandle.Crimelabstafftoldsexcrimes
investigatorsthattheydidnothavethepersonnelcapacitytotestallSAKs.Testingwasalimited
resource,tobeusedonlywhenneeded,asoneforensicscientistexplained:
Ifyougaveusalegitimatereasonforwhyyouneededittested,itgottested.Itgottestedeven
ifitprovedwewereright,itsallvictim(theDNAinthekitisonlythevictims).Tomethatwas
oneoftheadvantagesofworkingcloselywiththeinvestigatorsinasmallsettinglikethis
becausewegottoknowthem,wegottoknowhowtheyworked,theygottoknowhowwe
worked,andwhatwecoulddoandwhatwewouldntdo.
59
Theresearchteamwasnotabletoverifytheexistenceofanywrittendocumentationstipulatingtheserequirements,though
bothpolicedepartmentlabpersonnelandpoliceconfirmedthatthiswastheirjointunderstandingofSAKsubmissionpractices.
FBIguidelinesregardingCODISuploadinghavechangedovertimesuchthatreferencesamplesarenolongerrequired(though
agenciesmustremoveaprofileimmediatelyifitlaterbecameknownthatthattheprofilewasfromaconsensualpartner).
Despitethischange,policeofficialscontinuetostatethatreferencessamplesarerequiredpriortosubmittingakitfortesting.
120
Crimelabstaffemphasizedtosexcrimeunitinvestigatorsthattestingwasnotsomethingthatcouldor
shouldbedonewitheverykit.Labpersonnelwereacutelyawarethattherewouldalwaysbeurgent
casesthatwoulddemandimmediatetesting,sotheyalwayshadtojuggletheirworkload:
Theinvestigatorsunderstandthatthelabis...itsalimitedresource,andattimestheyneedto
callin,Idontwanttocallthemfavors,buttheyunderstandthattheyneedtousetheresource
wiselybecauseatsomepointinthefuturetheyreallyneedtoprioritizesomething....This
onesnotasbigofapriorityassomethingelse;Imgoingtogoontoanextcase.Andthenallof
asuddeniftheyhavesomethingthatreallyseemslikeitsareallybadcrimethenthatstheone
thatgetssubmitted.
Duetoitsownresourceconstraints,thelabwasconsistentlygivingthemessagethattheycouldnot
keepupwithdemandandthevolumeofcaseworkgeneratedbyacityofthiscrimerate,andthiswas
particularlyprevalentintheyearswhenthelabwasseekingaccreditation(20022005,andparticularly
from20042005,accordingtostakeholdersinbothorganizations).Asoneforensicscientistnoted,
one[DNAscientist]wasprettymuchentirelyonaccreditation...functionally,thoseyears,itslikewe
[had]lostaposition.Similarly,anothernoted,duringthen(thepushforaccreditation)itwasmore
impossiblethanusual[tokeepup]...andyeah,wetoldthem(thepolice)that.Thesequalitativedata
areconsistentwiththequantitativemodelingresultspresentedearlierinthischapter.SAKsubmissions
dippedintheseyears,eventhoughthelabhadCODISaccessduringthistime(whichpresumablywould
haveincreasedsubmissions,giventhegreaterutilityoftestingtoinvestigators);however,the2004
federalrequirementforindependentaccreditationwaschallenging,giventhelabsstaffinglevels,and
themessagedontsenduseverythingbecameparticularlyprevalentduringthistime.Labpersonnel
emphasizedthatthisbotheredthemasforensicscienceprofessionals,knowing,asonesaid,whatDNA
coulddotohelpthesecases,thesevictims,buttheydidnothavethecapacitytotestallSAKs.
121
Athirdandfinalcommonthemeinthecommunicationbetweencrimelabpersonnelandpolice
centeredonwhichSAKsshouldbetested.Policecrimelabstaffemphasizedthattheyhadlimited
resourcesforSAKtesting,andtheydidnotwanttheirtimewastedonkitsthatshouldntbetested.
Onestakeholderexplainedwhats/heviewedasthelabsongoingdilemma:
Soyoubringinthiskit(onebelievednotworthyoftesting)andinthemeantime[we]gota
horrificmurder,aserialrapist,limitedresources,limitedpersonnelandwereexpendingour
energyonthis...andifthevictimskindofshady...Justbringustherealones.
Itappearsthenthattheattitudesandbeliefsamongcrimelabpersonnelweresimilartothoseofthe
policeinthatvictimssuspectedofprostitution,adolescentvictims,andvictimsofnonstrangerrape
werenotdeemedcredibleand/orworthyofinvestigationalandtestingresources.Ofcourse,wecannot
disentanglewhethertheattitudesexpressedbycrimelabpersonnelwereduetomessagestheyhad
receivedfromthepoliceovertheyears(eitheratanindividuallevelorattheinstitutionallevel,given
thattheywerepartofsameorganization),orwhetherthesewere,moreorless,theirownbeliefsabout
victims.Attheveryleast,bothcrimelabpersonnelandpoliceconfirmedthattheyregularlydiscussed
howsomevictims,somecaseswerenotworthyoftheinvestmentoftesting.
Giventhesecommunicationswiththecrimelabpersonnel,howdidpolicehearthesemessages?
Howdidtheyinterprettheseremarks,inlightoftheirownresourceconstraintsandbeliefsabout
victims?AsshowninFigure3.5(nextpage),policepersonnelsaidthattheyheardthesemessagesas
confirmationthatnotallSAKscouldorshouldbetested.Investigatorscharacterizedthelabsmessage
thatkitsshouldnotbesubmittedunlesstheywerecomplete(i.e.,policehadobtainedreference
samplesfromsuspectsand/orrecentconsensualpartners)aspushback,themputtingmoreworkonus
tokeeptheirworkdown.Anotherinvestigatorexpressedfrustrationaboutthelabspracticeonthis:
gettingthereferencesample,Idonthavetimetogochasethatdown,cantwegetstartedwithout
out?Theysaidno.Policeofficialsnotedthatthesemessages,overtime,likelyhadanegativeeffecton
122
SAKssubmission,particularlyamonglowperforminginvestigators,whomayhaveshruggedoffthelabs
messagesandmayhaveuseditasanexcuse,asonesupervisorexplained:
Theyhearthat(thereferencesamplesrequirement)andsaytothemselves,thistakestoomuch
time...efforttosubmitakit,andthey(thelab)nevertakethemanyway,sowhybother?
AlsoasshowninFigure3.5,policeheardthelabsmessagesaboutlimitedcapacity,asonepoliceofficial
stated,theytoldustheycantdoitall,sowedidntsenditall.Therefore,thecommunicationwiththe
labreinforcedthepolicedepartmentsmindsetthattheycouldntorwouldntbeabletopursueall
cases,andtheattitudesexpressedbycrimelabpersonnelregardingcertainkindsofvictims/cases,
bolsteredpoliceopinionsaboutwhichcasesshouldbepursuedandwhichonesshouldnt.
FIGURE3.5IntraOrganizationalCommunicationBetweenthePolice&CrimeLab
123
The Police, the Crime Lab and the Prosecutors Office: Inter-Organizational
Communication
Fromourinterviewswithstakeholdersacrossthesethreeorganizations,weidentifiedthree
commonthemesinthefrontlinecommunicationsbetweensexcrimeunitinvestigators,crimelabstaff,
andassistantprosecutingattorneys(APAs)regardingsexualassaultcasesandSAKsubmissions.First,
theAPAsnotedthatwhentheyneedakittesteddependedonthecircumstancesoftheparticularcase
athand.Forexample,considerthisseriesofquotesfromvariousAPAsthatoutlinedifferentscenarios:
Sometimeswewantit(thekit)...processedevenbeforeweissueacase.Thathappens
occasionallywherewereallyneedtoknowwhatsinthatkit.Therearesomekitsthatwemay
needtestedbeforewecangoforwardwithprosecutionbecauseIDmaybeanissue.
Themajorityofourcasesareissuedwellbeforeweknowwhatsinthesexualassaultkit.
Itshouldbeanautomaticthing(sendingthekittothelab),ifIvechargedsomebody,theresno
reasonwhythatkitshouldntbesenttothelab.IfIchargedthisdefendant,thatshouldbethe
firstthingthathappens.
[Ifacaseisgoingtotrial]...thatkitneedstobetested...becausewhatsgoingtohappenis
thedefenseisgoingtothrowout,welltheydidnteventesttheevidence...Ineverknowwhat
thedefenseisgoingtobe.
Thesequoteshighlightthatitdependsthereisntaonesizefitsallapproachbecausesometimesa
kitneedstobetestedinordertoidentifythesuspecttobecharged;sometimeskittestingisnecessary
torebutadefenseofnocontactbetweenthevictimandoffender;sometimesthekittestingsimply
confirmsthattherewascontactsothatthecasecanproceedtoestablishotherelementsofthecrime
(e.g.,lackofconsent).Thus,themessagefromtheAPAswasthatSAKtestingneededtobeevaluated
onacasebycasebasis,dependingonthecircumstancesofthecase.
124
Second,APAsnotedthatwhentheyneededakittested,theyoftenneededitimmediately.If
testingwasbeingconductedtoidentifyanoffender,therewasapressingpublicsafetyconcernthat
necessitatedquickturnaround.Iftestingwastobepresentedincourt,theywerefacingamandated90
daycourttimelineandjudgeswhowouldnotgivecontinuancesforSAKtesting.60AsoneAPAexplained:
Therearetimeswhenthelabhasdroppedeverythingtheyweredoingandrushedtoprocess
rapekitsbecauseyoujustneedacaserushed.Thejudgeisntgoingtogivetheadjournment.
Indeed,policeandcrimelabpersonnelalsoexpressedfrustrationatthejudiciaryforwhatfelttothem
likeimpossibledeadlinesandexpectations.But,intheend,theinvestigators,forensicscientists,and
APAsknewthattheyweregoingtohavetojuggletheirworkloadstoaccommodateurgenttesting
requests.OneAPAdescribeditassuch:
They(thepoliceandlab)knewtherewasalwaysthepossibilityofthecall(Q:Thecall?Whatdo
youmean?)Thecallfor:findit(thekit),testit,weneeditnow.Thatwasthereality...weall
hadtoaccommodatethetrialschedule.
ThethirdcommonthemeincommunicationsfromAPAstothepolice(andtoalesserextent,to
crimelabpersonnel)isthattheymightnotapproveawarrantrequestandmoveforwardwithacaseif
therewereseriousconcernsaboutvictimscredibility.Inthesetwoquotesbelow,APAsnotedthat,
dependingonthecircumstancesofthecase,itcanbedifficulttodecidehowtoproceed:
60
Briefly,theprocessforfelonycasesinthisjurisdictionisasfollows:1)arrestwarrant;2)arraignmentonthewarrant;3)
preliminaryexaminationtoshowprobablecausethatacrimewascommittedandthedefendantcommittedit;4)ifthe
prosecutionmeetsitsburdeninthepreliminaryexamination,thecaseisboundoverfortrial;and5)arraignment,whichstarts
the90daytimelineandallothersteps(e.g.,pretrialmotions,subpoenas,jurytrials)mustbecompletedwithin90days.
125
Whenyoulookatit(thewarrantrequest)andyouseesomebody(avictim)whohasaprior
(criminaloffense)...Whetherornotweshouldgoforward[canbe]hardtodecide.
Ifyouhaveawomanwho'sgoingouttoabaranddrinkingandthengoesbacktoahousewith
aguy,[itdoesnt]automaticallymean...shewantstohavetosex...Evenifshewentback
andconsideredit,andsaidno,knowingnotechnicallymeansno,butthat'snotpeople'strue
belief.Thecaseisgettingreallydifficult.
TheAPAsalsonotedthattheydifferamongthemselvesastowhethertheywouldissueawarrantwhen
thevictimscredibilitywasquestionable,andthatbeforetheSexualAssaultTeamwasformed,there
wasevenmorevariabilityacrossAPAs,asthisquoteillustrates:
Wehavedifferentopinions[aboutprosecutingwhencredibilityisquestionable].Therewere
othersthatsaid,Iwouldnttrythat,donttrythat.Sowedifferamongstourselves.
Thoughtheexamplesquotedabovehighlighthowthegeneralpublicmightdoubtvictimswhohave
priorcriminalrecordsand/orthosewhohadbeendrinkingatthetimeoftheassault,theAPAs
emphasizedthattheirconcernsaboutcredibilitywerenotrootedincommonstereotypesaboutvictims;
indeed,theynotedthatsomevictimsarespecificallytargetedbyrapistsbecausetheywillnotbeseenas
crediblebythecriminaljusticesystemandthepublicasawhole.Rather,theyemphasizedthattheir
concernsaboutcredibilitywerebasedonwhethertherewereseriousconcernsabouttheveracityofthe
victimsstatementandwhethertheycould,consistentwiththeirresponsibilitiesasministersofjustice,
moveforwardonacaseinwhichthereweredoubtsaboutthefactsathand.
TheoverarchingmessagefromtheAPAsovertimewasitdepends:howtoproceedinasexual
assaultmustbeevaluatedonacasebycasebasis.Sohowthendidthepoliceinterpretthismessage,
giventheirownbeliefs,timepressures,andresourceconstraints?Figure3.6(nextpage)highlightshow
policeappeartohaveheardthesemessagesandhowthoseinterpretationsreinforcedtheirownbeliefs
126
andpracticesregardingsexualassaultinvestigationsandSAKsubmissions.Policeinvestigatorsnoted
thattheitdependsmessagewasconfusingandcontradictory,asthesetwoquotesillustrate:
Itwasntcleartouswhattheywantedwithtesting.Theytoldusdifferentthingsallthetime.
Therewasntaruleaboutit.Sometimeswegottodecide,sometimeswedecidedtogether,
sometimestheycalledusandsaid,testit,now....so,no,Ididntalwayssendeverythinginfor
testingbecauseIdidntknowwhetherthey(theprosecutors)wantedit...Ifigured,iftheyneed
it,theyllbecalling.
Nowweretryingtosortoutwhatwentwrongandmakingnewpolicies,buttruthis,partof
howthishappenedisthatitwasntclearwhowasdecidingandwhichonesshouldbetested.
FIGURE3.6InterOrganizationalCommunicationBetweenPolice,CrimeLab,&ProsecutorsOffice
127
Whenkitsweretobetested,theweneeditnowproblemcausedlongtermtensionsbetweenthese
threeorganizations,asoneAPAnoted:
[Thecrimelab]says,dontsendusarapekitunlessyouresureyouneedthekitanalyzed
becauseifyousendusoneanditturnsoutwevedonetheanalysisbuttheguysgoingtoplea,
thatstimeandeffortwespentonarapekitthatkeptusfromgettingtoonewherewegottaget
ittocourtandthejudgeisscreamingbecausewerenowlate.
Similarly,fromthepoliceperspective,thepushtogetkitstestedfortrialmadeitdifficultfor
investigatorsandcrimelabstafftomanageandprioritizetheirworkloads:
Therewereinstanceswheretheprosecutorsdidntwantthekittestedortheylltelluswhena
kitistobetested.[Sometimes]tomeitseemedliketheprosecutorswantedakittestedfor
leverageinacourtproceeding...togetthedefendanttocopaplea,whichmakesmorework
forallofus...Thenthey(theprosecutors)calltosay,wellwedontneeditnow.
Q:Howdoesthataffectyourrelationshipwithpeopleinthecrimelab?
Itsstrained...Therelationshipisstrainedallthewayaround,thelab,sexcrimes,prosecutors
office.Itsstrainedbecausenoonetrustseachother.Everyonepointsthefinger,everybody
placestheblame.Thiswholethingisflawedandnoonetrustseachother,noonewantstobeon
thesamepage.
Thechronicstrainbetweentheseorganizationsappearstohavebeenfurtherexacerbatedby
howpoliceinvestigatorsinterpretedAPAsdecisionsnottomoveforwardinsomecases.Fromthe
policepointofview,theAPAsbeliefsweresimilartotheirown,asonestakeholderdescribed:
They(theprosecutors)saidittoo...theycanpretendtheydidnt,buttheydid...theysaid
theycantwarrantacaseifshesbeentricking(engagedinprostitution)orusingcrackor
whatever...soweknewnottobotherwiththose.
128
Similarly,anotherpoliceofficialnoted,everyoneknewtheywouldnttakecertainkindsofcases,sowe
didntpresentthem...they(theprosecutors)werentlookingforextraworkeither.However,the
prosecutorsstronglydisagreedwiththenotionthattheirbeliefswereconsistentwiththepolice,asone
APAnoted:Ithinktheyheardwhattheywantedtohear...notwhatweactuallysaid.AnotherAPA
elaboratedonthedifferencesbetweenthepoliceandprosecutorialviewonvictimcredibility:
ItsacompletelydifferentthingtowriteoffacaseasCRPT(complainantrefusedtoprosecute)
becauseyouthinkthinkshemightpossiblybeeninvolvedinprostitution...tobasicallytossit
inthetrashandneverevenpresentittoanAPAforconsideration...thatstotallydifferentthan
makinganinformeddecisionthatyoucantwarrantafterathoroughinterviewwiththevictim.
..andconsiderationofalltheevidence(emphasisinoriginal).
Forpolicewhoweredisinclinedtobelievesomevictims,orperhapsmanyvictims,theAPAsmessages
aboutvictimcredibilityappeartohavebeenheardinwaysnotconsistentwithhowtheywereintended,
andnegativelyaffectedhowinvestigatorsapproachedcertainkindsofsexualassaultcases.
The Police and the Medical System: Minimal Inter-Organizational Communication

Asdescribedpreviouslyinthischapter,beforetheDetroitareaSANEprogramwasfounded,
therewasnoregularcommunicationbetweenthepoliceandmedicalcommunityregardingSAKs,which
weconfirmedinourstakeholderinterviews.Representativesfromboththepolicedepartmentandthe
medicalcommunitystatedthattherewasminimalcontact,butwhatcommunicationtherewasfocused
ontwocentralissues.First,medicalprovidersnotedthatpriortotheSANEprogram,sexualassault
casesweresimplynotapriorityinDetroitareahospitalemergencydepartments(ED),amessagethat
theycommunicatedtovictimsaswelltopolice:
Wewouldtellpolice,dontbotherwaitingaroundforit(thevictimandthekit)...wellgettoit
whenwegetittoit...wehavetodotherealemergenciesfirst.
129
Sexualassaultpatientstypicallydonotsustainserious/lifethreateninginjuries(seeTjaden&Thoennes,
2006),whichiswhatEDstaffconsideredtoberealemergencies.Thesecaseswerealsonotprioritized
becausemedicalprovidersdoubtedtheusefulnessoftheSAKtothepolice,asonephysiciannoted:
Itoldthem(thepolice)therewasntmuchwecoulddo,butweddothekit,notthatitwould
helpthem[intheirinvestigation]...itwasntapriorityforus,knowingthatitwasntgoingto
makeadifferenceanyway...gettingthem(victims)theMorningAfterPilland[STI]prophylaxis
wastheimportantthing.
Whenaskedtoclarifythecommentabouttheexam/kitnotmakingadifferenceanyway,this
stakeholderwentontoexplain:
Thekitjustestablishesthattherewascontact...sexualcontactbetweentwopeople...
doesnttellyouwhetheritwasrape...andthatswhatthepoliceneedtoknow...Idontthink
itsallthathelpfultothem.(Q:Justtoclarifyabitmore:Didtheytellyouitwasnthelpfulto
themordidyoucommunicatetothemitsnothelpful?)Probablyabitofboth...butmoreus
(doctors)givingthem(thepolice)ourprofessionalopinion.
Whenpressedforthebasisofthisopinion,thisstakeholdersaiditwascommonknowledgethatthe
sexualassaultkitsarenotuseful.
Second,medicalpersonnelemphasizedthatgiventheirownworkloadanddemands,theydid
nothavetimetoremindpolicetocomepickuptheSAKs.Stakeholderscommentedonhowtheyoften
forgottocallthepolicetoletthemknowtherewerekitstoberetrieved,asthesetwoquotesillustrate:
Therewasalsosortoflackofcommunicationwithlawenforcementwithregardstowhenakit
waseventheretobepickedupandso,orsomeoneyouknowwouldforgettocalltohavepolice
pickupandsokitswouldsitthereforyouknowlongperiodsoftimebeforepolicewouldpick
themup.
130
Imeantherewasntanyconsistentcommunication.ItwaswhateverERdoctororresidentdid
thekit,prettymuchatthatpointthekitwaslockedupwhereverthedesignatedplacewasand
thenlikethenursecalledthepolicelaterthatday,sometimesrightatthatpoint,sometimesyou
knowthenextdayoraweeklaterwhensomeonerealizedsomeoneforgot.
Medicalpersonnelchidedthepolice,tellingthemthattheydidnthavetimetobugthemabout
pickingupthekits,asonehospitalEDstaffmemberexplained:
WhenIdseethem(thepolice),sometimesIdsay,hey,youguysgottacheckinwithus,we
donthavetimetochaseyoudowntocomepickupthekits...sometimesIdjokewiththem...
hey,Idontwanttotestify...soyouknow,donthurryonthese.
Thisquoteunderscoresmedicalprovidersambivalenceaboutbeinginvolvedinlegalproceedings,which
mayhavebeenanotherreasonfortheirlackofengagementandcommunicationwithpoliceaboutSAKs.
Howwerethesemessagesheardandinterpretedbypolice?Priortotheimplementationofthe
SANEprogram,theindifferenceofthemedicalstaffaboutsexualassaultpatientswasobvioustopolice:
[thesecases]werentabigdealforthem,weknewthat,itwasobvious.AsnotedinFigure3.7(next
page),theircommunicationsthoughbriefandinfrequentreinforcedbeliefsthatthesecaseswere
notahighpriorityandtheutilityoftheSAKwasunclear,atbest;pointless,atworst.Ifphysicianswere
tellingpolicethatintheirprofessionalopiniontheSAKwasnothelpful,itseemslikelythatsuch
communicationsmayhavedeterredpolicefrompickingupkitsandsubmittingthemforanalysis.Asone
policeofficialstated,hey,ifthedocssaytheresnothinginitwhoarewetosayotherwise?Medical
systemstaff,inboththeirdisengagedbehavior,andtheirwords,communicatedthattheydidnotwant
tobepartofsexualassaultcases,whichalsoleftpolicequestioningwhethertheywouldevenbethere
forthem,fortheprosecutorsintheeventacasewenttotrial.Overall,themedicalsystemspervasive
indifferenceandoccasionalactiveresistancetoengaginginthesecasesappearstohavereinforced
policebeliefsthatsexualassaultisnotapriorityandthatSAKshavelimitedutility.
131
FIGURE3.7InterOrganizationalCommunicationBetweenthePolice&MedicalSystem
Bycontrast,whentheSANEprogramemergedin2006,policebeganhearingaradicallydifferent
message,namelythatmedicalforensicevidencecanbeinordinatelyhelpfultotheirinvestigationsand
thatSANEpractitionerswereoncalltohelpnotonlysexualassaultpatients,butotherpractitionersin
thecommunitywhowantedtheirprofessionalconsultation.Asoneinvestigatornoted:
Itwastotallydifferentafter[SANE]...goodserviceforvictims...andthenursesknowwhat
theyredoingandwealwaysknewwecouldreachthemforwhateverweneeded.
Thesedataareconsistentwiththequantitativeresultspresentedearlierinthischapter,whichshowed
thatSAKsubmissionssignificantlyincreasedpostSANE.Afteryearsofminimalcommunicationwiththe
medicalsystem,theopendoor,collaborativeapproachoftheSANEprogramappearstohavemadea
keydifferenceinthecommunity.
132
The Police and Victim Advocacy Organizations: Intra- & Inter-Organizational

Communication
Thepolicedepartmenthadalongstandingsystemsbasedvictimadvocacyprogram,staffedby
MSWleveladvocates.Inourstakeholderinterviews,policepersonnelandadvocatesnotedthatthey
wereinregularcommunicationandthattheadvocateshelpedtheinvestigatorsasneeded;however,
neitherthepolicenortheadvocatesprovideddetailsorexamples.Whenaskedabouthowtheyworked
together,whattheysaidtoeachother,howtheycoordinatedtheirefforts,thepoliceandtheadvocates
reiteratedthattheytalkedregularlyandworkedtogetherfrequently.Whenaskedwhetherthe
communicationsbetweenvictimsandtheadvocatesweresharedwiththepolice,wereceived
conflictinganswers,butultimately,policepersonnelclarifiedthatadvocatesdidprovideinformationto
investigatorsregardingwhatvictimshadsaidtothem.Whenaskedtogiveexamplesastohow,when,
andwhythisoccurred,boththepoliceandtheadvocatesdeclinedtoprovidethisinformation.Assuch,
wewerenotabletoidentifyconsistentthemesinthecommunicationbetweenthepoliceandtheir
victimadvocacyprogrambecauseneithergroupprovideddataastothenatureoftheirworktogether.
Oneformofcommunicationthatmightbeexpectedtooccurisadvocacyonbehalfofaclient
checkingonthestatusofacase/kit,nudginginvestigatorsifthingswerenotmovingalong(ifavictim
wantedtopursuecriminalinvestigationandprosecution),explainingtoinvestigatorsthatavictimdid
notwanttopursuethematter,challengingpolicepersonnelregardinghowtheyrespondedtoavictim,
andsoforth(seeMartin,2005).Whenweaskedthesystemsbasedadvocateswhethertheyengagedin
suchactionsfortheirclients,programpersonnelwereequivocal:
Willadvocatesaddressanofficerandsay,thisisinappropriate?Somewouldandsome
wouldnt.Maybetheneweronesandtheyoungeroneswouldbealittleintimidated.
133
Whenweaskedpolicepersonnelwhetheradvocatesengagedinclientadvocacy,onesaid,Noofcourse
not,theyworkforus.Othermembersofthecollaborativecommentedonthechallengesposedbythe
organizationalrelationshipbetweenthepolicevictimadvocacyprogramandthepolicedepartment
itself,asonestakeholderexplained:
Howcouldthis(thelargenumbersofuntestedSAKs)happenandthey(thepoliceadvocacy
program)didntknow?...Victimshadtobecallingthemafterwards...saying,Ihaventheard
anything,nobodysgottenbacktome.Whatsgoingon?Andthatsagainaperfectexampleof
whenthecomplaintisaboutyouremployer,howfarcanyoupushit?
AsshowninFigure3.8(nextpage),wedonotknowwhatmessageswerebeingcommunicated
fromthesystemsbasedadvocatestothepolice,andwhateffect,ifany,thathadonpolicepracticesin
sexualassaultcases.Basedonthedataavailable,itappearsthatthelackofadvocacyonbehalfof
clientsmayhavebolsteredpolicebeliefsthattheirapproachtosexualassaultinvestigationswas
reasonableandthatsomecases/kitswerenotworthpursuing.
Withrespecttocommunitybasedadvocacy,wewerealsounsuccessfulingleaningmuchinsight
intotheinteractionsbetweenthoseadvocatesandpolicepersonnel,thoughfordifferentreasons.As
notedpreviously,therewasonlyonepaidstaffpositionthroughoutmostofthe2000s,andasone
stakeholdernoted,Onestaffpositioncantdomuch...certainlycantchangewhatwashappeningin
[thepolicedepartment]allthoseyears.Theabsenceofastrongcommunitybasedadvocacyprogram
mayhaveindirectlycontributedtothelongstandingproblemswithhowpolicerespondedtosexual
assaultcases,asthesetwoquotesillustrate:
Therewerenocommunityornonprofitgroups[providingsexualassaultservices],itwasjust
easytoturnablindeyeandnotpayattention...[Ifthere]hadtherebeenastrongcommunity
basedsexualassaultprogramthiscouldnothavegoneonforasmanyyearsasitdid.
134
Victimswillsaywhyisntmycasegoingforward?Whyisnobodycallingmeback?Whats
happening?Andthatspartofthecommunityadvocateistoberattlingchainsandtosay,what
theheck,whatsgoingonhere?[Ifwehadcommunityadvocacyprograms]theywouldvebeen
hearingfromvictimsandtheywouldvebeenaskingquestionsof[thepolice]onaconstant
basis.
FIGURE3.8Intra&InterOrganizationalCommunicationBetweenthePolice&VictimAdvocacy
AsshowninFigure3.8,theabsenceofmessagesfromthecommunitybasedadvocacymovement
appearstohavebeendetrimental,suchthatpolicebeliefsandbehaviorswentlargelyunchallengedfor
decades.
135
Summary & Conclusions: Why So Many Unsubmitted SAKs in Detroit

Thepolicedepartmentstruggledwithchronicresourcedepletionforyears,whichmadeit
challengingforthesexcrimesunittomaintainstableleadership,supervision,andtraining.Police
stakeholdersnotedthatbecauseofthesebudgetandstaffingcuts,investigationqualitywas
compromisedandcuttingcornersbecamenormative.However,chronicresourceproblemsarenot
theonlyreasonwhysexualassaultinvestigationswereroutinelyshortshrifted.Consistentwithprior
researchonchronicscarcity,therewasclearevidenceofpolicetreatingvictimsindehumanizingways:
inboththestakeholderinterviewsandinactualpolicereports,lawenforcementpersonnelregularly
expressednegative,stereotypingbeliefsaboutsexualassaultvictims.Victimswhowereassumedtobe
prostituteswereconsideredtobeatfaultforwhathadhappenedtothem.Adolescentswereoften
assumedtobelying,tryingtoavoidgettingintotroublewiththeirfamiliesbyconcoctingafalsestory
aboutbeingraped.Friends/acquaintanceshadgotwhattheygotbecausetheyhadchosentoassociate
withtheperpetrator.Thefactthatallofthesevictimshadenduredalengthy,invasivemedicalforensic
examseemedtocarrylittletonoweight.BecausesomanyofthesesurvivorswereAfricanAmerican
women,manyofwhomwerelivinginpoverty(basedonDetroitcensusinformation),itisdifficultto
assesstheextenttowhichpolicealsoheldthesebeliefsforwomenofdifferentraces/ethnicityand
socioeconomicstatuses.Attheveryleast,thepoliceappearedtohavenocompunctionexpressingsuch
opinionsaboutAfricanAmericansexualassaultsurvivors.
Giventhesefindings,thenotionthatresourcedepletionisthesolecauseofthisproblemisnot
empiricallysupported.However,itislikelythatresourcestrugglesandnegativeattitudesbecamea
reinforcingnegativedynamic:becausethepolicedepartmentwaschronicallyunderstaffed,theyfelt
theycouldnotinvestigateallreportsthoroughly,sosomecaseshadtobeweededout,amindset
commoninconditionsofchronicscarcity.Sowhichonesshouldbeweededout?Casesthatwere
perceivedashard,complicated,timeconsumingand/oronesinwhichinvestigatorsstereotypesmade
136
themquestionthevictimscredibility.Withoutconsistentsupervisionandtrainingtochallengethese
practices,labelingacaseasadealgonebad,orotherwisedismissingitasnotreallyrape,wouldlikely
notbechallenged.
Policeoftenreferredtothisprocessasprioritizingortriagingcases.Thetermprioritize
meanstoarrangeordealwithinorderofimportanceandtriagemeanstoassignorderbasedon
urgency.Ouranalysescannotpinpointthedefiningfeaturesofcasesthatweredeemedimportantor
urgent,asthisisanalysisofwhatwasdeemednotimportant.However,thevastmajorityofcases
reportedeachyearappeartohavebeendeemednotimportant.Moreover,prioritizeandtriageimply
that,eventually,intheend,alltaskswillbecompleted.Therewasnoindicationthatpolicepersonnel
hadthatmindset.Rather,basedonthesedata,itseemsmoreaccuratetosaythatpolicewere
selectingcases,someofwhich(theminority)wereconsideredworthyoflimiteddepartmental
resourcesandtherest(themajority)werenot.Thecasesnotselectedwerenotpursuedandthekits
associatedwiththosecaseswereshelved.
IntheirinteractionswithotherDetroitorganizations,policeheardwidelydifferingopinions
aboutSAKsandtheirutilitytosexualassaultinvestigations.Fromthemedicalsystem,theyweretold
thatsexualassaultcaseswerenotapriorityandtheSAKitselfwaslikelynotgoingtobehelpfultothe
investigation.Inotherwords,medicalproviders,whowereostensiblytheexpertsonthekitsandtheir
contents,weretellingpolicethattheSAKswerenearlyuseless.Attheotherend,theprosecutorswere
tellingpolicethatkitswereessential,oftendropeverythingessential,butcircumstancesvariedacross
casesandtestingneededtobeevaluatedonacasebycasebasis.Inbetweenthosetwoextremes,
investigatorscolleaguesinthecrimelabwereconsistentlytellingthemthattheydidnothavethe
capacitytotestallSAKstestingwasalimitedresource,nottobeusedonbreadandbuttercases,as
onestakeholdernoted.LabpersonnelwerealsotellingthepolicethatSAKsubmissionrequiredmore
137
thansimplydroppingoffthekitfortestingadditionalinvestigativelaborwasnecessarytotrackdown
suspectsand/orconsensualpartnerstoobtainDNAreferencessamples.
Thesemessages,whenheardthroughthepolicesownfiltersofresourcescarcityandnegative
attitudestowardsvictims,reinforcedorcouldbeinterpretedtoreinforcetheirpracticesthatnotall
SAKscouldorshouldbetested.Intheworstcasescenario,itappearsthatmanypoliceheardthese
messagesas:SAKtestingisextraworkanditprobablywontmatteranyway,andIdonthavetimetodo
this,andIdontbelievethevictimandnoelsedoeseither,sowhyinvestinthiscase?Inthebestcase
scenario,someinvestigatorsheardthesemessagesas:SAKtestingcanbeusefulandwehavetojuggle
ourworkloadandbepreparedtostopwhatwedoingatanymomenttoreprioritizeforanurgent
casewhichmeanssomethingelsewontgetdone.Victimadvocacy,bothsystemsbasedand
communitybased,waslargelysilent,anddidnotactivelychallengepoliceonthesebeliefsandpractices.
Fromasystemicperspective,theseorganizations,thoughinterdependenttoeachother,didnot
worktogetherasacohesivewhole,withacommonvisionandplanforSAKtesting,specifically,and
postassaultservicesforsurvivorsmoregenerally.Suchbunkersandsilosbehavioriscommonamong
organizationsworkingunderconditionsofchronicscarcity.Intraandinterorganizational
communicationbecomeslessfrequentandmorestrained,asindividualsstruggletoempathizewith
theircolleaguesandunderstanddifferentpointsofvieworsimplychoosenottounderstanddifferent
pointsofview.Alsoconsistentwithpriorresearchonchronicscarcity,thedepletedresourceswithinall
oftheseorganizationsmadeitdifficulttoassistallvictims,butperhapsmoretroubling,itallowed
negativestereotypesandbeliefstogounchecked,sothatmanysurvivorsweretreatedinrevictimizing,
dehumanizingways.Theseinstitutionalpractices,repeatedincaseaftercaseforthirtyyears,resultedin
substantialnumbersofunsubmittedSAKsontheshelfinpoliceproperty.
138

SincethediscoveryoftherapekitsinAugust2009,figuringoutif,when,andhowtheywouldbe
testedweighedheavilyonlocallevelofficialsandstatelevelpolicymakers.Theurgencywaspalpableat
thebeginningofthisactionresearchproject400SAKshadbeentestedsofar(intheOVWfundedThe
400Project),buttherewerestillthousandstogo.TheDetroitcollaborativereachedouttoother
jurisdictionsthathavehadlargenumbersofuntestedSAKstoseektheirguidanceondevelopinga
testingplan.NewYorkCityandLosAngeleshadhadthefinancialresourcestoforklifttheirkitsall
wereboxedupandshippedenmassetovendorlaboratories.TheDetroitcommunityhadnowherenear
theresourcesrequiredforsuchaplan.PoolingfundsfromtheDetroitSAKARPbudget,thestatepolice
departmentsNIJDNABacklogReductionGrants,andtheresourcesofauniversitybasedforensic
laboratory(whichwasseparatelyfundedbyNIJ),theprojectwouldbeabletotest1,600kitslessthan
20%ofthekitsthatneededtobetested.61Therefore,thechallengebeforetheDetroitcollaborative
wastofigureoutwhattodoifacommunitycanttestalltheirrapekits(atleastinitially).Asone
memberofcollaborativeremarked,Whatsthatexpression,Howdoyoueatanelephant?Onebiteata
time.Ok,sohowdowetakeabiteoutofthiselephant?Wheredowestart?
GiventhatallSAKscouldnotbetestedinthecontextofthisactionresearchproject,theDetroit
collaborativeneededtodevelopaplanfortestingonlysomekits,whichopenedupdifficult
conversationsaboutwhichkitsshouldbeselected,whichinturnheightenedfrustrationsthatselecting
onlysomecaseswaswhatgotDetroitintothisprobleminthefirstplace(seeChapter3:WhySoMany
61
ThisestimateisbasedontheassumptionthatallSAKswithalaboratoryIDnumberhadinfactbeentested(seeChapter2:
HowManyUnsubmittedSAKsinDetroit).Itwasnotclearatthetime(orattheendofproject),howmanyofthoseSAKshadin
factbeentestedforDNA.Therefore,alessoptimisticestimateisthatthefundsavailableintheactionresearchprojectwould
test<10%ofthekitsthatneededanalysis.
139
UnsubmittedSAKsinDetroit).Asonememberofthecollaborativenoted,Weneedtofindasolution
andpickingonlysomekitsisthesameoldsameold...itsnotasolution.Weneedtosolvethis
problem(emphasesinoriginal).Thecollaborativeteamwentroundandround,butalwayscameback
tothehardrealitythatrightnow,inthecontextofthisactionresearchproject,theydidnothavethe
resourcestotestallkits,soitwouldbenecessarytodevelopaplantotestonlysome.So,whichones?
StakeholdershadvastlydifferentideasaboutwhichSAKsshouldandshouldnotbeselectedfor
testing.Forexample,somemembersofthecollaborativearguedthatstrangerperpetratedcrimes
shouldbeselectedfortesting,asDNAtestinghasthepotentialtoidentifytheoffenderand,some
added,strangerrapesreflectedthemostseriousthreattopublicsafety.Othermembersoftheteam
notedthatDNAcouldbehelpfulinnonstrangerrapecasestoo,andthattheseassaultsposedjustas
muchathreattopublicsafety.Somecollaborativemembersadvocatedforskippingoverkitsassociated
withcasesthatwerebeyondthestatuteoflimitationsinfavorofcasesthatwerestillactionable.
However,theprosecutorsnotedthattestingkitsassociatedwithcasesthatwerepresumedtobeSOL
expireddidhavemeritbecauseoncethefactsofthecasewerefullyreviewed,itwaspossiblethatthere
wouldbecircumstancesthatwouldmakethecasestilleligibleforprosecution.Furthermore,evenifthe
SOLhadexpired,iftheoffenderhadreperpetrated,itmightbepossibletoentertheevidenceofthe
priorassaultintoacurrentcase,perfederal404brulesofevidence.
ThesediscussionsanddebateshighlightedthatSAKtestingcouldhavedifferentialutilityto
victimsandtothecriminaljusticesystemdependingonthecircumstancesofthecase(e.g.,victim
offenderrelationship,statuteoflimitations,etc.).ACODIShitinastrangerperpetratedcrimemight
revealtheidentityoftheperpetratorandpossiblyrevealapatternofserialoffending;aCODIShitina
nonstrangerperpetratedcrimecouldconfirmidentityandalsopossiblyrevealserialoffending;aSOL
expiredcasecouldhittoacurrentcase,perhapsofferingthevictimoftheoldcaseachancetotestifyin
courtforthependingcase(ifdesired).TheDetroitteamcouldarticulatemanypossiblescenarios
140
regardingtheutilityofSAKtesting,butnooneknewhowoftenhowprobableanyofthesescenarios
mightbeinpractice.HowcommonisitthatanonstrangercaseyieldsaCODIShit?Howcommonisit
thataCODIShitshowsapatternofserialsexualoffending?HowcommonisitthatapresumedSOL
expiredcaseproducesaCODIShit?NooneknewDetroitorganizationshadnotbeenabletotrackthis
information,datafromNewYorkCityandLosAngeleswerenotavailable,andtherewerenopublished
studiesinthescientificliteraturethatcouldanswerthesequestions.
Therefore,theresearchteamencouragedtheDetroitcollaborativetoempiricallytestthese
ideasaboutSAKutilityunderdifferentcasecircumstances.Inotherwords,wecouldfocusthetestingto
becompletedintheactionresearchprojectonthesekeyquestionsregardingtherelativeutilityofSAK
testingunderdifferentcasecharacteristics.Forthe1,600SAKswecouldtestrightnow,wewouldbe
abletodeterminetheprobabilitythattestingwouldresultinaCODIShit(andpossibleserialsexual
offenderassociations)fordifferentkindsofsexualassaultcases.Suchinformationcouldthenbeusedto
developlongtermtestingplansforDetroit,aswellasotherjurisdictions.Forexample,iftheprobability
ofaCODIShitwassignificantlyhigherforcertainkindsofcasesvs.others,thenthatinformationcould
provideempiricallybasedguidelinesforprioritization(ifprioritizationwasnecessaryduetoresource
constraints);alternatively,iftheprobabilityofaCODIShitwasstatisticallynodifferentforcertaincases
vs.others,thenthatwouldsuggestthatsuchfactorswouldnotbegoodcriteriaforprioritizingcases.62
ThepurposeofthischapteristosummarizehowtheDetroitSAKARPdevelopedatestingplan
toevaluatetheutilityofSAKtestingunderdifferentcasecircumstancesandtopresenttheresultsfrom
62
Somemembersofthecollaborative,particularlythoseinthemedical/nursingandvictimadvocacyprofessions,notedthat
therearemanywaystodefinetheutilityofSAKtesting,andthatCODIShitsareonlyoneway.Theutilitytovictimsandtheir
health,wellbeing,andrecoverymustalsobeconsidered,evenifthesefactorsaremoredifficulttoconceptualizeandassess.
Theothermembersofthecollaborativeagreedandtheresearchteamdidexploreoptionsformultipleindicatorsofutility
includingsomethatwouldtapintothesehealthrelatedoutcomes.However,givenconcernsabouttheresearchteamhaving
contactwithvictimspreadjudication(seediscussioninChapter3:WhySoManyUnsubmittedSAKsinDetroitandChapter5:
Developing&EvaluatingaVictimNotificationProtocol),itwasnecessarytofocusthiscomponentoftheprojectonCODIShits
asanindexofSAKutility.Assuch,thisprojectpresentsonlyoneofmanypossibleoutcomesregardingtheutilityofSAKtesting.
141
thattesting.WewillbeginwithadescriptionofhowtheSAKtestingplanwasdevelopedand
implemented,includingdetailsregardinghowthe1,600wereselectedfromtheSAKscountedinthe
census(seeChapter2:HowManyUnsubmittedSAKsinDetroit).Then,wewillpresentdescriptive
findingsaboutthecases/SAKsthatweretestedinthecontextofthisproject(i.e.,victim,assailant,and
assaultcharacteristics).Thenwewillturnourattentiontotheforensictestingoutcome,beginningwith
anoverviewoftheprocessofDNAtestingandthetypesofCODIShitsthatcanresultfromDNAtesting.
Withthatbackgroundestablished,wewillthenpresentdescriptivefindingsregardingCODIShitrates
andserialsexualassaultrates(fortheoverallsampleandbyTestingGroup).Then,wewillpresentthe
resultsfromaseriesofstatisticalmodelsthatexaminedtheeffectofvictimoffenderrelationshipand
statuteoflimitationsstatusontestingoutcomes.Supplementalanalysesthatexaminedtheeffectof
victim,assailant,andassaultcharacteristicsonforensictestingoutcomeswillalsobesummarized.
Finally,wewillpresenttheresultsfromanexperimentaldesignstudythatcomparedtwodifferent
methodsofDNAtesting.FulldetailsregardingallofthesestatisticalanalysescanbefoundinAppendix
B:ProjectMethodology.
The Process of Developing a SAK Testing Plan
TheDetroitcollaborativestruggledtodevelopatestingplan,giventhecommunityslimited
financialresourcesandthestakeholdersdifferingopinionsaboutwhichkitsmeritedtesting.Therefore,
thegrouphadtoworkthroughanumberofcomplexissues,including,butnotlimitedto:
ShouldallunsubmittedSAKsbetestedIfresourceswerentanissue,shouldallkitsbetested?
Istestingallkitsthelongtermgoal?Thegrouphadtoconsiderthefinancialcostsandpotential
benefits(tovictims,tosociety),thelogisticalcomplicationsofsuchanendeavor,andthe
possibilityofunforeseenunintendednegativeconsequencesforsurvivors.
142
WhatfinancialresourcesareavailabletotestSAKsHowmuchmoneyisavailablefortesting
andhowmanykitswillthosefundscover?Thecollaborativehadtoexaminewhatsourcesof
fundingwerecurrentlyavailableforSAKtestinganddeterminehowmanykitscouldbetested.
WhichkitsshouldbeselectedfortestingIfallSAKscouldnotbetested,thenwhichones
shouldbeselected?Shouldkitsbeselectedatrandom?Shouldkitsbeprioritizedbycertain
criteria(e.g.,victimoffenderrelationship,statuteoflimitations,otherfactors?).Arethere
empiricallybasedguidelinesforprioritization?
WhatDNAtestingmethodwillbeusedtotestthekitsOverthepastseveraldecades,there
havebeensignificantnewdevelopmentsinDNAtesting,sowhichspecifictestingmethod
shouldbeused?Whatarethelegalimplicationsofthatchoice(i.e.,challengestoaparticular
testingmethodincourt?)
WhichlaboratoriesshouldtesttheSAKsCanthetestingbeconductedinhouseatthestate
policeforensicscienceslaboratoriesorwillthekitsneedtobeoutsourcedtoprivatelabs?
Figure4.1TheStepByStepProcessofCreatingtheDetroitSAKARPTestingPlan(following
pages)describeshowtheDetroitcollaborativeresolvedtheseissuestocreateatestingplan.Figure4.1
listseachissuethathadtoberesolved,thediscussionanddebateabouthowbesttoaddresseachissue,
anddecisionsmadebythecollaborative(andwhytheydecidedwhattheydid).Laterthisinchapter,we
willsummarizethechallengestheteamencounteredimplementingthisplan(seeFigure4.3).
143
FIGURE4.1TheStepbyStepProcessofCreatingtheDetroitSAKARPTestingPlan
ISSUE1
SHOULDALLOF
THESAKsBE
TESTED?
DISCUSSION.Therewere
differencesofopinion
betweenandwithin
disciplinesregarding
whetherallSAKsshouldbe
tested,butgenerally,the
prosecutorsadvocatedfor
testingallSAKsbecausethe
resultscouldidentifyserial
sexualoffendersandwould
helppopulateCODIS.
Representativesfromlaw
enforcement&forensic
sciencesexpressedconcerns
aboutthefinancial
investmentand/orlogistical
difficultiesofsuchaneffort.
Somepolicepersonnel
arguedthatnotallshouldbe
tested(e.g.,ifthepolice
reportindicated
complainantrefusedto
prosecute,thenresources
shouldntbeusedtotesta
kitassociatedwithan
unwillingvictim).
Communitybasedadvocates
emphasizedthatsurvivors
choicesmustberespected;
becauseitmaybepractically
impossibletoascertain
survivorswishesbefore
testing,itisimportantto
considerwhetherthere
couldbeanyunintended
consequencesoftestingall
SAKs.
DECISION.Thecollaborativedidnot
reachconsensusonthispoint,and
eventuallythetopicfadedfrom
discussion,largelybecauseitwasa
mootpoint(Detroitdidnothavefunds
totestallSAKsatthattime).
Thecollaborativealsodidnotcontinue
todebatetheissueoftestallbecause
theElectedProsecutorstatedin
multipleforums/venuesthatallSAKs
shouldbetested(i.e.,thelongtermgoal
wastohaveallSAKstested).Giventhe
Prosecutorsstatedpositiononthis
issue,itwasunclearwhetherthe
multidisciplinaryteamneededtodebate
thisissuegiventhattheymaynotbethe
entitytomakesuchadecision.
However,theissueoftestingkits
associatedwithcomplainantrefusedto
prosecute(CRTP)caseswasresolved.
Theresearchfindings(seeChapter3)
regardinghowvictimsweretreatedby
thepoliceraisedquestionsaboutthe
accuracyoftheCRTPdesignationand
whetherittrulyreflectedthevictims
sentiments(orwhetheritwaswhatthe
policewrotetocloseoutacasethat
theydidnotfeelwasmeritorious).
Thecommunitybasedadvocatesalso
highlightedthatvictimsmightchange
theirmindsaboutprosecution,given
timeandspacetohealfromthetrauma.
Therefore,thecollaborativeagreedthat
theCRTPdesignationshouldnotbea
criterionforSAKtesting.
144
ISSUE2
HOWMANYSAKs
CANBETESTED
NOWINTHE
SCOPEOFTHIS
ACTION
RESEARCH
PROJECT?
DISCUSSION2.TheARP
budgetincludedsome
fundsfortestingandthe
statepolicehadaseparate
NIJDNABacklogReduction
Grant.Thisgrant
supportedSAKtesting
throughouttheentirestate
&statepolicepersonnel
expressedconcerns
regardinghowmuchof
theirfundingcould/should
beallocatedtoDetroitkits.
DECISION2.Seniorleadershipfromthe
prosecutorsofficeandstatepolicehad
separatemeetingsaboutthisissue(i.e.,
itwasnotdiscussedintheregular
collaborativeteammeetings).Itwas
laterannouncedthatthisissuehadbeen
resolved.PoolingfundsfromtheDetroit
SAKARPbudget,thestatepolice
departmentsNIJDNABacklog
ReductionGrants,andtheresourcesofa
universitybasedforensiclaboratory
(whichwasseparatelyfundedbyNIJ),
theprojectwasabletotest1,600SAKs
(1,595actuallytested).
ISSUE3
WHATDNA
TESTING
METHOD
SHOULDBEUSED
TOTESTTHE
SAKs?
DISCUSSION3.Forensic
scientistsfromthestate
policecrimelabsuggested
thatthecollaborative
considernewerDNA
technologiesthatcouldbe
fasterandpotentiallyless
expensive(whichcould
increasethenumberofkits
thatcouldbetestedinthe
project).
Thecollaborativeaskedthe
statepoliceforensic
scientiststoprovidea
briefingtotheteamonany
newtechnologiesthat
oughttobeconsidered.
DECISION3.Thestatepoliceforensic
scientistsbriefedthegrouponY
screeningmethods(analternativeto
thetraditionalserologyscreeningstep
thatprecedesDNAtesting)and
selectivedegradationmethods(an
alternativemethodforpreparing
samplesforDNAtesting)(seesection
BackgroundContext:AnOverviewof
DNATestingandCODISfordetails).
Thecollaborativedecidednottousethe
Yscreeningmethod(inthisproject)
becausefollowuptestingmightbe
neededtodiscernwhichspecificrape
kitssamplesyieldedwhichspecificDNA
testingresults(detailsnecessaryfor
court).AlthoughYscreeningcouldsave
testingtime,therewerelingering
concernsabouttheneedforfollowup
testing(andthetimerequiredforthat).
Thecollaborativedecidedtoevaluate
theefficacyofselectivedegradation
methodsonalimitednumberofSAKs
(mostwouldbetraditionaltesting).
145
ISSUE4

HOWSHOULD
SAKsBE
SELECTEDFOR
TESTING?
DISCUSSION4A.Random
selectionwasconsidered
asamethodofselecting
the1,600kits.Yetconcerns
wereraisedaboutwhether
priorityshouldbegivento
casesthatwereatriskfor
expiringstatuteof
limitations(SOL).
DECISION4A.Theresearchteamadvised
thegroupnottodrawasimplerandom
sampleof1,600kits.Instead,the
researchersrecommendstratified
randomsampling,wherebythe
multidisciplinaryteamwouldoutlinekey
criteriaforselection(suchasSOLrisk)
andthencaseswouldberandomly
selectedwithinthoseparameters.
DISCUSSION4B.Theteam
debatedwhatcriteria
shouldbeusedtoselect
cases.SOLriskwasa
criterionreadilyagreed
upon,buttherewere
strongdifferencesof
opinionregardingother
possibleselectioncriteria.
Somebelievedthat
strangerperpetratedcases
shouldbeprioritizedwhile
othersnotedthatnon
strangerperpetratedwere
justasserious.Theideaof
prioritizingcases
associatedwithserial
offenderswasalso
discussed.
DECISION4B.Thecollaborativecould
notcometoconsensusonselection
criteria.Giventhesedifferent
viewpoints,theresearch/evaluation
teamconductedaRapidAssessment
Processevaluation(RAP;Beebe,2001)
togatherdatafromlocal,state,and
nationalsexualassaultstakeholders
regardingtheirbeliefs,assumptions,and
questionsaboutthepurposeandutility
ofSAKtestingsothattheselectionof
the1,600SAKswouldbeempirically
driven.Thisevaluationidentifiedmany
possibleselectioncriteria;theresearch
teamsuggestedthattheSAKARP
projectcouldempiricallyevaluate
multipleoptions,andthenthoseresults
couldinformthedevelopmentofalong
termtestingplan.
DISCUSSION4C.The
testingplanforthisproject
neededtoincludecasesat
riskforexpiringstatuteof
limitations.
DECISION4C.Theresearchersaskedthe
prosecutorstoprovideacutoffyear
forSOLrisk.Itwasdecidedthat
incidentsthatoccurredintheyear2002
orlaterwouldmostlikelystillbe
eligible,sothisbecamethecutoff.
Theresearchersrecommendedthat
mostSAKstobetestedshouldbe
sampledcasesfrom20022009,butone
sampleofpre2002casesshouldbe
testedtoexploretheutilityoftesting
casespresumedtobeSOLexpired.
146
ISSUE5

WHICHSAKs
WOULDBE
SELECTEDFOR
TESTING?
DISCUSSION5A
DECISION5A
TESTINGGROUP1:
STRANGERRAPECASES.
TESTINGGROUP1:
STRANGERRAPECASES.
Stakeholdersdisagreed
whetherstranger
perpetratedassaults
shouldhavepriorityover
nonstrangerassaults.
Somestakeholdersfeltthat
SAKtestingwouldbemost
usefulinstrangercases
becauseitcanrevealthe
identityoftheoffender&
somebelievedthatthese
casesposethelargest
threatstopublicsafety.
TestingGroup1wouldconsistof450
randomlyselectedstrangerassailant
casesfrom20022009.Inthesecases,
theidentityoftheassailantisunknown
sothebestchanceofsolvingthecase
wouldlikelybeDNAtestinganda
resultingCODIShit.Statisticalanalyses
wouldbeconductedtodeterminethe
probabilitythatastrangerSAKyieldsa
CODIShit.Additionalanalyseswould
examinewhatfactorspredictwhethera
casewillyieldaCODIShit(i.e.,victim,
assailant,casecharacteristics).
DISCUSSION5B
DECISION5B
TESTINGGROUP2:NON
STRANGERRAPECASES.
TESTINGGROUP2:NONSTRANGER
RAPECASES.
Stakeholdershaddiffering
opinionsabouttheutility
ofSAKtestinginnon
strangersexualassaults.
Somefeltthattestingwas
notaprudentuseof
funding,astheidentityof
theassailantwasalready
known.Prosecutorsand
advocatesnotedthatthese
casescouldidentifyserial
offendersand/orrefute
defenseclaimsofno
contactbetweenthevictim
&allegedoffender.
randomlyselectednonstranger
assailantcasesfrom20022009.Non
strangerincludesawidearrayof
victimoffenderrelationships,ranging
fromknownbysight(oronlybyfirst
name/nickname)tointimatepartner
sexualassault.Statisticalanalyseswould
beconductedtodeterminethe
probabilitythatanonstrangerSAK
yieldsaCODIShit.Additionalanalyses
wouldexaminewhatfactorspredict
whetheracasewillyieldaCODIShit.
147
ISSUE5
WHICHSAKs
WOULDBE
SELECTEDFOR
TESTING?
DISCUSSION5C
DECISION5C
TESTINGGROUP3:
PRESUMEDSOLEXPIRED
CASES.
TESTINGGROUP3:PRESUMEDSOL
EXPIREDCASES.
Stakeholdershaddifferent
opinionsabouttheutility
oftestingSAKspresumed
tobeSOLexpired.
Prosecutorsarguedthat
testingcouldreveal
associationstocurrent
cases,allowingthe
evidenceoftheprior
crimestobeadmittedto
thenewcase,per404(b)
federalrulesofevidence.
Advocatesnotedthat
survivorscouldfeel
coercedtoparticipatein
theprosecutionofother
cases,eventhoughtheirs
couldnotbeprosecuted.
randomlyselectedcasesthatoccurred
before2002(i.e.,thesecasesare
presumedtobeSOLexpired),including
bothstrangerandnonstrangercases.
Statisticalanalyseswouldbeconductedto
determinetheprobabilitythata
presumedSOLexpiredSAKyieldsaCODIS
hit.Additionalanalysesexaminewhat
factorspredictwhetheracasewillyielda
CODIShit.
DISCUSSION5D
DISCUSSION5D
TESTINGGROUP4:DNA
TESTINGMETHOD.
TESTINGGROUP4:DNATESTING
METHOD.
Stakeholderswereopento
theideaoftryingnewer
DNAtestingmethods,
providedthatthe
alternativemethodsdid
nottakeadditional
time/testingstepsandthat
thealternativeapproaches
wouldhavethesame
accuracyastraditional
methods.
randomlyselectedSAKsfrom20022009
thatwerethenrandomlyassignedtoone
oftwodifferentmethodsforpreparing
samplesforDNAtesting:traditional
extractionmethodsvs.selective
degradationmethods.
RatesofCODISentryandtime/costsof
testingwouldbecomputed&compared
acrosstestingcondition.
148
ISSUE6
WHICH
LABORATORY
SHOULDTESTTHE
SAKs?
DISCUSSION6.Thelocal
policedepartmentscrime
laboratorywasclosedin
2008andafterthat,the
statepolicecrimelabs
wereresponsiblefor
forensictestingforDetroit
cases(inadditiontoall
othercitiesinthestate).
Thestatepolicecrimelabs
didnothavethecapacity
totest1,600SAKswithin
thetimelineofthisproject,
sokitsneededtobe
outsourcedtoother
vendors.
DECISION6.Thestatepolicehada
contractualrelationshipwithawell
establishedprivatelaboratory.NIJhada
contractualrelationshipwitha
universitybasedforensiclaboratoryand
couldleveragethatrelationshiptohelp
supportthetestingofSAKsassociated
withthisproject.
TheassignmentofTestingGroupsto
vendorlaboratorieswasbasedon
availability,budget,andtestingcapacity.
TestingGroup4(DNATestingMethods)
hadtobedoneattheprivatelaboratory
astheyweretheonlyprojectvendor
thathadthetechnologyavailablefor
selectivedegradationmethods.Testing
Group1(StrangerRape)wasalsosentto
theprivatelaboratory;TestingGroup2
(NonStrangerRape)andTestingGroup
3(PresumedSOLExpired)weresentto
theuniversitybasedforensiclaboratory.
Statepoliceforensicscientists
conductedsitevisitsatbothvendors
(priortotheshipmentofkits)toensure
thattheirlaboratoryprocessesmet
specificstandardsandqualityassurance
procedures.
149
Implementing the Detroit SAK Testing Plan

Overview
Figure4.2(below)summarizestheDetroitSAKARPtestingplan.Thecollaborativedecidedto
formfourTestingGroups,eachonedesignedtoaddressspecificresearchquestionsregardingtheutility
ofSAKtestingunderdifferentcasecircumstances.ThisdesignallowsustoexaminetheutilityofSAK
testingforstrangerperpetratedsexualassaults(TestingGroup1),nonstrangerperpetratedsexual
assaults(TestingGroup2),andsexualassaultcasesthatarepresumedtobebeyondthestatuteof
limitations(TestingGroup3).WealsowantedtoexaminewhethernewerDNAtestingtechniques,such
asselectivedegradationmethodscouldofferfaster,lessexpensivetestingoptions,withoutsacrificing
accuracy(seesectionBackgroundContent:AnOverviewofDNATestingandCODISformoredetails
aboutselectivedegradationmethods).Therefore,TestingGroup4wasarandomizedexperiment
comparingSAKstestedwithtraditionalDNAmethodsvs.theselectivedegradationmethod.
FIGURE4.2AnOverviewoftheDetroitSAKActionResearchProjectTestingPlan
1,600KITS
EachTestingGroupwilladdressauniquesetofquestionstoinformfuture
testingpracticesinDetroitandothercommunities
TESTINGGROUP1
(targetn=450)
UtilityofTestingSAKs
inStrangerSexual
AssaultCases
inNonStranger
SexualAssaultCases
inPresumedSOL
ExpiredCases
TESTINGGROUP2
(targetn=450)
TESTINGGROUP3
(targetn=350)
TESTINGGROUP4
(targetn=350)
ComparisonofDNA
TestingMethods
150
Defining the Sampling Criteria for SAK Selection

Toselect1,600SAKsfortesting,weusedfoursamplingcriteria,whichwerebasedonboth
scientificaimsaswellaspracticalconstraints.First,wewantedtofocusonSAKsthathadneverbeen
tested(testingstatuscriterion).AsnotedinChapter2:HowManyUnsubmittedSAKsinDetroit,we
couldmakeareasonableinferenceastothetestingstatusofeachSAKinpoliceproperty,andbasedon
thosedata,the400kitstestedaspartoftheOVWfundedThe400Projectwereexcludedfromour
samplingframe(i.e.,theywereexcludedfromthepoolofcasesfromwhichwewouldselectthe1,600
tobetestedinthisproject).Inaddition,theSAKsthathadapolicedepartmentcrimelaboratoryID
numbers(seeChapter2)werealsoexcludedfromthesamplingframe;thoughwecouldnotverify
whetherthesekitshadbeentestedforDNA,theyhadbeensubmittedfortestingandmayhavehadat
leastsometestingperformed.
Second,wewantedtheSAKstobetestedinthisprojecttobeassociatedwithnonadjudicated
cases(adjudicationstatuscriterion)(i.e.,casesthatwerestillpotentiallyactionable,pendingSOL
determination).AsdescribedinChapter2:HowManyUnsubmittedSAKsinDetroit,wewerenotableto
determinetheadjudicationstatusforeverykitinpolicepropertyduringthecensusbecausedoingso
requiredlaborintensivesearchingofpoliceandcourtrecords.Therefore,wewouldneedtoaddastep
inthesamplingdesigntoscreeneachcasebeingconsideredforinclusionintheprojectforadjudication
status(i.e.,ratherthantryingtodetermineadjudicationstatusforallSAKs/cases,wewouldtrackdown
adjudicationinformationforonlythosecasesrandomlyselectedforpossibleinclusionintheproject).
Weoperationallydefinedadjudicatedasacaseinwhichtherehadbeenaconviction,acquittal,or
guiltyplea;acaseinwhichanarresthadbeenmade,butchargeswerenotfiledorweredroppedprior
totrial/pleawascodedasnonadjudicated.Inpractice,determiningadjudicationstatuswaschallenging
andburdensome,giventheincompletenessanddisorganizationoflegalrecordsinDetroit(seebelow
formoredetails).
151
Third,thecollaborativedecidedthattheprojectshouldexploretheutilityoftestingpresumed
SOLexpiredcases,butmostofthe1,600tobeselectedshouldbewithinthestatuteoflimitations(SOL
criterion).BasedonananalysisofMichiganlawregardingstatuteoflimitationsforcriminalsexual
conductcrimes,theprosecutorsofficeestablished2002asareasonablecutpointforstatuteof
limitations(i.e.,casespriorto19802001werelikelybeyondSOL;cases20022009werelikelywithin
SOL).Duringthecensus,thedate/yeartheSAKwascollectedwasrecorded,sothesamplingframecould
beeasilysortedbydateforSOLdetermination.
Finally,theprojectteamdecidedtoexaminetheutilityofSAKtestingforbothstrangerandnon
strangerperpetratedsexualassaults(victimoffenderrelationshipcriterion).Again,itwasnotfeasible
inthecensustodeterminethatinformationforeachcase;therefore,wewouldneedtoaddastepin
thesamplingdesigntoreviewcasesforvictimoffenderrelationshipifsuchinformationwasrelevant
fortheformationaparticulartestinggroup(i.e.,itwasrelevantfortheformationofTestingGroup1
[Stranger]andTestingGroup2[NonStranger],butitwasnotrelevantfortheothertwoGroups).
Consistentwithpriorresearch(Campbelletal.,2009;Campbelletal.,2012),wedefinedstranger
assaultsasthoseinwhichthevictimdidnotknowtheoffenderinanyway(TestingGroup1);allother
assaultswerecodedasnonstranger(TestingGroup2).Withinthenonstrangergroup,wesubcoded
victimoffenderrelationshipintothreecategories:knownbysight/nickname/streetname;
friend/associate/familymember(butnotintimatepartner);current/pastintimatepartner.Itwas
sometimesdifficulttodistinguishstrangervs.knownbysight/nickname/streetname,andsothe
prosecutorsofficestaffandtheresearchteamconferredtoestablishconsistentoperationalizations.63
63
Caseswerecodedasastrangeriftherewasnoinformationthatthepolicecouldworkfrominordertoidentifythepossible
assailant.Forexample,ifthevictimsstatementwasthattheassailantwascalledJohnandsheknewthathehangsoutat
thepartystore(andthespecificlocationofthatpartystorewasgiventothepolice),thecasewouldbeplacedinTesting
Group2(NonStranger),subcodedasknownbysight/nickname/streetname.Bycontrast,acaseinwhichthevictims
statementsaidthatshethoughtsheheardsomeoneelseatthecrimescenerefertoherattackerasJonesbutshehadno
physicaldescriptionofherattackerandnootherdetailsabouthisidentityandnootherleadstogoon,thatcasewouldbe
placedinTestingGroup1(Stranger).
152
Itisimportanttonotethatthesamplingcriteriadidnotincludefactorssuchasvictimage(e.g.,
samplingonthebasisofwhetherchild/adolescentoradult)orassaultcharacteristics(e.g.,useofforce,
useofweapon),meaningthatacasewouldnotbeincluded/excludedfromthesamplebasedonsuch
factors.Therefore,allfourTestingGroupsincludevictimsofvaryingages,races/ethnicities,assault
experiences,etc.,andtheevaluationofthetestingplanwouldexaminewhethersuchfactorswere
significantlyassociatedwithCODIShits/serialsexualoffending.
Selecting SAKs for the Testing Groups

Withthesefoursamplingcriteriaestablished,wefirstsampledtheSAKsforTestingGroup1
(Stranger)andTestingGroup2(NonStranger).64Usingthecensusresults,wegeneratedalistofall
unsubmittedSAKs(testingstatuscriterion),fromtheyears20022009(SOLcriterion),andputthat
listinrandomizedorder,stratifiedbyyear(i.e.,arandomizedlistofSAKsfor2002,2003,etc.).Staff
fromtheprosecutorsofficestartedwiththefirstSAKIDnumberoneachlistandthencompiledpolice
andcourtrecordstodeterminewhetherithadbeenadjudicated(adjudicationstatuscriterion).As
notedpreviouslyinChapter2:HowManyUnsubmittedSAKsinDetroit,onlysomeDetroitcriminal
justicesystemrecordsarecomputerized(mostarepaperfiles),sothiswasadifficultandtime
consumingtask.65Ifthecasehadbeenpreviouslyadjudicated,itwassetasideandthenextcaseonthe
64
ItwasnotfeasibletosampleSAKsforallfourTestingGroupsatoncebecauseadditionalscreeningforadjudicationstatusand
victimoffenderrelationshipwasnecessary,andwedidnothavethestaffingtoscreenlargenumbersoffilesallatonce.
65
Forexample,thecourtrecordsthatwerecomputerizedrequiredassailantnameanddateofbirth(DOB)assearchfields,but
theSAKswereidentifiedbyvictimnameandvictimDOB;toobtainassailantnameandDOB,projectstaffhadtorequestthe
policereport(whichwasusuallystoredoffsiteinremotestorage).Ifprojectstaffwereabletofindenoughinformationto
completeacomputerizedsearchofcourtrecords,thesearchresultswerenotalwaysconclusiveduetomissinginformationin
thedatabase.Ininstancesinwhichacomputerizedsearchwasnotpossible(orthesearchhadbeeninconclusive),projectstaff
thenhadtolocatearrestlogbooks,warrantingpaperwork,and/orprosecutioncasefilestodetermineadjudicationstatus.It
couldtakemultipledays,involvingstafffromseveralorganizations,tofindandcompiletherecordsnecessarytocompletethe
adjudicationstatusscreeningofonecase.
153
listwasreviewed.66Ifthecasehadnotbeenpreviouslyadjudicated,thenitwasreviewedforvictim
offenderrelationship(victimoffenderrelationshipcriterion)andplacedintoeitherTestingGroup1
(Stranger)orTestingGroup2(NonStranger).Stafffromtheprosecutorsofficecontinuedworking
downtherandomizedlistsforeachyearuntiltheyhadidentifiedapproximately56eligiblecasesper
year,ineachTestingGroup,foratargetoverallsamplesizeof450SAKsforeachTestingGroup.The
targetsamplesizeof450wasbasedonbudgetconstraints,andaposthocpoweranalysisindicatedthat
thissamplewouldbesufficientfortheplannedanalyses.Inpractice,thefinalsamplesizeforTesting
Group1(Stranger)was445(ratherthan450)becausefivekitsturnedoutnottobeeligibleoncethey
wereopenedatthelab(e.g.,thekitdidnotcontainanybiologicalsamplesfromasexualassault).
Unfortunately,wedidnotlearnthisuntilitwastoolatetosendfivereplacementkitsfortesting.The
finalsamplesizeforTestingGroup2(NonStranger)was449(ratherthan450)becausewediscovered
thatoneSAKhadbeenlistedtwiceonthelistofsampledcasesanditwastoolatetosenda
replacementkitfortesting.Table4.1(nextpage)summarizesthesamplingcriteriaasimplemented
foreachofthefourTestingGroups.
AfterTestingGroups1and2hadbeensampled,weselectedSAKsforTestingGroup3
(PresumedSOLExpired).Alistofallunsubmitted/untestedSAKs(testingstatuscriterion)from1980
to2001(SOLcriterion)wasgeneratedfromthecensusresults.ForthisTestingGroup,wedidnot
stratifybyyear(forsimplicity).Basedontheconsiderabledifficultiesweencounteredscreeningfor
adjudicationstatusinTestingGroups1and2(adjudicationstatuscriterion),theteamconferredabout
theplausibilityofcompletingthisscreeningforthepresumedSOLexpiredcases,giventhattherecords
fortheseolderkitswouldbeevenhardertolocateiftheystillexistedatall.Thescreeningofthe
66
ThestateappellatedefendersofficehadrecentlyreceivedadifferentNIJgranttosupportSAKtestingforpreviously
adjudicatedcases,sothecasesthatscreenedoutoftheSAKactionresearchprojectweresetasideforthatotherproject.
154
TABLE4.1SummaryofSamplingCriteria(asImplemented)fortheSAKTestingGroups(N=1,595)
TestingStatus
Criterion
AdjudicationStatus
Criterion
SOLCriterion
TestingGroup1
(StrangerRape)
TestingGroup2
(NonStrangerRape)
TestingGroup3
TargetN=450
TargetN=450
TargetN=350
TargetN=350
ActualN=445
ActualN=449
ActualN=351
ActualN=350
Unsubmitted/untested
Nonadjudicated
Nonadjudicated
<NOTSCREENED>
<NOTSCREENED>
WithinSOL
(20022009)
VictimOffender
Relationship
Criterion
StrangerPerpetrated
WithinSOL
(20022009)
NonStranger
Perpetrated
BeyondSOL
(19802001)
<NOTAPPLICABLE>
(includesbothstranger
&nonstranger)
TestingGroup4
(DNATestingMethod)
WithinSOL
(20022009)
<NOTAPPLICABLE>
(includebothstranger
&nonstranger)
kits/casesforTestingGroups1and2indicatedthatmostkitswereassociatedwithnonadjudicated
cases(i.e.,mostcasesfitourdesiredsamplingcriterion).Itiscertainlypossiblethattheolderkits(pre
2001)mightbedifferentfromthe20022009kitswithrespecttothisissue,butgiventhatpatternof
findings,wedecidednottoscreenforadjudicationstatusforTestingGroup3,underthatassumption
thatmostwouldbenonadjudicated,butcertainlynotall.67Forthisgroup,noadditionalscreeningwas
neededforstrangervs.nonstrangerperpetratedassaults(bothwereincluded)(victimoffender
relationshipcriterion).ToselectthecasesforthisTestingGroup,wedrewasimplerandomsampleof
350kitsfromthelistofunsubmitted/untestedSAKsfrom1980to2001.Inpractice,thefinalsamplesize
forthisGroupwas351kits.Whenoneoftheselectedkitswasopenedatthelab,itcontainedbiological
samplesfromtwodifferentvictims(hence+1,n=351,not350).
TestingGroup4(DNAMethod)wassampledlast.Usingthecensusresults,wegeneratedalistof
allunsubmittedSAKs(testingcriterion),fromtheyears20022009(SOLcriterion)excludingany
67
WewilldiscusshowweaccountedforthedifferencesinthesamplingdesignsinouranalysesforeachGroupindetailinthe
sectionEvaluatingtheDetroitSAKTestingPlanInferentialFindingsontheEffectofVictimOffenderRelationshipandStatute
ofLimitationStatusonForensicTestingOutcomes(seealsoAppendixB:ProjectMethodology).
155
casesthathadalreadybeensampledforTestingGroup1or2.Giventhedifficultiesofaccessingpolice
andcourtrecords,theTestingGroup4caseswerenotscreenedforadjudicationstatus(adjudication
statuscriterion).Forthisgroup,noadditionalscreeningwasneededforstrangervs.nonstranger
perpetratedassaults(bothwereincluded)(victimoffenderrelationshipcriterion).Therefore,we
generatedalistofallunsubmittedSAKs,fromtheyears20022009(forsimplicity,nostratificationby
year),andrandomlyselectedasampleof350kits.Afterkitswereselected,theywererandomly
assignedtotwotestingconditions(traditionalvs.selectivedegradation).Toconducttherandom
assignmentofSAKstotestingcondition,asimplerandomsampleof175SAKswasdrawn(without
replacement)usingtheRsoftwarepackagefromthetotalsampleof350kits.
Challenges Implementing the Testing Plan and Coordinating Post-Testing Activities
InadditiontothechallengesencounteredduringtheSAKscreeningandselectionprocess
(describedabove),thecollaborativehadtoresolveadditionallogisticsdifficultiespreparingthekitsfor
shipment,coordinatingposttestingreview,andcommunicatingthetestingresultstotheproper
authoritieswithinthecriminaljusticesystemandtomembersofthecollaborativeproject.Figure4.3
(followingpages)summarizestheseissues,discussions,anddecisions.Althoughotherjurisdictionsmay
notfacethesamestaffingshortagesthatDetroitencountered,whichaddedsignificantdelaystothe
processofpreparingandshippingSAKs,manyoftheseissueswilllikelyberelevanttoother
communitiesthataretestinglargenumberofrapekitsallatonce.
156
FIGURE4.3Challenges&SolutionsEncounteredImplementingtheDetroitSAKARPTestingPlan
ISSUE1
HOWSHOULD
KITSBEPULLED
FROMPOLICE
PROPERTYAND
PREPAREDFOR
SHIPMENTTO
THEVENDOR
LABORATORIES?
DISCUSSION1A.Oncekits
hadbeenselectedfor
testing,theyneededtobe
pulledfrompoliceproperty
andpreparedforshipping
tothetestingvendors.
Becausethepoliceno
longerhadtheirowncrime
laboratory,thestatepolice
hadresponsibilityforthe
testing,whichmeantthat
eachkithadtobegivena
statepoliceforensicscase
numberpriortoshipping.
DECISION1A.Therewasonlyonestaff
memberinthelocalpolicedepartment
(theforensicsciencescoordinator)
allocatedforthistask.Therewere
repeateddiscussionsinthecollaborative
teammeetingsregardinghowtomake
thisprocesslessburdensome(e.g.,
assigningadditionalstaff,streamlining
thesubmissionprocesssothatkitscould
godirectlyfromthelocalpolicetothe
testingvendor).However,thevast
majorityofthe1,600kitswereprepared
forshippingbyonlyoneindividual.
DISCUSSION1B.Preparing
olderSAKsforTesting
Group3(PresumedSOL
Expired)wasmore
challengingduetothepoor
conditionsofsomeofthe
SAKs(e.g.,tornenvelopes,
missingtagnumbers,etc.).
Someolderkitshadhad
serologytests,sotheblood
cardshadtoberetrieved
priortoshippingtheSAKs.
DECISION1B.Onepoliceforensic
sciencecoordinatorwasresponsiblefor
overseeingtheshippingofall1,600
SAKs,andthe350SAKsdesignatedfor
TestingGroup3(PresumedSOLExpired)
werethemostlaborintensive.
Thecollaborativediscussedoptionsfor
assigningadditionalpersonneltohelp
withthistask,butbecausestaffingre
allocations(eventemporaryones)were
difficulttosecure,thepreparationof
thekitswashandledbyonlyoneperson.
157
ISSUE2
HOWSHOULD
KITSBESHIPPED
TOTHEVENDOR
LABORATORIES?
ISSUE3
HOWSHOULD
POSTTESTING
RESULTSBE
REVIEWED&
VERIFIED?
DISCUSSION.Tomaintain
properchainofcustody,
eachkithadtohaveanID
numberfromboththelocal
policedepartmentandthe
statepoliceforensic
sciencedivision,andthena
manifestneededtobe
preparedforeach
shipment.Thekitsneeded
tobeshippedbyovernight
service(onlyonspecific
days)andreceiptofthekits
hadtobeacknowledgedby
thevendors.
DECISION.Stafffromfourorganizations
wereinvolvedinthisprocess:local
police(forensicsciencecoordinator),
statepolice(forensicscientists),the
prosecutorsoffice(therecipientofthe
grantandresponsiblepartyforthe
paymentofeachshipment),andeach
testingvendor.Therewerefrequent
miscommunicationsbetweenparties,
particularlyastheseprocedureswere
beingimplementedforTestingGroups1
and2,butbytheendoftheproject,
staffhadfoundwaystostreamline
communication.However,itwasnot
possibletocompletethistaskwithout
allfourorganizationsinvolvement.
DISCUSSION.Thekitswere
inlocalpolicedepartments
propertyfacilities&
shippeddirectlytotesting
vendors,buttheresults
weregiventothestate
policeforensicscience
divisionasstatepolice
forensicscientistswere
responsibleforreviewing
andcertifyingtheresults.
DECISION.Thestatepoliceforensic
sciencedivisionneededtodevelopa
staffingplanforthetechnicalreviewof
all1,600SAKswithintheproject
timeline.Distributingstafftimefor
testingcurrentcases(fromthroughout
thestate)&forreviewinglargenumbers
oftestingresultsfromthisprojectwas
crucial(albeitchallengingasbothtesting
vendorstendedtoreleaseresultsin
largebatches).
Forensicscientistswereoffered
overtimepaytocompletethescientific
technicalreviews.Thecostsofthat
overtimewereabsorbedbythestate
policedepartmentsbudget.
158
ISSUE4
WHICH
INDIVIDUALS/
ORGANIZATIONS
SHOULDRECEIVE
THETESTING
RESULTS?
DISCUSSION4A.OnceSAK
testingresultswere
reviewedbystatepolice
forensicsciencepersonnel
andenteredintoCODIS,
theresultsneededtobe
releasedtotheproper
criminaljusticesystem
authorities.
DECISION4A.Thelocalpolice
departmentandprosecutorsofficehad
todesignatespecificpeopleand/orunits
thatshouldbenotifiedre:testing
results/CODIShits.Proceduresforthat
notification(whatexactinformation
wouldbeshared,bywhatmechanism.)
hadtobediscussedandagreedupon.
DISCUSSION4B.Because
thistestingwasoccurring
inthecontextofa
multidisciplinary
collaborationandresearch
project,thegroupneeded
toconsiderwhat
informationcouldbe
releasedtotheother
agenciesinpartnershipand
totheresearchteam.
DECISION4B.Specificcaseinformation
(e.g.,caseIDnumber,victim/offender
names)couldnotbereleasedto
membersofthecollaborativewhowere
notdirectlyinvolvedinthetesting,
investigation,andpossibleprosecution
ofthecases.However,allmembersof
thegroupwereinterestedintracking
howtestingwasproceedingandthe
numberofCODIShitsthatwere
emanatingfromtesting.
Therefore,theresearchteam
constructedaCODIShitstrackingtool
(seeFIGURE4.7).Thestatepolice
forensicsciencepersonnelprovidedbi
monthlyupdatestothemultidisciplinary
teamforeachtestinggroupon:(a)the
numberofSAKstestedbyeachvendor
lab;(b)thenumberofSAKsreceivedby
thestatecrimelabafterhavingbeen
tested;(c)thenumberofSAKsreviewed
bythestatecrimelab;(d)thenumberof
profilesuploadedintoCODIS;(e)the
numberofCODIShitsincludingoffender
hits,forensichits,andoffenderand
forensichits.
159
ISSUE5
WHATSTEPS
SHOULDBE
TAKENFORTHE
CASESTHAT
RESULTEDINA
CODISHIT?
DISCUSSION.Thestate
policeforensicsciences
divisionalertedboththe
localpolice&prosecutors
regardingeveryCODIShit.
Whenthefirstbatchesof
hitsarrived,thepolice&
prosecutorsbegan
implementingtheirusual
processforfollowingupon
CODIShits(e.g.,pullingthe
originalpolicefiles,
identifyingnextstepsfor
theinvestigation).
However,theseefforts
werenotcoordinatedand
boththelocalpolice
departmentandthe
investigativeunitwithin
theprosecutorsoffice
begansimultaneousefforts
reviewingeachcase.
DECISION.Thecollaborativedecided
thatthecases/hitsassociatedwithThe
400Project(whichwasstillinprogress)
wouldbehandledbythelocalpolice;
thecases/hitsassociatedwiththis
projectwouldbehandledbythe
prosecutorsofficeinvestigators.
Therewasstillnosingledata
managementsystemthattrackedSAKs
astheymovedfromtestinginto
investigationandprosecution.Asthe
ARPwasending,themultidisciplinary
teamwaslookingintooptionsfor
computerizedtrackingofSAKsfromthe
momenttheyarecollectedbymedical
personnelthroughforensictestingand
thenintoinvestigationandprosecution.
Thisduplicativeeffortwas
discoveredquicklyduring
thereportoutportionof
amultidisciplinaryteam
meeting,whichprompted
thegrouptodevelopa
coordinatedplan.
160
Evaluating the Detroit SAK Testing PlanDescriptive Findings About the

Cases/SAKs Tested in this Project
BeforedelvingintotheforensictestingoutcomesfortheSAKstestedinthisproject(next
sectionofthischapter),wewanttofirstgroundthereaderinsomecontextaboutthesecases/SAKs
whatdoweknowaboutthesesurvivors,theirattackers,andtheassaultstheysustained?Foreachof
1,595SAKstested,werequestedthecorrespondingpolicefilesothatwecouldcodedemographicand
assaultcharacteristicsforeachcase.However,asnotedpreviouslyinthisreport,thepolicedepartment
didnothaveasophisticateddatamanagementsystemallreportswerehardcopypaperfiles,manyof
whichwereoffsiteinremotestorage.Thesexcrimesunithadmovedmultipletimesoverthe30year
periodinwhichthesecasesspanned(ashadthelocationofremotestorage),68soitistobeexpected
thatsomerecordswouldhavebeenlostovertime.Missingdatawasparticularlyproblematicfor
TestingGroup3(PresumedSOLExpired),butthepolicedepartmentmadeeveryefforttolocateas
manyfilesaspossiblefortheresearchteam.
Table4.2(followingpages)presentsdescriptivemeansandpercentagesforvictim,assailant,
andassaultcharacteristicsfortheoverallsampleandwithineachTestingGroup.Thevalidsamplesize
(N)andnumberofmissingcasesarenotedforeachvariable.Thevastmajorityofthesevictimswere
female(98%)andAfricanAmerican(81%).Thesamplewasquiteyoungatthetimeoftheassault,24
yearsoldonaverage,andaboutonefifth(21%)werechildren/adolescentsundertheageof16when
theywereraped.ThevictimswhosekitsweresampledforTestingGroup2(NonStrangerRape)were
younger(morewereundertheageof16)thanthesurvivorsinotherGroups.Priorstudieshavefound
thatadolescentsaremorelikelytobesexuallyassaultedbynonstrangersthanstrangers(Adams,
68
Forexample,inthe30monthsofthisactionresearchprojectalone,thesexcrimesunithadtomovetothreetimestothree
differentlocations(i.e.,threedifferentbuildings).
161
Girardin,&Faugno,2001;Jones,Rossman,Wynn,Dunnuck,&Schwartz,2003),whichmayexplainthis
ageeffectwithinthisTestingGroup.
Nearlyalloftheperpetratorsinthissampleweremale(99.5%)andmostwereAfricanAmerican
(92%).Theassailantswere,onaverage,aboutfouryearsolderthantheirvictims(28yearsoldon
average)andapproximately25%were21yearsoldoryoungeratthetimetheycommittedthisassault.
TheassailantsinTestingGroup2(NonStrangerRape)wereyounger(morewereundertheageof22)
thantheassailantsintheotherGroups.Asnotedabove,sexualassaultsagainstadolescentsaremore
likelytobecommittedbysomeonetheyknow(ratherthanstrangers)(Adamsetal.,2001;Jonesetal.,
2003),andifthoseassailantsarealsomoretypicallyteens/youngadults,thatmayexplainthese
findings;however,suchexplanationsarespeculativeandmeritfurtherresearch.
TheassaultsassociatedwiththeseSAKsoccurrednearlytenyearsago(9.48average),witha
rangeof4yearsagoto25yearsago.Asexpected,theassaultsinTestingGroup3(PresumedSOL
Expired)occurredlongeragothanthoseintheotherGroups(18yearsonaverage).Victimoffender
relationshipwasaselection/stratificationvariableforTestingGroup1(StrangerRape)andTesting
Group2(NonStrangerRape).Therefore,bydesign,100%ofthecasesinTestingGroup1were
strangerperpetrated.AllofthecasesinTestingGroup2werenonstrangerperpetrated,mostofwhich
(58%)werecommittedbyfriends,associates,orfamilymembersofthevictim.Inapproximately20%of
thecasesinthisgroup,thevictimknewtheassailantbysight/streetname(e.g.,Johnfromtheparty
storeat[specificlocationnamed]).InTestingGroup3(PresumedSOLExpired)andTestingGroup4
(DNATestingMethod),victimoffenderrelationshipwasnotaselection/stratificationcriterion;each
groupwasrandomlysampled,whichyieldedmorestrangerrapecasesinTestingGroup3(Presumed
SOLExpired).AsnotedinthehistoricalanalysisinChapter3:WhySoManyUnsubmittedSAKsin
Detroit,DetroitdidnothavefullaccesstoCODISuntil2006,andmanystakeholdersinthepolice
departmentandinthepolicedepartmentcrimelabnotedthatstrangerrapecaseswerelesslikelytobe
162
submittedinthepreCODISera(becausetheinvestigationalutilityoftheSAKwaslimitedwithout
CODIS/referencesamples).TheseSAKsubmissionpracticeslikelyexplainthehighernumberofstranger
perpetratedassaultsamongtheolderkits(TestingGroup3:PresumedSOLExpired).
Nearlyonequarteroftheseassaults(22%)weregangrapessuchthatthevictimwassexually
assaultedbymultipleoffenders(withinthesameincident).Multipleperpetratorassaultsweremore
likelyinTestingGroup1(StrangerRapes).Priorresearchhasalsofoundthatgangrapesaremore
commoninstrangerperpetratedassaults(Gidycz&Koss,1990;Kossetal.,1988;Porter&Alison,2006;
Ullman,2007).Withrespecttoalcoholanddruguseatthetimetheassault,wecoulddiscernthat29%
oftheassaultsoccurredinthecontextofsubstanceuse,butthedocumentationinthepolicefileswas
notsufficientlydetailedforustobeabletoparseoutvictimvs.assailantuse(ordistinguishalcoholvs.
druguse)(i.e.,ourcodingreflectsalcoholordrugusebyeithervictimorassailant).Theassaultsin
TestingGroup1(StrangerRapes)andTestingGroup4(DNATestingMethod)weremorelikelytooccur
inthecontextofalcohol/drugusethantheassaultsintheothertwoTestingGroups.Wedonothavea
clearunderstandingastowhysubstanceusewashigherinTestingGroup4(whichwasrandomly
selectedfrom20062009),butforTestingGroup1(StrangerRape),thereispriorliteraturesuggesting
higherratesofvictimalcohol/druguseamongvictimsofstrangerrape(Kossetal.,1988;Ullman&
Brecklin,2000;howeverLogan,Cole,&Capillo,2007alcoholusebyvictimswasmorecommoninnon
strangerassaults).WhenwewerecodingthecasesinTestingGroup1(StrangerRape),therewere
manyinstancesinwhichwomenhadbeeninbars,atparties,indrughousesandwerethenabductedby
strangers(likelybecausetheirsubstanceusemadethemvulnerable,seeLisak,2008).
Whereasthepolicefilesoftenlackedprecisedetailsaboutsubstanceuse,therecordswere
moreconsistentwithrespecttowhethertheassailantusedaweaponand/orphysicalforceagainstthe
victim.Nearlyonethirdoftheassaults(30%)involvedtheuseofaweapon(e.g.,gun,knife,object
wieldedasweapon),andconsistentwithpriorresearch,weaponusewasmorecommonamong
163
strangerperpetratedassaults(TestingGroup1;42%)(seeKossetal.,1988;Riggsetal.,2000;Ullman&
Siegel,1993).Mostassaults(71%)involvedsomedegreeofphysicalforcebytheassailants(e.g.,
grabbingandthrowingthevictim,holdingdownthevictim,stranglingthevictim)(71%).Physicalforce
wasmoretypicalintheassaultsinTestingGroup1(StrangerRape)andTestingGroup3(PresumedSOL
Expired).Priorresearchsuggeststhatphysicalforceisquitecommoninstrangerrapes(Kossetal.,1988,
Ullmanetal.,2006;Riggsetal.,2000),butcanbejustasprevalentinintimatepartnersexualassaults
(Loganetal.,2007;Ullman&Siegel,1993).
Aftertheassault,mostofthesevictimssoughtmedicaltreatmentveryquickly:62%hadthe
medicalforensicexamandSAKthesamedayastheassault,26%hadtheexamthenextday,and12%
hadtheexambeyondoneday.ThevictimsinTestingGroup2(NonStrangerRapes)werelesslikelyto
seeksamedaycarethansurvivorsintheothertestingGroups.Priorresearchhasyieldedmixedresults
withrespecttotherelationshipbetweenvictimoffenderrelationshipandmedicalhelpseeking:Millar,
Stermac,andAddison(2002)foundthatvictimsofstrangerrapeweremorelikelytoseekimmediate
treatment,butLoganetal.(2007)foundnoassociationbetweentypeofrapeandwhenthevictim
soughtmedicalcare.
164
TABLE4.2TheDetroitSAKs:Victim,Assailant,andAssaultCharacteristics
VICTIMCHARACTERISTICS
OverallSample
Victim
Gender
ValidN
MissingN
Female
Male
1565
30
97.6%
2.4%
Victim
Race
ValidN
MissingN
1542
53
AfricanAmerican 80.6%
Victim
Age
TestingGroup2
TestingGroup3
TestingGroup4
(StrangerRape)
(NonStrangerRape)
(DNATestingMethods)
ValidN
MissingN
Female
Male
ValidN
MissingN
AfricanAmerican
444
1
97.3%
2.7%
ValidN
MissingN
Female
Male
446
3
97.3%
2.7%
ValidN
MissingN
Female
Male
327
24
98.8%
1.2%
ValidN
MissingN
Female
Male
348
2
97.1%
2.9%
441
4
ValidN
MissingN
440
9
ValidN
MissingN
316
35
ValidN
MissingN
345
5
AfricanAmerican
81.4%
16.8%
1.6%
.2%
AfricanAmerican
AfricanAmerican
Caucasian
Hispanic/Latina
82.3%
17.4%
.3%
81.2%
18.3%
.6%
Caucasian
Hispanic/Latina
Asian
MultiRacial
17.9%
1.1%
.3%
.1%
Caucasian
Hispanic/Latina
Asian
MultiRacial
78.2%
19.0%
1.6%
.9%
.2%
ValidN
1506
ValidN
442
ValidN
439
ValidN
280
ValidN
345
MissingN
89
MissingN
MissingN
10
MissingN
71
MissingN
Mean:
24.44
Mean
27.37
Mean
21.93
Mean
25.49
Mean
23.04
Std.Dev.
Range
11.36
Std.Dev.
11.79
Std.Dev.
10.58
Std.Dev.
12.01
Std.Dev.
10.23
288
Range
272
Range
359
Range
588
Range
355
<16yearsold
11.8%
<16yearsold
30.3%
<16yearsold
20.7%
<16yearsold
22.6%
>16yearsold
88.2%
>16yearsold
69.7%
>16yearsold
79.3%
>16yearsold
77.4%
<16yearsold
21.3%
>16yearsold 78.7%
TestingGroup1
Caucasian
Hispanic/Latina
Asian
Caucasian
Hispanic/Latina
165
TABLE4.2(cont.)
ASSAILANTCHARACTERISTICS
Assailant
Gender
Assailant
Race
OverallSample
ValidN
MissingN
Female
Male
ValidN
MissingN
1522
73
.5%
99.5%
1487
108
AfricanAmerican 91.8%
Caucasian
Hispanic/Latino
ArabAmerican/
Chaldean
Asian
MultiRacial
Assailant
Age
TestingGroup1
TestingGroup2
TestingGroup3
TestingGroup4
(StrangerRape)
(NonStrangerRape)
(DNATestingMethods)
ValidN
MissingN
Female
Male
ValidN
MissingN
433
12
0.0%
100.0%
410
35
ValidN
MissingN
Female
Male
ValidN
MissingN
441
8
.9%
99.1%
432
17
ValidN
MissingN
Female
Male
ValidN
MissingN
313
38
0.0%
100.0%
310
41
ValidN
MissingN
Female
Male
ValidN
MissingN
335
15
.9%
99.1%
335
15
AfricanAmerican
92.2%
5.6%
2.0%
.2%
AfricanAmerican
90.5%
5.1%
3.7%
.5%
AfricanAmerican
91.9%
7.1%
.3%
.3%
AfricanAmerican
92.8%
5.7%
.9%
.3%
.3%
Caucasian
5.8%
1.9%
.2%
Hispanic/Latino
Asian
.2%
.1%
Caucasian
Hispanic/Latino
ArabAmerican/
Chaldean
Asian
Caucasian
Hispanic/Latino
ArabAmerican/
Chaldean
.2%
Asian
MultiRacial
.3%
.3%
Caucasian
Hispanic/Latino
Asian
MultiRacial
ValidN
1244
ValidN
283
ValidN
418
ValidN
286
ValidN
257
MissingN
351
MissingN
162
MissingN
31
MissingN
65
MissingN
93
Mean
28.60
Mean
29.38
Mean
28.07
Mean
28.68
Mean
28.49
Std.Dev.
9.69
8.46
Std.Dev.
10.75
Std.Dev.
8.84
Std.Dev.
10.01
Range
865
Std.Dev.
Range
1563
Range
865
Range
1058
Range
861
25.9%
<22yearsold
17.7%
<22yearsold
32.8%
<22yearsold
22.7%
<22yearsold
27.2%
82.3%
>22yearsold
67.2%
>22yearsold
77.3%
>22yearsold
72.8%
<22yearsold
>22yearsold
74.1%
>22yearsold
166
TABLE4.2(cont.)
ASSAULTCHARACTERISTICS
HowLong
AgoAssault
Occurred
(Years)
(asof
12/31/13)
OverallSample
ValidN
MissingN
Mean
Std.Dev.
Range
ValidN
MissingN
Victim
Offender
Relationship
Stranger
BySight/
Nickname
Friend/Associate/
FamilyMember
Current/Past
IntimatePartner
Suspectknown,
butwedonot
knowrelationship
Assault
Involved
Multiple
Perpetrators
(gangrape)
ValidN
MissingN
No
Yes
1427
168
9.48
4.96
425
1468
127
TestingGroup1
TestingGroup2
TestingGroup3
TestingGroup4
(StrangerRape)
(NonStrangerRape)
(DNATestingMethods)
ValidN
MissingN
Mean
Std.Dev.
Range
ValidN
MissingN
43.7% Stranger
10.5%
35.7%
8.0%
2.1%
1522
73
77.9%
22.1%
ValidN
MissingN
No
Yes
430
15
7.77
2.42
411
445
0
100%
444
1
68.7%
31.3%
ValidN
MissingN
Mean
Std.Dev.
Range
ValidN
MissingN
432
17
7.47
2.28
412
449
0
BySight/Nickname
Friend/Associate/
FamilyMember
Current/Past
IntimatePartner
Suspectknown,but
wedonotknow
relationship
19.6%
57.7%
15.8%
6.9%
ValidN
MissingN
No
Yes
444
5
82.4%
17.6%
ValidN
MissingN
Mean
Std.Dev.
Range
ValidN
MissingN
285
66
17.98
3.14
1025
287
64
ValidN
MissingN
Mean
Std.Dev.
Range
ValidN
MissingN
Stranger
BySight/
Nickname
Friend/Associate/
FamilyMember
Current/Past
IntimatePartner
42.5%
12.2%
37.6%
7.7%
Stranger
BySight/
Nickname
Friend/Associate/
FamilyMember
Current/Past
IntimatePartner
ValidN
MissingN
No
Yes
286
65
78.0%
22.0%
ValidN
MissingN
No
Yes
280
70
6.55
2.33
411
287
63
25.8%
10.8%
54.7%
8.7%
348
2
83.9%
16.1%
167
TABLE4.2(cont.)
ASSAULTCHARACTERISTICS
OverallSample
TestingGroup1
TestingGroup2
TestingGroup3
TestingGroup4
(StrangerRape)
(NonStrangerRape)
(DNATestingMethods)
Alcohol
and/or
Drugs
Involvedin
theAssault
ValidN
MissingN
No
Yes
1434
161
70.8%
29.2%
ValidN
MissingN
No
Yes
423
22
68.3%
31.7%
ValidN
MissingN
No
Yes
444
5
82.4%
17.6%
ValidN
MissingN
No
Yes
288
63
82.3%
17.7%
ValidN
MissingN
No
Yes
279
71
69.2%
30.8%
Weapon
Usedinthe
Assault
ValidN
MissingN
No
Yes
1429
166
70.3%
29.7%
ValidN
MissingN
No
Yes
426
19
58.0%
42.0%
ValidN
MissingN
No
Yes
436
13
81.7%
18.3%
ValidN
MissingN
No
Yes
288
63
59.7%
40.3%
ValidN
MissingN
No
Yes
279
71
82.1%
17.9%
Physical
ForceUsed
inthe
Assault
ValidN
MissingN
No
Yes
1428
167
29.2%
70.8%
ValidN
MissingN
No
Yes
425
20
25.2%
74.8%
ValidN
MissingN
No
Yes
436
13
32.3%
67.7%
ValidN
MissingN
No
Yes
288
63
22.6%
77.4%
ValidN
MissingN
No
Yes
279
71
37.3%
62.7%
Time
Between
Assaultand
Medical
Forensic
Exam&
Sexual
AssaultKit
ValidN
MissingN
1247
348
ValidN
MissingN
393
52
ValidN
MissingN
365
84
ValidN
MissingN
248
103
ValidN
MissingN
241
109
Assaultandexam
sameday
61.6%
26.4%
Assaultandexam
sameday
62.6%
25.4%
Assaultandexam
sameday
53.2%
29.6%
Assaultandexam
sameday
67.7%
25.4%
Assaultandexam
sameday
66.6%
24.1%
Examnextday
afterassault
Exambeyondone
day
12.0%
Examnextday
afterassault
Exambeyondone
day
12.0%
Examnextday
afterassault
Exambeyondone
day
17.3%
Examnextday
afterassault
Exambeyondone
day
6.9%
Examnextday
afterassault
Exambeyondone
day
9.5%
168
Evaluating the Detroit SAK Testing PlanDescriptive Findings on the Forensic

Testing Outcomes
Background Context: An Overview of DNA Testing and CODIS
TosetthestageforthestatisticalresultsoftheforensicoutcomesfromtheDetroitSAKtesting
plan,wewillbeginwithanoverviewoftheprocessbywhichrapekitevidenceisanalyzedandentered
intoCODISandthetypeofCODIShitsthatcanemanatefromthattesting.Figure4.4belowsummarizes
thismultistageprocess.
FIGURE4.4StagesofSAKDNATesting,CODISEntry,andCODISResults
0.Serology
Screening
DNA
Testing
Rate
1.DNA
Testing
CODIS
Entry
Rate
2.CODIS
Entry
CODIS
Hit
Rate
3.CODIS
Hit
Types:
Serial
SAHit
Rate
4.SerialSexual
Assault
1.Offender
2.Forensic
3.OffenderForensic
Rapekittestingbeginswithaserologyscreening(Step0).Forensicscientistsexaminethe
evidenceintheSAK(e.g.,thevaginal,oral,andanalswabstakenfromthevictimsbody)todetermine
whethertherearebodilyfluidspresent(e.g.,semen,saliva,and/orblood).Iftherearebodilyfluids
present,thentheDNAwithinthosesamplescanbeextractedandanalyzed.Thereforeakitwillpass
fromserologyscreening(Step0)toDNAtesting(Step1)ifthereareprobativesamplesinthekitfor
analysis(i.e.,sampleswithbiologicalfluidsthatcanbeanalyzedforDNA).69Theprobabilitythatakitwill
passfromStep0toStep1canbequantifiedastheDNATestingRate.
69
Analternateapproach,theYscreeningmethod,skipsthetraditionalserologyscreeningstage(Stage0).Insteadof
screeningeachswab(oral,vaginal,anal)forsemen/saliva/blood,theforensicanalysttakesasmallcuttingfromeachswabto
determineifmaleDNAispresent(i.e.,ratherthanscreeningbodilyfluidsspecifically,theYscreeningmethodchecksformale
169
TherehavebeenmajortechnologicaladvancesoverthepasttwodecadesregardinghowDNA
testing(Step1)isperformed(seehistoricalanalysisinChapter3:WhySoManyUnsubmittedSAKsin
Detroit).Briefly,theseadvanceshavefocusedonthedevelopmentoffastermethodsforDNA
extraction,testingtechniquesthatcananalyzesmallersamplesofevidence,andautomationforquicker
turnaroundtime.Inthisactionresearchproject,wewantedtocontributetothegrowingliteratureon
DNAtestingbysystematicallystudyingtwodifferenttechniquesforidentifyingandisolatingspermcells
foranalysis(i.e.thedevelopmentoffastermethodsforDNAextractionnotedabove).Inthisproject
wecomparedtraditionalmethodswithselectivedegradationmethodsforidentifyingandisolating
spermcells.
ForthetraditionalmethodofDNAtesting(Step1),ifsemenispresent(fromtheserology
screen,Step0),theforensicscientistwilluseadifferentialextractionmethodtoseparatethesperm
fromtheothercellsinthesample,mostnotably,thevictimscells/DNA,whichisalsointhesample
(typicallyfromcellsofthevaginalwall).Theextractionmethodusesbothchemicalandmechanical
methodsofseparatingthespermfromtheothercellsinthesample.Oncespermcellsareisolated,then
theycanbechemicallybrokentoextracttheDNAtherein(usingPCRSTRtechnology;seeChapter3).
FortheselectivedegradationmethodofDNAtesting(Step1),theforensicscientistusesa
fasteractingchemicaltechniqueforisolatingthesperm.AfteraninitialremovalofnonspermDNA,
chemicalsareaddedthatdestroytheremainingnonspermcellsinthesample(i.e.,thecellsthatare
mixedwiththespermcells),leavingonlythespermcells(hencethetermselectivedegradation).The
combinedchemicalmechanicalseparationmethodsusedinthetraditionalapproachoftenleave
behindtracesofothercells/DNA,sothatthefinalsampletobeanalyzedisamixtureofmultipleDNA
DNAgenerally).IfthereismaleDNAinthesamples,thenthekitwillproceedwithrestoftestingtoidentifyandisolatesperm
cells(Step1).IfthereisnomaleDNAinthesamples,thenkitdoesnotproceedtotesting(unlessspecificcasebycase
circumstancessuggestthatadditionaltestingiswarranted).WiththeYscreeningmethod,theremayneedtobefollowup
testingtodeterminewhichspecificbodilyfluid(semen/saliva/blood)wasfoundonwhichswab,asthosedetailsmayneedtobe
presentedincourt(e.g.,thevaginalswabhadmaleDNAfromsemen).
170
sources(victimandsuspect[s]),whichmakesitmorechallengingandtimeconsumingfortheanalystto
interpretthefindings.Withselectivedegradation,thesamplethatwillbeanalyzedforDNAiscleaner
inthatmethodminimizesmixturesbydestroyingnonspermDNAthatismixedwiththespermcells;if
therearemultiplemaleassailants,themixtureofthosetwoDNAsamplesisstillintact,asthemethod
doesnotdestroysperm(fromanysource).Oncethespermcellsareisolated,thenthetestingcan
proceedperusual(PCRSTRmethods).
Foreithermethod,ifthetestingyieldsaDNAsamplethatmeetsminimumstaterequirements
oncompletenessandeligibilityforentryintoCODIS,theDNAprofileisuploadedintoCODIS(Step2).
TheprobabilitythatakitwillpassfromStep1toStep2canbequantifiedastheCODISEntryRate.
WhenaprofileisenteredintoCODIS,itiscomparedtoexistingDNAsamples,whichare
organizedintwoindexingsystems.TheoffenderindexcontainsknownDNAprofilesfrom
arrestees/convictedoffenders,obtainedattheirqualifyingoffense(i.e.,apriorcriminaloffensethat
metlegalrequirementsforCODISentry).TheforensicindexcontainsDNAsamplesobtainedatcrime
scenesthatmightmatchtosamplesintheoffenderindexormightmatchtofuturesamplesuploaded
intoCODIS.Figure4.5(below)isasimplifieddepictionofthestructureofCODIS.
FIGURE4.5AnOverviewoftheStructureofCODIS
171
IfanewlyenteredDNAprofilematchesanexistingDNAsampleinCODIS,itisreferredtoasa
hit/CODIShit(Step3).TheprobabilitythatakitwillpassfromStep2toStep3canbequantifiedas
theCODISHitRate.Dependingonwhetherthematchistoasampleintheoffenderindexorforensic
index(orboth),itcanbesubclassifiedintodifferenttypesofCODIShits:
OffenderHit=
ThenewprofilematchestheDNAofanoffenderprofilealreadyinCODIS(i.e.,thematchis
toasampleintheoffenderindex).
TheoffendermaybeaserialSEXUALoffenderifboththequalifyingoffenseandthenew
profileoffensearebothsexualassaults.
ForensicHit=
ThenewprofilematchestheDNAfromanunknownforensicsamplecollectedatacrime
scene(i.e.,thematchistoasampleintheforensicindex).
Theoffender(identitystillunknown)maybeaserialSEXUALoffenderifthequalifying
crimesceneevidenceandthenewprofileoffensearebothsexualassaults.
Offender
ForensicHit=
ThenewprofilematchesDNAthathasbeenlinkedtomultiplecases(oftentermedcase
tocaseassociations).Therearemanyscenariosthatwouldqualifyasanoffenderforensic
hit;threecommonexamples:
AnewprofilehitstoDNAintheoffenderindexandtherehavebeenmultipleprior
hitstothesameprofileinothercriminalcases;
AnewprofilehitstoDNAthathadbeenenteredintotheforensicindex(first),
whichwaslatersolvedwhenasubsequententryintotheoffenderindex
matchedtheDNA;thenewprofilematchestobothcases;
Anewprofilehascasetocaseassociationstoothernewprofiles,whichmatch
DNAalreadyinCODIS.WhenSAKsaretestedinlargebatches/volumes,itis
possiblethattherewillbecasetocaseassociationstootherSAKsfromthesame
batchofkits.
Theoffenderhasmultiplecriminalcasesinwhichhis/herDNAhasbeenlinked:the
qualifyingoffense,otheroffenses,andtheoffenseassociatedwiththenewprofile.
TheoffendermaybeaserialSEXUALoffenderifatleasttwoofthelinkedcasesaresexual
assaults.
172
SerialsexualassaultscanbeidentifiedviaCODISbyexaminingthequalifyingoffensetype,
qualifyingcrimesceneevidencetype,ortheoffensetypeofcasetocaseassociations.IftheDNAfroma
SAKmatchestoothersexualassaultoffenses(byanyofthescenariosdescribedaboveoffenderhitto
apriorsexualassault;forensichittoapriorsexualassault;casetocaseassociationstoothersexual
assaults),thehitrevealsapatternofserialsexualoffending(Step4).Theprobabilitythatakitwillpass
fromStep3toStep4canbequantifiedastheSerialSexualAssaultHitRate.Figure4.6(below)
depictstypesofCODIShits,highlightinghowserialsexualassaultscanbeidentifiedthroughCODIS.
FIGURE4.6TypesofCODISHitsandIdentificationofSerialSexualAssaultsinCODIS
WiththisbackgroundaboutthestagesofDNAtestingandCODISestablished,wewillpresent
descriptivedata(countsandpercentages)regardinghowmanySAKsprogressedthrougheachstage,
resultinginhowmanyCODIShits(andwhattypeofhitsandhowmanyserialsexualassaults).Wewill
presentthesedescriptivedatafortheoverallsampleof1,595kitsandthenwithineachofthefour
TestingGroups.Lookingahead,thefollowingsectionofthisreportwillpresentstatisticalmodelsthat
adjustforsamplingdifferencesbetweentheTestingGroupsandcomparefindingsacrosstheGroups.
173
Descriptive Results: CODIS Hits & Serial Sexual Assaults in the Overall Sample
NumberandTypeofCODISHitsintheOverallSample.TheCODIShitresultsforthetotal
sampleofSAKstestedintheDetroitSAKARParesummarizedinFigure4.7(below),currentthrough
December31,2013.70ThenumberspresentedthisinFigurearetheactualcountsforeachtypeofhit,
combinedacrossthefourTestingGroups;thepercentageshavenotbeenweightedtoaccountforthe
differencesinthesamplingdesignsacrossthefourGroups(seesectionEvaluatingtheDetroitSAK
TestingPlanInferentialFindingsontheEffectofVictimOffenderRelationshipandStatuteofLimitation
StatusonForensicTestingOutcomesforweighteddata).
FIGURE4.7NumberofCODISHitsintheOverallSample(N=1,595)
TARGETSAMPLESIZE
1600
ACTUALNUMBEROF KITSTESTED
1595
ProfilesEnteredintoCODIS
=49%
785
CODISHits
=28.5%
455
=58%
OffenderHits
339
=74.5%
ForensicHits
27

=6%
Offender&ForensicHits
89
&
=19.5%
70
NewprofilesareenteredintoCODISeveryweek,whichcanchangethesearchresults(i.e.,aSAKDNAprofilethatdoesnot
resultinahitwhenitisfirstenteredintoCODIScouldhaveahitlaterweeks,months,yearswhenanewsampleisentered).
Inourinterviewswithnationalforensicscienceexperts(seeAppendixB:ProjectMethodology),stakeholderssaidthatmosthits
occur23weeksafteraprofileisfirstenteredintoastateleveldatabase.Therefore,wewaited3weeksafterthelastSAKwas
enteredintoCODISbeforewetabulatedthesecounts.
174
Ofthe1,595SAKstested,therewere785eligibleCODISprofiles(a49%unweightedCODISentry
rate),whichresultedin455CODIShits(a58%unweightedconditionalCODIShitrate).Mostofthe455
CODIShitswereoffenderhits(74.5%)(i.e.,theDNAintheSAKmatchedaprofileintheoffenderindexof
database);6%wereforensichits(i.e.,theDNAmatchedtoanunknownidentityforensicsample);and
19.5%wereoffenderforensichits(mostofwhichwerehitstootherSAKstestedinthisproject).For
eachofthese455CODIShits,weexaminedthequalifyingoffense(i.e.,foroffenderhits,thecrimethe
offenderwasarrestedfor/convictedofthatresultedinhis/herDNAbeingenteredintoCODIS;for
forensichits,thetypeofcrimescenefromwhichtheDNAwasobtained)andthestateinwhichthe
qualifyingoffenseoccurred.The455CODIShitshittocrimes(including,butnotlimitedtosexual
assaults)in23states:Alabama,Alaska,Arizona,California,DistrictofColumbia,Georgia,Florida,Illinois,
Indiana,Kentucky,Louisiana,Maryland,Michigan(inothercounties),Minnesota,Missouri,NewMexico,
NorthCarolina,Ohio,Oregon,Pennsylvania,Tennessee,Texas,Wisconsin.
NumberofSerialSexualAssaultsintheOverallSample.Asnotedpreviously,thereare
multiplewaystoidentifyserialsexualoffendingviaCODIS.ForoffenderCODIShits,iftheSAKDNA
matchesaprofileinwhichthequalifyingoffensewasasexualassault,thenthehitrevealsapatternof
serialsexualassault(forforensichits,iftheSAKDNAmatchesaprofileinwhichthequalifyingcrime
sceneevidencewasasexualassault;foroffenderforensichits,ifatleasttwoofthecrimesthathave
beenlinkedtogetherbyDNAaresexualassaults).AshelpfulasCODISdatacanbeinidentifyingserial
rapes,itisimportanttonoteakeylimitationofthisdatasource:offendersmayhavecommittedother
sexualassaultoffensesthatarenotreflectedinCODIS(e.g.,therewasnorapekit,therapekitwasnot
analyzed).Therefore,CODISdataarelikelyanunderestimatethetruescopeofserialsexualassaults.
Withthatlimitationinmind,weexaminedeachofthe455CODIShitstodeterminehowmany
hitswereserialsexualassaults.Figure4.8(nextpages)summarizesthatanalysis.Overall,127serial
sexualassaultswereidentified:ofthe339offenderhits,36wereserialsexualassaults(typicallybecause
175
FIGURE4.8NumberofSerialSexualAssaultsintheOverallSample(N=1,595)
CODISHits
455
OffenderHits
339
ForensicHits
27
89
36
Serial
SexualAssaults
303
NonSerial
SexualAssaults*
19
Serial
SexualAssaults
8
NonSerial
SexualAssaults*
72
Serial
SexualAssaults
17
NonSerial
SexualAssaults*
32=CODISqualifying
offensewassexual
9=CODISqualifyingforensicsamplewas
fromasexualassault
assault
4=casetocase
associationstoanother
DetroitARPSAK**
8=casetocaseassociationstoanother
DetroitARPSAK**
fromasexualassaultANDcasetocase
associationtoanotherDetroitARPSAK**
*BasedonCODISdataonly;offendersmayhavepreviousarrests/convictionsforsexualassaultoffensesin
theircriminalhistoryrecordsthatareNOTinCODIS;thereforethesecomputations,whicharebasedsolely
oninformationinCODIS,aremostlikelyanunderestimateofthetrueextentofserialsexualoffending.
**Therewere51Twinsand9Tripletsinthisdataset(theDetroitARP)(i.e.,60totalcasetocase
associationswithinthisdataset;4+8+2+39+7=60).ATwiniswhentwoSAKsmatchthesameoffender;
aTripletiswhenthreeSAKsmatchthesameoffender.
7=CODISqualifyingoffensewassexual
assaultANDcasetocaseassociationto
anotherDetroitARPSAK**
DetroitARPSAK**
anothersexualassaultcase(nonDetroitARP)
19=casetocaseassociationstoothersexual
assaultcases(nonDetroitARP)
1=CODISqualifyingforensicsamplewasnot
asexualassault,butitlatermatchedan
offenderprofileinadifferentcrimethatwas
asexualassault,whichlatermatcheda
DetroitARPSAK
176
thequalifyingoffenseinCODISwasasexualassault);ofthe27forensichits,19wereserialrapes
(usuallybecausethequalifyingcrimescenewasalsoasexualassault);ofthe89offenderforensichits,
72wereserialsexualassaults(mostlycasetocaseassociationswithintheDetroitARPdataset).
Descriptive Results: CODIS Hits & Serial Sexual Assaults, by Testing Group
NumberandTypeofCODISHits,ByTestingGroup.Weusedastratifiedsamplingdesigninthis
projectsothatwecouldexamineCODIShitratesandserialsexualoffendingamongdifferentkindsof
sexualassaultcases.PractitionersinDetroit(aswellasotherstakeholdersweinterviewedatthestate
andnationallevel)haddifferingbeliefsabouttheusefulnessofSAKtestingforstrangervs.nonstranger
rapesandforkitsassociatedwithcasesthatmightbebeyondthestatuteoflimitations.Wedesigned
thetestingplaninthisprojecttoinformthesedebates,andtothatend,weexaminedthenumberof
CODIShitswithineachTestingGroup(seeFigures4.9,4.10,4.11,and4.12,followingpages).71
Beginningwiththestrangerandnonstrangerperpetratedsexualassaults(Figures4.9and4.10,
nextpage),itisimportanttopauseandconsiderwhataCODIShitmaymeanwhatinformationitmay
beprovidinggiventhenatureofthevictimoffenderrelationshipintheassault.Inastrangerrape,the
assailantsidentityisunknown,sothe104offenderCODIShitsinTestingGroup1meanthat
potentially104rapesweresolvedbyDNAtesting.ThepotentiallycaveatiscriticalbecauseaCODIS
hitisnotconfirmatoryitisinformationthatcanbeusedintheinvestigation(andforstrangerrapes,it
givesinvestigatorsapromisingleadontheoffendersidentity).72InTestingGroup1,therewerealso10
71
InthefollowingsectionofthisreportEvaluatingtheDetroitSAKTestingPlanInferentialFindingsontheEffectofVictim
OffenderRelationshipandStatuteofLimitationStatusonForensicTestingOutcomes,wewillpursuethisissueofsimilarities
anddifferencesbetweenthefourTestingGroupsinmoredetail.
72
WehadacaseinTestingGroup1inwhichtheoffenderhitwasnottotheassailant,buttothevictimsconsensualpartner,
whohappenedtohaveacriminalrecordandaprofileinCODIS(completelyunrelatedtothesexualassault).Thevictimhadhad
consensualsexwithin72hours(orso)oftheassault,andresultingoffenderhitwastothevictimsboyfriend,nottotheman
whohadsexuallyassaultedher.Therefore,thecasewasnotsolvedbytheoffenderhit,whichhighlightsthatCODIShitsare
investigativeleadsthatrequirefollowup.
177
FIGURE4.9NumberofCODISHits,TestingGroup1(Stranger)(n=445)
FIGURE4.10NumberofCODISHits,TestingGroup2 (NonStranger)(n=449)
TESTINGGROUP1:STRANGERRAPESAKs
445
TESTINGGROUP2:NONSTRANGERRAPESAKs
449
=54%
239
=35%
CODISHits
156
=40%
180
=23%
CODISHits
103
=65%
OffenderHits
104

=67%
ForensicHits
10

=6%
42
&
=27%
=57%
OffenderHits
84
=83%
ForensicHits
5
=5%
14
&
=13.5%
178
forensicCODIShits,indicatingthattheoffendersidentitywasnotknowninbothaprioroffenseandin
thecurrentrape,butDNAevidenceisbeingcompiledandpreservedforpotentialprosecutionlater,if
theoffendersidentityissolved.InthisGroupofstrangerperpetratedrapes,therewerealso42
offenderforensicCODIShits,meaningthattheidentityoftheoffenderwasrevealed(possibly)via
testingandtherearelinkstomultipleothercrimescommittedbythesameoffender.
Turningtononstrangersexualassaults,stakeholdersdisagreedabouttheutilityofSAKtestingif
theidentityoftheperpetratorwasalreadyknown.Somepractitionersarguedthattestingisstill
importanttoconfirmidentityandestablishsexualcontactbetweenthevictimandassailant(i.e.,to
rebutnocontactdefense).Mostofthe103CODIShitsinTestingGroup2wereoffenderhits(i.e.,
testingconfirmedidentity)(seeFigure4.10,priorpage),buttherewerealsofiveforensicCODIShits,
meaningthematchingforensicsamplesinCODISwereassociatedwithunknownoffenders(i.e.,the
matchwastoDNAintheforensicindex).Bytestingnonstrangersexualassaultkits(inwhichthe
identityoftheoffenderwasknown),therewerefiveothercriminalcasesthatmayhavebeensolved
(pendingfurtherinvestigation).The14offenderforensichitsinthisGroupindicatethattheseknown
offendershadcommittedmultiplecrimes,whichwerenowlinkedtogetherbyDNA.
Local,state,andnationalstakeholdersalsohaddifferingopinionsaboutwhetherolderkits
thosepresumedtobebeyondthestatuteoflimitationsshouldevenbetested,asareasonableuseof
publicfunds.ThosewhoadvocatedfortestingpresumedSOLexpiredSAKsnotedthatthecasesmight
stillbeeligibleforprosecutionafteranindepthreviewofthecaseanditscircumstances,butifnot,
thenperhapstheevidenceofthesexualassaultcouldbeusedincourtifassailantreoffended.Inother
words,someofthekeygainsfortestingolderkitswouldmanifestintheprosecutionphase,but
examiningthatstagewasbeyondthetimelineandscopeofthisstudy.However,theutilityofthesekits
tolaterprosecutionsassumesthattestingwouldyieldCODIShits,whichissomethingthatcouldbe
examinedwithinthisactionresearchproject.
179
AsshowninFigure4.11(nextpage),73ofthe90CODIShitswereoffenderhits.TheSAKsinthis
Groupwerenotstratifiedbyvictimoffenderrelationship,butwedidcollectthisinformationfromthe
policefiles(seeEvaluatingtheDetroitSAKTestingPlanDescriptiveFindingsAbouttheCases/SAKs
TestedinthisProject);fromthoseadditionaldata,wewereabletodiscernthat24ofthe73offender
hitswerestrangerrapes(i.e.,24mayhavebeensolvedbyDNAtesting).73Therewerealso5forensic
hitsinthisGroup,indicatingwhatmaybealonghistoryofunapprehendedoffending(giventhe
age/dateofthekitstestedinthisTestingGroup),buttheevidenceispreservedintheeventthecaseis
latersolved.Intwoofthesefiveforensichits,therewasinsufficientinformationinthepolicefilestobe
abletodeterminevictimoffenderrelationship,butinthethreecasesinwhichwecoulddetermine
whethertheassaultwasstrangervs.nonstranger,twoforensichitsweretononstrangerassaults(i.e.,
twoothercrimesmayhavebeensolvedbytestinganonstrangerkit)(theothercasewasastranger
rape).Ofthe12offenderforensichitsinthisGroup,eightwerestrangerrapesthatwerepotentially
solvedbytesting;again,whetherprosecutionispossibleinthoseeightcaseswillhavetobedetermined
afterathoroughreviewoftheevidence.
ThefocusofTestingGroup4(DNATestingMethod)isanexperimentalcomparisonofCODIS
entryratesfortwoDNAtestingmethods;therefore,CODIShitresultsarenotaprimaryfocusforthis
group,butforcompleteness,wehavesummarizedthosedatainFigure4.12(nextpage).74Similartothe
distributionsintheotherTestingGroups,mostoftheCODIShitsinTestingGroup4wereoffenderhits
(78),andbasedoninformationinthepolicefiles,wewereabletodeterminethatmostwerenon
strangercases(54)and11werestrangerperpetrated(i.e.,11possiblerapessolved)(13hadmissing
dataonvictimoffenderrelationship).Ofthesevenforensichitsinthissample,fourwerenonstranger
73
In15ofthese73hits,itwasnotpossibletodeterminevictimoffenderrelationship,eitherbecausethepolicefilewasmissing
orbecausethedocumentationthereinwassosparsethatwewereunabletomakeadeterminationastotherelationship.
74
Bywayofpreview,thestatisticalanalysesinsectionEvaluatingtheDetroitSAKTestingPlanInferentialFindingsonthe
EffectofVictimOffenderRelationshipandStatuteofLimitationStatusonForensicTestingOutcomeswilluseTestingGroup4
asacomparisonsampletocontrastCODIShitsratesforpresumedSOLexpiredvs.nonexpiredcases.
180
FIGURE4.11NumberofCODISHits,TestingGroup3
(PresumedSOLExpired)(n=351)
FIGURE4.12NumberofCODISHits,Testing Group4
(DNATestingMethod)(n=350)
TESTINGGROUP3:PRESUMEDSOLEXPIRED
351
TESTINGGROUP4:DNATESTINGMETHOD
350
=49%
173
=26%
CODISHits
90
=55%
193
=30%
CODISHits
106
=52%
OffenderHits
73

=81%
ForensicHits
5

=5.5%
12
&
=13%
=55%
OffenderHits
78
=73.5%
ForensicHits
7
=7%
21
&
=20%
181
rapes,againsuggestingthattheremayhavebeenfourothercasesinCODISthatweresolvedbytesting
arapekitinwhichtheidentityoftheassailantwasknown.Ofthe21offenderforensichits,ninewere
strangerperpetrated,ninewerenonstranger,andthreecouldnotbedetermined.Thus,thepatternof
resultsinthisTestingGroupappearsimilartoTestingGroup3(PresumedSOLExpired)whichisthe
otherGroupinwhichcaseswerenotstratifiedbyvictimoffenderrelationship.MostCODIShitswere
offenderhits,withasizablenumberofhitsinstrangerrapescasesthatwerepotentiallysolvedby
DNAtesting.Theforensichitsincludedmatchestononstrangerperpetratedcrimes,sothattheidentity
oftheperpetratorinothercriminalcasesmayhavebeensolvedbytesting.Theoffenderforensichits
showapatternofmultiplecrimes,bothstrangerperpetratedandnonstrangerperpetrated.
NumberofSerialSexualAssaults,ByTestingGroup.Wealsoexaminedthenumberofserial
sexualassaultswithineachofthefourTestingGroups(seeFigures4.13,4.14,4.15,and4.16,following
pages).InTestingGroup1(strangerrapes)(Figure4.13,nextpage),10ofthe104offenderhitswere
serialsexualassaults(mostbecausethequalifyingoffenseinCODISwasalsoasexualassault).Among
theforensichitsinthisGroup,7(of10)wereserialsexualassaults,meaningthattheseperpetratorshad
committedmultiplerapesbuthavenotyetbeenidentified.IntheoffenderforensicCODIShits,34(of
42)wereserialsexualassaults,mostwerecasetocaseassociationstootherDetroitSAKARPSAKs.
InTestingGroup2(nonstrangerrape)(Figure4.14,followingpage),fourofthe84offenderhits
wereserialsexualassaults.Theidentityoftheseoffenderswasnotinquestionthevictimknewthe
assailantbutbytestingtheSAK,itbecameclearthattheassailanthadcommittedpreviousrapes.As
oneDetroitstakeholdernotedearlyonintheproject,itchangesthingsfromahesaid,shesaidcase
toahesaid,shesaid,shesaidcase...thattakesawayalotofdoubtinthemindsofthejury[toknow]
thatthiswasntaonetimethingormiscommunicationorwhateverthedefensetriestoargue.Among
thefiveforensichitsinthisgroup(i.e.,testingtheknownperpetratorSAKmayhavesolvedtheidentity
ofoffendersinfiveothercases),threewereserialsexualassaults.Inoneinstance,theunknown
182
FIGURE4.13NumberofSerialSexualAssaults,TestingGroup1(Stranger)(n=445)
CODISHits
156
OffenderHits
104
10
Serial
SexualAssaults
6=CODISqualifying
offensewassexual
assault
4=casetocase
DetroitARPSAK**
94
NonSerial
SexualAssaults*
ForensicHits
10
7
Serial
SexualAssaults
3
NonSerial
SexualAssaults*
fromasexualassault
DetroitARPSAK**
42
34
Serial
SexualAssaults
8
NonSerial
SexualAssaults*
DetroitARPSAK**
1=CODISqualifyingforensicsamplewasnot
asexualassault,butitlatermatchedan
offenderprofileinadifferentcrimethatwas
asexualassault,whichlatermatcheda
DetroitARPSAK
183
FIGURE4.14NumberofSerialSexualAssaults,TestingGroup2(NonStranger)(n=449)
CODISHits
103
OffenderHits
84
4
Serial
SexualAssaults
4=CODISqualifying
offensewassexual
assault
0=casetocase
DetroitARPSAK**
80
NonSerial
SexualAssaults*
ForensicHits
5
3
Serial
SexualAssaults
2
NonSerial
SexualAssaults*
fromasexualassault
DetroitARPSAK**
associationtoanotherDetroirARPSAK**
14
11
Serial
SexualAssaults
3
NonSerial
SexualAssaults*
DetroitARPSAK**
184
forensicsamplewasfromarape(i.e.,astrangerrape),andtheassailantreoffendedthistime
someoneknowntohimandwhenthatkitwastested,itmatchedtobothrapes.Ofthe14offender
forensichitsinthisGroup,most(11)wereserialsexualassaults,largelyduetocasetocaseassociations
tootherSAKsinthisactionresearchproject.
ThenumberofserialsexualassaultsinTestingGroup3(PresumedSOLExpired)ispresentedin
Figure4.15(followingpage).Amongthe73offenderCODIShitsinthisgroup,therewere19serial
sexualassaults,allduetomatchesintheoffenderindexinwhichthequalifyingoffenseshadbeen
sexualassaults.OfthefiveforensichitsinthisGroup,threewereserialsexualassaults;twoofwhich
wereinstancesinwhichthequalifyingforensicsamplewasalsoarape.Unfortunately,thepolicefiles
forthesecasesweremissing,sowewereunabletodeterminevictimoffenderrelationshipsinthese
cases.Mostofthe12offenderforensichitswerealsoserialsexualassaults(7),mostlyduetocaseto
caseassociationswithintheactionresearchprojectsample.
InTestingGroup4(DNATestingMethod),therewerefewerserialsexualassaultswithinthe
offenderCODIShitsascomparedtotheotherTestingGroups:ofthe78offenderhitsinthisGroup,
threewereserialsexualassaults(allmatchestocasesinwhichthequalifyingoffensewasanothersexual
assault)(seeFigure4.16,followingpages).OfthesevenforensichitsinthisGroup,sixwereserialsexual
assaults;ofthosesix,threewerenonstrangerrapesthathadcasetocaseassociationswithother
DetroitSAKs,whichhittounknownforensicsamplesinCODIS.Inotherwords,inthreecases,testinga
nonstrangerrapekityieldedcasetocaseassociationswithothersexualassaults,whichlinkedtoa
previouslyunknownforensicsamplefromadifferentcrime(i.e.,potentiallysolvingthatothercrime,
andshowingthattheoffenderhadmultipleotheroffenseslinkedbyDNA).Ofthe21offenderforensic
hitsinthisGroup,20wereserialsexualassaults,mostduetocasetocaseassociationswithintheaction
researchprojectdataset.
185
FIGURE4.15NumberofSerialSexualAssaults,TestingGroup3(PresumedSOLExpired)(n=351)
CODISHits
90
OffenderHits
73
19
Serial
SexualAssaults
19=CODISqualifying
offensewassexual
assault
0=casetocase
DetroitARPSAK**
54
NonSerial
SexualAssaults*
ForensicHits
5
3
Serial
SexualAssaults
2
NonSerial
SexualAssaults*
fromasexualassault
DetroitARPSAK**
12
7
Serial
SexualAssaults
5
NonSerial
SexualAssaults*
DetroitARPSAK**
186
FIGURE4.16NumberofSerialSexualAssaults,TestingGroup4(DNATestingMethod)(n=350)
CODISHits
106
OffenderHits
78
3
Serial
SexualAssaults
3=CODISqualifying
offensewassexual
assault
0=casetocase
DetroitARPSAK**
75
NonSerial
SexualAssaults*
ForensicHits
7
6
Serial
SexualAssaults
1
NonSerial
SexualAssaults*
fromasexualassault
DetroitARPSAK**
21
20
Serial
SexualAssaults
1
NonSerial
SexualAssaults*
DetroitARPSAK**
187
Inthenextsectionofthisreport,wewilldelvedeeperintothedatatounderstandthe
similaritiesanddifferencesbetweenthefourTestingGroups.Ourprimaryaiminthisnextsetof
analysesistodevelopandevaluatestatisticalmodelsthataccountforthedifferencesinthesampling
designsacrosstheGroups(i.e.,weighting)todetermineiftheratesofCODISentry,CODIShits,and
serialsexualassaultsaresignificantlydifferentbetweentheTestingGroups.Theseanalysescanhelp
informdebatesaboutwhetherallSAKsshouldbetested(includingnonstrangerandpresumedSOL
expiredSAKs)andwhetheritispossibletocreateempiricallyinformedprioritizationguidelinesforSAK
testingtotrytomaximizethelikelihoodofobtainingCODIShits.
Evaluating the Detroit SAK Testing PlanInferential Findings on the Effect of VictimOffender Relationship and Statute of Limitation Status on Forensic Testing Outcomes
Overview
KeyAnalyticGoals.IndevelopingtheDetroitSAKtestingplan,ourgoalwastocraftadesign
thatcouldinformpolicyandpracticedebatesabouttestall/testsomeSAKs.75Asnotedpreviouslyin
thischapter,therearemanyfactorstoconsiderwhendevelopingSAKtestingpoliciespublicsafety,
socialjustice,survivorshealthbutthisresearchfocusedononlyoneoutcome:forensictesting
outcomes,specificallythenumberofCODIShits.Ifajurisdictiondidnothavetheresourcestotestall
SAKs,isitpossibletodevelopempiricallyinformedtestingplans,wherebySAKscouldbeprioritizedby
thelikelihoodofyieldingaCODIShit?Forexample,iftheprobabilityofaCODIShitwassignificantly
higherforsomecasesvs.others,thenthatinformationcouldprovideempiricallybasedguidelinesfor
75
TheDetroitSAKtestingplanalsosoughttoevaluateanewmethodforDNAtesting(selectivedegradation)becausewhatever
prioritizationssystemsmayormaynotbeusedbyacommunity,itisimportanttoconsideriftherearealternativeDNAtesting
methodsthatcouldofferfaster,lessexpensivetestingoptions,withoutsacrificingaccuracy.Themethodsandresultsofthat
componentofthetestingplan(TestingGroup4DNAMethod)willbepresentedinthefollowingsectionofthischapter.
188
prioritization;alternatively,iftheprobabilityofaCODIShitwasstatisticallynodifferentforcertaincases
versusothers,thenthatwouldsuggestthatsuchfactorswouldnotbegoodcriteriaforprioritizingcases.
TheDetroitSAKtestingplanevaluatedtwoprimaryselectioncriteriavictimoffenderrelationship(i.e.,
strangervs.nonstrangerperpetratedsexualassaults)andstatuteoflimitationstatusasthesewere
thetwofactorsthatstakeholdersatthelocal,state,andnationalleveldisagreedaboutthemost(e.g.,
prioritizethestrangercases,skiptheSOLexpired).Thesetwopiecesofinformationarerelatively
easytoaccesspretesting:SOLcanberoughlydeterminedbythedateofthecase/kitandvictim
offenderrelationshipistypicallyrecordedinpolicefilesasastandarddatafield.Assuch,thesedata
pointscouldconceivablybeusedtoscreenandprioritizeSAKsfortesting.
Inadditiontothesetwofocalvariables,stakeholdersalsoexpressedinterestinassessing
whetherotheraspectsofthevictim,assailant,and/orassaultmightbeusefulscreeningcriteria(e.g.,
weaponuseintheassault,thetimebetweentheassaultandwhentheSAKwascollected).Stakeholders
generatedawishlistofpossiblescreeningvariablestoconsider,manyofwhichweredatafieldsthat
maynotbeeasily/quicklyaccessible.Giventhatgoalwastoevaluatepossibleselectioncriteriathat
couldbetime/costefficienttoimplement(andtrackingdowndatafieldsthatarehardtoaccesswould
notbeefficient),wehadtofocusthiscomponentoftheevaluationonalimitednumberofvariablesthat
wererelativelyaccessible.76Intheend,wewereabletocodesomeinformationaboutthevictim(e.g.,
gender,race,age),assailant(e.g.,gender,race,age),andtheassault(e.g.,multipleperpetrators
involved,alcohol/druguse,weaponuse,timebetweenassaultandexam)toexplorewhetherthese
factorsweresignificantlyassociatedwithforensictestingoutcomes.
76
Werecognizethatwhatinformationisrelativelyaccessiblevariesacrossjurisdictions.Incommunitieswithwelldeveloped
informationmanagementsystems,moreinformationmaybeavailableabouteachSAK/caseinordertomakeaSAKtesting
decision.Giventhatthiswasaresearchproject,wedecidedtoinvestefforttocodethefilessothatwecouldevaluatethe
predictiveutilityofvictim,assailant,andassaultcharacteristicsonforensictestingoutcomes.Inpractice,screeningonsuch
variablesiftheresultsshowedthatsuchscreeningwaswarrantedwouldbedifficulttoimplementinDetroit,butmightbe
morefeasibleinotherjurisdictions(iftheresultssuggestedthatsuchvariableswereinfluentialtoforensictestingoutcomes).
189
DevelopingStatisticalModelstoTesttheEffectofVictimOffenderRelationshipandSOL
StatusonForensicTestingOutcomes.Thebroadestfocalpopulationforthisstudyconsistedofall
previouslyuntestedDetroitSAKsanddrawingconclusionsaboutthispopulationwastheobjectivefor
someanalyses.OursamplingdesignallowedustoobtainarepresentativesampleofSAKsfromitby
combiningdatafrommultiplegroups,withdifferentsamplingweightsappliedtoSAKsfromeachTesting
Grouptoaccountforthecomplexsamplingdesignandensurethattheresultswouldbettergeneralize
tothefocalpopulation(seeAppendixB:ProjectMethodology).However,wealsoneededasampling
strategythatpermitteddrawingconclusionsaboutmorenarrowlydefinedsubpopulations.Figure4.17
(below)illustratesthefocalpopulationofinterestandthesubpopulationsrepresentedbyeachGroup.
FIGURE4.17TheDetroitSAKTestingGroupsandPopulationsRepresented
Group3
SRSdesign
UntestedDetroitSAKs(19802009)
PresumedSOLexpired
SOLunexpired
(19802001)
(20022009)
Group4T
Group4 Randomassignment
SRSdesign
Group4
D
Nonadjudicated
Strangerrapes
Group1
SBYdesign
Nonadjudicated
Nonstrangerrapes
Group2
SBYdesign
Figure4.17:RelationshipofTestingGroupstoSubpopulationsoftheDetroitSAKCollection.Thelargerectanglerepresents
thepopulationofpreviouslyuntestedDetroitSAKscollectedbetween1980and2009.Itisdividedintotwosmallerpartsbased
onwhetherornotthestatuteoflimitations(SOL)haspresumablyexpired.Thesubpopulationofnonadjudicated,SOL
unexpiredSAKswasfurtherdividedbyvictimoffenderrelationship(therectangleshaveunequalsizesbecausenonstranger
rapesoccurmoreoftenthanstrangerrapes).Groups14aresamplesdrawnfromspecificsubpopulationsusingdifferent
samplingstrategies(SBY=stratifiedsamplingbyyear;SRS=simplerandomsampling).Groups1and2represent
subpopulationsofthenonadjudicated,SOLunexpiredSAKs,whileGroup4representsthebroaderSOLunexpired
subpopulation.Group3representsthepresumedSOLexpiredsubpopulation.Groups13allreceivedtraditionalDNAtesting,
whileSAKsinGroup4weredividedintotwosubgroupsbyrandomlyassigningthemtoreceiveeithertraditionalDNAtesting
(Group4T)orDNaseselectivedegradationDNAtesting(Group4D).SAKsinshadedgroups(Groups1and4)weretestedbya
privatevendorlaboratory,SAKsinunshadedgroups(Groups2and3)weretestedbyaforensiclaboratoryaffiliatedwitha
university.Personnelfromthestatepoliceforensicsciencedivisionconductedqualitycontrolvisitstoensurebothfacilities
usedequivalentDNAtestingproceduresandmetapplicablestandards.
190
Toevaluatetheeffectofvictimoffenderrelationshiponforensictestingoutcomes,we
examinedwhetherCODISresultsvariedbetweenSAKsassociatedwithstrangerversusnonstranger
rapes.Becauseanysuchdifferenceswouldbemostrelevantifthestatuteoflimitationshasnotyet
expired,wedrewsamplestosupportthiscomparisonstrictlyfromthesubpopulationofnon
adjudicated,SOLunexpiredSAKs.TestingGroup1representsthesubpopulationofnonadjudicated,
SOLunexpiredSAKsresultingfromstrangerrapes,andTestingGroup2representsthesubpopulationof
nonadjudicated,SOLunexpiredSAKsresultingfromnonstrangerrapes.
SamplingforTestingGroups1and2attemptedtoobtainequalnumbersofSAKsfromboth
strangerandnonstrangerrapesforeachyearbetween2002and2009,eventhoughSAKsfromdifferent
yearswerenotnecessarilyequallycommon.Thisdisproportionatestratifiedsamplingapproach
guaranteedthatdatafromeachsubgroupofSOLunexpiredSAKsdefinedbythecombinationofvictim
offenderrelationshipandyearwouldbepresentinthefinalsample.Withappropriateweighting,data
fromTestingGroups1and2canbecombinedtorepresenttheSOLunexpiredsubpopulation(see
AppendixB:ProjectMethodologyforadditionaldetails).
TestingGroup3(PresumedSOLExpired)wasobtainedbydrawingasimplerandomsamplefrom
thesubpopulationofpresumedSOLexpiredSAKs,withoutstratifyingbyyear.Datafromthisgroupcan
beanalyzedonitsown,orweightedandcombinedwithdatafromTestingGroup4(DNAMethod/SOL
Unexpired)tofacilitatecomparingoutcomesforSOLexpiredSAKstothoseofSOLunexpiredSAKs.
TestingGroup4wasplannedasacomparisonofforensictestingoutcomesbetweentwoDNA
testingmethods,butitcouldalsoserveasacomparisontoTestingGroup3forexaminingtheeffectof
SOLstatus.TestingGroup4wasobtainedbydrawingasimplerandomsamplefromthesubpopulation
ofSOLunexpiredSAKs,thendividedintotwosubgroupsbyrandomlyassigningequalnumbersofSAKs
toreceiveeithertraditionalDNAtesting(Group4T)orDNaseselectivedegradationDNAtesting(Group
4D).
191
Inourstatisticalmodels,weusedcontinuationratiomodels(Agresti,2002;Hosmer,Lemeshow,
&Sturdivant,2013)toquantifyhowmanySAKsproceededthrougheachstageofDNAprocessing
(Figure4.4replicatedbelowforreference).
FIGURE4.4StagesofSAKDNATesting,CODISEntry,andCODISResults
0.Serology
Screening
DNA
Testing
Rate
1.DNA
Testing
CODIS
Entry
Rate
2.CODIS
Entry
CODIS
Hit
Rate
3.CODIS
Hit
Types:
Serial
SAHit
Rate
4.SerialSexual
Assault
1.Offender
2.Forensic
3.OffenderForensic
Thecontinuationratiomodelsexamined:(1)CODISentryrate,whichistheproportionoftestedSAKs
thatyieldedDNAprofilessuitableforuploadintoCODIS;(2)CODIShitrate,whichistheproportionof
CODISentriesfromDetroitSAKsthatyieldhitstootherCODISrecords;77and(3)serialassaultrate,
whichistheproportionofCODIShitsthatareassociatedwithserialsexualassaults.
TheCODISentryrateasdefinedaboveisanunconditionalrate,whichmeansthatthe
denominatorfortheproportionisthetotalnumberofSAKstested.Incontrast,theothertesting
outcomes(CODIShitrateandserialsexualassaultrate)areconditionalratesbecausethedenominator
includesonlythesubsetofSAKsthatmeetparticularconditions,suchashavingyieldedaCODISentryor
aCODIShit.Forexample,CODIShitrateisconditional,wherebythedenominatoristhenumberof
CODISentries;serialsexualassaultrateisalsoconditional,wherebythedenominatoristhenumberof
CODIShits.Inourpresentationofthemodelresults,wewillalsodiscussunconditionalversionsofthe
77
ThisincludeshitstootherDetroitSAKstestedduringthisproject(i.e.,casetocasehits)inadditiontohittorecordsinCODIS
thatexistedpriortoenteringDNAprofilesextractedfromtheDetroitSAKs.
192
varioushitratesmentionedaboveandwillexplicitlycallthemunconditionalrateswhendoingso.
Distinguishingbetweenconditionalandunconditionalratesiscrucialtocorrectlyinterpretingthe
findings:thereisalargesubstantivedifferencebetweenstating,forinstance,thattheconditionalserial
sexualassaulthitrateis20%(meaning20%ofallCODIShitsobtainedfromDetroitSAKsareassociated
withserialsexualoffenders)andthattheunconditionalserialsexualassaultrateis20%(meaningthat
20%ofallDetroitSAKstestedwereassociatedwithserialsexualoffenders).78
Tounderstandtheeffectofvictimoffenderrelationshipandstatuteoflimitationsstatuson
forensictestingoutcomes,wequantifiedandcomparedtheCODISentryrates,CODIShitrates,and
serialassaultratesforstrangerversusnonstrangerassaultsandforSOLexpiredversusSOLunexpired
assaults.Wefirstpresenttheresultsintermsoftheconditionalratesthataredirectlyestimatedbythe
continuationratiomodels,thentranslatetheresultsintooddsratios(ORs),estimatesofrelativerisk
(RR),anumberneededtosubmit(NNS)statistic,andunconditionalratestofacilitatetheinterpretation
oftheresults(seeAppendixB:ProjectMethodology).79
DevelopingStatisticalModelstoTesttheEffectofOtherVariablesonForensicTesting
Outcomes.Toassesswhethercharacteristicsofvictim,assailant,andassaultcouldbeusefulfor
empiricallybasedSAKprioritization,weusedlogisticregressiontodeterminewhethersuchfactors
predictwhethereachSAKyieldedaCODIShit(coded0=no,1=yes)(i.e.,forsimplicity,wedidnottest
continuationratiomodelsforentry,hit,andserial,asabove;wefocusedonthekeyoutcomeCODIS
hit).80WeanalyzedthedatafromTestingGroups13separatelybecausetheeffectofagivenpredictor
mayvaryacrosssubpopulationsofDetroitSAKs;theseanalyseswerenotconductedforTestingGroup4
78
Thesearebothhypotheticalvalues,notouractualfindings.
79
Definitionsandexamplesofeachofthesestatisticalindiceswillbepresentedinthefollowingsection,usingdatafrommodel
resultsasworkingexamples.
80
Themean(average)ofthisbinaryvariableacrossasetofSAKsistheunconditionalCODIShitrate,whichisequaltothe
proportionofSAKsthatyieldedhits.
193
becausethepurposeofthatcomponentoftheprojectwastocomparedifferentDNAmethods(and
serveasacomparisontoTestingGroup3forevaluatingtheimpactofSOLstatus).
Withrespecttocharacteristicsofvictim,therewaslimitedvariabilityingenderand
race/ethnicity(seeTable4.2priorpages),sowefocusedonageasapossiblepredictor,dividedintotwo
levelsbasedontheageofconsentinMichigan(015yearsvs.16+years).Therewasalsoinsufficient
variabilityinassailantgenderandrace/ethnicity(seeTable4.2),sowetestedonlyage.Thereweretoo
fewassailantslessthan18yearsoldtodivideassailantagebywhethertheywouldbeconsideredminors
forlegalpurposes(017yearsvs.18+years),soweinsteaddividedassailantsintothreegroups(021
years,22+years,orunknown),whichwasamoresensiblegroupinggiventhedistributionofthis
variable.Theoldergroupservedasthereferencelevelforeachagevariable.
WithrespecttocharacteristicsofassaultthatcouldinfluencetheprobabilityofaCODIShit,we
wereconstrainedtoasetofvariablesthatwerereliablyaccessibleinthepolicefiles(seeprior
discussion).WecodedbinaryindicatorsofwhetherthecasefilesassociatedwitheachSAKexplicitly
documentedthattheassaultinvolvedmultipleperpetrators,alcoholordruguse,useofaweapon,and
useofphysicalforce.Wealsoconsideredtheeffectofthetimebetweentheassaultandthemedical
forensicexam,whichwascodedintothreecategories(0days[sameday,referencelevel],1dayafter
assault,or2+daysafterassault).
Becauseallpredictorsinthemodelarecategorical,thepairwisedifferencesintheoddsofa
CODIShitbetweenSAKsinthereferencelevelandeachoftheotherlevelsweretheprimaryfocusof
theanalysis.WereporttheexactpvaluesassociatedwithWaldtestsforeachofthesecomparisons,but
focusmoreoninterpretingthe95%confidenceintervals(CIs)forthecorrespondingoddsratios(ORs)
becausethesemeasuresofeffectsizearemoreinformativethanthesignificancetestsbasedonthe
194
conventional=.05criterionfortheTypeIerrorrate(seeAppendixB:ProjectMethodology).81TheOR
describesboththesize(i.e.,magnitude)anddirectionofaneffect.Sizeisencodedinthedistanceofan
estimatedORfromtheneutralvalueof1(whichrepresentsequaloddsfortwogroupsbeingcompared),
withvaluesfartherfrom1indicatingstrongereffects.Thedirectionofaneffectisinferredfromwhether
theORislargerorsmallerthan1:largerORsmeanthatthegroupdescribedbyaneffecthashigherodds
ofachievingtheoutcomeofinterestthanthereferencegroup,whilesmallerORsindicatethatithas
loweroddsinstead.
ThepointestimateoftheORisthesinglemostlikelyvaluefortheeffectsize,buttheCIs
describetherangeofeffectsizesthataremostplausiblegiventheobservedvariabilityinthesample
data.Thetrue,unknowneffectsizeismostlikelytoliewithinthatrange.Inthisstudy,wideCIsindicate
moreuncertaintyaboutthesizeandpossiblythedirectionofthedifferenceintheoddsofobtaininga
CODIShit;narrowCIstellusthatwehaveapreciseestimateandallowustobemorecertainabout
effectsizeanddirection.CarefullyconsideringtheimplicationsofthevaluesspannedbyaCIcantellus
considerablymorethanwhetheranobservedeffectcouldbeduetochancesamplingvariation.See
AppendixB:ProjectMethodologyforcompletestatisticaltablesandmodelfitinformationforthe
predictionmodels.
Victim-Offender Relationship Effect on Forensic Testing Outcomes: Stranger & NonStranger Sexual Assaults
Doforensictestingoutcomes(i.e.,theCODISentryrates,CODIShitrates,andserialassault
rates)differbetweenSAKsfromnonstrangerassaultsandstrangerassaults?Toanswerthisquestion
wecombinedthedatafromTestingGroups1(StrangerRape)and2(NonStrangerRape),whichare
bothsamplesfromthesubpopulationofpreviouslyuntested,nonadjudicated,SOLunexpiredDetroit
81
TypeIerrorsoccurwhenwefalselyconcludethereisaneffectwheninrealitythereisnone;TypeIIerrorsarewhenwe
falselyconcludethatthereisnoeffectwhentherereallyisone(wefailtodetectit).
195
SAKsforwhichvictimoffenderrelationshipcouldbedetermined.PreviouslyunpublisheddatafromThe
400Projectindicatedthat62.5%oftheSAKsinthatsubpopulationwereassociatedwithnonstranger
assaultsand37.5%werestrangerassaults.BecauseTestingGroups1and2containedapproximately
equalnumbersofSAKs(N=445and449,respectively),combiningthemyieldsadisproportionate
stratifiedsamplefromthesubpopulationofinterest.Therefore,weanalyzedaweighteddatasetto
properlyaccountforthisstratification(seeAppendixB:ProjectMethodology).
ConditionalRates.Figure4.18(nextpage)showsthekeyresultsforthecomparisonoftesting
outcomesbetweennonstrangerandstrangerassaults.Inthisanalysis,allSAKssubmittedfortesting
startatStage1(DNATesting),sotheCODISentryrateinFigure4.18isanunconditionalestimate.
However,thehitandserialassaultratesareconditionalestimatesthatdependonanSAKhavingalready
reachedStages2(CODISEntry)and3(CODISHit)respectively.
TheconditionalCODISentryratesfornonstrangerandstrangerSAKswere40.1%(95%CI
[35.7,44.7])and53.7%(95%CI=[49.0,58.4]),respectively(Figure4.18,firstpanel)(seealsoTable4.3).
ThisisastatisticallysignificantdifferenceintheoddsofgeneratingaCODISentry(p<.001),butthe
oddsratio(OR=1.73,95%CI=[1.33,2.26])isinfactonlyasmalltomediumsizepositiveeffect
(Rosenthal,1996).82TheoddsofaCODISentrywereonlyabout1.73timeshigherforstrangerassaults
thanfornonstrangerassaults.
82
AnOR=1.00indicatesnodifferenceatall;eithersmallerorlargervaluesindicateadifferenceinoddsbetweentwogroups,
withvaluescloserto1.00consideredweaker.Thus,Rosenthal(1996)describesORsof0.67or1.50assmalleffects,0.40or
2.50asmediumeffects,0.25or4.00aslargeeffects,and0.10or10.00asverylargeeffects.
196
FIGURE4.18TheEffectofVictimOffenderRelationshiponCODISEntry,CODISHits,andSerial
SexualAssaultRates(TestingGroup1[Stranger]andTestingGroup2[NonStranger])
Proportion (Probability)
1.00
CODIS Entry Rate
CODIS Hit Rate
Serial Assault Rate
OR = 1.734, p < 0.001

95% CI = [1.330, 2.261]
OR = 1.405, p = 0.094
95% CI = [0.939, 2.102]
OR = 2.294, p = 0.008
95% CI = [1.238, 4.248]
0.80
0.60
0.40
0.20
0.00
Non-stranger
Stranger
Non-stranger
Stranger
Non-stranger
Stranger
Victim-Offender Relationship
Figure4.18:VictimOffenderRelationshipEffectonCODISEntry,CODISHit,andSerialAssaultRatesAmongSOLUnexpired
DetroitSAKs.TheCODISentryratesareunconditionalestimates(theproportionofSAKstestedthatyieldedaDNAprofile
suitableforuploadintoCODIS).TheCODIShitratesandserialassaultratesareconditionalestimates(respectively,the
proportionsofCODISentriesthatyieldmatchestootherCODISrecords,andtheproportionofCODIShitsthatareassociated
withaserialsexualoffender).Theseresultsgeneralizetothesubpopulationofuntested,nonadjudicated,SOLunexpired
DetroitSAKsforwhichvictimoffenderrelationshipdataareavailable.TheyarebasedonanalysesofN=894SAKs(449non
strangerassaultsand445strangerassaults).TheseestimateswereobtainedfromacontinuationratiomodelofSAK
progressionacrossStages14,weightedtoreflectthedisproportionatestratifiedsamplingdesign.Thedotsmarkthepoint
estimatesfortherates,thewhiskersdepictthecorresponding95%CIs.Theoddsratios(OR)andassociated95%CIsquantify
thesimpleeffectofvictimoffenderrelationshipontheratenamedineachpanel.
TABLE4.3RatesEstimatedfromContinuationRatioModelExaminingEffectof
VictimOffenderRelationshiponTestingOutcomes
Outcome
CODISentryrate
StrangerRate[95%CI]
NonStrangerRate[95%CI]
0.537[0.490,0.584]
0.401[0.357,0.447]
0.653[0.589,0.711]
0.572[0.499,0.643]
0.327[0.257,0.405]
0.175[0.113,0.260]
Stage12
CODIShitrate
Stage23|Entry
Serialassaultrate
Stage34|Hit
197
Wecandelvedeeperintotheseresultsbyconsideringtherelativerisk(RR)andthenumber
neededtosubmit(NNS)statistics.Relativerisk(RR=ps/pns)reflectshowmuchmoreoftenanevent
happens(e.g.,aCODISentry)foronegroup(e.g.,strangerassaults)relativetoanothergroup(e.g.,non
strangerassaults).RRisaratio,constructedbydividing,forexample,theCODISentryrateforTesting
Group1bytheCODISentryrateforTestingGroup2,whichtellsushowmuchmorelikelythatoutcome
(CODISentry)isforonegroupversustheother.Intheseanalyses,RR=1.34,whichtellsusthatstranger
assaultsareonlyabout1.34timesmorelikelytoyieldCODISentriesthannonstrangerassaults,whichis
notasubstantialdifference.
Numberneededtosubmit(NNS=1/[pspns])isanotherstatisticthathelpsusunderstandthe
implicationsofdifferentialratesintwogroups.Iftwogroups(strangerandnonstranger)have
differentialratesofproducinganoutcome(aCODISentry),thenifwetestthesamenumberofstranger
SAKsandnonstrangerSAKs,thestrangerSAKswillproducemoreCODISentries(whichiswhattheOR
andRRstatisticspresentedabovetoldus).TheNNShelpsusseethesedifferentialratesinadifferent
way,bystartingwiththeCODISentriesandwalkingbacktoconsiderhowmanySAKswouldhaveto
betestedtoactuallyyieldmoreCODISentriesfromthegroupwiththehigherentryratethanfromthe
groupwiththelowerentryrate.TheNNSstatisticasks:toobtainaoneunitdifferenceintheoutcome
variablebetweenthetwogroupstoobtainexactlyonemoreCODISentryhowmanySAKswould
havetobetestedtoseethatoneunitdifferenceintheoutcome?Ingeneral,thelargertheNNSthe
moreitmovesawayfrom1.00themorecasesthatareneededtoobtainaoneunitdifference
betweenthegroups;ifittakesasubstantialnumberofSAKstoyieldjustaoneunitdifference,thenthat
suggeststhetwogroupsarefairlysimilaranditmaynotmakepracticalsensetotrytotreatthem
differentially.IntheseanalysesregardingCODISentryrates,theNNS=7.35,whichtellsusthat
submitting7.35strangerassaultSAKsforforensictestingwould,onaverage,yieldonemoreCODIS
entrythanwecouldexpecttogetfromsubmittingthesamenumberofnonstrangerassaultSAKsfor
198
forensictesting(i.e.,3.95versus2.95CODISentries).Inpractice,thetime,effort,andlaborofscreening,
submitting,andtesting7.35SAKstogetaoneunitbumpintheCODISentryratemaynotbean
efficientuseofresources.
TurningtotheresultsregardingconditionalCODIShitrates,amongSAKswithCODISentriesthe
CODIShitratewas57.2%(95%CI=[49.9,64.3])fornonstrangerSAKsand65.3%(95%CI=[58.9,71.1])
forstrangerSAKs(Figure4.18,secondpanelandTable4.3priorpages).Thissmallpositiveeffect(OR=
1.41,95%CI=[0.94,2.10])isnotastatisticallysignificantdifferenceintheoddsofahit(p=.094).
Indeed,suchadifferencecouldbeobservedentirelybychanceduetorandomsamplingvariability.The
confidenceintervalsuggeststhedirectionoftheeffectismorelikelytobepositivethaneithernegative
ornonexistentbecausethelowerboundwouldrepresentonlyatinynegativeeffectofstrangerassault
onthehitrateandliesveryclosetotheneutralvalueof1.00thatmarksnoeffect.Meanwhile,the
upperboundoftheintervalwouldrepresentatmostsmalltomediumpositiveeffect.Itishighly
unlikelythatvictimoffenderrelationshipexertsalargeinfluenceonthehitrate.TheRR=1.14indicates
thatCODISentriesfromstrangerassaultsareamere1.14timesmorelikelytoyieldahitthanCODIS
entriesfromnonstrangerassaults.Similarly,theNNS=12.35suggestswewouldneedtouploadCODIS
entriesforatleast12.35strangerassaultSAKstoobtainonemoreCODIShitthanwecouldexpectifwe
uploadedthesamenumberofentriesfromnonstrangerassaultSAKs(i.e.,8.06versus7.06hits).
Finally,thethirdpanelofFigure4.18showsthattheconditionalserialassaultratefornon
strangerSAKswithCODIShitswas17.5%(95%CI=[11.3,26.0];itwas32.7%(95%CI=[25.7,40.5])for
strangerSAKs.Thedifferencebetweenthosetworatesrepresentsamoderate,statisticallysignificant,
positiveeffectofvictimoffenderrelationshipontheoddsofdetectingaserialassault (OR=2.29,95%CI
[1.24,4.25],p=.008).Whilethisstronglyestablishesthattheeffectispositive,thesizeoftheeffectis
unclear.ThewideCIshowsthatitcouldbeanywherefromaverysmalleffecttoalargeeffect,soits
practicalimportanceisnotyetwellestablished. TheRRstatisticshowsthatCODIShitsfromstranger
199
assaultsare1.87timesmorelikelytobeserialassaultsthanCODIShitsfromnonstrangerassaults.For
thisoutcome,theNNS=6.58meansthatwewouldneedatleast6.58CODIShitsfromstrangerassaults
todetectonemoreserialassaultthanwewouldexpectinasimilarnumberofCODIShitsfromnon
strangerassaults(i.e,2.15versus1.15serialassaults).
UnconditionalRates.ToclarifytheimplicationsoftheconditionalratesdepictedinFigure4.18,
considertheunconditionalratesshowninFigure4.19aswell(nextpage).TheDNAtestingrateis100%
forbothstrangerandnonstrangerassaultsbecauseallSAKsaresubmittedfortestingatStage1.CODIS
entryratesremainthesameacrossbothfiguresbecausetheyrepresentthefirsttransitionbetween
stagesandsomeSAKsdonotgenerateDNAprofilessuitableforentryintoCODIS.Incontrast,the
unconditionalhitandserialassaultratesaredifferentacrossthetwographsbecausetheratesinFigure
4.19allusethetotalnumberofSAKssubmittedfortestingatStage1asthedenominatorratherthan
thenumberofkitsreachingthepreviousstage(asinFigure4.18).
TheunconditionalCODIShitratefornonstrangerSAKswas22.9%,whilethecorrespondingrate
forstrangerSAKswas35.1%(Figure4.19).Intermsofrelativerisk,thatmeanstestinganSAKfroma
strangerassaultis1.53timesmorelikelytoyieldaCODIShitthantestinganSAKfromanonstranger
assault.ThecorrespondingNNSstatisticrevealsthatwewouldneedtosubmit8.20strangerassault
SAKsforforensictestingtoobtainjustonemoreCODIShit(2.88hits)thanwewouldfindbysubmitting
asimilarnumberofnonstrangerSAKs(1.88hits).
Finally,theunconditionalserialassaultratefornonstrangerSAKswas4.0%,ascomparedto
11.5%forstrangerSAKs(Figure4.19).WhiletestinganSAKfromastrangerassaultSAKisalmostthree
times(RR=2.88)morelikelytodetectaserialsexualassaultthantestingonefromanonstranger
assault,theNNSstatisticstillshowsthatwewouldneedtosubmitatleast13.33strangerassaultSAKs
forforensictestinginordertodetectjustonemoreserialassault(1.53serialassaults)thanwewould
expecttofindbysubmittingasimilarnumberofnonstrangerSAKs(0.53serialassaults).
200
FIGURE4.19UnconditionalCODISEntry,CODISHit,andSerialSexualAssaultRates,byVictim
OffenderRelationship(TestingGroup1[Stranger]andTestingGroup2[NonStranger])
Stage Reached
S = 1.000
1. DNA test
N = 1.000
S = 0.537
2. CODIS entry
N = 0.401
S = 0.351
3. CODIS hit
4. Serial assault
N = 0.229
S = 0.115
N = 0.040
0.00
0.20
0.40
0.60
0.80
1.00
Unconditional Proportion (Probability)
Figure4.19:UnconditionalRatesofCODISEntry,CODISHit,andSerialAssaultRatesAmongSOLunexpiredDetroitSAKsby
VictimOffenderRelationship.AllratesshownhereareunconditionalestimatesoftheproportionofSAKstestedthatreached
eachstagelisted.Theseresultsgeneralizetothesubpopulationofuntested,nonadjudicated,SOLunexpiredDetroitSAKsfor
whichvictimoffenderrelationshipdataareavailable.TheyarebasedonanalysesofN=894SAKs(449nonstrangerassaults
and445strangerassaults).TheseestimateswerecalculatedfromtheresultsofacontinuationratiomodelofSAKprogression
acrossStages14,weightedtoreflectthedisproportionatestratifiedsamplingdesign.N=nonstranger;S=stranger.
PredictingUnconditionalCODISHitRatesTestingGroup1(StrangerRapes).Theanalyses
presentedinthepriorsectionsuggestthatvictimoffenderrelationshipdoesnothaveastatistically
significanteffectonconditionalCODIShitrates.However,stakeholdersatthelocal,state,andnational
levelwereinterestedinexploringwhetherotherfeaturesofthevictim,assailant,andassaultmight
predictCODIShitrates.Giventhatsucheffects(e.g.,weaponuse)couldvarywithindifferentvictim
offenderrelationships,weconductedtheseanalysesseparatelyforTestingGroup1(StrangerRapes)
andtestingGroup2(NonStrangerRapes).83
83
Forsimplicity,wefocusedonunconditionalhitrates,onlyfortheCODIShitoutcome(i.e.,notCODISentryorserialsexual
assaults).Becausewelackedsufficientpopulationlevelinformationtoadjustforthedisproportionatestratificationbyyear,
thismodeltreatedthedataasasimplerandomsamplefromthesubpopulationofinterest.
201
Ofthe445SAKsinTestingGroup1(StrangerRapes),atotalof54(12.3%)oftheseSAKswere
excludedfromtheseanalysesduetomissingdataoneithervictimage(N=3)orexamtiming(N=52).
Table4.4belowsummarizesthelogisticregressionresultsbasedontheremaining391SAKs(see
AppendixB:ProjectMethodologyformoredetailsonthismodel).
TABLE4.4LogisticRegressionPredictingCODSHitsWithin TestingGroup1(StrangerRapes)
Predictor
OR
OR95%CI
Intercept(forreferencegroup)
0.358
[0.192,0.655]
0.001
Victimage<16yearsa
0.519
[0.224,1.093]
0.100
Assailantageb
<22years
0.888
[0.390,1.921]
0.769
Unknown
1.337
[0.813,2.198]
0.251
Multipleperpetratorsc
0.968
[0.588,1.578]
0.896
Useofalcohol/drugsc
0.985
[0.609,1.579]
0.950
Useofweaponsc
1.674
[1.072,2.624]
0.024
Useofphysicalforcec
1.190
[0.715,2.011]
0.508
Examtimingd
1dayafterassault
0.930
[0.549,1.557]
0.785
2+daysafterassault
0.576
[0.257,1.198]
0.157
Note:SampleincludedN=391SAKs,afterremovingSAKswithmissingdataforvictimageorexamtiming(assumedtobe
missingcompletelyatrandom).OR=oddsratio(ameasureofeffectsize);CI=confidenceinterval;p=statisticalsignificancep
a
b
c
d
valueobtainedfromaWaldtestofthecoefficient.Referencelevelsare: 16+years; 22+years; no;and 0daysafterassault
(sameday).
202
ThemodelinterceptallowedustocalculatetheunconditionalCODIShitrateforareference
groupconsistingofSAKsfromoldervictims(age16+years)whowereeachassaultedbyasingle,adult
assailant(age22+years)withoutanyinvolvementofalcohol/drugs,weapons,orphysicalforce,andwho
receivedaforensicmedicalexamthesamedaytheywereassaulted.Thatreferencegrouphitratewas
26.4%(95%CI=[16.1,39.6]).TheinterceptpredictorinTable4.4merelytestswhethertheoddsofa
CODIShitforthatreferencegroupdifferfromahypotheticalvalueof1.00(i.e.,a50%hitrate),whichit
does(OR=0.36,p=.001).
AlltheotherpredictorsinTable4.4testdeviationsfromthisreferencegrouprateforSAKsthat
differfromthereferencegroupwithrespecttoaspecificcharacteristic(e.g.,thevictimwasyounger
thanthoseinthereferencegroup,e.g.,age<16years),afteradjustingfortheeffectsoftheother
predictors.Wecommentbelowonthepredictorsthatwerestatisticallysignificantorhaveconfidence
intervals(CIs)thatstronglyindicatethelikelydirectionoftheeffectbecausethevalueof1iscloseto
eithertheupperorlowerbound.TheCIsfortheotherpredictorsshowthattheyareunlikelytoexert
anythingmorethansmalleffects,withoutclarifyingthelikelydirectionofthoseeffects.
Ofallthepredictorsinthemodel,theonlysignificantpredictorwasweaponuse(OR=1.67,95%
CI=[1.07,2.62],p=0.024).TheoddsofobtainingaCODIShitwereapproximately1.67timeshigherfor
SAKsinwhichweaponswereusedintheassault(asmalleffect),buttheeffectcouldbeaslittleas1.07
timeshigher(functionallyequalduetoatinyeffectsize)orasmuchas2.62timeshigher(amedium
effect).OurbestestimatefortheeffectofvictimagesuggeststhattheoddsofaCODIShitdropby
almosthalf(OR=0.519)whenthevictimisyounger(<16years).Thissmalleffectwasnotsignificant
(p=.100).However,the95%CIshowsthattheoddsofahitareatmostessentiallyequalregardlessof
victimage(OR=1.09,atinypositiveeffect),butitisaslikelythattheyaresubstantiallylowerforyoung
victims(OR=0.22,alargenegativeeffect).Thatlowerboundimpliesthattheoddsofahitcouldbe
almost4.5timeshigherforoldervictimsthanforyoungervictims.Whenthemedicalforensicexams
203
occurredtwoormoredaysaftertheassault(ratherthanonthesameday),theoddsofaCODIShit
droppedbyalmosthalf(OR=0.58),butthissmalleffectwasalsonotsignificant(p=0.157).Aswiththe
victimageeffect,theCIsuggeststhatoddsofahitareatmostessentiallyunchangedwithexams
occurringthislate(OR=1.20,averysmallpositiveeffect),butitisequallylikelythattheyare
substantiallylowerwhenexamsareconducted2ormoredaysaftertheassault(OR=0.26,alarge
negativeeffect).Thus,theoddsofaCODIShitcouldbealmost3.9timeshigherforSAKsinwhichthe
examwasconductedthedayoftheassaultthaniftheexamwasconductedtwoormoredayslater.
PredictingUnconditionalCODISHitRatesTestingGroup2(NonStrangerRapes).Ofthe449
SAKsinTestingGroup2(NonStrangerRapes),weexcluded86(19.1%)ofthoseSAKsfromthelogistic
regressionmodelduetomissingdataoneithervictimage(N=10)orexamtiming(N=84).Table4.5
(followingpage)summarizesthelogisticregressionmodelbasedontheremaining343SAKs.
27.1%(95%CI=[15.8,41.7]).Again,theinterceptpredictorinTable4.5merelytestswhethertheodds
ofaCODIShitforthatreferencegroupdifferfromahypotheticalvalueof1.00(i.e.,a50%hitrate),
whichitdoes(OR=0.37,p=.004).
TheotherpredictorsinTable4.5testwhethertheoddsofahitforSAKsthatdifferfromthe
referencegroupwithrespecttoaspecificcharacteristic(e.g.,thevictimwasyoungerthanthoseinthe
referencegroup,e.g.,age<16years)arehigherorlowerthanthoseinthereferencegroup,after
adjustingfortheeffectsoftheotherpredictors.Onceagain,wecommentonlyonthepredictorsthat
werestatisticallysignificantorhaveCIsthatstronglyindicatethelikelydirectionoftheeffect.
204
WithinthissampleofnonstrangerSAKs,theoddsofaCODIShitdropbymorethanhalf(OR=
0.40,p=.015)foryoungervictims(<16years)(amediumsizeeffect).The95%CIsuggeststhatthereis
atleastaverysmallnegativeeffect(OR=0.83),butquitepossiblyalargeone(OR=0.19).TheCIimplies
thattheoddsofahitcouldbeanywherefrom1.2to5.3timeshigherforoldervictimsthanforyoung
victims.Thus,thedirectionoftheeffectisclear,butitssizeremainsimpreciselyestimated.
TABLE4.5LogisticRegressionPredictingCODSHitsWithinTestingGroup2(NonStrangerRapes)
Predictor
OR
OR95%CI
0.372
[0.188,0.716]
0.004
Victimage<16yearsa
0.403
[0.189,0.825]
0.015
Assailantageb
<22years
1.014
[0.522,1.939]
0.966
Unknown
2.424
[0.820,6.829]
0.097
1.197
[0.601,2.306]
0.598
Useofalcohol/drugsc
0.642
[0.354,1.138]
0.136
Useofweaponsc
2.155
[1.134,4.058]
0.018
Useofphysicalforcec
0.712
[0.390,1.312]
0.270
Examtimingd
1dayafterassault
1.689
[0.957,2.977]
0.069
2+daysafterassault
0.554
[0.212,1.278]
0.191
a
b
c
d
(sameday).
205
Weaponusewasalsostatisticallysignificant(OR=2.16,95%CI=[1.13,4.06],p=0.018).The
oddsofobtainingaCODIShitwereabout2.16timeshigherforSAKswhenweaponswereusedinthe
assault(asmalltomediumeffect),butcouldbeaslittleas1.13timeshigher(essentiallyequalduetoa
tinyeffectsize)orasmuchas4.06timeshigher(alargeeffect).Again,thedirectionoftheeffectisclear,
butitssizeandpracticalimportancearenotasclearlyestablishedbythesedata.
Whenassailantagewasunknown,theoddsofahitincreasedmoderately(OR=2.42)overwhen
theassailantwasanadult(>21years),butthiseffectwasnotsignificant(p=0.097).However,the95%
CIsuggeststhatiftheassailantageisunknown,theoddsofahitcouldrangefromalittlelower(OR=
0.82,averysmallnegativeeffect)tosubstantiallyhigherthanthosefortheadultassailants(OR=6.83,a
largepositiveeffect).Onbalance,thissuggeststhatwhileitisstillpossiblethatunknownassailantage
actuallydecreasesthehitrateabitordoesnotaffectitatall,itismorelikelythatthehitrateincreases
whenassailantageisunknown.
Whentheassaultinvolvedalcoholordruguse,therewasasmalldecreaseintheoddsofahit
(OR=0.64)comparedtowhenalcohol/drugswerenotinvolved.Thiseffectwasnotsignificant(p=
0.136).However,the95%CIsuggeststhatifalcohol/drugswereinvolved,thentheoddsofahitareat
mostslightlyhigherthanwhenalcohol/drugswerenotinvolved(OR=1.14,atinypositiveeffect),but
couldalsobemoderatelylower(OR=0.35,amediumsizenegativeeffect).Thelowerboundimpliesthat
theoddsofahitcouldbeabout2.82timeshigherwhentherewasnoevidencethatalcohol/drugswere
involvedintheassault.
Whenmedicalforensicexamsoccurredonedayaftertheassault(insteadofonthesameday),
theoddsofahitincreasedmoderately(OR=1.69),buttheeffectwasnotsignificant(p=0.069).
However,the95%CIsuggeststhatiftheexamoccurredonedayaftertheassault,thattheoddsofahit
couldrangefromessentiallyequaltowhentheexamoccurredonthesameday(OR=0.96,atiny
negativeeffect)tomoderatelyhigher(OR=2.98,amediumsizepositiveeffect).Ourbestestimate
206
showsthereisasmall,nonsignificantdecreaseintheoddsofahit(OR=0.55,p=0.191)whenmedical
forensicexamsoccurredtwoormoredaysaftertheassault(ratherthanonthesameday).However,the
95%CIsuggeststhatiftheexamoccurredtwoormoredaysaftertheassault,thattheoddsofahitare
atmostslightlyhigherthanwhentheexamoccurredonthesameday(OR=1.28,averysmallpositive
effect),butcouldalsobesubstantiallylower(OR=0.21,alargenegativeeffect).Thatlowerbound
impliesthattheoddsofaCODIShitcouldbeupto4.72timeshigherwhentheexamoccurredonthe
samedayinsteadoftwoormoredayslater.Weconcludethatmedicalforensicexamsconductedtwoor
moredaysaftertheassaultprobablyhadanegativeeffect(decreasingthehitrate),butthesizeand
practicalimportanceofthedecreaseisnotwellestablishedduetotheimpreciseCI.
Statute of Limitations Effect on Forensic Testing Outcomes: Presumed SOL-Expired and

Unexpired SAKs
Doforensictestingoutcomes(i.e.,theCODISentryrates,CODIShitrates,andserialassault
rates)differbetweenSOLexpiredandSOLunexpiredSAKs?Answeringthisquestionrequiredcombining
datafromTestingGroups3(PresumedSOLExpired)and4(DNAMethod/SOLUnexpired),whichare
bothsamplesfromthesubpopulationofpreviouslyuntested,SOLunexpiredDetroitSAKs(seeFigure
4.17).PreviouslyunpublisheddatafromThe400Projectindicatedthat63.6%oftheSAKsinthat
subpopulationresultedfromSOLexpiredassaultsand36.4%resultedfromSOLunexpiredassaults.
TestingGroups3and4containedapproximatelyequalnumbersofSAKs(N=351and350,respectively),
soweonceagainweightedthedata,treatingitasadisproportionatestratifiedsamplefromthe
subpopulationofinterest(seeAppendixB:ProjectMethodology).
AlloftheTestingGroup3SAKsandhalfoftheTestingGroup4SAKsweretestedwithtraditional
DNAtestingmethods;theremainingGroup4SAKsweretestedwiththeDNaseselectivedegradation
method.ResultsreportedbelowintheEvaluatingtheDetroitSAKTestingPlanInferentialFindingson
207
theEffectofDNATestingMethodonForensicTestingOutcomessectionconvincedusthatthis
differenceinDNAtestingmethodsposesnothreattothevalidityofconclusionsdrawnfromanalyzing
thiscombineddataset.
ConditionalRates.Figure4.20(below)showsthestatuteoflimitationseffectontesting
outcomes.AllSAKssubmittedfortestingstartatStage1(DNATesting),sotheCODISentryrateinFigure
4.20isanunconditionalestimate,buttheCODIShitandserialassaultratesareconditionalestimates
thatdependonanSAKhavingalreadyreachedStages2(CODISEntry)and3(CODISHit)respectively.
FIGURE4.20TheEffectofStatuteofLimitationsonCODISEntry,CODISHits,andSerialSexualAssaultRates
(TestingGroup3[PresumedSOLExpired]andTestingGroup4[DNAMethod/SOLUnexpired])
1.00
CODIS Entry Rate
CODIS Hit Rate
Serial Assault Rate
OR = 1.265, p = 0.121
95% CI = [0.940, 1.702]
OR = 1.124, p = 0.579
95% CI = [0.741, 1.705]
OR = 0.792, p = 0.458
95% CI = [0.425, 1.476]
0.80
0.60
0.40
0.20
0.00
Expired
Unexpired
Expired
Unexpired
Expired
Unexpired
Statute of Limitations Status

Figure4.20:StatuteofLimitationsEffectonCODISEntry,CODISHit,andSerialAssaultRatesAmongDetroitSAKs.TheCODIS
entryratesareunconditionalestimates(theproportionofSAKstestedthatyieldedaDNAprofilesuitableforuploadinto
CODIS).TheCODIShitratesandserialassaultratesareconditionalestimates(respectively,theproportionsofCODISentries
thatyieldmatchestootherCODISrecords,andtheproportionofCODIShitsthatareassociatedwithaserialsexualoffender).
TheseresultsgeneralizetothepopulationofuntestedDetroitSAKs(regardlessofadjudicationstatusorvictimoffender
relationship).TheyarebasedonanalysesofN=701SAKs(351SOLexpiredassaultsand350SOLunexpiredassaults).These
estimateswereobtainedfromacontinuationratiomodelofSAKprogressionacrossStages14,weightedtoreflectthe
disproportionatestratifiedsamplingdesign.Thedotsmarkthepointestimatesfortherates,thewhiskersdepictthe
corresponding95%CIs.Theoddsratios(OR)andassociated95%CIsquantifythesimpleeffectofSOLontheratenamedin
eachpanel.
208
ThefirstpanelinFigure4.20showsthat49.3%(95%CI=[44.1,54.5])oftheSOLexpiredand
55.1%(95%CI=[49.8,60.3])oftheSOLunexpiredSAKsyieldedCODISentries(seealsoTable4.6,
below).Thisisaverysmallandstatisticallynonsignificanteffect(OR=1.27,95%CI=[0.94,1.70],p=
.121).TheoddsofgeneratingaCODISentryaremostlikelyonlyabout1.27timeshigherfortheSOL
unexpiredSAKs.TheasymmetricalCIhereagainsuggestsbutdoesnotdefinitelyprovethatthe
directionoftheSOLeffectispositivebecausethelowerboundfortheORliesjustbelow1.00.Italso
circumscribesthemaximumplausiblesizeoftheeffect:SOLcouldexertatinynegativeeffect,noeffect
atall,oratmostitcouldexertasmall,positiveeffectthatslightlyraisestheCODISentryrateforSOL
unexpiredSAKs.Therelativeriskcalculation(RR=pu/pe)showsthatunexpiredSAKsareonly1.12times
morelikelytoyieldanentrythanpresumedexpiredSAKs.TheNNScalculation(NNS=1/[pupe])tells
uswewouldneedtosubmit17.24SOLunexpiredSAKstogetonemoreCODISentrythanwecould
expectfromsubmittingthesamenumberofSOLexpiredSAKs(i.e.,9.50versus8.50entries).
TABLE4.6RatesEstimatedfromContinuationRatioModelExaminingEffectofStatuteof
LimitationsonTestingOutcomes(TestingGroup3[PresumedSOLExpired]andTesting
Group4[DNAMethod/UnexpiredSOL])
Outcome
CODISentryrate
PresumedSOLExpired
SOLUnexpired
Rate[95%CI]
Rate[95%CI]
0.493[0.441,0.545]
0.551[0.498,0.603]
0.520[0.446,0.594]
0.549[0.478,0.619]
0.322[0.234,0.425]
0.274[0.196,0.367]
Stage12
CODIShitrate
Stage23|Entry
Serialassaultrate
Stage34|Hit
209
ThesecondpanelinFigure4.20showsthattheSOLeffectisevensmallerontheconditional
CODIShitrate,whichwas52.0%(95%CI=[44.6,59.4])forSOLexpiredSAKsand54.9%(95%CI=[47.8,
61.9])forSOLunexpiredSAKs.Thishasatriviallysmall(andnonsignificant)effectontheoddsofa
CODIShit(OR=1.12,95%CI=[0.74,1.70],p=.579).TheCIfortheORgivesusnoclearsignalaboutthe
directionoftheeffect,butitplainlyshowsthatSOLisunlikelytoexertmorethanasmalleffecteither
way.CODISentriesforSOLunexpiredSAKsareonlyRR=1.06timesmorelikelytoyieldahitthanCODIS
entriesforSOLexpiredSAKs.ItwouldtakeNNS=34.48CODISentriesfromSOLunexpiredSAKstoyield
onemorehitthanexpectedfromthesamenumberofSOLexpiredSAKs(18.93vs17.93hits).
Finally,thethirdpanelofFigure4.20showsthattheconditionalserialassaultrateis32.2%(95%
CI=[23.4,42.5])forSOLexpiredSAKswithCODIShitsand27.4%(95%CI=[19.6,36.7])forSOL
unexpiredSAKswithCODIShits.Thisisaverysmall,statisticallynonsignificantnegativeeffectofSOL
(OR=0.79,95%CI=[0.43,1.48],p=.458)ontheoddsofdetectingaserialassault.TheCIisconsistent
withpossibleeffectsizesrangingfromamedium,negativeeffectonthelowendtoasmall,positiveone
atthehighend,withoutstronglysuggestingthelikelydirectionoftheeffect. TheRRstatisticshowsthat
CODIShitsfromSOLunexpiredassaultsare0.85timeslesslikelytobeserialassaultsthanCODIShits
fromSOLexpiredassaults.TheNNS=20.83meansthatexamining20.83CODIShitsfromSOL
unexpiredassaultswouldlikelydetectonelessserialassaultthanwewouldexpectinasimilarnumber
ofCODIShitsfromSOLexpiredassaults(i.e,5.71versus6.71serialassaults).
UnconditionalRates.Figure4.21(nextpage)furtherclarifiestheconditionalratesdepictedin
Figure4.20bytranslatingtheresultsintounconditionalratesateachstageoftheforensictesting
process.Asbefore,theDNAtestingrateis100%regardlessofSOLstatusbecauseallSAKsare
submittedfortestingatStage1andtheCODISentryratesremainthesameacrossbothfiguresbecause
theyrepresentthefirststagetransition.TheCODIShitandserialassaultratesdifferacrossthetwo
210
graphsbecausetheyusedifferentdenominators(Figure4.21usesthetotalnumberofSAKssubmitted
fortestingatStage1,whileFigure4.20usesthenumberofkitsreachingthepreviousstage).
Stage Reached
FIGURE4.21UnconditionalCODISEntry,CODISHit,andSerialSexualAssaultRates,byStatuteofLimitations
(TestingGroup3[PresumedSOLExpired]andTestingGroup4[DNAMethod/SOLUnexpired])
U = 1.000
1. DNA test
E = 1.000
U = 0.551
2. CODIS entry
E = 0.493
U = 0.303
3. CODIS hit
4. Serial assault
E = 0.256
U = 0.083
E = 0.083
0.00
0.20
0.40
0.60
0.80
1.00

Figure4.21:UnconditionalCODISEntry,CODISHit,andSerialAssaultRatesAmongDetroitSAKsbyStatuteofLimitations
Status.AllratesshownhereareunconditionalestimatesoftheproportionofSAKstestedthatreachedeachstagelisted.These
resultsgeneralizetothepopulationofuntestedDetroitSAKs(regardlessofadjudicationstatusorvictimoffenderrelationship).
TheyarebasedonanalysesofN=701SAKs(351SOLexpiredassaultsand350SOLunexpiredassaults).Theseestimateswere
calculatedfromtheresultsofacontinuationratiomodelofSAKprogressionacrossStages14,weightedtoreflectthe
disproportionatestratifiedsamplingdesign.E=SOLexpired;U=SOLunexpired.
TheunconditionalCODIShitrateforSOLexpiredSAKswas25.6%,whilethecorrespondingrate
forSOLunexpiredSAKswas30.3%(Figure4.21).Intermsofrelativerisk,thatmeanstestinganSOL
unexpiredSAKisonly1.18timesmorelikelytoyieldaCODIShitthantestinganSOLexpiredSAK.The
correspondingNNSstatisticrevealsthatonaveragewewouldneedtosubmit21.28SOLunexpiredSAKs
toforensictestingtoobtainjustonemoreCODIShit(6.45hits)thanwewouldfindbysubmittinga
similarnumberofSOLexpiredSAKs(5.45hits).
Finally,theunconditionalserialassaultrateforbothSOLexpiredandSOLunexpiredSAKswas
8.3%(Figure4.21).TheRR=1.00,whichmeansthatwewillbeequallylikelytodetectaserialsexual
211
assaultregardlessofSOLstatusoftheSAKsubmitted.Becausetheratesareidentical,theNNS=
(infinity),indicatingthatitisimpossibletodetectmoreserialsexualassaultsbyselectivelytestingbased
onSOLstatus.
PredictingUnconditionalCODISHitRatesTestingGroup3(PresumedSOLExpired).The
analysespresentedinthepriorsectionsuggestthatSOLstatusdoesnothaveastatisticallysignificant
effectonconditionalCODIShitrates.However,givenstakeholdersinterestinwhetherotherfeatures
ofthevictim,assailant,andassaultmightpredictCODIShitratesamongpresumedSOLexpiredcases,
weconductedalogisticregressionanalysistoexplorethesequestions.TestingGroup3containeda
simplerandomsampleof351untested,presumedSOLexpiredSAKs.Weexcluded107(30.5%)ofthose
SAKsfromthelogisticregressionmodelduetomissingdataoneithervictimage(N=71)orexamtiming
(N=103).Table4.7(nextpage)summarizesthelogisticregressionmodelbasedontheremaining244
SAKs(seeAppendixB:ProjectMethodologyformoredetails).
15.6%(95%CI=[5.8,37.7]).TheinterceptpredictorinTable4.7merelytestswhethertheoddsofa
CODIShitforthatreferencegroupdifferfromahypotheticalvalueof1.00(i.e.,a50%hitrate),whichit
does(OR=0.16,p<0.001).
TheotherpredictorsinTable4.7testwhethertheoddsofahitforSAKsthatdifferfromthe
referencegroupwithrespecttoaspecificcharacteristic(e.g.,thevictimwasyoungerthanthoseinthe
referencegroup,e.g.,age<16years)arehigherorlowerthanthoseinthereferencegroup,after
212
TABLE4.7LogisticRegressionPredictingCODSHitsWithinTestingGroup3(PresumedSOLExpired)
Predictor
OR
OR95%CI
0.156
[0.058,0.377]
<0.001
Victimage<16yearsa
0.817
[0.332,1.901]
0.649
Assailantageb
<22years
1.799
[0.776,4.136]
0.167
Unknown
1.168
[0.349,3.408]
0.786
0.612
[0.256,1.355]
0.245
Useofalcohol/drugsc
0.613
[0.231,1.451]
0.290
Useofweaponsc
2.095
[1.131,3.929]
0.020
Useofphysicalforcec
1.655
[0.741,4.050]
0.240
Examtimingd
1dayafterassault
1.011
[0.494,2.007]
0.976
2+daysafterassault
1.432
[0.404,4.538]
0.553
a
b
c
d
(sameday).
adjustingfortheeffectsoftheotherpredictors.Again,wecommentonlyonthepredictorsthatwere
statisticallysignificantorhaveCIsthatstronglyindicatethelikelydirectionoftheeffect.
Ofallthepredictorsinthemodel,theonlystatisticallysignificantpredictorwasweaponusein
theassault(OR=2.09,95%CI=[1.13,3.93],p=0.020):theoddsofobtainingaCODIShitwereabout
2.09timeshigherforSAKswhenweaponswereusedintheassault(asmalltomediumpositiveeffect),
butcouldrangefromaslittleas1.13timeshigher(essentiallyequalduetoatinypositiveeffect)up3.93
timeshigher(alargepositiveeffect).Thisclearlyshowsthattheeffectispositive,butitssizeremains
impreciselyestimated,limitingourabilitytodrawstrongconclusionsaboutitspracticalimportance.
213
Whenassailantswereyounger(age<22years),ourbestestimateindicatesasmallbutnonsignificant
increaseintheoddsofahit(OR=1.80,p=.167).ThecorrespondingCIsuggeststhattheoddsofahit
couldrangefromalittlelower(OR=0.78,asmallnegativeeffect)whentheassailantisyoungerto
substantiallyhigher(OR=4.14,alargepositiveeffect).Thissuggeststhattheeffectisprobablypositive
(increasingthehitrate),butisinconclusivebecausetheCIiswideenoughtoencompassthepossibility
thatthehitrateactuallydecreasesordoesnotchangeatallwhenassailantsareyoung.
Theremainingpredictorsinthemodelwerenotstatisticallysignificant.TheCIsforvictimage,
unknownassailantage,andexamsoccurringonedayaftertheassaultshowthatthesepredictorsare
unlikelytoexertanythingmorethansmalltomoderateeffects,withoutclarifyingthedirectionofthe
effect.Meanwhile,theCIsforinvolvementofmultipleperpetrators,useofalcohol/drugs,useofforce,
andexamsoccurringtwoormoredaysaftertheassaultspanwiderangesofeffectsizes.Althoughthey
showthatmoderatetolargeeffectsizesarepossibleinatleastonedirection,noneoftheseCIsstrongly
establisheitherthedirectionorthesizeoftheeffect.
Evaluating the Detroit SAK Testing PlanInferential Findings on the Effect of DNA
Testing Method on Forensic Testing Outcomes
Overview
Inthisactionresearchproject,wealsowantedtoexaminewhethernewerDNAtesting
techniques,suchasselectivedegradationmethods,couldofferfaster,lessexpensivetestingoptions,
withoutsacrificingaccuracy.Asdescribedpreviously,theselectivedegradationmethodisusedinthe
firststepofDNAtesting(seeFigure4.4)wherebytheforensicscientistusesafasteractingchemical
techniqueforisolatingthespermanddestroyingtheremainingnonspermcellsinthesample.With
selectivedegradation,thesamplethatwillbeanalyzedforDNAiscleanerinthatmethodminimizes
mixturesbydestroyingnonspermDNAthatismixedwiththespermcells;iftherearemultiplemale
214
assailants,themixtureofthosetwoDNAsamplesisstillintact,asthemethoddoesnotdestroysperm
(fromanysource).InthiscomponentoftheDetroitSAKtestingplan,weusedanexperimentaldesignto
compareSAKstestedwithtraditionalmethods(controlgroup)vs.selectivedegradationmethods
(experimentalgroup)(i.e.,theSAKsintestingGroup4(DNAMethod)wererandomlyassignedtothese
twoDNAtestingmethodconditions).
Inthissection,wewillreporttheresultscomparingthesetwodifferentDNAtestingmethods.
First,wewillexaminewhetherthereisadifferenceintheCODISentryratesyieldedbytraditionalversus
selectivedegradation(DNase)testing.Second,wewilltestwhethertheCODISentryratesforthesetwo
methodsarefunctionallyequivalent(i.e.,isthedifferencebetweenthemsmallerthana5%marginof
equivalence).Finally,wewillcomparewhetherthetwomethodsdifferwithrespecttothecostof
consumablesuppliesusedduringtestingandtheamountofpersonneleffortrequiredtotest,interpret,
andreviewthekits.Selectivedegradationhasthepotentialforautomation,whichwasnotusedinthis
experiment;thus,thepotentialtimesavingsofthismethodareunderestimatedinthisstudy.
Effect of DNA Testing Method on DNA Testing Rates and CODIS Entry Rates
WequantifiedandcomparedDNAtestingrateandCODISentryratesforuntested,SOL
unexpiredSAKstestedviatraditionalvs.selectivedegradation(DNase)testingmethods.Mostofthe
resultsinthissectionwereobtainedfromcontinuationratiomodelsofhowTestingGroup4SAKs
progressedthroughStages02oftheforensictestingprocessshowninFigure4.4.Therestwere
obtainedfromequivalencetests,whicharedescribedbelow.
TheDNAtestingrateistheproportionofSAKssubmittedfortestingthatpassedascreening
proceduredesignedtodeterminewhethertherewassufficientbiologicalevidencepresentintheSAKto
215
warrantevenrunningaDNAtestatall.84BecauseSAKswererandomlyassignedtothetwotesting
methodsandtheactualDNAtestoccursatStage1,thereisnoapriorireasontoexpectadifferencein
theDNAtestingrate.Wetestedforitonlytoensurethatanyunexpecteddifferencewouldbeexplicitly
observed,reportedinourresults,andpreventedfromcontaminatingthecrucialcomparison,which
shouldfocusonlyonconditionalCODISentryrates.
Intheseanalyses,theCODISentryrateisaconditionalratherthanunconditionalestimate.It
correspondstotheproportionofSAKsthatwereactuallytested(i.e.,thatpassedthescreeningatStage
0)thatyieldedDNAprofilessuitableforuploadintoCODIS.Focusingontheconditionalestimate
ensuresthatwehaveacleancomparisonbetweenthetwoDNAtestingmethodsunderconditions
wherelaboratorypersonnelbelievethereisenoughbiologicalevidencepresentintheSAKsthatitis
actuallypossibletoextractanassailantsDNAprofile.
ThereisnomechanismbywhichtheDNAtestingmethodcouldaffecteithertheCODIShitrate
ortheserialassaultrate.ThoseoutcomesdependonthematchbetweenthecontentsofnewCODIS
entriesandotherrecordsalreadystoredinCODIS.Anyeffectonthesetworatesmustbemediatedby
eitherwhatgetsenteredorwhatwasalreadypresentinthesystem.StringentcriteriacontrolwhatDNA
profilesmaybeuploadedintoCODIS,sonewCODISentriesresultingfromeitherkindoftestmustmeet
thesamecriteria.Thateliminatesonepotentialpathwaytoinfluencinghitorserialassaultrates.The
onlyotherpathwayiseliminatedbythefactthatneitherkindofDNAtestcouldpossiblyaffectthe
contentofpreviouslyenteredrecords.Therefore,wedidnotextendourmodeltoestimatethoserates.
Wepresenttheresultsintermsoftheconditionalratesthataredirectlyestimatedbythemodels,then
translatetheresultsintooddsratios(ORs),estimatesofrelativerisk(RR),anumberneededtosubmit
84
TheresultsofthisscreeningwereonlyrecordedasaseparatevariableforTestingGroup4,soanalysesforTestingGroups13
couldnotquantifyandcompareDNAtestingrates.SAKsinthosegroupsthatdidnotcontainsufficientbiologicalevidenceto
permitDNAtestingatallalwaysstoppedatStage1,alongwithSAKsthatcontainedsuchevidenceandweretested,butfailed
toyieldDNAprofilessuitableforentryintoCODIS.
216
(NNS)statistic,andunconditionalratestofacilitateinterpretation.Theseresultswillbepresentedfor
twosetsofanalyses:(1)onefromabasemodelthatomitscovariates;and(2)anothermodelthat
accountsforacovariate(thepresenceofspermintheSAK)actingasamoderatorofthetestingmethod
effect.85
TheDNaseselectivedegradationtestingmethodisdesignedtoworkbyselectivelydegrading
DNAnotcontributedbyaspermcell.Thus,ithasthegreatestpotentialtooutperformtraditionaltesting
whenspermispresent.Italsohassomepotentialtoperformworsewhenspermisnotpresentbecause
otherassailantDNAthatispresentmaybedegraded.Accountingforthispotentialmoderatormay
thereforebeimportantinunderstandingwhentraditionalversusselectivedegradationtestingmayyield
differentresults.Toexaminemorestringentlywhetherthetestingmethodinfluencesforensicoutcomes
afterweaccountforthepresenceorabsenceofsperm,weextendedthemodeltoincludeastageby
testingmethodbypresenceofsperminteractioneffect,thenlookedatthesimplemaineffectoftesting
methodoneachrateseparatelywhenspermwasabsentandwhenitwaspresent.
WealsodirectlyexaminedwhetherthetwogroupsyieldconditionalandunconditionalCODIS
entryratesthatarefunctionallyequivalent.Conventionalstatisticaltests,suchasthecontinuationratio
model,adoptanullhypothesisthatthereisnodifferencebetweengroups(i.e.,theiroutcomesare
exactlyequal).Whentheanalysisprovidesstrongevidencerefutingthatassumption,wecanconclude
thereisindeedadifference.However,CarlSagans(1995,p.221)famousquotetheabsenceof
evidenceisnotevidenceofabsenceremindsusthatfailingtofindasignificantdifferencebetween
groupsisnotthesamethingasprovingthattheydonotdiffermeaningfully.Anonsignificantfinding
fromaconventionaltestyieldsonlyanabsenceofevidencewithrespecttothehypothesisthattwo
groupshaveequivalentoutcomes.
85
Inthiscontext,amoderatorisatypeofcovariatethatmodifieshowstronglyafocalpredictoraffectstheoutcomeof
interest.Here,weconsiderthepossibilitythatthetestingmethodeffectonDNAtestingratesandCODISentryratedependson
whetherornotspermispresentintheSAK.
217
Establishingthatgroupshaveequivalentoutcomesrequiresgeneratingcredibleevidenceof
absencewithrespecttogroupdifferences,whichisthepurposeofequivalencetests.Thesestatistical
methodsadoptthenullhypothesisthattheoutcomesforthegroupsarenotequivalent(i.e.,the
differenceislargeenoughtobeimportant)(Barker,Luman,McCauley,&Chu,2002;daSilva,Logan,&
Klein,2009;Rogers,Howard,&Vessey,1993;Stegner,Bostrom,&Greenfield,1996;Wellek,2010).Only
whentheanalysisprovidesstrongevidencerefutingthatassumptioncanweconcludethatgroupsare
equivalent.Explicitlydefiningequivalenceinadvanceiscrucialforthesetests.Wesetthemarginof
equivalenceat=5%becausefeedbackfromforensicsciencestakeholders(twoatthestateleveland
twoatthenationallevel)suggestedthatCODISentryratesforthetwogroupsthatarewithin5%ofeach
other(0.05<<0.05,where=pTpD)wouldwarrantconsideringthetwoDNAtestingmethods
functionallyequivalent.
ConditionalRates(ModeratorOmitted).Figure4.22(nextpage)showstheeffectofDNA
testingmethodonthetestingoutcomeswhenweomitthepotentialmoderatorfromthemodel.The
firstpanelshowsthat,asexpected,thereisverylittledifferenceintheDNAtestingrates(OR=1.15,
95%CI=[0.73,1.81],p=0.560),whichwere68.6%(95%CI=[61.3,75.0])forSAKsinthetraditionaltest
groupand71.4%(95%CI=[64.2,77.7])intheDNasetestgroup(seealsoTable4.8,followingpages).
TheRR=1.04showsthatSAKsintheDNasegroupwereonly1.04timesmorelikelytopassthe
screeningandactuallyreceiveDNAtestingthanSAKsinthetraditionaltestgroup.Wewouldneedto
submitNNS=35.71SAKsfromtheDNasegrouptodetectjustonemoreSAKwithsufficientbiological
evidencetowarrantDNAtesting(25.5SAKs)thanwewouldexpectamongthesamenumberofSAKs
fromthetraditionalgroup(24.5SAKs).Randomassignmenttothetwogroupsthusappearstohave
eliminatedsystematicdifferenceswithrespecttothepresenceofbiologicalevidence.
218
FIGURE4.22TheEffectofTestingMethodonCODISEntry,CODIS
Hits,andSerialSexualAssaults(TestingGroup4[DNAMethod)
1.00
DNA Testing Rate
CODIS Entry Rate
OR = 1.146, p = 0.560
95% CI = [0.725, 1.811]
OR = 0.785, p = 0.441
95% CI = [0.421, 1.465]
0.80
0.60
0.40
0.20
0.00
Traditional
DNase
Traditional
DNase
DNA Testing Method

Figure4.22:TestingMethodEffectonDNATestingandCODISEntryRatesAmongSOLunexpiredDetroitSAKs.TheDNA
testingrateisanunconditionalestimate(theproportionofSAKsthatpasstheStage0serologyscreeningtoreachtheactual
DNAtestatStage1).TheCODISentryrateisaconditionalestimate(theproportionofSAKstestedthatyieldedaDNAprofile
suitableforuploadintoCODIS).Theseresultsgeneralizetothesubpopulationofuntested,SOLunexpiredDetroitSAKs
(regardlessofadjudicationstatusorvictimoffenderrelationship).AfterselectingN=350SAKsfromthatsubpopulation,n=
175SAKswererandomlyallocatedtoeachofthetwotestingmethods.Theseestimateswereobtainedfromacontinuation
ratiomodelofSAKprogressionacrossStages02(unweightedduetothesimplerandomsamplingdesign).Thedotsmarkthe
pointestimatesfortherates,thewhiskersdepictthecorresponding95%CIs.Theoddsratios(OR)andassociated95%CIs
quantifythesimpleeffectoftestingmethodontheratenamedineachpanel.
TheconditionalCODISentryrateforSAKsinthetraditionalgroupwas80.8%(95%CI=[72.8,
86.9]),ascomparedto76.8%(95%CI=[68.5,83.4])intheDNasegroup(seeTable4.8).Theeffectsizeis
verysmallandnonsignificant(OR=0.79,95%CI=[0.42,1.47],p=0.441).TheRR=0.95indicatesthat
SAKsallocatedtoDNasetestingareabout0.95timeslesslikelytoyieldCODISentriesthantraditional
testing.Furthermore,theNNS=25.00meansthattesting25.00SAKscontainingbiologicalevidencevia
DNasetestingwouldlikelyyieldonelessCODISentrythanwewouldexpectfromtraditionaltestingofa
similarnumberSAKscontainingbiologicalevidence(i.e,19.20versus20.20CODISentries).
219
Thecontinuationratiomodeldoesnotofferstrongevidenceforadifferencebetweentesting
methods,sowecomputedanequivalencetesttoevaluatewhethertheconditionalCODISentryrates
areactuallyequivalent.The4.0%differencebetweenthosetworatesispreciselyestimated(90%CI=
[3.4,4.7]),withallplausiblevaluesfallingcomfortablywithinthe5%marginofequivalence.The
conditionalentryratesarethereforefunctionallyequivalentwhenweomitthepresenceofsperm
moderatorfromthemodel.
TABLE4.8RatesEstimatedfromContinuationRatioModelComparingDNATesting
Methods(TestingGroup4[DNAMethod])
Outcome
DNAtestingrate
TraditionalRate
SelectiveDegradation(DNase)Rate
[95%CI]
[95%CI]
0.686[0.613,0.750]
0.714[0.642,0.777]
0.808[0.728,0.869]
0.768[0.685,0.834]
Stage01
CODISentryrate
Stage12|Tested
UnconditionalRates(ModeratorOmitted).Figure4.23(followingpage)showsresultsofthe
experimentintermsofunconditionalrateswhenweomitthepotentialmoderatorfromthemodel.
Here,thescreeningrateis100%regardlessofDNAtestingmethodbecauseallSAKssubmittedfor
testingarescreenedatStage0;theDNAtestingrateremainsthesameasinFigure4.22becauseit
representsthefirststagetransitioninthemodel.Thus,onlytheCODISentryratesdifferacrossFigures
4.22and4.23becausetheyusedifferentdenominators(Figure4.23usesthetotalnumberofSAKs
submittedatStage0,whileFigure4.22usesthenumberofkitsreachingStage1).
TheunconditionalCODISentryratewas55.4%forthetraditionalgroupand54.9%forthe
DNasegroup,yieldingaRR=0.99.Thisistrivialdifferenceinthetworates,suggestingthatweare
equallylikelytoobtainCODISentriesregardlessofDNAtestingmethod.Itissosmallthatsubmitting
220
200.00SAKstoforensictestingislikelytoyieldjustonelessCODISentryunderDNasetestingthanwe
wouldgetifweusedtraditionaltestinginstead.
Stage Reached
FIGURE4.23UnconditionalDNATestingandCODISEntryRates,byDNATestingMethod
D = 1.000
0. Serology screening
T = 1.000
D = 0.714
1. DNA test
T = 0.686
D = 0.549
2. CODIS entry
T = 0.554
0.00
0.20
0.40
0.60
0.80
1.00

Figure4.23:UnconditionalDNATestingandCODISEntryRatesAmongSOLunexpiredDetroitSAKsbyTestingMethod.All
ratesshownhereareunconditionalestimatesoftheproportionofSAKstestedthatreachedeachstagelisted.Theseresults
generalizetothesubpopulationofuntested,nonadjudicated,SOLunexpiredDetroitSAKs(regardlessofadjudicationstatusor
victimoffenderrelationship).AfterselectingN=350SAKsfromthatsubpopulation,n=175SAKswererandomlyallocatedto
eachofthetwotestingmethods..TheseestimateswerecalculatedfromtheresultsofacontinuationratiomodelofSAK
progressionacrossStages02(unweightedduetothesimplerandomsamplingdesign).D=DNase;T=traditional.
ConditionalRates(ModeratorIncluded).Figure4.24(followingpage)showstheeffectof
testingmethodonthetestingoutcomeswhenweincludethepresenceofspermasamoderatorinthe
model.Thetoppanelsshowtheresultswhenspermwasabsent;thebottompanelsshowtheresults
whenspermwaspresent.ComparingFigures4.22and4.24highlightsthefactthateachpanelinthe
formerisessentiallyaweightedaverageofthetopandbottompanelsfromthelatter.86Intheabsence
ofsperm,therewaslittledifferenceintheDNAtestingrates(OR=1.32,95%CI=[0.52,3.32],p=.556),
whichwere15.4%(95%CI=[08.5,26.3])forSAKsinthetraditionaltestgroupand19.4(95%CI=[11.3,
86
TheweightswouldbebasedontherelativenumbersofSAKsthatcontainversusdonotcontainsperm.
221
FIGURE4.24SimpleMainEffectsofTestingMethodonDNATestingand
CODISEntryRatesbySpermAbsent/Present(TestingGroup4[DNAMethod])
DNA Testing Rate
CODIS Entry Rate
OR = 1.320, p = 0.556
95% CI = [0.524, 3.323]
OR = 2.000, p = 0.489
95% CI = [0.275, 14.567]
0.40
0.20
0.00
DNA Testing Rate
1.00
0.60
0.40
CODIS Entry Rate
Sperm Present
0.80
Sperm Present
0.60
Sperm Absent
0.80
Sperm Absent
1.00
0.20
OR = 1.000, p = 1.000
95% CI = [0.766, 1.305]
OR = 0.692, p = 0.317
95% CI = [0.333, 1.437]
0.00
Traditional
DNase
Traditional
DNase
DNA Testing Method

Figure4.24:SimpleMainEffectsofTestingMethodonDNATestingandCODISEntryRatesAmongSOLunexpiredDetroit
SAKs,DependingonWhetherSpermWasAbsentFromorPresentintheSAK.TheDNAtestingrateisanunconditional
estimate(theproportionofSAKsthatpasstheStage0serologyscreeningtoreachStage1(theactualDNAtest).TheCODIS
entryrateisaconditionalestimate(theproportionofSAKstestedthatyieldedaDNAprofilesuitableforuploadintoCODIS).
Theseresultsgeneralizetothesubpopulationofuntested,SOLunexpiredDetroitSAKs(regardlessofadjudicationstatusor
victimoffenderrelationship).AfterselectingN=350SAKsfromthatsubpopulation,n=175SAKswererandomlyallocatedto
eachofthetwotestingmethods.TheSAKsweresortedbythepresence/absenceofSperminthesamples.Theseestimates
wereobtainedfromacontinuationratiomodelofSAKprogressionacrossStages02thatcontaineda3waystagebytesting
methodbypresenceofsperminteractioneffect(unweightedduetothesimplerandomsamplingdesign).Thedotsmarkthe
pointestimatesfortherates,thewhiskersdepictthecorresponding95%CIs.Theoddsratios(OR)andassociated95%CIs
quantifythesimpleeffectoftestingmethodontheratenamedineachpanel.
222
31.2])intheDNasetestgroup.TheRR=1.26showsthatSAKsintheDNasegroupwereonly1.26times
morelikelytopassthescreeningandactuallyreceiveDNAtestingthanSAKsinthetraditionaltest
group.WewouldneedtosubmitNNS=25.00SAKsfromtheDNasegrouptodetectjustonemoreSAK
withsufficientbiologicalevidencetowarrantDNAtesting(4.85SAKs)thanwewouldexpectamongthe
samenumberofSAKsfromthetraditionalgroup(3.85SAKs).Thereisnodetectabledifferencebetween
thegroupsinDNAtestingratewhenspermispresent(OR=1.00,95%CI=[0.77,1.31],p=1.00)
becausetheDNAtestingratequitepredictablyincreasedto100.0%forbothgroups(spermisafterall
oneofthekindsofbiologicalevidencethatwarrantmovingontoStage1).Randomassignmenttothe
twogroupsappearstohaveeliminatedsystematicdifferenceswithrespecttothepresenceofbiological
evidenceasintended.
Whenspermwasabsent,theconditionalCODISentryratewas20.0%(95%CI=[05.0,54.1])for
SAKsinthetraditionaltestgroupand33.3%(95%CI=[12.9,62.8])intheDNasegroup(seeTable4.9,
followingpages).Thewideconfidenceintervalsfortheseestimates(toprightpanel,Figure4.24)are
likelyduetothelowDNAtestingrates,whicheffectivelyreducedthesamplesizeandincreasedthe
uncertaintysurroundingtheestimatedproportionsandthecorrespondingeffectsize.Thus,although
thedifferenceintheoddsofaCODISentrywasnominallymoderateandnonsignificant(OR=2.00,95%
CI=[0.28,14.57],p=.489),itcouldplausiblybesubstantiallylargerandfavoreithergroup.Thereisfar
toomuchsamplingvariabilitytobecertaineitherwaybecausetheCIspansarangeconsistentwith
everythingfromalargenegativeeffecttoaverylargepositiveeffect.TheRR=1.67indicatesthatwhen
SAKscontainingbiologicalevidencebutnospermareallocatedtoDNasetesting,theyareabout1.67
timesmorelikelytoyieldCODISentriesthansimilarSAKsallocatedtothetraditionaltest.Furthermore,
theNNS=7.52meansthattesting7.52SAKscontainingbiologicalevidencebutnospermviaDNase
testingwouldlikelyyieldonemoreCODISentrythanwewouldexpectfromtraditionaltestingofa
similarnumberSAKscontainingbiologicalevidencebutnosperm(i.e,2.50versus1.50CODISentries).
223
DirectlyexaminingthedifferenceintheconditionalCODISentryrates(=pTpD,ratherthan
thedifferenceintheoddsofCODISentries)whenspermwasabsentturnsupaconflictingfinding.Both
theobserveddifferenceof13.3%andtheentire90%CI=[15.7,10.9]falloutsidethemarginof
equivalence,indicatingthattheratesarenotequivalentbecauseDNaseyieldsameaningfullyhigher
conditionalCODISentryratethantraditionaltestingwhenthereisnosperminanyoftheSAKstested.
Thislooksatfirstglancelikestrongevidencefornonequivalence;unfortunately,bothoftheseratesare
estimatedfromverysmallnumbersofSAKs(1012SAKspergroup).Werecommendviewingthisresult
withcautioninlightofitsdiscrepancywiththeresultsfromthecontinuationratiomodel,whichsuggest
thatnosuchdifferencewasclearlydiscerniblefromthesamedata.
Whenspermwaspresent,theconditionalCODISentryratewas86.4%(95%CI=[78.5,91.7])for
SAKsinthetraditionaltestgroupand81.4%(95%CI=[73.0,87.6])intheDNasetestgroup.Thiswould
beconsideredasmall,statisticallynonsignificant,negativeeffect(OR=0.69,95%CI=[0.33,1.44],p=
0.317)ontheoddsofaCODISentrythatcouldreflectnothingmorethanrandomsamplingvariation.
However,thewideCIfortheORisconsistentwitharangeofplausibleeffectsizesfallinganywhere
betweenalargenegativeeffectandasmallpositiveone. TheRR=0.94indicatesthatSAKsallocatedto
DNasetestingareabout0.94timeslesslikelytoyieldCODISentriesthanSAKsallocatedtothe
traditionaltest.Furthermore,theNNS=20.00meansthattesting20.00SAKscontainingspermvia
DNasetestingwouldlikelyyieldonelessCODISentrythanwewouldexpectfromtraditionaltestingofa
similarnumberSAKscontainingsperm(i.e,17.28versus16.28CODISentries).
TheequivalencetestdirectlyexaminingthedifferenceinconditionalCODISentryratesalso
turnedupanunusualresultwhenspermwaspresent.Theobserveddifferenceof4.9%infavorof
traditionaltestingliesrightbelowtheupperboundforthemarginofequivalence.The90%CI=[4.3,5.6]
thereforespansthatboundarywithalmosthalfoftheintervaloneachside.Thismeanswedonothave
224
unambiguousevidenceforequivalence.Thetworatesmightdifferbylessthan5%,ortheymightdiffer
byalittlemorethanthat.
TABLE4.9RatesEstimatedfromContinuationRatioModelComparingDNATesting
MethodsbySpermAbsent/Present(TestingGroup4[DNAMethod])
Outcome
Traditional
DNase
Rate[95%CI]
Rate[95%CI]
SpermAbsent
DNAtestingrate
Stage01
CODISentryrate
Stage12|Tested
0.154[0.085,0.263]
0.194[0.113,0.312]
0.200[0.050,0.541]
0.333[0.129,0.628]
SpermPresent
DNAtestingrate
Stage01
CODISentryrate
Stage12|Tested
1.000[1.000,1.000]
1.000[1.000,1.000]
0.864[0.785,0.917]
0.814[0.730,0.876]
UnconditionalRates(ModeratorIncluded).Figure4.25(followingpage)translatestheresults
fromFigure4.24intounconditionalrates.Thescreeningrateremains100%regardlessoftesting
methodandthepresenceofspermbecauseallSAKssubmittedfortestingarescreenedatStage0;the
DNAtestingrateremainsthesameasinFigure4.24becauseitrepresentsthefirststagetransitioninthe
model.Thus,onlytheCODISentryratescandifferacrossFigures4.24and4.25becausetheyuse
differentdenominators(Figure4.25usesthetotalnumberofSAKssubmittedatStage0,whileFigure
4.24usesthenumberofkitsreachingStage1).
TheunconditionalCODISentryratewas3.1%forthetraditionalgroupand6.5%fortheDNase
groupwhenspermwasabsent,yieldingaRR=2.10,suggestingthatweareabouttwiceaslikelyto
obtainCODISentrieswhenusingtheDNasetest.Submitting29.41SAKstoforensictestingislikelyto
225
yieldjustonemoreCODISentryunderDNasetestingthanwewouldgetifweusedtraditionaltesting
instead(1.91versus0.91entries).Theequivalencetestshowsthatthe3.4%differenceinthe
unconditionalCODISentryratewhenusingDNasetestingissmallenoughtobeconsideredequivalent
(90%CI=[3.9,2.9])totherateexpectedfromtraditionaltestingwhenspermwereabsent.Wehave
moreconfidenceinthisresultthanwedidforthecorrespondingconditionalratesbecauseitisbasedon
largersamplesizes(6265SAKspergroup).Italsoseemsprudenttonotethatbothratesarelowinthe
firstplace.
FIGURE4.25UnconditionalDNATestingandCODISEntryRatesbyDNATestingMethodandSperm
Absent/Present(TestingGroup4[DNAMethod])
Sperm Absent
D = 1.000
Stage Reached
1. DNA test
2. CODIS entry
T = 1.000
D = 0.194
T = 0.154
D = 0.065
T = 0.031
Sperm Present
D = 1.000
T = 1.000
D = 1.000
1. DNA test
T = 1.000
D = 0.814
2. CODIS entry
T = 0.864
0.00
0.20
0.40
0.60
0.80
1.00
Figure4.25:UnconditionalDNATestingandCODISEntryRatesAmongSOLunexpiredDetroitSAKsbyTestingMethod,
separatelyforPresenceofSperm.AllratesshownhereareunconditionalestimatesoftheproportionofSAKstestedthat
reachedeachstagelisted.Theseresultsgeneralizetothesubpopulationofuntested,SOLunexpiredDetroitSAKs(regardlessof
adjudicationstatusorvictimoffenderrelationship).Theseestimateswerecalculatedfromtheresultsofacontinuationratio
modelofSAKprogressionacrossStages02(unweightedduetothesimplerandomsamplingdesign).D=DNase;T=traditional.
226
TheunconditionalCODISentryrateswereidenticaltotheconditionalrateswhenspermwas
present(86.4%forthetraditionalgroupand81.4%fortheDNasegroup)becausetheDNAtestingrate
of100.0%meanstheyusethesamedenominator.Therefore,theRRandNNSstatisticsandthe
equivalencetestresultsfortheunconditionalratesmatchthosereportedabovefortheconditional
rateswhenspermispresent.
Comparisons of Testing Costs and Personnel Effort Between Testing Methods

Inthisexperiment,wealsoexaminedwhetherthetwomethodsdifferwithrespecttocostand
personneleffortrequired.Assumingequalperformancewithrespecttoforensicoutcomes,onemight
preferthetestingmethodthatcostslessormoreenablespersonneltoworkmoreefficiently.The
availabledataoncostandpersonneleffortweresometimesrecordedasaggregatevaluesforbatchesof
multipleSAKsandothertimesrecordedseparatelyforeachSAK.Thisinconsistencyinthelevelofdetail
recordedforcedustoaggregatecostandpersonneleffortdata.Assuch,wereportonlydescriptive
summariesbecausewearenotawareofanystatisticalmethodthatcanadequatelyquantifythe
samplingvariationexpectedaroundtheestimatesgiventheinconsistentwaythedatawererecorded.
Table4.10(followingpage)summarizescostofconsumablesuppliesusedduringtesting,the
amountoflaboratorypersonneltimespentontestingandreviewingtheresults,andtheamountof
statepoliceforensicsciencedivisionpersonneltimespentreviewingtheDNAtestresults.Therewas
littledifferenceintotalsuppliescosts($16.20total,$0.09perSAK)betweenthetwomethods,which
couldbeentirelyattributabletothedifferenceincostsattheStage0screening.
Theaggregatetimespentbylaboratorypersonnelontraditionaltestingwas848.50hours,
whichwas192.50morehoursthanlaboratorypersonnelspentonDNasetesting(656.00hours)forthe
samenumberofSAKs(N=175each).Mostofthedifferencewastimespentontheactualtesting
(181.75hours)ratherthanreviewingresults(10.75hours).So,DNasetestingappearstohavesavedan
227
averageof1.10hoursoflaboratorypersonneltimeperSAKrelativetotraditionaltesting.87Cumulated
acrossalargecollectionofSAKs,thismayyieldsubstantialsavingsonpersonnelcosts,butwecannot
assesshowmuchsamplingvariationonemightexpectintheseestimatesfromthecurrentdata.The
trueaveragetimesavingsforlaboratorypersonnelcouldbehigherorlower,butwecannotprovidea
validCIbecausethedatawerealreadypartiallyaggregatedwhenwereceivedthem.
TABLE4.10CostandPersonnelEffortComparisons(TestingGroup4[DNAMethod])
Traditional
DNase
Difference
Variable
(N=175)
(N=175)
(TD)
Totalconsumablesuppliescost($)
52,986.76
52,970.56
16.20
6,406.12
6389.92
16.20
46,580.64
46,580.64
0.00
302.78
302.69
0.09
SAKsnegativeatscreening(noDNAtest)
SAKspositiveatscreening(DNAtest)
MeanconsumablesuppliescostperSAK($)
Vendorlaboratorypersonneleffort(hours)
Totaltestingtime
780.75
599.00
181.75
67.75
57.00
10.75
848.50
656.00
192.50
MeantestingtimeperSAK
4.46
3.42
1.04
MeanreviewingtimeperSAK
0.39
0.33
0.06
Meantesting+reviewingtimeperSAK
4.85
3.75
1.10
Totalreviewingtime
Totaltesting+reviewingtime
Statepoliceforensicsciencepersonnel
effort(hours)
Totalreviewtime
MeanreviewtimeperSAK
119.90
109.05
10.85
0.69
0.62
0.07
Note:SampleincludedN=350SAKs(175/group).Costestimatesincludewaste,controls,andreprocessing.SAK=sexual
assaultkit.
87
Laboratorypersonnelspentanaverageof848.50hours/175SAKs=4.85hours/SAKontraditionalDNAtesting;theyspentan
averageof656.00hours/175SAKs=3.75hours/SAKonDNasetesting.Therefore,DNasetestingmethodsavedanaverageof
4.853.75=1.10hours/SAKinlaboratorypersonneltime.
228
Theaggregatedifferenceinstatepoliceforensicsciencedivisionpersonneltimespenton
reviewingtestresultswassubstantiallysmaller(atotalof10.85hours).Reviewingresultsfrom
traditionaltestingtookstatepoliceforensicsciencedivisionpersonnelatotalof119.90hours,while
reviewingresultsfromDNasetestingtookatotalof109.05hours.Onaverage,thatsuggestsadopting
DNasetestingcouldsaveabout0.07hoursofpersonneltimeperSAK.Again,wecannotassesshow
muchsamplingvariationonemightexpectforthatfigure.Thetrueaveragetimesavingsforstatepolice
forensicsciencedivisionpersonnelcouldbehigherorlower,butwecannotprovideavalidCIbecause
thedatawerealreadypartiallyaggregatedwhenwereceivedthem.
Summary & Conclusions: Empirical Findings Regarding the Utility of SAK Testing
TheDetroitSAKTestingPlanwasdesignedtoprovideempiricaldataregardingifandhowSAKs
couldbeselected/prioritizedfortesting.Basedonourinterviewswithstakeholdersatthelocal,state,
andnationallevel,wefocusedontwokeyvariablesthatcouldbeusedasselectioncriteria:victim
offenderrelationship(stranger/nonstranger)andstatuteoflimitationsstatus.Atissuewaswhether
CODISentryrates,CODIShitrates,andserialsexualassaultratesvariedasafunctionofthesevariables.
Theresultsfromtheconditionalandunconditionalrateanalysessuggestthatthereisa
significant,thoughsmall,differenceinCODISentryratesasafunctionofvictimoffenderrelationship
(higherentryratesforstrangerperpetratedassaults).However,therearenodifferencesinCODIShit
ratesforstrangerandnonstrangersexualassaults.Inotherwords,strangerandnonstrangerSAKsdo
notsignificantlydifferintheirCODIShitrates.ManystakeholdersinDetroit,aswellasthoseatthestate
andnationallevel,feltthatstrangerperpetratedassaultsshouldbeprioritizedovernonstranger
assaultspresumablybecausetheywouldyieldmoreCODIShits.Thesedatadonotsupport
prioritizationoftestingonthebasisofvictimoffenderrelationshipCODIShitratesarenotstatistically
differentforstrangerandnonstrangerassaults.However,therateofserialsexualassaultswas
229
significantlyhigherforstrangerperpetratedassaults.Meaning,CODIShitratesarestatistically
equivalent,butoncethereisahit,thestrangerperpetratedassaultsweremorelikelytobehittingto
othersexualassaults(serialsexualassaults).Yet,itisimportanttoconsiderhistoricalbiasesinSAK
submissionpractices,whichmayhelpexplaintheseeffectsregardingvictimoffenderrelationshipand
serialsexualassaults.Theextenttowhichlawenforcementpersonnelhavebeensystematicallynot
submittingnonstrangerSAKsfortesting(e.g.,becausetheresnopoint,identityisknown)meansthat
CODISisnotaspopulatedwithDNAsamplesfromoffenderswhoassaultthoseknowntothem.Inother
words,offenderswhospecializeinnonstrangerserialsexualassaultsmaybeunderdetected,given
historicalpracticesinSAKsubmissions(whichaffecthowCODISispopulated).Ataminimum,these
resultssuggestsaneedforfurtherresearchonserialsexualassaultsandvictimoffenderrelationshipto
explorewhetherourfindingsofhigherserialratesamongstrangerrapesarereplicated,orwhether
serialsexualassaultratesarecomparableamongstrangerandnonstrangerassaultsinother
jurisdictions.
TurningtotheresultsregardingtheeffectofSOLstatusonforensictestingoutcomes,the
resultsfromtheconditionalandunconditionalrateanalysessuggestthatthereisnosignificant
differenceinCODISentryrates,CODIShitrates,andserialsexualassaultratesasafunctionofthe
statuteoflimitationsstatus.Inotherwords,SAKsassociatedwithcasesthatarepresumedtobe
beyondthestatuteoflimitationshavestatisticallyequivalentCODISratesasdocasesstillwithinthe
statuteoflimitations.Again,somestakeholdersinDetroit(aswellasthoseatthestateandnational
level)advocatedforprioritizingSAKsbystatuteoflimitations,suchthatcasesstillwithinthestatute
shouldbegivenhighertestingprioritythanthosethatarepresumedtobeexpired.Froman
investigationandprosecutionpointofview,suchprioritizationcouldbewarranted(thoughadditional
researchwouldbeneededtotestthatassumption),butwithrespecttoforensicoutcomes,theseresults
suggestthatskippingolderkitsisnotwarrantedbecausetheratesofCODIShitsarenotstatistically
230
differentforSOLexpiredandunexpiredSAKs.Putanotherway,theseresultsclearlyindicatethatthere
ismeritintestingpresumedSOLexpiredSAKs,intermsofCODISentry,CODIShits,andidentificationof
serialsexualassaults.
Wealsoexaminedwhetherotherfeaturesofthevictim,assailant,and/orassaultmightpredict
forensictestingoutcomes,andthereforecouldbeusedapossibleselection/prioritizationcriteria.We
testedwhethervictimage,assailantage,andassaultcharacteristics(multipleperpetrators,alcoholor
druguse,useofaweapon,useofphysicalforce,timebetweenassaultandexam)predictedwhethera
SAKwouldyieldaCODIShit(analysesconductedseparatelywithinTestingGroups1(Stranger),2(Non
Stranger),and3(PresumedSOLExpired)).Theseanalysesdidnotyieldmanysignificantpredictors.
Weaponusewassignificantinallmodels,indicatingthatiftheassailanthadusedaweaponinthe
assault,theSAKwasmorelikelytoyieldaCODIShit.InTestingGroup2(NonStranger),assaults
committedagainstyoungervictims(<16)werelesslikelytoyieldaCODIShit.Recallthatouranalysesin
Chapter3:WhySoManyUnsubmittedSAKsinDetroitshowedthatlawenforcementpersonnelwereless
likelytobelieveadolescentrapevictimsandthatdisbeliefaffectedtheirdecisionnottosubmitaSAKfor
testing.TheextenttowhichthisbiasinSAKsubmissionpracticeshasbeenoccurringovertimeandin
otherjurisdictions,thenoffenderswhospecializeinthisformofassault(adolescentvictimsthey
know)maybelesslikelytobeinCODIS(hence,fewerCODIShits).Thissuggeststheneedforadditional
researchonoffenderspecialization,particularlywithrespecttovictimage.
However,onthewhole,mostofthevariablesinthesepredictionmodelswerenotstatistically
significant.Furthermore,therangeofplausibleeffectsizesassociatedwitheachpredictorwastypically
ratherbroad.Evenwherewehavesomegroundstoconcludethattheprobabledirectionofaneffect
(i.e.,whetheritincreasesordecreasesthehitrate)isreasonablyclear,thesemodelsprovideonly
impreciseestimatesofhowstrongtheeffectsare.Althoughsomeofthesepredictorscouldexertlarge
effectsonthehitrate,furtherresearchwillbenecessarytoestablishmoreconclusivelywhetherthey
231
reallydoso.Itisalsoentirelyplausiblethateveryeffectweexaminedisquitesmall.Assuch,we
concludethatthemodelsdonotprovidesufficientevidencetoadvocateforselectivelytestingSAKs
basedonvictimage,assailantage,examtiming,oranyoftheassaultcharacteristicsweexamined.88
Finally,TestingGroup4wasanexperimentalcomparisonbetweentwoDNAtestingmethods:
traditionalvs.selectivedegradation(DNase).Theresultsindicatednosignificancebetweenthetwo
groups,indicatingthattheselectivedegradationmethodhadnodecrementinperformancerelativeto
customarymethods.FollowuptestsofequivalenceestablishedthattheratesofDNAtestingandCODIS
entrywerefunctionallyequivalentacrossthetwomethods.Comparisonsofmaterialscostswerealso
equivalentacrossthetwogroups,buttheselectivedegradationmethodofferedasubstantialsavings
staffinginstafftimeininterpretation/review:1.10hoursoflaboratorypersonneltimeperSAKrelative
totraditionaltesting.CumulatedacrossalargecollectionofSAKs,thismayyieldsubstantialsavingson
personnelcosts.Thesedatasuggestthatselectivedegradationisapromisingmethodthatcouldoffer
forensiclaboratoriessignificantpersonnelsavings,withoutsacrificingoutcomes.Theseresultsmerit
replicationinotherlabs/settingspriortobroadbasedimplementation.
88
Onlyunambiguousevidenceofmoderatetolargeeffectscouldjustifysuchselection,buttheseanalysesdonotprovidesuch
evidence.
232

TestingpreviouslyunsubmittedSAKsraisescomplexissuesregardingif,when,andhowvictims
oughttobenotifiedaboutwhathadhappenedtotheirkit,whatiscurrentlyhappening,andwhatmight
happeninthefuture.Thereare,ofcourse,legalissuestobesortedoutregardingthetestingresultsand
theviabilityoffutureprosecution,butfundamentally,victimnotificationisreopeningsurvivors
memoriesandfeelingsregardingasignificanttraumaticeventintheirlives,onethatmayhave
happenedafewyearsago,oradecadeago,ordecadesago.Althoughthereisnopriorresearch
specificallyonvictimnotificationsincasesofunsubmittedSAKs,thereisanextensivebodyofliterature
ontraumaanditsaftermath,whichoffersacautionarytaleaboutwhatthisprocesscouldmeanfor
survivors.Activatingtraumaticmemoriestriggersneurobiologicalandphysiologicalresponsesinthe
brainandbodythatareemotionallyandphysicallydistressing(Banks,2002;Foaetal.,2000;Garfinkel&
Liberzon,2009;Roozendaaletal.,2009;Rubinetal.,2008;Sher,2010;Southwicketal.,2005).
Recoveryfromtraumaticeventsisoftenalongprocess,andmanyvictimsstruggleforyearswith
symptomsofdepression,posttraumaticstressdisorder(PTSD),suicidality,substanceabuse,and
addictionrelapse(seeCampbell,Dworkin,&Cabral,2009;Herman,2002;Steenkamp,Dickstein,Salters
Pedneault,Hofman,&Litz,2012forreviews).Assuch,victimnotificationcanposesignificantemotional
andphysicalrisksforsurvivors,whichmayormaynotbemitigatedbytheinformationprovided(e.g.,
testingresults,CODIShits,suspectidentification)andopportunitiesforactionand/orclosure(e.g.,
investigationandprosecution).
Giventhiscontext,amultidisciplinaryeffortwasessentialforplanningandconductingvictim
notifications.Asonestakeholderinthisprojectnoted,weseethingsdifferently(thedifferent
disciplines/professions),weretraineddifferently,havedifferentjobs,sometimesdifferentpriorities
233
honestly,butItrulybelieveweallwanttodowhatsbestforvictims.Continuingonwiththatlineof
thinking,anotherstakeholdercommented,TherearesomanywaystolookatthisandIonlyknowmy
way...putusalltogethertotalkaboutit,Icanstarttoseeitfromdifferentpointsofviewandwecan
figureoutawaytogo.Drawingupontheirdifferentprofessionalexperiences,lessonslearnedfrom
otherjurisdictions,theresearchliteratureontheneurobiologyoftrauma,consultationswithlegal
ethicistsandtraumaexperts,and,asonememberofthecollaborativestated,agutcheckeverysingle
daytorememberwhatthismeansforthem(survivors),theDetroitcollaborativecreated,implemented,
andevaluatedavictimnotificationprotocolforunsubmittedSAKs.Thepurposeofthischapteristo
summarizehowtheteamcreatedtheprotocol,describetheprotocolitself,andthenpresentthe
evaluativefindingsregardingitsefficacyforlocatingandnotifyingsurvivors.
The Process of Developing the Victim Notification Protocol
TheDetroitcollaborativediscussedvictimnotificationfromtheinceptionoftheproject,butthe
imminenttestingresultsfromthefirstgroupofSAKspromptedthegrouptoformalizetheirideasintoa
protocol.Therewereastaggeringnumberofcomplexissuestoconsider,including,butnotlimitedto:
WhenandwhyavictimshouldbenotifiedAtwhatpointshouldvictimsbenotifiedthattheir
kitshadnotpreviouslybeentested?Shouldtheyhavebeennotifiedevenbeforethekitwas
sentfortesting?Shouldvictimsbenotified...onlyifDNAisfoundintheSAK,onlyiftheresa
CODIShit,onlyiftheresstillachanceforprosecution?Shouldvictimsbenotifiedregardlessof
thetestingresults?Whatiftheresnothingtotellvictims(e.g.,thetestingrevealednoDNA,
noCODIShit,statuteoflimitationshasexpired)butmaybenothingreallyisameaningful
updateforsomevictims?Whatifsomesurvivorsdonotwantthispartoftheirlivesreopened?
234
WhoshoulddecidewhetheravictimwillbenotifiedIsthereawayforsurvivorstobeableto
makethischoiceforthemselves(ifandwhentobenotified)?Ifnot,thenwhoshouldmakethe
decisiontonotifyavictim?Police,prosecutors,advocates,amultidisciplinaryteam?
WhoshouldnotifyvictimsWhichprofession,whichspecificindividualsshouldbetaskedwith
findingvictimsandgivingthemsuchemotionallycharged,complexinformation?Shoulditbe
thepolicetheoriginalinvestigatingofficer(ifpossible),anewofficerwhohadnotbeen
involvedintheoriginalcase,ormaybeanindependentlawenforcementofficernotaffiliated
withthefocalpolicedepartmentatall?Giventhatthismightbeatraumaticeventforvictims,
perhapsnotificationshouldbeconductedbyadifferentdisciplineentirely,suchasanadvocate
orperhapsaforensicnurse?Whoeverwillbeconductingthenotifications,howcanweprepare
theseprofessionalsforthiscomplextask?
HowshouldvictimsbecontactedIsthereabestwaytocontactvictims?Whatstheright
waytodeliverthisinformationByletter?Byphone?Inperson?Whatarethesafety,privacy,
andconfidentialityissuesofdifferentmethodsofnotification?
Whatshouldnotificationpersonnelsaytovictims(whetherbyletter,phone,orinperson)
Whatinformationcanandshouldbesharedwithvictimsandwhen?Ifvictimsarehighly
traumatizedbythenotification,whatwilltheybeabletoabsorbandremember?Shouldthere
beanapology,andifso,whatshouldthenotifyingpersonnelsayandaretherelegalliability
implicationsofanapology?
235
Towhatextentshouldnotificationsbestandardized(i.e.,followingspecificstepbystep
protocols)Giventhateachvictimisuniqueandeachcaseisunique,whatsthebestwayto
balancecasebycaseconsiderationswithensuringthatallnotifyingpersonnelcoverkey
componentsandallvictimsreceivecoreservices?
Howcannotifyingpersonnel(specifically)andthecommunity(generally)supportvictims
Whatresourcesandsupportsmightvictimsneedimmediately,shortterm,and/orlongterm?
Whatisthebestwaytoconnectsurvivorstoservicesandresources?Howcanweprepare
providerstoworkwithvictimswhosekits/casesmightbereopenedaftermany,manyyears?
TheDetroitcollaborativedecidedtohaveatwodayretreatinordertoallocatesufficienttimeto
discusstheseissuesindepthandtocraftapilotprotocolawelldeveloped,butstillprovisionalplan
thatwouldbeimplementedwithasmallnumberofcasesandevaluatedbytheresearch/evaluation
teamtodeterminewhatstrategiesandmethodswerewellreceivedbysurvivors.Then,basedonthose
data,thecollaborativewouldrevisetheprotocolaccordingly.Forthatretreat,theresearch/evaluation
teamsynthesizedthegroupsdiscussionsthusfar(i.e.,discussionsthatoccurredpriortotheretreat)
andorganizedthemintoadecisiontree,outliningthedifferentchoicepointsandtheprosand
considentified(sofar)ofeachoption.Thisdecisiontreewasthenusedasaguidingframeworkfor
theretreatandwasinstrumentalinfacilitatingthedevelopmentofaprotocol.Thisdecisiontreecanbe
foundinAppendixC:VictimNotificationResources,anditmayprovideausefulframeworkforother
communitiesastheydevelopvictimnotificationprotocols.Figure5.1TheStepByStepProcessof
CreatingtheDetroitVictimNotificationProtocol(followingpages)describeshowtheDetroit
collaborativedevelopedthepilotprotocol.Figure5.1listseachissuethathadtoberesolved,the
discussionanddebateabouthowbesttoaddresseachissue,anddecisionsmadebythecollaborative
team(andwhytheydecidedwhattheydid).
236
FIGURE5.1TheStepbyStepProcessofCreatingtheDetroitVictimNotificationProtocol
DISCUSSION1A.AsSAKs
begantobesubmittedfor
testing,thecollaborative
begantoactivelyplanfor
howtonotifyvictims
regardingthetesting
results.Thegroupwanted
tolearnaboutthe
experiencesofThe400
Projectandother
organizationstoavoid
reinventingthewheel.
ISSUE1
HOW&WHEN
SHOULDTHE
COLLABORATIVE
STARTPLANNING
FORVICTIM
NOTIFICATION?
DECISION1A.Thecollaborativeengaged
inamultiweeklearningprocessbefore
developinganyprotocolregardingthe
victimnotificationprocess.
PresentationsweregivenbyThe400
Project,frontlinepractitioners
(prosecution,lawenforcement,and
advocacywhohaddoneCODIShit
notificationbefore),&representatives
fromanationalvictimadvocacy
organization.Thecollaborativealso
reviewedwebinarsfromothernational
victimorganizations
Theresearch/evaluationteamreviewed
theliteratureontheneurobiologyof
traumaanditsimplicationforvictim
notificationandpresentedasummary
tothegroup.
DISCUSSION1B.Itwas
difficulttodevotesufficient
timetodiscussvictim
notificationindepthatthe
regularlyscheduled
meetings,butthegroup
neededtodevelopa
notificationplanquicklyas
theSAKswouldsoonbe
returningfromtesting.
Thisprocessshedsomelightonthe
issues,questions,&multidisciplinary
differencesthatwouldpresentcropup
duringthisproject.
DECISION1B.Theteamdecidedtohave
atwodayweekendretreatinorderto
havethesufficienttimeandspaceto
developvictimnotificationprotocols.
237
ISSUE2
WHATSHOULD
BEADDRESSED
ANDDISCUSSED
DURINGTHE
TWODAYVICTIM
NOTIFICATION
RETREAT?
DISCUSSION2A.The
materialsfromthe
presentationsonvictim
notificationneededtobe
synthesizedintochoice
pointsanddecisiontree
flowchartsthatcouldbe
usedtoguidediscussionat
theretreat.
DECISION2A.Thedecisiontree
flowchartsdevelopedidentified9key
questionstobediscussedattheretreat.
(SEE9QUESTIONSBELOW)
DISCUSSION2B.
Recognizingthatthe
multidisciplinaryteamhad
varyingopinionsaboutthe
victimnotificationprocess,
developingaplanforhow
allvoiceswouldbeheard
attheretreatwas
essential.
DECISION2B.Thenationalviolence
againstwomenadvocacyfoundation(a
memberofthecollaborative)identifieda
professionalfacilitatortoguidethe
retreat.Thefacilitatorhadsubstantive
experienceworkingwithtrauma
survivors&processexperiencehelping
communitiesdevelopserviceprograms.
DISCUSSION2C.Thegroup
wantedtoleavetheretreat
withadraftoftheprotocol
&thereforeneededto
reachanagreementon
eachdecision.
DECISION2C.Toexpeditedecision
making,representativesfromeach
disciplinewerehighrankingindividuals
whohadtheauthoritytomakedecisions
onthebehalfoftheirorganizations.The
facilitatorguidedthegrouptoreach
consensus.
9KEYQUESTIONSTOCONSIDERWHEN
DEVELOPINGAVICTIMNOTIFICATIONPLAN:
1. Howwillweapproachvictimnotification?
2. Whenandwhydowenotifyvictims?
3. Whomakesthedecisionthatavictimwillbe
notified?
4. Whatdowehopetoachieveinthefirst
contactwiththevictim?
5. Whoshouldcontactthevictimfirst?
6. Howshouldfirstcontactbemade?
7. Whatinfoshouldwegivevictimsduring
thefirstcontact?
8. Whathappensafterinitialcontact?In
subsequentcontacts?
9. Howshouldvictimnotificationstaffbe
trained?
238
PLANNING
QUESTION#1
WHATISOUR
APPROACHTO
VICTIM
NOTIFICATION
SeeAppendixC,
VictimNotification
RetreatPlanning
Resources,for
moredetails
DISCUSSIONQ1:The
collaborativeidentified
threepossibleapproaches
forvictimnotification:
1) astandardized
approach:allcasesare
handledthesameway
2) acasebycase
approach:notification
methodsaretailored
toeachindividualcase
3) ahybridapproach:
eachcaseishandledis
informedbyasetof
guidingprinciples,but
howtheguiding
principlesareexecuted
wouldbedecidedona
casebycasebasis.
DECISIONQ1:TheexperiencesofThe
400Projectindicatedthatitwould
difficulttoimplementastandardized
protocolbecauseeachcaseisunique.
However,thegroupwasnot
comfortablewithanentirelycaseby
caseapproachbecausetheywantedto
ensurethatnotificationstafffollowed
keyguidingprinciplestoensurethatall
victimsreceivedcoreinformation,
services,referrals,etc.
Thecollaborativedecidedonahybrid
approach,wherebyspecificnotification
methodsmightvarycasebycase,but
generalprinciplesregardinghow
notificationshouldunfoldwouldbe
consistentacrosscases/survivors.
239
PLANNING
QUESTION#2
WHEN&WHY
DOWENOTIFY
VICTIMS
DISCUSSIONQ2:Thegroup
discussedwhenandwhyto
notifyvictimsby
consideringthenatureof
theinformationthatthey
couldberelayingtovictims
(i.e.,possibletesting
outcomes,suchasnoDNA
found,DNAwasfoundbut
therewasnoCODIShit).
TESTINGRESULTS&
NOTIFICATIONPROCEDURES:
DNAprofile Active outreach

&CODIShit
tovictims
CODISprofile
nohit(yet)
NoDNA
profile
Activeoutreach
tovictims
Noactive
outreachto
victims,butmake
testingresults
available
DECISIONQ2:Thecollaborativedecided
thatwhen&whytonotifyvictims
dependsonthenatureofthetesting
results(seeinsetforhownotification
wouldbehandleddifferentlybasedon
thetestingresults).
Acrossallofsituations,survivorswould
havethechoicetooptin/optout.In
casesselectedforactiveoutreach(i.e.,
directcommunicationtospecific
individuals),survivorscouldoptinand
agreetotalk/meetwithnotifying
personnel;ortheycouldoptoutand
refusetotalk/meet,andthatdecision
wouldberespected.
Insituationsinwhichtherewouldnot
beactiveoutreachtovictims,victims
whowantedtoknowtheirresultscould
optinbymakingcontactwiththe
project;orcouldoptoutbychoosing
nottoinitiatecontactwiththeproject.
Theoptin/optoutsystemwouldneed
tobepublicizedandwouldrequirea
complexplanningprocess,sothe
collaborativedecidedtohandlethisata
latertime.
240
PLANNING
QUESTION#3
WHOMAKESTHE
DECISIONTHATA
VICTIMWILLBE
NOTIFIED?
PLANNING
QUESTION#4
WHATDOWE
HOPETO
ACHIEVEINTHE
FIRSTCONTACT
WITHTHE
VICTIM?
DISCUSSIONQ3:While
therewereguidelinesfor
whichcaseswouldreceive
activeoutreach,thegroup
notedthattherearemany
timesinwhichactive
outreachwouldneedtobe
carefullyconsidered(e.g.,
therewasaDNA/CODIShit,
butthecaseisSOLexpired)
andwantedtocreatea
mechanismforreview&
decisionmaking.
DECISIONQ3:Thecollaborativedecided
toformamultidisciplinaryVictim
NotificationReviewTeam(NRT),which
wouldbetaskedwithreviewingtesting
results,decidingwhethertonotify
victims,anddiscussinghowbestto
approacheachvictim,givenwhatwas
knownabouttheircasesandthe
circumstancesoftheirlives.Theteam
wouldconsistoffrontlinepractitioner
representativesfromlawenforcement,
theprosecutorsoffice,community
advocacy,systemsadvocacy,andSANE.
DISCUSSIONQ4:Some
membersofthe
collaborativefeltthatthe
firstmeetingshouldfocus
onlocatingthevictim,
verifyingthatthecorrect
personhadbeenfound,
andtryingtosetupa
followuptimeforamore
indepthdiscussionofthe
issuesathand.Othersfelt
thatthefirstcontact
could/shoulddomore,
suchassharingcase
informationwiththe
victim,providingthevictim
withservices/referralsfor
services,etc.
DECISIONQ4:Intheend,thegroup
decidedthattheinitialcontactwiththe
victimshouldbesimpleandbrief.The
notifyingstaffshouldprovideanapology
tothevictimthatacknowledgedthat
thevictimsSAKhadnotbeentested
previously&shouldalsoconveyconcern
forthesurvivorandher/hiswellbeing.
Thefirstcontactshouldattempttoset
upatimeforamoreextended
discussionaboutthetestingresultsand
optionsfornextsteps.However,ifthe
victimexpressedreadinessand
willingnessforamoreindepth
discussionatthetimeoffirstcontract,
thatdecisionshouldberespectedand
thenthenotifyingstaffwouldhavea
moreextendedfirstmeeting.
241
DISCUSSIONQ5&6:Alegal
investigator,acommunity
basedadvocate,asystem
basedadvocate,and/or
forensicnursewere
consideredaspossiblefirst
contactprofessionals.
Combinationsofthese
disciplineswerealso
considered.
PLANNING
QUESTIONS
#5&6
WHOSHOULD
CONTACTTHE
VICTIMFIRST
ANDHOW
SHOULDTHE
FIRSTCONTACT
BEMADE?
PLANNING
QUESTION#7
WHAT
INFORMATION
SHOULDBE
GIVENTO
VICTIMSATTHE
FIRSTCONTACT?
DECISIONQ5&6:Itwasdecidedthata
singlelegalinvestigatorshouldmakethe
firstcontactbecauseof:1)safety
concerns(forvictimandnotifyingstaff);
2)concernofoverwhelmingthevictim
(byhavingmultiplepeoplepresent);&
3)determiningwhetherthecorrect
individualhadbeenidentified/located.
Thegroupdebatedthe
pros&consofphone,in
person,and/orletterfirst
contact.
Thefirstcontactwouldbemadein
person,orpossiblybyphone(ifvictim
hadaphone).Iftheseinperson/phone
contactswereunsuccessful,thena
generalpleasecontactusletterwould
besenttothevictimslastknown
address(withnodetailsaboutthecase).
DISCUSSIONQ7:
Informationthatcouldbe
sharedwiththevictim
included:theoriginalpolice
report,theoriginalvictim
statement(s),information
abouttheperpetrator
(includinginformation
recentlylearnedasaresult
ofaCODIShit),and/or
communityresourcesfor
victims.
DECISIONQ7:Giventhatthefirst
contactshouldbesimpleandbrief,the
retreatgroupdecidedthatvictims
shouldbeprovidedwithname/number
ofthelegalinvestigatormakingcontact,
thename/numberofacommunity
basedadvocate,thenumberfora24
hourhotline,andaresourcebrochureof
Detroitareavictimserviceprograms.
Thereleaseofanyotherinformation
wouldneedtobedecidedonacaseby
casebasis.
Aftertheretreat,staffatthe24hour
hotlinewerebriefedonthetypesof
callsthatmightbecominginandthe
researchteamdevelopedaresource
brochureofDetroitareavictimservice
programs.
242
PLANNING
QUESTION#8
WHATSHOULD
HAPPENAFTER
THEINITIAL
CONTACTWITH
THEVICTIM?
decidedthatthefirst
contactshouldbesimple
andbrief,focusingon
establishingaconnection
foralonger,subsequent
meeting.Assuch,the
groupneededtodecide
whatshouldhappenin
thosesubsequentcontacts.
DECISIONQ8:Wheneverpossible,
followupcontactwiththevictimshould
bemadebybothalegalinvestigatorand
communitybasedadvocate.These
meetingscouldoccuratthecommunity
advocacyorganizations,oratalocation
ofthevictimschoosing.Thecontentof
thesemeetingswouldbedeterminedon
acasebycasebasis,dependingonthe
testingresults,thesurvivorsquestions
andconcerns,thenextsteps,etc.
Thegroupagreedthatthetoneofthe
subsequentcontactsshouldemphasize
victimscontrolandchoicewhetherto
optin/optoutoffurthercontact.
PLANNING
QUESTION#9
HOWSHOULD
VICTIM
NOTIFICATION
PERSONNELBE
TRAINED?
agreedthatallnotification
staffshouldbetrainedona
victimcentered,trauma
informedapproachbefore
thefirstnotification
occurred.
DECISIONQ9:Theretreatgroup
brainstormedalistoftopicstobe
coveredintraining,including:sexual
assaulttrauma,strategiesforworking
withsurvivorsinastateofcrisis,
conveyinginformationabout
DNA/CODIS,workingwithspecial
populations(e.g.,thosewithdisabilities,
immigrantcommunities,etc.).
Afulldaytrainingwasscheduled,butIt
wasnotpossibletocoveralltopics
discussedattheretreatinaoneday
event,sothetrainingfocusedonthe
psychologicalaspectsofvictim
notificationandstrategiesforworking
withindividualswhohavejusthada
traumaticmemoryreopened.
243
ISSUE3
SHOULDTHERE
BEAPUBLIC
APOLOGY
REGARDINGTHE
UNSUBMITTED
SAKs?
DISCUSSION3.Duringthe
victimnotificationretreat,
therewasextensive
discussionastowhether
thereshouldbeapublic
apologytovictimsbecause
SAKshadnotbeen
consistentlytestedfor
manyyears.Therewas
markeddisagreement
surroundedwhethera
publicapologywas
warranted,andifso,which
organization(s)wouldissue
theapology.Somelocal
policeexpressedconcern
aboutthelegalrisksofan
apology/admissionof
wrongdoingwhileothers
didnotfeeltherehadbeen
widespreadwrongdoing.
Otherorganizationsdidnot
wanttobeassociatedwith
anapologybecausethe
localpolicedepartment
wasultimatelyresponsible
forSAKsubmission.
DECISION3.Anationallegalethicistwas
consultedaftertheretreattolearn
moreaboutthelegalramificationsofa
publicapologyanditspotentialimpact
onsurvivors.Thisconsultation
suggestedthatitwouldbeunlikelythat
theorganizationsinvolvedwouldface
negativelegalrepercussionsandthata
publicapologywouldlikelybe
meaningfulandreparativeforsurvivors.
However,intheendtherewasnot
broadbased,multiorganizational
supportforaissuingapublicapology.
Thecollaborativepartnershipagreed(as
notedpreviously)thatindividualvictims
shouldreceiveanapologywhen
contacted/notifiedbythelegal
investigators/communitybased
advocates.However,thegroupcould
notreachconsensusaboutapublic
apology,andsoonewasnotissued
duringthetimeofthisproject.
244
ISSUE4
HOWSHOULD
THEVICTIM
NOTIFICATION
PROTOCOL
DEVELOPEDAT
THERETREATBE
EVALUATED?
DISCUSSION4A.Thegroup
wantedtotryoutthe
protocolwitharelatively
smallnumberofcases,
approximately2530,and
thenreassesshowitwas
workingandconsider
changes.
DECISION4A.Theresearch/evaluation
teamdevelopedtrackingtoolsthatwould
chartthesteps/effortstakenbythelegal
investigatorstofindvictimsfor
notification.Theinvestigators
perceptionsofthecontactwouldalsobe
recorded.
DISCUSSION4B.The
researchteamwantedto
collecttheadvocates
perspectivesregardinghow
thenotificationprocess
unfolded.VAWA
confidentiality
requirementsdidnotallow
theadvocatestoprovide
casebycasefeedback.
DECISION4B.Theresearchteamworked
withtheStategovernmentviolence
againstwomenagencytodevelopa
mechanismwherebyadvocatescould
givegeneralizedfeedbackregardingtheir
perceptionsofhowthefirst2530
notificationcasesunfolded.
DISCUSSION4C.The
research/evaluationteam
wantedtointerview
victimsregardingtheir
notificationexperiences;
however,theprosecutors
officestaffexpressed
strongconcernsaboutthis.
DECISION4C.Ifacasecouldstillbe
adjudicated,anyindividualhaving
contactwiththevictimcouldbecalled
asawitness.Althoughtheresearchers
wouldnotbeabletotestify(per
researchconfidentiality),explainingthat
refusaltoajudge/jurywouldcomplicate
matters.TheProsecutordecidedthat
theresearcherscouldnothavecontact
withvictimsuntiltheadjudication
processwascomplete.
Note:ThecollaborativeagreedthattheNRTmeetingscouldbeobservedbytheresearch/evaluationteamfor
evaluationpurposes,andpertheIRB,norevealinginformationfromthosediscussionscouldbedisclosed.
245
The Detroit SAK Action Research Project Pilot Victim Notification Protocol
Guiding Principles
Thecoreguidingprincipleofthisprotocolisthatnotificationsshouldbevictimcenteredand
traumainformed.DrawingontheworkoftheSexualViolenceJusticeInstitute(2008),victimcentered
means:
Thevictimisatthecenterofalldecisionsregardingrecoveryandanyinvolvementwiththe
criminaljusticesystem;
Victimschoice,safety,&wellbeingisthefocus;
Theneedsofthevictimareeveryonesconcernandacollectiveeffort(notjustthetaskofone
discipline,suchasvictimadvocacy).
ConsistentwiththeworkoftheNationalCenteronDomesticViolence,Trauma,andMentalHealth
(2011),traumainformedmeans:
Attendingtovictimsemotionalsafety,aswellastheirphysicalsafety;
Strengtheningvictimscapacitytorecoverfromthetraumaticeffectsofabuseandviolenceby
providinginformation,resources,services,andsupport;
Educatingvictims,serviceproviders,andthegeneralcommunityabouttheimpactoftraumaon
survivorshealthandwellbeing.
Tothatend,allDetroitcollaborativeteammembersandallpersonnelinvolvedinvictimnotifications
participatedinafullday,victimcentered,traumainformedtrainingthatcovered:
Theneurobiologyoftrauma(generalandspecifictosexualassault);
Strategiesforworkingwithindividualsincrisis/activetraumaresponse;
StrategiesforconveyingcomplexinformationaboutDNA,CODIS,andthelegalsysteminways
thatareaccessibleforsurvivors(aFAQdocumentwaslatercreatedbasedonthesediscussions);
Strategiesforselfcarefornotifyingpersonnel.
246
The Detroit Pilot Victim Notification Protocol

Figure5.2(nextpage)summarizesthepilotprotocolthatwascreatedinthevictimnotification
retreat(andtobeevaluatedbytheresearchteam,seenextsection).Inthecontextofthisaction
researchproject,victimnotificationswouldbeinitiatedonlyifthetestingresultshadyieldedaCODIShit
(seeFigure5.2,secondoval).ThenumberofSAKsthatresultsinaCODIShitwasunexpectedlyhigh
giveninitialprojectionsfromThe400Project(seeChapter4:Developing&EvaluatingaSAKTesting
Plan),andtheorganizationsinvolvedinthisprojectsimplydidnothavesufficientresourcestoexpand
victimnotificationtoincludeothertestingoutcomes(e.g.,testingdidnotproduceaCODIShit,testing
didnotyieldDNA,etc),thoughthismayoccurinthefuture.Thepilotprotocolstipulatedthatthe
prosecutorsofficewouldreviewthecasesthathadCODIShits,andfromthatpool,selectcasesto
presenttothemultidisciplinaryNotificationReviewTeam(NRT).
Atthevictimnotificationplanningretreat,thecollaborativedecidedthattheNRTshouldconsist
oflocallevelpractitionersfromthepolice,prosecutorsoffice,systembasedadvocacy,community
basedadvocacy,andSANE/forensicnursing.Specificindividualsfromeachoftheseorganizationswere
selectedbythecollaborativetoserveontheNRT,basedontheirexperienceworkingwithsexualassault
survivors.ThoughtheDetroitSAKARPhadmembersfromstatelevelandnationallevelorganizations,
theinitialdecisionattheretreatwasthatonlyDetroitbasedpractitionersshouldserveonNRT,asthese
weretheprofessionalswhowouldbeworkingdirectlywithsurvivors.Torespectvictimsprivacy,only
thoseindividualswhowereneedtoknowshouldbepresentintheNRTmeetings.Inpractice,the
compositionofNRTchangedandrepresentativesfromstatelevelandnationallevelorganizationsdid
participate,giventhattherewerecomplexvictimadvocacyissuespertainingtoconfidentialitythat
neededtobediscussedandthegroupfeltthatbroaderbasedinputwouldbehelpful(seeFigure5.3.,
Challenges&SolutionsEncounteredImplementingtheDetroitSAKARPVictimNotificationProtocol).
247
FIGURE5.2TheDetroitSAKARPPilotVictimNotificationProtocol
248
TheprosecutorsofficedecidedwhichcaseswouldbepresentedtoNRT,thoughtherewas
agreementamongmembersofthecollaborativethattimesensitivecases(e.g.,nearstatuteof
limitationsexpirations,identifiedoffenderswhowereabouttobereleasedfromprison,etc.)should
receivepriority.Theprosecutorsalsoagreedthattheywouldpresentbothstrangerperpetratedand
nonstrangerperpetratedcasestotheNRTsothattheteamcouldconsiderifandhownotifications
mightunfolddifferentlydependingonwhetherthevictimknewher/hisassailant.TheNRTwastasked
withdecidingwhethertonotifyandconsideringhowthefactsandcircumstancesofthecasemight
shapehowbesttoreachouttothevictim.Forexample,iftheassaultwasanintimatepartnersexual
assaultthatoccurredinthecontextofdomesticviolence,theNRTwouldthoroughlydiscusssafety
considerationsforthenotification.Tobeclear,NRTwasnottaskedwithdecidingwhetheracasewould
beprosecutedthatdecisionwassolelythatoftheprosecutorsoffice.
AftertheNRThaddecidedtonotifyavictim,lawenforcementpersonnelaffiliatedwiththe
prosecutorsofficenotthefocalpolicedepartmentwouldbegintryingtolocatethevictim(see
Figure5.2,thirdoval).Manylargeprosecutorsofficeshaveaninvestigativeunit,staffedbylaw
enforcementofficerswhoworkwiththeprosecutorsoncasesthatmightbemovingforwardfor
prosecution.Inthisproject,thedecisiontohavetheseinvestigators(i.e.,thoseaffiliatedwiththe
prosecutorsoffice)conductthenotificationswasbasedontwokeyfactors.First,thefocalpolice
departmentwasstillinvestigatingcasesemanatingfromThe400Projectandtheydidnothavesufficient
staffingavailabletoworkonnewoldcases.Second,theprocessofreviewingcasefilesassociated
withpreviouslyunsubmittedSAKsclearlyindicatedthatmanyvictimsdidnothaveapositiveexperience
withthefocalpolicedepartment(seeChapter3:WhySoManyUnsubmittedSAKsinDetroit).Indeed,
manyvictimsweretreatedinvictimblaming,insensitivewaysthatlikelyexacerbatedtheirtraumaand
distress.Therefore,theProsecutordecidedthatdifferentinvestigators,notaffiliatedwiththat
department,oughttoconductthesenotifications.
249
Oncetheprosecutorsofficelawenforcementinvestigatorshadaprobablelocationforthe
victim,theywouldmakeaninpersonattempttomeetwiththevictim(orbyphone,ifinpersonwas
notfeasible)withthegoalofestablishingabrief,supportivefirstcontact(seeFigure5.2,fourthoval).
Theinvestigatorswouldfirstverifythattheyhadinfactidentified/locatedthecorrectperson,andifso,
theywouldthenintroducethemselvesandexplainthattheywerepartofagroupinDetroitworkingon
oldcriminalcases.Theinvestigatorswouldaskvictimswhethertheyrememberedmakingareportto
thepolice(e.g.,Doyouremembermakingareporttothepolicein2004?)(ratherthanannouncing
thattheyweretherebecauseofareportedsexualassault);theinvestigatorswouldthengoonto
explainthattheyhadadditionalinformationonthecasethatmayallowthecriminaljusticesystemto
proceedwithaninvestigationand/orcriminalcharges(seeAppendixC:VictimNotificationResourcesfor
aFAQdocumentwithsamplelanguageastohowthisinformationcouldbeexplained).Victimswould
thenbeaskediftheywereinterestedinhavingafollowupmeetingwiththeinvestigatorsandan
advocatetodiscussnextsteps.Allvictimswouldbeprovidedwithacomprehensivepacketof
communityresources(seeAppendixC:VictimNotificationResources).
Ifthesurvivorwasinterested,asecondmeetingwouldbescheduled,usuallyatthecommunity
basedadvocacyorganization,foranextendeddiscussionaboutpossiblenextstepsandcommunity
resourcesandservices(seeFigure5.2,fifthoval).Atthatmeeting,theinvestigatorsandadvocatewould
meetwiththesurvivorandanysupportpersonss/hewishedtobepresenttodiscussthetesting
resultsandnextstepsinmoredetail.Attheconclusionofthatdiscussion,survivorswouldbeofferedan
opportunitytomeetprivatelywiththeadvocateinaconfidentialsetting.89Forvictimswhowere
interestedincontinuedcontactwiththecriminaljusticesystem,theadvocateswouldoffertheir
continuedassistanceandinvolvement,ifdesiredbythesurvivors.
89
Akeyreasonwhycommunitybasedadvocateswereselectedtobepartofthissecondmeetingisthattheycouldoffer
survivorsconfidentialcommunication(i.e.,theadvocateswouldnotbeabletodiscloseanythingthesurvivorsaidtothe
investigatorsorprosecutors,withoutwrittenauthorizationfromthesurvivor).
250
Throughouteachstepofthenotificationprotocol,investigatorsandadvocateswouldhavethe
flexibilitytomodifystrategiesasneededtoberesponsivetothespecificneedsofthevictim(seeFigure
5.2,sixthoval).Forexample,ifthevictimwantedmoreinformationandamoreextendeddiscussionat
thefirstcontact,thentheinvestigatorswouldtrytoloopintheadvocateeithertojointhemataspecific
locationortobeoncallifneeded,andthentheywouldcontinuewiththenotificationmeeting.
Challenges Implementing the Pilot Victim Notification Protocol

Figure5.3Challenges&SolutionsEncounteredintheImplementationoftheDetroitSAKARP
VictimNotificationProtocol(followingpages)summarizesthekeychallengesthecollaborativefaced
implementingthisprotocolandthesolutionstheydevelopedtorespondtothoseproblems.Overall,
theprotocolwasimplementedinamannerhighlyconsistentwithitsoriginaldesign(i.e.,highfidelity);
however,thegroupfacedmanypracticalchallengestryingtomanagesomanynotifications,giventhat
thecircumstancesofeachcasewereindeedunique.Thetestingresultshighlightedavarietyof
complexitiesthatneededtobenegotiatedonacasebycasebasis(e.g.,whatiftheCODISprofilewasto
aconsensualpartnerofthevictim,onewhohappenedtobeinCODISforadifferentcrime?)
Coordinatingstaffschedulesforthetwostagenotificationwaschallenging,andthegroupstartedto
developmeetingfatiguewheretheNRTmeetingsstartedtofeellikeonemorething.Intheend,
thegrouptookabreakfromNRTmeetings,andthendecidedtoreinstitutethemwithanewmission
andmorestreamlinedprocess.
251
FIGURE5.3Challenges&SolutionsEncounteredImplementingtheDetroitSAKARPVictim
NotificationProtocol
DISCUSSION.TheNRT
DECISION.Themajorityofcasesbeing
ISSUE1
WHATSTHE
DIFFERENCE
BETWEENTHE
DECISIONTO
NOTIFYANDTHE
DECISIONTO
CONSIDER
PROSECUTION
OFTHECASE?
teamwastaskedwith
decidingwhethertonotify
victimsoftheirSAKtesting.
Thisdecisionwasoften
intertwinedwithwhether
thecasemightbe
prosecuted(e.g.,shoulda
victimbenotifiedifthe
casewasSOLexpiredand
therewasnooptionfor
prosecutionorifthe
perpetratorwasalreadyin
prisonforadifferent
crime?).Theprosecution
decisionwouldbemade
solelybytheprosecutors
office,makingitdifficultto
considernotificationissues
independentof
prosecution.
broughttotheNRTforreviewwere
beingconsideredforprosecution.This
causedsomeconfusionbecausethere
wasnonotificationdecisiontobemade
aboutcasesthatwouldbeprosecuted
becausesuchcasesrequired
notification.
DISCUSSION.Theprotocol
stipulatedthatcases
neededtobereviewedby
theNRTpriortoanyvictim
contact.However,asthe
testingresultscamein,the
prosecutorsofficefeltthat
somecasesmerited
immediateattention.
DECISION.Theprosecutorsofficemade
thedecisiontonotifythevictim,which
proceededintheusualmanner(i.e.,first
contactbylegalinvestigatortosetupa
followupmeetingwiththecommunity
basedadvocates),andtheNRTwas
informedaboutthesecasesattheirnext
meeting.
TheNRTwasnotalwaysabletoadvise
thelegalinvestigatorsdoingthe
notificationsregardinghowbestto
approacheachvictimasplanneddueto
theminimalinformationprovidedbythe
oldpolicereports.
WhiletheNRTmeetingsdidnotunfold
asplanned,theconversationshelped
establishedtrustamongmembersand
forgepositiverelationshipsamongthe
professionalswhowouldbeworking
togetheronthesecases.
ISSUE2
WHATIFACASE
WASTIME
SENSITIVEAND
THEREWASNT
TIMETOHAVEIT
REVIEWEDBY
NRTBEFORE
ACTIONNEEDED
TOBETAKEN?
252
ISSUE3
HOWMANY
VICTIM
NOTIFICATIONS
SHOULDBE
ACTIVEATANY
GIVENPOINTIN
TIME?
ISSUE4
WHATSTHE
BESTWAYTO
PROTECTTHE
SAFETYOF
NOTIFYING
PERSONNEL?
ISSUE5
WHATSTHE
BESTWAYTO
COORDINATE
THESECOND
CONTACT
MEETINGWITH
THEADVOCATE?
DISCUSSION.Thetimingof
NRTmeetings&victim
notificationsneededtobe
consideredinconjunction
withavailabilityof
investigatorsand
prosecutorstopursuethe
case(a400Project
Recommendation).
DECISION.NRTmeetingswerescheduled
whentheprosecutorsoffice
investigatorsandprosecutorswereready
totakeonanewbatchofcases(pending
thevictimswillingnesstoparticipate).As
such,theNRTmeetingswerenot
regularlyscheduled,andwerecalledad
hocwhenstaffwerereadytobeginnew
notifications.
DISCUSSION.Althoughthe
planwasforonlyonelegal
investigatortomakein
personcontact,itwas
oftennecessarytohave
multipleinvestigatorsat
thenotification,inorderto
protectthephysicalsafety
ofthenotifyingstaffand
thevictims.
DECISION.Forsomeinperson
notificationsinsomeneighborhoods,it
wasnecessarytohavemultiple
investigatorstoensurethesafetyofthe
victim&theinvestigatorsthemselves.
Onewouldfocusonthesurvivor,the
other(s)wouldattunetothe
surroundingsandpossiblethreatsto
theirsafety.
DISCUSSION.Coordinating
thesecondcontactwith
communitybased
advocatesbecamemore
challengingastheirtime
wasbeingspentfollowing
upwithvictimswhohad
beenpreviouslynotified.
DECISION5E.Thecommunitybased
advocacyorganizationsworkedwiththe
stategovernmentviolenceagainst
womenagencytodevelopstaffingplans
wherebyagenciescouldoffermultiple
guaranteedblocksoftimeeachweekin
whichadvocateswouldbeavailablefor
thosesecondcontactmeetings.
253
ISSUE6
WHATIFTHE
VICTIMWANTED
EXTENDED
CONTACT&
INFORMATION
ATTHEFIRST
CONTACT
MEETING?
ISSUE7
WHATIFTHE
PERSON
IDENTIFIEDIN
THECODISHIT
WASA
CONSENSUAL
PARTNEROFTHE
VICTIM?
DISCUSSION.Thefirst
shortcontact/second
longercontactnotification
wasplannedtogivevictims
achancetoabsorbthe
shock/surprisetheymay
haveexperiencedandto
preparethemselvesforthe
moredetailed/
complicateddiscussionof
optionsandnextsteps.If
victimswantedmore
informationatthefirst
notification,then
investigatorswould
provideit.However,the
investigatorswereunclear
exactlyhowmuchdetail
theyshouldcoveratthe
firstmeeting,ifthevictim
wantedmoreinformation.
DECISION.Becausetheinvestigators
wereunclearhowmuchinformation
theycouldprovideatthefirstcontact
meeting(andbecausetheyknewthe
overallgoalwastomakesurethatan
advocatecouldbepresentforthe
victim),theywerehesitanttoprovide
extensivedetailstovictimsatthefirst
notification,whichwasfrustratingto
somesurvivors.Whenthisproblemwas
broughttotheattentionoftheARP
collaborative,thegroupdecidedthat
theinvestigationsshouldhaveall
availableinformationonhandandthat
theadvocateswouldestablish
guaranteedblocksoftimetheycouldbe
availabletocomeoutimmediatelyto
joinanotificationmeeting,ifdesiredby
thevictim.
DISCUSSION.The
offendernamedinan
offenderCODIShitmayor
maynotbetheperpetrator
ofthesexualassault.Itis
possiblethattheperson
identifiedinthehitwas
actuallyaconsensual
partnerofthevictim.The
legalinvestigatorsdoing
thenotificationhadto
discussthisissuewith
victimsduringnotification.
DECISION.Thelegalinvestigators
addressedthisissueonacasebycase
basisinthenotifications(i.e.,the
specificcircumstancesofthecase
determinedhowtheydiscussedidentity
andpossibleconsensualpartners).
254
ISSUE8
WHATIFTHE
DNATESTING
REVEALEDA
MIXTUREOF
MULTIPLEDNA
PROFILES?
ISSUE9
HOWSHOULDTHE
COLLABORATIVE
PROTECTTHE
CONFIDENTIALITY
OFTHEVICTIMS
WHOHADBEEN
NOTIFIED?
DISCUSSION.Attimesthe
SAKresultsrevealeda
mixtureofmultipleDNA
sources,someofwhichhad
thepossibilitytobefroma
consensualpartner.The
legalinvestigatorsdoing
thenotificationhadto
discussthisissuewith
somevictimsduringthe
notification,suggesting
takingabuccalswabfrom
saidconsensualpartnersto
preventuploadofthe
individualsDNAprofiles
intoCODIS.
DECISION.Thelegalinvestigators
addressedthisissueonacasebycase
basisinthenotifications(i.e.,the
specificcircumstancesofthecase
determinedhowtheydiscussedidentity
andpossibleconsensual
partners/mixturesamples).
DISCUSSION.Collaborative
bimonthlymeetings
includeddiscussionof
confidentialinformation
regardingvictimsandtheir
notificationprocess.
Membersofthegroupnot
directlyinvolvedwith
victimswerepresentfor
theseupdates
DECISION.Allmembersofthe
collaborativeteamhadsigned
confidentialityagreementstonot
discloseinformationexchangedduring
teammeetings,butitwasagreedupon
thatvictimnotificationupdatesshould
belimitedtothosedirectlyinvolved
withvictims.
Thebimonthlymeetingswouldbesplit
intotwoparts;thesecondhalfofthe
meetingwouldbeattendedonlyby
thosewhoweregrantedaccesstothe
privilegedvictimnotification
information.Thesesessionswereheld
inperson;teammemberswerenot
allowedtojoinbyphonefor
confidentialitypurposes.
255
ISSUE10
DISCUSSION.Thedifferent
disciplinesinvolvedinthe
victimnotificationprocess
haddifferingrules
regardingthedisclosureof
confidentialinformation.
Thecommunitybased
advocatescouldNOTshare
informationwiththelegal
investigatorsaboutthe
notifications,which
frustratedthe
investigators.
HOWCAN
MEMBERSOFTHE
COLLABORATIVE
SHARE
INFORMATION
ABOUTTHE
NOTIFICATIONS,
WITHOUT
VIOLATING
CONFIDENTIALITY?
DECISION.Investigatorswantedtoknow
iftheirnotificationstrategieswere
effective,butadvocatescouldnot
disclosesuchinformation.Ahighly
experiencedadvocatewhoworkedfor
theStategovernmentviolenceagainst
womenagencyjoinedtheNRTmeetings
toprovidegeneralguidanceto
investigators.Thisallowedlocal
advocatestoupholdtheirconfidentiality
policiesandinvestigatorstoaskgeneral
questionstothestateleveladvocate.
ISSUE11
HOWCANTHE
MEMBERSOFTHE
COLLABORATIVE
JUGGLEMULTIPLE
TEAMMEETINGS
ARETHENRT
MEETINGSREALLY
NECESSARY?
DISCUSSION.Allofthe
NRTteammembersalso
regularlyattendedthebi
monthlyARPcollaborative
teammeetings,andsome
individualsexpressed
frustrationaboutthe
amountoftimethatwas
beingspentinmeetings
aboutthecases,whichleft
lesstimetoworkonthe
casesthemselves.
DECISION.Thisissuewasraisedright
beforethethirdNRTmeeting,andwhile
thethirdmeetingwasheld,theARP
decidedtoterminateallfutureNRT
meetings.Theydecidedtocontinueto
followthesameprotocol(i.e.,legal
investigatorscoordinatingnotifications
withcommunitybasedadvocates),but
therewouldbenomore
multidisciplinarymeetingstoreview
casesandprovideinputonnotification
decisions.Thedecisionif/whytonotify
avictimwouldbemadebythe
prosecutorsoffice.
However,aftertheevaluationresults
werepresented(seenextsection),
whichrevealedthatmultidisciplinary
collaborationandcommunicationwere
important,thegroupreconsideredthis
decisionanddecidedtoreenvisionthe
NRTmeetingsandcontinuethemona
monthlybasistocheckinoncurrent
casesandpendingnotifications.
256
Evaluating the Pilot Victim Notification Protocol

Developing the Evaluation Design
Theresearch/evaluationteamfacedmanychallengesdevelopinganevaluationplan,notthe
leastofwhichwasthatthevictimnotificationprotocolwasnotstandardizedbydesignandexplicitly
permittedcasebycaseflexibility.Anonstatic,changingevaluandisnotuncommoninevaluation,
particularlyincomplex,developmentalprojects(seePatton,2011),andanevaluatorcanstillmake
reasonableinferencesaboutaprogramanditsefficacy,providedthatthereareclearguidingprinciples,
processes,andanticipatedoutcomes(andtherewereforthisprotocol)(seeAppendixB:Project
Methodologyforextendeddiscussionondevelopmentalevaluationtheoryandpractice).
Itwasnotpossiblefortheevaluatorstobepresenttoobservethenotificationsandcollectreal
time,prospectivedata.AsnotedinFigure5.1,therewereconcernsaboutprotectingthephysicalsafety
oftheinvestigators(giventheratesofviolentcrimeinmanyDetroitneighborhoods),somuchsothat
theadvocatesdidnotattendthefirstcontact;thesesameconcernsextendedtotheresearchteamas
well.However,evenifitwerepossiblefortheevaluatorstoobservethenotification,wewouldnot
have,asdoingsowouldhaveviolatedthesurvivorsprivacy.Thefocusofthenotificationneededtobe
onthesurvivorsandtheirwellbeingnotdatacollection.Itmighthavebeenpossibletoascertain
victimswillingnesstohavearesearcherattendthesecondnotificationmeeting,butthecollaborative
partnersandtheresearch/evaluationteamconcurredthatdoingsowasinadvisable,asagainitmight
haveshiftedthefocusfromtheneedsofthesurvivors.Therefore,datacollectionwouldneedtobe
retrospectiveaccountsfromtheinvestigators,advocates,andperhapsthevictimsaswell.
Wedevelopeddatatrackingtoolsfortheinvestigatorstorecordthespecificstepsandstrategies
theyusedtolocatevictimsandtheirperceptionsofwhathappenedinthenotificationmeetings(see
AppendixD:DataCollectionInstruments).Ideally,wewouldhavelikedtohavehadcomparabledata
collectionfromtheadvocates,linkedtotheinvestigatorsdatasowecouldtriangulateinformation
257
regardingwhathappenedinthenotificationmeetings.However,becauseadvocatesurvivor
communicationisconfidential,itwasnotpossibletocollectdatainthatmanner,asitwouldhavebeen
possibletoidentifywhichvictimstheadvocateshadworkedwithandconnecttheiridentitiestotheir
data.Toaddressthischallenge,theresearch/evaluationteammetwiththeadvocacyorganizationsand
theirstatefunderstoexploreoptionsforcollectingdatafromtheadvocatesinwaysthatwouldnot
violateconfidentiality.Weagreedthattheadvocateswouldrecordtheirperceptionsofwhathappened
ineachcasewheretheyhadbeeninvolvedinvictimnotification,butthatthevictimsnamesandother
identifyinginformationaboutthecasewouldnotbeincluded(seeAppendixD:DataCollection
Instruments);furthermore,theadvocateswouldprovidethedatatotheresearchteaminonebatchof
deidentifiedcasessothatitwouldnotbepossiblefortheresearchteamtoinfer(basedonwhichcases
theinvestigatorshadrecentlycompleted)whichvictimscorrespondedtowhichrecords.Assuch,itis
notpossibletocomparetheaccountsoftheinvestigatorsandadvocatesregardingwhathappenedin
anyindividualnotification,butthismethodologicallimitationcouldnotbeavoidedgiventhe
confidentialityrequirementsoftheorganizationsinvolvedinthisproject.
Wewantedtointerviewvictimsaftertheirnotificationstoassesstheirperceptionsandsolicit
theirfeedbackonhowtoimprovethenotificationprotocol.However,theprosecutorsofficestaffhad
concernsthattheevaluatorscouldpossiblybecalledaswitnessesinthecase(ifwehadcontactwith
victimspreadjudication).Althoughtheevaluatorscouldnottestifyastothecontentoftheirwork,the
processofexplainingthatrefusaltoajudge/jurywouldcomplicatematters.Therefore,theElected
Prosecutordecidedthattheresearcherscouldnothavecontactwithvictimsuntilpostadjudication.At
thetimethisevaluationwasbeingconducted,therewereonlytwocasesinwhichthevictimshadbeen
notifiedandtheadjudicationwascomplete(becauseinthoseinstances,thecaseswouldnotbe
prosecuted);allothernotificationswerestillintheprocessofinvestigationandpossibleprosecution.As
such,datacollectionfromsurvivorswasnotpossibleinthecontextofthisevaluationanditstimeline.
258
The Decision to Notify: Selecting Cases for Victim Notification

DuringtheoneyearevaluationperiodofNovember,2012toDecember,2013,therewerethree
multidisciplinarynotificationreviewteam(NRT)meetings.ThefirstmeetingtookplaceinNovember
2012.Sixpractitionersattendedthemeeting,representingfourorganizations:theprosecutorsoffice,
thepolicedepartment,theSANE/communitybasedadvocacyprogram,andanationaladvocacy
organizationpartner;fourmembersoftheresearch/evaluationteamobservedthemeeting.Sevencases
werepresentedfordiscussion:4casesfromTestingGroup1(StrangerRapeCases)and3casesfrom
TestingGroup2(NonStrangerRapeCases).90ThesecondNRTmeetingtookplaceinApril,2013with
ninepeopleinattendance,representingthefiveorganizations(thefourlistedaboveforthefirstNRT
meeting,plusthepolicedepartmentsvictimadvocacyprogram);onememberofthe
research/evaluationteamobservedthemeeting.Eighteencaseswerepresentedfordiscussion:13
casesfromTestingGroup1(strangerperpetrated);5casesfromTestingGroup2(nonstranger
perpetrated).ThethirdandfinalNRTmeetingthattookplaceduringtheevaluationperiodwasin
August,2013.Sixpeopleattendedthemeeting,representingfourorganizations:theprosecutorsoffice,
thepolicedepartment,theSANE/communitybasedadvocacyprogram,andastategovernment
violenceagainstwomenagency;91twomembersoftheresearch/evaluationteamobservedthemeeting.
Fifteencaseswerepresentedfordiscussion:11casesfromTestingGroup1(strangerperpetrated)and4
casesfromTestingGroup2(nonstrangerperpetrated).
90
MorecasesfromTestingGroup1(StrangerRapeCases)werepresentedforreviewatthisNRTmeeting(andsubsequentNRT
meetings)becausethelabvendorprocessingtheGroup1kitsfinishedtestingandreportedtheirresultsbackbeforethelab
vendorprocessingGroup2kits.Thus,thehighernumberofstrangerrapecases(relativetononstrangercases)reviewedinthis
evaluationisanartifactofwhentestingresultsbecameavailable(intheoveralltimelineoftheproject)andshouldnotbe
interpretedtoreflectdifferentialemphasis,concern,orperceivedseriousnessofthesetypesofcasesrelativetoothers.
91
AsnotedinFigure5.1,thecommunitybasedadvocatescouldnotdiscusstheircaseswithoutviolatingclientconfidentiality;
therefore,thecollaborativeinvitedahighlyexperiencedadvocatewhoworkedfortheStategovernmentviolenceagainst
womenagencytojointhethirdNRTmeeting.Thisadvocatewasnotdirectlyinvolvedinanyofthecases,butcouldprovide
generalguidancetoinvestigatorsregardingvictimnotification.Thiswouldallowthelocaladvocatestoupholdtheir
confidentialitypoliciesandtheinvestigatorswouldbeabletoaskgeneralquestionstothestateleveladvocate.
259
Atotalof40cases(allofwhichhadCODIShits,seeFigure5.2,secondoval)werepresentedby
theprosecutorsofficetothemultidisciplinaryNRT.TherewereonlythreevictimstheNRTdecidednot
tonotifyallthreewerefromTestingGroup2(NonStrangerRapeCases)andallweredomestic
violencerelatedintimatepartnersexualassaults.Inonecase,thestatuteoflimitationshadexpired;in
thesecondcase,thevictimhadnotappearedincourtwheninitialchargeshadbeenpursuedyearsago;
inthethirdcase,thevictimhadnotappearedforawarrantinginterview(twice)attheprosecutors
officewhentheinitialchargeswerebeingconsideredyearsago.Inthelattertwocases,NRTmembers
felttherewereclearindicationsinthecaserecordsthatthesurvivorshadnotwantedtopursue
prosecutionyearsago,andgivenspecificdetailsinthecaserecords,recontactingthemnowcould
compromisetheirsafety.
Fourcaseswereselectedfornotificationatthediscretionoftheprosecutorsoffice(i.e.,these
caseswerenotbroughtbeforeNRTforreview;seeFigure5.2,thesixthovalregardingcasebycase
flexibility).Inallfourcases,theprosecutorsfeltimmediatenotificationwasnecessarybasedonthe
CODIShitinformationandoriginalcasefilematerials(e.g.,theCODIShitwastoanoffendercurrentlyin
prison,butabouttobereleasedonparole).Inthesefourinstances,therestofthevictimnotification
protocolwasfollowed(e.g.,loopingincommunitybasedadvocacy),pertheestablishedprotocol(see
Figure5.2).ThesefourcaseswerelaterpresentedtoNRT(afterthefact)sothatallmembersofthe
teamwereawarewhathadhappened;inallfourcases,theNRTunanimouslyagreedthatthevictims
shouldhavebeennotifiedandthatemergencyactionhadbeenwarranted.
Locating Victims: Strategies and Success Rates

Overall,ofthe41casesselectedfornotification(37casesNRTdecidedtonotify+4cases
selectedbyprosecutorsofficefornotification/notreviewedbyNRT),theinvestigatorswereableto
locate31victims(8notificationswerestillpendingatthetimetheevaluationconcluded).Ofthe33
260
caseswithcompletedata,31hadbeenlocatedandnotified(95%findrate).Figure5.4(below)shows
thenumberofcasesthatwereselectedfornotificationinwhichvictimsultimatelywerelocated.
FIGURE5.4NumberofCasesSelectedforNotificationinwhichtheVictimsWereUltimatelyLocated
4Cases
Selectedfor
Notificationby
ProsecutorsOffice
40Cases
ReviewedbyNRT
3Cases
DecidedNottoNotify
37Cases
DecidedtoNotify
8Cases
Pending
2Cases
Unable
toFind
4Victims
Notified
27Victims
Notified
31(TOTAL)
VICTIMS
NOTIFIED
261
Tolocatevictims,investigatorsbeganbysearchingthreeprimarydatabasestotrytofindthe
victimscurrentresidence:
LEIN(LawEnforcementInformationNetwork)database=restrictedusedatabaseforpolice
officers;containscriminalrecords,drivingrecords,vehicleregistrations,andwantedpersons.
LEINinterfaceswiththeNationalCrimeInformationCenter(NCIC)toprovideinformationabout
missingpersons;
TLOOnlineInvestigativeServicesdatabase=privatedatabaseavailablebysubscription;
compilespublicandprivatesourcesforindividualslastknownaddresses,phonenumbers,
possiblerelatives,bankruptcies,andsocialsecuritynumbers;
NEXIS/LEXISdatabase=privatedatabaseavailablebysubscription;compileslegalandpublic
records,includinglocal,national,andinternationalnewspapers,magazines,tradejournals,wire
andsocialmediasources.
Atleastoneofthesedatabaseswassearchedforall31victimswhowerenotified(typically,two
databasesearcheswereperformedpervictim,thoughsometimesallthree).TheLEINdatabasewasthe
primarygotoresource(searchedinn=25cases,yieldingactionableinformation88%ofthetime).
Afterobtainingaddresses/phonenumbersforthevictimspossiblewhereabouts,the
investigatorsinitiatedcontacttoverifywhethertheyhadinfactfoundtherightperson.Onaverage,
investigatorsmadethreecontactattempts(range=116contacts)oneachcase,leadinguptoand
includingthecontactinwhichtheyconfirmedtheyhadfoundthecorrectindividual.Investigators
averagedtwoinpersonvisitspercase(range=06attempts).Sometimesinvestigatorsweresuccessful
locatingvictimsonthefirstattempt,buttypically,investigatorshadtoeitherreturntothesamehome
onmultipleoccasionsbeforetheywereabletospeakwiththevictimorvisitaseriesofaddresses,ruling
themoutonebyone.Typically,investigatorsmadeaninpersonvisittoalastknownaddressandlefta
262
businesscardiftheywerenotabletospeakwithanyoneatthelocation.Theinvestigatorswerethen
abletoconfirmtheidentityofthevictimifandwhens/hecalledback.Investigatorsaveragedone
phonecallpercase(range=012calls).Forvictimswhosemostcurrentaddresseswereoutsidethe
Detroitmetroareaoroutofstate,phonecontactwasthesolestrategyused.
Therewasconsiderablevariabilityinthenumberofinpersoncontactattemptsandphone
contactattemptsthatwerenecessarybeforetheinvestigatorsweresuccessfulconnectingwiththe
victim(06and012,respectively).Tounderstandthisvariability,wereexaminedthedatatodetermine
whethertherewerediscerniblepatternsintheeffortrequiredtolocatesurvivors.AsshowninFigure
5.5(nextpage),inmostcases,thevictimswereabletobelocatedwithrelativelylowinvestigational
effort:65%werelocatedbyconductingdatabasesearches,plus04phonecalls,plus01inpersonvisit
tooneaddress.Theremaining35%ofcasesrequiredmoreextensiveinvestigationalefforttolocateand
notifythevictims:16%requiredanadditional12inpersonvisits(to12addresses);and16%required
extensiveefforts(612phonecallsand46inpersonvisitstomultipleaddresses).92
Itisnoteworthythatthemajorityofvictimscouldbefoundwithrelativelyloweffortbecause
communitiesstrugglingwithlargenumbersofunsubmittedSAKsmaybeconcernedaboutthelabor
expenditureoflocatingandnotifyingvictims.ThesedatafromDetroitsuggestthatmostvictimscanbe
locatedwithouttremendousinvestigativeeffort(i.e.,deskworkofdatabasesearchesand04phone
calls,andthen01inpersonvisitstooneaddress).Furthermore,inthisproject,theaveragelengthof
timebetweenwhentheassaultoccurredandthetimeofnotificationwasnineyears;therefore,this
evaluationsuggeststhatitispossibletoreconnectwithvictimsafterprolongedperiodsoftimeand
doingsowillnot(inmostcases)requireextensiveinvestigativeeffort.
92
Jumpingaheadtothelaterstagesofthevictimnotificationprotocol,thefivevictimsinthehighinvestigationaleffort
groupingwerenomoreorlesslikelythanthesurvivorsintheothereffortgroupingstoagreetoparticipateintheprosecution
oftheircases(3yes,2no),indicatingthattheadditionaleffortexpendedtofindthesesurvivorswasusefulinthatmostwanted
toengageinfurtheractionontheircases.
263
FIGURE5.5EffortRequiredtoLocateVictimsforNotification
65%(n=20)REQUIREDRELATIVELYLOWEFFORT
DatabaseSearches
+
04PhoneCallsand/or01InPersonVisitto1Address
Forexample
Databaserevealed
possiblelocaladdress
Visitedaddress;
locatedvictim
19%(n=6)REQUIREDMODERATEEFFORT
DatabaseSearches
+
01PhoneCallsand/or23InPersonVisitsto12Addresses
Forexample
Database
revealed
possibleaddress
Visitedaddress;
victimnolonger
livedthere
Victimcalled
investigators
Calledvictimsparole
officer;receivednew
possibleaddress
Visitedaddress;spokewithvictims
mother&leftbusinesscard
16%(n=5)REQUIREDHIGHEFFORT
DatabaseSearches
+
012PhoneCallsand/or46InPersonVisitstoMultipleAddresses
Forexample
Databaserevealedseveral
possibleaddressesand
phonenumbers
Visited4th
address;
locatedvictim
Visited1st
address;
vacanthome
Called12differentphone
numbers;nonereachedthe
victimandmanydisconnected
Visited2nd
address;victimno
longerlivedthere
Visited3rd
address;victimno
longerlivedthere
264
The First Notification Contact: Victims Emotional Reactions to the Investigators

IntheDetroitvictimnotificationprotocol,thefirstcontactwiththevictim(Figure5.2,fourth
oval)wouldideallyoccurinperson(orbyphone)sothattheinvestigatorcouldberesponsivetothe
victimsquestionsandemotionalreactions.Theprimarygoalsofthisfirstcontactmeetingwereto
explainthattheSAKhadnotbeentestedatthetimethevictimhadfiledapolicereport;giveanapology
tothevictimforthefactthatithadnotbeentested;explainthatithadnowbeentested(andingeneral
terms,describethetestingresults);andofferthevictimasecond,followupmeetingtodiscussthe
informationinmoredetailwiththeinvestigatorsandacommunitybasedvictimadvocate.
Ofthe31notificationsthatoccurredduringtheevaluationperiod,mostofthefirstcontact
meetingstookplaceatthevictimscurrentresidence(n=21).Inthreeinstances,theinvestigatorsmade
preliminaryphonecontactwiththevictim,whothenrequestedthattheytabletheconversationand
meetinperson,atalocationofthesurvivorschoosing(oneaskedtomeetatherplaceofemployment,
oneaskedtomeetatanearbyrestaurant,oneaskedtomeetatthecommunitybasedadvocacy
organization).Inseveninstances,theinitialnotificationoccurredbyphone(twoofthesesevenphone
contactswereincasesinwhichthevictimnolongerlivedintheDetroitmetroarea).Mostofthese
initialnotificationswereconductedbytwoinvestigators(n=23,74%)andtheconversationslastedan
averageof18minutes(rangedfrom545minutes).
Theinvestigatorsrecordedtheirperceptionsregardingwhathappenedduringthefirstcontact
victimnotificationmeetings.Specifically,weaskedthemtodescribehowvictimsreactedwhenthey
weretoldthattheirkithadnotbeentestedyearsago,thatithadrecentlybeenfoundandtested,and
thatthetestinghadyieldedaCODIShit.Althoughwedonotknowthesurvivorsinternalthoughtsand
feelings(aswecouldnotinterviewsurvivorsuntilpostadjudication),theirexternalreactions,as
witnessedbytheinvestigators,followedoneofthreegeneralpatterns:
265
Strongnegativeemotionalreaction(n=5)Thesevictimsdidnotwanttodiscusstheircasewiththe
lawenforcementinvestigators.Someofthesevictimswerehesitantandsuspiciousoftheinvestigators
intentionincontactingthemandinquiredastowhythecasewasbeinglookedintoaftersomanyyears.
Othersexplainedthattheynowhadotherconcernsintheirlife(e.g.,healthconcerns)thatsuperseded
theirassaultfromsomanyyearsago(e.g.,theyhadmovedon,ordidntcareaboutitanymore,or
consideredthenotificationahassle).Onesurvivorterminatedthediscussionabruptlybecauses/he
becameveryupsetandwasworrieds/wouldgetsicktohis/herstomach.93
Strongpositiveemotionalreaction(n=9)Thesevictimswerehappyandexcitedtobenotifiedthat
theircasewasnowbeinginvestigatedandthattheymighthavetheopportunitytotestifyagainsttheir
assailant.Thatsaid,thesevictimsdidnotalwaysinitiallyreactpositivelytothelawenforcement
investigators.Severalvictimswerenervouswhentheinvestigatorsarrived,astheythoughttheywerein
troubleorweregoingtobearrested,buttheirreactionchangeddramaticallywhentheylearnedthat
theiroldsexualassaultcasewasbeingreinvestigated.Manyvictimscriedbecausetheyweresohappy
thatsomethingwasfinallybeingdonewiththeircase;thatitwill[finally]moveforwardaftereight
years.Onevictimstatedshe/hesawthenewsonthekitsandthoughthers/hiswasthrownout.After
beingnotifiedthatthekithadnotbeenthrownout,andthattheperpetratorwasserving40yearsin
prison(foradifferentoffense),thevictimwasgladhewaslockedupandcanthurtanyoneelse.94
Absenceofastrongnegativeorstrongpositiveemotionalreaction(n=17)Thesevictimsdisplayed
someemotionduringthenotification,buttheydidnotexhibiteitherastrongnegativeorstrongpositive
response.Manywerematteroffactintalkingwiththeinvestigators,thoughsomeshowedsomesigns
ofemotionaldistress(e.g.,briefepisodesofcrying).
93
Thequotesherearefromtheinvestigatorsnotesregardingwhatvictimssaidatthenotification.
94
Asthisexampleillustrates,theNRTdecidedtonotifyvictimswhoseassailantswerealreadyinprisonforotheroffensesin
hopethatsuchnewswouldgivethemsomepeaceofmind.
266
Giventhesedifferentreactions,weexaminedwhethersomevictimswereparticularlylikelyto
havepositiveornegativereactions(ortheabsenceofstrongpositive/negativereactions).Duetothe
smallsamplesizeinthisevaluation,quantitativeanalyseswerenotfeasible,butfromaqualitative
perspective,datavisualizationmatricescanbeausefulstrategyforidentifyingpatterns(seeMilesetal.,
2014).Tothatend,weexaminedhowvictimsreactionsvariedasafunctionofthreevariables:
TheamountoftimethathadpassedsincetheassaulttothetimeofnotificationIngeneral,sexual
assaultvictimsdistresslevelsdecreaseovertime(seeFrazieretal.,2004;Koss&Figueredo,2004;
Steenkamp,Dickstein,SaltersPedneault,Hofman,&Litz,2012),but,priorresearchhasnotexamined
howsurvivorsreacttoanunexpectedreactivationofthetraumaticmemory.Therefore,weexamined
therelationshipbetweentimesinceassaultandvictimsnotificationreactions.Inthissample,the
averagelengthoftimefromtheassaulttothenotificationwasnineyears,fromwhichweformedthree
analysisgroups:1)lessthannineyears;2)nineyearsago;and3)morethannineyearsago.
VictimageatthetimeoftheassaultPriorresearchhassubstantiatedthatchild/adolescentvictims
differfromadultsinpostassaultmentalhealthsequelae,copingbehaviors,andhelpseeking(see
Campbell,Greeson,&FehlerCabral,2013;FehlerCabral&Campbell,2013;Finkelhor,Wolak,&Berliner,
2001;Fryetal.,2013;Jonesetat.,2003;Martin,Houston,Mmari,&Decker,2012;Zinzowetal.,2012).
Furthermore,previousstudieshavefoundthatcriminaljusticesystempersonneloftendoubtthe
credibilityofyoungervictims(seeCampbelletal.,2013;Campbelletal.,2012;Shaw&Campbell,2013)
(aswasreplicatedinthisproject,seeChapter3:WhySoManyUnsubmittedSAKsinDetroit).Therefore,
wewantedtoexplorewhethervictimsreactionstonotificationvariedasafunctionoftheirageatthe
timeoftheassault.Fortheseanalyses,victimagewasdividedintothreecategories:1)under16yearsold
(Michiganlawdefinesageofconsentat16);2)between16and24yearsold;and3)over24yearsold.95
95
Theageatassaultforthese31victimstendedtobeyoung(consistentwithgeneralfindingsofsample),withnaturally
occurringgroupingsformingatage16and24.
267
ThevictimoffenderrelationshipAlthoughthepsychologicalliteraturedoesnotfindmarked
differencesinthetraumaticimpactofstrangerperpetratedsexualassaultscomparedtothosecommitted
bysomeoneknowntothevictim(i.e.,botharehighlytraumatic)(Campbell,Dworkin,&Cabral,2009;
Domino,2012;Ullman,2010),criminaljusticesystempersonneltendtoresponddifferentlytovictimsof
strangerrape(e.g.,strangerrapevictimsencounterlesssecondaryvictimization,lessvictimblaming,are
providedmoreassistance)(Campbell,2008;Campbelletal.,2009;Frazier&Haney,1996;Kerstetter,
1990;Spohn&Spears,1996).Giventhislongestablishedfindingintheliterature,wewantedtoexamine
ifvictimoffenderrelationshipwasassociatedwithvictimsnotificationreactions.Fortheseanalyses,we
comparedstrangerrapevictimstothosewhoknewtheirassailants(bysight,friends/acquaintances,
intimatepartners,formerintimatepartners).
Figure5.6(nextpage)depictstherelationshipsbetweenthesethreevariables(timesince
assaulttonotification,victimageattimeofassault,andvictimoffenderrelationship)andsurvivors
reactionstothenotifications.InFigure5.6,therowsrepresentthetimesincetheassault;thecolumns,
victimsageattimeoftheassault;andthecelldivisions,victimoffenderrelationship.Thesymbols
insidethecellsrepresentthevictimsemotionalreactions:stronglypositive(greenplussign),strongly
negative(redstrikethroughsign),absenceofastrongreaction(yellowopencircle).Thenumberof
symbolsineachcellreflectshowmanyvictimshadthatpatternofassociations(e.g.,intheuppermost
leftcell,therewasonlyonesurvivorwhowasassaultedwhens/hewaslessthan16andwasnotified
lessthan9yearsaftertheassault;thisindividualhadbeensexuallyassaultedbyastranger,ands/he
exhibitedapositivereactiontothenotification).UsingMiles,Huberman,andSaldanas(2014)methods
forcreatingandanalyzingdatadisplaymatricesforqualitativedata,ourapproachhereisdescriptive
innature,identifyingpossibleassociationwithinthedatainahypothesisgenerationapproachthatcan
informfutureresearchwithlarger,morerepresentativesamplesthanwhatwehadinthisevaluation.
268
FIGURE5.6VictimsEmotionalReactionstotheFirstContactNotifications
Under16YearsOld
(attimeofassault)
0%
Less
Than9
Years
Since
Assault
0%
1624YearsOld
(attimeofassault)
23%
Known
Known
9Years
Since
Assault
StrangerN/A
StrangerN/A
More
Than9
Years
Since
Assault
29%
Known
N/A
8%
Stranger0%
Known
N/A
N/A
0%
Known
Known
N/A
Stranger0%
Over24YearsOld
(attimeofassault)
13%
Stranger0%
Known
Stranger11%
Known
Known
0%
33%
0%
StrangerN/A
N/A
Stranger40%
Stranger25%
Legend
Known:14%
Stranger:16%
Victimexhibitedastrongpositiveemotionalreaction
Victimdidnotexhibitastrongpositiveorstrongnegativeemotionalreaction
Victimexhibitedastrongnegativeemotionalreaction
XX%Percentofvictimswhoexhibitedastrongnegativeemotionalreaction
269
Figure5.6revealsseveralinterestingpatternsinvictimsemotionalreactions.Perhapsnot
surprisingly,howvictimsreactedwasassociatedwithhowlongagotheassaultoccurred:29%ofvictims
whowerenotifiedover9yearsaftertheassaulthadastrongnegativereactiontothenotification;by
comparison,only8%ofthevictimsnotifiedatthenineyearpointand0%ofthevictimsnotifiedless
than9yearsaftertheassaulthadstrongnegativereactions.Thelongertheperiodoftimebetween
whentheassaultoccurredandwhenthekitwastestedandthevictimwasnotified,themorelikelya
survivorwouldhaveanegativereactiontothenotification.Astheinvestigatorsnoted,manyofthe
survivorswhowereangryatthenotificationexpressedthathadmovedonwiththeirlives.
Victimsemotionalreactionsmayhavevariedtosomeextentbyhowoldtheywereatthetime
oftheassault:strongnegativereactionsweresomewhatmoretypicalamongwomenwhohadbeen16
24atthetimeoftheassault(23%,comparedto13%ofvictimsover24yearsoldand0%ofvictims
under16).Priorresearchsuggeststhatlateadolescentsmaybeparticularlylikelytohavenegative,
victimblamingexperienceswithlawenforcementpersonnel(Campbelletal.,2012),andtheresults
previouslypresentedinChapter3:WhySoManyUnsubmittedSAKsinDetroitindicatedthatyounger
victims(adolescents)wereoftenaccusedbythepoliceofmakingfalsereports.Thereforeitseems
possiblethatvictimsnotificationreactionscouldberelatedtohowtheyhadbeentreatedinitiallyatthe
timeofthereport;however,giventhenatureofthedatawehaveinthisevaluation,wecannotmake
definitiveconclusionsaboutsuchanassociation.
Finally,Figure5.6suggeststhatvictimsemotionalreactionstothenotificationdidnotvaryasa
functionofvictimoffenderrelationship.Thoseassaultedbyknownperpetratorsdidnothavemarkedly
differentemotionalreactionstothenotificationasthosewhohadbeenrapedbyastranger(14%of
knownperpetratorshadstrongnegativereactionsvs.16%forstrangerperpetrators).However,only
sevenofthenotificationsweremadeknownperpetratorvictims,sotheremaynotbeenoughcasesin
thisevaluationtodetectapattern.
270
The Second Notification Contact: Victims Decisions Regarding Further Involvement with
the Criminal Justice System
Inthefirstcontactmeeting,theinvestigatorsaskedvictimsiftheywereinterestinginhavinga
followupsecondmeetingtodiscussthecaseinmoredetail,reviewoptions,andmeetwitha
communitybasedadvocate(seeFigure5.2,fifthoval).Inthreecases,itbecameclearatthefirstcontact
thattherewasnoneedforasecondfollowupmeeting:inonecase,thestatuteoflimitationshad
alreadyexpiredandthenotificationwasdonesimplytoinformthevictimaboutthetestingresults;in
thesecondcase,theCODIShitturnedouttobeamatchtothevictimsconsensualpartner(notthe
assailant),sotherewasnolegalcase(asofyet)tobepursued;inthethirdcase,thevictimwasina
grouphomeforindividualswithseriousmentalandphysicaldisabilities,andshehadnomemoryofthe
assault(andsonolegalcasecouldbeconsidered).Inallthreeinstances,theinvestigatorsprovided
victimswiththeircontactinformation(iftheyhadanyfollowupquestionsorconcerns)andabookletof
communityresources(seeAppendixC:VictimNotificationResources).Giventhecircumstancesofthese
threecases,thesamplesizefortheevaluationshiftedfromN=31toN=28(i.e.,therewereonly28
casesinwhichasecondcontactwaspossibleandlegalactioncouldbeconsidered).
Eighteenofthe28victimswhohadfirstcontactnotifications(andtherewasreasontohavea
followupmeeting)hadasecondarymeeting(64%)(seeFigure5.2,fifthoval).96Itisdifficulttoevaluate
whetherthe64%returnrate(i.e.,victimswantingafollowupmeeting)isgoodbecausethereareno
otherstudiesintheliteratureofthissort.Typically,ifresearchers(orserviceproviders)aretryingtore
connectwithvictims,itisshortlyaftertheassault(<1week),usuallywiththegoalofrecruitingvictims
intostudiesorintotreatmentprograms(i.e.,therewasanincentiveforagreeingtofurthercontact)
(Campbell,Sprague,Cottrill,&Sullivan,2011).ThetwostageapproachusedintheDetroitvictim
96
Thecountof18isbasedonthedatabaseusedbythecommunitybasedadvocatestodocumenttheirinteractionswith
victims.Asnotedpreviously,thesedatacannotbelinkedtospecificvictimsinordertomaintainvictimadvocateconfidentiality.
However,18entriesweremadeinthisdatabase,indicatingthatatleast18ofthe28victimsinitiallycontacteddidhavea
followupmeetingwiththeinvestigatorsandadvocate.
271
notificationprotocolissomewhatanalogous,inthesensethatthegoalwastoscheduleasecond
contactquicklybutquitedifferentinthatthelengthoftimesincetheassaultwasmuchlonger(andit
isdebatablewhetherparticipationinthelongprocessofprosecutionisanincentive;seeKonradi,2007).
Campbelletal.s(2011)reviewofstudiesthathavetriedtoreconnectwithrapesurvivorsshortlyafter
anassaultfoundhighlyvariablesratesofsuccess,rangingfrom11%to68%ofvictimsengagingina
secondcontact.Withthat,albeitlimited,basisofcomparison,the64%rateobtainedinthisprojectis
quitehigh,particularlyinlightofthefactthattheassaultoccurredonaveragenineyearsago.
Atthesecond,followupmeeting,theinvestigator(s)andadvocateexplainedthenextstepsin
theinvestigation(e.g.,followuponmedicalrecords,stepstobetakentolocatetheperpetrator,the
overallcriminaljusticesystemprocess).Then,dependingonthespecificcase,thevictimwasaskedto
confirmhis/herstoryaswrittenintheoriginalpolicereport,wasprovidedwithaphotolineupto
identifytheperpetrator,and/orwasaskedforpermissiontoretrievemedicalrecords.Theactionstaken
ineachnotificationvarieddependingonthespecificdetailsofthecaseandtheappropriatenextsteps.
Theinvestigatorsandadvocatesalsodocumentedwhatvictimssaidregardingtheirinterestandintent
toparticipatefurtherinthepossibleprosecutionoftheircases.Theinvestigatorsdidnotpressfora
youmustdecidenowifyouwanttoprosecutedetermination,buttheydidassessvictimswillingness
forcontinuedparticipationintheinvestigationprocess.
Overall,overall16the28victimsdecidedthattheydidwanttohavecontinuedparticipationin
theinvestigationandpossibleprosecutionoftheircases(57%).Again,giventhattherearenoprior
studiesintheliteratureonvictimnotificationinpreviouslyunsubmittedSAKs,itisdifficulttoevaluate
whetherthisrateof57%isgood.Inefforttoprovidesomecontext,wesearchedtheliteraturefor
studiesthathaveexaminedcrimevictimsengagementinthejusticesystemafterprolongedperiodsof
inactivityinthecase;wedidnotfindanysuchresearch(forthecrimeofsexualassaultspecificallyorfor
anytypeofcrime).Therefore,weexaminedtheliteratureonsexualassaultvictimsexperienceswith
272
lawenforcementtoseeifthatbodyofworkmightprovidesomeinsightintowhatvictimsmightdecide
regardingreengagement.Campbells(2008)reviewofthatliteraturefoundthatmostvictims(80%+)
experiencesecondaryvictimizationintheirinteractionswithpolice(i.e.,negative,victimblaming
treatment),andtheseencountersleavesurvivorsfeelingdepressed,anxious,andreluctanttoseekany
furtherhelp.Ifvictimsdonotwanttoseekfurtherhelp,thenitmaybereasonabletoinferthatthey
wouldnotwantfurthercontactwiththecriminaljusticesystem.Assuch,thesedatawouldsuggestthat
arelativelylowpercentage(perhaps20%)wouldbewillingtoreengage.Giventhepervasivenessof
secondaryvictimizationdocumentedinthepolicereportsreviewedforthisproject(seeChapter3:Why
SoManyUnsubmittedSAKsinDetroit),itseemslikelythatmanyvictimswouldnotwanttoreengage,
andthereforethe57%ratedocumentedinthisevaluationisquitegood.However,wecannotdiscern
exactlywhyvictimschosetoreengage,thoughthevictimcentered,traumainformedapproachusedby
notifyingpersonnelmayhavebeenapositivecontributingfactor.Thefactthattheinvestigatorswere
notaffiliatedwiththefocalpolicedepartmentmayhavealsobeeninfluentialtosurvivors,butwedid
notcomparenotificationexperiencesandreengagementratesasafunctionoftheorganizational
affiliationoftheinvestigativepersonnel(allnotificationswereconductedbyinvestigatorsaffiliatedwith
theprosecutorsoffice).
Mostvictimsdecidedthattheydidwanttocontinueinvolvementwiththecriminaljustice
systemtopursuepossibleprosecutionoftheircases,butasizableminoritydidnot.Giventhis
variability,wewantedtotrytounderstandthesepatternsinthedata.Asmightbeexpected,victims
emotionalreactionsduringthenotification(seesectionabove)wererelatedtotheirwillingnessto
engageinfurthercontactwiththecriminaljusticesystem.AsshowninTable5.1(nextpage),allvictims
whohadastrongnegativeemotionalreactiondidnotwantanyfurthercontact(100%ofthosewhohad
anegativeemotionalreaction).Mostvictimswhohadastrongpositivereactionweremorelikelyto
agreetofurthercontactwiththecriminaljusticesystem(78%ofthosewhohadapositiveemotional
273
reaction).Survivorswhodidnothaveastrongpositiveornegativeemotionalreactionalsotypically
decidedthattheydidwantcontinuedcontactwiththecriminaljusticesystem(64%ofvictimswhodid
nothaveastrongemotionalreaction).
TABLE5.1VictimsDecisions,byEmotionalReactionsatFirstContact
DidNOTContinue
InvolvementwithCase
DecidedtoContinue
InvolvementwithCase

StrongNegative
EmotionalReaction(n=5)
StrongPositive
AbsenceofaStrong
12/28=43%
16/28=57%
Wealsoexaminedhowvictimswillingnesstohavecontinuedcontactwiththecriminaljustice
systemwasrelatedtothetimesincetheassault,thevictimsageatthetimeoftheassault,andvictim
offenderrelationship.Figure5.7(nextpage)summarizesthesefindings.Timesincetheassaultwas
associatedwithvictimswillingnesstohavecontinuedcontactwiththecriminaljusticesystem.
Specifically,62%ofvictimswhohadbeenassaultedmorethan9yearspriortothenotificationwere
unwillingtoengagefurtherwiththeinvestigator(s),whereasonly18%ofvictimswhowerenotified9
yearsaftertheassaultand50%victimsnotifiedlessthan9yearsaftertheassaultwereunwillingto
continuetheirparticipation.Inotherwords,themoretimethathadelapsedsincetheassaultandthe
notification,victimswerelesswillingtoreengagewiththecriminaljusticesystem.
274
FIGURE5.7VictimsDecisionsRegardingContinuedParticipationintheCriminalJusticeSystem
Under16yearsold
(attimeofassault)
0%
Lessthan
9years
since
assault
1624yearsold
(attimeofassault)
62%
Known
Over24yearsold
(attimeofassault)
31%
Known
Known
N/A
100%
N/A
50%
Stranger0%
StrangerN/A
Known
9years
since
assault
Stranger0%
Known
Known
N/A
N/A
N/A
18%
StrangerN/A
More
than9
years
since
assault
62%
Stranger0%
Known
Known
Known
0%
StrangerN/A
Stranger25%
100%
100%
Stranger60%
Stranger33%
Legend
Known:86%
Stranger:29%
Victimwaswillingtoengagefurtherwiththeinvestigator(s)
Victimwasnotwillingtoengagefurtherwiththeinvestigator(s)
XX%Percentofvictimsnotwillingtoproceedwithmeetings/investigation/prosecution
275
Victimswhowerebetween16and24yearsoldatthetimeoftheassaultwerelesswillingto
interactfurtherwiththeinvestigator(s)ascomparedtovictimswhowereolderoryoungeratthetime
oftheassault(62%unwillingfor1624yearolds;31%unwillingforvictimsovertheageof24;0%for
victimsunder16yearsold).Again,victimsinthisagegroup(1624)oftenhavenegativeexperiences
withthepolice(i.e.,highratesofsecondaryvictimization,seeCampbelletal.,2012),andthismight
explaintheirdecisionnottoengagefurtherpostnotification.
Finally,thenatureoftherelationshipbetweenthevictimandtheperpetratorseemedtorelate
tovictimswillingnesstoparticipatefurther.Mostofthevictimsassaultedbysomeonetheyknew(86%)
wereunwillingtoengagefurther,whereasonly29%ofvictimsassaultedbyastrangerwereunwillingto
havecontinuedcontact.Forthosewhowerevictimsofnonstrangerrape,theDNAtestingandCODIS
hitdidnotrevealnewinformation(itconfirmedwhathadpreviouslybeenknownabouttheidentityof
theassailant),andforreasonswedonotknow,theydidnotwanttoreengagetothesameextentas
didvictimsofstrangerrape.Giventhatonlysevennotificationsinvolvedaknownperpetrator,itislikely
thattherewerenotenoughcasesinthesampletoidentifypatternsregardingwhynonstrangerrape
victimsdidnotwanttoengagefurtherwiththecriminaljusticesystem.
Investigators and Advocates Reflections on the Victim Notifications
Inadditiontorecordingvictimsreactionstothenotification,investigatorsandadvocates
reflectedupontheirownexperiencesparticipatinginthenotificationprocess.Overall,theinvestigators
overwhelminglydescribedtheirexperiencesaspositiveandsuccessful,feelingthattheyhadhelped
starttheprocessofbringingjusticeforthevictim.Oneinvestigatorexplainedthats/he,feltasmall
senseofaccomplishment.Itfeltgoodtobetheindividualthatmadethenotificationandstartthe
process;similarly,anotherinvestigatorexpressedthat,itfeltgoodtohavethevictimsconfidence
restoredinthepoliceattheendofourinitialmeeting.Investigatorsalsonotedthattheywerehappyto
276
seeasmanyvictimsdoingwellintheirlivesastheydid;asonedescribed,she(thevictim)wasina
goodplacewithherfamily,children,schooling,andhadaplanwithherlifenow.Boththeinvestigators
andadvocatesfelttremendousempathyforthevictimsandwerehumbledbybearingwitnessto
survivorsrelivingtheexperiencesomanyyearsaftertheassault.
Intheirreflections,theinvestigatorsandadvocatesnotedseveralimportantlessonslearned
aboutvictimnotification(seealso,Figure5.9attheendofthischapter).Withrespecttolocating
victims,investigatorsreportedthathittingthestreetsandtalkingtothevictimsinpersonworkedbest
becausesomanyvictimswouldlikelynothavebeenwillingtotalkviaphone.Onecriticaldetailwasthe
decisionfortheinvestigatorstoarriveinanunmarkedcarandcivilianattire(i.e.,notapoliceuniform).
Oneinvestigatorreportedthatavictimbecamenervousforaminuteandwasworried[theinvestigator]
wasinapolicecar...[thevictim]commentedthat[s/he]didnotwanttheneighborstoseepoliceatthe
house[s/he]wasstaying.Similarly,anothervictimdidnotwant[his/her]otherfamilymembersto
knowanddecidedtotellthem[theinvestigator]wasaJehovahWitness.Iftheinvestigator(s)had
arrivedinmarkedvehiclesand/oruniforms,therewouldhavebeenconsiderableunwantedattention
anduncomfortablequestionsforthevictims.Sometimestheinvestigatorshelpedsurvivorscomeup
withacredibleexcusetheycouldgivefriends,family,neighborsastowhythey(theinvestigators)
weretheretoseethem.Theinvestigatorsalsoemphasizedtheimportanceofmeetingon[thevictims]
terms;victimsmayrequestaspecifictimeorlocationforsafetyreasonsanditisimportanttohonor
theirwishes.Forexample,onesurvivorrequestedtomeetataneutrallocation,andtheinvestigators
laterfoundoutthesuspectcomesandgoesfrom[thevictims]houseandmaybestayingtherefrom
timetotime.Meetingwithvictimsontheirterms,inperson,andtravelingindiscreetvehiclesand
attireseemedtobeofutmostimportance.
Asnotedpreviously,victimshadmarkedlydifferentemotionalreactionstothenotification,and
itwaschallengingsometimesforthenotificationstafftoknowhowbesttorespond.Theadvocates
277
notedthatoveralltheinvestigatorsdidanexcellentjobremainingpatient,givingthesurvivortimeto
reflectandrecallresponse[s]acknowledgingcomplexity,and[the]importanceof[thesurvivors]
needs.However,theadvocatesnotedthatsometimesinvestigatorsmove[d]tooquicklythrough[the]
questionsanddidnotseemtogivepauseorshow[the]typicallevelofsensitivitytothesurvivors
reaction.Tellingvictimstocalmdownorattemptingtotouchthevictiminordertocomforthim/her
wasnotwellreceivedbythesurvivors.Theinvestigatorsandadvocatesnotedthatitwasparticularly
importanttobeattunedtothepossibilitythatinknownperpetratorsexualassaults,theremaybea
historyofdomesticviolence/intimatepartnerviolence.Inonecase,anadvocatenotedthatthevictim
appearedsurprised[theinvestigator]understooddomesticviolenceissuesandwhy[s/he]maynothave
wantedtofollowthrough(withthecase)backthen.Theinvestigatorshadhadtrainingonthedynamics
ofdomesticviolence,whichwasinstrumentalforthembuildinganempathicconnectionwithsurvivors.
Theinvestigatorsandadvocatesnotedthatthelogisticsoftheprotocolweresometimes
challengingtoimplement.Jugglingschedulestolineupadateandtimewhenthevictim,investigators,
andadvocatewereallavailablewassometimesdifficult,butwhenthatproblembecameapparent,the
leadershipoftheadvocacyorganizationssteppedintoaddressitimmediately.Coordinatingwiththe
advocateswasalsodifficultifvictimswantedtojumprightinatthefirstcontactwhentheadvocate
wasntthereandstartdiscussingthecaseinthemoredetail;theinvestigatorsfelthandcuffedbythe
protocolbecausethecollaborativehaddecidedthatfirstcontactshouldbebriefandthatsurvivors
shouldhavetheoptionofanadvocatepresentformoreextendeddiscussionsaboutnextsteps.Again,
thecollaborativediscussedthesesituationsandclarifiedthattheoverarchinggoaloftheprotocoland
theprojectgenerallywastorespectvictimschoices,soifsurvivorswantedtoproceedimmediately,
theyoughttobeinformedabouttheoptionofhavinganadvocate,butthatiftheywantedtocontinue
withoutone,thatwastheirchoicetodoso.
278
Perhapsthemostdifficultpartofthenotificationprocessfortheinvestigatorsandadvocates
wastryingtoanswersurvivorsquestionsastowhywhynow,whynotthen,whydidnoonebelieve
themyearsago,whyweretreatedthewaytheywerebackthen?Oneinvestigatorhighlightedacasein
whichthesurvivorwashappytobenotifiedandwaswillingtoparticipateinfurtherintheinvestigation
ofthecase,butwasangryandsaddenedandwantedtoknowwhynoonebelieved[her/him]in2005.
Theinvestigatorsacknowledgedthevictimshurtandanger,reiteratedtheirapologies,butultimately,
thereisnogoodanswerforwhathadhappenedandthepainthesurvivorshadbeenputthrough.In
thisevaluation,wedidnotcomparevictimnotificationsconductedbythefocalpolicedepartmentwith
thoseconductedbytheinvestigationunitintheprosecutorsoffice,sowecannotdetermineifandhow
victimsreactdifferentially,dependingonwhichorganizationcontactedthem;however,itdoesappear
thatvictimsoftenappreciatedthattheinvestigatorswerenotaffiliatedwiththepoliceandthatthis
opportunityforafreshstartwiththeircaseswaswelcomedbymany.
Summary & Conclusions: A Victim-Centered, Trauma-Informed Approach to

Victim Notification
TheDetroitcollaborativeusedamultidisciplinaryteamapproachtodevelopavictimcentered,
traumainformednotificationprotocol,whichwasimplementedwithasmallsampleof41cases.Most
ofthesevictimsweresuccessfullylocatedandnotified(95%todate).Again,therearenopublished
studiesforcomparisonpurposes,butobjectively,thisrateisquitehigh.Typically,mostsurvivors(~65%)
couldbefoundwithrelativelylowinvestigationaleffortbutthatstillrequiredhittingthestreets,as
oneinvestigatorputit.However,thehighrateofsuccessfullocationsinthisprojectwasbecausestaff
investedextraefforttofindasmanysurvivorsaspossible,whichsometimesrequired12+phonecalls
and6+inpersonvisitstovariouslocationstryingtofindthevictim.Itisdifficulttoknow,without
comparativedata,howthesurvivorsinDetroitmightcomparetothoseinotherjurisdictionswith
279
respecttotheirlocateability,butcertainlytherewereanumberofhardtofind,transientindividuals
whocouldbefoundwithsufficientinvestigationaleffort.
Withrespecttowhathappenedinthenotificationmeetingsthemselves,theevaluationfindings
confirmedmanyofthehopesandworriesdiscussedbythemultidisciplinarycollaborative
throughouttheproject:thereisnoonebestwaytodovictimnotification,eachcaseisunique,and
victimsreactdifferentlytothisnewssomewerehappyandrelieved,otherswereangryandsad,and
nearlyallweretroubledbythefactthatnothinghadhappenedintheircasessomanyyearsago.
Despitethesecasebycasevariations,theevaluationfindingshighlightedsomepossiblepatternsinthe
notificationdatathatcanbepursuedinfutureresearch.Victimswerelesslikelytoreactpositivelyand
toreengagethelongerthetimebetweentheassaultandthenotification,whichisnotsurprisingand
certainlyunderscorestheimportanceoftimely(i.e.,atthetimetheassault)testingofSAKsand
investigatingreportedsexualassaults.Survivorswhowere1624yearsoldatthetimeoftheassault
weresomewhatmorelikelytohaveastrongnegativereactiontothenotificationandwerealso
somewhatlesslikelytowanttohavecontinuedcontactwiththecriminaljusticesystem.Giventhat
priorresearchhasfoundthatvictimsinthisagegroupoftenhavedifficultexperienceswithpolicewhen
reportingasexualassault,itstandstoreasonthatthesesurvivorsmayhavehaddifficultencounters
yearsago,andassuch,theyweredisinclinedtoreengage.Atthetimethisprojectwasconducted,only
asmallnumberofnotificationshadbeenconductedwithvictimsofnonstrangerrape,whogenerally
werenotaslikelytowanttocontinuecontactwiththecriminaljusticesystempostnotification.
Whenwepresentedthesefindingstothecollaborativepartners,thenotifyingpersonnel
discussedhowtheprotocolcouldbemodifiedtobemoreresponsivetotheneedsofvictimswho
tendedtoreactmorenegativelyanddecidednottoengageinfurthercontactontheircases.In
particular,thegroupdiscussedhowtoimproveschedulingandinformationsharingamongpractitioners
(withintheboundsofeachdisciplinesconfidentialityrequirements)However,intheend,the
280
collaborativepartnersreadilyacceptedthefactthatsomesurvivorsmaychoosenottopursuetheir
cases,andrespectingthatchoiceisessentialinavictimcenteredapproach;asonememberoftheteam
summarized,ourjobistosupportvictimsinwhatevertheydecidetodo,inwhateversbestforthem,
andtoletthemknowthatourdoorisalwaysopentothemiftheyeverchangetheirmindsoriftheyever
needus.
281
CHAPTER 6: Discussion
DetroitisoneofagrowingnumberofU.S.citiesthathavelargenumbersofuntestedrapekitsin
policeproperty(seewww.endthebacklog.org).Inresponsetothisgrowingnationalscaleproblem,the
NationalInstituteofJusticefundedtwoactionresearchprojectstostudytheunderlyingcausesofthis
problemandtodevelopevidencebasedresponsestrategiesthatcanbeusedtohelpjurisdictionswith
largenumbersofunsubmittedSAKs.Tothatend,theDetroitSAKARPhadfourprimarygoals:
1) ToobtainanaccuratecountofthenumberofSAKsinpolicepropertythoughacompletecensus;
2) ToidentifytheunderlyingfactorsthatcontributedtowhyDetroithadsomanyunsubmittedSAKs;
3) TodevelopanempiricallybasedplanfortestingSAKsandtoevaluatetheefficacyofthatplan;
4) Tocreateavictimnotificationprotocolandevaluatetheefficacyofthatprotocol.
Thepurposeofthischapteristwofold.First,foreachoftheseprimaryprojectgoals,wewillsummarize
keyfindings,discusslessonslearned,andsuggeststrategiesforimprovingpolicyandpractice.Wewill
alsodescribethegroupprocesswithinthecollaborativeandoffersuggestionsforstrengtheninginter
grouprelationsinmultidisciplinaryteams.Second,wewilloutlinewhathaschangedinDetroits
responsetosexualassaultsincethediscoveryoftheuntestedkitsinAugust,2009andduringthe30
monthsofthisproject.Actionresearchprojectsaresupposedtosupportempiricallybasedproblem
solving,sowewillexaminetheextenttowhichtheDetroitSAKARPservedasacatalystforchange.
Major Findings of the Detroit SAK Action Research Project

Goal 1: Determine How Many Unsubmitted SAKs in Detroit Through a Complete Census
KeyFindingsfromtheSAKCensus.WhenthisactionresearchprojectbeganinApril2011,the
firsttaskwastoobtainanaccuratecountofhowmanySAKswereinpolicepropertyandtoascertain
282
thetestingstatusofeachkit.Ostensibly,thisshouldhavebeenastraightforwardtaskofcullingand
crossreferencingdatabases.Itwasnot.Thepolicepropertydatabaseinventoriedwhatitemshadbeen
enteredintostorage,buttestingstatuswastrackedinaseparateExcelspreadsheet.Recordsregarding
theadjudicationstatusofthecasesassociatedwitheachkitwerelargelypaperrecords,dispersed
acrossmultipleorganizations(police,prosecution,court,corrections).Complicatingmatters,police
officialshadnotbeenforthcomingwithinformationaboutthestatusofthekitssincetheirdiscoveryin
2009,whichmadeotherstakeholdersskepticalabouttheaccuracyandcompletenessoftheirrecords.
Giventhisstateofrecordsmanagementandgrowingconcernsabouttransparency,theDetroitSAKARP
didamanualcensuseachkitwasretrievedfrompolicepropertyandcounted,onebyone.
Themanualcensuswasalaborintensiveendeavor,whichtooknearlyfourmonthstocomplete
(15weeks)andrequiredapproximately2,365hoursofstafftime.Theprocessofcountingthekitswas
arduous,butoneofthekeygoalsofactionresearchprojectsistodocumenttheontheground
challengesoftacklingcomplexcriminaljusticeproblems.Tothatend,theresearchteamfollowed
alongasthecollaborativeconductedthecensus,summarizingtheprocessintoastepbystepguide
thatcanbeusedbyotherjurisdictions(seeFigure2.1TheStepbyStepProcessofPlanningand
ConductingtheDetroitSAKARPCensus).TheDetroitSAKARPstruggledtodefinethegoalsofthe
census:Justcountthekits?Countandascertaintestingandadjudicationstatus?Countandcompile
policereportsandotherdocumentation?Atthebeginningofthecensus,alloftheseaimswere
considered,butthegroupfocusedoncountingallSAKandcollectingonlybasicinformation(victim
name,DOB,examdate,assaultdate)foreachkit.
Theresultsofthiscensusindicatedthattheinitialestimatesmadeatthetimeofthediscovery
(~10,000to11,000SAKs)wereremarkablyaccurate:aftercrosscheckingthecensuscountwiththe
policepropertydatabaseandweedingouttheSAKsthatturnedoutnottocontainsexualassault
evidence(i.e.,aSAKboxhadbeenusedtoholdotherevidence),therewere11,219SAKsinpolice
283
property(asofNovember1,2009).Inthisproject,wefocusedonidentifyingthetestingstatusofthese
SAKs,andbasedontherecordsavailable,2,512SAKshadlaboratoryIDnumbers,indicatingthatthey
hadbeensubmittedtothepolicedepartmentcrimelabfortesting;however,attheconclusionofthe
ARP,westilldidnothaveconfirmationastothetestingstatusofthesekits.Thescopeofthisaction
researchprojectfocusedonthe8,707thathadneverbeensubmittedfortesting.
Thoughweareconfidentintheseresults,therearesomelimitationswiththiscomponentof
projectthatneedtobeacknowledged.Withregardstoconductingthecensusitself,manydifferent
peoplewereinvolvedincountingthekits(e.g.,assistantprosecutors,lawschoolstudents,interns,ARP
projectstaff),andthoughtheyweregivenclearinstructionsandconsistentsupervision,thereisalways
thepossibilitythatindividualsunderstoodandimplementedthedirectionsindifferentways.Giventhat
thescopeofthecensuswasnarrowedrelativelyquicklytocountingthekitsandcollectingbasic
informationthatwasclearlydemarcatedontheoutsideoftheSAKs,wearereasonablyconfidentthat
thedatacollectionwasperformeduniformly.Becausetheorganizationsinvolvedinthisprojectdidnot
havesufficientITresources,datacollectionwasdoneonpaperforms,whichwerethentransferredto
anExcelsheet(i.e.,thegroupdidnothaveenoughlaptopstoallowfordirectcomputerizeddataentry).
However,theARPprojectcoordinatorturnedthisproblemintoanasset:censusstaffwouldturnin
smallstacksofpaperrecords,whichshethenimmediatelyenteredintothecomputer;any
inconsistencies,illegibleentries,orotherproblemscouldthenberesolvedbeforethekitswerereturned
intoproperty.
ThecountoftheSAKswasdoneinathorough,systematicway,butwhetherthatfinalnumber
trulyreflectstheactualnumberofSAKsinpolicecustodyisstillsubjecttodebatebecausewecouldonly
countwhatwewereprovided.ToensurethatpolicepropertypersonneldidindeedretrieveeverySAKin
custody,thereweremeetingsbetweenseniorleadershipoftheprosecutorsofficeandthepolice
department,stressingtheimportanceofacomplete,exhaustivecount.Thepropertystoragefacilities
284
weretouredbytheARPprojectcoordinator,otherARPstaff,andtheresearchteamsothatallparties
understoodthesystem,couldaskquestions,andidentifypossibleproblems.Despitetheseextensive
efforts,weknowthatsomeSAKswerenotidentifiedinthecensus.AftertheARPhadconcluded,we
wereinformedthatwhenthepolicedepartmentmainheadquartersmovedtoanewbuilding,the
propertydivisionfoundanadditional37SAKsaspartofanaudittheywereconductinginrelationtothat
move.These37SAKsarenotinthecensusresultspresentedinthisreport(astheywerediscovered
afterthefact).Wehighlightthisanecdotebothtoacknowledgealimitationinourcensusresults,and
alsotoemphasizethedifficultiesoffindingallSAKsforacensus,particularlyinlarge,urbanjurisdictions,
sothatothercommunitiesfacingthisproblemareawareofhowthoroughtheireffortswillneedtobe
toobtainacomprehensivecount.
LessonsLearnedfromtheSAKCensusandImplicationsforPolicyandPractice.Throughoutthe
multimonthprocessofconductingthecensus,theresearch/evaluationteamkeptdetailednotes
regardingkeylessonslearned;inaddition,threefocusgroupswereconductedattheendofthe
projecttocollectteammembersfeedbackregardingprojectsuccesses,challenges,andlessonslearned
(seeAppendixB:ProjectMethodology).Thecouldhave,wouldhave,shouldhavesregardingthe
censusaresummarizedinFigure6.1(followingpages).Asonememberofthecollaborativenoted,I
hopeourMondaymorningquarterbackingwillmakelifeeasierforanotherjurisdictionthathastodoa
completecensus.PerhapsthemostimportantlessonlearnedthattheDetroitSAKARPcanofferother
jurisdictionsistherecommendationtostartsmall:reviewasampleofkits(evenasfewas10wouldbe
informative)andwalk(them)through,starttofinish:figureoutwhatinformationisavailablefrom
whatsource,whatinformationneedstobetrackedinthecensus(vs.whatinformationcanwaittobe
compiledlater),who(whichstaffmembersfromwhichorganizations)haveaccesstothatinformation
andcanparticipateinthecensus,andhowmanystaffhoursitwilltaketocompletethecensus.
FIGURE6.1
285
Lessons Learned:
Conducting a SAK Census
The take-home lessons from the Detroit SAK ARP based on

their experiences conducting a census of SAKs in police property.
What we should have done, in twenty-twenty-hindsight, if only wed known.
1. Bring everyone to the table
2.
A review of kits will bring up complex legal,

psychological, and evidentiary issues. Have a
broad-based multidisciplinary team to ensure
that diverse perspectives will be considered at
every stage of the process.
Establishing the goals of the

project at the beginning will
help streamline the process
and save staffing time.
Consider including representatives from:

o Police
o Prosecution
o Forensic sciences
o Medical/nursing
o Systems-based advocacy
o Community-based advocacy
3. Clarify language and agree

on terms to be used
The extent to which team members

can be consistent and precise in
their language, planning and
conducting the census will result in
fewer opportunities for
miscommunication.
Formulate goals
Narrowly defined goals might

include simply determining the
number of untested SAKs.
More expansive goals might focus on
determining the number of SAKs and review
each for SOL-risk, prosecution potential, and
other such complex issues.
Some Terms to Consider:

Submitted= the SAK was submitted to a laboratory for
testing
Tested= the SAK was tested (for older kits, clarify
whether that testing included DNA analysis
Adjudicated= the SAK is associated with a case that
has already been adjudicated (note: this should be
cross-checked with testing status, as some cases may
have been adjudicated without the testing of the kit)
Backlogged= the SAK was submitted to a laboratory
for testing, but is still in queue at the lab to be tested
286
4. Find ALL the locations where SAKs
POSSIBLE SAK LOCATIONS:
are being stored

To obtain an accurate count of all
unsubmitted SAKs, the team needs to
identify all possible locations where such
kits could be found.
Police property storage rooms/facilities

Police property overflow storage,
Property rooms at police precinct/offices
Specialized storage cold rooms
Hospital emergency rooms
SANE programs
5.Find out what information is computerized (and whats not!)

If the kits date back many years and/or if the law enforcement agency has been
under-resourced for many years some records may not be computerized.
Determine which organizations maintain which databases, what info is already
shared between organizations, and what information not currently shared could
be shared with others.
6. Loop in IT support
Depending on what information is/is not computerized,
the role of IT staff may vary. At a minimum, it is likely the
team will need help crosschecking and merging files/fields.
7. Start small
Review a small sample of kits, even as few as 10 kits/cases. This will help clarify goals, identify
available resources, determine what information is hard to track down, and gauge what
resources/staffing will be necessary in the future. Use the following steps to take one SAK and
walk through it, start to finish:
Take one SAK and walk through it, start to finish
Determine info on the outside of the kit
Pull records for each test-run SAK and

identify problems with this process
Assess whether such info is important & should

be included in the database
Determine what other info needs to be tracked

down according to the predetermined goals
Decide what fields from these additional

sources should be included in the central
database
Determine where other info may be found
Track how many staffing hours went into

walking through each SAK/case
287
10. Develop a central

database
9. Touch it once
This lesson intends to prevent
later backtracking and
duplicative efforts. Develop a
work flow process that
streamlines efforts and saves
staff time and effort.
A shared &accessible database

for all organizations on the team
to use can help ensure all can
see the information and that
efforts are not duplicated.
If possible, all info should be
entered directly into the
database instead of being
recorded on hard copies.
11. Evaluate
resources available
and develop a
staffing plan
The start small test run
should provide info about the
staff needed for a census.
The team should develop a
timeline for completion of the
census and draw upon staff
from multiple organizations to
reach this goal.
12. Remember: The devil is in the details!

The census might involve physically retrieving each SAK from police property so that
info on the kits can be recorded. Key logistic details to be considered include:
Training and supervising staff/volunteers who will be handling the kits

Developing procedures for retrieving/returning the kits from property
Finding the physical space to review the kits
Securing computer access to enter info directly into the database
Providing other supplies (masks, gloves, etc.) for staff/volunteers
Maintain proper legal chain of custody and police property procedures,
having police or forensic staff present if necessary
13. Support the staff and volunteers

Develop a staffing plan that allows for rotated
duties and scheduled breaks, and identifies
resources for assisting staff with the possibility of
vicarious trauma. Its important to address issues of
burnout and vicarious trauma in ways that dont
single out individuals or make them feel stigmatized.
288
Once the census has been completed

14. Expect that the final count will create controversy
There will be disagreements and challenges regarding the
results of the census. A multidisciplinary input throughout the
process can help minimize disagreements.
15. Reexamine
16. Prepare for media
local policies &

state statutes
regarding
evidence retention
inquiries and public

scrutiny
A census will help a jurisdiction gauge

the amount of evidence in its possession
as well as provide details about the
nature of that evidence. It may be
helpful to review local policies regarding
evidence retention as well as examine
state statues to explore whether policy/
legislation changes may be warranted.
Develop a proactive media plan

that can manage
requests, promote
transparency, &
It is suggested
simultaneously
that jurisdictions
protect the
wait to release
integrity of the
the findings until
census process.
they are
complete.
17. Be aware that the final count may not be final

Its possible that even after the census has been finalized, more information will become
available that will require an adjusting of the numbers (i.e., more kits will be discovered
and the final count will be greater than anticipated). Its also possible that after the kits
have been submitted and are opened at the lab, the evidence inside some kits may
not be sexual assault-related evidence.
289
TheprocessofconductingtheDetroitSAKARPcensus,aswellastheresultsthemselves,have
severalimplicationsforpolicyandpracticeregardingthetrackingandstorageofSAKsfromthepointat
whichtheyarecollectedbyahealthcareproviderandreleasedbyvictimsforlawenforcementretrieval
towhentheyaretakenintopolicecustody,andbeyond:
1) PolicedepartmentsmaynotbeawarethattheyhavelargenumbersofunsubmittedSAKs
iftheirpropertystoragepoliciesdonotsegregateSAKsfromothertypesofevidence.
InbothNewYorkCityandDetroit,theproblemofuntestedSAKsbecameevidentwhen
therespectivepolicedepartmentsreorganizedtheirpropertyroomstopullSAKsfrom
generalstorage(i.e.,intermingledwithothertypesofcrimesceneevidence)and
grouped/segregatedthem.However,NewYorkimmediatelyidentifiedthelarge
numbersofuntestedSAKsasaproblem,whileDetroitdidnot.Thepropertyroomre
organizationinDetroitoccurredin2002,butSAKscontinuedtoaccumulateforseven
moreyears,untiltheywerespottedgroupedinstorageboxesduringaproperty
auditin2009bymembersofotherorganizations.
Therefore,policedepartmentsthatdonotsegregateSAKsmaynotbeawarethatthey
haveaproblemofunsubmittedSAKs.StoragepracticesthatseparateSAKsfromother
crimesceneevidenceandstorerapekitstogethermaymakeiteasiertorecognizethe
problemandmonitoreveningeneralwaySAKaccumulation.IfSAKsthathavebeen
testedareseparatedfromthosethathavenot,itwouldbeeveneasiertoassessthe
extenttowhichtherewasaproblem.
2) PolicedepartmentsmaynotbeawarethattheyhavelargenumbersofunsubmittedSAKs
iftheirpropertycomputersystemdoesnotinterfacewithrecordsthattrackSAKtesting.
InDetroit,itwasalsodifficulttoidentitythattherewasaproblemofunsubmittedSAKs
becausethepolicepropertydatabasedidnottracktestinganddidnotinterfacewith
otherrecordsthatdidtracktesting.AseparatestandaloneExcelspreadsheetwasused
bythepolicedepartmentcrimelaboratorystafftorecordwhichSAKshadbeentested
andwhichhadnot,buttherewasnosysteminDetroitthatallowedforcomputerized
290
monitoringofhowmanySAKshadbeenenteredintoevidenceandhowmanyhadbeen
submittedforDNAtesting.
Therefore,policedepartmentsthatdonothaveITsystemsthatcrossreference
propertyandtestinginformationmayhavedifficultyidentifyingwhethertheyhavea
problemwithunsubmittedSAKs.Informationtechnologysystemsthatallow
practitionersfromthevariousdisciplinestoascertainthelocationandstatusofeachkit
fromthepointitwascollectedinthemedicalforensicexamthroughtestingwouldbe
helpful.
3) PolicedepartmentsmaybeawarethattheyhavelargenumbersofunsubmittedSAKs,but
maynotviewthisasproblematic,givendepartmentalnormsregardingsexualassault
investigations.
InDetroit,therewasnoalarmin2002whenthepropertyroomreorganization
identifiedthousandsofSAKsinevidence.Afterthekitswerediscoveredin2009,an
internalaffairsinvestigationintothematter(whichsurfacedduringtheARP)concluded
thattherewerejustifiablereasonswhySAKshadnotbeentested,butthatconclusion
wasbasedonasacursoryreviewofanonrandomselectionofaverysmallnumberof
SAKs.Theinternalaffairsreportbaseditsconclusionsonthedocumentationinthe
policereports,whichtheARPresearchclearlyshowedwereoftenvictimblaming.
Therefore,thissuggeststhattheremaybedeeperissuestoexamineregardinga
departmentscultureregardingsexualassaultinvestigationsarethesecasesan
organizationalpriority?Hasthedepartmentinvestedsufficientstaffingresourcesfor
sexualassaultinvestigations?Howdoofficers/detectivesviewvictims?Havetheyhad
adequatetrainingabouttraumaanditsimpactonvictims?Statelevelandnationallevel
technicalassistanceisnecessarytoworkwithlawenforcementleadershiptodevelop
comprehensivepoliciesregardingsexualassaultinvestigations.
291
4) PolicedepartmentsmaybeawarethattheyhavelargenumbersofunsubmittedSAKs,but
donotwanttopubliclydisclosetheproblem.
InDetroit,theproblembecamepublicquickly,giventhatoutsideorganizations(the
prosecutorsofficeandstatepolice)werepartoftheauditinwhichthekitswere
discovered.Assuch,thepolicedepartmentdidnothaveachoiceastowhetherto
disclosethattheyhadlargenumbersofkits.However,inotherjurisdictions,itseems
possiblethatpoliceofficialsmaybeawarethattheyhaveaproblemanddonotwantto
disclosethatinfearofbecomingthenextDetroit(orHouston,Memphis,NewYork,
LosAngeles,Phoenix,SanAntonio,etc.)
Therefore,stateandnationalpolicymakersneedtoconsiderhowbesttopromotethe
disclosureofthisproblem,sothatpolicedepartmentswillcomeforwardandpublicly
reporthowmanyuntestedSAKstheyhaveincustody.Itseemsprobablethatone
reasonwhypoliceagenciesmaybereticenttodisclosethisinformationisfearoflegal
action(e.g.,lawsuits,ashashappenedinMemphis).Therefore,stateandfederalpolicy
makersmaywishtoaddressthisissueproactivelyintheirmedia/communications.
Anotherprobablereasonwhylawenforcementagenciesmaybereluctanttodisclose
thisproblemisthattheymaynotknowhowtoaddresstheproblemand/ormaynot
havetheresourcestoremedyit.Stateandnationalpolicymakersandtechnical
assistanceprovidersmaywishtoconsiderhowtodevelopcomprehensiveresource
materialsforjurisdictionsonuntestedSAKs,including,butcertainlynotlimitedto,the
stepbystepguidescreatedinthisproject.Additionally,increasingpublicawareness
aboutfederalprogramsforSAKtesting(e.g.,NIJDNABacklogReductionGrants,theNIJ
FBIpartnershipstotestSAKs)mayalsobehelpful.
292
Goal 2: Examine Why Detroit Had So Many Unsubmitted SAKs in Police Property
KeyFindingsRegardingtheUnderlyingContributingFactors.Howdoesapolicedepartment
accumulate11,000+SAKs,mostofwhichwereneversubmittedfortesting?Toanswerthisquestion,
weconductedanindepthhistoricalanalysisoftheprimaryDetroitorganizationsthatserverapevictims
(police,crimelab,prosecution,medical/SANE,andvictimadvocacy)tounderstandtheirstaffing,
leadership,policies,andpracticesoverthethirtyyearsthatthesekitswereaccumulating.We
interviewedcurrentandformeremployeesinallorganizations,andexaminedpubliclyavailable
documentsandinternalorganizationalrecordstoassesstheresourcesavailableforservingrapevictims
(ingeneral)andtestingSAKs(specifically).Foradditionalcontext,wecollectedsimilarmetricsinfour
U.S.citiesthatarecomparabletoDetroitinsize/population,racialcomposition,and/orcrimerates
(Philadelphia,Dallas,Baltimore,andNewOrleans)togaugewhetherwhatwewereseeinginDetroit
wastypicalrelativetoothercommunities.
TheresultsofthishistoricalcontextualanalysisindicatedthatallorganizationsinDetroitthat
serverapevictimshavestruggledfordecadeswithchronicunderstaffingandresourcedepletion.Inthe
policedepartmenttherewasconstantinstabilityinleadership,suchthatanewChiefwasappointed,
onaverage,everytwoyears,whichisaturnoverratethatfarexceedsnationalnormsandisatypical
amongotherurbancitieswithcomparablecrimerates.Theturnoverstrickleddowntothesexcrimes
unit,whichalsochangedsupervisorsregularly,makingitdifficulttoidentifyandcorrectproblematic
policiesandlowperformingstaff.Overthisthirtyyearperiod,thesexcrimesunitalsosustainedtwo
50%cutsintheirstaffinglevels,andforthepastseveralyears,theyhavenotbeenabletooffer24/7
servicetothecommunity(i.e.,sexualassaultsthatoccuronoffhoursarehandledbypatrol,andthen
forwardedtosexcrimeslater).ThepolicecrimelabwasresponsibleforallDNAtesting(plusallother
typesofforensictesting)forthecityofDetroit,buttheytypicallyhadonly23DNAscientistsonstaff,a
numbersubstantiallylowerthannationalnormsandlowerthaninotherurbancitieswithcomparable
293
crimerates.Theprosecutorsofficehadfarmorestabilityinleadership,butgivencountylevelbudget
cuts,thenumberoftrialattorneysdeclinedsignificantlyovertime.SexualAssaultNurseExaminer
(SANE)programsarewidelyconsideredtobebestpracticeforpostassaultmedicalcare,butDetroitdid
nothavesuchaprogramuntil2006,sothevastmajorityofSAKswerecollectedbymedicalpersonnel
whohadminimaltraininginforensicevidencecollection.Withrespecttovictimadvocacyservicesfor
survivors,thepolicedepartmenthadalongstandingsystemsbasedvictimadvocacyprogram,but
communitybasedadvocacyservices(e.g.,rapecrisiscenter)werelargelynotavailable.
Withthiscontextestablished,wethenreexaminedtheSAKcensusresultstoexplorewhether
thehistoricalchangeswedocumented(e.g.,whenthepolicecrimelabgainedaccesstoCODIS,when
thesexcrimesunithadstaffingcuts)mightexplaintheratesofSAKsubmissionsovertime.For
example,stakeholdersnotedthatbeforetheadventofCODIS,theutilityofSAKtestingtoan
investigationwaslimited(e.g.,astrangerrapeDNAprofiledoesnothavethesameinvestigativeutility
withoutadatabaseofcomparisonsamplesagainstwhichitcanbecompared).Indeed,wefoundthat
thereweresignificantlylowerratesofSAKsubmissioninthepreCODISeraandsignificanthigherrates
whenthepolicedepartmentcrimelabhadfullaccesstoCODIS.Wealsoexaminedhowchangesinthe
policedepartmentsresourcesandpoliciesovertimemighthaveaffectedSAKsubmissionrates.For
instance,giventhatthesexcrimesunitexperiencedtwo50%staffingcuts,wetestedwhetherSAK
submissionratesdeclinedpostbudgetcuts.Therewasnosignificantassociationbetweenstaffinglevels
andSAKsubmissionrates,whichsuggeststhattheproblemofunsubmittedSAKsisnotsimplyaperson
powerissuethereareotherreasonswhypolicedonotsubmitSAKsthatmustbeidentified(see
below).WealsosawnosignificantrelationshipbetweenthepolicedepartmentspolicychangeinSAK
submissions(i.e.,bothknownoffenderandunknownoffenderSAKsshouldbesubmittedfortesting)
andactualsubmissionrates.Giventhatwewerenotabletoverifytheexistenceofawrittenpolicy
regardingSAKsubmissions,wesuspectthatmayhavebeenaverballycommunicatedpractice,which
294
didnotappeartodirectlyimpactfrontlinedecisionmaking.Wealsosawnochangeinsubmissionrates
asafunctionofthepolicedepartmentcrimelaboratorysreceiptoffederalDNAbacklogreduction
funds.Wewerenotabletodeterminehowthosefundswereallocated(i.e.,whatpercentageofthe
monieswereusedforSAKsspecificallyvs.testingDNAinothertypesofcrimes),butitappearsthatthis
fundingdidnotchangepracticeregardingSAKsubmissions(thoughthefundsmayhavehelpedkeep
submissionratesfromdropping).Onehistoricalchangethatdidaffectsubmissionrateswasthe
establishmentofaSANEprogram:oncethecommunityhadaresourceforqualitymedicalforensic
examsandforensicconsultation,SAKsubmissionratessignificantlyincreased.Practitionersacross
multipledisciplinesnotedthattheSANEswereinstrumentalineducatingmembersofother
organizationsabouttheutilityofforensicevidenceinsexualassaultcases.
ThesestatisticalanalysessuggestedthatalthoughDetroitorganizationswereclearlystruggling
withchronicscarcity,therewasnotadirectrelationshipbetweenpolicedepartment/policecrimelab
resourcesandSAKsubmissionrates.So,whatotherfactorswereinfluentialtolawenforcementand
theirdecisionswhethertosubmitaSAKfortesting?Toexplorethisquestion,weconductedadditional
qualitativeinterviewswithstakeholdersinallorganizationstounderstandthehistoryoftheirfront
practicesanddecisionmakinginsexualassaultcases.Wealsoreviewed1,268sexualassaultpolice
reportsassociatedwithunsubmittedSAKstoseehowpoliceinvestigatedthesecasesandhowthey
characterizedtheassault,thevictim,andtherapekitintheirwrittendocumentation.
Intheseinterviews,policepersonnelreadilyacknowledgedanindirecteffectofbudgetand
staffingcutsoninvestigationqualitysuchthatcuttingcornersbecamenormative.Caseswereclosed,
oftenlabeledascomplainantrefusedtoprosecute[CRTP]afterminimalinvestigationaleffort:Okay,I
madeacouplephonecalls,Ithrewmycardinthedoorandthatstheendofthat.Closeit.Becausethe
policedepartmentwaschronicallyunderstaffed,lawenforcementpersonnelfelttheycouldnot
investigateallreportsthoroughly,sosomecaseshadtobeweededout,amindsetcommonin
295
conditionsofchronicscarcity(Lipsky,2009;Mullainathan&Shafir,2013;Rouxetal.,2012;Tetlock,
2000).Thecasesthatwereweededouttypicallywereonesinwhichthepolicequestionedthevictims
credibility.Inboththestakeholderinterviewsandintheactualpolicereports,lawenforcement
personnelexpressednegative,victimblamingbeliefsaboutsexualassaultvictims.Rapesurvivorswere
commonlyassumedtobeprostitutesandthereforewhateverhadhappenedtothemifanything,as
policeoftencommentedintheirreportswastheirownfault.Adolescentswereassumedtobelying,
tryingtoavoidgettingintotroublebyconcoctingafalsestoryaboutbeingraped.Friendsand
acquaintanceshadgotwhattheygotbecausetheyhadchosentoassociatewiththeperpetrator.
Thesefindingsareconsistentwithpriorstudiesonrapevictimsexperienceswithlawenforcementand
thecriminaljusticesystemresponsetosexualassault(e.g.,Campbell,2008;HumanRightsWatch,2013;
Maier,2008,2011;Maddoxetal.,2011;Martin,2005;Monroeetal.,2005;Patterson,2011a,2011b),
andhighlighthowvictimblamingaffectsallaspectsoftheinvestigation,includingrapekittesting(see
alsoPatterson&Campbell,2012;Shaw&Campbell,2013).Withoutconsistentsupervisionandtraining
inDetroittochallengethesepractices,labelingcaseaftercaseasadealgonebad,orotherwise
dismissingitasnotreallyarape,wentunchallengedandunsubmittedSAKscontinuedtoaccumulate.
Theresultsofthisprojectalsorevealedhowthesenegativedynamicswithinthepolice
departmentwerereinforcedbyotherunits/organizationsinDetroit.Forexample,fromtheircolleagues
inthepolicedepartmentcrimelab,lawenforcementpersonnelheard,overandoveragain,thatthelab
didnothavethecapacitytotestallSAKstestingwasalimitedresource,tobeusedonlywhenyou
reallyneedit.LabpersonnelalsotoldthepolicethatSAKsubmissionsrequiredadditionalinvestigative
labortotrackdownsuspectsand/orconsensualpartnerstoobtainDNAreferencessamplesamessage
thatthepoliceheardas,theworkyoucantevenkeepupwith,yeah,nowitrequiresevenmorework.
Detectivesnotedthatitwasntworththeadditionaleffortandcashinginachit...withthelab
becausehospitalERdoctorswereconsistentlytellingthemthattheSAKswerenotgoingtobehelpfulto
296
theirinvestigation.WhydoctorsinDetroithospitalsbelievedthisfordecadesisunclear,asitnotappear
tobegroundedinanymedical,socialscience,orforensicresearch.Thesemessagesfromtheforensic
scienceandmedicalcommunities,whenheardthroughthepolicesownfiltersofresourcescarcityand
negativeattitudestowardsvictims,reinforcedtheirpracticesthatnotallSAKscouldorshouldbetested.
Bycontrast,themessagefromtheprosecutorsofficewasthatSAKtestingwasvaluable,often
criticalforprosecution.But,giventhelabslimitedcapacityfortesting,policefelttheyhadtohold
slotsforthecasestheprosecutorsneededandthereforetheyshouldnotsubmitallSAKsfortesting.
VictimadvocacyorganizationsmightalsobeexpectedtoemphasizetheimportanceofSAKtesting,but
inDetroit,advocateswerelargelysilent.Theadvocacyprogramwasinternaltothepolicedepartment
anditsdifficulttocalloutyouremployer,andDetroithadonlyoneexternal,communitybasedvictim
advocatebecauseoflackoffunds.
Itisimportanttoacknowledgethatthereareotherorganizationsthatundoubtedlyhaddirector
indirectinfluenceonsexualassaultcaseprocessinginDetroitthatwewerenotabletostudyaspartof
thisresearchproject.Forexample,theworkoftheprosecutors,inparticular,isstronglyinfluencedby
thejudiciary(e.g.,courtscheduling,timelines,continuances,allowableevidence,etc.).Theresearch
teammadeconsiderableeffortstoreachouttothesestakeholderstounderstandtheirperspectiveon
sexualassaultcaseprocessing,butalljudgesinthisjurisdictiondeclinedtoparticipateinaninterview.
Likewise,oureffortstounderstandtherootcausesofthechronicunderfundingofDetroitorganizations
waslimitedbythefactthatpublicofficials/fundersatthecity,county,andstateleveldeclinedtobe
interviewed.Therefore,oursystemicresearchonthistopicismissingtheperspectivesofsomekey
componentsofthecriminaljusticeandsocialservicessystems.
Despitethesemethodologicallimitations,thedatawedohaveclearlyindicatethatDetroit
sexualassaultorganizationssufferedchronicresourcescarcity,andovertime,scarcitycreatesscarcity
asMullainathanandShafir(2013)aptlynoted(p.67),suchthatcumulativedepletioncanhavefar
297
reachingeffects.Intraandinterorganizationalcommunicationbecamelessfrequentandmore
strained,andbunkersandsilosratherthansystemiccollaborationbecamethenorm(seeFried,1982;
Kramer,1990;Lipsky,2009;Rouxetal.,2012;Walsh,1961).Practitionersempathyforeachotherand
forthosetheyweretaskedtohelperoded,sothatmanysurvivorsweretreatedinrevictimizing,
dehumanizingways.Intheend,therapeofthousandsofindividualswassomethingthepoliceeither
couldntdoanythingaboutorwouldntdoanythingabout.Thismindset,repeatedincaseaftercase
forthirtyyears,resultedinsubstantialnumbersofunsubmittedSAKsontheshelfinpoliceproperty.
LessonsLearnedfromStudyingtheUnderlyingContributingFactorsandImplicationsforPolicy
andPractice.ThisparticularprojectgoaltoexaminetheunderlyingreasonswhyDetroithadsomany
unsubmittedSAKswasspecificallyrequestedintheoriginalNIJRFPforthisactionresearchproject(see
Chapter1:Introduction).Manyofthefindingsfromthiscomponentoftheactionresearchproject
helpedinformedpolicyandpracticeschangesinDetroit(seeEvidenceofInstrumentalUse:Changesin
Policy&Practice),but,overall,theexperienceofconductingthisresearchandsharingitwiththe
collaborativeactuallyheightenedtensionsanddiscord.Asonememberoftheteamnoted:
Imnotsurethiswasthemosthelpfulpartoftheproject...lookingbackatwhatwasandwhat
usedtobejuststirreditallupagain...theanger,defensiveness,mistrust,fingerpointing...
weweretryingtocometogethertomoveforwardandthistookusback...Igetthatitwas
required,Ijustdontthinkitwashelpful...wewouldhavefiguredoutwhatneededtobe
changedanyway,withoutdiggingupoldbattlesandoldfeuds.
Afundamentaltenetofsocialscienceresearch(ingeneral)andtheactionresearchparadigm(in
particular)isthatunderstandinghowandwhyaproblemdevelopedisessentialforplanningitssolution
andpreventingitfromhappeningagain(Coieetal.,1993;Kelly,2006;Klofasetal.,2010;Schensul,
2009;Seidman,1988;Trickett,2011).Thatsaid,Kennedys(2012)reflectionsontheBostonGunProject
298
alsonotedthateffortstoseekouttherootcausesmaynotbeashelpfulasonemightthinkfor
movingacommunityforwardintryingtostopcrimeandsolvesocialproblems:
Imabornandbredrootcauser,cameupbelievinginfixingtheeconomy,fixingeducation,
supportingfamilies,eradicatingracism:Healthecommunityandthecrimewilltakecareof
itself.ThatnotiondidntsurvivethefirstfiveminutesinNickersonGardens.
Butyoudontneedajobnottoshootpeople.Youdontevenneedtoliveastraightlifenotto
shootpeople.Youjustneednottoshootpeople.
Mostofusdontthinkupstreamlikethisinourordinarylives.Wedbedeadifwedid.Whenfall
comesanditgetscold,youputonasweater,buildafire,buildahouse.Youdontgotoworkon
therootcausesofwinter.
Itis,astheysay,anempiricalquestionastowhethertheDetroitcollaborativewouldindeedhave
figuredoutwhatneededtobechangedanywaywithoutthisresearchcomponent.Atveryleast,this
studywashelpfulinempiricallydocumentingriskfactorsforthedevelopmentofthisproblem,which
othercommunitiescanuseasaguideforamorefocusedtakingstockexplorationofhowandwhy
theyhavesomanyunsubmittedSAKs.Theseriskfactorsmayormaynotapplyinalljurisdictionsand
futureresearchisneededtodeterminetheapplicabilityandgeneralizabilityofthesefindingstoother
communities.
Pendingreplicationinotherjurisdictions,theresultsofthisprojectcanhelpinformpolicyand
practiceinregardstotheearlydetectionofthisproblem,andperhapsevenmoreimportantly,the
preventionofthisprobleminothercommunities.Theresultssuggestthefollowingriskfactors:
299
1) Chronicinstabilityinpolicedepartmentleadershipcanbeariskfactorforthe
accumulationofuntestedSAKsbecauseinconsistentsupervisionmakesitdifficultto
identifyandremedyproblematicfrontlinepractices.
InDetroit,theoversightofthesexcrimesunitanditspracticesregardinginvestigations
andSAKsubmissionwashighlyunstable,makingitdifficultforanyonesupervisorto
recognizethattherewasaproblemandtoinstitutecorrectiveaction.
Therefore,communitiesthathavehadfrequentturnoversinpoliceleadership(atthe
highestlevelsandinthesupervisionofthesexcrimesunit/sexcrimesstaff)shouldtake
stockoftheiruntestedSAKsandinvestigationalpractices.Trainingandtechnical
assistanceforlawenforcementagenciesonstrategiesforbufferingleadership
transitionsmaybeusefulaswell.
2) Understaffinginpolicedepartmentsexcrimesunitscanbeariskfactorforthe
accumulationofuntestedSAKs,asitcontributestotriaging/selectingcasesfor
investigation.
InDetroit,lawenforcementpersonnelreadilyacknowledgedthatunderstaffingledto
compromisedinvestigationalqualitysuchthatofficers/detectivesweededoutcases
becausetheydidnothavesufficienttimeandpersonneltodoafullinvestigation.
Therefore,lawenforcementagenciesneedtostaffsexcrimesunitscommensuratewith
theircrimesratesandthetime/effortthatthesecasesrequire.Asstateleveland
nationallevelprogramsandpoliciesaredevelopedregardinguntestedSAKs,funding
specificallyforinvestigationsisalsocritical.
3) ThelackofawrittenpolicyonSAKsubmissionscanbeariskfactorfortheaccumulation
ofunsubmittedSAKs;withoutclearexpectationsregardingkitssubmissions,individual
discretion(whichisofteninfluencedbystereotypesandvictimblamingbeliefs)
determineswhetherakitwillbesubmittedfortesting.
InDetroit,therewasnowrittenpolicy(thatweabletofind)regardingSAKsubmissions.
Thedecisiontosubmitakitwaslargelyatthediscretionoftheofficerincharge(OIC),
300
thoughsometimesSAKtestingwasspecificallyrequestedbytheprosecutorsoffice.
Giventhatdiscretion,officerstypicallydecidednottosubmitakit,oftenbecausethey
didnotbelievethevictimsaccountoftheassault.
Therefore,lawenforcementagenciesthatdonothavewrittenpoliciesregardingSAK
testingshouldsetsuchexpectationsandcodifythemintotheirstandardoperating
procedures.Communicationandcollaborationwithprosecutorsisalsocriticalin
establishingsuchpoliciessothattheseinterconnectedcomponentsofthecriminal
justicesystemhaveclear,consistentproceduresforSAKprocessing.
4) ChronicresourcedepletionisnottheonlyreasonwhySAKsarenotsubmitted;negative,
victimblamingattitudestowardvictimsaffectpolicedecisionsregardingsexualassault
investigationsandSAKsubmissions.
InDetroit,therewasclearevidencethatpolicedecidednottosubmitSAKsbecausethey
didnotbelievethevictim,sotheinvestmentoftime,effort,andresourcesinthecase
wasnotwarranted(intheiropinion).Thiswasparticularlycommonforassaults
committedagainstadolescents/youngerwomenandfornonstrangersexualassaults.
Therefore,trainingforlawenforcementonthedynamicsofsexualassault,the
neurobiologyoftraumaanditsimplicationsforvictimbehavior,andoffenderfocused
investigationsisnecessary.Thereareavarietyofempiricallybasedtrainingresources
onvictimcentered,offenderfocusedapproachestosexualassaultinvestigations(see
Shaw&SARAProject,2011forareview)thatneedtobewidelydisseminated.
5) Insufficientlaboratorystaffing(commensuratewithacommunityscrimerateandneed
forforensicservices)isariskfactorforuntestedSAKs,aslawenforcementpersonnel
decidenottosubmitSAKsiftheirlaboratorydoesnothavethecapacitytokeepupwith
demand.
InDetroit,thepolicedepartmentcrimelabrepeatedlytolddetectivesthattheydidnot
havethecapacitytotestallSAKs,sotheydidnotsubmitallSAKsforDNAtesting.The
staffinglevelsinthepolicedepartmentcrimelabwerewellbelownationalaverages.
301
Therefore,increasedfundingfortestingandbuildinglaboratorycapacityisessential.
Theresultsofthisresearchalsohighlightwhatcanbeproblematicabouthavingcrime
labsembeddedwithinapolicedepartment:bothunits(labandinvestigations)canbe
negativelyaffectedbybudgetconstraints,organizationalleadershipproblems,and
organizationalnormsregardingsexualassaultinvestigations.
6) Thelackofstrongcommunitybasedadvocacyservicesforsurvivorscanbeariskfactorfor
theaccumulationofuntestedSAKs,giventhattheseorganizationsadvocateforindividual
survivorsandpromotesystemicreforminthecommunityresponsetorape.
InDetroit,therewereminimalcommunitybasedvictimadvocacyservices.Detroitdid
nothavearapecrisiscenterandservicesforrapevictimswereembeddedindomestic
violenceprograms,police/systemsbasedprograms,and,inlateryears,thesexual
assaultnurseexaminerprogram.Therewasnotastrongpushbackonbehalfof
survivorsregardingcommonpracticesinsexualassaultinvestigations.
Therefore,increasedfundingforcommunitybasedadvocacyservicesisessential.State
andnationaltechnicalserviceprovidersshouldconsiderhowtostrengthenadvocacy
servicesincommunitiesthatdonothaverapecrisiscenterstoensurethattheneedsof
sexualassaultsurvivorsareadequatelyaddressed.
7) Sexualassaultnurseexaminer(SANE)programscanbeanimportantresourcetothelegal
communityandcanhelppreventtheaccumulationofunsubmittedSAKs.
InDetroit,therateofSAKsubmissionssignificantlyincreasedaftertheestablishmentof
aSANEprogram.Historically,ERphysicianshadbeentellingpolicethatSAKswerenot
useful,buttheforensicnurseswereabletocounterthatpositionandofferthe
communityeducationandresourcesonthevalueofforensicevidence.
Therefore,theseresultshighlighttheimportanceofSANEprogramsforpostassault
care,forensicevidencecollection,andcommunitycollaboration.Forcommunitiesthat
donothaveSANEprograms,theseresultsemphasizetheneedforthecreationof
strong,sustainableprograms.
302
Goal 3: Develop and Evaluate a SAK Testing Plan

KeyFindingsfromtheSAKTestingResults.Atthebeginningofthisactionresearchproject,
DetroitdidnothavesufficientfundingtotestallpreviouslyunsubmittedSAKsinpoliceproperty.
However,stakeholdersdisagreedastowhetherallSAKsshouldbetested,eveniffundswereavailable.
IsitusefultotestaSAKiftheassailantisalreadyknown(nonstranger)?Doesitmakesensetotesta
SAKifthecaseisbeyondthestatuteoflimitations?Thesesamequestionswereraisedinourresearch
interviewswithstateandnationalstakeholders,sotheDetroitSAKTestingplanwasdesignedtogather
dataaboutthesefundamentalquestions.PoolingfundsfromtheDetroitSAKARPbudget,thestate
policedepartmentsNIJDNABacklogReductionGrants,andtheresourcesofauniversitybasedforensic
laboratory(whichwasseparatelyfundedbyNIJ),theprojectcouldtest1,600SAKs(1,595actually
tested).ThekitswereplacedintofourTestingGroups,eachonedesignedtoaddressspecificresearch
questionsregardingtheutilityofSAKtestingunderdifferentcasecircumstances.Thisdesignallowedus
toexaminetheutilityofSAKtestingforstrangerperpetratedsexualassaults(TestingGroup1),non
strangerperpetratedsexualassaults(TestingGroup2),andsexualassaultcasesthatarepresumedtobe
beyondthestatuteoflimitations(TestingGroup3).Wealsostudiedwhetheremergingtechnologiesfor
DNAtesting(e.g.,selectivedegradation)couldofferfaster,lessexpensivetestingoptions,without
sacrificingaccuracy.SAKsinTestingGroup4wererandomlyassignedtotestingwithtraditionalDNA
methodsversustheselectivedegradationmethod.Whereastherearemanywaystodefinetheutility
ofSAKtesting,inthisprojectwefocusedonforensictestingoutcomes.ThefourTestingGroupswere
comparedontheirratesofCODISentries,CODIShits,andserialsexualassaulthits.
Turningfirsttoourdescriptivefindingsfortheoverallsample,the1,595SAKsyielded785CODIS
eligibleprofiles(49%oftheSAKstested),455CODIShits(28.5%oftheSAKstested;58%oftheprofiles
entered),and127serialsexualassaults(8%oftheSAKstested;28%oftheCODIShits).Itisdifficultto
knowwhethertheseratesaretypicalbecausethepublishedliteratureonSAKtestingisextremely
303
sparse.Todate,thereareonlytwootherprojectsthathavereportedSAKtestingresultsinsufficient
detailtomaketentativecomparisonstotheresultsinDetroitevenpossible:Petersonandcolleagues
(2012)studyofasampleof1,320SAKstestedinLosAngelesandNelsons(2013)reportof830SAKs
testedinNewOrleans.Thefindingsacrosstheseprojects(Detroit,LosAngeles,andNewOrleans)are
summarizedinTable6.1(below).
TABLE6.1SAKCODISEntry&CODISHitRates:ResultsfromDetroit,LosAngeles,andNewOrleans
Project
NumberofSAKsTested&SamplingDesign
NumberofCODISEligible
Profiles&Unweighted
CODISEntryRate
NumberofCODIS
Hits&Unweighted
CODISHitRate*
Detroit
1,595
785
455
Randomlysampled(withstratificationbyVictim
OffenderRelationshipandSOLExpiration)from
8,391untested/unsubmitted(400ProjectSAKS)
49%
58%
LosAngeles
(Petersonetal.2012)
1320
699
347
Randomlysampledfrom10,895
backlogged/untestedSAKs
53%
50%
NewOrleans
(Nelson,2013)
830
256
83
Samplingdesignnotspecified;includesoldcases
andcurrentcases
31%
32%
*NeithertheLosAngelesnorNewOrleansstudiesreportedratesofserialsexualassault.
Thesamplingdesignsacrossthesethreestudiesaremarkedlydifferent,socomparisonsmustbe
consideredwiththatkeycaveatinmind.TheLosAngelesdataweredrawnfromasampleof10,895
backlogged/untestedSAKs,whichPetersonetal.(2012)definedas:kithadbeencollectedinthe
courseofaninvestigationandwentuntestedbythecrimelaboratory...Thegreatmajorityofbacklogged
caseshadnotbeensubmittedtotheirrespectivecrimelaboratory,butwasbeingheldinalaw
enforcementpropertystoragefacilityand/orfreezer(p.36).TheDetroitdataweredrawnfroma
sampleofunsubmitted/untestedSAKs(nobackloggedcases);giventhatthemajorityoftheLosAngeles
SAKswerealsounsubmitted/untested,acomparisonbetweenthesetwoprojectsisreasonably
informative,thoughthesamplingdesignsarenotequivalent.AsshowninTable6.1,DetroitandLos
304
AngeleshadsimilarunweightedratesofCODISentryandCODIShits.Bycontrast,theNewOrleans
projectincludedbotholdcasesandcurrentcases(submissionandadjudicationstatusunknown),
andtheunweightedratesinthatstudyarelowerthantheresultsfrombothDetroitandLosAngeles.
However,becausethesamplingframefortheNewOrleansprojectisfundamentallydifferent(i.e.,
currentcaseswereincluded),comparisonstoeitherDetroitorLosAngelesarenotasinformative.
OneotheroptionforcomparisoninthepublishedliteratureisRomanandcolleagues(2009)
studyofCODISentryratesandCODIShitratesforburglarycrimes.Inthatproject,1,079burglarycases
inwhichDNAevidencehadbeenobtainedatthecrimesceneweretested,whichyieldeda55%CODIS
entryrate(SAKsinDetroit=49%,LosAngeles=53%,NewOrleans=31%)and43%CODIShitrate(SAKs
inDetroit=58%,LosAngeles=50%,NewOrleans=32%).Propertycrimesdifferfromviolentcrimes
againstapersoninsomanywaysthatcomparisonstothisstudyaremoretenuous,butattheveryleast,
CODISentryratesappeartobesimilarforSAKevidenceandpropertyevidence(exceptNewOrleans).
Basedontheselimitedoptionsforcontextandcomparison,itisdifficulttoknowhowbestto
characterizetheresultsinDetroit,otherthantonotethatunweightedCODISentryratesappearsimilar
tootherstudies,andunweightedCODIShitratesweresimilartotheoneotherstudymostlikeDetroits
withrespecttopopulationandsamplingdesign.Fromapracticepointofview,455CODIShitsmeans
thatlawenforcementpersonnelandprosecutorshad455newcases,127ofwhichwereserialrapes.
Fromthatvantagepoint,455CODIShitsand127serialsexualassaultsisquitehigh,particularlyfora
communityasresourcedepletedasDetroithasbeenandcontinuestobe.
TheDetroitSAKTestingPlanalsoallowedustoexaminetheutilityofSAKtestingunderdifferent
casecircumstances(e.g.,strangervs.nonstrangerrape).Ourfindingsbearoutwhatweheardfrom
manystakeholdersatthelocal,state,andnationallevel:testingcanbeusefulinavarietyofcase
circumstances,notjustinthemostcommonlyconsideredscenarioofstrangerrape.Forexample,
TestingGroup1wasasampleentirelyofstrangerperpetratedassaults,andfromthattesting,there
305
were104offenderhitsand42offenderforensichits,meaningthatin146cases,theidentityofthe
assailantmayhavebeensolved(pendingfurtherinvestigation)byDNAtesting.Fiftyone(51)ofthe
CODIShitsinthisgroupwereserialsexualassaults,meaningtherewasDNAevidenceofapriorsexual
assault.TestingGroup2wasasampleofentirelynonstrangerperpetratedassaultsandfromthat
testing,therewere103totalCODIShits,mostofwhichwereoffenderhits.PractitionersinDetroit(as
wellasotherstakeholdersatthestateandnationallevel)notedthattestingnonstrangerSAKscould
helpsolveothercrimesandidentifypatternsofserialrape.Thedatafromthisprojectprovidesome
empiricalsupportforthoseassertions:therewerefiveforensichitsinthisTestingGroup,meaningthat
testingknownoffenderSAKsrevealedtheprobableidentityoftheoffenderinfivepreviouslyunsolved
crimes.Therewere18serialsexualassaultsinthisTestingGroup,andmostoftheoffenderforensichits
inthisgroupwereduetocasetocaseassociationswithotherSAKstestedinthisactionresearch
project.Thesedatasuggestthatwhentheseoffenderswerenotapprehendedforsexualassaultsthey
committed(perhapsbecausethekitshadnotbeentested),theywentontorapeothervictims.
TestingGroup3wasasampleofolderSAKs,presumedtobeSOLexpired.Therewere90
CODIShitsinthisgroup(mostwereoffenderhits)and29serialsexualassaultswereidentifiedthrough
DNAtesting.Whetherthecasesassociatedwiththesehitscanbeprosecutedisstilltobedetermined,
butataminimum,theseresultshighlightthatoffendersoftenhavelonghistoriesofcrime,including
repeatedsexualassaults.Testingtheseolderkitsgivesprosecutorsanopportunitytopresentevidence
ofthesepriorcrimes,intheeventthereisacasethatcanbechargedwithinthestatuteoflimitations.97
Withthesedescriptivefindingsestablished,wethentestedaseriesofstatisticalmodelsthat
comparedtheprobabilitiesofCODISentryrates,CODIShitrates,andserialsexualassaulthitrates,
accountingforthesamplingdifferencesbetweenthefourGroups.Usingcontinuationratiomodels(see
97
TheprimarypurposeofTestingGroup4wastocomparetwoDNAtestingmethods;thedescriptiveresultsfromthisgroup
arepresentedinChapter4:Developing&EvaluatingaSAKTestingPlan,butarenotrepeatedhereinthissummary.
306
AppendixB:ProjectMethodology),theresultsfromtheconditionalandunconditionalprobabilityrate
analysesshowednosignificantdifferenceinCODIShitratesasafunctionofeithervictimoffender
relationshiporSOLstatus.Inotherwords,SAKsassociatedwithcasesthatwerestrangerperpetrated
hadstatisticallyequivalentCODIShitratesascasesperpetratedbynonstrangers;similarly,ratesdidnot
significantlydifferbystatuteoflimitationsstatus.SomestakeholdersinDetroit(aswellasthoseatthe
stateandnationallevel)advocatedforprioritizingSAKsbyvictimoffenderrelationshipand/orskipping
overSAKsassociatedwithcasesthatarepresumedtobebeyondthestatuteoflimitations,butthese
resultsdonotsupportsuchaplanbecausetheratesofCODIShitsdonotsignificantlydifferasafunction
ofthesevariables.Putanotherway,theseresultsindicatethatthereismeritintestingbothstranger
andnonstrangerSAKs,andpresumedSOLexpiredandnonexpiredSAKs,intermsofexpectedyields
forCODISentries,CODIShits,andidentificationofserialsexualassaults.
Ouranalysesregardingwhetherotherfeaturesofthevictim,assailant,and/orassaultmayaffect
CODIShitratesdidnotfindmanysignificantpredictors.TherewerenodifferencesinCODIShitratesas
afunctionofvictimorassailantdemographics,whetherthereweremultipleperpetrators(i.e.,gang
rape),whetherdrugsand/oralcoholwereinvolved,orwhethertheassailantusedphysicalforce.The
rateofCODIShitsalsodidnotdifferdependingonwhentheSAKwascollectedpostassault(i.e.,the
timebetweentheassaultandthemedicalforensicexam).Theonlyvariablethatpredictedwhethera
SAKwouldyieldaCODIShitisweaponuse:iftheassailantusedaweaponintheassault,theSAKwas
morelikelytoyieldaCODIShit.Giventhatnearlyallofthevariablesinthesepredictionmodelswere
notstatisticallysignificantandtherangeofplausibleeffectsizesassociatedwitheachpredictorwas
ratherbroad,weconcludethatthereisnotsufficientevidencetoadvocateforselectivelytestingSAKs
basedonvictimage,assailantage,examtiming,oranyoftheassaultcharacteristicsweexamined.
IntheDetroitSAKTestingPlan,thefourthTestingGroupinvolvedanexperimentaldesign
comparingforensicoutcomesfortwodifferentmethodsofDNAtesting:traditionalvs.selective
307
degradation.Fortheseanalyses,wefocusedonratesofDNAtestingandCODISentryratesonly.The
resultsindicatednosignificancebetweenthetwogroups,indicatingthattheselectivedegradation
methodhadnodecrementinperformancerelativetocustomarymethods.Comparisonsofmaterials
costswerealsoequivalentacrossthetwogroups,buttheselectivedegradationmethodsaved1.10
hoursofstafftime,perSAK.Itisimportanttonotethatthisexperimentwasconductedwithout
automatedtesting,andgiventhatselectivedegradationissuitableforautomation,thesetimesavings
couldbeevengreater.Thesedatasuggestthatselectivedegradationisapromisingmethodthatcould
offerforensiclaboratoriessignificantpersonnelsavings,withoutsacrificingoutcomes.Theseresults
meritreplicationinotherlabs/settingspriortobroadbasedimplementation.
TherearethreekeylimitationsaboutourSAKtestinganalysesthatmustbenoted.First,as
notedpreviously,ourfocaloutcomewasCODIShits,whichisoneofmanypossiblewaystodefinethe
utilityofSAKtesting.Assuch,thisstudycannotaddresshowSAKtestingmightaffectinvestigationand
prosecutionoutcomes.WealsodonotknowwhatitmeanttothesesurvivorstohavetheirSAKsfinally
testedandhowtheoutcomesofthattestingaffectedtheirhealth,wellbeing,andrecovery.Second,
missingdatawereaproblemforsomeanalyses.Wehadnomissingdataonthekeydependentvariable
(forensictestingoutcomes),butbecausesomeentirepolicefilesweremissingandotherswereso
incompletewithrespecttoinformationabouttheassault,thenumberofindependentvariableswe
couldconsiderwaslimited.Thelevelofdetail/nuancethatwecouldcaptureinourcodingwasalso
constrained;forexample,ourassessmentofdrug/alcoholuseintheassaultwasacompositecodeof
whethertheassailantorvictimusedeitheralcoholordrugsinthegeneraltimeframeoftheassault.
Priorresearchsuggeststhatitisimportanttoparseoutwhatsubstanceswereused,bywhom,and
when(relativetowhentheassaultoccurred)(seeBrecklin&Ullman,2001;Loganetal.,2007;Ullman&
Brecklin,2002).Wewerenotabletoassessthesedistinctionsandsowedonotknowwhether,with
bettermeasurement,alcohol/drugusewouldpredictCODIShits.Finally,ouranalysesinTestingGroup
308
4regardingcostsformaterialsandstafftimeshouldbeinterpretedwithcaution,giventhatthesedata
werenotcollectedinthelevelofdetailwewouldhavepreferred.Wehadtocompareaggregateddata
acrossthetwogroups,whichiscertainlyareasonableanalyticapproach,giventheconstraintswefaced,
butweacknowledgethatthisisnotasrefinedananalysisasispreferredincosteffectivenessstudies
(Boardman,Greenberg,Vining,&Weiner,2010;Levin&McEwan,2001).Thislimitationisakeyreason
whywerecommendreplicationoftheTestingGroup4analysespriortolargescaleimplementation.
LessonsLearnedfromtheSAKTestingResultsandImplicationsforPolicyandPractice.Asthe
collaborativewascreatingandimplementingtheDetroitSAKTestingPlan,theresearch/evaluationteam
trackedkeychallengesencountered(seeFigure4.1TheStepByStepProcessofCreatingtheDetroit
SAKARPTestingPlanandFigure4.3Challenges&SolutionsEncounteredImplementingtheDetroitSAK
TestingPlan)andlessonslearned.Figure6.2(followingpages)summarizeskeytakehomelessonsfor
otherjurisdictionsthatneedtocreateatestingplanforunsubmittedSAKs.Aswithallother
componentsofthisproject,multidisciplinaryinputwasessential,asrapekittestingcanhavedifferential
meaningandsignificanceacrossthedisciplines(e.g.,police,prosecution,forensicsciences)andtothe
survivorsthemselves.Forexample,akitthatdoesnotyieldaDNAprofilesuitableforCODISuploadis
placedasidebytheforensicscientistandhis/herworkiscomplete,butthattestingresultmaymean
thatpolicehavenonewleadstopursueandthesurvivorstillhasnoclosure.Havingalldisciplines
involvedintheSAKtestingprocesshelpsthegroupreflectonwhatdifferenttestingoutcomesmay
meanforthecriminaljusticesystemanditsworkofholdingperpetratorsaccountableandforsurvivors
andtheirrecovery.Itisalsoimportanttoemphasizethatalthoughthestatisticalanalysesofthisproject
suggestthatthereismeritintestingSAKsacrossawidevarietyofcasecircumstances,wedonotknow
whatproblemsandchallengesmightbeencounteredwhenimplementingaforklifttestingapproach.
Additionalresearchisneededtounderstandthebenefitsandchallengesofsuchatestingplan.
309
FIGURE6.2
Lessons Learned:
Developing a SAK Testing Plan
their experiences developing and evaluating a SAK testing plan.
Where do you start? How do you eat an elephant? One bite at a time.

If a multidisciplinary team was formed to plan & execute the SAK census,
then those same individuals/ organizations are well-positioned to guide the development
of a testing plan. If the census was completed without the multidisciplinary team, then
forming one for the testing phase is paramount. SAK testing raises complex legal,
psychological, and evidentiary issues; representatives from police, prosecution, forensic
sciences, medical/nursing, systems-based advocacy, and community-based advocacy,
help ensure that diverse perspectives are considered.
2.
Discuss the purpose & utility of

SAK testing
Explore how different team members think
about the purpose and value of SAK testing.
Its likely that these opinions will be deeplyrooted in their profession & their disciplines
roles & responsibilities to society. It is not
necessary to come to complete agreement
on all issues; the team may agree to
disagree on some issues and still move
forward.
OPINIONS MIGHT INCLUDE:

Testing is most useful in stranger assault cases.
Testing is less useful in non-stranger cases because
the identity of the assailant is already known.
Testing can be useful in non-stranger cases to
identify patterns of serial non-stranger assaults.
Cases that are likely SOL-expired should not be
tested to conserve limited testing resources.
Cases that are likely SOL-expired should be tested
in the event a CODIS hit links to a current case.
310
3. Test all SAKs vs. test some SAKs

The decision whether to test all SAKs or some SAKs will be influenced by both values (i.e., whether
team members believe all kits should be tested, see above) and by practical matters (i.e., funds
available to test SAKs). Testing all kits at once, often referred to as the forklift approach, is often not
feasible. The Start Small recommendation can likely be helpful for communities in which testing of
all kits is ideal but not practical.
4. Funding &
resource availability
How many kits can be tested in the

immediate future will be determined by
current resource availability. However,
developing a long-term testing plan
consistent with the jurisdictions ultimate
decision regarding how many kits should be
testedis important if current resources are
not commensurate with that aim. It is quite
likely that jurisdictions will need to apply for
grants (e.g., federal grants, such as NIJs DNA
Backlog Reduction Grants; local/state
foundation grants) and/or engage in
fundraising to secure more resources for
testing.
5. What should we call it?:

Talking about language
Unless testing all kits, teams will have to decide
which kits will be tested and in what general
order. Here, language matters a great deal as
words like, prioritize, triage, select, tier,
sample, etc. have different connotations. For
example, the word prioritize might imply that
kits will be processed in a particular order that
is based on their inherent value. Have an
explicit conversation about these issues to
avoid conflict later.
THESAURUS
DICTIONARY
6. Develop a process for selecting which SAKs will be tested

If it is not possible to send all SAKs for testing at once, then a process must be
developed for selecting which kits will be tested and in what general order.
Three main strategies include:
Select SAKs randomly (this approach could be good when starting small)
Select SAKs after a thorough review of all case material
Select SAKs based on a shorter list of selection criteria (i.e., information readily
available and accessible to speedy decision making), such as SOL expiration
311
7. Determine the specific criteria for selecting SAKs
Whether SAKs will be selected after a thorough case review or by

shorter selection criteria, detailed decision rules must be created
that specify the circumstances under which a SAK will be selected
for testing.
8. Considerations for SOL as selection criteria

Statutes of limitations (SOLs) often vary as a function of the nature of the crime.
As such, there could be multiple SOL-risk cut-off dates. The extent to which a
jurisdiction can employ more sensitive criteria (e.g., if [this] and [that] then
selection date is ____) or whether they will have to use a general across-theboard date (that should work for most cases) likely depends on the number of
cases to be screened and the resources available for screening. It is also crucial
to budget for the time that it will take to test the kit and to have the testing
results reviewed/uploaded into CODIS, etc.
9. Budget sufficient time and

resources for selecting SAKs
Starting small can help develop estimates of
how long it will take to identify cases for
selection; the time needed for this process will
likely be based on the selection criteria. For
perspective: The 1,600 SAKs tested in this
research project were selected based on three
criteria (adjudication status, victim-offender
relationship, and statute of limitations)& it took
approximately 2,958 staffing hours to review
materials and determine case selection
eligibility for these SAKs.
10. Budget extra time for older kits

Very old kits may require extra time to
prepare for testing/shipping due to peeling
labels, missing labels, re-sealing, repackaging, etc. Forensic science staff may
need extra time to review older kits and
address any problems that need to be
resolved before the laboratory can accept
the kit for testing.
Refer back to the Lessons Learned:

Developing a Census document
for reminders on how to Start Small,
Touch It Once, Develop a Central
Database, and Support Staff &
Volunteers. All of these lessons are
also important for develop testing
processes.
312
11. Track &

share testing
results
It is helpful to track the testing
results and share those results with
the full multidisciplinary team.
Case-specific results may not be
appropriate to share widely (e.g.,
in case X, victim name Y, we
found . . .). However, aggregate
data may be quite useful to the
group to track CODIS hits and the
nature of those hits (e.g., case-tocase serial offenders).
12. What happens after testing?

Devote appropriate attention to developing a plan for
what happens after testing; starting small will likely help
with this decision. The following are some key issues to
consider:
Who should be informed re: testing results?
How will post-testing investigations be coordinated?
How will case-to-case CODIS hits be handled?
How will current caseloads be handled with these
new/old cases being re-opened?
Can a flexible process be developed to respond to highly
time-sensitive cases?
13. When testing results start coming in, expect the unexpected
Given the dearth of empirical research on untested SAKs, it is difficult to know whether testing
results are typical or atypical. It might be helpful for jurisdictions to connect with other
communities who have tackled these issues to compare findings and strategize solutions.
14. Re-examine & refine testing policies & protocols

While reviewing existing SAK testing procedures may cause defensiveness at times, it is also
possible that jurisdictions will want to make immediate changes to their testing policies.
Regardless of the reaction, it is important to revisit the policies regularly as new
information/insights will develop throughout the course of resolving the previously-untested
SAKs. Take special care to revise selection criteria as needed as criteria may not be as clearcut or easy to enforce as originally conceived.
15. Consider whether legislative changes are necessary

The process will very likely suggest legislative changes that might be necessary to
remedy problems, including, but not limited to: requirements for mandatory kit
submissions and timelines for submissions and testing; procedures for retaining kits
before and after testing; procedures for handling kits if victims are unsure about
possible involvement with the criminal justice system; and tracking mechanisms for
identifying where a kit is in the process of submission/testing.
313
TheresultsfromtheDetroitSAKtestingplanhaveseveralimplicationsforpolicyandpracticein
regardstotestinglargenumbersofunsubmittedSAKs:
1) ItmaybeusefultoselectarelativelysmallrandomsampleofSAKs,testallofthosekits,
andusethoseresultstohelpplanalongtermtestingplan.
IntheDetroitSAKARP,wehadthebenefitoftheOVWfundedThe400Project,which
randomlyselectedandtested400SAKs.Fromthosedata,wewereabletoestimatethe
percentageofSAKsthathadbeenpreviouslytestedandadjudicated.Thedatawere
alsohelpfulinidentifyingchallengesinpreparingkitsforshipment,workingwithvendor
laboratories,andtrackingdownpolicefiles,medicalrecords,andotherdocumentation.
Therefore,communitiesthathavelargenumbersofSAKsmaywishtoconsiderapilot
projectsimilartoThe400ProjecttogetafeelforthevolumeofCODIShitstheymay
needtoprepareforandthestaffingtheymayultimatelyneedtoaddressthisproblem.
Stateandnationalfundersmaywishtoconsiderdevelopingpilotgrantprojects(with
accompanyingtechnicalassistance)toencouragecommunitiestodevelopand
implementthesestarterscaleprojects.
2) TheresultsoftheactionresearchprojectsuggestthatscreeningSAKsfortestingbasedon
informationprovidedintheoriginalpolicereportsmaynotbeusefulandinfactmay
underestimatetheseriousnessoftheassaults/offenders(aswellasunjustifiablydiscount
thecredibilityofthevictimsreports).
IntheDetroitSAKARP,wefoundthatpolicereportsoftenreflectedofficersstereotypic
beliefsaboutrapevictims,andfewinvestigativestepsweretakeninthecase.Iftesting
decisionshadbeenbasedontheofficersassessmentsofthemeritsofthecaseas
reflectedinthoserecords,thenitislikelymanySAKswouldnothavebeensubmitted
fortesting(because,forexample,thedetectivenotedthatitwasntreallyarape).
However,therateofCODIShitsandserialsexualassaultsdocumentedinthisproject
indicatesthatsuchassessmentswerelikelybiasedandtestingthosekitswaswarranted.
Therefore,itisnotrecommendedthatcommunitiesbasescreeningandprioritizationon
theoriginaldocumentationinthepolicereport.Itmaybemorefruitfultothe
314
investigationtotestallSAKsandallowtheforensicoutcomestoguidenextstepsfor
investigation,prosecution,andvictimnotification.
3) TheresultsoftheactionresearchprojectsuggestthatscreeningSAKsfortestingbasedon
selectioncriteriasuchasvictimoffenderrelationshipandstatuteoflimitationsmaynot
behelpfulvisvismaximizingCODIShits.
IntheDetroitSAKARP,weexaminedwhetherprioritizingSAKsbyvictimoffender
relationshiporstatuteoflimitationsstatusmightbehelpfulinmakingthemostof
limitedfinancialresources(i.e.,tryingtomaximizethenumberofCODIShitsperkits
tested).WedidnotfindsignificantdifferencesinCODIShitratesasafunctionofthese
screeningcriteria.
Therefore,ifcommunitieshavelimitedfinancialresourcesandaretryingtoprioritize
SAKsfortesting,thesedataindicatethatselectingonthebasisofvictimoffender
relationshiporSOLstatusmaynothaveaneffectonCODIShitrates.Communitiesmay
wishtogivespecialconsiderationtoSAKsassociatedwithcasesthatarenearthe
statuteoflimitationsexpiration,buttoskipoverpresumedSOLexpiredcasesisnot
empiricallysupported,basedontheresultsofthisproject.Ourfindingssuggestthat
thereismeritintestingallSAKs.
4) TheresultsofthisactionresearchprojecthighlighttheutilityoftestingSAKsacrossa
varietyofcasecircumstances,butwedonotknowwhatproblemsorchallenges
communitiesmayfacewithaforklift/testallplan.
TheDetroitSAKARPdidnotevaluatetheutilityofaforkliftapproach;thisproject
evaluatedwhethertherearedifferencesinforensicoutcomesasafunctionofdifferent
screeningvariables(therearenot).
Therefore,additionalresearchisneededincommunitiesthatadoptaforklift
approachinordertodocumentthechallenges,solutions,andforensictestingoutcomes
associatedwiththatparticulartestingstrategy.
315
Goal 4: Develop and Evaluate a SAK Victim Notification Protocol

KeyFindingsfromtheEvaluationoftheVictimNotificationProtocol.Thetaskofcreatingand
implementingavictimnotificationprotocolwasperhapsthemostdauntingandhumblingchallenge
theDetroitcollaborativefacedinthisactionresearchproject.Inearlyteamconversationsaboutvictim
notification,representativesfrompoliceandprosecutionemphasizedthatthiswasnotsomethingnew
forthemtheyroutinelyhadtorecontactcrimevictimswhennewinformationbecameavailablein
theircasesandtoascertaintheirwillingnesstoreengagewiththecriminaljusticesystem.However,the
circumstancesunderwhichthesenotificationswouldoccurgaveallmembersofthecollaborativepause,
asonememberoftheteamexplained:
Thisisntanycrime,itsrape...[whichis]invasiveanddehumanizing...andthereasonwhy
werecontactingthem(thesurvivors)isdifferenttoo...somethingthatshouldhavebeendone
before,wasnt...thesekitswerenttestedandthewomenthoughttheywouldbe...worse,
[manyofthesesurvivors]weretold[bythepolice]thatitwasjustadealgonebad...Thatsthe
contexthereandcontextiseverything.
TheDetroitcollaborativehadatwodayplanningretreattodevelopavictimcentered,trauma
informednotificationprotocol(seeAppendixC1:VictimNotificationRetreatPlanningGuideandFigure
5.1TheStepByStepProcessofCreatingtheDetroitVictimNotificationProtocol).Amultidisciplinary
teamwouldreviewcasesthathadCODIShitsanddiscussifandhowtonotifysurvivors,giventhe
uniquecircumstancesofeachcase.Thenotificationswouldproceedinatwostageprocess,whereby
thegoalsofthefirstcontactweretoexplainthattheSAKhadnotbeentestedatthetime,butnowit
hadbeentested;offeranapologytothesurvivorthather/hisSAKhadnotbeentested;andrequesta
followupmeetingtodiscusstheissuesinmoredetail,afterthesurvivorhadhadachancetocatchher
316
breath.Atthefollowupmeeting,aninvestigatorandcommunitybasedadvocatewouldprovidemore
detailedinformation,discussoptions,andconnectsurvivorstocommunityservices.
Intheevaluationofthispilotprotocol,41caseswereselectedforvictimnotification,andthe
investigatorswereabletofind31survivors(2caseswereclosedoutbecausetheinvestigatorshad
exhaustedallpossibleleadstryingtofindthevictims;8caseswerestillpendingatthetimethe
evaluationdatacollectionperiodclosed).Stakeholdersatthelocal,state,andnationallevelexpressed
concernanddoubtaboutwhetheritwouldbepossibletoreconnectwithsurvivorssomanyyears
laterandtheamountoftimeandeffortitwouldtaketodoso.However,inthisevaluation,wefound
thatmostsurvivors(65%)couldbefoundwithrelativelylowinvestigationaleffort:searchesof
computerizeddatabases(e.g.,LEIN),plus04phonecalls,and01inpersonvisitsto1address.However,
toobtainthehighfindrateinthisproject,theinvestigatorshadtoinvestadditionaleffortforthe
remaining35%ofthecases,whichsometimesrequired12+phonecallsand6+inpersonvisitsto
variouslocationstryingtofindthevictim.Survivorswhowerehardertofindwantedtoparticipateinthe
prosecutionoftheircasesinacomparableratetothosewhowereeasiertofind,sothesefindings
suggesttheextraeffortwaswarranted.Assuch,thisevaluationsuggeststhatvictimslocateability
shouldnotbeaselectioncriterionforeitherSAKtestingorvictimnotification.
Thefirstcontactwiththesurvivorswasmadebyinvestigators(affiliatedwiththeprosecutors
office,notthefocalpolicedepartment)andtypicallyoccurredatthevictimshomes.Somesurvivors
hadstrongnegativereactions(16%)(e.g.,anger,refusaltotalktoinvestigators),morehadstrong
positivereactions(29%)(e.g.,happiness,relief),andmost(55%)didnotexhibitastrongemotional
reactiontheywereopentohearingwhattheinvestigatorshadtosay,butwerereservedandcautious.
Mostsurvivors(64%)agreedtoafollowupmeetingwiththeinvestigatorsandanadvocatetodiscuss
optionsinmoredetail,andintheend,most(57%)alsodecidedthattheywantedtoparticipatein
furtherinvestigationandprosecutionofthecases.Thisrateofreengagementishigherthanexpected,
317
giventhesignificantsecondaryvictimization(victimblamingtreatment)survivorsexperiencedfrom
lawenforcementpersonnelatthetimetheymadetheiroriginalreport(seeChapter3:WhySoMany
UnsubmittedSAKsinDetroit).Priorstudieshaveshownthatmostvictimswhoexperiencesecondary
victimizationfromthelegalsystemaredisinclinedtoseekanyfurtherhelp(80%onaveragedonotwant
anyfurthercontact/help)(seeCampbell,2008forareview),sothefactthat57%didwantfurther
contactis,inourview,anencouragingfinding.
Tounderstandwhysomevictimsmayhavehadnegativereactionsanddecidednottore
engage,theevaluationteamexaminedwhethertheseresponsesvariedasfunctionofvictimage,time
sinceassault,andvictimoffenderrelationship.Victimswerelesslikelytoreactpositivelyandtore
engagethelongerthetimebetweentheassaultandthenotification,whichhighlightstheimportanceof
timely(i.e.,atthetimetheassault)testingofSAKsandinvestigationofreportedsexualassaults.
Survivorswhowere1624yearsoldatthetimeoftheassaultweresomewhatmorelikelytohavestrong
negativereactionstothenotificationandwerelesslikelytowanttohavecontinuedcontactwiththe
criminaljusticesystem.Giventhatpriorresearchhasfoundthatvictimsinthisagegroupareat
particularlyhighriskforsecondaryvictimization(Campbelletal.,2012;Greeson,Campbell,&Fehler
Cabral,2014a,2014b),whichwasalsosubstantiatedinthisactionresearchproject(seeChapter3:Why
SoManyUnsubmittedSAKsinDetroit),itseemsprobablethatthesegirls/youngwomenhadhaddifficult
encountersyearsago,andassuch,theyweredisinclinedtoreengage.Inthisevaluation,onlyasmall
numberofnotificationshadbeenconductedwithvictimsofnonstrangerrape,butpreliminaryfindings
suggestedthattheywerenotaslikelyasvictimsofstrangerrapetocontinuecontactwiththecriminal
justicesystempostnotification.
Becauseakeymethodologicallimitationinourevaluationistheabsenceofdirectdata
collection,wedonotknowtheexactreasonswhyvictimsmayormaynothavewantedtoengage
further(proxydatawerecollectedfromtheinvestigatorsandadvocates,seeAppendixB:Project
318
MethodologyandAppendixD:DataCollectionInstruments).Asnotedpreviously,therewerestrong
concernsamongtheprosecutorsthatiftheresearchershadcontactwithsurvivorspreadjudication,we
couldbecalledaswitnessesinthecaseandourrefusaltotestify(perIRBrequirements)couldhavehad
anunintendednegativeeffectonthecase(e.g.,defensecounselsuggestingthatwewerehiding
something).TheethicalprinciplesoftheAmericanEvaluationAssociation(1995/2004)statethat
evaluatorsmustconsidernotonlyimmediateoperationsandoutcomesoftheevaluation,butalsothe
broadassumptions,implications,andpotentialsideeffects(ResponsibilitiesforGeneralandPublic
WelfarePrinciple)(emphasesadded).Assuch,wehadtopauseandcarefullyconsiderhowbestto
designtheevaluationofthisprotocol.
Weturnedtotheliteraturetoseeiftherewasprecedentforinterviewingrapesurvivorspre
adjudication.Therewerenumerousexamplesofinterviewingsurvivorsaboutthelawenforcement
investigation(e.g.,Campbell,2005,2006;Greesonetal.,2014a,2014b;Monroeetal.,2005Patterson,
2011a,2011b),andnoneofthestudiesthatdidprospectivedatacollectionreportedthatthe
researchershadbeencalledaswitnessesinlatercourtproceedings.Farfewerprojectshavefocusedon
theprosecutionphase,andKonradis(1996a,1996b,2007)studywastheonlyoneexamplewefoundin
thepublishedliteratureinwhichinterviewswereconductedwithsurvivorspreadjudication.Again,
therewasnoindicationthattheresearcherswerecalledtotestifyorthatthestudyhadnegative
consequencesoncaseprocessingorcaseoutcomes.Therefore,basedontheseexamples,the
research/evaluationteamnotedthatalthoughbeingcalledasawitnesswaspossible,itseemed
improbableandthatunlikelihoodneededtobeweighedagainstthebenefitsofhearingdirectlyfrom
survivorssotheycouldexpresstheirthoughtsandfeelingsaboutthenotifications.Theprosecutors
certainlyagreedwiththeimportanceofgivingsurvivorsopportunitiestovoicetheirexperiencestheir
concernwasthetimingofsuchinterviewsandtheneedtoguardagainstcomplicationsinpendinglegal
casescasesthatwerebynomeanstypical,giventhattheSAKshadnotbeentestedforyearsand
319
anynumberofchallengesfromthedefensehadtobeconsidered.Interviewingsurvivorspost
adjudicationwouldstillprovideanopportunitytosharetheirexperiences,butwithouttheriskof
unintendedconsequencesonthecourtcase.However,givenhowlongtheadjudicationprocesstakes,
itwasnotpossiblewithinthetimelineofthisactionresearchprojecttointerviewsurvivors.Therefore,
theevaluationofthevictimnotificationprotocolutilizedproxysourcesofinformationandtheresearch
teamintendstoconductafollowupinterviewstudywithsurvivorsafterthecompletionoftheaction
researchprojectinordertoobtaintheirfirsthandaccounts.
LessonsLearnedAboutVictimNotificationandImplicationsforPolicyandPractice.
Throughouttheprocessofcreatingandimplementingthevictimnotificationprotocol,the
research/evaluationteamtrackedkeychallengesencountered(seeFigure5.3Challenges&Solutions
EncounteredintheImplementationoftheDetroitSAKARPVictimNotificationProtocol)andlessons
learned.Figure6.3(followingpages)summarizesimportanttakehomelessonsforotherjurisdictions
thatmaybefacedwithdevelopingavictimnotificationprotocol.Oneofthemostcriticallessons
learnedinthisprojectwastheimportanceofamultidisciplinaryperspectiveonvictimnotification,given
thatitinvolvesreactivatingmemoriesofamajortraumaticeventandcommunicatingcomplexforensic
andlegalinformation.Theinputfromallcoredisciplinespolice,prosecution,crimelab,nursing/SANE,
andadvocacywasessentialforcreatingacomprehensiveplanthataddressedsurvivorspsychological
andinformationalneeds.
320
FIGURE6.3
Lessons Learned:
Conducting Victim Notifications

their experiences developing a victim notification protocol.
How we tried to do the right thing after the right thing wasnt done a long time ago.
Notification involves complex legal,

psychological, and evidentiary issues.
While only some disciplines may be involved
in conducting the actual notification (e.g., law
enforcement, advocacy), having a broad-based
multidisciplinary team to ensure that diverse
perspectives will be considered at every stage of
the process is important.
Consider including
representatives from:
Police
Prosecution
Forensic sciences
Medical/nursing
Systems-based
advocacy
Community-based
advocacy
2. Clarify what decisions are and are not

open to multidisciplinary input
There are many ways in which the multidisciplinary review, input, and debriefing
sessions may be done (e.g., an oversight steering committee, a separate working
group, etc.) and how often it needs to be done (weekly, monthly, quarterly, ad-hoc).
However this process pans out, it is helpful to clarify which individuals/organizations
have decision making authority over what issues. While some decisions may be made
through collaborative discussion and consensus, others may be made solely
by one discipline.
However, even if decision-making authority is held by
only one individual/ organization, collaborative input can help inform

that decision.
321
3. Connect with local, state, and national

colleagues who have expertise in victim notification
Reaching out to other jurisdictions to hear about their lessons
learned and education within the group on such lessons can be
instrumental. In addition, individuals within the multidisciplinary
team may already have experience doing notifications as part of
routine CODIS-hit follow-up protocols. It is important to include
those individuals in the process of developing notification protocols.
4. Allocate sufficient time

for
developing protocol
While there is no standard amount of time

needed to develop notification protocols, it will
likely be a lengthy process as many complex
decisions must be made. It might be helpful to
consider setting aside additional
meetings/retreats to discuss victim notification
in sufficient depth.
5. Expect disagreements
among stakeholders
Team members are likely to have
different opinions as to when, how, and
why victims should be notified about
SAK testing. Consider using an outside
facilitator, like the Detroit SAK ARP did,
to allow the group to come reach
consensus on complex issues
6. Acknowledge that there is no right way to do victim notification

Given the limited research to-date on the problem of unsubmitted SAKs, empirically-based
best practices for victim notification have not yet been identified.
SOME GUIDING PRINCIPLES FROM RELATED RESEARCH:
Notification could be traumatic for many victims and therefore their emotional needs
must be taken into consideration
Providing victims with options and choices helps facilitate their recovery from trauma
Providing victims with options and choices can have a positive indirect effect on their
participation in the criminal justice system
322
7. Establish protocol to
8. Be sensitive to differing rules
protect the confidentiality

of the victims who had
been notified
regarding the disclosure of

confidential information
Set clear expectations up-front regarding what can and

cannot be divulged to the multidisciplinary team. While
investigators might want to know how their notification
interactions were perceived by victims, advocates
cannot disclose their communications with survivors
(unless specifically authorized to do so by the victim). The
Detroit collaborative invited an experienced advocate
who was not directly involved in any of the cases to
provide general guidance to investigators regarding
victim notification and to help the group understand
confidentiality differences across professions.
Updates on victim notification

should be limited to those directly
involved with the victims, regardless
of what confidentiality agreements
have been signed.
It is also important to offer victims
the option of working with
practitioners with whom they can
have confidential conversations (i.e,
community based advocates).
9. Provide comprehensive training
to all staff who will be involved in

conducting victim notifications
10. Know your community,

know your population
Victim notification strategies should be
tailored to the specific needs, resources,
and realities of a community.
Consider the diversity of a community
with respect to race/ethnicity, social
class, religion, cultures, and customs as
well as the relationship that the
community historically has with police.
These training should emphasize the

importance of self-care & a victim-centered,
trauma-informed approach to working with
sexual assault survivors.
11. Start smalland be flexible

Develop draft notification protocols, implement
them with a small number of cases, evaluate
them (either formally or informally through
multidisciplinary team debriefings), make
changes, and implement/re-evaluate the
revised procedures. Because each sexual
assault case is unique, the victim
notification protocols should be
flexible to accommodate unusual
and/or urgent situations.
323
12. If multiple practitioners will be working together to conduct a victim

mm
be briefed (ahead of time) about the case
notification,
then all need to
It is important that the practitioners who will be
directly involved in the case are briefed ahead of
time so that they can be aware of key issues that
may come up during notification. At the same time,
victims privacy must also be respected, and team
members need to be mindful not to stereotype based
on preliminary information.
Consider briefing on:

testing results, good/bad
experiences the victim may
have had during the initial
investigation [if known],
whether suspect is currently
incarcerated, etc.
13. Be prepared to take action when practitioners are ready

The timing of victim notification should be considered in light of when legal
investigators, prosecutors, and/or advocates are ready and available to work
on the caseand to keep the victim informed regarding what steps are being
taken. It is important to try to avoid a delay/lag after notification so that the
victim does not need to be re-notified. If there is a delay/lag for some reason, it
is very important to keep the victim informed regarding what is happening with
her/his case.
14. Attune to safety

concerns for notifying
staff and for victims
15. Develop
Depending on how/where notification

may take place, the physical safety of the
notifying staffand victimscould be at
risk. Developing safety plans before
notification is critical. For example,
sending notifying officers out in pairs is
recommended. In addition, victim
notification will take an emotional toll on
staffand victimsso ensuring that these
needs are addressed is equally important.
resources for
victims & providers
Developing an FAQ brochure and a

packet of community resource
information will likely be useful. This
information could include: community
referrals, the name/number of a
community advocate, a 24-hour crisis line
number, and a plausible excuse/story to
tell others who might question their
interaction with police (e.g., tell them you
were a witness to another crime).
324
TheDetroitSAKARPsexperienceofdevelopingandimplementingaplanforvictimnotification,
aswellastheevaluationresultsregardingtheefficacyofthatprotocol,haveseveralimplicationsfor
policyandpracticeregardinghowtoworkwithsurvivorswhoseSAKsweretestedlongafterthemedical
forensicexamwasconductedandthepolicereportwasfiled:
1) Victimnotificationraisescomplexlegalandpsychologicalissuesforsurvivors,soa
multidisciplinaryapproachisnecessaryforboththecreationandimplementationof
victimnotificationprotocols.
InDetroit,earlyconversationsaboutvictimnotificationfocusedonhowtoconvey
complexinformationtosurvivorsaboutDNAtestingandthestepsofprosecution.
Stakeholdersfromadvocacy,nursing/SANE,andtheresearchteamencouragedthe
grouptothinkaboutthepsychologicalimpactofreopeningtraumaticmemoriesand
howthevictimnotificationexperienceasawholeshouldsupportsurvivors.
Therefore,multidisciplinarystakeholdersneedtobeinvolvedindesigninga
communitysplanforvictimnotificationsevenifonlyonediscipline(ordisciplines)will
beprimarilyinvolvedincontactingsurvivors.Multidisciplinaryperspectivesare
particularlyimportantfordetermininghowtoprotectsurvivorssafety,privacy,and
confidentiality.ExistingSART(SexualAssaultResponseTeam)resourcesneedtobe
examinedfortheextenttowhichtheyaddresstheneedsofcommunitiesthatneedto
formmultidisciplinarycollaborationstoaddressuntestedSAKs.
2) Comprehensivetrainingisnecessaryforallvictimnotificationstaffregardingthevarious
legal,forensic,psychological,andculturalissuesthatmayariseinthenotifications.
InDetroit,somestakeholdersemphasizedthatbecausetheyhaddonenotification
before,theydidnotneedadditionaltraining,butastheresultsoftheactionresearch
projectcontinuedtounfold,itbecameclearthatallpractitionersfromallorganizations
couldbenefitfromtrainingonvictimcentered,traumainformedpractices.
325
Othercommunitiesthataredevelopingvictimnotificationprotocolsmayormaynot
havetheresourcesandexpertisetodevelopempiricallybasedtrainingprograms;
therefore,stateandnationaltechnicalassistanceprovidersmaywanttoconsiderhow
todevelopanddisseminateempiricallybasedtrainingmaterials(e.g.,trainingvideos,
webinars,TAhotlines).
3) Samplevictimnotificationprotocols,trainingmaterials,andresourcematerialsneedtobe
compiledanddisseminatedsothatmultidisciplinaryteamscanreviewthesematerialsand
considerwhichstrategiesmightbemosteffectiveintheirowncommunity.However,
whatworksinonecommunitymayormaynotbeagoodchoiceinanotherjurisdiction.
InDetroit,thediscussionsattheplanningretreat(andthelaterNotificationReview
Teammeetings)consideredwhatmaybeuniquechallengesfacedinthisparticularcity,
givencharacteristicsofitsresidents,neighborhoods,etc.Thedecisionsmadeforthe
Detroitvictimnotificationprotocolreflectthecollectivewisdomofmanyindividuals
whohavelivedandworkedinthiscommunityfordecades.
Therefore,multidisciplinaryteamsneedtoconsiderifandhowtotailorvictim
notificationprotocolstofitthespecificneedsoftheircommunity(e.g.,whetherphone
notificationsmightbefeasible,whetherhavingadvocatesaccompanyinvestigatorsis
possible,etc.).Itmaybeusefultodevelopapilotprotocolthatisimplementedwitha
smallnumberofcasestoevaluatewhetherthestrategiesareeffective.Stateand
nationaltechnicalassistanceprovidersmaywanttoconsiderhowtoprovideevaluation
resourcestocommunitiessothattheycanconductsmallscale,utilizationfocused
evaluationsoftheirpilotprograms.
326
4) Findingsurvivorstakesdedicatedeffort,butforexperienced,welltrainedinvestigators,
thismaynotbeadauntingtask,asmanyvictimsmaybelocatedwithrelativelyminimal
investigationaleffort.
InDetroit,legalinvestigatorsaffiliatedwiththeprosecutorsofficeweretaskedwith
findingssurvivorsfornotification;theseindividualswerehighlyexperienceddetectives
whohadconsiderablepriorexperienceworkinginthiscommunity.
Therefore,multidisciplinaryteamsneedtocarefullyconsiderwhoisselectedto
find/locatesurvivors;highlyexperienceddetectives/retireddetectivesmaybe
particularlywellsuitedforthisroleastheiryearsofexperienceinthecommunitycanbe
instrumentalinfindingsurvivors.Trainingspecificallyonvictimcentered,trauma
informedstrategiesforworkingwithrapesurvivorswouldstillbebeneficial,aseven
experienceddetectivesmaynothavehadexposuretotheseperspectives(seeabove).
5) Survivorswillhavediversereactionsandnotificationpersonnelneedtorespectvictims
choices.
InDetroit,thevictimnotificationtrainingemphasizedhowtoworkwithindividualsin
crisis(whethertheirreactionswerestronglypositive,negative,orseeminglyneutral/in
shock).Victimsmayhavedifferentpreferencesforif,when,andhowtheywantto
discussthecaseandtheiroptions.
Therefore,theDetroitvictimnotificationprotocolemphasizedcasebycaseflexibility,
solongasthecorevictimcentered,traumainformedprincipleswererespected.
Stakeholdersmayfeelaneedtoencourageornudgevictimstoagreetoprosecute
(givenwhattheyknowaboutthecase,theperpetratoretc.),somultidisciplinaryteams
needstohaveexplicitexpectationsforhowtoaddressthesepotentialtensions.
Trainingandresourcematerialsregardingvictimnotificationneedtoexplicitlyaddress
thisissueofhowtobalancevictimschoices/wisheswithwhatserviceprovidersmay
feelisthebestcourseofactioninacase.
327
Creating a Multidisciplinary Team to Address the Problem of Unsubmitted SAKs:Overall

Project Lessons Learned and Implications
ToaccomplishthesefourgoalsoftheDetroitSAKARP,thecollaborativepartnershadtolearn
howtocometogetherandworkasateam,whichwaschallengingattimes,givenlongstandingconflicts
withinandbetweenorganizations.Onestakeholderrelayedthefollowinganecdoteasacaseinpoint
exampleofhowfracturedthegroupwasatstartoftheproject:
Whatwasitlikeinthebeginning?Well,thiswastelling:rememberwhenthewholeteamhad
toflytoDCforthekickoffmeetingwithNIJ?Ithinktherewerelikesevenofusthatwenton
thattripandwetooksevendifferentflights...thatsimpressivewhenyouthinkaboutit,notin
agoodwaythough.Imeanreally,howmanyflightsaretherefromDetroittoDC?Andyet,
everyonemanagedtobeonadifferentplane.
Neartheendoftheproject,NIJhostedanotherinpersonmeetinginDCandstakeholdersremembered
thattripfarmorefondly,asonememberoftheteamrecounted:
Iremembersittinginthefoodcourtattheairport,waitingforourflight.(Note:nearlythe
entireteamwasonthesameflight.)Allofustogether...teasingeachotheraboutwhatwed
pickedfromthefoodcourt,wholikedthegreasyChinese,whowasgoodandgotasalad.
Laughing,tellingstories,passingaroundiPhoneslookingatpicturesofourkidsandgrandkids.
Whathappenedinbetweenthosetwotrips?Howdidthegroupcometogetherandcommitto
solvingalongterm,complexproblem?Inadditiontochartingthecollaborativesprogressonthefour
keysubstantivegoalsoftheproject(summarizedabove),theresearch/evaluationteamalsotracked
overallgroupprocessinordertodocumentthechallengesworkingacrossdisciplines.Figure6.4
(followingpages)summarizesthehighlightsfromthatjourney,keylessonslearned,andadviceforother
multidisciplinaryteamstacklingtheproblemofunsubmittedSAKs.
328
FIGURE6.4
Lessons Learned:
Conducting an Action Research Project
The take-home lessons from the Detroit SAK ARP based on their
Experiences participating in a multidisciplinary action research project.
Multidisciplinary is easy to say, but hard to do.
1. There needs to be a champion
2. Forming a multidisciplinary
(or champions for the cause)
team is important, but

building a team is
even more
important
Addressing the problem of untested SAKs is a

long and arduous process, and many individuals
and/or organizations will feel pressures to make
the problem go away as quickly as possible. The
work of confronting, resolving, and then
preventing this problem from happening again is
a long-term endeavor; as such, there needs to
be a champion (or champions) of the cause. An
individual (or multiple individuals) who have
long-term standing and respect in the
community need to take on the issue and keep
key organizations and systems engaged in the
work of solving the problem.
Forming a multidisciplinary working group is

essential; however, it may not be
reasonable to expect that these
individuals/organizations have much
experienceor interestin working
together (at least at first). The results of this
action research project highlight that years
of poor systemic collaboration can be a
contributing factor to the development of
this problem. Therefore, allocating time to
learn about the jobs of each team member
and the work of each organization is critical.
Team building activitiesparticularly in the
beginning of the projectto help the group
coalesce and work together effectively is
also useful.
329
3. A project coordinator position is essential to the success of the project

Previous action research projects have shown that a designated project
coordinatorsomeone responsible for keeping the group on task,
providing day-to-day leadership, managing conflict and interpersonal
difficulties, and sustaining the collaborationis critical for the success of
the project.
Group Facilitator Options:
4. The challenges of
group facilitation
Group facilitation is a challenging task
and a multidisciplinary team may need
to try different strategies before finding
one that works. Managing
stakeholders with divergent points can
be a very challenging task. A skilled
group facilitator is needed to create a
setting in which individuals can openly
share their points of view.
The project coordinator*

An outside facilitator: a professional facilitator who
is not affiliated with any of the member
organizations
A team member who is not the project
coordinator*
*If team meetings will be facilitated by someone
inside the group, then the group needs to consider
if/how the internal facilitators can share their own
views in the discussion without biasing the process.
5. Who has decision-making authority?

The individual representatives in a multidisciplinary team may or may not have the
authority to make decisions on behalf of their organizations. As such, a teams provisional
decisions may not be enacted immediately, as it will take time for those
recommendations to be considered by senior executives. Furthermore, it is possible that
the decisions/recommendations of the team will not be approved by senior executives,
and then the team must re-evaluate their plan of action.
In a multidisciplinary team consisting of multiple organizations, there may be some
organizations that are more influential than others. This could be one single organization
that has more authority than any other (e.g., authority granted by an outside entity, by
law, by community norms, etc.). The multidisciplinary team needs to discuss these issues
explicitly so that all parties are clear as to what decisions individual members can make,
which helps clarify the extent to which the decisions of the team are advisory or binding.
330
6. New cases dont stop coming in: Balancing the workload

Communities that have large number of untested sexual assault kits will
struggle with the issue of how to attune to these old cases (many of which
will become current cases as they are re-opened and investigated and
prosecuted) while simultaneously responding to incoming cases. All
disciplinesprosecution, law enforcement, forensic sciences,
medical/nursing, and advocacywill face this challenge. Some jurisdictions
may be able to form specialized cold case units, whereby some staff focus
exclusively on these old cases and others maintain responsibility for
incoming cases.
7. Prepare for change because

change is constant
Addressing the problem of untested SAKs is a long-term endeavor, so it is to be
expected that there will be (perhaps considerable) staff turnover over the duration of
the initiative (in both upper leadership and among front-line
practitioners). Other changes in funding, resources, and community
context should also be expected. Having multidisciplinary champions
of the cause is important for weathering these changes and
keeping the community invested in the development
of long-term solutions.
8. The action research paradigm

can be confusing to practitioners
Community practitioners may not have much experience
participating in social science research projects, and if they
have, its more likely that they experienced traditional
research/evaluation (i.e., the researchers are focused primarily
on data collection). The action research paradigm is a very
different approach to research, and community partners need
more education about this kind of workbefore they agree to
partner in such an endeavor. Once involved in an ARP, on-going
education/conversation about roles and responsibilities is
important so that the collaborative has appropriate expectations
for the researchers (and vice versa).
331
Firstandforemost,thereneedstobeachampionorchampionsforthecause.Addressing
theproblemofuntestedSAKsisalongandarduousprocess,andmanyindividualsand/ororganizations
willfeelpressures(internaland/orexternal)tomaketheproblemgoawayasquicklyaspossible.An
individual(ormultipleindividuals)whohaslongtermstandingandrespectinthecommunityneedsto
takeontheissueandkeepkeyorganizationsandsystemsengagedintheworkofsolvingtheproblem.
IntheDetroitSAKARP,thereweretwokeychampionsofthecause(oneatthelocallevel,oneatthe
statelevel),andmembersofthemultidisciplinaryteamconsistentlymentionedthatthecombined
effortsofthesetwopeoplewereenormouslyinfluentialinkeepingstakeholdersengagedinlongterm
solutionstothisproblem;asonememberoftheteamnoted:
Wehadtwochampions...weneededboth,wereallydidbecauseyouwontgetanythingdone
withoutsomeonedemandingaccountabilityandaction...[bothofthem]alwaysbroughtit
backtosocialjustice,justiceforsurvivors.Whatsthebestthingtodoforsurvivors?Thatkept
usmotivatedandfocused.
Second,effectivecommunitychampionsmaybeabletobringeveryonetothetable,but
stakeholdersmaynotknoweachother,understandtheworkoftheotherorganizations,andappreciate
theperspectiveoftheirdisciplines.ThiswascertainlythecaseintheDetroitSAKARP,asonememberof
theteamnoted,Wehadtostartoffbyhavingeveryoneteacheveryoneelseontheteamaboutwhat
theydidbecausehonestly,wedidntknowwhoeachotherwasandwhatexactlytheydid.Inthefirst
fourmonthsoftheproject,theorganizationbyorganizationreportoutportionoftheteammeeting
wasa101andthena201andthena301foreveryoneelse...ok,todayfolks,wereallgoingtolearn
thestepbystepprocessofDNAtesting,howawarrantrequestworks,andsoon.Tobeclear,the
Detroitcommunitywasnotsofracturedthatstakeholderswerewhollyunawareofeachothersjobs
theydidworktogethertoprocesscurrentcasesbuttheyquicklyrealizedthattheywereoften
332
coastingalongonassumptionsandthoseassumptionswerenotalwaysaccurate.Asonememberof
thecollaborativehighlighted:
Itmightseembasic,evenawasteoftime(tohaveeveryoneexplainwhattheydo)butit
wasnt.Absolutelynot.IwouldtellanymultidisciplinaryteamworkingonSAKstodothesame
thing.Dontassumeyouunderstandwhattheothergroupsdointheirjobs.Assumptionsare
probablywhatgotyouintothemessyourein...gobacktobasicsandlearnabouteachjob.
Eachdisciplineseesthingsdifferently...youneedtounderstandthosedifferences.
Fromthatbasekeyleadershipandmultidisciplinaryunderstandingthecoredaytodaywork
hasabetterchanceofcreatingsustainablechange.Tothatend,theDetroitcollaborativeteam
membersemphasizedthattheremustbeadesignatedprojectcoordinator,someonedowninthe
weeds,payingattentiontothedetails.ThemembersoftheDetroitSAKARPhadhighpraiseforthe
projectscoordinator,whohaddirectaccesstothehighestleadershipinkeyorganizations(prosecutor,
lawenforcement,forensicsciences)andwasabletocutthroughorganizationalredtapeandresistance
togettasksaccomplishedrelativelyquickly.Thedaytodayoperationsoftheteamalsorequired
negotiatinggroupconflict,facilitatingdecisionmaking,balancingworkloads,andjugglingpersonnel
changesandotherorganizationaltransitions.Theprojectcoordinatorwasinstrumentalinguidingthe
groupthroughthosechallengesaswell.98
98
Figure6.4alsohighlightsthattheDetroitcollaborativefoundtheactionresearchparadigmtobeconfusingattimesand
recommendedthatifothercommunitiesareconsideringaresearchcomponentintheirSAKinitiatives,thentheyshouldlearn
moreaboutthisapproachandclarifyexpectationsthroughouttheproject.Thefollowingsectionofthisreportwillincludean
extendeddiscussionoftheactionresearchparadigmanditsapplicationtoSAKinitiatives.
333
Whats Different Now: The Outcomes of the Detroit SAK Action Research Project
Evaluating the Success of an Action Research Project
Theactionresearchparadigmisbecomingamorecommonlyusedstrategyforaddressinglarge
scale,complexcrimeandsocialjusticeissues(seeKlofas,Hipple,&McGarrell,2010).Thisapproachis
quitedifferentfromtraditionalresearchinthatitrequiresbuildingacollaborativepartnershipbetween
researchersandpractitionerstodevelopempiricallyinformed,datadrivensolutionstocommunity
problems(McEwen,2003;Kennedy,2012;Klofasetal.,2010;Rosenbaum&Roehl,2010).Giventhose
aims,itisreasonableask:Diditwork?Didtheprojectachievetheaimsoftheactionresearch
paradigm?Inpriorcriminaljusticeactionresearchprojects,theprimaryoutcomethemeasureof
successwaswhethertherewasasignificantdecreaseincrime.Forexample,theBostonGunProject
waswidelyheraldedasasuccessduetosignificantdecreasesinyouthhomicideandnonfatalgun
violence(Kennedy,2012;Kennedy,Braga,Piehl,&Waring,2001).However,theDetroitSexualAssault
KitActionResearchProjecthadfundamentallydifferentaims.Theoverarchinggoalwasnotareduction
incrime(becauseprojectactivitieswerenottargetedtowardsthepreventionofsexualassault);rather,
thekeyaimwastochangethecriminaljusticesystemresponsetosexualassaultandtheprocessingof
rapekits.Inotherwords,thechangeneededwaswithinthecriminaljusticesystem(primarily)andin
theinterorganizationallinkagesbetweenagenciesthatassistrapevictims.Giventhoseaims,what
wouldsuccesslooklike?
Forguidanceontheconceptualizationandassessmentofsuccess,weturnedtotheevaluation
literature,asthissamebasicquestionisposedinthosecirclestoo:Didtheevaluationinfluencepolicy
andpractice?Inthefieldofprogramevaluation,scholarsdistinguishbetweendifferentformsof
influence(i.e.,differentwaysinwhichanevaluationmightaffectpractice),basedonwhataspectsofthe
projectarepickedup(i.e.,attendedto)andhowtheyareacteduponandbywhom(Henry&Marks,
2003;Kirkhart,2000;Leviton&Hughes,1981;Patton,1998;2008;Preskill&Caracelli,1998;Rossietal.,
334
2003;Weiss,1980,1988).Inparticular,Kirkharts(2000)theoryofevaluationinfluenceprovidesauseful
modelforoperationalizingtheoutcomesthisparticularactionresearchproject;Figure6.5(below)
summarizescorecomponentsofthisconceptualmodel.99
FIGURE6.5EvaluatingtheOutcomesofActionResearch:AnEvaluationInfluenceModel
ProcessBasedInfluence
Changesin
Attitudes
Toward
Evaluation
Changesin
Knowledge
About
Evaluation
Changesin
Behavior
Regarding
Evaluation
ResultsBasedInfluence
ConceptualUse
ofFindings
InstrumentalUse
ofFindings
Direct
Instrumental
Use
Indirect
Instrumental
Use
First,Kirkhart(2000)describedhowtherecanbeprocessbasedinfluence:notallof
evaluationsinfluenceemanatesfromtheformativeorsummativereportingofresults.Sometimesthe
primaryinfluencecentersaroundtheprocessofconductingtheevaluationitself(p.10).Processbased
influence(alsotermedprocessuse)referstochangeswithinprogramstaff(e.g.,knowledge,skill
development)andorganizations(e.g.,cultureshifts)thatstemfromparticipatinginanevaluation(Amo
&Cousins,2007;Patton,1998;2008;Shaw&Campbell,2014).Thistypeofuseisoftenreferredtoas
thedevelopmentofevaluativethinking.Toassessprocessuse,researchers/evaluatorsexamine
whetherthosewhoparticipatedintheevaluationprojectshowchangesintheirfeelingsandattitudes
99
Kirkharts(2000)theoryoffersamultidimensionalconceptualizationofinfluencethatconsiders:source(whatgetspicked
upfromtheevaluation),intentionality(thedegreetowhichinfluenceintendedorunintended),andtime(whetherthe
influencewasimmediate,endofcycle,orlongterm).Forsimplicity,wefocusedouranalysisonthesourcecomponentwhat
aspectsoftheprojectwerepickedup,bywhom,andusedtowhatends.
335
aboutevaluation(e.g.,reducedanxietyaboutevaluation),newknowledgeaboutevaluationmethods
(e.g.,increasedfamiliaritywiththehowsandwhysofresearch),andemergentbehaviorsregarding
institutionalizingevaluationwithintheorganization(e.g.,changinginternalprocesses,suchasrecord
keeping,tofacilitateongoingevaluation).InthecontextofthisSAKactionresearchproject,atissueis
whetherthereweretherechangesovertimeinthesedimensionsofevaluativethinkingamong
membersoftheDetroitcollaborative.
Second,Kirkhart(2000)notedthattherecanalsoberesultsbasedinfluence,whichrefersto
howthesubstantivefindingsoftheevaluationprojectinfluencepracticeandpolicy.Withinthisgeneral
formofinfluence,evaluatorshavedistinguisheddifferentsubtypesofinfluence,dependingontheways
inwhichthefindingsareutilizedbystakeholders.Conceptualuseiswhentheevaluationfindings
changehowstakeholdersthinkaboutaproblemorissueinafundamentalway(e.g.,anewinsight,a
deeperunderstanding,newquestions)(Patton,2008;Preskill&Caracelli,1988;Weiss,1980,1998;
Weiss,MurphyGraham,&Birkeland,2005).Thesubstantivefindingsoftheprojectitselfmaychallenge
stakeholdersideasandbeliefs,promptingthemtorethinkdeeplyheldideologies.Conceptualuseis
oftenlikenedtoanahamoment,onethatmightleadtospecificobservableactionstepstakenfor
change(seeinstrumentaluse,below)oronethatputsanewideaonslowsimmerthatlater(perhaps
evenyearslater)contributestopolicychange(seeKingdon,1995;Weiss,1980,1998,2004).Inthe
currentcontext,iftherehadbeenconceptualuseofthefindings,thenthestakeholdersparticipatingin
theSAKactionresearchprojectwouldhavechangedthewayinwhichtheythinkaboutrape,sexual
assaultinvestigations,and/orSAKtesting.
Instrumentaluseiswhentheevaluationfindingsareusedtoguideadecision,actionstep,or
otherchangeinitiatives(Patton,2008;Rossietal.,2003;Weiss,1980,1998).Insomeformsof
instrumentaluse,thereisadirectlinkbetweenthesubstantiveresultsandaprogrammaticchange(e.g.,
becausewefoundthis,wedecidedtodothat).However,itisalsopossiblethattheevaluationresults
336
becomeacatalystforfarreachingchanges.Inotherwords,theevaluationcreatedasnowballeffect,
wherebyaseriesofchangesandinitiativesfollow,andthelinkbetweenanyonechangeandthe
evaluationfindingsmaybemoreindirect.Inthecontextofthisactionresearchproject,thequestionis
whethertheevaluationfindingswereusedtocreatedirectchangesregardingDetroitsSAKtesting
practices,andwhethertheprojectcontributedtoindirect,furtherreachingchangesinpolicyregarding
sexualassaultinvestigationsandtheuseofmedicalforensicevidence.
Evidence of Process Use: Changes in Stakeholders Knowledge, Attitudes, and Beliefs

about Research and Evaluation
Investingthetimeandefforttoparticipateinanactionresearchprojectwouldbeexpectedto
developprogramstaffsevaluativethinking.Asnotedabove,processuseisassessedbyexamining
changesinstakeholders:1)attitudesaboutevaluation;2)knowledgeaboutevaluation;and3)
emergentbehaviorsregardinginstitutionalizingevaluationwithintheorganization.IntheDetroitSAK
ARP,therewasstrongevidenceofprocessuseamongstakeholdersinallthreeofthesesubdomains.
Withrespecttochangesinattitudestowardresearchandevaluation,thereweretwonotable
shiftsthatoccurredwithinmembersofthecollaborativeoverthecourseoftheactionresearchproject.
Atthebeginningoftheproject,somestakeholdersexpressedconcernsthattheresearchwouldtrash
us,makeuslookbad,andthatitwouldbeanotherexampleinalonghistoryofDetroitbeingroughed
upandchewedupandspitout.Whatseemedtoeasetheirconcernsandshiftattitudesovertime
wastheresearchteamsemphasisonprotectingidentity,confidentiality,andprivacy(seeChapter1:
Introduction).Thesearecoreprinciplesthatmustbeupheldinsocialscienceresearch,butbecausethe
stakeholdersinthisprojectdidnothavemuchpriorexperiencewithresearch,theyexpected,basedon
theirhistory,thatthefindingswouldlandinthe[news]papers,namesattached,allthat.Theideathat
researchcouldprovideaprotected,privatespacetostudyandunderstandproblemswasnoveltomany,
337
andastheycametoseetheseprinciplesinpracticeovermanymonths,therewasmarkedshiftintheir
attitudes.Asonememberofthecollaborativenoted,Whenwesaw[thatthis]wasntgoingtobe
anotherhackjob...[itwas]arealstudywithrules(abouthowresearchisconducted)andyou(the
researchers)werefollowingtherulesandtherulesprotectedus...thatmadeabigdifference.100
StakeholdersattitudesabouttheusefulnessofresearchandwhetheritcouldtrulyhelpDetroit
withitsproblemofunsubmittedSAKsalsoshiftedsubstantiallyovertime.Intheearlymonthsofthe
project,somemembersoftheteamexpressedaquidproquoattitudeabouttheresearchcomponent:
toleratetheevaluationpiecebecauseitcomeswithmoneytogetkitstested.Therewasskepticism
aboutwhetherresearchcouldtrulyhelp:whydoweneedresearchonthis?Theproblemissimplethe
kitswerenttestedandthesolutionissimpletoogetthemtested.Notallparticipating
individuals/organizationsexpressedthispointofview,butitwasastrongundercurrentatthebeginning
oftheproject.AttitudesbegantoshiftwhenthegroupwasstrugglingtocreatetheSAKtestingplan
(seeChapter4:Developing&EvaluatingaSAKTestingPlan).Detroitdidnothavefundsavailabletotest
allofthekits,stakeholdersstronglydisagreedaboutwhichkitstotestandwhy,andthegroupwas
grindingtowardastandstill.Asonememberoftheteamrecalled:
Irememberthosemeetingsabouttesting...whatamess...[then]you(theleadresearcher)
saidsomethinglike,wecouldtestalltheseideas,likeintheresearchproject.Itwasalightbulb
forus.Weweregoingaroundincirclesandyousteppedinandsaid,enoughstopgoingin
circles,theresawayout,andhereitis.Ithinkthatswhenalotofussawthatwereallydid
needsomedataandinformationandresearchtogetoutofthismess.
100
Foronespecificorganization,therewasamarkednegativechangeintheirattitudestowardtheresearch/evaluationprocess
afterreadingadraftofthefinalreport,astheyhadstrongconcernswithsomecontent(asnotedthroughout).However,the
quotesabovehighlighttheimportanceofstandardizedrulesandproceduresforresearch,whichwerefollowedconsistently,up
toandincludingthepreparationanddistributionofthisfinalreport(e.g.,sharingresultsinprogress,givingopportunityto
reviewadraftofthereport,notingdissentthroughoutthereport).
338
Stakeholdersalsopointedtotheteamsworkonvictimnotificationasanothermomentinwhichtheir
attitudesaboutresearchchangedsubstantially.Thegroupwasstrugglingtofigureouthowbestto
notifyvictimsandfeelingparalyzedbywhatwastherightwaytodothis,andasonememberofthe
collaborativerecalled,theresearchersinterjectedtosuggestempiricallybasedsolutions:
Thatwasbigmoment,when(theleadresearcherandcoinvestigator)toldustheresresearch
outtherethatcanhelpus.Theresnohowtomanual...(but)theresresearchabouthow
traumaaffectsvictims,whatitmightbelikeforvictimstohavetraumareactivated...andwe
canusethatresearchtohelpusfigureoutwhattodo...usuallywehavetojustmakeabest
guess...thiswashelpfultohavesomeinformationtoguideus.
Inthedevelopmentofboththetestingandvictimnotificationprotocols,theresearchteamwasableto
offerthecollaborativespecificsuggestions,strategies,andinformation,whichchangedstakeholders
attitudesaboutwhetherresearchcouldbehelpfultoaddressingtheproblemofunsubmittedSAKs.
Processusealsoreflectschangesinknowledgeaboutevaluation(e.g.,increasedunderstanding
ofscientificterms,scientificmethods,etc.).Inourformalandinformalinterviewswithstakeholders,
teammembersdidnotspecificallycommentonthisspecificformofprocessuse,otherthannotingthat
theylearnedalotabouthowresearchworks.However,we(theresearchers)noticedagreatdealof
changeintheteammembersknowledgeaboutevaluation,whichwedocumentedinourfieldnotes.
Overthecourseoftheproject,stakeholdersstartedpickingupthelanguageandterminologyof
evaluationandusingthemcorrectlyintheirdiscussionsanddebates.Belowisanexcerptfromourfield
notesfromameetinginwhichaseniorpolicedepartmentofficialchallengedtheteamtothinkabout
samplingmorecarefully:
339
When(policecommandstaffmember)askedtodayarenttheresamplingissuesweneedto
considerhere?Wecantjustpickstuffhereandthere,wehavetobesystematicaboutit,right
(researcher)?IthoughtIwasgoingtofallover.S/hebroughtitupwedidnt.We
(researchers)havebeenworkingsteadilyoverlastfewmeetingstoteachthegroupabout
samplingandwhatitisandwhyitmatters,anditlookslikeitsunkinand(astakeholder)
reflecteditbacktous.Tagthisprocessuse,knowledgechange.
Stakeholdersincreasingknowledgeaboutevaluationwasalsoevidentinthewayinwhichthey
engagedandchallengedtheresearchteamabouttheprojectsfindings.Foreachmaingoalofthis
project,theresearchteamconductedapowerpointpresentationofdraftfindingsforreviewand
comment.Overthecourseoftheproject,wenoticedthattheirquestionsbecamefarmoretechnically
informedandmethodologicallysophisticated,asevidentinthisexcerptfromourfieldnotes:
TodayIdidthepowerpointonthevictimnotificationfindingstothecollaborative...lotsof
questions,goodquestionsabouthowwedidtheanalysesandgeneratedthefindings.Things
like:whatsthebasisofcomparisonforthatnumber?Arethereotherstudiesthathavefound
thatrate?Howdidyoucalculatethatrate?Didyoudoatestofsignificanceonthat?...At
first,Ididntnoticeanythingunusualabouttheirquestions,IthinkbecauseImusedtofielding
thesekindsofquestionsallthetimefromacademicaudiences.Andthenithitmewaita
minute,Imnotpresentingtoanacademicaudience.Imgettinggrilledonmethodsby
practitioners.Itwasawesome.Tagthisprocessuse,knowledgechange.
Processusealsoreflectsemergentbehaviorsregardinginstitutionalizingevaluation.As
stakeholdersbecomemorefamiliarwithwhatevaluationcouldoffertheirorganizations,thenitstands
toreasonthattheymightstarttakingstepstobuilddatasystemsintotheirpractice.Forthisspecific
subtypeofprocessuse,wesawintentionforbehavioralchange.Specifically,twoorganizations
involvedinthecollaborativenotedthattheyarepayingfarmoreattentiontorecordkeepinganddata
managementasaresultofparticipatinginthisproject.Oneorganizationwasseekingfinancial
340
resourcestohireafulltimedataanalysttohelptrackSAKs(aswellasotherissuespertinenttothework
ofthatorganization).Stakeholderswithinthatorganizationnotedthattheyhadcometoseethevalue
inbeingabletoaccessnumbers,trackprogress,andsharethatinformationwithmembersoftheirstaff
andwiththecommunityatlarge,andfeltitwouldbevaluabletoinstitutionalizethatwithintheir
organization.
Evidence of Conceptual Use: Changes in Stakeholders Attitudes and Beliefs About
Sexual Assault and SAK Testing
Conceptualusereflectschangesinstakeholdersunderlyingunderstandingandbeliefsabouta
problem.Theevaluationfindingsmaychallengeteammemberstothinkabouttheissuesathandina
differentway.IntheDetroitSAKARP,thereweretwonotablemanifestationsofconceptualuse.As
notedpreviously,manystakeholderswereskepticalabouttheutilityofresearch,andovertime,their
attitudeschanged(processuse)whentheysawhowthefourTestingGroupscouldhelpanswertheir
questions.Theanswerstothosequestionsi.e.,theresultsoftheSAKtestinghadastrongimpacton
stakeholdersunderlyingbeliefsaboutrapekittesting.ThetestingyieldedahighrateofCODIShits,far
morethanmosthadexpected,whichchangedhowmanythoughtaboutthesekitsandthepurposeof
SAKtesting.Asonememberofthecollaborativeexplained:
Remember,theseweretheones(kits)nottested,becausewhatever,theydidntmatter,the
victimswerelying,thevictimswerethisorthat,andthenbam,everymonth,(forensicscience
teamrepresentative)bringstheupdate(CODIShittrackingchart)tothemeeting...wevegot
thismanyhits,thismanyserials.Alloftheseshouldhavebeennothing,nohits,right?Because
thesewerethetossaways,theonesnotworthit,therewasntanythingtofind...wetested
themandguesswhat,therearesomeseriouslydangerouspeopleoutthere.Stilloutthere...If
thatdoesntchangethewayyouthinkaboutthis,Idontknowwhatwould...itwasahuge
wakeupcall...theproofwasinthatchart,everymonth,showingthenumbers,showingthat
thewaywevebeenthinkingaboutthiswaswrong.Flatoutwrong.(emphasesinoriginal)
341
SeveralothermembersofthecollaborativenotedthattheresultsfromTestingGroup2(NonStranger
Rapes)inparticularchallengedtheirbeliefsaboutboththeseriousnessofknownoffenderrapesandthe
importanceoftestingSAKsforthesecrimes,asthesetwoquotesillustrate:
[TestingGroup2resultswere]notwhatIwasexpecting.Thatmanyhitsfortheknown
offenders?...Iwasoneofthosepeoplethinking,nopointintestingthose.Saveourmoney,no
point...butlookatthehits,lookattheserialsinthatgroup...Iadmitit,Ithoughtaboutthese
ashesaid/shesaids..itprobablywasntrape,itwasamisunderstandingaboutwhowanted
what...itsnotamisunderstandingwhenyouseethepattern...thistotallychangedhowI
thinkabouttheserapes(nonstranger).
Goingin(tothisproject),Isaid,itsthestrangerrapesthatneedtesting...thatsthethreatto
publicsafety...wedonthavealotofmoney,sothatswherewegottoputit...andIguess
everyoneelse(knownoffenderassaults)justneedstoworkthingsoutontheirown,youknow?.
..[butthen]youlookatthechart(CODIShittrackingchart)andwhatcanIsay?...Iguesswe
needtofindmoremoneybecausetheseneedtestingtoo(nonstrangerSAKs).
Stakeholdersalsoexhibitedconceptualchangeinregardtotheirunderstandingofthetraumaof
rapeandhowitaffectssurvivors.Whentheteamwasdevelopingthevictimnotificationprotocol,the
advocacypartnersintheprojectpushedthegrouptotakeavictimcentered,traumainformedpointof
view(seeChapter5:Developing&EvaluatingaVictimNotificationProtocol).But,asoneadvocate
noted,itbecameprettyclearthatpeopledidntknowwhatweweretalkingabout...whatthat
meant.Atthevictimnotificationplanningretreat(seeChapter5),thegroupdecidedthattheresearch
teamandadvocacyorganizationsshouldworktogethertodevelopatrainingforallorganizations
involvedintheARPonthetraumaofsexualassaultandhowstakeholderscouldtakeavictimcentered
approachtotheirworkwithsurvivors.Thattrainingincludedareviewoftheliteratureonthe
neurobiologyoftrauma,withanemphasisonhowstresshormonesaffectvictimsbehavioral
342
presentationsandtheirmemoryrecalloftheassault.Manystakeholderscommentedthatthistraining
wasthefirsttimetheyhadeverreceivedformalinstructionontheimpactoftraumaandithada
profoundimpactonhowtheythoughtaboutsexualassault,asoneteammemberdescribed:
Ivebeendoingthis(sexcrimeslawenforcement)foralong,longtime.Iveneverheard
anythinglikethis...nevergotanytrainingonthis...somanythingsIthoughtmeantvictim
waslying,andthenIlearntthatsometimesvictimsarescatteredandhavetroublemakingsense
andsoundreallyflatandoutofitbecausethatsthetraumaoftherape...thetrauma
hormones...Ithoughtbacktooldcases,whatIsaw,whatIthinkIsaw,andIwasoff,byalot.
Membersofthelawenforcementcommunitynotedthat,tothem,aclearsignofafalsereportwasa
victimgivingasketchyversionoftheassault,disorganized,wanderingallaround,liketheyremaking
itupastheyregoingalong.Researchontheneurobiologicalimpactoftraumaclearlyshowsthat
victimsmemoriesofassaultsarefragmentedandtherefore,recallcanbeslowanddifficultand
disorganized(Kossetal.,1995,1996;Roozendaaletal.,2009;Rubinetal.,2008).Thatinformation
challengedofficersbeliefs,asonenoted:
Sowaitaminute.Youmeanwhatwereseeingisntbullshit?Ivealwaysbelieveditwasbullshit
...Well,howboutthat.Realtraumalookslikebullshit.
Thegoalofthistrainingwastoinformteammembersabouttheunderlyingmechanismsofhowrape
affectsvictimshealthandwellbeingsothattheycoulduseavictimcentered,traumainformed
approachtovictimnotifications.Theinformationpresentedbytheadvocacyorganizationandthe
researchteamchallengedmanymembersbeliefsabouthowvictimsoughttobehaveandhowthey
reallydobehaveandwhy.Asonememberoftheteamnoted,This[goes]beyondvictimnotification.
..thisisusefulformycurrentcases,rightnow,today...itscompletelydifferentwayoflookingatthis.
343
Evidence of Instrumental Use: Changes in Policy & Practice
Instrumentalusereferstochangesinpracticeandpolicy,stemmingfromanevaluation.Some
formsofinstrumentalusearedirect:becauseoffindingX,changeYwasmade(i.e.,adirectoneto
onecorrespondence).Otherinstancesofinstrumentalusemaybeindirect:theevaluationprojectsets
intomotionaseriesofreflections,thoughts,andideasthatbecomesacatalystforchange.Therewas
strongevidenceofinstrumentalusefromthisprojectregardingsexualassaultingeneralandSAKtesting
specifically)(seeFigure6.6followingpages).
1) IncreaseSupportServicesForSexualAssaultSurvivors:
Theadvocacyorganizationswantedtoensurethatsurvivorswhowouldbenotifiedhad
comprehensivesupportservices(e.g.,hotline,advocacy,counseling).Thestate
governmentviolenceagainstwomenagencythatparticipatedinthecollaborative
initiatedanefforttolinkallDetroitareavictimserviceorganizationssothatallwere
awarethatnotificationswereabouttobegin,sothatnomatterwhichagencyavictim
mighttoturnforsupport,allorganizationswerereadytohelp.(DirectInstrumental)
Theresultsofthehistoricalcontextanalysis(Chapter3:WhySoManyUnsubmittedSAKs
inDetroit)showedthatDetroitdidnothavesufficientcommunitybasedadvocacy
services.Statefunderswereawareofthatproblem,buttheevaluationfindings
highlightedthefarreachingnegativeeffectsithadonsystemicfunctioning.Inresponse,
thestategovernmentVAWagencyaddedfournewcommunitybasedadvocate
positionstoDetroitorganizationsduringthe30monthsoftheactionproject(supported
byOVWfunds).ThestategovernmentVAWagencyalsoinvestedintraining,mentoring,
andcapacitybuildingwiththeexecutivedirectorsandstaffofDetroitserviceagencies
regardingvictimcentered,traumainformedcare.(DirectandIndirectInstrumental)
2) ExpandTrainingForAllDetroitSexualAssaultServiceProviders:
Theresultsofthehistoricalcontextanalysis(Chapter3:WhySoManyUnsubmittedSAKs
inDetroit)showedthatserviceprovidershavenothadagreatdealofprofessional
344
trainingaboutsexualassaultanditsimpactonvictims.Therefore,theresearchteam,
thenationalVAWfoundationwhowasapartnerinthisproject,andthelocalvictim
advocacyorganizationspartneredtocreateaonedaytrainingonvictimcentered,
traumainformedcareforallmembersoftheDetroitSAKARPandallvictimnotification
staff.(DirectInstrumental)
TheresultsfromthereviewofpolicefilesassociatedwithunsubmittedSAKs(inthis
project,seeChapter3:WhySoManyUnsubmittedSAKsinDetroit)showedthatthevast
majorityofthesereportedassaultswerenotthoroughlyinvestigatedandthatofficers
madeanumberofvictimblamingstatementsintheirreports.Toaddressthisproblem,
membersoftheresearchteamconductedaseparatelawenforcementonlytrainingfor
thelocalpolicedepartmentontheneurobiologyofsexualassaultanditsimplications
foravictimcenteredresponsetosexualassault.Inaddition,thestatelevel
prosecutorsassociation(whichwasamemberofthecollaborative)wasintheprocess
ofdevelopingamultidaytrainingforlawenforcementonoffenderfocusedsexual
assaultinvestigations(asaseparateinitiative),whichwassupportedbyOVWfunds,
administeredbythestategovernmentVAWagency.ContentfromtheNIJSAKARP
informedthecontentofthattrainingandmembersoftheDetroitSAKARPcollaborative
(andtheresearchteam)wereselectedastrainersforthisevent.(IndirectInstrumental)
3) PursueTestingofallPreviouslyUnsubmittedSAKs:
TheresultsoftheSAKcensusshowedthatthevastmajorityofSAKsinpoliceproperty
hadnotbeensubmittedfortesting(seeChapter2:HowManyUnsubmittedSAKsin
Detroit).GiventhatDetroitisoneofmanycitiesinWayneCounty,MI,theprosecutors
officeaskedtheother41lawenforcementagenciesinthecountytovoluntarilyassess
howmanyunsubmittedSAKstheyhadincustodytoinformcountyleveltestingplans
andresources.(DirectandIndirectInstrumentalUse)
TheDNAtestingandCODIShitresultsofthe1,600(1,595)SAKstestedinthescopeof
thisprojectwerenotyetfinalized,butpreliminarydataindicatedhighratesinall
TestingGroups,includingTestingGroup2(NonStrangerRapes)andTestingGroup3
(PresumedSOLExpired),whichsomestakeholdersexpectedwouldhavesubstantially
345
lowerCODIShitrates.KeystakeholdersinDetroitSAKARPworkedwiththeGovernors
OfficeandtheMichiganAttorneyGeneralsOfficetosecurefundingtotestasmany
remainingpreviouslyunsubmittedDetroitSAKsaspossible.TheMichiganAttorney
GeneralsOfficeallocated$4millionfromsettlementfundsfortestingDetroitSAKs.The
statepoliceforensicsciencedivisionnegotiatedacontractwithaprivatevendor
laboratory,and7,393previouslyuntestedDetroitSAKshavebeensubmittedfortesting
thusfar.(DirectandIndirectInstrumentalUse)101
4) SupportInvestigation,Prosecution,andVictimAdvocacyForAllTestedSAKs:
ThepurposeoftheDetroitSAKARPwastoassessthescopeoftheproblem,research
thecontributingfactorsthatgaverisetotheproblem,anddevelopandevaluateplans
forSAKtestingandvictimnotification.Thenextstepsofinvestigatingandprosecuting
thecasesthatemanatefromthattestingwerenotwithinthescopeoftheARP.
However,giventhatthetestingresultsindicatedthattherewouldbeasubstantial
numberofcasesthatwouldneedtobereopenedandexamined,thecommunity
neededtodevelopalongtermplanforinvestigation,prosecution,andvictimadvocacy.
TheprosecutorsofficeappliedforandreceivedafederalOVWGrantstoEncourage
Arrestawardtosupportamultidisciplinarycoldcaseunittoinvestigateandprosecute
thesecases,withaccompanyingvictimadvocacyservices.(IndirectInstrumentalUse)
TheOVWGrantstoEncourageArrestawardwascriticalinestablishingacoldcaseunit;
however,movingforwardwiththeinvestigationandprosecutionofthecasesassociated
withtheSAKstestedinthisprojectandthosesubmittedfortestingwithfundsfrom
theMichiganAttorneyGeneralsOfficewillrequireadditionalfinancialresources.To
101
OneofkeyresearchquestionsintheDetroitSAKARPwaswhethertherewasempiricalsupportfortestingallSAKsor
whetheritispossibletocreatedatadrivenprioritization/selectionguidelines.AsnotedinChapter4:Developing&Evaluatinga
SAKTestingPlanandpreviouslyinthisChapter,theresultssuggestthatthereismeritintestingallSAKsreleasedbyvictimsto
lawenforcement,regardlessofvictimoffenderrelationshipandstatuteoflimitationsstatus,asratesofCODIShitsare
statisticallyindistinguishableamongthesedifferentcases/SAKs.Thoughtheresultsofthisactionresearchprojectsuggestthat
testingallSAKsinDetroitwaswarranted,thedecisiontoseekfundingfromtheGovernorsOfficeandAttorneyGeneralsOffice
fortestingallremainingkitswasmadebeforethestatisticalanalyseshadbeencompleted.Inthatsense,therewasnotDIRECT
instrumentaluseoftheprojectsfindings;however,thepreliminaryresultswereinfluentialtokeylocalandstatepolicymakers
(i.e.,indirectinstrumentaluse).Theresearcherswerenotinvolvedinanyoftheeffortstosecurefundingtotestallremaining
kits,indeferencetotheongoingactionresearchprojectandtheresearchquestionspertainingtotestall/testsome.
346
thatend,theprosecutorsofficepartneredwithastatefoundationandthelocalcrime
commission,bothnonprofit501(c)(3)s,toraiseawarenessandgarnerthefinancial
resourcesnecessarytoseethatallsuspectsidentifiedthroughtestingareinvestigated
andprosecutedtothefullestextentofthelaw.(IndirectInstrumentalUse)
5) PreventtheProblemofUnsubmittedSAKsfromHappeningAgain:
WithinthefirstthreemonthsoftheDetroitSAKARP,thepolicedepartmentmadea
policychangetosubmitallSAKs(incurrentcases)fortesting.Thisdecisionwas
informedbydiscussionswithinthecollaborativemeetingsaswellasprivate
conversationsamongtheleadershipofkeyorganizations(police,prosecution,state
policecrimelab).(DirectandIndirectInstrumentalUse)
ArecurringchallengeintheDetroitSAKARPwasthelackofinformationmanagement
andtrackingofSAKsfromthepointatwhichtheywerecollectedbyahealthcare
providerandreleasedbythevictimforretrievalbylawenforcement,towhentheywere
submittedfortesting,towhenthetestinghadbeencompleted.Whereasbuildingnew
ITinfrastructureisakeylongtermgoalofmanyorganizationsinvolvedinthisproject,a
moreimmediateactionstepwastakenbytheprosecutorsofficetosecurefundingfora
pilotSAKtrackingproject.AllMichiganSAKsnowhaveabarcodeontheoutsideofthe
box,buttodate,nomunicipalitiesorlawenforcementjurisdictionsinMichiganhave
theinfrastructurefortrackingkits(andnostatewidetrackingsystemexistseither).The
prosecutorsofficeformedapartnershipwithUnitedParcelService(UPS)todevelop
andimplementapilottrackingprojectforDetroitSAKs(currentcases/SAKs)(Directand
IndirectInstrumentalUse)
TopreventtheproblemofuntestedSAKsinthefuture,legislativechangemaybe
necessary.Tothatend,multipleorganizationswhoparticipatedintheDetroitSAKARP
workedtogethertosupportlegislationrequiringallcurrentSAKstobesubmittedfor
testing(providedthevictimhasauthorizedthereleaseoftheSAKfortesting).The
SexualAssaultKitEvidenceSubmissionAct(PA227)wasunanimouslypassedbyboth
housesoftheMichiganlegislatureandwassignedbytheGovernorintolawonJune26,
2014.(IndirectInstrumentalUse).
347
FIGURE6.6
Moving Forward:
Changing SAK Policy and Practice

Key changes in Detroit and Michigan regarding SAKs,
sexual assault investigations, and victim services.
The NIJ project was a huge catalyst for change.
1. Increased state funding for

communitybased sexual assault
advocates
State agencies have increased funding to
Detroit-area non-profit organizations to
expand community-based advocacy for
sexual assault victims notified as part of the
Detroit SAK ARP. These funds will also be used
to provide advocacy services in new sexual
assault cases.
2. Training for
sexual assault
practitioners
Training on a victim-centered, traumainformed response to sexual assault has
been and will continue to be
conducted with police, prosecutors,
medical/nursing, and victim advocacy.
Law enforcement personnel have also
had in-depth training on offenderfocused sexual assault investigations.
3. County-wide assessment of the scope of the problem of untested SAKs

The problem of untested SAKs is not exclusive to the city of Detroit. The prosecutors office has
reached out to the 41 other police agencies in Wayne County, as well as local hospitals, to
determine the extent of untested sexual assault kits county-wide.
348
4. Funding to test all remaining SAKs

After the scope of the problem in Detroit had been clearly established, the Governors
Office and the Michigan Attorney Generals Office allocated $4 million to test all
remaining previously untested SAKs in Detroits police property.
7. Development of a
5. Federal Grants to
SAK tracking project
Encourage Arrest Award

A grant was awarded to the prosecutors office to help fund
a Cold Case Sexual Assault Unit that would investigate and
prosecute the cases emanating from the testing conducted
in the Detroit SAK ARP.
6. Creating Community
Partnerships
The prosecutors office has
partnered with the Michigan
Womens Foundation and Detroit Crime Commission, both
non-profit 501(c) (3)s, to raise awareness and garner the
financial resources necessary to see that all suspects
identified through testing are investigated and prosecuted
to the fullest extent of the law.
8.
The Detroit SAK ARP

highlighted how there is no
mechanism to track a SAK
from when it was collected
by a medical provider to
when it was tested by a
forensic scientist. In order to
ensure that SAKs do not fall
through the cracks and go
untested in the future, it is
crucial that all kits are
accounted for throughout
the collection and testing
processes. The prosecutors
office is partnering with UPS
to develop a pilot electronic
tracking project for all rape
kits collected in the city of
Detroit.
Legislative Reform
State legislation has been drafted and enacted that requires all sexual assault kits
released to law enforcement to be submitted for testing (MI Act 227 of Public Acts 2014).
Additional bills are currently pending in the legislature that would, among other things,
create a statewide electronic tracking system for SAKs and provide victims with
electronic access to the status of their SAKs.
349
Summary & Conclusions

ThegoaloftheDetroitSAKARPwastodevelopdatadrivensolutionsfortheunsubmittedSAKs
inpolicepropertyinthisjurisdiction,andinsodoing,toprovideempiricallybasedrecommendationsfor
othercommunitiesstrugglingwiththissameproblem.Thisprojectdocumentedthechallenges
communitiescanexpecttofaceintheprocessofconductingacensus,developingatestingplan,and
creatingandimplementingavictimnotificationprotocol.Thelessonslearnedfromthisprojecthave
beentranslatedintoplanningtools,sampleprotocols,resourcebooklets,anddatacollectionsheetsthat
otherjurisdictionscanmodifyfortheirownuse.Asonestakeholdernotedatthebeginningofthe
project,Iwishtherewasaroadmaporahowtoguideforthis.Ourexperiencesinthisprojectsuggest
thatbecausethehistory,context,andresourcesofeachcommunityareunique,thereisnooneright
waytotacklethisproblem,butthatitistherightthingtotackletheproblem.TherateofCODIShits
andserialsexualassaultsdocumentedinthisprojectunderscoresthepressingneedforreforminhow
thecriminaljusticesystemrespondstosexualassault.
Theactionresearchparadigmprovidedausefulmechanismforbringingmultidisciplinary
practitionerstogetherwithresearcherstodevelopshorttermandlongtermresponsestrategies.The
projecthelpedbringtolightlongtermfrustrationsandfrictionsinthecommunity,butalsoagreater
awarenessoftheinterdependenceoftheseorganizationsandhowwhentheypooltheirresourcesand
expertise,theyaremoresuccessfulincreatingchange.Actionresearchprojectsrequireatremendous
amountoftime,effort,andtrust,andinsodoing,theycanhelpbolsterthesenseofurgencythats
neededforchangeandsupportlongtermplanningofchangeinitiatives.Inclosing,onememberofthe
collaborativesummarizedthesentimentsofmanywiththisreflectionontheexperienceofbeinginthis
projectandwhatitaccomplished:
350
Thiswasalong,hardproject...Ithinkitdidalotofgoodinrebuildingbrokenrelationships[in
Detroit]andmakingnewones...Ihopewhatwedidherewillhelpustodobetterinthefuture.
Ihopewhatwedidwillhelpothercommunities.Mostofall,Ihopewehelpedpreventthisfrom
happeningagainhereandinothercitiesandotherstates...Everykitisaperson...itsnota
box,itsaperson.Wehavearesponsibilitytofulfilltoeachandeveryoneofthem.
351
APPENDIX A: References
Abbey,A.,&McAuslan,P.(2004).Alongitudinalexaminationofmalecollegestudents'perpetrationof
sexualassault.JournalofConsultingandClinicalPsychology,72,747756.
Abbey,A.,Wegner,R.,Pierce,J.,&JacquesTiura,A.J.(2012).Patternsofsexualaggressionina
communitysampleofyoungmen:Riskfactorsassociatedwithpersistence,desistance,and
initiationovera1yearinterval.PsychologyofViolence,2,115.
Adams,J.A.,Girardin,B.,&Faugno,D.(2001).Adolescentsexualassault:documentationofacute
injuriesusingphotocolposcopy.JournalofPediatricandAdolescentGynecology,14,175180.
Adler,P.A.,&Adler,P.(2002).Thereluctantrespondent.InJ.F.Gubrium&J.A.Holstein(Eds.),
Handbookofinterviewresearch:Contextandmethod(pp.515536).ThousandOaks,CA:Sage.
Alkin,M.C.(2004).Comparingevaluationpointsofview.InM.C.Alkin(Ed.),Evaluationroots:Tracing
theoristsviewsandinfluences(pp.311).ThousandOaks,CA:Sage.
AmericanEvaluationAssociation(1995/2004).Guidingprinciplesforevaluators.AmericanJournalof
Evaluation,28,237238.
Agresti,A.(2002).Categoricaldataanalysis(2nded.).NewYork,NY:JohnWiley&Sons,Inc.
Amo,C.,&Cousins,J.B.(2007).Goingthroughtheprocess:Anexaminationoftheoperationalizationof
processuseinempiricalresearchonevaluation.NewDirectionsforEvaluation,2007,526.
Angen,M.J.(2000).Evaluatinginterpretiveinquiry:Reviewingthevaliditydebateandopeningthe
dialogue.Qualitativehealthresearch,10,378395.
Aquino,J.,Enzmann,D.,Schwartz,M.,Jain,N.,&Kraft,S.(2013).descr:Descriptivestatistics(Version
1.0.2)[Computerprogram,Rpackage].Fortaleza,CE,Brazil.Retrievedfromhttp://CRAN.R
project.org/package=descr
Ashworth,P.(1993).Participantagreementinthejustificationofqualitativefindings.Journalof
PhenomenologicalPsychology,24,316.
Atkinson,P.,Coffey,A.,Delamont,S.,Lofland,J.,&LoflandL.(2001).Handbookofethnography.
ThousandOaks,CA:Sage.
352
Azzam,T.,Evergreen,S.,Germuth,A.A.,&Kistler,S.J.(2013).DataVisualizationandEvaluation.New
DirectionsforEvaluation,2013,732.
Baker,B.R.(2009).Detroitpolicecrimelabclosure:Impactonstatepoliceforensicsciencedivision
backlog.FromStateNotes:TopicsofLegislativeInterest.Retrievedfrom:
http://www.senate.michigan.gov/sfa/Publications/Notes/2009Notes/NotesJanFeb09.pdf
Banks,A.(2002).Polypharmacyofposttraumaticstressdisorder.MedicalPsychiatry,17,151174.
Barker,L.E.,Luman,E.T.,McCauley,M.M.,&Chu,S.Y.(2002).Assessingequivalence:analternativeto
theuseofdifferencetestsformeasuringdisparitiesinvaccinationcoverage.AmericanJournalof
Epidemiology,156,10561061.
Bashford,M.(2013,March).HowNewYorkCitytackleditsbacklog.WebinarpresentedfortheNational
CenterforVictimsofCrime.
Beebe,J.(2001).Rapidassessmentprocess:Anintroduction.WalnutCreek,CA:AltaMira.
Beichner,D.,&Spohn,C.(2005).Prosecutorialchargingdecisionsinsexualassaultcases:Examiningthe
impactofaspecializedprosecutionunit.CriminalJusticePolicyReview,16,461498.
Bergoffen,D.B.(2005).Howrapebecameacrimeagainsthumanity:Historyoferror.InA.D.Schrift
(Ed.),Modernityandtheproblemofevil(pp.6680).Bloomington,IN:IndianaUniversityPress.
Bernard,H.R.(2011).Researchmethodsinanthropology(5thed.).Lanham,MD:Altamira.
Black,M.C.,Basile,K.C.,Breiding,M.J.,Smith,S.G.,Walters,M.L.,Merrick,M.T.,Chen,J.,&Stevens,
M.R.(2011).TheNationalIntimatePartnerandSexualViolenceSurvey(NISVS):2010Summary
Report.Atlanta,GA:NationalCenterforInjuryPreventionandControl,CentersforDisease
ControlandPrevention.
Boardman,A.,Greenberg,D.,ViningA.,&Weimer,D.(2010).Costbenefitanalysis(4thed.).NewYork:
PrenticeHall.
Bonanno,G.A.(2004).Loss,trauma,andhumanresilience:Haveweunderestimatedthehuman
capacitytothriveafterextremelyaversiveevents?AmericanPsychologist,59,2028.
Bonanno,G.A.(2005).Resilienceinthefaceofpotentialtrauma.CurrentDirectionsinPsychological
Science,14,135138.
353
Bradburn,N.M.,Sudman,S.,&Wansink,B.(2004).Askingquestions:thedefinitiveguideto
questionnairedesignformarketresearch,politicalpolls,andsocialandhealthquestionnaires.
SanFrancisco,CA:JohnWiley&Sons.
Brecklin,L.R.,&Ullman,S.E.(2001).Theroleofoffenderalcoholuseinrapeattacks:Ananalysisof
NationalCrimeVictimizationSurveydata.JournalofInterpersonalViolence,16,321.
Britten,N.(2007).Qualitativeinterviews.InC.Pope&N.Mays(Eds.),Qualitativeresearchinhealthcare
(3rdedition)(pp.1220).Oxford,UK:Blackwell.
Bronfenbrenner,U.(1979).Theecologyofhumandevelopment(pp.1642).Cambridge,MA:Harvard
UniversityPress.
Bronfenbrenner,U.(1986).Ecologyofthefamilyasacontextforhumandevelopment:Research
perspectives.DevelopmentalPsychology,22,723742,
Bronfenbrenner,U.(1995).Developmentalecologythroughspaceandtime:Afutureperspective.InP.
Moen,G.H.Elder,Jr.,andK.Luscher(Eds.),Examininglivesincontext:Perspectivesonthe
ecologyofhumandevelopment(pp.619647).Washington,DC:APABooks.
Buchbinder,E.(2010).Beyondchecking:Experiencesofthevalidationinterview.QualitativeSocial
Work,doi:1473325010370189.
Butler,J.M.(2005).ForensicDNAtyping:Biology,technology,andgeneticsofSTRmarkers.Waltham,
MA:Elsevier.
Butler,J.M.(2010).FundamentalsofforensicDNAtyping.Boston,MA:AcademicPress/Elsevier.
Butler,J.M.(2012).AdvancedtopicsinforensicDNAtyping:Methodology.Waltham,MA:Elsevier.
Campbell,R.(1998).Thecommunityresponsetorape:Victimsexperienceswiththelegal,medical,and
mentalhealthsystems.AmericanJournalofCommunityPsychology,26,355379.
Campbell,R.(2002).Emotionallyinvolved:Theimpactofresearchingrape.NewYork:Routledge.
Campbell,R.(2005).Whatreallyhappened?Avalidationstudyofrapesurvivorshelpseeking
experienceswiththelegalandmedicalsystems.Violence&Victims,20,5568.
Campbell,R.(2006).Rapesurvivorsexperienceswiththelegalandmedicalsystems:Dorapevictim
advocatesmakeadifference?ViolenceAgainstWomen,12,116.
354
Campbell,R.(2008).Thepsychologicalimpactofrapevictimsexperienceswiththelegal,medical,and
mentalhealthsystems.AmericanPsychologist,68,702717.
Campbell,R.(2009).Science,socialchange,andendingviolenceagainstwomen:Whichoftheseisnot
liketheothers?ViolenceAgainstWomen,15,434439.
Campbell,R.,Adams,A.E.,Wasco,S.M.,Ahrens,C.E.,&Sefl,T.(2009).Traininginterviewersfor
researchonsexualviolence:Aqualitativestudyofrapesurvivorsrecommendationsfor
interviewpractice.ViolenceAgainstWomen,15,595617.
Campbell,R.,Adams,A.E.,Wasco,S.M.,Ahrens,C.E.,&Sefl,T.(2010).Whathasitbeenlikeforyouto
talkwithmetoday?:Theimpactofparticipatingininterviewresearchonrape
survivors.ViolenceAgainstWomen,16,6083.
Campbell,R.,Ahrens,C.E.,Sefl,T.,Wasco,S.M.,&Barnes,H.E.(2001).Socialreactionstorape
victims:Healingandhurtfuleffectsonpsychologicalandphysicalhealthoutcomes.Violence&
Victims,16,287302.
Campbell,R.,Baker,C.K.,&Mazurek,T.(1998).Remainingradical?Organizationalpredictorsofrape
crisiscenterssocialchangeinitiatives.AmericanJournalofCommunityPsychology,26,465491.
Campbell,R.,Bybee,D.,Ford,J.K.,Patterson,D.,&Ferrell,J.(2009).AsystemschangeanalysisofSANE
programs:Identifyingmediatingmechanismsofcriminaljusticesystemimpact.Washington,DC:
NationalInstituteofJustice.
Campbell,R.,Bybee,D.,Kelley,K.D.,Dworkin,E.R.,&Patterson,D.(2012).TheimpactofSexualAssault
NurseExaminer(SANE)programservicesonlawenforcementinvestigationalpractices:A
mediationalanalysis.CriminalJustice&Behavior,39,169184.
Campbell,R.,Bybee,D.,Shaw,J.L.,Townsend,S.M.,&Karim,N.(2014).Theimpactofsexualassault
nurseexaminer(SANE)programsoncriminaljusticecaseoutcomes:Amultisitereplication
study.ViolenceAgainstWomen,20,607625.
Campbell,R.,Dworkin,E.,&FehlerCabral,G.(2009).Anecologicalmodeloftheimpactofsexual
assaultonwomensmentalhealth.Trauma,Violence,&Abuse,10,225246.
Campbell,R.,Greeson,M.R.,Bybee,D.,Kennedy,A.,&Patterson,D.(2010).Adolescentsexualassault
victimsexperienceswithSANESARTsandthecriminaljusticesystem.Washington,DC:National
InstituteofJustice.
355
Campbell,R.,Greeson,M.R.,&FehlerCabral,G.(2013).Withcareandcompassion:Adolescentsexual
assaultvictimsexperiencesinSexualAssaultNurseExaminerprograms.JournalofForensic
Nursing,9,6875.
Campbell,R.,Greeson,M.R.,FehlerCabral,G.,&Kennedy,A.(2014).Pathwaystohelp:Adolescent
sexualassaultvictimsdisclosureandhelpseekingexperiences.ViolenceAgainstWomen.
Campbell,R.,Gregory,K.A.,Patterson,D.,&Bybee,D.(2012).Integratingqualitativeandquantitative
approaches:Anexampleofmixedmethodsresearch.InL.A.Jason&D.Glenwick(Eds.),
Innovativemethodologicalapproachestocommunitybasedresearch:Theoryandapplications
(pp.5168).WashingtonDC:AmericanPsychologicalAssociation.
Campbell,R.,Patterson,D.,Bybee,D.,&Dworkin,E.(2009).Predictingsexualassaultprosecution
outcomes:Theroleofmedicalforensicevidencecollectedbysexualassaultnurseexaminers
(SANEs).CriminalJustice&Behavior,36,712727.
Campbell,R.,Patterson,D.,&FehlerCabral,G.(2010).Usingecologicaltheorytoevaluatethe
effectivenessofanindigenouscommunityintervention:Astudyofsexualassaultnurse
examiner(SANE)programs.AmericanJournalofCommunityPsychology,46,263276.
Campbell,R.,Patterson,D.,&Lichty,L.F.(2005).Theeffectivenessofsexualassaultnurseexaminer
(SANE)program:Areviewofpsychological,medical,legal,andcommunityoutcomes.Trauma,
Violence,&Abuse:AReviewJournal,6,313329.
Campbell,R.,&Raja,S.(2005).Thesexualassaultandsecondaryvictimizationoffemaleveterans:Help
seekingexperiencesinmilitaryandciviliansocialsystems.PsychologyofWomenQuarterly,29,
97106.
Campbell,R.,Sefl,T.,Barnes,H.E.,Ahrens,C.E.,Wasco,S.M.,&ZaragozaDiesfeld,Y.(1999).
Communityservicesforrapesurvivors:Enhancingpsychologicalwellbeingorincreasing
trauma?JournalofConsultingandClinicalPsychology,67,847858.
Campbell,R.,Shaw,J.,&Gregory,K.A.(2014).Givingvoiceandnumbers,too:Mixedmethods
researchincommunitypsychology.InM.Bond&M.Shinn(Eds.),Handbookofcommunity
psychology.Washington,DC:AmericanPsychologicalAssociation.
Campbell,R.,Sprague,H.A.,Cottrill,S.,&Sullivan,C.M.(2011).Longitudinalresearchwith
sexualassaultsurvivors:Amethodologicalreview.JournalofInterpersonalViolence,6,433461.
Campbell,R.,&Townsend,S.M.(2011).Definingthescopeofsexualviolence.InC.Renzetti,J.Edleson,
&R.Bergen(Eds.)Sourcebookonviolenceagainstwomen(2nded.)(pp.95109).Thousand
Oaks,CA:Sage
356
Campbell,R.,Wasco,S.M.,Ahrens,C.E.,Sefl,T.,&Barnes,H.E.(2001).Preventingthesecondrape:
Rapesurvivorsexperienceswithcommunityserviceproviders.JournalofInterpersonal
Violence,16,12391259.
Caringella,S.(2008).Addressingrapereforminlawandpractice.NewYork:ColumbiaUniversityPress.
CentersforDiseaseControlandPrevention(2004).Sexualviolenceprevention:Beginningthedialogue.
Atlanta:Author.
Charmaz,K.(2006).Constructinggroundedtheory:Apracticalguidethroughqualitativeanalysis.
Chelimsky,E.(1997).Thepoliticalenvironmentofevaluationandwhatitmeansforthedevelopmentof
thefield.InE.Chelimsky&W.R.Shadish(Eds.),Evaluationforthe21stcentury:Ahandbook(pp.
5368).ThousandOaks,CA:Sage.
Clarke,P.(2008).Whencangrouplevelclusteringbeignored?Multilevelmodelsversussinglelevel
modelswithsparsedata.JournalofEpidemiologyandCommunityHealth,62,752758.doi:
10.1136/jech.2007.060798
ClayWarner,J.,&McMahonHoward,J.(2009).Rapereporting:Classicrapeandthebehaviorof
law.Violenceandvictims,24,723743.
Coie,J.D.,Watt,N.F.,West,S.G.,Hawkins,J.D.,Asarnow,J.R.,Markman,H.J.,Ramey,S.L.,Myrna
S.B.,&Long,B.(1993).Thescienceofprevention:aconceptualframeworkandsomedirections
foranationalresearchprogram.AmericanPsychologist,48,1013.
Cole,J.,&Logan,T.(2008).Negotiatingthechallengesofmultidisciplinaryresponsestosexualassault
victims:Sexualassaultnurseexaminerandvictimadvocacyprograms.ResearchinNursing&
Health,31,7685.
Corbin,J.,&Strauss,A.(2008).Basicsofqualitativeresearch(3rded.).LosAngeles,CA:Sage.
Cordray,D.S.(2000).Enhancingthescopeofexperimentalinquiryininterventionstudies.Crime&
Delinquency,46,401424.
Corti,L.(2007).Reusingarchivedqualitativedatawhere,how,why?.ArchivalScience,7,3754.
Cousins,J.B.,&Chouinard,J.A.(Eds.).(2012).Participatoryevaluationupclose:Anintegrationof
researchbasedknowledge.IAP.
Cowan,E.J.,&Koppl,R.(2010).AneconomicperspectiveonUnanalyzedevidenceinlawenforcement
agencies.Criminology&PublicPolicy,9,411419.
357
Creswell,J.W.(2010).Mappingthedevelopinglandscapeofmixedmethodsresearch.InA.Tashakkori
&C.Teddlie(Eds.)TheSagehandbookofmixedmethodsinsocialandbehavioralresearch(2nd
ed.)(pp.4568).ThousandOaks,CA:Sage.
Creswell,J.W.(2012).Qualitativeinquiryandresearchdesign:Choosingamongfiveapproaches(3rd
ed.).ThousandOaks,CA:Sage.
Creswell,J.W.,&Clark,V.P.(2011).Designingandconductingmixedmethodsresearch(2nded.).
Creswell,J.W.,Clark,V.L.,Gutmann,M.L.,&Hanson,W.E.(2003).Advancedmixedmethodsresearch
designs.InA.Taskakkori&C.Teddlie(Eds.),Handbookofmixedmethodsinsocialand
behavioralresearch(p.209240).ThousandOaks:Sage.
Cross,T.,Walsh,W.,Simone,M.,&Jones,L.(2003).Prosecutionofchildabuse:Ametaanalysisofrates
ofcriminaljusticedecisions.Trauma,Violence&Abuse,4,323340.
daSilva,G.T.,Logan,B.R.,&Klein,J.P.(2009).Methodsforequivalenceandnoninferioritytesting.
BiologyofBloodandMarrowTransplantation,15,120127.doi:10.1016/j.bbmt.2008.10.004
Davidson,E.J.(2005).Evaluationmethodologybasics:Thenutsandboltsofsoundevaluation.
Davies,P.,Francis,P.,&Greer,C.(Eds.).(2007).Victims,crimeandsociety.London:Sage.
Decker,S.E.,Naugle,A.E.,CarterVisscher,R.,Bell,K.,&Seifert,A.(2011).Ethicalissuesinresearchon
sensitivetopics:Participantsexperiencesofdistressandbenefit.JournalofEmpiricalResearch
onHumanResearchEthics:AnInternationalJournal,6,5564.
DepartmentofJustice(2006).Nationaltrainingstandardsforsexualassaultmedicalforensicexaminers.
Washington,DC:Author.
DepartmentofJustice(2013).Anationalprotocolforsexualassaultmedicalforensicexaminations:
Adults&adolescents(2nded.).Washington,DC:Author.
DicksonSwift,V.,James,E.L.,&Liamputtong,P.(2008).Whatissensitiveresearch?.Undertaking
SensitiveResearchintheHealthandSocialSciences:ManagingBoundaries,EmotionsandRisks,
110.
Domino,J.L.(2012).TheroleofthevictimperpetratorrelationshipinthedevelopmentofPTSDand
depressionfollowingsexualassault.(Unpublisheddoctoraldissertation).AuburnUniversity,
Montgomery,AL.
358
Donaldson,S.I.,Christie,C.A.,&Mark,M.M.(Eds.).(2009).Whatcountsascredibleevidenceinapplied
researchandevaluationpractice?.ThousandOaks,CA:Sage.
DuMont,J.&White,D.(2007)Theusesandimpactsofmedicolegalevidenceinsexualassaultcases:A
globalreview.Geneva,Switzerland:TheWorldHealthOrganization
DuMont,J.,White,D.,&McGregor,M.J.(2009).Investigatingthemedicalforensicexaminationfrom
theperspectivesofsexuallyassaultedwomen.SocialScience&Medicine,68,774780.
doi:10.1016/j.socscimed.2008.11.010
Eagly,A.H.,&Riger,S.(2014).Feminismandpsychology:Critiquesofmethodsandepistemology.
AmericanPsychologist(inpress)
Edney,J.J.(1981).Paradoxesonthecommons:Scarcityandtheproblemofequality.Journalof
CommunityPsychology,9,334.
Edwards,H.,&Gotsonis,C.(2009).StrengtheningforensicscienceintheUnitedStates:Apath
forward.StatementbeforetheUnitedStateSenateCommitteeontheJudiciary.
EklandOlson,S.,&Martin,S.(1988).Organizationalcompliancewithcourtorderedreform.Lawand
SocietyReview,22,359384.
Emerson,R.M.,Fretz,R.I.,&Shaw,L.L.(1995).Writingethnographicfieldnotes.Chicago:Universityof
ChicagoPress.
Emerson,R.M.,&Pollner,M.(1988).Ontheusesofmembers'responsestoresearchers'
accounts.HumanOrganization,47,189198.
Erickson,F.(1986).Qualitativemethodsinresearchonteaching.InM.C.Wittrock(Ed.),Handbookof
researchonteaching(pp.119161).London:Macmillan.
Evergreen,S.D.H.(2014).Presentingdataeffectively:Communicatingyourfindingsformaximum
impact.ThousandOaks,CA:Sage.
FehlerCabral,G.,&Campbell,R.(2013).Adolescentsexualassaultdisclosure:theimpactofpeers,
families,andschools.AmericanJournalofCommunityPsychology,52,7383.
Fetterman,D.M.(2010).Ethnography:Stepbystep(3rded.).ThousandOaks,CA:Sage.
Finkelhor,D.,Wolak,J.,&Berliner,L.(2001).Policereportingandprofessionalhelpseekingforchild
crimevictims:Areview.ChildMaltreatment,6,1730.
359
Fisher,B.A.,Cullen,F.T.,&Turner,M.G.(2000).Thesexualvictimizationofcollegewomen.
Washington,DC:NationalInstituteofJustice.
Fisher,R.P.&Geiselman,R.E.(2010).hecognitiveinterviewmethodofconductingpoliceinterviews:
Elicitingextensiveinformationandpromotingtherapeuticjurisprudence.InternationalJournal
ofLawandPsychiatry,33,321328.
Foa,E.B.,KeaneT.M.,&FriedmanM.J.(Eds.).(2000).EffectivetreatmentsforPTSD:Practiceguidelines
fromtheinternationalsocietyfortraumaticstressstudies.NewYork:GuilfordPress.
Frazier,P.,&Haney,B.(1996).Sexualassaultcasesinthelegalsystem:Police,prosecutorandvictim
perspectives.LawandHumanBehavior,20,607628.
Frazier,P.,Tashiro,T.,Berman,M.,Steger,M.,&Long,J.(2004).Correlatesoflevelsandpatternsof
positivelifechangesfollowingsexualassault.JournalofConsultingandClinicalPsychology,
72,1930.
Fried,M.(1982).Endemicstress:Thepsychologyofresignationandthepoliticsofscarcity.American
JournalofOrthopsychiatry,52,419.
Frohmann,L.(1997a).Discreditingvictimsallegationsofsexualassault:Prosecutorialaccountsofcase
rejections.SocialProblems,38,213226.
Frohmann,L.(1997b).Convictabilityanddiscordantlocales:Reproducingrace,class,andgender
ideologiesinprosecutorialdecisionmaking.Law&SocietyReview,31,531556.
Fry,D.A.(2007).ARoomofOurOwn:sexualsurvivorsevaluateservices.NewYork:NewYorkCity
AllianceAgainstSexualAssault
Fry,D.A.,Messinger,A.M.,Rickert,V.I.,OConnor,M.K.,Palmetto,N.,Lessel,H.,&Davidson,L.L.
(2013).Adolescentrelationshipviolence:Helpseekingandhelpgivingbehaviorsamongpeers.
JournalofUrbanHealth,91,320334.
Funk,R.E.(2006).Queermenandsexualassault:whatbeingrapedsaysaboutbeingaman.InC.
Kendall&W.Martino(Eds.),Genderedoutcastsandsexualoutlaws:Sexualoppressionand
genderhierarchiesinqueermenslives(pp.131146).Binghamton,NY:HarringtonParkPress.
Gaines,L.,Worral,J.,&Southerland,M.(2003).Policeadministration(2nded.).Boston,MA:McGraw
Hill.
Gardiner,J.C.,Luo,Z.,&Roman,L.A.(2009).Fixedeffects,randomeffectsandGEE:Whatarethe
differences?StatisticsinMedicine,28,221239.doi:10.1002/sim.3478
360
Garfinkel,S.N.,&Liberzon,I.(2009).AReviewofneuroimagingfindings.PsychiatricAnnals,39,370
381.
Gelman,A.,&Hill,J.(2007).Dataanalysisusingregressionandmultilevel/hierarchicalmodels.New
York,NY:CambridgeUniversityPress.
Glaser,B.G.(2007).Doingformaltheory.InA.Bryant&K.Charmaz(Eds.),TheSAGEHandbookof
groundedtheory(pp.97113).ThousandOaks,CA:Sage.
Glaser,B.G.,&Strauss,A.L.(1967).Thediscoveryofgroundedtheory:Strategiesforqualitative
research.Chicago:Aldine.
Gidycz,C.A.,&Koss,M.P.(1990).Acomparisonofgroupandindividualsexualassaultvictims.
PsychologyofWomenQuarterly,14,325342.
Grauerholz,L.(2000).Anecologicalapproachtounderstandingsexualrevictimization:Linkingpersonal,
interpersonal,andsocioculturalfactorsandprocesses.Childmaltreatment,5,517.
Greene,J.C.(2007).Mixedmethodsinsocialinquiry.SanFrancisco,CA:JosseyBass.
Greenwood,D.J.,&Levin,M.(2006).Introductiontoactionresearch(2nded.).ThousandOaks,CA:
Sage.
Greeson,M.R.,Campbell,R.,&FehlerCabral,G.(2014a).ColdorCaring?AdolescentSexualAssault
Victims'PerceptionsofTheirInteractionsWiththePolice.ViolenceandVictims,29,636651.
Greeson,M.G.,Campbell,R.,&FehlerCabral,G.(2014b).Nobodydeservesthis:Adolescentsexual
assaultvictimsperceptionsofdisbeliefandvictimblamefrompolice.JournalofCommunity
Psychology.
Guest,G.,Bunce,A.,&Johnson,L.(2006).Howmanyinterviewsareenough?Anexperimentwithdata
saturationandvariability.FieldMethods,18,5982.
Hanley,J.A.,Negassa,A.,&Forrester,J.E.(2003).Statisticalanalysisofcorrelateddatausinggeneralized
estimatingequations:Anorientation.AmericanJournalofEpidemiology,157,364375.
Hardin,J.W.,&Hilbe,J.M.(2002).Generalizedestimatingequations.Retrievedfrom
http://www.crcnetbase.com/isbn/9781584883074doi:10.1201/9781420035285.fmatt
HarrellJr.,F.E.(2014).Hmisc:HarrellMiscellaneous(Version3.144)[Computerprogram,Rpackage].
Nashville,TN:Author.Retrievedfromhttp://CRAN.Rproject.org/package=Hmisc
361
Harvey,MR.(1996).Anecologicalviewofpsychologicaltraumaandtraumarecovery.Journalof
TraumaticStress,9,323.
Heise,L.L.(1998).Violenceagainstwomen:Anintegrated,ecologicalframework.ViolenceAgainst
Women,4,262290.
Hemmens,C.T.T.,Brody,D.C.,&Spohn,C.C.C.(2012).Criminalcourts:Acontemporaryperspective.
Henderson,S.,&Segal,E.H.(2013).Visualizingqualitativedatainevaluationresearch.NewDirections
forEvaluation,2013,5371.
Henry,G.T.,&Mark,M.M.(2003).Beyonduse:Understandingevaluationsinfluenceonattitudesand
actions.AmericanJournalofEvaluation,24,293314.
Herman,JL.(1992).ComplexPTSD:Asyndromeinsurvivorsofprolongedandrepeatedtrauma.Journal
ofTraumaticStress,5,377391.
Herman,J.L.(2002).PosttraumaticStressDisorder:AComprehensiveText?TreatingPsychological
TraumaandPTSD.AmericanJournalofPsychiatry,159,14521453.
HesseBiber,S.N.(Ed.)(2007).Handbookoffeministresearch:Theoryandpraxis.ThousandOaks,CA:
Sage.
Heyl,B.S.(2001).Ethnographicinterviewing.InP.Atkinsons,A.Coffey,S.Delamont,J.Lofland,&L.
Lofland(Eds),Handbookofethnography(pp.369383).London:Sage.
Hill,M.R.(Ed.).(1993).Archivalstrategiesandtechniques.SagePublications.
Hosmer,D.W.,Lemeshow,S.,&Sturdivant,R.X.(2013).Appliedlogisticregression(3rded.).Hoboken,
NJ:JohnWiley&Sons,Inc.
Hox,J.J.(2010).Multilevelanalysis:Techniquesandapplications(2nded.).NewYork,NY:Routledge.
Hudzik,J.,&Cordner,G.(1983).Planningincriminaljusticeorganizationsandsystems.NewYork:
McMillan.
HumanRightsWatch(2009).Testingjustice:TherapekitbackloginLosAngelesCityandCounty.New
York:Author.
HumanRightsWatch(2010).IusedtothinkthelawwouldprotectmeIllinoissfailuretotestrapekits.
NewYork:Author.
362
HumanRightsWatch(2013).Capitoloffense:PolicemishandlingsexualassaultcasesintheDistrictof
Columbia.NewYork:Author.
IAFN(2009).Forensicnursing:Scopeandstandardsofpractice.SilverSpring,MD:AmericanNurses
Association.
Jewkes,R.,Sen,P.,&GarciaMoreno,C.(2002).Sexualviolence.InE.G.Krug,L.L.Dahlberg,J.A.Mercy,
A.B.Zwi,&R.Lozano(Eds.),Worldreportonviolenceandhealth(pp.147181).Geneva:World
HealthOrganization.
Jobling,M.A.,&Gill,P.(2004).Encodedevidence:DNAinforensicanalysis.NatureReviewsGenetics,5,
739752.
Johnson,K.(1985).Ifyouareraped:Whateverywomanneedstoknow.HolmesBeach,FL:Learning
Publications,Inc.
Johnson,J.C.(1990).Selectingethnographicinformants.ThousandOaks,CA:Sage.
Johnson,J.,Hall,J.,Greene,J.C.,&Ahn,J.(2013).Exploringalternativeapproachesforpresenting
evaluationresults.AmericanJournalofEvaluation,doi:1098214013492995.
Johnson,D.,Peterson,J.,Sommers,I.,&Baskin,D.(2012).Useofforensicscienceininvestigatingcrimes
ofsexualviolence:Contrastingitstheoreticalpotentialwithempiricalrealities.ViolenceAgainst
Women,18,193222.
Jones,J.S.,Rossman,L.,Wynn,B.N.,Dunnuck,C.,Schwartz,N.(2003).Comparativeanalysisofadult
versusadolescentsexualassault:Epidemiologyandpatternsofanogenitalinjury.Academic
EmergencyMedicine,10,872877.
Kelley,K.D.,&Campbell,R.(2013).Movingonordroppingout:Policeprocessingofadultsexualassault
cases.Women&CriminalJustice,23,118.
Kelly,J.G.(1966).Ecologicalconstraintsonmentalhealthservices.AmericanPsychologist,21,535553.
Kelly,J.G.(1968).Towardsanecologicalconceptionofpreventiveinterventions.InJ.W.Carter,Jr.(Ed.),
Researchcontributionsfrompsychologytocommunitymentalhealth(pp.7599).NewYork:
BehavioralPublications.
Kelly,J.G.(1970).Antidotesforarrogance:Trainingforcommunitypsychology.AmericanPsychologist,
25,524531
Kelly,J.G.(1971).Qualitiesforthecommunitypsychologist.AmericanPsychologist,26,897903.
363
Kelly,J.G.(2006).Becomingecological:Anexpeditionintocommunitypsychology.NewYork:Oxford
UniversityPress.
Kelly,J.G.,Ryan,A.M.,Altman,B.E.,&Stelzner,S.P.(2000).Understandingandchangingsocial
systems.InJ.Rappaport&E.Seidman(Eds.),Handbookofcommunitypsychology(pp.133159).
NewYork,NY:KluwerAcademic/PlenumPublishers.
Kemmis,S.,&McTaggart,R.(2005).Participatoryactionresearch:Communicativeactionandthepublic
sphere.InN.K.Denzin&Y.S.Lincoln(Eds.),TheSagehandbookofqualitativeresearch(3rded.)
(pp.559604).ThousandOaks,CA:Sage.
Kennedy,D.M.(2012).Dontshoot:Oneman,astreetfellowship,andtheendofviolenceininnercity
America.NewYork:Bloomsbury.
Kennedy,D.,Braga,A.A.,Piehl,A.M.,&Waring,E.J.(2001).Reducinggunviolence:TheBostongun
projectsoperationceasefire.WashingtonDC:NationalInstituteofJustice.
Kerstetter,W.(1990).Gatewaytojustice:policeandprosecutorresponsetosexualassaultagainst
women.JournalofCriminalLaw&Criminology,81,267313.
Kilpatrick,D.G.,Resnick,H.S.,Ruggiero,K.J.,Conoscenti,L.M.,&McCauley,J.(2007).Drugfacilitated,
incapacitated,andforciblerape:Anationalstudy.Washington,DC:NationalInstituteofJustice.
Kingdon,J.W.(1995).Agendas,alternatives,andpublicpolicies(2nded.).NewYork:HarperCollins.
Kirkhart,K.E.(2000).Reconceptualizingevaluationuse:Anintegratedtheoryofinfluence.New
directionsforevaluation,2000,523.
Klofas,J.,Hipple,N.K.,&McGarrell,E.(Eds.).(2010).Thenewcriminaljustice:Americancommunities
andthechangingworldofcrimecontrol.NewYork,NY:Routledge.
Konradi,A.(1996a).Preparingtotestify:Rapesurvivorsnegotiatingthecriminaljusticeprocess.Gender
&Society,10,404432.
Konradi,A.(1996b).Understandingrapesurvivors'preparationsforcourt:Accountingfortheinfluence
oflegalknowledge,culturalstereotypes,personalefficacy,andprosecutorcontact.Violence
AgainstWomen,2,2562.
Konradi,A.(2007).Takingthestand:Rapesurvivorsandtheprosecutionofrapists.Westport,CT:
GreenwoodPublishingGroup.
364
Koss,M.P.,Dinero,T.E.,Seibel,C.A.,&Cox,S.L.(1988).Strangerandacquaintancerape:Arethere
differencesinthevictim'sexperience?.Psychologyofwomenquarterly,12,124.
Koss,M.P.,&Figueredo,A.J.(2004).Changeincognitivemediatorsofrape'simpactonpsychosocial
healthacross2yearsofrecovery.JournalofConsultingandClinicalPsychology,72,1063.
Koss,M.P.,Figueredo,A.J.,Bell,I.,Tharan,M.,&Tromp,S.(1996).Traumaticmemorycharacteristics:a
crossvalidatedmediationalmodelofresponsetorapeamongemployedwomen.Journalof
AbnormalPsychology,105,421.
Koss,M.P.,&Harvey,M.R.(1991).Therapevictim:Clinicalandcommunityinterventions.Thousand
Oaks,CA:SagePublications.
Koss,M.P.,Tramp,S.,&Tharan,M.(1995).Traumaticmemories:Empiricalfoundations,forensicand
clinicalimplications.ClinicalPsychology:ScienceandPractice,2,111132.
Kramer,R.M.(1990).Whenthegoinggetstough:Theeffectsofresourcescarcityongroupconflictand
cooperation.InE.J.Lawler&B.Markovsky(Eds.)Advancesingroupprocesses:Theoryand
research.Greenwich,CT:JAIPress.
Krueger,R.A.&Casey,M.A.(2008).Focusgroups:Apracticalguideforappliedresearch(4thed.).
Krug,E.G.,Mercy,J.A.,Dahlberg,L.L.,&Zwi,A.B.(2002).Theworldreportonviolenceandhealth.The
lancet,360,10831088.
Langhout,R.D.(2003).Reconceptualizingquantitativeandqualitativemethods:Acasestudydealing
withplaceasanexemplar.AmericanJournalofCommunityPsychology,32,229244.
Ledray,L.E.(1999).Sexualassaultnurseexaminer(SANE)developmentandoperationsguide.
Washington,DC:OfficeforVictimsofCrime,U.S.DepartmentofJustice.
Ledray,L.E.,Burgess,A.W.,&Giardino,A.P.(2011).Medicalresponsetoadultsexualassault:A
resourceforcliniciansandrelatedprofessionals.St.Louis,MO:STMLearning.
Lee,R.M.(2000).Doingresearchonsensitivetopics.ThousandOaks,CA:Sage.
Levin,H.M.,&McEwan,P.J.(2001).Costeffectivenessanalysis:Methodsandapplications(2nded.).
Leviton,L.C.,&Hughes,E.F.(1981).ResearchOntheUtilizationofEvaluationsAReviewand
Synthesis.EvaluationReview,5,525548.
365
Liamputtong,P.(2007).Researchingthevulnerable:Aguidetosensitiveresearchmethods.London:
Sage.
Liamputtong,P.(2011).Focusgroupmethodology:Principleandpractice.London:SagePublications.
Lincoln,Y.S.,&Guba,E.G.(1985).Establishingtrustworthiness.Naturalisticinquiry,289,331.
Lipsky,L.V.(2009).Traumastewardship:Aneverydayguidetocaringforselfwhilecaringforothers.San
Francisco:BerrettKoehlerPublishers.
Lisak,D.(2008).Understandingthepredatorynatureofsexualviolence.Retrievedfrom:http://www.
innovations.harvard.edu/showdoc.html.
Lisak,D.,&Miller,P.M.(2002).Repeatrapeandmultipleoffendingamongundetectedrapists.Violence
&Victims,17,7384.
Logan,T.K.,Cole,J.,&Capillo,A.(2007).Differentialcharacteristicsofintimatepartner,acquaintance,
andstrangerrapesurvivorsexaminedbyaSexualAssaultNurseExaminer(SANE).Journalof
InterpersonalViolence,22,10661076.
Lonsway,K.A.,&Archambault,J.(2012).Thejusticegapforsexualassaultcases:Futuredirectionsof
researchandreform.ViolenceAgainstWomen,18,145168.
Lord,V.B.,&Rassel,G.(2000).Lawenforcement'sresponsetosexualassault:Acomparativestudyof
ninecountiesinNorthCarolina.Women&CriminalJustice,11,6788.
Lovrich,N.P,Pratt,T.C.,Gaffney,M.J.,Johnson,C.L.,Asplen,C.H.,Hurst,L.H.&Schellberg,T.M.
(2004).NationalforensicDNAstudyreport.Washington,DC:NationalInstituteofJustice.
Lumley,T.(2004).Analysisofcomplexsurveysamples.JournalofStatisticalSoftware,9,119.
Lumley,T.(2010).Complexsurveys:AguidetoanalysisusingR.InM.P.Couper,G.Kalton,J.N.K.Rao,
N.Schwarz&C.Skinner(SeriesEds.),WileySeriesinSurveyMethodology,Retrievedfrom
http://onlinelibrary.wiley.com/book/10.1002/9780470580066doi:10.1002/9780470580066
Lumley,T.(2012).Survey:analysisofcomplexsurveysamples(Version3.282)[Computerprogram,R
package].Seattle,WA:Author.Retrievedfromhttp://CRAN.Rproject.org/package=survey
Lynch,V.A.(2006).Forensicnursingscience.InR.M.Hammer,B.Moynihan,&E.M.Pagliaro(Eds.),
Forensicnursing:Ahandbookforpractice(pp.140).Sudbury,MA:JonesandBartlett.
366
MacQueen,K.M.,McLellanLemal,E.,Bartholow,K.,&Milstein,B.(2008).Teambasedcodebook
development:Structure,process,andagreement.InG.GuestandK.M.MacQueen(Eds.),
Handbookforteambasedqualitativeresearch(pp.119136).Lanham,MD:Altamira/Rowman
&Littlefield.
Maddox,L.,Lee,D.,&Barker,C.(2011).PoliceempathyandvictimPTSDaspotentialfactorsinrapecase
attrition.JournalofPoliceandCriminalPsychology,26,112117.
Maier,S.L.(2008).Ihaveheardhorriblestories:Rapevictimadvocatesperceptionsofthe
revictimizationofrapevictimsbythepoliceandmedicalsystem.ViolenceAgainstWomen,14,
786808.
Maier,S.L.(2011).Sexualassaultnurseexaminersperceptionsoftherevictimizationofrape
victims.JournalofInterpersonalViolence
Martin,P.Y.(2005).Rapework:Victims,gender,andemotionsinorganizationandcommunitycontext.
NewYork,Routledge.
Martin,C.E.,Houston,A.M.,Mmari,K.N.,&Decker,M.R.(2012).Urbanteensandyoungadults
describedrama,disrespect,datingviolenceandhelpseekingpreferences.MaternalandChild
HealthJournal,16,957966.
Martin,P.Y.,&Powell,M.(1995).Accountingforthesecondassault:legalorganizationsframingof
rapevictims.LawandSocialInquiry,20,853890.
McDonald,S.(2005).Studyingactionsincontext:aqualitativeshadowingmethodfororganizational
research.QualitativeResearch,5,455473.
McEwen,T.(2003).Evaluationofthelocallyinitiativeresearchpartnershipprograms.Washington,DC:
TheNationalInstituteofJustice.
McGarrell,E.F.,Hipple,N.K.,Corsaro,N.,Bynum,T.S.,Perez,H.,Zimmermann,C.A.,&Garmo,M.
(2009).Projectsafeneighborhoods:anationalprogramtoreduceguncrime:finalprojectreport.
USDepartmentofJustice,OfficeofJusticePrograms,NationalInstituteofJustice.
McGarrell,E.F.(2010a).Accumulatinglessonsfromprojectsafeneighborhoods.InJ.M.Klofas,N.K.
Hippie,&E.F.McGarrell(Eds.),Thenewcriminaljustice:Americancommunitiesandthe
changingworldofcrimecontrol(pp.3950).NewYork:Routledge.
McGarrell,E.F.(2010b).Strategicproblemsolving,projectsafeneighborhoods,andthenewcriminal
justice.InJ.M.Klofas,N.K.Hippie,&E.F.McGarrell(Eds.),Thenewcriminaljustice:American
communitiesandthechangingworldofcrimecontrol(pp.3950).NewYork:Routledge.
367
McNeish,D.(2014).Modelingsparselyclustereddata:Designbased,modelbased,andsinglelevel
methods.PsychologicalMethods,Advanceonlinepublication.doi:10.1037/met0000024
McWhorter,S.K.,Stander,V.A.,Merrill,L.L.,Thomsen,C.J.,&Milner,J.S.(2009).Reportsofrape
reperpetrationbynewlyenlistedNavypersonnel.Violence&Victims,24,204218.
Meehan,A.(2000).Theorganizationalcareerofgangstatistics:Thepoliticsofpolicinggangs.The
SociologicalQuarterly,41,337370.
Mertens,D.M.(2008).Transformativeresearchandevaluation.NewYork:Guilford.
Mertens,D.M.,&HesseBiber,S.(2013).Mixedmethodsandcredibilityofevidenceinevaluation.New
DirectionsforEvaluation,2013,513.
MatthewB.,Miles,Huberman,A.M.,&Saldaa,J.(2014).QualitativeDataAnalysis:AMethods
Sourcebook.ThousandOaks,CA:Sage.
Miller,R.L.,Forte,D.,Wilson,B.D.M.,&Greene,G.J.(2006).Protectingsexualminorityyouthfrom
researchrisks:Conflictingperspectives.AmericanJournalofCommunityPsychology,37,341
348.
Millar,G.,Stermac,L.,&Addison,M.(2002).Immediateanddelayedtreatmentseekingamongadult
sexualassaultvictims.Women&Health,35,5364.
Minkler,M.(2012).Communityorganizingandcommunitybuildingforhealthandwelfare(3rded.).
Piscataway,NJ:RutgersUniversityPress.
Minkler,M.,&Wallerstein,N.(2008).Communitybasedpartcipatoryresearchforhealth:Fromprocess
tooutcome(2nded.).NewYork:Wiley.
Monroe,L.M.,Kinney,L.M.,Weist,M.D.,Dafeamekpor,D.S.,Dantzler,J.,&Reynolds,M.W.(2005).
Theexperienceofsexualassault:Findingsfromastatewidevictimneedsassessment.Journalof
InterpersonalViolence,20,767776.
Morell,J.A.(2010).Evaluationinthefaceofuncertainty:Anticipatingsurpriseandrespondingtothe
inevitable.NewYork:GuilfordPress.
Morgan,D.L.,&Krueger,R.A.(1997).Thefocusgroupkit.ThousandOaks,CA:Sage.
Morse,J.(1994).Emergingfromthedata:Thecognitiveprocessesofanalysisinqualitative
inquiry.InJ.Morse(Ed.)Criticalissuesinqualitativeresearchmethods(pp.2343).Thousand
Oaks,CA:Sage.
368
Mullainathan,S.,&Shafir,E.(2013).Scarcity:Whyhavingtoolittlemeanssomuch.NewYork:Henry
HoltBooks.
Murphy,E.,&Dingwall,R.(2001).Theethicsofethnography.InP.Atkinsons,A.Coffey,S.Delamont,J.
Lofland,&L.Lofland(Eds),Handbookofethnography(pp.339351).London:Sage.
NationalCenteronDomesticViolence,Trauma&MentalHealth(2011).Creatingtraumainformed
services:Tipsheetseries.Chicago,IL:Author.
NationalCenterforVictimsofCrime.(2008).SexualAssault.Retrievedfrom
http://www.ncvc.org/ncvc/main.aspx?dbName=DocumentViewer&DocumentID=32369
Nelson,M.S.(2010).MakingsenseofDNAbacklogs,2010mythsvs.reality.Washington,DC:The
Nelson,M.S.(2013).AnalysisofuntestedsexualassaultkitsinNewOrleans.Washington,DC:The
Neville,H.A.,&Heppner,M.J.(1999).Contextualizingrape:Reviewingsequelaeandproposinga
culturallyinclusiveecologicalmodelofsexualassaultrecovery.Applied&Preventative
Psychology,8,4162.
Parnis,D.,&DuMont,J.(2006).Symbolicpowerandtheinstitutionalresponsetorape:Uncoveringthe
culturaldynamicsofaforensictechnology.TheCanadianReviewofSociologyandAnthropology,
43,7393.
Patterson,D.(2011a).Thelinkagebetweensecondaryvictimizationbylawenforcementandrapecase
outcomes.JournalofInterpersonalViolence,26,328347.
Patterson,D.(2011b).Theimpactofdetectivesmannerofquestioningonrapevictimsdisclosure.
ViolenceAgainstWomen,17,13491373.
Patterson,D.,&Campbell,R.(2010).Whyrapesurvivorsparticipateinthecriminaljusticesystem.
JournalofCommunityPsychology,38,191205.
Patterson,D.&Campbell,R.(2012).Theproblemofuntestedsexualassaultkits:Whyaresomekits
neversubmittedtoacrimelaboratory?JournalofInterpersonalViolence,27,22592275.
Patterson,D.,Greeson,M.R.,&Campbell,R.(2009).Protectthyself:Understandingrapesurvivors
decisionsnottoseekhelpfromsocialsystems.Health&SocialWork,34,127136.
Patton,M.Q.(1998).Discoveringprocessuse.Evaluation,4,225233.
369
Patton,M.Q.(2002).Qualitativeresearch&evaluationmethods(3rded.).ThousandOaks,CA:Sage.
Patton,M.Q.(2008).Utilizationfocusedevaluation(4thed.).ThousandOaks,CA:Sage.
Patton,M.Q.(2011).Developmentalevaluation:Applyingcomplexityconceptstoenhanceinnovation
anduse.NewYork:GuilfordPress.
Paul,L.A.,Zinzow,H.M.,McCauley,J.L.,Kilpatrick,D.G.,&Resnick,H.S.(2013).Doesencouragement
byothersincreaserapereporting?Findingsfromanationalsampleofwomen.Psychologyof
WomenQuarterly,doi:0361684313501999.
Peak,K.J.&Glensor,R.W.(1996).Communitypolicingandproblemsolving:Strategiesandpractices.
UpperSaddleRiver,NJ:PrenticeHall.
Peterson,J.L.,Hickman,M.J.,Strom,K.J.,Johnson,D.J.(2013).Effectofforensicevidenceoncriminal
justicecaseprocessing.JournalofForensicSciences,58,7890.
Peterson,J.L.,Johnson,D.,Herz,D.,Graziano,L.,&Oehler,T.(2012).Sexualassaultkitbacklogstudy.
Washington,DC:TheNationalInstituteofJustice.
Peterson,J.,Sommers,I.,Baskin,D.,&Johnson,D.(2010).Theroleandimpactofforensicevidencein
thecriminaljusticeprocess.Washington,DC:TheNationalInstituteofJustice.
Pierce,S.J.,&Zhang,Z.(2011a).Projectednumbersofevidencekitsexpectedtoreachdifferentstagesof
forensicanalysis:Asummarybasedondatafromthe400Project.EastLansing,MI:Michigan
StateUniversity,CenterforStatisticalTrainingandConsulting.
Pierce,S.J.,&Zhang,Z.(2011b).Projectednumbersofevidencekitsexpectedtoreachdifferentstagesof
forensicanalysis,part2:Theimpactofalternativecasestatusscreeningscenariosontheresults
fromthe400Project.EastLansing,MI:MichiganStateUniversity,CenterforStatisticalTraining
andConsulting.
PoliceExecutiveResearchForum.(1997).Policechiefsurvey.Retrievedfrom:
http://www.policeforum.org/library/police
management/Police%20Executive%20Research%20Forum.pdf.
Porter,L.E.,&Alison,L.J.(2006).Examininggrouprape:Adescriptiveanalysisofoffenderandvictim
behaviour.EuropeanJournalofCriminology,3,357381.
Preskill,H.,&Caracelli,V.(1997).Currentanddevelopingconceptionsofuse:EvaluationuseTIGsurvey
results.AmericanJournalofEvaluation,18,209225.
370
Prior,L.(2012).Theroleofdocumentsinsocialresearch.InS.Delamont(Ed.),HandbookofQualitative
ResearchinEducation(pp.426438).Northampton,MA:EdwardElgarPublishing,Inc.
RDevelopmentCoreTeam.(2014).R:Alanguageandenvironmentforstatisticalcomputing(Version
3.1.1)[Computerprogram].Vienna,Austria:RFoundationforStatisticalComputing.
http://www.Rproject.org.
Rainguet,F.&Dodge,M.(2001).Theproblemsofpolicechiefs:Anexaminationoftheissuesintenure
andturnover.PoliceQuarterly,4,268288.
Raudenbush,S.,&Bryk,A.(2002).Hierarchicallinearmodels:Applicationsanddataanalysismethods
(2nded.).ThousandOaks,CA:SagePublications.
Raudenbush,S.,Bryk,A.,Cheong,Y.F.,&Congdon,R.(2011).HLM7:Hierarchicallinearandnonlinear
modeling.Chicago,IL:ScientificSoftwareInternational.
Renzetti,C.M.,&Lee,R.M.(1992).Researchingsensitivetopics.ThousandOaks,CA:Sage.
Resnick,H.,Acierno,R.,Holmes,M.,Dammeyer,M.,&Kilpatrick,D.(2000).Emergencyevaluationand
interventionwithfemalevictimsofrapeandotherviolence.JournalofClinicalPsychology,56,
13171333.
Riggs,N.,Houry,D.,Long,G.,Markovchick,V.,&Feldhaus,K.M.(2000).Analysisof1,076casesofsexual
assault.Annalsofemergencymedicine,35,358362.
Ritter,N.(2011).Theroadahead:Unanalyzedevidenceinsexualassaultcases.Washington,DC:The
Roehl,J.,Rosenbaum,D.P.,Costello,S.K.,Coldren,J.R.,Schuck,A.M.,Kunard,L.,&Forde,D.R.(2006).
Strategicapproachestocommunitysafetyinitiative(SACSI)in10U.S.cities:Thebuildingblocks
forprojectsafeneighborhoods.WashingtonDC:U.S.DepartmentofJustice.
Rogers,J.L.,Howard,K.I.,&Vessey,J.T.(1993).Usingsignificanceteststoevaluateequivalence
betweentwoexperimentalgroups.PsychologicalBulletin,113,553565.doi:10.1037/0033
2909.113.3.553
Roman,J.K.,Reid,S.E.,Chalfin,A.J.,&Knight,C.R.(2009).TheDNAfieldexperiment:arandomized
trialofthecosteffectivenessofusingDNAtosolvepropertycrimes.JournalofExperimental
Criminology,5,345369.
Roozendaal,B.,McEwen,B.S.,&Chattarji,S.(2009).Stress,memoryandtheamygdala.NatureReviews
Neuroscience,10,423433.
371
Rosenbaum,D.P.,&Roehl,J.(2010).Buildingsuccessfulantiviolencepartnerships:Lessonsfromthe
strategicapproachestocommunitysafetyinitiative(SACSI)model.InJ.M.Klofas,N.K.Hippie,&
E.F.McGarrell(Eds.),Thenewcriminaljustice:Americancommunitiesandthechangingworldof
crimecontrol(pp.3950).NewYork:Routledge.
Rosenthal,J.A.(1996).Qualitativedescriptorsofstrengthofassociationandeffectsize.JournalofSocial
ServiceResearch,21,3759.doi:10.1300/J079v21n04_02
Rossi,P.H.,Lipsey,M.W.,&Freeman,H.E.(2003).Evaluation:Asystematicapproach.Sage
Publications.
Roux,C.,Goldsmith,K.&Bonezzi,A.(2012).Ontheconsequencesofscarcity:Whenlimitedresources
promoteagenticresponses.Evanston,IL:NorthwesternUniversity,KelloggSchoolof
Management.
Rubin,D.C.,Berntsen,D.,&Bohni,M.K.(2008).Amemorybasedmodelofposttraumaticstress
disorder:EvaluatingbasicassumptionsunderlyingthePTSDdiagnosis.Psychologicalreview,115,
985.
Rubin,H.J.&Rubin,I.S.(2011).Qualitativeinterviewing:Theartofhearingdata(3rded.).Thousand
Oaks,CA:Sage.
Rush,P.D.(2010).CriminallawandthereformationofrapeinAustralia.InC.McGlynn&V.E.Munro
(Eds.),Rethinkingrapelaw:Internationalandcomparativeperspectives(pp.237250).NewYork,
NY:Routledge.
Saldana,J.(2003).Longitudinalqualitativeresearch:Analyzingchangethroughtime.WalnutCreek,CA:
AltaMira.
Sagan,C.(1995).Thedemonhauntedworld:Scienceasacandleinthedark.NewYork,NY:Random
House.
Sandelowski,M.(1993).Rigororrigormortis:Theproblemofrigorinqualitativeresearch
revisited.AdvancesinNursingScience,16,18.
Sandelowski,M.(1994).Focusonqualitativemethods:Theuseofquotesinqualitativeresearch.
ResearchinNursing&Health,17,479482.
Sandelowski,M.(1995).Samplesizeinqualitativeresearch.ResearchinNursing&Health,18,179183.
Sarkar,D.(2008).Lattice:MultivariatedatavisualizationwithR.NewYork,NY:Springer
Science+BusinessMedia,LLC.
372
Schensul,J.J.(2009).Community,cultureandsustainabilityinmultileveldynamicsystemsintervention
science.Americanjournalofcommunitypsychology,43,241256.
Schensul,J.J.,&Trickett,E.(2009).Introductiontomultilevelcommunitybasedculturallysituated
interventions.AmericanJournalofCommunityPsychology,43,232240.
Scherer,R.(2014).PropCIs:Variousconfidenceintervalmethodsforproportions(Version0.25)
[COmputerprogram,Rpackage].Hannover,Germany:Author.Retrievedfromhttp://CRAN.R
project.org/package=PropCIs
Scriven,M.(1997).Truthandobjectivityinevaluation.InE.Chelimsky&W.R.Shadish(Eds.)Evaluation
forthe21stCentury(pp.477500).ThousandOaks,CA:Sage.
Seal,D.W.,Eldrige,G.D.,&Kacanek,D.(2007).Alongitudinal,qualitativeanalysisofthecontextof
substanceuseandsexualbehavioramong18to29yearoldmenaftertheirreleasefrom
prison.SocialScience&Medicine,65,23942406.
Seidman,E.(1988).Backtothefuture,communitypsychology:Unfoldingatheoryofsocialintervention.
AmericanJournalofCommunityPsychology,16,422.
SexualViolenceJusticeInstitute(2008).Becomingvictimcentered.Minneapolis,MN:MNCASA.
Shaw,J.(2014).Justifyinginjustice:Howthecriminaljusticesystemexplainsitsresponsetosexual
assault.(Unpublisheddoctoraldissertation).MichiganStateUniversity,EastLansing,MI.
Shaw,J.andtheSARA(SexualAssaultResourceAnalysis)Project.(2011)Sexualassaultresponsetraining
forlawenforcement,medical,andadvocacyprofessionals.SexualAssaultReport,15,34,4244.
Shaw,J.,&Campbell,R.(2013).Sexualassaultkit(SAK)submissionamongadolescentrapecasestreated
inforensicnurseexaminerprograms.Inpress:JournalofInterpersonalViolence.
Sher,L.(2010).Neurobiologyofsuicidalbehaviorinposttraumaticstressdisorder.ExpertReviewof
Neurotherapeutics,10,12331235.
SieberJ.E.(2004).Ethicalprinciples.InM.S.LewisBlack,A.Bryman,&T.F.Liao(Eds.),Encyclopediaof
socialscienceresearchmethods(pp.323326).ThousandOaks,CA:Sage.
SingletonJr,R.A.,Straits,B.C.,&Straits,M.M.(1993).Approachestosocialresearch(2nded.).New
York,NY:OxfordUniversityPress.
373
Smith,D.D.(1974).Thesocialconstructionofdocumentaryreality.SociologicalInquiry,44,257268.
Smith,D.E.(2005).Institutionalethnography:Asociologyforpeople.Lanham,MD:AltaMira.
Southwick,S.M.,Rasmusson,A.,Barron,J.,&Arnsten,A.(2005).Neurobiologyandneurocognitive
alterationsinPTSD:Afocusonnorephinephrine,serotonin,andthehypothalamicpituitary
adrenalaxis.InJ.J.Vasterling&C.R.Brewin(Eds.)NeuropsychologyofPTSD:Biological,
Cognitive,andClinicalPerspectives.NewYork:GuilfordPress
Spohn,C.,&Holleran,D.(2001).Prosecutingsexualassault:Acomparisonofchargingdecisionsinsexual
assaultcasesinvolvingstrangers,acquaintances,andintimatepartners.JusticeQuarterly,18,
651688.
Spohn,C.&Spears,J.(1996).Theeffectofoffenderandvictimcharacteristicsonsexualassaultcase
processingdecisions.JusticeQuarterly,13,647679.
Spohn,C.&Tellis,K.(2012).Thecriminaljusticesystemsresponsetosexualviolence.ViolenceAgainst
Women,18,169192.
Stake,R.E.(1997).Advocacyinevaluation:Anecessaryevil?InE.Chelimsky&W.R.Shadish(Eds.),
Evaluationforthe21stcentury:Ahandbook(pp.5368).ThousandOaks,CA:Sage.
Stake,R.E.(2010).Qualitativeresearch:Studyinghowthingswork.NewYork,NY:GuilfordPress.
Stake,R.E.(2013).Multiplecasestudyanalysis.NewYork,NY:GuilfordPress.
Starks,H.,&Trinidad,S.B.(2007).Chooseyourmethod:Acomparisonofphenomenology,discourse
analysis,andgroundedtheory.QualitativeHealthResearch,17,13721380.
Steenkamp,M.M.,Dickstein,B.D.,SaltersPedneault,K.,Hofmann,S.G.,&Litz,B.T.(2012).
TrajectoriesofPTSDsymptomsfollowingsexualassault:Isresiliencethemodaloutcome?.
JournalofTraumaticStress,25,469474.
Stegner,B.L.,Bostrom,A.G.,&Greenfield,T.K.(1996).Equivalencetestingforuseinpsychosocialand
servicesresearch:Anintroductionwithexamples.EvaluationandProgrammingPlanning,19,
193198.doi:10.1016/01497189(96)000110
Stevens,A.P.(2001).Arrestingcrime:ExpandingthescopeofDNAdatabasesinAmerica.TexasLaw
Review,79,921960.
Stewart,D.W.,Shamdasanu,P.N.,&Rook,D.(2006).Focusgroups:Theory&practice(2nded.).
374
Streiner,D.L.,&Sidani,S.(Eds.).(2011).Whenresearchgoesofftherails:whyithappensandwhatyou
candoaboutit.NewYork,NY:GuilfordPress.
Stringer,E.T.(2013).Actionresearch(4thed.).ThousandOaks,CA:Sage.
Strom,K.J.&Hickman,M.J.(2010).Unanalyzedevidenceinlawenforcementagencies:Anational
examinationofforensicprocessinginpolicedepartments.Criminology&PublicPolicy,9,381
404.
Subramanian,S.V.,&O'Malley,A.J.(2010).Modelingneighborhoodeffects:Thefutilityofcomparing
mixedandmarginalapproaches.Epidemiology,21,475478.
Sudman,S.,Bradburn,N.M.,&Schwarz,N.(1996).Thinkingaboutanswers:Theapplicationof
cognitiveprocessestosurveymethodology.SanFrancisco:JosseyBass.
Swarthout,A.G.,Swarthout,K.M.,&White,J.W.(2011).Whatyourdatadidnttellyouthefirsttime
around:Advancedanalyticapproachestolongitudinalanalyses.ViolenceAgainstWomen,17,
309321.
Tetlock,P.E.(2000).Copingwithtradeoffs:Psychologicalconstraintsandpoliticalimplications.InA.
Lupid,M.D.McCubbins,&S.L.Popkin(Eds.),Elementsofreason:Cognition,choice,andthe
boundsofrationality.Cambridge,UK:CambridgeUniversityPress.
TheWeinerReport(2002).DNAjustice:Casessolvedatlast.NewYork:Author.
Tjaden,P.,&Thoennes,N.(2006).Extent,nature,andconsequencesofrapevictimization:Findingsfrom
theNationalViolenceAgainstWomenSurvey.Washington,DC:NationalInstituteofJustice.
Tofte,S.(2013,August6).GuestblogSarahTofte:Policeoftenabandonedrapekitsand
investigations.Cleveland.com.RetrievedAugust6,2013,fromhttp://www.cleveland.com/rape
kits/index.ssf/2013/08/guest_blog_sarah_tofte.html
Torrance,H.(2012).Triangulation,respondentvalidation,anddemocraticparticipationinmixed
methodsresearch.Journalofmixedmethodsresearch,doi:1558689812437185.
Trickett,E.J.(1984).Towardadistinctivecommunitypsychology:Anecologicalmetaphorforthe
conductofcommunityresearchandthenatureoftraining.AmericanJournalofCommunity
Psychology,12,261279.
Trickett,E.J.(1996).Afutureforcommunitypsychology:Thecontextsofdiversityandthediversityof
contexts.AmericanJournalofCommunityPsychology,24,209234.
375
Trickett,E.J.(2011).FromwaterboilinginaPeruviantowntolettingthemdie:Culture,community
intervention,andmetabolicbalancebetweenpatienceandzeal.AmericanJournalof
CommunityPsychology,47,5868.
Trickett,E.J.,Kelly,J.G.,&Vincent,T.A.(1985).Thespiritofecologicalinquiryincommunityresearch.
InE.C.Susskind&D.C.Klein(Eds.)Communityresearch:Methods,paradigms,andapplications
(pp.283333).NewYork:Praeger.
Ullman,S.E.(2007).Comparinggangandindividualrapesinacommunitysampleofurbanwomen.
ViolenceandVictims,22,4351.
Ullman,S.E.(2010).Talkingaboutsexualassault:Societysresponsetosurvivors.Washington,DC:
AmericanPsychologicalAssociation.
Ullman,S.E.,&Brecklin,L.R.(2002).Sexualassaulthistory,PTSD,andmentalhealthserviceseekingina
nationalsampleofwomen.JournalofCommunityPsychology,30,261279.
Ullman,S.E.,Filipas,H.H.,Townsend,S.M.,&Starzynski,L.L.(2006).Theroleofvictimoffender
relationshipinwomenssexualassaultexperiences.JournalofInterpersonalViolence,21,798
819.
Ullman,S.E.,&Siegel,J.M.(1993).Victimoffenderrelationshipandsexualassault.Violenceand
Victims,8,121134.
U.S.CensusBureau.(n.d.).AmericanFactFinder.Retrievedfrom:
http://factfinder2.census.gov/faces/nav/jsf/pages/searchresults.xhtml?refresh=t
UnitedStatesDepartmentofHealthandHumanServices.(1979).TheBelmontReport:ethicalprinciples
andguidelinesfortheprotectionofhumansubjectsofresearch.Washington,DC:US
DepartmentofHealthandHumanServices.
UnitedStatesDepartmentofJustice,FederalBureauofInvestigation.(2000).CrimeintheUnited
States2000.Retrievedfromhttp://www.fbi.gov/aboutus/cjis/ucr/crimeintheu.s/2000
Valliant,R.,Dever,J.A.,&Kreuter,F.(2013).Practicaltoolsfordesigningandweightingsurveysamples.
InS.E.Fienberg&W.J.Linden(SeriesEds.),StatisticsforSocialandBehavioralSciences,
Retrievedfromhttp://link.springer.com/openurl?genre=book&isbn=9781461464488
Walsh,V.C.(1961).Scarcityandevil.EnglewoodCliffs,NJ:PrenticeHall.
376
Warnes,G.R.,Bolker,B.,Lumley,T.,&Johnson,R.C.(2013).gmodels:VariousRprogrammingtoolsfor
modelfitting(Version2.15.4.1)[Computerprogram,Rpackage].Albany,GA:Author.Retrieved
fromhttp://CRAN.Rproject.org/package=gmodels
Weiss,C.H.(1980).Knowledgecreepanddecisionaccretion.ScienceCommunication,1,381404.
Weiss,C.H.(1998).Havewelearnedanythingnewabouttheuseofevaluation?AmericanJournalof
Evaluation,19,2133.
Weiss,C.H.(2004).Rootingforevaluation:Acliffnotesversionofmywork.InM.C.Alkin(Ed.),
EvaluationRoots:TracingTheoristsViewsandInfluences.ThousandOaks,CA:Sage.
Weiss,C.H.,MurphyGraham,E.,&Birkeland,S.(2005).Analternateroutetopolicyinfluencehow
evaluationsaffectDARE.AmericanJournalofEvaluation,26,1230.
Wellek,S.(2010).Testingstatisticalhypothesesofequivalenceandnoninferiority,(2nded.).Retrieved
fromCRCnetBasedatabaseathttp://www.crcnetbase.com/isbn/9781439808184
White,J.W.,&Kowalski,R.M.(1998).Violenceagainstwomen:Anintegrativeperspective.Perspectives
onHumanAggression,205229.
White,J.W.,Koss,M.P.,&Kazdin,A.E.(Eds.)(2011).Violenceagainstwomenandchildren:Volume1:
Mappingtheterrain.Washington,DC:AmericanpsychologicalAssociation.
Wolcott,H.F.(2005).Theartoffieldwork(2nded.).WalnutCreek,CA:AltaMira.
Yin,R.K.(2008).Casestudyresearch:Designandmethods(4thed.).ThousandOaks,CA:Sage.
Yin,R.K.(2009).CaseStudyResearch:DesignandMethods,Essentialguidetoqualitativemethodsin
organizationalresearch.ThousandOaks,CA:Sage.
Yin,R.K.(2011).Applicationsofcasestudyresearch.ThousandOaks,CA:Sage.
Young,W.,Bracken,A.C.,Goddard,M.A.A.,&Matheson,S.(1992).TheNewHampshiresexualassault
medicalexaminationprotocolprojectcommitteesexualassault:reviewofanationalmodel
protocolforforensicandmedicalevaluation.ObstetGynecol,80,878883.
Zinzow,H.M.,Resnick,H.S.,Barr,S.C.,Danielson,C.K.,&Kilpatrick,D.G.(2012).Receiptofpostrape
medicalcareinanationalsampleoffemalevictims.AmericanJournalofPreventiveMedicine,
43,183187.
377
APPENDIX B: Project Methodology

ThepurposeofthisAppendixistodescribethedatacollectionmethodsandanalysesusedin
thisactionresearchproject.Inordertomakethisreportasaccessibleaspossibletodiverseaudiences
ofresearchersandpractitioners,wedidnotprovideextensivemethodologicaldetailsineachchapter;
instead,wepresentedoverviewsofthedatasourcesandanalyticapproachesusedtogeneratethe
findings.InthisAppendix,wewillnowprovidethosedetailstraditionallycoveredintheMethods
sectionofresearchreports/manuscripts,including:102
1.
Guidingevaluationtheoryfortheproject(DevelopmentalEvaluationTheory)
2.
Guidingresearchdesignfortheproject(MultiStage,SequentialExploratoryMixedMethodsDesign)
3.
Datacollectionmethods(includingsampling,procedures,andreliabilityassessments)
4.
Dataanalysismethods(qualitative&quantitative)
TheintendedaudienceforthisAppendixistheresearchcommunity,andthereforeit
presupposesaworkingknowledgeofqualitativeandquantitativeresearchdesignandanalysis.
PractitionersmayalsofindtheinformationinthisAppendixtobehelpfulinthatitdetailsthenatureof
datacollection,thespecificdatasourcesobtained,andthechallengesofconductingcommunitybased
actionresearchprojects.Togroundthereadersinthegoalsoftheprojectandthedatacollection
methods,wehavereproducedTable1.1(OverviewofSAKActionResearchProjectDataCollection
Methods)andTable1.2(OverviewoftheQuantityofDataCollectionintheSAKActionResearchProject)
(followingpages)fromChapter1:Introduction,whichsummarizethefourmethodsusedtoassesseach
ofthefourmainprojectgoalsandthequantityofdatacollected(organizedbymethod).
102
Allresearch/evaluationdatacollectioninstrumentscanbefoundinAppendixD:DataCollectionInstruments.
378
TABLE1.1OverviewofSAKActionResearchProjectDataCollectionMethods
Goal1:
ArchivalRecords
FocusGroups
DocumenttheDiscoveryoftheUnsubmittedSAKsandEventsThatOccurredPriortotheBeginningofCensus
ConductaCensusof
SAKsinPoliceProperty aboutthediscoveryofthe
unsubmittedSAKsandthe
interorganizational
communicationsthereafter
aboutthediscoveryofthekitsand
theeventsthereafter
Reviewinterorganizationalmemos
re:thediscoveryofthekits&intra
organizationalinvestigations/audits
Reviewmediareportsaboutthe
discoveryofthekits
(NA)
DocumenttheProcessoftheSAKCensus
Observetheprocessof
conductingthecensusfor
keyquestions,issues,and
decisions
aboutthesuccesses,challenges,and
lessonslearnedfromconductinga
census
propertydatabase
forensiclaboratorytesting
spreadsheet
Discusssuccesses,
challenges,and
lessonslearned
re:conductinga
census
Goal2:
DocumentHistoricalContextinDetroitSexualAssaultOrganizations
Identifythe
UnderlyingFactors
Re:WhyDetroitHas
UnsubmittedSAKs
aboutpolicies,practices,and
resourcesavailableineach
organizationovertime
eachdisciplinere:dailyoperations/
servicesprovided;staffinglevels;
andreportingstructure,trainingand
supervision
Interviewpublicofficialsinfour
comparablecitiesre:services&
staffinglevels
staffing,&resourcesinDetroit
organizations(19892009)
staffing,&resourcesfromsexual
assaultorganizationsinfour
comparablecities(Philadelphia,
Dallas,Baltimore,&NewOrleans)
(NA)
379
TABLE1.1(continued)
Goal2(continued)
ArchivalRecords
FocusGroups
ExamineFrontLineServices&DecisionMakinginSexualAssaultCases
aboutstandardoperating
procedures&decisionmaking
insexualassaultcases
eachdisciplinere:decisionmaking
processesinsexualassaultcases
Goal3:
DocumentProcessofDevelopingTestingPlan
DevelopSAKTesting
PlanandEvaluate
Efficacy
Observetheprocessof
developingaSAKtestingplan
re:keyquestions,issues,and
decisions
Reviewcriminalsexualassault
policereports(19892009)re:
investigationalpracticesand
decisionmaking
(NA)
InterviewDetroitstakeholdersand
(NA)
stakeholdersinnationalorganizations
concernedwithcriminaljustice,
forensicsciences,andviolenceagainst
womenregardingthepurpose&utility
ofSAKtesting
Discusssuccesses,
challenges,and
lessonslearned
re:developinga
testingplan
(NA)
(NA)
EvaluateTestingPlan
Observegroupdiscussionsre:
testingresultsandthe
implicationsofthefindings
Reviewpolicefilesassociated
withSAKstestedinthisproject
forvictim,assailant,andcase
characteristics
Documentnumberandtypeof
CODIShitsassociatedwithSAKs
testedinthisproject
380
TABLE1.1(continued)
Goal4:
DocumentProcessofDevelopingVictimNotificationProtocols
DevelopVictim
NotificationProtocols
andEvaluateEfficacy
Observetheprocessof
developingvictimnotification
protocolsre:keyquestions,
issues,anddecisions
InterviewDetroit
stakeholdersand
stakeholdersinnational
organizationsconcernedwith
criminaljustice,forensic
sciences,andviolenceagainst
womenregardinghowand
whenvictimsshouldbe
notifiedabouttestingresults
ArchivalRecords
(NA)
FocusGroups
Discusssuccesses,
challenges,andlessons
learnedre:victim
notificationandthe
utilityofaformalized
VictimNotification
ReviewTeam
EvaluateVictimNotificationProtocols
Documenttheprocessesand
decisionsoftheVictim
NotificationReviewTeam
(NA)
Trackinvestigatorseffortstolocate (NA)
victimsfornotificationand
notifications
Documentcommunitybased
advocates(deidentified)
notifications
381
TABLE1.2OverviewoftheQuantityofDataCollectedintheDetroitSAKARP
METHODS
DATAQUANTITY
Ethnographic
Observations
N=81observations,~186hoursofobservation
n=53collaborativeteammeetings,~106hoursofobservation
n=18impromptumeetings,~32hoursofobservation
n=6shadowingobservationsofstakeholdersconductingtheirjobs,~18hoursofobservation
n=1planningretreat,~12hoursofobservation
n=3VictimNotificationReviewteammeetings,~18hoursofobservation
Individual
Interviews
N=42formalinterviewswithDetroitstakeholders(16onetime/crosssectionalinterviews;
26longitudinalinterviews10peopleinterviewedtwotimes,2peopleinterviewthreetimes)
N=187informalinterviewswithDetroitstakeholders(30people,numberofinterviewsvaried)
N=5interviewswithnationalcriminaljustice/forensicsciencestakeholders
N=5interviewswithnationalviolenceagainstwomenorganizationstakeholders
N=35interviewswithpublicofficialsincomparablescities
ArchivalRecords
N=2databasesreviewedre:thenumberofunsubmittedSAKs:policepropertydatabase
(~11,000entries)andpoliceforensicsciencestestingspreadsheet(~2,500entries)
N=5intraandinterorganizationalrecords(andN=6mediareports)re:discoveryofthe
unsubmittedSAKsinAugust,2009
N=93(publicallyavailableandinternal)fromDetroitorganizationsre:leadership,staffing,
resources,servicesprovided,andpolicies&proceduresovertime
N=33records(publicallyavailable)fromorganizationsincomparablecitiesre:leadership,
staffing,andresources
N=1,268policereportsreviewedre:investigationalpracticesanddecisionmakinginsexual
assaultcasesandcodedforvictim,assailant,&casecharacteristics
N=1,595SAKDNAtestingresultsreports
N=31investigatorrecordsandN=18communitybasedadvocaterecordsre:victim
notificationsconductedinthisproject
FocusGroups
N=3focusgroupsre:successes,challenges,andlessonslearned
382
Guiding Evaluation Theory: Developmental Evaluation Theory

Actionresearchprojectsare,bytheirverynature,dynamicenterprises(Klofas,Hipple,&
McGarrell,2010)anditwastobeexpectedthattheDetroitcollaborativewouldworkina(sometimes)
nonlinearprocessasinformationemergedanddecisionschangedinlightofnewknowledge.This
flexibilityundoubtedlybenefitstheprojectinmanyrespects,butitdoesposesomechallengesinthe
contextofanevaluation.Traditionally,evaluationresearchassumesafarmorestaticevaluand(that
whichistobeevaluated)theinterventionisset,processesareinplace,andtheseentitiesmustremain
unchanged(asmuchaspossible)soasnottocompromisethedesignandvalidityoftheevaluation
(Davidson,2005;Rossi,Lipsey,&Freeman,2003;Scriven,1997).However,thepresumptionofan
unchangingevaluandisunrealisticinmanyfieldresearchprojects(Cordray,2000;Morell,2010;Streiner
&Sidani,2011),andassuch,newtheoreticalmodelsarenecessarythatnotonlytoleratesuchambiguity
butactuallyembracethedynamicnatureofanintervention.Inotherwords,astheevaluanditself
changes,theevaluationcanaccommodatethosechangestocontinuetocapturetheprocessesand
outcomesoftheintervention.
Consistentwiththeseideas,Patton(2011)proposedacomprehensivetheoryofevaluation
practicetermeddevelopmentalevaluation.ConsistentwithPattonslongstandingpositionthat
evaluationismosthelpfultopractitionersandpolicymakerswhenitisapproachedfromautilization
perspective(seePatton,2008forareview),thedevelopmentalapproachwascreatedtoguide
evaluationsunderconditionsofcomplexity,whenaninterventionorinitiativeisinherently
complicatedandchanging,butinformationaboutitsimpactiscriticalforintendedusers.AsPatton
(2011)described:
383
Complexenvironmentsforsocialinterventionsandinnovationsarethoseinwhichwhattodoto
solveproblemsisuncertainandkeystakeholdersareinconflictabouthowtoproceed.Informed
bysystemsthinkingandsensitivetononlineardynamics,developmentalevaluationsupports
socialinnovationandadaptivemanagement(p.1)
Thedevelopmentalapproachrecognizesthatprogramsoftenchangeandoftenshouldchange
throughoutimplementationtoimproveuponidentifiedweaknesses.Assuch,developmental
evaluationsareaniterativeprocessofmixedmethodsdatacollectionwherebythedataareanalyzed
quicklyandreportedbacktostakeholdersregularly,sothattheprogramcanberevisedandimproved.
Patton(2011)notedthatthisfocusonthereallifechallengesofsocialproblemsolvingmakes
developmentalevaluationanaturalfitwithinanactionresearchparadigm.Asstakeholderswrestlewith
thechallengesofidentifyingsolutionstotheproblem,developmentalevaluationprovidesaframework
forcapturingtheprocessesandoutcomesintheactionresearchmodel.
Developmentalevaluationisnotastepbystepguidebook,butratheraframeworkforreflective
practice(Patton,2011).First,theevaluationbeginswithdefiningthefocalquestions,andthismustbea
collaborativeendeavorwithrepresentationfromallstakeholders.Theevaluatorsroleissomewhat
secondaryintheprocessasthecommunityitselfmustidentifyitsownquestions,buttheevaluator
shouldhelpstakeholdersunderstandtheimplicationsofselectingvariousquestions.Mostcritically,the
evaluatormusteducatetheteamastothedifferencebetweenevaluationquestionsthatseektoassess
themerit,worth,andsignificanceofanevaluation(whichtypicallyrequireafarmorestaticevaluand
thanislikelyinadevelopmentalproject)andevaluationquestionsthatseektocaptureprocessand
understandhowandwhyinterventions/initiativesworkastheydo.Theevaluatorisresponsiblefor
helpingtheteamselectquestionsthatareappropriateforthenatureoftheproject.
Second,oncetheguidingquestionsareestablished,theevaluatorworkswiththestakeholders
toidentifythekindofdatanecessarytoaddressthosequestions.Developmentalevaluationtheorydoes
384
notdictateapreferreddesignormethodology,butbecausequalitativemethodsareparticularlyadept
atcapturedynamicprocesses,theyarecommonlyusedwithinthisevaluationtheory.Though
qualitativemethodsareoftencriticizedfortheirlimitationswithrespecttogeneralizabilitytoother
localesandsituations(seeStake,1995;Yin2009forreviewsanddebates),Stake(2010)arguedthat
qualitativeinvestigationsthatfocusonmechanismshowandwhythingshappen(ordonothappen)
caninfacthavebroadreachbyidentifyingcoreprocessesthattranscenduniquecontextualconditions.
Inotherwords,qualitativemethodscanbeinstrumentalinidentifyingtransportablelessons
applicableacrossarangeofcontextsprovidedthatdatacollection(andanalysis)movebeyonda
descriptiveleveltoseekoutexplanatorymechanisms.Toincreasethelikelihoodofbroaderapplicability,
Stake(2010)andPatton(2011)recommendtheuseofmultipledatasourcesincluding,butnotlimited
to,interviewing,observationaltechniques,andarchivalrecordreview.
Third,theevaluator(perhapswithassistancefromthestakeholdersthemselves)willgatherthe
necessarydata,whichareanalyzedforpatternsandthemes.Thisprocessisalsohighlycollaborativeas
thestakeholdersareinvitedtoengagethedatatohelpidentifyimplications,lessons,andanactionplan
thatmaynotbeobvioustoanevaluator,whichstrengthenstheapplicabilityoftheconclusions.Finally,
thefindingsaresharedwithstakeholdersandtheirconstituents,feedbackissolicited,andanothercycle
ofreflectivepracticemaybeinitiatedtorefineorexpandupontheknowledgegleanedinthe
evaluation.Throughoutallsteps,theevaluatormustattunetohistoricalandcontextualfactorsthat
shapethenatureofintervention/initiative,aswellastothepotentiallydivergentandcompetingpoints
ofviewoftheinitiative.
DevelopmentalevaluationprovidedausefultheoreticalmodelfortheDetroitSAKAction
ResearchProject.Thecoreelementsofwhatmakesinterventions/initiativescomplexnonlinearity,
interactingelements,uncertaintyandconflictwereallpresentinthisinitiative.Atthebeginningofthis
project,thecollaborativewasuncertainashowbesttoprocessapproximately10,000kitsinamanner
385
thatsimultaneouslybalancestheneedsforpublicsafety,socialjustice,fiscalresponsibility,and
victim/survivorwellbeing.Itwastobeexpectedthatmanyideaswouldbeproposed,debated,perhaps
putintoaction,andperhapsabandonedinfavorofrevisedideas,whichisthenatureofanaction
researchmodel.Suchanapproachrequiredaflexibleandnimbleevaluationmodelthatcancapturethe
evolvingprocessofsocialproblemsolving.
Guiding Research Design: Sequential Exploratory Mixed Methods Design

Mixedmethodsresearchdesignsarecommonlyusedindevelopmentalevaluationtheoryas
theyprovideflexible,diversetoolsforcapturingeventsastheyunfoldinrealtime.Inthisproject,we
usedCreswellssequentialexploratorymixedmethodsdesignasanoverarchingmethodological
frameworkforunitingthefourdistinctgoalsofthisproject(conductingacensus,studyingunderlying
factors,developingatestingplan,anddevelopingvictimnotificationprotocols)(Creswell,2010;Creswell
&Clark,2011;Creswelletal.,2003).Thismultistagedesignbeginswithqualitativedatacollectionand
analysistoexplorethecontextofthesetting,itspeople,history,andsuccessesandchallenges.From
thoseexploratorydata,refinedresearchquestionsandhypothesesaregenerated,whicharethen
evaluatedinasecondstageofquantitativedatacollectionandanalysis(seeFigureB.1,below).
FIGUREB.1SequentialExploratoryMixedMethodsDesigns
386
Inthepurestapplicationofthisdesign,themethodsaredistinctwithrespecttotimingthe
qualitativeworkisfinished(i.e.,allanalysescompleted)andtheresultsgeneratedfromthatmethodare
thenusedtoplanthesubsequentquantitativestudy,butinpractice,itismoretypicaltousepreliminary
findingsfromthequalitativeworktoguidesubsequentwork,whiletheinitialresearchcontinues(see
Campbell,Shaw,&Gregory,2014forareview).Thisdesigncanbeextendedintoamultistage
application,wherebythecycleofqualitative,followedbyquantitativedatacollectionrepeats.Inthis
instance,thefollowupquantitativefindingssuggestedanideaforanewqualitativestudy(whichcould
befollowedbyanotherquantitativestudy,andsoon)(seeCampbell,Gregory,Patterson,andBybee,
2012foramultistageexample).Ineitherthetraditionalormultistageapplicationofthisdesign,
qualitativemethodstendtohavegreateremphasis(hencethecapitallettersinFigureB.1),meaning
thatthepurposeofthequantitativemethodsistoverifykeyfindingsgeneratedfromthenarrativedata.
Applyingthisdesigntothecurrentproject,webeganwithextensivequalitativedatacollection
(ethnography,archivalrecords,andindividualinterviews)tounderstandthehistoryandcurrent
operationsofeachoftheorganizationsparticipatinginthecollaborative.Forourfirstprojectgoal
(conductingthecensus),thisqualitativeworkemphasizedunderstandingthekeyeventssurrounding
thediscoveryoftheunsubmittedkits,theinterorganizationalcommunicationsthereafter,andhow
thoseeventsmightaffecttheplanningandexecutionofthecensus.Foroursecondprojectgoal
(studyingtheunderlyingfactorsoftheproblem),thefirststageofqualitativedatacollectionfocusedon
capturingchangesinleadership,staffing,andresourceswithineachorganizationovertime.Forthe
thirdprojectgoal(developingatestingplan)andfourthgoal(developingvictimnotificationprotocols),
weconductedqualitativeinterviewswithDetroitstakeholdersandnationalcriminaljustice/forensic
scienceandviolenceagainstwomenstakeholderstogatherdifferentpointsofviewregardingthe
purposeandutilityofSAKtestingandhowandwhenvictimsshouldbenotifiedabouttesting.Thisfirst
387
stageofqualitativedatacollectiontookapproximatelyoneyeartocomplete,thoughsomemethods
(e.g.,ethnography,informalinterviewing)lastedthroughouttheentiredurationoftheproject.
Preliminaryfindingsfromthesequalitativedatainformedsecondstagequantitativedata
collectionforallprojectgoals.Thecensusresultswerequantitativelygraphed,whichshowedthatthe
numberofunsubmittedSAKsvariedfromyeartoyear.Astowhysubmissionratesmighthave
fluctuatedovertime,thequalitativedatacollectedforthesecondprojectgoal(studyingunderlying
factors)hadidentifiedseveralkeychangesinpolicy,practice,andresourcesthatmighthaveaffected
SAKsubmissions.Bringingthesetwogoalsandtwodatasourcestogether,wedevelopedquantitative
hypothesesregardingratesofSAKsubmissionovertime,andusedmultilevelmodeling(MLM)totest
thosepredictions.Forourthirdprojectgoal(developingtestingplan),thefirststagequalitativework
identifiedstakeholdersbeliefsabouttheutilityofSAKtesting(e.g.,testingismoreusefulinstranger
rapecases),soweformedfourTestingGroupstotestthoseassumptionsempirically.Usingavarietyof
quantitativetechniques(e.g.,continuationratiomodeling,logisticregression,andequivalencetests),we
statisticallycomparedthenumberandtypeofCODIShitsfromdifferentkindsofsexualassaultcases.
Forthefourthprojectgoal(developingavictimnotificationprotocol),thefirststagequalitativework
highlightedthatitwouldlikelybeverychallengingtofindsurvivorsandnotifythemsomanyyearsafter
theassault;therefore,wedevelopedsimplequantitativetrackingtoolsforinvestigatorstocompleteto
capturethespecificstrategiesthatweremoreorlesssuccessfulinfindingvictims.
Forsomecomponentsoftheproject,weconductedanotherroundofqualitativedatacollection
andanalysisbecausethequantitativeresultshadraisedstillmorequestionstobeanswered.For
example,thequantitativemodelingyieldedmixedresultsastowhyratesofSAKsubmissionvariedover
time.Tounpackthosefindings,weconductedadditionalqualitativedatacollection(longitudinal
interviewsandmorearchivaldatacollection)toexploreinmoredepththefrontlineresponsetosexual
assaultcases.Withrespecttothevictimnotificationprotocols,weconductedadditionalqualitative
388
datacollectiontoassessinvestigatorsandadvocatesperceptionsregardingthenotificationsandhow
survivorsreactedtobecontactedabouttheirtestingresults.Takentogether,thesemultiplestageof
qualitativeandquantitativedatacollectionandanalysisprovidedacomprehensiveassessmentofhow
andwhyDetroithassomanyunsubmittedSAKsandhowthiscommunitycametogethertobeginto
addressthisproblemandprocessthesekits.
Data Collected in the Detroit SAK Action Research Project

Ethnographic Observations
Sampling.TheresearchteammemberswereparticipantobserversinallSAKcollaborative
meetings,whichincluded:53regularbimonthlyteammeetings,18impromptumeetings(e.g.,break
outmeetingstodiscussissuesinmoredetailthathadbeenraisedattheallteammeeting),onetwoday
victimnotificationplanningretreat,andthreeVictimNotificationReviewTeam(NRT)meetings.The
researchteamconductedanadditionalsixobservationsshadowingstakeholdersfromlawenforcement,
forensicsciences,andprosecutiontounderstandtheirworkwithsexualassaultsurvivorsinmoredepth
(McDonald,2005).Overall,theresearchteamobservedapproximately186hoursofinteractionamong
stakeholdersthroughoutthedurationoftheproject.Toensurethatallprojectmeetingswere
observed,themembersofthecollaborativeagreedtoinformtheresearchteamaboutallmeetings
(includingimpromptumeetings)andtoallowobservationalaccess(inpersonorbyphone).Therewere,
ofcourse,additionalmeetingsbetweenstakeholdersaspartoftheirnormaldaytodayworktogether
andonoccasion,projectmatterswerediscussed.Theresearchersroutinelycheckedwithkey
stakeholdersfromeachorganizationviainformalinterviewstocapturethisinformation,and
stakeholdersoftenvoluntarilycontactedtheresearcherstoletusknowaboutsuchevents.Assuch,we
arereasonablyconfidentthatwehavedocumentednearlyalloftheprojectrelatedinteractions
betweenthecollaborativepartners.
389
DataCollectionProcedures&ReliabilityAssessments.Thevastmajorityofcollaborativeteam
meetingswereobservedbytworesearchers:44of53bimonthlymeetings,15ofthe18impromptu
meetings,andtwoofthethreeVictimNotificationReviewTeammeetings(82%overall;theremaining
meetingswereobservedbyonlyoneresearcher).Foreachpairedobservation,oneresearchteam
member(typicallytheprojectsCoInvestigator)wasdesignatedobserveronlyandthatindividual
transcribedthediscussionsatthemeetingsastheywereoccurring.Consistentwithrecommendedbest
practicesintheethnographicliterature(e.g.,Emersonetal.,1995;Fetterman,2010;Wolcott,2005),
audiorecordingswerenottakenduetoconcernsregardingparticipantsreactivitytobeingrecorded.
Thoughtheresearchobserverwasnotabletocaptureeverystatementmadeinthemeetingverbatim,
themajorityofthediscussionwascapturedindirectquotes.AllmembersoftheSAKcollaborativewere
briefedindividuallyandasagroupregardingIRBproceduresforethnographicobservationssothatthey
understoodthattheirremarkswouldbewrittendown.
Thesecondresearchteammemberhadaparticipantobserverrole(typicallytheprojects
PrincipalInvestigator),consistentwiththeactionresearchparadigm.Thisresearcherparticipatedinthe
substantivediscussionsinthemeetings,andalsotooknotes,thoughnotintherunningtranscriptstyle
ofthefirstobserverrole.Thesecondobservercapturedimportantverbatimquotes,taggedsubstantive
decisionsanddisagreements,andnotedimportantinterpersonaldynamicsinthemeeting.103
Tomonitorthequalityofdatacollection,thePIauditedthetwosetsofnotes(i.e.,the
observeronlyrunningtranscriptsandtheparticipantobservernotes)toverifytheaccuracyof
verbatimquotesandkeysubstantivedecisions/debates.Inthefirstsixmonthsofdatacollection,the
notesfromallmeetingsobservedbytworesearcherswereaudited,andgiventhatdiscrepancieswere
extremelyuncommon,reliabilitycheckswerethenconductedperiodically(overall,80%ofthemeetings
103
Forthemeetingsobservedbyonlyonememberoftheresearchteam,thatindividualfunctionedintheobserveronlyrole
(seedetailsaboveregardingthenatureofthenotestaken,datacoding,etc.)
390
thathadtwoobservershadreliabilitychecks).Therewerenodiscrepanciesbetweenthetwosetsof
notesregardingsignificantgroupdecisions;discrepanciesinexactwordingofquotesoccurred
occasionallyandthosedataweretaggedandwerenotincludedintheanalyses.
InaccordwiththemethodsoutlinedbyEmerson,Fretz,andShaw(1995),fieldnoteswere
writtenandpreliminarycodingwascompletedwithin72hoursofeachobservation.Trackingsheets
weremaintainedtomonitordates/timelinesforthecompletionofthesetasks.Theobserveronly
researcherwasresponsibleforconductingopencodingofthetranscriptnotes,taggingtextthat
pertainedtoemergingthemesrelatedtoeachprojectgoal(Corbin&Strauss,2008).Theobserver
onlyresearcherwasalsoresponsibleforcreatingandmaintainingatimelinedocumentthat
summarizedtheprojectssequencesofeventsandkeydecisionsmadebythegroup.Theparticipant
observerresearcherwasresponsibleforwritingmoretraditionalfieldnotes,whichconsistedofthree
mainsections(documentedforeachmeeting):1)thickdescriptionsofthemeeting,supplementedwith
verbatimquotes;2)contentmemoingregardingemergingconcepts,hypotheses,andfindings,aswellas
ideasforadditionaldatacollection;and3)reflexivememoingregardingtheresearchersown
experiencesofconductingtheproject(seeLincoln&Guba,1985).
Individual Interviews
Sampling.Acombinationofpurposiveandsnowballsamplingmethodswereusedtoidentify
andrecruitparticipantsforindividualinterviews.Withinthefirstthreemonthsoftheproject,allcore
membersofthecollaborativewereaskedtoparticipateinaformal,oneononeinterview(i.e.,
purposivesampling;100%participationrate).Inthoseinterviews,weaskedparticipantstonominate
otherkeyindividualswithintheirorganizationsthatweshouldalsointerview,giventheirknowledgeand
expertise(i.e.,snowballsampling;100%participationrate).Overthe30monthsofproject,we
conductedatotalof42formalinterviewswithDetroitstakeholders,spanningallorganizationsandall
391
staffinglevelswithineachorganization(i.e.,frontlineworkerstoupperadministration):16werecross
sectionalinterviews(i.e.,theparticipantwasinterviewedonlyonce)and26werelongitudinalinterviews
(10individualswereinterviewedtwice,approximatelyoneyearapart;2individualswereinterviewed
threetimes,eachapproximately9monthsapart).Participantrecruitmentandinterviewingcontinued
untilweachievedsaturation,wherebythesamethemeswererepeated,withnonewthemesemerging
amongparticipants(Guest,Bruce,&Johnson,2006;Sandelowski,1995;Starks&Trinidad,2007).At42
interviews,wehadestablishedclear,discerniblepatternsinourdatawithrespecttoourfocalquestions
regardingtheresourcesavailableineachorganizationovertimeandthefrontlinepracticesand
communicationswithinandbetweenorganizationsregardingsexualassaultcaseprocessing.
Inethnographicresearch,itistypicalthatresearchershavecontactwithindividualsoutsideof
formalmeetings/settings,andtheseinformalinterviewsprovideanotherwayofdocumentingevents
throughoutaproject.UnderIRBconsentforethnographicobservation,theresearchershadongoing
informalinterviewswithrepresentativesfromeachparticipatingorganization:187informalinterviews
with30differentpeople(numberofinterviewsperpersonvariedfromthreeto39).Theseinterviews
continuedthroughouttheentiredurationoftheproject.
Wealsoconductedlimitedscaleinterviewingwithnationalstakeholdersfromcriminal
justice/forensicscienceandviolenceagainstwomenorganizations.Toidentifypotentialparticipants,
weconductedextensiveliterature/onlinesearchesonthetopicsofSAKtestingandvictimnotificationto
identifyspecificorganizations/stakeholderswhoareinvolvedinthiswork.Wepurposelyselectedfive
individualsfromcriminaljustice/forensicscienceorganizationsandfivestakeholdersfromviolence
againstwomenorganizations(100%participation).Thesesamplesizesweresufficienttoachieve
saturation,likelyduetothefactthatscopeofourinquiryfortheseinterviewswasquitefocused(see
AppendixD:DataCollectionInstruments).
392
ForourworkonunderstandingthehistoryandcontextofDetroitsexualassaultorganizations,
wecollectedleadership,staffing,andresourcedatafromfourcomparablecities(Dallas,Philadelphia,
Baltimore,andNewOrleans).Aspartofthatdatacollection,weinterviewedN=35publicofficialsin
thesecities(100%participation);again,thissamplesizewassufficienttoachievesaturation.
DataCollectionProcedures&ReliabilityAssessments.TheinterviewswiththeDetroit
stakeholdersandthenationalstakeholdersweresemistructuredqualitativeassessments(seeAppendix
D:DataCollectionInstruments).ConsistentwithPattons(2002)recommendationsforqualitative
interviewing,wesoughttoconveyanonjudgmentalstance(bothverballyandnonverbally)towards
participantsthoughts,emotions,andexperiencessothattheinvestigatordoesnotsetouttoprovidea
particularperspectiveormanipulatethedatatoarriveatpredisposedtruths(p.51).Ifstakeholders
providedinformationthatwepersonallyagreedwithorthatwesuspectedothersinthecollaborative
mightalsoagreewith,wedidnotdiscloseoursharedpointofview.Iftheyprovidedinformationthat
wedisagreedwith,wedidnotexpressthateither;instead,weaskedadditionalclarifyingquestions
aboutthecontent.Thoughtherearedebatesinthequalitativeliteratureaboutneutralityversusexplicit
engagementofdivergentpointsofview(e.g.,Greene,2007;Mertens,2008),wedidnotfeelthelatter
methodswouldhavebeeneffectiveinpromotingdisclosure,particularlydisclosureofinformationthat
somegroupsmightdisagreewithand/orfindobjectionable(seePatton,2002;Rubin&Rubin,2011).
Becausewewereseekinginformationthatspannedthreedecades,wedrewuponcognitive
interviewingtechniques(seeFisher&Geiselman,2010forareview),whichemphasizegrounding
participantsintime/date/setting/contextcuespriortoaskingfocalcontentquestionsinorderto
improvetheaccuracyofmemoryrecall.TheinterviewguidescanbefoundinAppendixD:Data
CollectionInstruments,butgiventhequalitativenatureofthiswork,thespecificcognitivecuesand
additional/supplementalquestionswerecraftedonacasebycasebasis,dependingontheparticipant
beinginterviewedandhowthatindividualhadansweredthefocalquestions.
393
TheformalinterviewswithDetroitstakeholdersexaminedwhetherthereweresystematicgaps
instaffmembersKSAs(knowledge,skills,andabilities)thatcontributedtotheproblemofuntested
kits.Furthermore,becausethereisampleliteraturesuggestingthatlegalandmedicalsystempersonnel
oftenhavenegativeandvictimblamingattitudestowardrapevictims,weexaminedhowsuchbeliefs
mighthaveaffecteddecisionmakingregardingkitprocessing.Yet,itisunlikelythatasystemicfailureof
thismagnitudecanbeadequatelyexplainedbyindividuallevelphenomena(Patton,2011);assuch,we
exploredhoworganizationallevelfactorswithineachstakeholderorganizationmayhavecontributedto
theproblem.Wediscussedhowpoliciesandproceduresinplaceatthetimemayhavecontributedto
thestockpilingofuntestedkits.Furthermore,weexaminedwhatresourceswereavailableandnot
availabletoallstakeholdergroupsacrosstheyearswhensomanykitswentuntested.Finally,because
previousresearchsuggeststhatlegalorganizationsdifferwithrespecttotheirnormsandexpectations
regardingtheimportanceofsexualassaultcasesandtheirrelativeprioritytoothercrimes(Frohmann,
1997,1998a,1998b;Martin,2005;Martin&Powell,1994),weaskedparticipantsabouttheir
organizationscultureregardingtheprocessingofsexualassaultcases.
TheinformalinterviewswithDetroitstakeholderswereoftenamixoffactualprojectupdates
withdisclosuresthatwereclearlyprivate(e.g.,ventingfrustrations,personalreflections);therefore,
consistentwithstandardpracticeinethnographicresearch(seeFetterman,2010;Wolcott,2005),only
factualinformationwasrecorded,unlesstheresearcherspecificallyaskedpermissiontomakenoteof
theothercontent(whichwasonlydonetodocumentthattherewerefrustrations,personalissues
invoked,etc.butthespecificnatureofthoseissueswasnotrecorded).
Theinterviewswithnationalcriminaljustice/forensicsciencestakeholdersandviolenceagainst
womenorganizationstakeholdersfocusedonassessingparticipantsperceptionsofthepurposeand
utilityofSAKtestingandvictimnotificationfrommultiplepointsofview(investigator,prosecution,
forensicsciences,victim/survivors).Stakeholderswerealsoaskedabouttypicalstaffinglevels/resources
394
(nationally)tohelpputourfindingsfromDetroitincontext.Theinterviewswithpublicofficialsinthe
fourcomparablecitiesassessedfactualinformationregardingservicesprovidedintheircommunity,
leadershipchangesovertime,andstaffinglevels.
TheformalinterviewswithDetroitstakeholdersweredigitallyrecorded(withparticipants
permission)andtranscribedverbatim.Allotherinterviews(informalinterviewswithDetroit
stakeholders,interviewswithnationalstakeholders,interviewswithpublicofficialsincomparablecities)
werenotrecordedandtranscribed(duetoresourcerestrictions);consistentwiththedatacollection
proceduresfortheethnographicobservations(describedabove),theinterviewerkeptdetailednotesof
theinterview,supplementedwithverbatimquotations.
ThePrincipalInvestigatorconductedthevastmajorityoftheinterviews:33oftheformal
interviewswithDetroitstakeholders(theremaining9weredonebytheCoInvestigator),allinformal
interviews,andallnationalstakeholderinterviews.ThePIhasextensiveexperienceconducting
communitybasedqualitativeinterviewsandhaspublishedextensivelyonthismethodology(e.g.,
Campbell,Adams,Wasco,Ahrens,&Sefl,2009,2010).Reliabilityassessmentsarenotgermaneto
qualitativeinterviewing,butconsistentwithrecommendationsintheliterature,peerdebriefing
occurredonanongoingbasisforqualityassurance.LincolnandGuba(1985)definepeerdebriefingas
aprocessofexposingoneselftoadisinterestedpeerinamannerparallelingananalyticalsessionand
forthepurposeofexploringaspectsoftheinquirythatmightotherwiseremainonlyimplicitwithinthe
inquirersmind(p.308).Inthecontextofthisproject,givenitscomplexconfidentialityandprivacy
concerns(seeChapter1:Introduction),disclosuretoatrulydisinterestedpartywouldhavebeen
inadvisable,buttheprojectsCoInvestigator,researchassociates,andstatisticianswereabletofulfill
thisrolebydiscussingtheinterviews,challengingtheinterviewersassumptions,identifyinggapsinthe
knowledgebase,andsuggestingareasforimprovementandnewinquiry.
395
Archival Records
Sampling.Manydifferentkindsofarchivalrecordswerecollectedforthisproject.Toascertain
thenumberofSAKsinpoliceproperty,werequestedaccesstothepolicedepartmentsproperty
database.AlistofthefieldspertainingtoallSAKsininventoryfrom1980throughNovember2009was
madeavailabletothecollaborativepartnersandtheresearchteam(~11,000entries).Theserecords
detailedwhenevidencewasenteredintoproperty,butdidnotprovideanyinformationastowhether
theSAKhadbeentested.Todeterminethatinformation,werequestedaccesstothepoliceforensic
sciencestestingspreadsheet(oncetheexistenceofthisdocumentwasknown,seeChapter3:WhySo
ManyUnsubmittedSAKsinDetroit)(~2,500entries).Thisspreadsheetverifiedwhetherakithadbeen
submittedtothelab(i.e.,whetherithadaBNumber),butitwasdifficulttoascertainfromthefields
providedwhetherinfacteachkithadbeentestedforDNA.Giventhefrequentdiscussionsregarding
theseissuesinthecollaborativeteammeetings,wearereasonablyconfidentthatalldocuments
pertainingtothenumberofSAKsinpolicepropertywereidentifiedandshared,butthatthe
completenessandutilityoftheserecordsislimited,giventhenatureofhowthedatawerecollectedand
maintainedovertimebycommunitystakeholders.
AlsoaspartofoureffortstounderstandthescopeoftheproblemofunsubmittedSAKsin
Detroit,werequestedinterandintraorganizationalrecordsfromtheagenciesthathadbeendirectly
involvedintheeventssurroundingtheAugust2009discoveryoftheunsubmittedSAKs.Fiveinternal
documentswereobtained(includinganintraorganizationinternalaffairsaudit);104somewere
voluntarilyprovided,somewereobtainedthroughFreedomofInformationAct(FOIA)requests.Itis
impossibletoknowwhetherallrelevantdocumentsonthisissuewereinfactprovidedtotheresearch
team,particularlybecausethereisevidencethatoneorganizationwithheldinformationforaprolonged
periodoftime(fromtheresearchteamandfromotherorganizations).Wespecificallyinquiredabout
104
SixmediareportsregardingthediscoveryoftheSAKswerealsoreviewed.
396
thecompletenessofourrecordsinourmemberchecks,andwearereasonablyconfidentthatevenifwe
didnotobtainallrecords,weweresuccessfulinsecuringtheprimarydocuments/exchangesbetween
organizations.
ForourworkunderstandingtheunderlyingreasonswhyDetroithassomanyunsubmittedSAKs,
wesoughttolearnthehistoryandcontextofallkeyDetroitsexualassaultorganizations(i.e.,police,
policeadvocacyprogram,policeforensiclab,prosecution,stateforensiclab,sexualassaultforensic
examprogram,communitybasedadvocacyorganizations).Werequestedrecordsfromeachagency
regardingtheirleadership,staffing,resources,servicesprovided,andpoliciesandproceduresfrom1980
to2009(e.g.,yearlybudgets,staffingrosters,SOP[standardoperatingprocedure]documents,internal
memosoutliningchangesinpolicy/practice).Fiveorganizations/programsprovidedrawdata(i.e.,
individualrecords/spreadsheets,internalmemos)andtwoorganizationsprovidedsomerawdataand
sometabulated/aggregateddata,followingalistofspecificquestionsprovidedbytheresearchers.
Overall,93individualdocumentsandaggregatedtallyreportswereprovidedtotheresearchteam.The
qualityandcompletenessoftheserecordsvariedconsiderably,aswouldbeexpectedinathirtyyear
analysis,whichprecludedexactyearbyyeardescriptivesforeachorganization;however,wewereable
tosubstantiatereliablerangesoffigures(e.g.,XXXXstaffmembersfortheyearsXXXXXXXX)forall
organizationsandwereabletodocumentthepoliciesandpracticesofallorganizationsovertime.We
alsoreviewedN=35publicallyavailabledocuments(total)fromfourcomparablecities(Dallas,
Philadelphia,Baltimore,andNewOrleans)regardingtheirleadership,staffing,andresourcesovertime.
AlsoaspartofoureffortstounderstandwhyDetroithassomanyunsubmittedSAKs,wealso
examinedsexualassaultpolicereportsasawayofdocumentingactualonthegroundpractices.There
hasbeencriticismintheliteratureregardingpolicereportsasanofficialdatasourceregarding
incidentsofsexualassault,giventhatsuchreportsareasocialconstructionoftheevents,toldbythe
police,throughtheirpointofview,andtherefore,arenotnecessarilyreflectiveofvictimslived
397
experiences(Meehan,2000;Prior,2012;Smith1974).However,forthepurposesofthisresearch,the
policepointofviewwaspreciselywhatwesoughttocapturewhatistheirofficialversionofwhat
happenedinthesecasesinwhichSAKswerenotsubmittedtoalaboratoryforanalysis?Therefore,we
requestedaccesstothepolicereportscorrespondingtothe1,595SAKsthatweretestedaspartofthis
project;328reportscouldnotbefoundbypolicepersonnel(20.5%),sowewereablereview1,268
reports.Theextentofmissingdatawasaconcerntous(andtootherorganizations),andthepolice
madesustainedeffortstolocateasmanyfilesaspossible,butintheend,itappearsthatasubstantial
numberofreportswereeithernevertaken/recordedbypolicepersonnelorwerelostovertime.
Oncethese1,595SAKsweretested,weobtainedcompleteforensictestingresultsonallkits
(100%completedata).Theforensiclaboratoriesprovidedinformationregardingwhethereachkithada
CODISeligibleprofile,whethertherewasaCODIShit,andifso,thenatureofthehit(offender,forensic,
oroffender&forensichit).
Finally,forourworkevaluatingthevictimnotificationprotocols,weobtainedN=31records
fromtheinvestigatorswhoconductedthenotificationsregardingthesteps/strategiestheyusedtofind
thesurvivorsandtheirimpressionsofhowthenotificationunfolded(100%completedata).The
investigatorsrecordedeverystrategytheyusedtofindeachvictimandhowmanytimeseachstrategy
wasused(e.g.,strategy=phonecall;numberoftimesstrategyusedinCaseA=4callsto3different
phonenumbers).Tocollecttheinvestigatorsandadvocatesimpressionsofwhatoccurredduringthe
notifications,theresearchteamprovidedtheinvestigatorswithquestionprompts(seeAppendixD:
DataCollectionInstruments)andnotificationpersonnelrecordedtheirnotesinastandardizeddatabase
within48hoursofeachnotification(100%completedata).
DataCollectionProcedures&ReliabilityAssessments.Requestsforrecordsweremadeorally
andinwritingtoallorganizations,supplementedwithalimitednumberofFreedomofInformationAct
(FOIA)requests(3).Typically,theorganizationsprovidedahardcopyorelectroniccopyoftherequested
398
information,105buttherewere12recordsthatwewereallowedtoview,butnottocopy.Inthose
instances,wetranscribedverbatimkeysectionsofthedocumentthatwererelevanttoourresearch.
Eachrecordwasreviewed(typicallybytwomembersoftheresearchteam)todeterminewhat
informationshouldbeextractedforlateranalysis.Becauseitbecameincreasinglyunwieldytomanage
somanydocuments,wecreatedanewdocument(inWord)andcopiedtext/screenshotsofthearchival
documentsintothatrunningfile(taggedbydateandsource).Fromthisprocess,weemergedwith
anotherqualitativedatafileofarchivalinformationthat,inadditiontotheethnographicfieldnotesand
interviewtranscripts,wascodedandanalyzed(seebelow)forourspecificresearchquestions.
Anotherprimarydatacollection/codingtaskweconductedwitharchivalrecordswasreviewing
andextractinginformationfromthesexualassaultpolicereportsandtheforensictestingresults.The
datacodingsheetsforthespecificvictim,assailant,andcasecharacteristics(andhowtheywerecoded)
canbefoundinAppendixD:DataCollectionInstruments.Weconstructedoperationaldefinitionsfor
eachcode,followingthepriorworkofCampbell,Bybee,Kelley,Dworkin,andPatterson(2012).Ideally,
thereportswouldhavebeencodedbytworesearcherstocomputeinterraterreliability;however,
practically,thiswasnotfeasiblegiventimeconstraints.Giventhatthecodingsheetsandcoding
proceduresusedinthisstudywerebasedonpriorwork(Campbelletal,2012),whichhadyieldedhigh
interraterreliabilityestimatesandkappacoefficients,wefeltthatitwasreasonabletoallowsingle
coders,providedthattherewerealternativemechanismsinplacetomonitordataquality.Tothatend,
wecodedingroupmeetingsof24researchers;eachcoderhadapileoffilestoreviewandanytime
therewasambiguityastohowtocodeaspecificfile/specificcode,theteampausedtheirworkto
discussthecase/variableandarriveatgroupconsensus.
105
HardcopiesofdocumentswerescannedintoPDFfilesandpasswordprotected.
399
Finally,thedatapertainingtovictimnotificationwerecodedbytworesearchers.Thefirstcoder
reviewedtherecordsprovidedbytheinvestigatorsandadvocatestodeveloppreliminarycodes/themes
regardinghowtheylocatedvictimsandhowthenotificationsunfolded.Thecodingwasindependently
reviewed/checkedbyasecondcoderandthefindingswerereviseduntilconsensuswasreached.
Focus Groups
Sampling.Inthisproject,threefocusgroupswereconductedinthefinalmonthsoftheaction
researchprojectasamethodforgatheringlessonslearnedaboutconductingthecensus,developing
thetestingplan,creatingvictimnotificationprotocols,andoverallprocessingoftheproject.Ninety
minutesattheendofthelasttworegularcollaborativeteammeetingswerereservedforfocusgroup
datacollection,andthirtyminuteswerereservedattheendofthelastvictimNotificationReviewTeam
(NRT)meeting.Thefirstfocusgroupwasattendedby13individuals,spanningfiveorganizations;the
secondfocusgroupattendedby15individuals,spanningfiveorganizations;thethird(NRTspecific)was
attendedby6individuals,spanningthreeorganizations.
DataCollectionProcedures&ReliabilityAssessments.Inpreparationforthefocusgroups,all
membersofthecollaborativeprojectweregivensummariesofourethnographicfieldnotesand
interviewsregardingkeylessonslearnedintheproject.Stakeholderswereaskedtoreviewthe
materialsandidentifyquestionsandareasfordiscussionpriortothefocusgroups.Foreachfocus
group,theprojectPIservedasthefocusgroupfacilitatorandanothermemberoftheresearchteam
tookdetailednotes,intherunningtranscriptstyleusedinallotherteammeetings(seeabove).ThePI
followedafocusgroupscript(providedinAppendixD:DataCollectionInstruments),andconsistentwith
standardpracticesforthismethod,attentionwaspaidtoensuringparticipationfromallattendees(and
monitoringtolimitparticipationfromsomesothatotherscouldengagemore)(Krueger&Casey,2008;
Liamputtong,2011;Stewart,Shamdasani,&Rook,2006).
400
Qualitative Data Analyses

Data Preparation & Triangulation Assessments
Throughoutdatacollection,weconductedbasicopencodingofthefieldnotesandinterview
transcriptstomonitoremergingthemesanddatasaturation(seeabove).Whenwewerereadytomove
forwardwithmoredetailedcodingandanalysis,webegandevelopingandexecutingtheprocedures
describedbelow(DataCoding&AnalysisProcedures),butsoonrealizedweneededtopauseand
carefullyexaminetheextenttowhichwehadtriangulationofinformationacrossdatasources.Inother
words,ifPersonXsaidYinhis/herformalinterview,wasthatinformation(Y)confirmed/triangulated
byanotherpersonwithinthatorganization,apersonfromadifferentorganization,anobservation,
and/oranarchivalrecord?Beforebuildingmorecomplexcodingandanalysesaroundthatinformation,
howcertainareweabouttheaccuracyofthatinformation?Assessingtriangulationacrossdatasources
isoftennotedasanimportantcomponentofmixedmethodsresearch/evaluation(Creswell&Clark,
2011;Donaldson,Christie,&Mark,2009;Greene,2007;Mertens&HesseBiber,2013),thoughitisnot
oftendoneinpractice(seeCampbell,Shaw,&Gregory,2014forareview),mostlikelybecauseitisa
timeconsumingandchallengingprocess.However,inthecontextofthisproject,empiricallyassessing
thecredibilityofthedatawasnecessarybecausetheorganizationsinthecollaborativepartnershiphave
hadalonghistoryofmistrust,andstakeholdersexpressedconcernsaboutwhetherindividualsmight
usetheresearchinterviews(andotherformsofdatacollection)totrytobiasormisinformtheresearch
team.Assuch,wedevelopedatriangulationassessmentprocess,bothforourownpeaceofmindthat
wewereworkingwithcredibledataandtoassurestakeholdersthattheresearchfindingswere
trustworthy.
Webeganbycreatinganewmastercopyofallthequalitativedatafiles,whichincluded:the
ethnographicobservations(runningtranscriptnotesandtraditionalfieldnotesforallmeetings),formal
interviewtranscripts,notesfrominformalinterviews,andthedocumentintegratinginformationacross
401
thevariousarchivalrecords.106Wecomparedwhatinformationwasprovidedbywhichindividuals,
throughwhattypeofdata(observation,interview,archivalrecord);then,usingavisuallyintuitive
greenyellowredcolorcodingsystem,wecodedtheextenttowhicheachkeypieceofinformationhad
beentriangulated:
darkgreen=
multipleindividualswithinanorganization,individualsoutsidethatorganization,and
multipledatatypesconfirmedinformation
green=
multipleindividualswithinanorganizationandindividualsoutsidethatorganization
confirmedinformation,ormultipledatatypesconfirmedinformation;
yellow=
informationconfirmedbymultiplepeoplewithinsameorganization(no
outside/alternatedatasourceconfirmation)
red=
informationprovidedbyonlyoneperson/onedatasource
Attheconclusionofthistask,wewereabletoreviewallofourdatafilesandascertainata
glance(bythecolorcoding)theextenttowhicheachkeypieceofinformationtobeanalyzedwas
credible.Wetriedtoclearupasmuchyellowandredinformationaspossiblebyseekingout
additionaldatafromstakeholders/archivalrecords.Iftherewereinconsistenciesbetweenstakeholders
accounts(e.g.,informationprovidedinobservationsorinterviews)andarchivalrecords,wedidfollow
upinformalinterviewersand/oremailswithstakeholdersinanattempttoclarifytheinformation.
Whendiscrepanciescouldnotberesolved(e.g.,stakeholdersrememberthingsoneway,buttherecords
suggestotherwise),weprivilegedthearchivalrecordsandusedthatinformationforouranalyses.In
instanceswhereinformationprovidedinthestakeholderinterviewscouldnotbeverifiedbyarchival
records(e.g.,recordsdidnotexist,couldnotbefound),wesoughttriangulationofthedatabycross
checkingwithotherinterviewsfrommembersofthesameorganizationandwheneverreasonable,with
106
Weretainedthepreliminaryopencodingconductedduringdatacollectioninthesemasterfiles.
402
membersfromotherorganizationswhomightbeexpectedtohavereliableinformationaboutthe
matterathand.However,thereweresomeinstancesinwhichinformationwasprovidedbyonlyone
individualandcouldnotbeverifiedbecausethatpersonwastheonlyoneoroneofonlyasmallnumber
ofpeoplewhocouldhaveknownthatinformation,andoureffortstoreachthoseotherindividualswere
unsuccessful.Excludingallyellowandreddataseemedunnecessarilyrestrictive,andbecauseitwas
arelativelyrareoccurrence(bythetimewefinishedthistriangulationtask),wedecidedtoretainthem
(stillcolorcoded)andcontinuewithadditionalcodingandanalyses,knowingthatwewouldneedtore
evaluatetheirinclusioninthefinalanalyses/report.
Data Coding & Analysis Procedures
Workingfromthesemasterfilesthatcontainedthematicopencodesandtriangulationcodes,
webeganmorefocusedcodingofthequalitativedata.Thecodingandanalysisteamconsistedoffive
individuals:thePrincipalInvestigator,theCoInvestigator,andthreeresearchassociates.Thoughitis
commoninqualitativeresearchfortheindividualwhocollectedthedatatoalsobethesoledataanalyst
(seeCharmaz,2006;Corbin&Strauss,2008;Lincoln&Guba,1985),wefeltateamapproachwasuseful
inthisprojectsothatthefindingsforeachcomponentoftheprojectwascrosscheckedandverifiedby
multipleanalysts(seeMacQueen,McLellanLemal,Bartholow,&Milstein,2008).ThePIandCoIboth
workedoneverysetofanalyses,withtheassistanceofone(ofthethree)otheranalysts(i.e.,three
analystswereinvolvedineachsetoffindingsreportedinthisdocument).
Datacodingandanalysisproceededinathreephaseprocess.First,consistentwithCorbinand
Strausss(2008)conceptofopencodingandMiles,Huberman,andSaldanas(2014)conceptofdata
condensation,thequalitativedatacollectedinthisproject(ethnographicobservations,interview
transcripts,andarchivaldocuments)wereindependentlyreviewedbytwoanalysts,who(separately)
taggedandlabeledconceptstodefineanddevelopcategoriesbasedontheirpropertiesand
403
dimensions.Afterevery23newobservations,interviews,and/orrecords,thenewdatawerechecked
againsttheexistingcodestoseeiftherewerenewcodestobecreatedandnewpatternsemerging.
Codingwasrevisedtoaccommodatethenewdata,andifwarranted,datacollectionprotocols(e.g.,
interviewquestions,archivalrecordstorequest)wererevisedtoseekoutadditional/clarifying
information.
Thesecondphaseofcodingfocusedonorganizingcomparisonsandcontrastsofthedata,akin
toCorbinandStrausss(2008)conceptofaxialcodingandMiles,Huberman,andSaldanas(2014)
datadisplayphase.Twocodersconstructedmicroleveltablesthatorganizeddatafromdifferent
sourcesbytheindividualthemes(e.g.,datafromobservations,interviews,andarchivalrecords
regardingthethemepolicestereotypesre:adolescentsexualassaultvictims).Then,thesetableswere
combinedintomacroleveltablesthatexaminedtheassociationsbetweenthemes(e.g.,police
attitudesre:adolescentsexualassaultvictimsandpoliceinvestigationaleffort).Athirdanalyst(the
PI)thencreatedadditionalmacroleveltablestoexamineassociationsamongthemeswithinand
betweeneachoftheprojectsmajorgoalstoexplorecrosscuttingassociationsbetweencomponentsof
theproject.Fromthisprocess,weemergedwithmultipleorganizationsofthedatathatallowedusto
focusatthelevelofathematiccode,projectgoal,orcrosscuttingprojectaims.
Thethirdphaseofanalysisinvolvedconstructingmechanisticlinkagesbetweenthemesand
conceptsinthedata(seeMiles,Huberman,andSaldanas[2014]drawingandverifyingconclusions
phase).Forthisphase,wedrewuponEricksons(1986)analyticinductionmethod,whichisaniterative
procedurefordevelopingandtestingempiricalassertionsinqualitativeresearch.Anassertionrefersto
ahypothesizedpatterninthedata(e.g.,policeofficersheldnegativebeliefsaboutadolescentsexual
assaultvictimsandbecauseofthoseviews,theywerelesslikelytoinvestinvestigationaleffortintheir
cases,morelikelytoquestionyoungwomenscredibility,resultinginaSAKnotbeingsubmittedfor
testing).Twoanalystsworkedtogethertodevelopassertionsfromthedata,andthenoneanalystused
404
Ericksens(1986)methodstoevaluatetheadequacyoftheassertions.Todeterminewhetheran
assertionwassubstantiated/wellfounded,theanalystassembledconfirminganddisconfirming
evidence,lookingforfivetypesofevidentiaryinadequacy:(1)inadequateamountofevidence;(2)
inadequatevarietyinthekindsofevidence;(3)faultyinterpretativestatusofevidence(i.e.,doubts
abouttheaccuracyofthedataduetosocialdesirabilitybias);(4)inadequatedisconfirmingevidence
(i.e.,nodatawerecollectedthatcoulddisconfirmakeyassertion);and(5)inadequatediscrepantcase
analysis(i.e.,nocasesexistthatarecontrarytoakeyassertion)(Erickson,1986,p.140).Assertionswere
revisedoreliminatedbasedontheirevidentiaryadequacyuntilasetofwellwarrantedassertions
remain,similartoGlasersconstantcomparisonprocess(Glaser,2007;Glaser&Strauss,1967).
Thecodingandanalysisteamthenmettoreviewthefinalassertionsanddevelopdata
visualizationstrategiesforpresentingthefindings.Qualitativeresearchreportsareoftenheavily
narrative(e.g.,longdescriptionsofthethemesandrelationshipbetweenthemes,illustratedwith
extensivequotations),whichcanbecumbersomeforstakeholderstodigest,therebylimitingthe
utilizationofthefindings(Henderson&Segal,2013).Consistentwithemergingtrendsinqualitative
evaluation(seeAzzam,Evergreen,Germuth,&Kistler,2013;Evergreen,2014;Johnson,Hall,Greene,&
Ahn,2013forreviews),wewantedtodevelopmoreaccessiblepresentationsofthedata,sowe
followedHendersonandSegals(2013)recommendationsforcreatingvisualmapsthathighlightedthe
connectionsbetweenthemes(forexample,seeFigure3.4,PoliceProcessingofSexualAssaultCases
AssociatedwithUnsubmittedSAKs).Wethenselectedwhichquotes(previouslyassembledas
confirmingevidenceoftheassertion)wouldaccompanythevisualmapsinthisreport.Selectedquotes
werereviewedtoassesspotentialidentifabilitytomembersofthecollaborativeproject,andas
necessary,materialwaslightlyeditedtoremovedistinctivespeechmannerisms,turnsofphrase,etc.
(seeSandelowski,1994regardingtheneedforsucheditingtoprotectconfidentiality).Thequotesand
selectedcasestudyexamples(forexample,thosepresentedinChapter3:WhySoManyUnsubmitted
405
SAKsinDetroit)werealsocarefullyreviewedtoensurethattheyweretypicalofwhatwesawinthe
interviews/archivalrecords.Thoughqualitativeresearchsometimesseekstohighlighttheextremes
ofasample,experience,setting,etc.,inthisproject,wefeltitwasmoreusefultodepictcommon
problemsandregularlyoccurringpatternsandlanguageinthedata.
Validity (Trustworthiness) Assessments

Inqualitativeresearchandmixedmethodsdesignsthatutilizequalitativedataresearchers
mustexplicitlyaddressthevalidityofthedata,orinqualitativenomenclature,thetrustworthinessof
thedata,giventhatthedataarenarrativeandsubjecttotheresearchersinterpretations(Creswell,
2012;Mertens&HesseBiber,2013;Milesetal.,2014).Todate,numerousstrategieshavebeen
proposedforassessingtrustworthiness(seeCreswell,2012forareview),andthestandardsoutlinedby
LincolnandGuba(1985)remainthewellregardedgoldstandardinthefieldforhighquality,rigorous
work.Specifically,LincolnandGuba(1985)articulatedfourstandardsthatqualitativeresearchersmust
meettoestablishthevalidityoftheirwork:
Credibility=
Confidenceinthetruthofthefindings
Transferabilty=
Thefindingshaveapplicabilityinothercontexts
Dependability=
Thefindingsareconsistentandcouldberepeated
Confirmability=
Thefindingareshapedbytheparticipantsandnottheresearchersbias
InTableB.1(followingpages),wedefineeachofthesestandards(andsubcriteria,asappropriate)and
describehowwesoughttomeetthatstandardinthecontextofthisactionresearchproject.
406
TABLEB.1Validity(Trustworthiness)Assessments,UsingLincoln&Gubas(1985)Standards
STANDARD:CREDIBILITY
SubCriteria
Definition
HowMetThisStandardInThisProject
Prolonged
Engagement
Spendingsufficienttimeinthefieldtolearntheculture,
setting,andphenomenonofinterest.Spending
considerabletimeobservingvariousaspectsofthe
setting,talkingwithpeopleatalllevelsofeach
organizations,anddevelopingrelationshipsandrapport
withsettingmembers.
Theresearchteamengagedin30monthsofobservationaldatacollection.In
additiontotheseformalobservations,thePIspentconsiderabletimewith
stakeholdersinbetweenmeetings(informalinterviewbyphone,crossing
pathsatothersexualassaultrelatedmeetingsandconferences,traveltoNIJ
forARPcrosssitemeetings,etc.).
Persistent
Observation
Spendingsufficienttimeobservingthespecific
characteristicsandelementsthataremostrelevantto
theproblembeingstudied.Incontrasttoprolonged
engagement(whichhelpsresearchersdevelopbreadth
andscope),thiscriteriafocusesingainingindepth
understandingofkeyelementsofthesetting.
Theresearchteamconductedsixindepthobservationsofkeyproject
activities,pertainingtoeachofthefourmaingoalsoftheproject:
Census=onsiteobservationofthepolicepropertyroomandthe
counting/auditingprocess(6hours)
Underlyingfactors=shadowingstakeholdersfromlawenforcementand
prosecutiontounderstandtheirworkwithsexualassaultsurvivors
(6hours)
Testing=shadowingforensicsciencestakeholdersre:preparingkitsfor
submissionfortestingandtrackingtestingresults(6hours)
Victimnotification=observingthreeVictimNotificationReviewteam
meetings(18hourstotal)
Totalobservationaltime(meetings+shadowing)=~186hours
407
STANDARD:CREDIBILITY(continued)
SubCriteria
Definition
HowStandardMetInThisProject
Triangulation
Collectingmultipledatasourcestoensurethattheresearchis Fourdatacollectiontechniqueswereusedinthisproject:ethnographic
detailed,comprehensive,andwelldeveloped.Therearethree observations,individualinterviews,archivalrecords,andfocusgroups.
kindsoftriangulation:
Allthreeformsoftriangulationwereassessed:
MethodsTriangulation:Checkingconsistencyoffindings
generatedbydifferentdatacollectionmethods
MethodsTriangulation:Analysescheckedwhetherdatasources
providedconsistentfindingsacrossmethod
TriangulationofSources:Checkingconsistencyofdata
sourceswithinsamemethod
TriangulationofSources:Analysescheckedwhethersourceswithin
thesametheorganizationwereconsistent
AnalystTriangulation:Usingmultipleobservers,coders,
andanalyststocheckonselectiveperception,blind
spots,andbias.
AnalystTriangulation:Analyseswereperformedbyateamof
analystsandallanalyseswereconstructedandreviewedbyatleast
threeanalysts
PeerDebriefing
Workingwithdisinterestedpeerstotest/defendtheemergent
hypothesesandtohelpilluminateimplicitassumptions.
Giventhisprojectscomplexconfidentialityandprivacyconcerns(see
Chapter1:Introduction),disclosuretoatrulydisinterestedpartywould
havebeeninadvisable,buttheprojectsCoInvestigator,research
associates,andstatisticianswereabletofulfillthisrolebydiscussingthe
interviews,challengingtheinterviewersassumptions,identifyinggaps
intheknowledgebase,andsuggestingareasfornewinquiry.
NegativeCase
Analysis
Searchingforevidencewithinthedatathatdonotsupport
emergingpatternsorinterpretations.Helpsrevise,confirm,
broadenpatterns/findingsinthedata.
ThedataanalystsusedEricksens(1986)methodsforassessing
evidentiaryadequacy/inadequacy,whichincludessearchingfor
disconfirmingevidenceofanassertion(andthenrevisingtheassertion
accordingly).
408
STANDARD:CREDIBILITY(continued)
SubCriteria
Definition
Referential
Adequacy
Reserving/archivingaportionofthedata;theresearcher
analyzesremainingdatatodevelopfindings,thenreturnsto
thearchiveddatatotestvalidityofthefindings.
Thedataanalystsbegantheirworkwiththeethnographicobservations
andinterviewdata,reservingthearchivaldatatocrosscheckand
triangulatetheinformationgatheredbytheothermethods.Inaddition,
wewereallowedaccesstotheOVWfundedThe400Projectdatasetof
N=252policefiles.Wereviewedthesefilestoassessthereferential
adequacyofouranalysesre:underlyingfactorsastowhyDetroithasso
manyunsubmittedSAKs(andspecifically,theanalysespertainingto
frontlinepracticesinsexualassaultcases).
MemberChecks
Sharingpreliminaryfindingswithparticipantstoexplorethe
extenttowhichtheresearchersinterpretationsresonatewith
theirunderstandingoftheissues.
Foreachemergingsetofanalyses(e.g.,communitycontext,underlying
factors,testingresults,victimnotification),weconductedapowerpoint
presentationofthosefindingsforthecollaborativeteam.Sometimes
weconductedseparatewithinorganizationpresentationsoffindings
(beforepresentingtothefullcollaborative)whenwesuspectedthe
findingswouldbecontroversialsothatwecouldfocusonunderstanding
eachorganizationsuniqueperspectiveonthefindings.
Thereisdebateinthequalitativemethodsliterature107about
whethermembercheckscanestablishcredibilitybecause:
Forconfidentialityreasons,participantsmaynotbeprivytoallof
thedatauponwhichtheresearchersbasedfindings
Thememberchecksdidnotyieldconsistencybetweenthestakeholders
andtheresearchersregardingthefindingsontheunderlyingfactors,
Participantsmaytrytoinfluencethefindingstocreateamore
specificallytheindirecteffectsofotherorganizationsonpolicedecision
positive,flatteringportrayal,iftheresultsareseenasnegative
making.Whereasthepolicedidnotcontestthesefindings,oneother
Researchersandparticipantsmaylegitimatelydisagreeastowhat organizationdidnotagreewiththeresearchersfindings.Theexisting
isafairaccountofthephenomenaofinterest.
datawererecheckedandnewdatawerecollectedtopursuenuances
thathelpedreviseandclarifythefinalresultspresentedinthisreport.
107
SeeAngen,2000;Ashworth,1993;Buchbinder,2010;Emerson&Pollner,1988;Miles,Huberman,&Saldana,2014;Morse,1994;Sandelowski,1993;Torrance,2012.
409
STANDARD:TRANSFERABILITY
Definition
Theresearcherneedstoprovidethereaders/audiencewithsufficientdetailabout
whathappenedinthesetting/contextofinterestsothattheycanmakeinformed
assessmentastowhetherconclusionsthatcanbedrawnaretransferabletoother
settings,situations,etc.
ThePIkeptfieldnotesthroughouttheproject,detailingthick
descriptionsofthekeyeventsthattookplacethroughoutthisproject.
Wealsocollectedextensivedetailsaboutthehistoryandcontextin
Detroitaswellasinfourcomparablecities.
Thereisdebateinthequalitativemethodsliterature108astowhetherthisstandard
requiresthatfindingsmustbetransferabletobeconsideredtrustworthy;or
whetherunique,sitespecificfindings,describedinsufficientdetailsoastogauge
theircurrent/futuretransferabilitytoothersettingsissufficient.
Thecomparablecitiescomparisonshighlightthatsomeoftheresource
constraintsinDetroitaretypicalofothermajorurbanareas(particularly
thosewithhighcrimerates/highconcentrationsofAfricanAmerican
residents),butthatinsomekeyareas(e.g.,policeleadershipturnover,
DNAforensicscientiststaffinglevels,communitybasedadvocacy),
Detroithasfaceduniquechallenges.
STANDARD:DEPENDABILITY
Definition
Theresearcherneedstoarticulatethemethodsinsufficientdetailsothatthe
processbywhichfindingswerecreatedisclear,transparent,andreproducible.To
thatend,theresearchermustkeepanaudittrail,whichincludes:datacollection
procedures,codingprocedures,analyses,revisions,memberchecks,etc.
ThePIandCoIkeptanaudittrailthroughouttheproject,trackingall
datacollectionmethods,codingprocesses,codingdecisions,analysis
drafts,analysisrevisions,andmembercheckfeedback.
108
SeeStake(1995,2010,2013)andYin(2008,2011).
410
STANDARD:CONFIRMABILITY
Definition
Theresearcherneedstoexaminehowhis/herownidentity,lifeexperiences,social
location,biases,andperspectivesmayhaveshapedtheprocessoftheresearchand
theresultingfindings.Theresearcherneedstokeepreflexivefieldnotes
throughouttheprocesstoexaminetheseissuesandbracketthemfromthework
(and/orempiricallyexaminethemwithinthework,seeCampbell,2002asan
example).
ThePIkeptfieldnotesthroughouttheproject,whichincludedreflexive
memoing.Inpeerdebriefingwithothermembersoftheresearchteam,
weregularlyquestioned/challengedeachothersperspectives.Given
thatthisprojecthadanexplicitevaluationaim(seeChapter1:
Introduction),weemphasizedbracketingreflexivityratherthan
incorporatingitintotheresearchprocess.
Otherstrategiesforassuringconfirmabilityinclude:keepinganaudittrail(see
above)andconductingtriangulationassessments(seeabove).
Wealsokeptanaudittrail(seeabove)andconductedextensive
triangulationassessments(seeabove)toensurethatthefindingsdid
notreflectthebiasesoftheresearchteam.
411
Quantitative Data Analyses Modeling SAK Submission Rates Over Time
TheresultsoftheSAKcensus(Chapter2:HowManyUnsubmittedSAKsinDetroit)revealedthat
thenumberofunsubmittedSAKsfluctuatedovertime:someyearsthereweresubstantiallymore
unsubmittedSAKsthanSAKssubmittedfortesting,andinotheryears,theproportionswerenearly
equal(thoughtherewerenoyearsinwhichmoreSAKsweresubmittedthannot).Toexplorewhy
DetroithadsomanyunsubmittedSAKs(Chapter3),weidentifiedkeyhistoricalchangesinpolicy,
practices,andresources.Therefore,wewantedtoexaminewhetherthesekeyturningpoints/changes
identifiedthroughthequalitativedatawereassociatedwithSAKsubmissionratesovertime.Using
multilevellongitudinalquantitativemodeling,wetestedwhetherkeyhistoricaleventsidentifiedinthis
analysiswereassociatedwiththeobservedratesofSAKsubmissionsovertime.
Data Analytic Methods
Analysisfocusedonsystemlevelchangesthoughttoaffectthepercentofsexualassaultkits
thatweresubmittedtothecrimelab.Datawerecollectedforeachof10,817kitscollectedintheyears
from1980to2009.Systemchangesoccurredatvariouspointsduringthisperiod.Thedispositionof
eachkitwasrecorded(whetheritwassubmittedtothecrimelabforanalysisornot),alongwiththe
yearinwhichthekitwascollected.Severalpotentiallyinfluentialsystemicchangeswereidentified
acrossthespanofyears,including1)changesintheprocessbywhichDNAevidencewashandledbythe
crimelab,2)receiptofadditionalDOJfundsforDNAtesting,3)apolicedepartmentpolicychange
expandingthetypesofevidencethatshouldbesubmittedforforensictesting,4)establishmentofthe
SANEprogram,and5)staffcutstothepolicesexcrimesunit.Descriptivestatisticsforeachofthese
variablesareinTableB.2(nextpage).
412
TABLEB.2ModelingSAKSubmissionRatesOverTime,DescriptiveStatistics(N=10,817)
#kits
%ofkits
%oftotal submitted
DEPENDENTVARIABLE
Submissionofsexualassaultkittocrimelab
Submitted
2426
22.4
100
NotSubmitted
8391
77.6
PreDNAEra(19801993)
2261
20.9
13.6
DNAbutnoCODIS(19941997)
2751
25.4
12.9
DNAandCODISbutpolicecrimelabcannotupload(19982001)
2026
18.7
32.9
DNAandCODISandpolicecrimelabhasprovisionalupload(20022005)
2070
19.1
23.6
DNAandCODISandpolicecrimelabhasfullupload(20062009)
1709
15.8
35.5
BeforeDOJfunding
8464
78.2
20.0
AfterDOJfunding
2353
21.8
31.0
Beforepolicepolicychange
6089
56.3
16.6
Afterpolicepolicychange
4728
43.7
29.9
403
3.7
46.9
10414
96.3
21.5
925
8.6
29.2
9892
91.4
28.1
INDEPENDENTVARIABLES
"Eras"ofhandlingDNAevidencebycrimelab
DOJfundingforcrimelabDNAtesting(2005)
Policepolicychangeresubmittingevidencetocrimelab(2002)
SANEprogramestablished(2006)
YearSANEprogramestablished
NotyearSANEprogramestablished
Yearsofstaffcutsinpolicesexcrimesunit(2002and2008)
Staffcutyear
Notstaffcutyear
413
Theunitofanalysisisthekit,whichisnestedwithinyears.Onlythedependentvariable
dispositionismeasuredatthekitlevel;allcovariatesareatthelevelofyears.Toaccommodate
dependenciesamongkitscollectedinthesameyear,theanalysisusedmultilevelormixedeffects
regression,withindividualkitsatlevel1,nestedwithinyearsatlevel2.Modelswereestimatedwith
randominterceptsforkitsubmissionandfixedslopeeffectsforlevel2covariates.Becausethe
dependentvariablesubmissionisadichotomousvariable,alogitlinkfunctionwasused,producing
multilevellogisticregressionmodels.AnalyseswereconductedusingHLM7software(Raudenbush,
Bryk,Cheong,&Congdon,2011).
Abaselinenullmodelwasestimatedfirsttoverifytheadequacyofthevarianceoftherandom
interceptandtoobtainabaselineloglikelihoodvalueagainstwhichthefitofmorecomplexnested
modelscouldbecompared,asothercovariateswereadded.Thenafixedlineareffectfortime(inyears,
centeredatthemeanyearof1999)wasestimatedinarandomintercept/fixedslopemodel.Tocheck
forcurvilinearity,modelsincorporatingpolynomialfunctionsoftime(squaredandcubed)werealso
estimated,butneithertermmadeasignificantimprovementtomodelfit,indicatingthattheeffectof
timewasessentiallylinearinform.ThenullandfixedlineartimeslopemodelsaresummarizedinTable
B.3(nextpage).Therandomeffectsvarianceinthenullmodelwaslargeandsignificant,indicating
sufficientvariabilityamongyearlysubmissionratestowarrantfurtheranalysis.Theadditionofthe
linearslopetermmadeasignificantimprovementtomodelfit(likelihoodratio(LR)chisquare=15.86;p
<.0001);thefixedeffectfortimewaspositiveandsignificant(oddsratio(OR))=1.09),indicatingthaton
average,eachsubsequentyearwasassociatedwitha1.09increaseinsubmissionrate.Thesebaseline
modelsformedthebasisforsubstantivemodelsexaminingthecontributionoflevel2covariates.
Becauseallcovariatesaretimerelatedandthereforepotentiallycorrelatedwiththelineareffectof
time,eachcovariatewasinitiallyexaminedbothwith(conditional)andwithout(unconditional)thefixed
effectoftime.Subsequentmultivariablemodelswerederivedfromtheinitialbivariatemodels.
414
TABLEB.3BaselineModelsforSubmissionofSAKstoCrimeLab
FixedEffects
Model
Model
components
Log
0dds
Null
Model
Intercept
(random)
1.390 0.139 0.249
Fixed
Linear
Time
Slope
Model
Intercept
(random)
1.299 0.097 0.273
Time(years
linear)
0.084 0.015 1.087
se
RandomEffectVariance
Odds Confidence
ratio
interval
0.187
0.331
SD
df
chi
square
ModelFit
Log
#
LRchi
likelihood parameters square
<
.001 0.655 27 757.156 <.001 15351.1
0.0224
0.333 <.000 0.435 26 3999.27 <.001
1.054
1.122 <.001
15343.20
15.86 <.0001
415
Detailed Statistical Results
Firstweexaminetheeffectofindividualcovariates.Theeffectsofchangesintheprocessby
whichDNAevidencewashandledbythecrimelabaresummarizedinTableB.4(followingpages).The
toppanelexaminestheunconditionaleffects(i.e.,withoutadjustingforlineartimeeffects),andthe
bottompanelexaminestheconditionaleffects,adjustingforlineartime.Intheunconditionalmodel,
DNAeramadeasignificantcontributiontomodelfit(LRchisquare=12.54,p=0.14).Eachoftheeras
differedsignificantlyfromthefourthera,thecategorydesignatedasthereferencebecauseitprovided
theclearestillustrationoftheoveralleffect:thefirst2eraspreDNAandDNAwithoutCODISdidnot
differsignificantlyfromeachbutwerebothsignificantlylowerthanthereferencecategoryDNAand
CODIS,withthepolicecrimelabhavingprovisionaluploadrightstoCODIS.Ratesofkitsubmissioninthe
firsttwoeraswereonlyhalfthesizeofratesinthereferencecategory(OR=0.46and0.47,
respectively).Eras3DNAandCODISbutcrimelabcannotuploadand5DNAandCODIS,withthe
crimelabhavingfulluploaddidnotdiffersignificantlyfromeachotherbutwerebothsignificantly
higherthanthereferencecategory(OR=1.55and1.80,respectively).Theconditionalmodelinbottom
panelofTableB.4addsthelineareffectoftimetotheDNAeramodel.Ratherthanimprovingfit,the
additionoftimeworsenedmodelfit,relativetothefitoftheunconditionalmodel(LRchisquare=2.4,
p=1).Thisfinding,alongwiththereversedsignoftheeffectoflineartimeandtheinflationofstandard
errorsintheconditionalmodel,indicatesstrongcollinearitybetweenDNAeraandlineartime.
TableB.5(followingpages)containstheunconditionalandconditionalmodelsforthereceiptof
DOJfundingforDNAtesting.Inthetoppanel,theunconditionalmodelindicatesthatDOJfundinghasa
significant(p=.01)andpositive(OR=2.24)effectonkitsubmissionandthatitsignificantlyimproves
modelfitoverthenullmodel(LRchisquare=5.42,p=.02).YearsafterreceiptofDOJfundinghadrates
ofkitsubmissionthatweremorethandoubletheratesofyearspriortothefunding.Theadditionof
lineartimeintheconditionalmodel(bottompanel)significantlyimprovedmodelfit(LRchisquare=
416
10.32,p=.001).However,afteradjustingforlineartime,theeffectofDOJfundingreversedsignand
wasmuchsmallerandnolongersignificant.
TableB.6(followingpages)containsthemodelsfortheeffectsofachangeinpolicedepartment
policythatexpandedthetypesofevidencethatshouldbesubmittedtothecrimelab.Inthe
unconditionalmodel,theeffectofthispolicychangehadasignificant(p<.001)andpositive(OR=2.59)
effectonkitsubmissionanditsignificantlyimprovedmodelfitoverthenullmodel(LRchisquare=6.64,
p=.01).Theadditionoflineartime(bottompanel)madeasignificantimprovementtomodelfit(LRchi
square=8.68,p=.003).However,inthisconditionalmodel,thechangeinpolicedepartmentpolicywas
nolongersignificant.
TableB.7(followingpage)containsthemodelsforeffectsofstaffcutsinthepolicesexcrimes
unit,whichoccurredinboth2002and2008.Inboththeunconditionalandconditionalmodels,staffcuts
hadnosignificanteffectonkitsubmission.Resultswerevirtuallyidenticalforseparateanalyses
examiningtheeffectofeachseparateyearinwhichtherewerestaffcuts(2002and2004).Resultsof
theseseparateanalysesarenotshown.
TableB.8(followingpages)containsthemodelsforeffectsofestablishmentoftheSANE
program.Inthetoppanel,theunconditionalmodelshowsthatSANEhadasignificant(p=.043)and
positive(OR=3.71)effectonkitsubmissionandthatitsignificantlyimprovedmodelfitoverthenull
model(LRchisquare=5.40,p=.02).TheyearinwhichSANEwasestablishedhadarateofkit
submissionthatwasnearly4timestheaverageacrossallyears.Inthebottompanel,theadditionof
lineartimesignificantlyimprovedmodelfit(LRchisquare=14.22.p<.001).TheeffectofSANE
remainedpositive(OR=1.94),althoughitwasnolongersignificant,accordingtotheWaldtest(p=
.149).
Theseinitialbivariatemodelswereusedtodevelopandevaluatesubsequentmultivariable
models.BecauseDNAeras(changesintheprocessbywhichDNAevidencewashandledbythepolice
417
crimelab)appearedtofullyexplainthelinearincreaseinkitsubmissionovertime(i.e.,addingthelinear
timeeffectworsenedmodelfit,duetoitscollinearitywithDNAera;seeTableB.4),DNAerawasusedas
thebasisofafinalmodelofsystemschangesrelatedtokitsubmission.Thesemodelshadtherandom
interceptinblock1andtheDNAerasvariableinblock2.Block3ofeachsubsequentmodeladded,
oneatatime,avariablethathadbeenfoundinbivariatemodelstohavesignificantrelationshipswith
kitsubmission,unconditionalonlineartime.DOJfunding(TableB.5),policepolicychange(TableB.6),
andestablishmentofSANE(TableB.8)metthiscriterion.Thesemultivariablemodelsarepresentedin
TablesB.9throughB.11(followingpages).TableB.9showstheadditionofDOJfundingforDNAtesting
inblock3andthecategoriescomprisingtheDNAeravariableinblock2.DOJfundingwasnotsignificant
(p=.229)anddidnotimprovemodelfit.TableB.10addspolicepolicychangeinsubmissionofevidence
tothecrimelabinblock3.Changeinpolicepolicywasnotsignificant(p=.783)anddidnotimprove
modelfit.TableB.11addstheyeartheSANEprogramwasestablishedinblock3.Althoughthe
coefficientwaslargeandpositive,itdidnotreachtheconventionallevelofsignificance(OR=1.81,
p=.053).However,theadditionofSANEestablishmentsignificantlyimprovedmodelfit(LRchisquare=
4.24,p=.039).Thiswasconsideredthefinalmodel.
Thus,themodelcontainingtwocovariatesDNAeras,orchangesintheprocessbywhich
DNAevidencewashandledbythecrimelab,alongwiththeyearinwhichSANEwasestablished,was
determinedtobethebestfittingmodelexplainingshiftsinratesofkitsubmission.Althoughtwoother
potentialcovariatesDOJfundingforDNAtestingandthepolicedepartmentpolicychangein
submissionofevidencetothecrimelabhadsignificantunconditionalrelationshipswithkitsubmission
rates,theywerenolongersignificantwheneitherlineartimeorDNAeraswereaccountedfor.Yearsin
whichtherewerestaffcutsinthesexcrimesunitdidnotshowanyrelationshiptochangeinkit
submission,eitherunconditionalorafteraccountingforDNAeras.
418
TABLEB.4ModelsExaminingtheEffectsofDNAEra
419
TABLEB.5ModelsExaminingtheEffectsofDOJBacklogReductionGrantFunding
Table4
ModelsexaminingtheeffectsofreceiptofDOJf
Fixedeffects
Model
Unconditional
model
(Withouttime)
Log
Modelcomponents odds
Intercept
DOJfundingfor
DNAtesting(2005
2009)
Conditionalmodel Intercept
(Addinglinear
time)
DOJfundingfor
DNAtesting(2005
2009)
Time(Years
linear)
se
Odds
ratio
Confidence
interval
Randomeffectsvariance
1.550
0.139
0.212 0.1590.282
<.001
0.807
0.293
2.240 1.2264.093 0.011
1.222
0.126
0.295 0.2280.382
0.313
0.329
0.731 0.3721.439 0.350
0.099
0.022
1.104 1.0551.155
<.001
<.001
SD
0.568
0.431
df
26
chi
square
634.9 <.001
Modelfit
Log
#
LRchi
15351.13
15348.42
5.42 0.020
15343.26
10.32 0.001
25 381.240 <.001
420
TABLEB.6ModelsExaminingtheEffectsofthePolicePolicyChangeRegardingSAKSubmissions
Table5
ModelsexaminingtheeffectofchangeintheDetroitPoliceDepartmentpolicyexpandingtypeso
Model
Modelcomponents
Log
odds
se
Unconditional
model
Intercept
1.778 0.138
(Withouttime)
DPDpolicychangere
submissionofkits
0.953 0.204
Fixedeffects
Odds
Confidence
ratio
interval
Modelfit
Log
LRchi
Likelihood #Parameters square
0.169 0.1270.224 <.001 0.463 26 477.08 <.001
15351.13
2.590 1.7063.045 <.001
15347.81
6.64
15343.47
8.68 0.003
Conditional
model
Intercept
1.335 0.206
(Addinglinear
time)
DPDpolicychangere
submissionofkits
0.099 0.373
1.105 0.5122.382 0.792
Time(Yearslinear)
1.080 1.0161.147 <.015
0.077 0.029
chi
SD df square
p
0.01
0.260 0.1700.397 <.001 0.432 25 397.99 <.001
421
TABLEB.7ModelsExaminingtheEffectsofthePoliceSexCrimeUnitStaffingCuts
Table6
Modelsexaminingtheeffectofstaffcutsinthesexcrimesunit
Fixedeffects
Log
Odds Confidence
Model
FixedEffects
se
Odds
Ratio
Interval
Unconditional
model
(Withouttime)
Conditional
model
(Addinglinear
time)
chi
SD df square
p
Modelfit
Log
LRchi
Likelihood #Parameters square
Intercept
1435 0.143
0.238
0.1780.319 <.001 0.639 26 740.278 <.001
15351.1
Yearofstaffcuts
(2002&2008)
0.549 0.480
1.732
0.6464.641 0.263
15351.1
Intercept
1.288 1.102
0.276
0.2230.341 <.001 0.434 25 396.408 <.001
Yearofstaffcuts
(2002&2008)
0.111 0.350
0.895
0.4351.839 0.753
0.085 0.016
1.089
1.0541.126 <.001
15343
Time(Yearslinear)
0.01
16.22 <.001
422
TABLEB.8ModelsExaminingtheEffectsoftheEstablishmentoftheSANEProgram
Table7
ModelsexaminingtheeffectofSANE
Model
Modelcomponents
Unconditional
model
Intercept
(Withouttime)
YearSANEwas
established(2006)
Log
odds
se
1.434 0.130
1.310 0.614
Conditional
model
Intercept
(Addinglinear
time)
YearSANEwas
established(2006)
Time(Yearslinear)
Fixedeffects
Odds
Confidence
ratio
interval
0.238
chi
SD df square
p
0.1830.311 <.001 0.592 26
3.706 1.04813.099
624.42 <.001
0.04
1.329 0.094
0.265
0.2180.321 <.001
0.665 0.446
1.944
0.7754.877
0.077 0.015
1.080
1.0471.114 <.001
Modelfit
Log
likelihood
LRchi
#parameters square
15351.13
15348.43
5.40 0.020
15341.32
14.22 <.001
0.41 25 357.583 <.001
0.15
423
TABLEB.9ModelsExaminingtheEffectsofDNAErasandDOJBacklogReductionGrantFunding
Table8
ModelexaminingtheeffectsofDNA"eras"andDOJ
Block#
Logodds
Intercept
"DNAEras"inwhichsexualassaultkitwas
se
1.059 0.156
Fixedeffects
Odds
Confidence
ratio
interval
0.347
0.2510.579
Modelfit
p
<.001
PreDNAEra(19801993)
0.871 0.200
0.418
0.2760.633
<.001
0.849 0.208
0.428
0.2780.658
<.001
DNAandCODISbutpolicecrimelab
cannotupload(19982001)
0.343 0.305
1.409
0.9212.155
0.109
DNAandCODISandpolicecrimelab
hasprovisionalupload(20022005)
ReferenceCategory
hasfullupload(20062009)
0.490 0.206
1.633
1.0652.503
0.026
DOJfundingforDNAtesting(20052009)
0.384 0.310
0.681
0.3581.296
0.229
SD
0.249
df
Chisquare
137.204
p
<.001
22
Log
likelihood
#Parameters
LRchi
square
15351.13
15344.86
12.54
0.014
15344.97
0.11
424
TABLEB.10ModelsExaminingtheEffectsofDNAErasandPolicePolicyChangeRegardingSAKSubmissions
Table9
ModelexaminingtheeffectsofDNA"eras"andchangeintheDetroitPoliceDepartmentpolicyexpandingtypesofevidencetobesubm
Block#
LogOdds
Intercept
1.234
"DNAEras"inwhichsexualassaultkitwas
0.31
0.291
0.1530.554
p
<.001
PreDNAEra(19801993)
0.703
0.336
0.495
0.2470.994
0.048
0.675
0.341
0.509
.2511.033
0.06
0.478
0.241
1.614
.9792.660
0.06
se
Fixedeffects
Odds
Confidence
Ratio
Interval
Modelfit
#
LRchi
Parameters square
15351.13
15344.86
15344.83
12.54 0.014
ReferenceCategory
0.588
0.198
1.8
1.1942.712
0.007
DPDpolicychangeresubmissionofkits
0.077
0.277
1.08
0.6081.918
0.783
Chisquare
148.119
p
<.001
Log
Likelihood
SD
0.26
df
22
0.03
425
TABLEB.11ModelExaminingtheEffectsofDNAErasandtheEstablishmentoftheSANEProgram
Table10
FINALMODELModelexaminingtheeffectsofDNA"eras"andestablishmentofSANE
FixedEffects
Confidence
Odds
interval
Block# Independentvariables
se
Logodds
ratio
1
Intercept
1.157
"DNAEras"inwhichsexualassaultkit
wascollected
0.314
0.2420.409
<.001
PreDNAEra(19801993)
0.764
0.174
0.466
0.3250.667
<.001
0.751
0.181
0.472
0.3240.686
<.001
0.440
0.178
1.553
1.0742.246
0.021
0.127
Modelfit
Log
#
LRchi
15351.1
15344.9
12.54
0.014
15342.7
4.24
0.039
ReferenceCategory
0.437
0.193
1.548
1.0362.412
0.034
YearSANEwasestablished(2006)
0.596
0.291
1.814
0.9923.316
0.053
SD
0.231
df
Chisquare
119.564
p
<.001
22
426
Quantitative Data Analyses Modeling CODIS Hit Rates

IndevelopingtheDetroitSAKtestingplan,ourgoalwastocraftadesignthatcouldinform
policyandpracticedebatesabouttestall/testsomeSAKs.InthiscomponentoftheARP,wefocused
onwhetherforensictestingoutcomes,specificallythenumberofCODIShits,differbyspecificcase
characteristics(e.g.,victimoffenderrelationship,statuteoflimitations,victimdemographics,assailant
demographics,assaultcharacteristics).Wealsocomparedforensictestingoutcomesfortwodifferent
DNAtestingmethods:traditionalandselectivedegradation(DNase).
Data Analytic Methods

SamplesandSamplingProcedures.ThisquantitativeportionofthisDetroitSAKARPwas
designedtoexaminemultipleresearchquestionsregardingSAKforensictestingoutcomes,someof
whichpertainedonlytospecificsubpopulationsoftheunsubmittedDetroitSAKs.Asaresult,data
collectionwasdividedintofourseparateprocesses,eachofwhichproducedasampleofSAKs(Testing
Group)thatcouldbeanalyzedindependentlyorcombinedwithothertestinggroupsintoalarger
sampleforadditionalanalyses.Figure4.17(TheDetroitSAKTestingGroupsandPopulations
Represented,Chapter4)illustratedtherelationshipoftheTestingGroupstovarioussubpopulations;
TableB.12(nextpage)furtherdescribesthesizeofeachsampleweanalyzed,alongwiththepopulation
orsubpopulationitrepresentsandtheanalyticalpurposesitserved.
WealsousedanadditionalsamplecalledTestingGroup5inthisAppendixcomprisedofN
=250ofthe400SAKsoriginallysubmittedforforensicanalysisonThe400Project,(Pierce&Zhang,
2011a,2011b)supplementedbyadditionalvictimoffenderrelationshipandSOLstatusinformation
collectedaboutthoseSAKsforanotherstudy(Shaw,2014).Thiscrucialresourceallowedustocompute
samplingweightsthatimprovedourabilitytoaccountforthedisproportionatesamplingofcertainkinds
ofSAKsinTestingGroups14(detailsaredescribedbelow).
427
TABLEB.12SamplesUsedintheQuantitativeAnalysesExaminingForensicTestingOutcomes
Group
PopulationorSubpopulation
Purposes
SamplingDesign Weighting
445 Untested,nonadjudicated,SOL PredictCHR

unexpired,DetroitSAKsfrom
strangerassaults
Stratifiedbyyear None
449 Untested,nonadjudicated,SOL PredictCHR

unexpired,DetroitSAKsfrom
nonstrangerassaults
Stratifiedbyyear None
351 Untested,SOLexpiredDetroit
SAKs
Simplerandom
sample
None
350 Untested,SOLunexpiredDetroit Quantify&compareDTR& Simplerandom

sample
SAKs
CERtoexaminetesting
methodeffect
None
250 UntestedDetroitSAKs
None
PredictCHR
Providesamplingweights Simplerandom
sample
1+2
894 Untested,nonadjudicated,SOL Quantify&compareCER,

CHR,&SARtoexamine
unexpiredDetroitSAKsfrom
VOReffect
assaultswhereVORisknown
Stratifiedbyyear StratifiedbyVOR
&VOR
3+4
701 UntestedDetroitSAKs
StratifiedbySOL StratifiedbySOL
status
status
Quantify&compareCER,
CHR,&SARtoexamine
SOLeffect
Note:Group5wasnotnewdatacollectedunderthisstudy.Itwasasubsetof250SAKsfromThe400Projectthathavebeen
usedinanotherstudy(Shaw,2014);secondaryanalysesofthesedataprovidedthebestavailablesourceforcomputing
samplingweightsrequiredtoproperlyanalyzedatafromtestinggroups14.CER=CODISentryrates;CHR=CODIShitrates;
DTR=DNAtestingrates;SAR=serialsexualassaultrates;SOL=statuteoflimitations;VOR=victimoffenderrelationship.
Althoughitwouldhavebeenidealtocollectdataonadjudicationstatus,SOLstatus,andvictim
offenderrelationshipfortheentirecollectionofDetroitSAKsduringthecensus,thiswasnotpractical.
Preservingthechainofcustodythroughoutthecensuswasofparamountimportance.Stafffromthe
prosecutingattorneysofficehadtoconductthecensusinthepresenceofadesignatedpropertyofficer
andtheentirecensushadtobecompletedinlessthanfourmonths.Asaresult,planstocollectdata
thatwouldhavebeenextremelyusefulfordrawingsubsequentsamplesweresetasideandthecensus
428
wasfocusedstrictlyonobtaininganaccuratecountofSAKs.SamplingwithoutreplacementforTesting
Groups14wasdoneinthreestepsasshownbelow,followedbyanadditionalsteptoobtainthesample
weultimatelyusedtogeneratesamplingweights.ThePIandCoIconsultedwithmultiple
methodologists/statisticiansduringthecreationandimplementationofthissamplingplan,aswe
encounteredpracticalconstraints(e.g.,samplesize,distributionofSAKsacrossyears)thatnecessitated
compromisesinouroriginalplans(asnotedbelowandinChapter4).
TestingGroups1and2:AlistofIDsforSAKscollectedbetween20022009wasextractedfrom
thecensuslisttoensurethatallSAKsconsideredwouldbeSOLunexpired.Thislistwasplacedina
randomorderandthenprovidedtotheprosecutorsresponsibleforsamplingtheSAKs.Theywere
instructedtoreviewtheSAKsintheorderlisted,makingadecisionabouteachSAKbeforemovingonto
thenextoneonthelist.PreviouslyadjudicatedandpreviouslytestedSAKswereremovedfromthelist
andexcludedfrombothgroups.Untested,nonadjudicatedSAKswerethensortedbyvictimoffender
relationship:strangerassaultswentintoTestingGroup1,nonstrangerassaultswentintoTestingGroup
2.Thisprocesscontinuedaimingtoaccumulateastratifiedsampleofabout56SAKsperyear(two
randomlyselectedyearshadtargetsof58)ineachgroup.Thistargetsamplesizeof450wasbasedon
budgetconstraints.Thefinalnumberofkitsperyearineachgroupdidnotexactlymatchthetargets,
largelyduetohowmanyeligiblecasestheyhadperyear.Theprosecutorsadjustedthetargetstodeal
withshortfallsinoneyearbyincreasingthetargetnumberofSAKsforfollowingyears(distributingthe
extraSAKsequallyacrosstheremainingyears).Thiswasnotideal,butwewereunabletoweightthe
databyyearanyway.Whatseemedmostimportantisthatthesamplingschemeguaranteedthatdata
fromallyearsintherange2002to2009wereincludedinbothTestingGroups1and2.
TestingGroup3:Thisgroupwasselectednext.AlistofIDsforSAKscollectedbetween1980
and2001wasextractedfromthecensuslisttoensurethatallSAKsconsideredwouldbeSOLexpired.
Thelistorderwasrandomizedandthenitwasprovidedtotheprosecutorsresponsibleforsamplingthe
429
SAKs.TheywereinstructedtoreviewtheSAKsintheorderlisted,makingadecisionabouteachSAK
beforemovingontothenextoneonthelist.PreviouslytestedSAKswereremovedfromthelistand
excludedfromtheTestingGroup3sample,whileuntestedSAKswereaddedtoTestingGroup3
(regardlessofadjudicationstatus)untilatotalof350SAKshadbeensampled.Therewasno
stratificationbyyearbecausethedistributionwasskewed.Ideally,previouslyadjudicatedSAKswould
havebeenexcludedaswell,butthetimeandresourcesnecessaryforsuchscreeningwerewelloutside
thetimelineofthiscomponentoftheproject.Thisexpandedthedefinitionofthepopulation
representedbythesampletoincludebothadjudicatedandnonadjudicatedSAKs.
TestingGroup4:Thisgroupwasselectedlast.WewentbacktothecensuslistforSAKsfrom
20022009,thenexcludedIDnumbersthathadbeenscreenedoutbasedonadjudicationstatus(during
theTestingGroup1and2screening)oraddedtoeitherofthoseTestingGroups.Theorderofthe
remainingIDnumberswasrerandomized,thenlistofIDnumberswasgiventotheprosecutors.They
wereinstructedtoreviewtheSAKsintheorderlisted,makingadecisionabouteachSAKbeforemoving
ontothenextoneonthelist.PreviouslytestedSAKswereremovedfromthelistandexcludedfromthe
TestingGroup4sample,whileuntestedSAKswereaddedtoTestingGroup4(regardlessofadjudication
status)untilatotalof350SAKshadbeensampled.Therewasnostratificationbyyearbecausethe
distributionwasskewed.Ideally,previouslyadjudicatedSAKswouldhavebeenexcludedaswell,but
thetimeandresourcesnecessaryforsuchscreeningwereprohibitive.Thisexpandedthedefinitionof
thepopulationrepresentedbythesampletoincludebothadjudicatedandnonadjudicatedSAKs.After
caseswereselectedintoTestingGroup4,halfofthemwererandomlyassignedtoeachofthetwo
testingmethods.
TestingGroup5:Giventheveryspecificnatureofthephenomenonwewerestudying,we
determinedthattherewerenopreviouslypublishedstudiesfromdemographicallycomparablecities
thatalreadyprovidedadequate,relevantinformationaboutthefrequencyofspecifictypesofuntested
430
SAKs.Therefore,weneededlocaldatafromthepopulationofDetroitSAKs.Theonlyavailablesource
thatwefeltcouldprovidecrediblesamplingweightswasasubsetof250SAKsfromtheThe400Project
(Pierce&Zhang,2011a,2011b).For250ofthe400cases,The400Projectteamwasabletotrackdown
accompanyingpolicereportsandmedicalrecordstobeabletodeterminevictimoffenderrelationship
ortheexactdateeverySAKwascollected(fromwhichSOLstatuswasderived)(seeShaw,2014).This
subsampleof250SAKs(fromtheoriginal400)becamewhatwecallTestingGroup5here.The
samplewaslargeenoughtoprovideweightsstratifiedbySOLstatusandvictimoffenderrelationship,
butitwastoosmalltoprovidestableestimatesthatwerefurtherstratifiedbyyear.
MeasuresandCovariates.Below,wedescribethevariablesusedintheanalysesreportedin
Chapter4;thevariablesarelistedhereintheordertheyarementionedinthatchapter.
Stageofforensictestingreached:Thiswasanordinal,categoricalvariablerecordingthelast
stageofforensictestingreachedbyeachSAK.Thestages(showninFigure4.2)were0=serology
screening,1=DNAtesting,2=CODISentry,3=CODIShit,and4=serialsexualassault.Thestatistical
modelsinChapter4usedvariousrecodedversionsofthisvariableasthekeyoutcomes.The
continuationratiomodelsrequiredasetofstagespecificbinaryoutcomevariablesshowingwhethera
SAKprogressedpasteachstagetothenextone(0=no,1=yes);analyzingthoseindicatorsprovided
estimatesoftheDNAtesting,CODISentry,CODIShit,andserialassaultrates(theratebeingestimated
dependsonwhichstageyouexamine).AnySAKthatreachedoneofthelaterstageshadbydefinition
progressedpasteverypreviousstage.Thelogisticregressionmodelsinsteadusedasimplebinary
indicatorofwhetherornotanSAKhadgeneratedaCODIShit,whichoccurredforallSAKsreaching
Stages34.
Victimoffenderrelationship(VOR):Therawdataforthisvariableconsistedofseveralcategories
(1=stranger,2=knownbysight/nickname,3=friend/associate/familymember,4=current/past
intimatepartner,777=suspectknowntovictimbutnatureoftherelationshipnotclearbasedonthe
431
informationinthefile,or999=missingdata).Werecodeditintoabinaryindicator(0=nonstranger
[combiningrawcodes24,777],1=stranger)codingwhetherornottheassailantwasastrangertothe
victim.SAKswithmissingVORdatawereexcludedfromthesampleusedfortestingtheVOReffect.
Statuteoflimitations(SOL)status:Thiswasabinaryindicator(0=SOLexpired,1=SOL
unexpired)codingwhetherornotthesexualassaultassociatedwithaSAKoccurredbetween2002and
2009,whichwoulddefinitelyindicatethattheSOLforthatcrimehadnotyetexpired.SAKscollected
between1980and2001werepresumedtohaveexpired.
Victimage:Althoughwehadactualageinyearsasacontinuousvariable,wedividedvictimage
intotwocategoriesbasedontheageofconsentinMichiganbecausethisdistinctionistypicallymore
salienttothelegalprocessingofthecasethanexactvictimage.Theoldergroupwasthereferencelevel
(0=16+years,1=015years).
Assailantage:TheSAKsandassociatedcasedocumentationoftencontainedonlythe
approximateagesfortheassailants,asestimatedbythevictim.Thereweretoofewassailantslessthan
18yearsoldtodivideassailantageaccordingtowhethertheywouldbeconsideredminorsforlegal
purposes(017yearsvs.18+years),soweinsteaddividedassailantsintothreegroups(0=22+years
[adults],1=021years,2=unknown).Theoldergroupservedasthereferencelevelbecausemost
assailantswereadults.Unknownagewastreatedasavalidcategorybecausethenatureand
circumstancessurroundingsomesexualassaultsmakeitimpossibletoobtainassailantageinformation
priortoidentifyingasuspect.ThismayitselfbepredictiveofwhetheraCODIShitwillresultfrom
forensictesting.
Multipleperpetrators:Thiswasabinaryindicator(0=no,1=yes)codingwhetherornotthe
SAKandassociatedcaseinformationexplicitlydocumentedthatthereweremultipleperpetrators
involvedinthesexualassault.Incompleterecordsthatmightotherwisebeconsideredmissingdata
werecodedas0.
432
Useofalcoholordrugs:Thiswasabinaryindicator(0=no,1=yes)codingwhetherthepolice
reportexplicitlydocumentedthattheassaultoccurredinthecontextofalcoholordrugs.Notethatthis
doesnotdistinguishwhetheritwasthevictim,theassailant,orbothwhowereusingalcoholordrugsat
thetimeoftheassault;itonlydocumentsthatatleastoneofthemwasdoingso.Incompleterecords
thatmightotherwisebeconsideredmissingdatawerecodedas0.
Useofweapons:Thiswasabinaryindicator(0=no,1=yes)codingwhetherthepolicereport
explicitlydocumentedthataweaponwasusedintheassault.Incompleterecordsthatmightotherwise
beconsideredmissingdatawerecodedas0.
Useofphysicalforce:Thiswasabinaryindicator(0=no,1=yes)codingwhetherthepolice
reportexplicitlydocumentedthattheassailantusedphysicalforceinthesexualassault.Incomplete
recordsthatmightotherwisebeconsideredmissingdatawerecodedas0.
Examtiming:Examtimingreferstothelengthoftimeelapsedbetweenthesexualassaultand
themedicalforensicexaminationofthevictimthatgeneratedtheSAK.Examtimingwascalculatedby
comparingdates,thendividedintothreecategories(0=0days[sameday,referencecategory],1=1
dayafterassault,or2=2+daysafterassault).
DNAtestingmethod:EachSAKinTestingGroup4wasrandomlyassignedtooneoftwogroups
(0=traditionalDNAtesting,1=DNaseselectivedegradationtesting).Thisbinaryindicatorrecorded
thosegroupassignments.AllSAKsinTestingGroups13receivedtraditionaltesting.
Presenceofsperm:Thisbinaryindicator(0=no,1=yes)recordedwhetherlaboratory
personneldetectedspermintheSAKwhenscreeningitforbiologicalevidence.Itwasrecordedonlyfor
TestingGroup4.
Costofconsumablesupplies:ThecostofconsumablesuppliesusedinscreeningSAKsfor
biologicalevidenceandconductingDNAtestswasrecordedinUSdollarsbythelaboratorypersonnel.
433
Thesedatawereprovidedonlyinaggregateformandincludethecostsofwaste,controls,and
reprocessing.
Personneleffort:Theamountoftimespentonforensictestingwasrecordedinhours.Separate
estimateswereprovidedfor(1)laboratorypersonneleffortspentontestingtheSAKs(2)laboratory
personneleffortspentonreviewingtestresults,and(3)statepoliceforensicsciencedivisionstaffeffort
spentonreviewingtestresults.Someofthesedatawererecordedonlyattheaggregatelevelfor
batchesofSAKs,ratherthanrecordedseparatelyforeverySAK.
PreparingtheDataforAnalysis.Manycomplexstatisticalissuesneededtoberesolvedpriorto
theanalysisofthesedata,including:howtoappropriatelycombineandweightdatafrommultiple
groups,howtohandlenonindependentobservations,andhowtoaddresstheproblemofmissingdata.
Below,wedescribeourapproachtoresolvingeachofthesechallenges.
Samplingweights:Becausethesizesofthefocalsubpopulationsandthesamplingprocedures
usedvariedacrossthefourtestinggroups,combiningdatafrommultiplegroupsrequirestreatingthem
aspartofacomplexsamplingdesign.Herewedescribeourapproachtodevelopingthesampling
weightsthatpermittedustodrawmoreappropriateandgeneralizableinferencesfromouranalyses.
Obtaininganappropriatesourcefromwhichtocalculatesamplingweightsrelevanttoourfocal
populationandsubpopulationswasachallengingprospect.ThedemographiccompositionofDetroitis
unlikethatofotherlargecitiesthathaveaccumulatedcollectionsofuntestedSAKs(HumanRights
Watch2010,2011).Furthermore,thereisnopriorresearchthatreportshowmanyuntestedSAKsthere
areineachcellofa2x2tablebrokendownbywhetherthestatuteoflimitationshasexpiredand
victimoffenderrelationship.DatafromThe400Projectwereaninvaluableresourceforthispurpose.
Foranalyzingtheeffectofvictimoffenderrelationshiponforensictestingamonguntested,non
adjudicated,SOLunexpiredSAKs,TestingGroups1and2werecombinedintoasinglesample.The
unweightedgroupswerealmostexactlyequalinsize,whichwasinconsistentwiththerelativefrequency
434
ofstranger(40.9%,N=36)andnonstrangerassaults(59.1%,N=52)amongthe88SOLunexpiredSAKs
fromTestingGroup5forwhichvictimoffenderrelationshipwasknown.Thereforeweusedthose
proportionsassamplingweightsinadisproportionate,stratifiedsampledesign(Valliant,Dever,&
Kreuter,2013).TheweighteddatasetproperlyaccountedforthedisproportionateoccurrenceofSAKs
fromstrangerandnonstrangerassaults.
Foranalyzingtheeffectofstatuteoflimitation(SOL)statusonforensictestingratesamong
untestedDetroitSAKs,TestingGroups3and4werecombinedintoasinglesample.Theunweighted
groupswerealmostexactlyequalinsize,whichwasinconsistentwiththerelativefrequencyofSOL
expired(63.6%,N=159)andSOLunexpired(36.4%,N=91)SAKsamongthe250untestedSAKsfrom
TestingGroup5.Thereforeweusedthoseproportionsassamplingweightsinadisproportionate,
stratifiedsampledesign(Valliant,Dever,&Kreuter,2013).Theweighteddatasetproperlyaccountedfor
thedisproportionateoccurrenceofSOLexpiredandSOLunexpiredSAKs.
Groups13werealltreatedassimplerandomsamplesfromthesubpopulationsofinterest
showninTableB.12whenrunningthepredictionmodels.ThemodelsforTestingGroups1and2should
technicallyhavebeentreatedasstratifiedbyyearandweightedaccordingly,butTestingGroup5was
toosmalltoprovidereasonableestimatesofpopulationproportionsbyyear.TestingGroup3wasa
simplerandomsamplebydesign,sotherewasnoneedtoadjustforthesamplingdesignbefore
generalizingthefindingstothetargetsubpopulation.
NoneoftheanalysesthatusedonlyTestingGroup4datarequiredsampleweightingbecause
Group4wasasimplerandomsamplefromthesubpopulationofuntested,SOLunexpiredDetroitSAKs.
Itwasthereforeappropriatetogeneralizefindingstothatsubpopulationwithoutfirstadjustingforthe
samplingdesign.
Nonindependentobservations:Thepresenceofserialsexualoffendersinthepopulation
impliesthatSAKsfromthesameoffendercouldbenonindependent(forensictestingoutcomescould
435
becorrelated).Normally,thiswouldbecausetoselectstatisticalmethodsthatexplicitlyaccountfor
nonindependence.Weconsideredtwogeneraltypesofsuchmethods:generalizedlinearmixedmodels
(GLMMs;Gelman&Hill,2007;Hox,2010;Raudenbush&Bryk,2002)andgeneralizedestimating
equation(GEE)models(Hanley,Negassa,deB.Edwardes,&Forrester,2003;Hardin&Hilbe,2002).Both
wouldtreatSAKs(level1units)asobservationsclusteredwithinuniqueoffenders(level2units)and
permitestimatingfixedeffectsofpredictorscorrectlyadjustedfornonindependence,buttheydifferin
whetherinterpretationfocusesonsubjectspecific(GLMMs)ormarginal,populationaverageeffects
(GEE)(Gardiner,Luo,&Roman,2009;Subramanian&OMalley,2010).
WeultimatelyoptedagainstusingeitherGLMMsorGEEbecausetheamountofactual
clusteringinthedatawasverysmall:wehad1,595SAKs(eachpresumablyfromauniquevictim)
associatedwithatotalof1,563uniqueoffenders.Therewereonly29serialoffendersidentifiedbycase
tocasehitstootherSAKswithinthedataset(3associatedwith3SAKseach,plus26with2SAKseach;
theother1,534offenderswereeachassociatedwithonlyoneSAK).TheratioofSAKstounique
offendersshowsthattheaverageclustersizewasjust1.02SAKsperuniqueoffender.
Completeindependencewouldyieldanaverageclustersizeofexactly1,sothisisclearlyan
extremelysparselyclustereddataset.Hence,welookedatthemethodsliteratureonmodelingsparsely
clustereddata.Simulationsformodelswithbinaryoutcomessuggestthatvalidandreliableestimates
canbeobtainedfromGLMMswithaverageclustersizeofatleast5(Clarke,2008;McNeish,2014)and
fromGEEmodelswithaverageclustersizeofatleast2providedtherearemanyclusters(McNeish,
2014).Whenaverageclustersizesrangefrom2to5,GLMMsmayyieldbiasedestimatesofselected
parameters(e.g.,betweenclustervariance),butignoringclusteringalsocausessomeproblems(Clarke,
2008).Wewereunabletofindanymethodologicalresearchfocusingonsituationsasextremelysparsely
clusteredasthedatawewereanalyzing.Thesmallestaverageclustersizeexaminedinthesesimulations
was1.4(Clarke,2008).
436
Conductinganewsimulationtoexaminethatwasbeyondthescopeofthisproject.Giventhat
ouraverageclustersizeisverynearlyequalto1(whichwouldindicatenoclusteringatall),wesuspect
anybiascausedbyignoringtheclusteringisnegligibleandthatapplyingsimplermethodsispreferable
totryingtoapplyGLMMsorGEEunderconditionslikelytocauseestimationproblems.Hencethe
continuationratioandlogisticregressionmodelsdescribedbelowtreatthedataasindependent
observations.
Missingdata:Missingdatafrequentlyarisesinstatisticalanalysesandwascertainlyanissue
here.Dataweremissingintwocontexts;firstlywithdatafromThe400Projectthatwasusedinthe
contextofobtainingweightsinthedisproportionatelystratifiedanalysis,secondlywithmissingdata
fromsomeofthevictim,assailantandassaultcharacteristicsusedaspredictorsinthelogisticregression
models.InthecaseofthedatafromThe400Project,threeofthe250SAKsweremissingvictim
offenderrelationshipinformation.TheseSAKswereexcludedwhiledevelopingtheweightsforthe
disproportionatelystratifiedsamplingschemethatwereappliedwhenanalyzingthevictimoffender
relationshipeffect.
Therewerenomissingdataindatasetsusedforthecontinuationratiomodels.Forthelogistic
regressionmodels,wetreatedmissingdataonassailantageasanexplicitcategoryofthatpredictor.We
definedthefourassaultcharacteristics(multipleperpetrators,involvementofalcohol/drugs,useof
weapons,anduseofphysicalforce)asindicatorsofexplicitevidencethatthesecharacteristicswere
present.Lackofsuchevidencewastreatedthesameasexplicitdocumentationthattheywerenot
present(theywerecodedas0=no).Finally,missingvictimageandexamtimingdatawerepresumedto
bemissingcompletelyatrandombecausecollectingthesevariablesisroutinepracticeinmedical
forensicexams.Stakeholdersfamiliarwithsuchsettingsassuredusthattheonlyreasonsuchdatawould
bemissingwouldbeduetohumanerroronthepartofhospitalstaff.Weusedlistwisedeletionto
excludeSAKswithmissingdataonthosevariables,conductingtheanalyseswiththeremainingSAKs.
437
SelectingAppropriateStatisticalTechniques&Software.Belowwesummarizethestatistical
modelingtechniquesweselectedtoaddresstheprimaryresearchquestionsandthesoftwareusedto
conductthoseanalyses.
Continuationratiomodels:Weusedcontinuationratiomodels(Agresti,2002;Hosmer,
Lemeshow,&Sturdivant,2013)toquantifyandcomparetheDNAtesting,CODISentry,CODIShit,and
serialassaultrates.Continuationratiomodelsarewellsuitedtomodelingordinaloutcomesresulting
fromsequentialselectionprocesses(Agresti,2002),sothismethodisperfectlymatchedtothegoalof
howSAKsprogressedthroughthevariousstagesoftheforensictestingprocess.Let indicatethestage
atwhicheachkitstops,sothatYhaspossibleoutcomesrangingfrom1toJ,whereJisthefinalstage
thatcanbereachedintheprocess.GiventhatanSAKreachesstagej,thecontinuationratiomodel
modelstheprobabilitythatthekitcontinuestogobeyondstagej,usingalogitlinkfunctiontoregressit
onpotentialpredictors withcommonregressioncoefficients ,butvaryingintercepts
forstagej
from1toJ1.WeonlyneedJ1interceptsbecausetheprobabilityofstoppingatthelaststageis1.00by
definition(otherwiseitwouldnotbethelaststageintheprocess).Thecontinuationratiologitmodel
canthenbeformallywrittenasfollows:
|
wherethelogisticfunctiontakesvalue 0,1 andhastheform

Thisillustratesthatthesemodelsareinfactanextensionofthemorefamiliarlogisticregressionmodel.
Indeed,onemethodforfittingthesemodelssimplyinvolvesrestructuringthedatasetthenapplying
standardlogisticregressionsoftware.Becausewealsoneededtoincorporatesamplingweights,we
438
usedageneralizedlinearmodelfunctioninRthatwasdevelopedtoworkwithcomplexsurveysampling
designs(Lumley,2004,2010,2012).
Wetooktheunusualstepofomittingthenormalintercepttermsothatbinarypredictors(like
SOLstatusandvictimoffenderrelationship)wouldyieldseparateestimatesforeachgroupdefinedbya
predictoratthefirststageconsideredineachmodel.Maineffectsforstageandthefocalpredictor
(victimoffenderrelationship,SOLstatus,andDNAtestingmethod,dependingonthespecificmodel
considered)werealwaysincludedinthemodel.Addingstagexpredictorinteractioneffectstoeach
modelmadeiteasiertocalculatethedifferentoutcomeratesofinterestundertheassumptionthatthe
effectofthepredictorvariedacrossstages.ForthefinalDNAtestingmethodanalysis,weaddedamain
effectforpresenceofsperm,plusallpossible2and3wayinteractionsbetweenpresenceofsperm,
stage,andthefocalpredictor.Estimatingcarefullydefinedcontrastssimplifiedtheprocessofobtaining
thestagespecificoutcomeratesforeachgroupdefinedbyafocalpredictor,brokendownbythe
presenceorabsenceofspermasneeded.
Contrastsandestimablelinearfunctionsofcombinationsofthemodelcoefficientswere
obtainedbypostprocessingthemodeloutputs.Theestimatesproducedinthemodelandthepost
processingresultsweretransformedtocalculateoddsratios(ORs)andprobabilities(i.e.,rates
expressedasproportions).Thesequantitieswerethenusedtocalculatetherelativerisk(RR)andthe
numberneededtosubmit(NNS)statistics.
Logisticregressionmodels:Weusedstandardlogisticregressionmodels(Agresti,2002;Hosmer,
Lemeshow,&Sturdivant,2013)topredictunconditionalCODIShitratesseparatelyinTestingGroups13
asafunctionofvictim,assailant,andassaultcharacteristics.Let beabinaryvariabletomodel
denotingwhetherornotanSAKyieldedaCODIShit,sothatYhaspossibleoutcomes0and1,where1
indicatesaCODIShit.ThelogisticregressionmodelmodelstheprobabilitythattheSAKyieldsaCODIS
439
hit,usingalogitlinkfunctiontoregressitonpotentialpredictors withintercept andregression

coefficients .Thelogisticregressionmodelcanthenbeformallywrittenasfollows:
1|
wherethelogisticfunctiontakesvalue 0,1 andhastheform

TheestimatesproducedinthemodelweretransformedintoORsandprobabilities(i.e.,ratesexpressed
asproportions)asneeded.Weretainedthenormalinterceptterminthesemodels,interpretingitasthe
CODIShitrateforareferencegroupinwhicheverypredictorwassettoavalueofzero.
Equivalencetests:ToconducttheequivalencetestsfortheDNAtestingmethodexperiment,
weadoptedthecommontwoonesidedtests(TOST)strategyataTypeIerrorrateof=0.05by
computingtwotailed90%CIsfortherawdifferencebetweentheCODISentryrates(whichare
proportions)forthetraditionaltestingandDNasetestinggroups(Barker,Luman,McCauley,&Chu,
2002;daSilva,Logan,&Klein,2009;Rogers,Howard,&Vessey,1993;Stegner,Bostrom,&Greenfield,
1996;Wellek,2010).Thisteststhenullhypothesis|pTpD|againstthealternatehypothesis|pTpD|
<,where=0.05wasthemarginofequivalencedefiningatolerablysmalldifferenceinperformance
(daSilva,Logan,&Klein,2009).Weusedtheratesestimatedfromtheassociatedcontinuationratio
modelforpTandpD.
AlthoughonecanalsodefineequivalencemarginsintermsofORs(daSilva,Logan,&Klein,
2009),thoseareveryabstractquantitiesandimplythattheactualwidthofthemargintranslatedback
intoasimpledifferencebetweentheproportionsdependsontheproportioninthereferencegroup.
Wedecidedthatdefiningthemarginsintermsofasimple,constantdifferencebetweentheCODISentry
rateswaseasierforstakeholderstounderstandandinterpret.Wesetthemarginofequivalencebased
onfeedbackfromourstakeholders(twoatthestatelevelandtwomoreatthenationallevel),whosaid
440
theywouldconsiderCODISentryratesforthetwogroupsequivalentiftheydifferedbynomorethan
5%(0.05<<0.05,where=pTpD).EachCIwasestimatedviaascoreintervalforthedifference
betweentwoindependentproportions(Agresti,2002).
Statisticalsoftware:WeusedR3.1.1(RDevelopmentCoreTeam,2014)toperformthe
analysespresentedinthisdocument.Risfree,opensourcestatisticalcomputingsoftwarethatiswidely
regardedashighqualitysoftwarebyprofessionalstatisticians.Weusedvarioussummaryfunctionsfrom
thedescrpackage(Aquino,Enzmann,Schwartz,Jain,&Kraft,2013).Weusedthesvydesignandsvyglm
functions(Lumley,2004,2010,2012)toweightthedataandrunthecontinuationratiomodels,
obtainingadditionalcontrastsviatheestimablefunction(Warnes,Bolker,Lumley,&Johnson,2013).We
graphedthemodelresultswiththedotplot(Sarkar,2008)andxYplot(HarrellJr.,2014)functions.For
theequivalencetests,weusedthediffscorecifunction(Scherer,2014).Forthelogisticregression
models,weusedtheglmfunctionfromthebaseRsoftware(RDevelopmentCoreTeam,2014).The
datafilesandsoftwarecoderequiredtoreplicatetheanalyseswillbearchivedintheNationalArchiveof
CriminalJusticeData(NACJD).
Detailed Statistical Results
Theparameterestimatesfromthecontinuationratiomodelsevaluatingtheeffectofvictim
offenderrelationshipandSOLstatue(seeChapter4)arepresentedinTablesB.13andB.14(following
pages).Wepresentthesemoredetailedtablesofthemodelresultstodocumentparameterestimates,
standarderrors,andmodelfitinformationthatwereomittedfromChapter4forthesakeofbrevity.
Similarly,TablesB.15B.17(followingpages)presentmoredetailedresultsofthelogisticregression
analysesthatexaminedthepredictiveutilityofvictim,assailant,andassailantcharacteristics(within
TestingGroups1,2,and3).TablesB.18andB.19(followingpages)presentthedetailedresultsfrom
TestingGroup4:theexperimentalevaluationoftwodifferenttestingmethods(traditionalvs.DNase).
441
TABLEB.13ContinuationRatioModelforVictimOffenderRelationshipEffect
ModelEffects
Parameter
Estimate
Stage
Stage1>2:CODISentryrate|Non
stranger
Stage2>3:CODIShitrate|Non
stranger
Stage3>4:Serialhitrate|Non
stranger
VictimOffenderRelationship(VOR)
Stage1>2:Strangereffecton
CODISentryrate
StagebyVORInteraction
Stage2>3:Strangereffectadj.for
CODIShitrate
Stage3>4:Strangereffectadj.for
Serialhitrate
DerivedContrasts
SE
Rate
0.402
0.096
0.401
0.291
0.151
0.572
1.552
0.260
0.175
0.550
0.135
0.634
pvalue
[0.554,
0.808]
[0.996,
1.797]
[0.127,
0.352]
<0.001
0.669
1.338
0.212
0.054
<0.001
[0.571,
0.693]
1.734
[1.330,
2.261]
<0.001
[0.502,
1.307]
[0.681,
2.570]
0.389
OR 95%
CIa
pvalue
0.448
0.280
0.339
0.569
Estimate
SE
Rate
Stage1>2:CODISentryrate|Stranger
0.149
0.095
0.537
Stage2>3:CODIShitrate|Stranger
0.631
0.136
0.653
Stage3>4:Serialhitrate|Stranger
0.722
0.171
0.327
Estimate
SE
Rateb
0.340
0.830
Stage2>3:StrangereffectonCODIS
hitrate
Stage3>4:StrangereffectonSerial
hitrate
[0.357,
0.447]
[0.499,
0.643]
[0.113,
0.260]
OR 95%
CI
0.244
SimpleMainEffects
Odds
Ratio
0.210
SpecificRates
Rate95%
CI
[0.334,
0.567]
[0.405,
0.720]
0.810
Rate95%
CI
[0.490,
0.584]
[0.589,
0.711]
[0.257,
0.405]
Odds
Ratioa
Rate95%
CIb
0.203
Odds
Ratio
1.405
0.311
2.294
1.323
0.409
0.118
<0.001
<0.001
OR 95%
CI
[0.939,
2.102]
[1.238,
4.248]
pvalue
0.094
0.008
Note:Theseresultsgeneralizetothesubpopulationofuntested,nonadjudicated,SOLunexpiredDetroitSAKsforwhichvictim
offenderrelationshipdataareavailable.TheyarebasedonanalysesofN=894SAKs(449nonstrangerassaultsand445stranger
assaults).TheseestimateswereobtainedfromacontinuationratiomodelofSAKprogressionacrossStages14,weightedtoreflectthe
disproportionatestratifiedsamplingdesign.Modelfitstatistics:totaldf=1572,residualdf=1566,nulldeviance=2179,residual
deviance=2060,AIC=2078.
a.
OddsratiosandcorrespondingCIsarenotreportedbecausethesecontrastscombinecoefficientsintovaluesthataremore
meaningfulwhentransformedbackintostagespecifictransitionratesforparticularsubsetsofSAKs.
b.
RatesandcorrespondingCIsarenotreportedbecausethesecontrastscombinecoefficientstodirectlyquantifythesimplemaineffect
ofVORontherateforaparticularstagetransition;oddsratiosareamoremeaningfulmetricforexaminingadifferencebetweenthe
ratesobservedintwosubsetsofSAKs.
442
TABLEB.14ContinuationRatioModelforStatuteofLimitations(SOL)Effect
ModelEffects
Parameter
Estimate
SE
Rate
Rate
95%CI
Odds
Ratio
OR95%
CI
p
value
Stage1>2:CODISentryrate|
Expired
Stage2>3:CODIShitrate|Expired
0.028
0.107
0.493
0.972
0.152
0.520
Stage3>4:Serialhitrate|Expired
0.744
0.226
0.322
[0.788,
1.198]
[0.805,
1.461]
[0.305,
0.740]
0.790
0.081
[0.441,
0.545]
[0.446,
0.594]
[0.234,
0.425]
Stage
StatuteofLimitations(SOL)Status
1.084
0.475
0.595
0.001
Stage1>2:Unexpiredeffecton
CODISentryrate
StagebySOLStatusInteraction
0.235
0.152
0.558
Stage2>3:Unexpiredeffectadj.
forCODIShitrate
Stage3>4:Unexpiredeffectadj.
forSerialhitrate
DerivedContrasts
0.118
0.259
0.470
0.468
0.348
0.385
Estimate
SE
Rate
[0.484,
0.630]
1.265
[0.940,
1.702]
0.121
[0.535,
1.476]
[0.316,
1.240]
0.648
OR95%
CIa
p
value
0.055
SpecificRates
[0.348,
0.596]
[0.240,
0.554]
0.888
Rate
95%CI
[0.498,
0.603]
[0.478,
0.619]
[0.196,
0.367]
Rate
95%CIb
Odds
Ratioa
0.626
Unexpired
Stage2>3:CODIShitrate|Unexpired
0.206
0.108
0.551
0.198
0.145
0.549
Stage3>4:Serialhitrate|Unexpired
0.977
0.218
0.274
Estimate
SE
Rateb
0.117
0.210
Odds
Ratio
1.124
0.233
0.314
0.792
SimpleMainEffects
CODIShitrate
Serialhitrate
0.180
0.172
<0.001
OR95%
CI
[0.741,
1.705]
[0.425,
1.476]
p
value
0.579
0.450
Note:TheseresultsgeneralizetothepopulationofuntestedDetroitSAKs(regardlessofadjudicationstatusorvictimoffender
relationship).TheyarebasedonanalysesofN=701SAKs(351SOLexpiredassaultsand350SOLunexpiredassaults).These
estimateswereobtainedfromacontinuationratiomodelofSAKprogressionacrossStages14,weightedtoreflectthe
disproportionatestratifiedsamplingdesign.Modelfitstatistics:totaldf=1263,residualdf=1257,nulldeviance=1751,residual
deviance=1715,AIC=1723.
a.
b.
RatesandcorrespondingCIsarenotreportedbecausethesecontrastscombinecoefficientstodirectlyquantifythesimple
maineffectofSOLstatusontherateforaparticularstagetransition;oddsratiosareamoremeaningfulmetricforexamininga
differencebetweentheratesobservedintwosubsetsofSAKs.
443
TABLEB.15LogisticRegressionPredictingCODISHitRatesAmongTestingGroup1(StrangerRape)
Parameter
Estimate
SE
zvalue
OR
OR95%CI
a
Victimage<16years
pvalue
1.027
0.312
3.287
0.358 [0.192,0.655]
0.001
0.656
0.400
1.643
0.519 [0.224,1.093]
0.100
Assailantageb
<22years
Unknown
0.118
0.404
0.293
0.888 [0.390,1.921]
0.769
0.290
0.253
1.147
1.337 [0.813,2.198]
0.251
0.033
0.251
0.131
0.968 [0.588,1.578]
0.896
Useofalcohol/drugs
0.015
0.243
0.063
0.985 [0.609,1.579]
0.950
0.515
0.228
2.259
1.674 [1.072,2.624]
0.024
0.174
0.263
0.662
1.190 [0.715,2.011]
0.508
1dayafterassault
0.072
0.265
0.273
0.930 [0.549,1.557]
0.785
2+daysafterassault
0.551
0.389
1.416
0.576 [0.257,1.198]
0.157
Multipleperpetrators
c
Useofweapons
c
Useofphysicalforce
d
Examtiming
missingcompletelyatrandom).OR=oddsratio(ameasureofeffectsize);CI=confidenceinterval;pvalue=pvalueobtained
fromaWaldtestofthecoefficient.Referencelevelsare:a16+years;b22+years;cno;andd0daysafterassault(sameday).
Modelfitstatistics:totaldf=390,residualdf=381,nulldeviance=491.5,residualdeviance=477.7,AIC=497.7.
444
TABLEB.16LogisticRegressionPredictingCODISHitRatesAmongTestingGroup2(NonStrangerRape)
Parameter
a
Victimage<16years
Estimate
SE
zvalue
0.988
0.340
OR95%CI
pvalue
2.906
0.372 [0.188,0.716]
0.004
0.908
0.375
2.424
0.403 [0.189,0.825]
0.015
Assailantage
OR
<22years
0.014
0.333
0.043
1.014 [0.522,1.939]
0.966
Unknown
0.886
0.533
1.661
2.424 [0.820,6.829]
0.097
0.180
0.341
0.527
1.197 [0.601,2.306]
0.598
0.443
0.297
1.491
0.642 [0.354,1.138]
0.136
0.768
0.324
2.369
2.155 [1.134,4.058]
0.018
0.340
0.308
1.103
0.712 [0.390,1.312]
0.270
c
Useofalcohol/drugs
c
Useofweapons
c
Useofphysicalforce
d
Examtiming
1dayafterassault
2+daysafterassault
0.524
0.289
1.815
1.689 [0.957,2.977]
0.069
0.591
0.452
1.308
0.554 [0.212,1.278]
0.191
a
b
c
d
fromaWaldtestofthecoefficient.Referencelevelsare: 16+years; 22+years; no;and 0daysafterassault(sameday).
445
TABLEB.17LogisticRegressionPredictingCODISHitRatesAmongTestingGroup3(PresumedSOLExpired)
Parameter
a
Victimage<16years
Estimate
SE
zvalue
OR
OR95%CI
pvalue
1.858
0.473
3.925
0.156
[0.058,0.377]
<0.001
0.202
0.442
0.456
0.817
[0.332,1.901]
0.649
0.587
0.424
1.383
1.799
[0.776,4.136]
0.166
Assailantage
<22years
Unknown
0.155
0.571
0.272
1.168
[0.349,3.408]
0.786
0.490
0.422
1.162
0.612
[0.256,1.355]
0.245
Useofalcohol/drugsc
0.489
0.463
1.057
0.613
[0.231,1.451]
0.290
0.740
0.317
2.337
2.095
[1.131,3.929]
0.019
0.504
0.429
1.174
1.655
[0.741,4.050]
0.240
1dayafterassault
0.011
0.356
0.031
1.011
[0.494,2.007]
0.976
2+daysafterassault
0.359
0.606
0.593
1.432
[0.404,4.538]
0.553
Useofweapons
c
Useofphysicalforce
d
Examtiming
a
b
c
d
fromaWaldtestofthecoefficient.Referencelevelsare: 16+years; 22+years; no;and 0daysafterassault(sameday).
446
TABLEB.18ContinuationRatioModelforDNATestingMethodEffect(ModeratorOmitted)
ModelEffects
Parameter
Estimate
Stage
SE
Rate
Rate
95%CI
Odds
Ratio
OR95%
CI
pvalue
[1.585,
3.003]
[2.676,
6.647]
<0.001
[0.725,
1.811]
0.560
Stage0>1:DNAtestingrate|
Traditional
Traditional
TestingMethod
0.780
0.163
0.686
1.439
0.232
0.808
Stage0>1:DNaseeffecton
DNAtestingrate
StagebyTestingMethod
Interaction
Stage1>2:DNaseeffectadj.for
CODISentryrate
DerivedContrasts
0.136
0.234
0.534
[0.613,
0.750]
[0.728,
0.869]
2.182
[0.420,
0.644]
1.146
4.217
<0.001
SpecificRates
DNase
DNase
SimpleMainEffects
CODISentryrate
0.378
0.392
0.407
[0.241,
0.596]
0.685
[0.318,
1.476]
0.335
Estimate
SE
Rate
pvalue
0.167
0.714
Odds
Ratioa
OR95%
CIa
0.916
1.197
0.212
0.768
Estimate
SE
Rateb
Rate
95%CI
[0.642,
0.777]
[0.685,
0.834]
Rate
95%CIb
0.242
0.314
<0.001
Odds
Ratio
0.785
<0.001
OR95%
CI
[0.421,
1.465]
pvalue
0.441
Note:Theseresultsgeneralizetothesubpopulationofuntested,SOLunexpiredDetroitSAKs(regardlessofadjudicationstatus
orvictimoffenderrelationship).AfterselectingN=350SAKsfromthatsubpopulation,n=175SAKswererandomlyallocatedto
eachofthetwotestingmethods.TheseestimateswereobtainedfromacontinuationratiomodelofSAKprogressionacross
Stages02(unweightedduetothesimplerandomsamplingdesign)thatomittedthepresenceofspermmoderator.Modelfit
statistics:totaldf=595,residualdf=591,nulldeviance=824.8,residualdeviance=680,AIC=688.
a.
b.
maineffectofDNAtestingmethodontherateforaparticularstagetransition;oddsratiosareamoremeaningfulmetricfor
examiningadifferencebetweentheratesobservedintwosubsetsofSAKs.
447
TABLEB.19ContinuationRatioModelforDNATestingMethodEffect(ModeratorIncluded)
ModelEffects
Parameter
Estimate
Stage
0sperm,Traditional
0sperm,Traditional
TestingMethod
DNAtestingrate|0sperm
Sperm
SE
Rate
Rate
95%CI
Odds OR95%CI
Ratio
1.705
0.344
0.154
0.180
1.386
0.791
0.200
[0.085,
0.263]
[0.050,
0.541]
0.278
0.471
0.569
[0.344,
0.769]
1.320
0.250
[0.093,
0.357]
[0.053,
1.179]
<0.001
[0.524,
3.323]
0.556
<0.001
21.271
0.355
1.000
[1.000, 1.73E+09 [8.63E+08,

1.000]
3.46E+09]
0.416
1.106
0.602
[0.148,
0.930]
Stage1>2:1spermeffecton
CODISentryrate
TestingMethodbySperm
Interaction
Stage1>2:DNaseeffecttimes
1spermeffectadj.forCODIS
entryrate
StagebyTestingMethodby
SpermInteraction
Stage1>2:1spermtimes
DNasetimesCODISentry
rateeffect
DerivedContrasts
18.039
0sperm,DNase
0sperm,DNase
1sperm,Traditional
1sperm,Traditional
1sperm,DNase
1sperm,DNase
0.080
Stage0>1:1spermeffecton
DNAtestingrate|Traditional
StagebyTestingMethod
Interaction
CODISentryrate
StagebySpermInteraction
SpecificRates
pvalue
1.520
[0.173,
13.244]
0.707
[0.000,
0.000]
<0.001
[0.290,
1.978]
0.571
0.911
0.000
[0.000,
0.000]
0.000
0.278
0.490
0.431
[0.225,
0.664]
0.760
0.784
1.174
0.313
[0.044,
0.820]
0.460
[0.046,
4.555]
0.504
Estimate
SE
Rate
pvalue
0.322
0.194
Odds
Ratioa
OR95%
CIa
1.427
0.693
0.613
0.333
19.566
0.086
1.000
1.846
0.278
0.864
19.566
0.085
1.000
1.477
0.242
0.814
Rate
95%CI
[0.113,
0.312]
[0.129,
0.628]
[1.000,
1.000]
[0.785,
0.917]
[1.000,
1.000]
[0.730,
0.876]
<0.001
0.258
<0.001
<0.001
<0.001
<0.001
448
TABLEB.19ContinuationRatioModelforDNATestingMethodEffect(ModeratorIncluded)[continued]
SimpleMainEffects
Rate
Odds OR95%CI
pvalue
Estimate
SE
Rateb
95%CIb
Ratio
Stage0>1:DNaseeffectonDNA
0.000
0.134
1.000
[0.766,
1.000
testingrate|1sperm
1.305]
0.693
1.001
2.000
[0.275,
0.489
CODISentryrate|0sperm
14.567]
0.369
0.369
0.690
[0.333,
0.317
CODISentryrate|1sperm
1.437]
Note:Theseresultsgeneralizetothesubpopulationofuntested,SOLunexpiredDetroitSAKs(regardlessofadjudicationstatus
orvictimoffenderrelationship).AfterselectingN=350SAKsfromthatsubpopulation,n=175SAKswererandomlyallocatedto
eachofthetwotestingmethods.TheSAKsweresortedbythepresence/absenceofSperminthesamples.Theseestimateswere
obtainedfromacontinuationratiomodelofSAKprogressionacrossStages02(unweightedduetothesimplerandomsampling
design)thatcontaineda3waystagebytestingmethodbypresenceofsperminteractioneffect(themoderator).Modelfit
statistics:totaldf=595,residualdf=587,nulldeviance=824.8,residualdeviance=338.2,AIC=354.2.
a.
b.
maineffectofDNAtestingmethodontherateforaparticularstagetransition;oddsratiosareamoremeaningfulmetricfor
examiningadifferencebetweentheratesobservedintwosubsetsofSAKs.
449
APPENDIX C: Victim Notification Resources
AppendixC1:
AppendixC2:
AppendixC3:
SamplePlanningToolforDevelopingaVictimNotificationProtocol:
VictimNotificationRetreatPlanningGuide
SampleFrequentlyAskedQuestions(FAQ)ResourcePacket:
VictimNotificationCommonQuestionsVictims/SurvivorsMayAsk:
AResourceGuideForPractitioners
SampleVictim/SurvivorCommunityResourcesBrochure:
Detroit/WayneCountyCommunityResources
The NIJ Detroit Sexual

Assault Kit (SAK) Action
Research Project (ARP)
Victim Notification Retreat
Planning Guide
Table of Contents
About this Guide.......... ..3
What is Victim Notification? 4
When Should Planning for Victim Notification Begin? .5
Why a Retreat? ............ 6
Pre-Retreat Discussions... .. 7
Preparing Discussion Topics 8
Retreat Ground-Rules .. 9
Other Tips for Retreat Planning 10
Decisions to Make at the Retreat 11
How should victim notification be approached? 12
When and why should victims be notified? 13
Who will make the decision that victim will be notified? 14
What should be achieved during the first contact? 15
Who should make the first contact with the victim? .. 16
How should the first contact with the victim be made? 17
What information should be given to victims at first contact?......... 18
What should happen after the initial contact? . 19
How should victim notification staff be trained? .. 20
A Discussion on Public Apologies 21
Evaluating the Protocol.. 22
Protecting Confidentiality After the Retreat 23
About this Guide

Developing a victim notification protocol is no easy task.
The NIJ Detroit SAK ARP discovered this first hand when
attempting to put together a victim notification
protocol for Detroit, MI. The collaborative team
ultimately decided to host a two-day retreat to in order
to have the sufficient time and focus to develop a
victim notification protocol.
The following pages provide definitions, discussion
topics, and possible decisions to consider when
planning their victim notification protocol.
There are many steps involved in planning a retreat. This
guide will cover:
1.
Pre-retreat discussions
2.
Preparing discussion topics
3.
Retreat ground rules
4.
Other helpful tips for retreat planning
5.
Key questions to discuss at the retreat
6.
Protecting confidentiality
What is Victim Notification?

Victim notification (VN) is the process of contacting and
informing sexual assault survivors about the status of
their unsubmitted sexual assault kits and/or the
investigation of their case.
There are two main circumstances that will arise with
victim notification:
1
2
COLD CASES
Contacting victims of a new CODIS hit that has emerged from an old case.
UNSUBMITTED KITS
Testing previously unsubmitted/untested SAKs and moving forward with
investigation and possible prosecution after some period of time.
When Should Planning for Victim Notification Begin?
It is useful to begin planning for victim notification before

SAKs are submitted for testing. Once kits have been
submitted, it then becomes critical to develop a plan
regarding how to notify victims regarding the testing results.
Initial planning should include: developing the team that will
be working on victim notification, reaching out to front-line
practitioners who have done victim notification before,
reviewing tools from other national victim organizations, and
compiling other resources about victim notification.
Why A Retreat?
Developing a victim notification protocol will raise many

complex issues. It may be difficult for a collaborative team
to devote sufficient time at regular meetings to discuss these
issues in-depth. A retreat provides sufficient time and focus
to explore topics of victim notification uninterrupted.
Also keep in mind that a notification protocol must be
developed by the time that SAKs return from testing.
Pre-Retreat Discussions
There are many topics that a collaborative team might want
to have a clear understanding on before the retreat. Having
certain elements of victim notification established
beforehand will save time and will give the group a solid
foundation upon which to begin planning. Consider:
1
2
TAKE INVENTORY
Establish a thorough understanding of how your community is already handling
victim notification (e.g., in cold cases). Is there already a clearly-defined way that cold
cases are being handled and how can that inform current efforts?
ESTABLISH GOALS
As with any major project, it is crucial to establish goals; this is especially
important when hosting a victim notification retreat as there will be limited time to
address a multitude of issues. Ensure that the goals established can be accomplished;
do not overwhelm retreat participants with unreasonable expectations.
Preparing Discussion Topics
Develop a list of key questions that will need to be

resolved/discussed at your retreat. This guide provides nine
(9) key questions, although your group might have other
imperative issues that should be discussed.
It may be helpful to establish the possible options for each
decision ahead of time; by doing this, your group will spend
less time thinking of different options and will spend more
time deciding which option is right for your community.
Retreat Ground-Rules
It is extremely important to establish ground-rules for your
retreat ahead of time. Below are some ground rules that
may provide a useful starting place for your planning.
Designate someone to be in charge of making sure the
group follows the agreed upon guidelines. When possible,
consider a professional retreat facilitator.
1. Treat others with equity; while titles and positions cannot truly be left at the
door, everyones voices can be heard and opinions respected.
2. Everyones participation is needed to accomplish the goals of the retreat;
be present and engaged for all sessions. Please silence all cell phones.
3. Listen as an ally; create a safe space for the sharing of varied opinions.
4. Listen carefully to each others contributions; avoid thinking about how to
express your own response or concern while someone else is sharing.
5. Only one person speaks at a time; do not interrupt one-another and avoid
side-bar discussions.
6. Honor time limits; share relevant information with the group, but spend just
as much time thinking and reflecting on what others are saying.
7. Agree on what important words and ideas mean; seek clarification when
you do not understand anothers point or terminology.
8. Say and, not but; add to others comments and ideas by using and
rather than but, a word that tends to dismiss others ideas.
9. Disagree respectfully; it is acceptable to challenge and question other
peoples ideas, but do so respectfully and politely.
10. Chase elephants, not rabbits; focus on making decisions about the large
topics at hand and challenge the group if it seems to be getting off-course.
11. Discuss the un-discussable issues; speak courageously about the issues that
are important to you and your community.
Other Useful Tips for Planning the Retreat

1. The retreat can be held almost anywhere, but due to the
subject matter, be sure that the discussion cannot be
overhead and that privacy/confidentiality are protected.
2. Decide how long the retreat will be. This guide works well with
a 1 to 2-day retreat, but can be modified for other lengths
of time.
3. Establish who will lead the retreat; consider a professional
retreat facilitator or a leader who can keep the group on task.
4. Be sure to bring extra materials and supplies, such as: markers,
pens, pads of paper, post-it notes, name tags, etc.
5. Refreshments and meals are crucial to a positive retreat
experience.
6. Present a written agenda/ schedule for the retreat; make sure
to include designated meal times and breaks to keep
attendees motivated.
7. Maintain variety: vary the time of each presentation, the type
of interaction, and the topics discusses throughout the retreat.
8. Encourage attendees to get refreshments, use the restroom,
and/or stretch and move around as needed.
9. Incorporate socializing into the different scheduled activities.
10. Decide how decisions will be made ahead of time:
consensus, majority rule, or what the leader says goes.
10
Decisions to be Made at the Retreat

As mentioned previously, it is important to prepare a list of
decisions that must be made during the retreat. The
following are nine recommended topics. Elaboration on
each may be found on the consecutive pages.
KEY QUESTIONS:
1.
How should victim notification be approached?
2.
When and why should victims be notified?
3.
Who will make the decision that victim will be notified?
4.
What should be achieved during the first contact?
5.
Who should make the first contact with the victim?
6.
How should the first contact with the victim be made?
7.
What information should be given to victims at first contact?
8.
What should happen after the initial contact?
9.
How should victim notification staff be trained?
11
Discussion Q1: How should VN be approached?

Standardized Approach all cases are
handled the same way
Pros- Less training required. Consistent
response. Easier to evaluate.
Cons- Method may not be the best approach
for each case.
OR
Case-by-Case Approach-- notification
methods are tailored to each individual case
How should
VN be
approached?
Pros- Attend to unique needs of each victim/

case.
Cons- Multiple methods used; requires more
training to address all possible situations;
difficult to evaluate because the methods
vary.
OR
Hybrid Approach each case is informed by a
set of guiding principles, but how these are
executed is decided on a case-by-case basis.
Pros- Establish guiding principles that allow for
case-by-case assessment to guide
notification.
Cons- Time and effort to establish guiding
principles
Discussion Points:
If we use a standardized approach, how can we try to ensure that it applies to
all/most cases?
If we use a customized approach, how can we prepare providers to know when and
why to use certain VN methods?
Can we develop a set of guiding principles for how to do VN AND customize what
strategies to use for each case/victim?
Does the type of case (stranger, non-stranger) impact this decision?
12
Discussion Q2: When and why are victims notified?

When and why are
victims notified and
how does this vary
by DNA results?
DNA with CODIS hit

Active outreach to victims
Discussion Points:
DNA with no CODIS hit yet
Are there scenarios in which we would

not notify victims of the testing results?
Active outreach to victims
Does the type of assault (stranger,

non-stranger) impact this decision?
At what point will we notify the victim?
Do we want to notify only if the case
will go forward to prosecution? How
do we prevent false hope?
No DNA found
No active outreach to
victims but make testing
results available.
What is the purpose of involving the

victim before, during, and/or after
case is reopened?
Different Types of SAK Testing Results
1
2
3
DNA WITH CODIS HIT: Testing identified a DNA profile and a perpetrator
was possibly identified through a CODIS hit.
DNA WITH NO CODIS HIT YET: Testing identified a DNA profile but DNA
was not (yet) linked to a possible perpetrator in the CODIS database.
NO DNA FOUND: Testing did not produce a DNA profile. Identity of
perpetrator will not be revealed through DNA.
13
Discussion Q3: Who makes the decision

that a victim will be notified?
Law Enforcement Investigator (solely)
Who makes the

decision that a
victim will be
notified?
Pros- Qualified SCU investigator has

experience handling SA cases and may know
what type of case is more likely to be
warranted.
Cons- Doesnt bring in the perspective of
other team members (prosecutors,
advocacy). Does not keep everyone
informed and up-to-date on case progress.
Prosecutor (solely)
Discussion Points:
Does the type of assault
(stranger, non-stranger)
impact this decision?
What do we think will lead to
better outcomes (for the
victim and for prosecution)?
What is needed to create a
team review process where
everyone is involved and wellinformed about each case?
If we decide on a team
approach, where will the
team be housed? Should
one agency take the lead?
If so, which one?
Pros- Qualified SAT prosecutor has experience

handling SA cases and may know what type
of case is more likely to be successfully
prosecuted.
other team members (LE, advocacy). Does
not keep everyone informed and up-to-date
on case progress.
Advocate (solely)
Pros- Qualified advocate has experience and
knowledge or victims emotional and mental
health needs.
other team members (LE, prosecutors). Does
not keep everyone informed and up-to-date
on case progress.
Multidisciplinary Team
Pros- Decisions on what will happen in each
case is made as a group (LE, Prosecutor,
Advocacy), taking expertise from each
discipline into consideration. Biases more
likely to be challenged and resolved. All
parties are informed and up-to-date on case
progress.
Cons- Requires additional time and effort to
work as a team. Challenging each others
perspectives may lead to conflict.
14
Discussion Q4: What should be achieved

during the first contact?
Initial Contact
(Foot in the Door)
Build rapport, assess situation
Information Sharing
Inform victim about kit, testing,
investigation
What should
be achieved
during the first
contact?
Inform Victims about Services

Inform victim about advocacy,
counseling, other services
Obtain Victim Statement

Interview victim again re: the
assault
Ask About Participation in
Investigation and Prosecution
Assess how willing or able victim
is to be involved in prosecution
Discussion Points:
How do we ensure that initial contact does NOT cause unnecessary trauma?
Are there benefits to making victim notification a process rather than a one-time
experience?
What type of follow-up and continued social services will be needed for victims?
Does the type of case (stranger, non-stranger) impact this decision?
15
Discussion Q5: Who should make the

first contact with the victim?
Legal Investigator Only
Pros- simple, less overwhelming than having
more than one person. Can provide
resources for advocacy when appropriate.
Able to discuss details about case.
Who should
make the first
contact with the
victim?
Cons- Victims distrust of LE, not willing to talk

to LE, or feel like they are in trouble.
Reactivation of trauma worse if prior
negative experience with LE. No other
services immediately available.
OR
Community-Based Advocate only
Discussion Points:
Does the type of assault
(stranger, non-stranger)
impact this decision?
What do we think will lead to
better outcomes (for the
victim and for prosecution)?
What is needed to create a
team review process where
everyone is involved and wellinformed about each case?
If we decide on a team
approach, where will the
team be housed? Should
one agency take the lead?
If so, which one?
Pros- Immediate crises intervention and

services for negative effects of reactivating
trauma. Independence from legal system is
less threatening.
Cons- Survivor may not want crisis
intervention immediately. If advocate is not
informed about case, she may not be able to
answer questions about case.
OR
Legal System-Based Advocate only
Pros- Immediate crises intervention and
services for negative effects of reactivating
trauma. May be able to answer questions
about case.
Cons- Survivor may not want crisis
intervention immediately. May not trust an
advocate that is affiliated with the criminal
justice system.
OR
Investigator AND social worker/advocate

Pros- Ability to answer questions about case
immediately. Crisis intervention and services
for negative effects of reactivating trauma.
Cons- Victims may feel overwhelmed by
multiple people. May not want crisis
intervention immediately.
16
Discussion Q6: How should the first contact be made?

Phone call
How should the

first contact be
made?
Pros: Not invasive as home contact. Fewer

resources used. Build rapport and answer
immediate questions.
Cons: Difficult to locate current #, no answer
or call backs. No follow-up from victim re:
prosecution and/or obtaining services.
Impersonal. Can come across as harassing.
Discussion Points:
When should we use one
strategy over another? Should
there be a process that
outlines which one to try first,
then second, etc.?
Under what
circumstances/types of cases
would we try which strategy
first, second, third, etc.?
In Person Contact
Pros: Personal approach, able to provide
social services immediately. Build rapport and
answer immediate questions.
Cons: Difficult to locate, transient population.
Distrust of LE; not opening door. Perception of
harassing victims. Outing victims in front of
those who live with them/privacy issues.
Requires more time and resources.
How can we work to ensure

that the victims privacy is
protected in each strategy?
Mailed Letter
Is there another strategy we

have not considered?
Pros: Not invasive as personal contact or

phone call. Gives victim time to process
information.
How generic or detailed

should these strategies be?
Who should be involved in
creating these strategies?
What is needed to make
locating victims easier? What
search engines or search
strategies were useful in other
VN projects?
Is there an appropriate time
to close a case after all
strategies are exhausted?
Who should the letters and
phone calls come from?
Investigators? Advocates?
Cons: Need to have current mailing address.

Not personal approach. Not able to address
initial concerns or questions from victim. More
likely to disregard letter altogether. Violation
of privacy if opened by another person.
Public Notice
Pros: Places choice in victims hands. Ability
to reach more people at one time. Ability for
system leader to offer public apology re:
handling of SAKs.
Cons: Requires phone line or website set up
and management of new system. May lead
to public outcry about problem.
17
Discussion Q7: What information should be

given to the victims at first contact?
Original police report
Pros: May give victim sense of control. Can
help victim remember what was reported at
the time of the assault.
Cons: Victim may see reports from officers
who were not trained in SA and therefore not
sensitive to survivors; revictimization.
Original victim statement

Pros: Sense of control. Can help victim
remember what was reported at the assault.
What
information
should be
given to the
victims at first
contact?
Cons: May bring up memories that can retraumatize survivor. Any inconsistencies found
in victim statements may confuse victim.
Could be challenges from defense attorneys.
Information about perpetrator

Pros: May provide closure/healing to victim to
know identity of perpetrator (if not already
known). Become hopeful about case and
may be more willing to participate.
Cons: Victim and/or family could retaliate
against perpetrator. Victim may inadvertently
use this info in ways that may backfire on her
case during trial. False sense of hope .
Community resources for victims
Discussion Points:
What do we do if victims ask for certain documentation? Do they have the right to
access these documents? Are there circumstances in which information may be more
harmful than helpful? If yes, who gets to make that decision, and how is it made?
Should we provide the same information to different types of victims?
18
Discussion Q8: What should happen after initial contact?
Follow-up contact
These meetings could occur at the
community advocacy organizations, or at a
location of the victims choosing. The
content of these meetings would be
determined on a case-by-case basis,
depending on the testing results, the
survivors questions and concerns, the next
steps, etc. The meetings should emphasize
victims control and choice whether to opt
in/opt out of further contact.
What should
happen after
initial
contact?
No follow-up contact
The only acceptable reason for no follow-up
contact is that the victim explicitly stated that
she does not want further contact with legal
investigators. This victim should have been
provided with a list of community resources
and may seek out these resources at her/his
own comfort.
Discussion Points:
How much detail may be provided to victims at first contact if victims are requesting
more information? What information is only appropriate to give at a later contact?
How will a second contact be coordinated with advocates in a timely fashion?
How should investigators approach the possibility that a person identified in CODIA
was a consensual partner? As follow-up, how do they suggest taking a buccal swab
from said consensual partners to prevent their DNA from being uploaded in CODIS?
19
Discussion Q9: How should victim

notification staff be trained?
SA training
Pros: LE, Prosecutors, and advocates will
learn about how DNA and CODIS affects SA
cases. Understand value of DNA in stranger
and non-stranger cases. Able to discuss DNA
and CODIS in laymans terms.
How should
victim
notification staff
be trained?
Cons: Requires additional training time and

resources.
Discussion Points:

resources.
Who needs to be trained re:

notification? Why?
DNA/CODIS training
Pros: LE, Prosecutors, and advocates will learn
about how DNA and CODIS affects SA cases.
Understand value of DNA in stranger and nonstranger cases. Able to discuss DNA and
CODIS in laymans terms.
Scripts and role-plays
What resources are needed

to conduct such training?
Where will those resources
come from?
Pros: Practice victim-centered approach for

notification. Become comfortable with various
situations that may come up during
notification.
Who should develop and

conduct this training?

resources.
Are case reviews useful? How

can we make them most
productive?
Should our training/planning
be tailored to the type of
case? (Stranger vs. Nonstranger)
Written checklist
Pros: Keep providers accountable and on
task. Everyone is on the same page about
best practices.
Cons: Another policy that may or may not
be followed and/or enforced properly. May
not allow for flexibility in the moment of
notification.
Team case reviews
Pros: Input from different
disciplines/approaches to learn from each
other. Analyze case from varying
perspectives.
Cons: Requires time to coordinate bringing
everyone to the table.
20
A Discussion of Public Apologies

It is very possible that some members of your collaborative
group will request that a public apology be issued due to
the fact that SAKs have not been consistently tested. This
topic might raise some concern for your group. Consider the
following pros & cons of issuing an apology:
Pros:
Cons:
Acknowledgement by the
community that SAKs had not
been consistently tested.
Must decide which organization

or organizations will issue the
apology.
Possibly meaningful and

reparative for survivors to have
a public apology.
Potential disagreement
regarding whether an apology
is warranted.
Negative legal ramifications

are unlikely.
Consider potential legal risks of

admission of wrong-doing.
21
Evaluating the Protocol

Consider trying out your newly developed protocol with a
relatively small number of cases and then re-assess how it
worked and what might need to be changed.
Develop tracking tools that can chart the steps/efforts taken
by the legal investigators to find victims for notification & to
record their perceptions of what happened in each
notification contact.
22
Protecting Confidentiality After the Retreat

Updates on the victim notification process should be limited
to those directly involved with the victims, regardless of what
other confidentiality agreements have been put into place.
Be sensitive to the fact that different disciplines involved in
the victim notifications have differing rules regarding the
disclosure of confidential information. For example,
community-based victim advocates are not allowed to
share information about their experiences with victims with
the group (without written authorization to do so).
23
Victim Notification
Common Questions Victims/Survivors May Ask
A Resource for Practitioners
U P D AT E D : F E B R U A R Y 2 0 1 3
Victim Notification
Common Questions Victims/Survivors May Ask
A Resource for Practitioners
Victim notification is the process of contacting victims/survivors about the status

of their cases and/or sexual assault kit (SAK) results. Typically, the term victim
notification is used when a substantial period of time has passed between when
the assault occurred and when law enforcement, advocacy, or other professionals
are attempting to re-contact the victim/survivor.
Victims/survivors will likely have many questions about their cases, the SAK
results, and what might happen next.
The purpose of this document is to provide law enforcement, advocacy, medical
personnel, prosecutors, and other professionals who may be involved in victim
notification with a list of common questions victims/survivors may ask and
sample answers that can be provided.
The answers have been written in clear, simple language so that they can be
understood by the lay person, unfamiliar with the forensic terminology, medical
terminology, or the steps and stages of the criminal justice system.
The answers provide general information that applies to most cases. The information provided in this document should be used in conjunction with casespecific information that can be provided to the victim/survivor
This document is a resource for practitioners conducting victim notification.
Please do not give this to victims/survivors. Instead, use it as a resource to guide
a conversation with victims/survivors.
A resource packet of community services is available to give to victims/survivors
(Detroit-Wayne County Community Services).
2been
Victim
Notification:
Common Questions
published
by the Department.
Opinions or points of view expressed are those of the author(s)
Questions
What is a sexual assault kit (SAK)?

A sexual assault kit is also called a SAK, or frequently referred to as a rape
kit. The SAK is used as part of a medical forensic examination to collect
evidence after a sexual assault. This might include swabs of any area where
there was contact between the victim and the perpetrator or hair samples.
When and where is the SAK done?
The SAK is collected after a sexual assault at a hospital or other medical
facility. It is usually collected within the first 96 hours after the assault, but
sometimes after that time frame. They try to do it as soon as possible so as
to collect as much evidence as possible.
Who does the SAK?
A health care provider, like a doctor or nurse, collects evidence from the
victims/survivors body. This is just ONE of the services a health care
provider can offer. The health care provider can also do a full exam to care
for any injuries, and provide the victim with emergency contraception (the
morning after pill) or other medication (perhaps to prevent contracting a
sexual transmitted disease [STD] or HIV).
Why is the SAK done?
The SAK is intended to help in the collection and preservation of potential evidence in a sexual assault case. Police and prosecutors may use this
evidence to help them investigate a case.
What is evidence?
Evidence is anything that can provide information as to what occurred.
One piece of evidence from the SAK that may be very valuable is biological evidence (such as blood, saliva, and/or semen) that can be tested to
identify the DNA from the perpetrator.
A Practitioner
Resource
been published by the Department. Opinions or points of view expressed
are those of the
author(s)3
What is DNA?
It stands for deoxyribonucleic acid. It exists in human cells, like in blood,
bone, teeth, and hair, and is like a blueprint for how each human should
be built. DNA is similar to fingerprints. Each person has unique fingerprints and each person, except for identical twins, has unique DNA unlike
anyone else.
Why is DNA so valuable/important?
DNA is like a fingerprint. No two people have the same fingerprint, and
no two people, except for identical twins, have the same DNA. If someones fingerprints are found somewhere, it can be used to prove that they
were there. Similarly, if someones DNA is found somewhere, it can be
used to prove they were there.
How does DNA get left behind?
DNA is in human cells. DNA is in saliva and skin cells. It is in sperm,
blood, and other bodily fluids. A perpetrator may leave these fluids or skin
cells on a sexual assault victims body. When these fluids or cells are left
behind, DNA is too and might be able to be collected.
What happens after a SAK is done?
The SAK may contain evidence of a crime, like DNA. It is important that
the healthcare provider handle the evidence with care. If the victim decided to make a police report about the sexual assault and signed a release,
the health care provider will turn over the completed SAK to a police
officer. The next step is to take the SAK to the crime lab so that trained
scientists can analyze it for DNA.
How does the crime lab analyze for DNA?
The crime lab looks at the swabs and other items in the kit. They go
through some of the swabs and samples to check for DNA. It is possible
that they wont find any DNA. If they do find DNA, they will create a
DNA profile.
4been
Victim
Notification:
Common Questions
published
by the Department.
What is a DNA profile and how is it used?

A DNA profile is a unique pattern of genes, specific to an individual.
DNA profiles can be created from biological evidence found at the crime
scene, such as blood, saliva, and/or semen. Once a DNA profile is created, it can identify an unknown suspect, confirm the presence of a known
suspect, and/or connect an offender to multiple crime scenes.
How are DNA profiles matched?
DNA profiles can be stored in a computer database. When a new DNA
profile is created, it can be entered into the computer database and the
database can scan to see if there is a match. In the United States, there is a
master computer database called CODIS.
What is CODIS?
CODIS stands for the Combined DNA Index System. There are two
ways in which DNA is put into CODIS. One way is when a DNA profile
from an unknown perpetrator is created from crime scene evidence and
entered into the database. Another way is when the DNA profile from a
known perpetrator is entered into the database.
How does CODIS work?
CODIS stores DNA profiles. When a new DNA profile is created following a crime, it is entered into CODIS. If there is a match between the
new DNA profile and an existing DNA profile in CODIS, it comes back
as a hit. The new DNA profile will also be stored in CODIS for future
searches.
What is a CODIS hit?
A hit is when a new DNA profile is entered into CODIS and there is a
match between this new DNA profile and some other DNA profile in the
database.
A Practitioner
Resource
are those of the
author(s)5
If there is a CODIS hit, does it mean that

the suspect has been identified?
Not necessarily. The hit could be a match between the DNA collected in
the SAK and the DNA of a known person. But it also could be a match
to a DNA profile of an unknown suspect in another case, from a crime
scene where the suspect has not yet been identified.
Why was the victims kit not analyzed before?
A task force has been formed to look into this issue and figure out why
this happened and to make sure it wont happen again. At this point, we
dont have all the answers, but we are very sorry that this happened in this
case.
What is going to happen next for this case?
The Detroit Police Department and Wayne County Prosecutors Office
are reviewing the information from CODIS. They will be working to
gather more evidence and the victim will be updated when more is known
about the case. The victim can also contact an advocate if he/she has
additional questions or would like to know the current status of the case
(provide advocate information).
6been
Victim
Notification:
Common Questions
published
by the Department.
References
Federal Bureau of Investigation. (2011). Frequently asked questions (FAQs) on the

CODIS program and the national DNA index system. Available at: http://
www.fbi.gov/about-us/lab/codis/codis-and-ndis-fact-sheet
Gaensslen, R.E., & Lee, H.C. (2002). Sexual assault evidence: National assessment
and guidebook. Washington DC: US Department of Justice.
National Center for Victims of Crime. (2007). DNA & crime victims: What victim assistance professionals need to know. Available at: http://www.victimsofcrime.org/docs/Resource%20Library/DNA%20PROVIDER%20
bro.pdf ?sfvrsn=0
National Human Genome Research Institute. (2012). Deoxyribonucleic acid
(DNA).Available at: http://www.genome.gov/25520880
A Practitioner
Resource
are those of the
author(s)7
WA Y N E C O U N T Y P R O S E C U T O R S O F F I C E
1 4 4 1 S T. A N T O I N E , R O O M 1 2 2 0
D E T R O I T, M I 4 8 2 2 6
313-224-7270
DetroitWayne County
Community Resources
U P D AT E D : F E B R U A R Y 2 0 1 3
Table of Contents
SEXUAL ASSAULT SERVICES
DOMESTIC VIOLENCE SERVICES
MENTAL HEALTH SERVICES
SUBSTANCE ABUSE SERVICES

Help Lines
13
Inpatient & Outpatient 14
13
HEALTH & HUMAN SERVICES
17
MEDICAL & DENTAL SERVICES
17
PREGNANCY & WOMENS HEALTH SERVICES
20
HIV TESTING SERVICES
21
DISABILITY SERVICES
21
LGBT SERVICES
22
CLOTHING & FURNITURE ASSISTANCE
23
FOOD ASSISTANCE
24
HOUSING ASSISTANCE
Network 27
Transitional Housing 28
Teen Housing 29
27
JOB TRAINING PROGRAMS
30
SEXUAL ASSAULT SERVICES
Wayne County SAFE (Sexual Assault Forensic Examiner)

Phone: (313) 964-9701
Crisis pager: (313) 430-8000
Address: 2727 2nd Ave, Suite 120, Detroit, MI 48201
Website: www.wcsafe.org
Services provided: Sexual assault medical forensic examinations, free group and
individual counseling, crisis intervention, court accompaniment, and advocacy.
Volunteer opportunities and community outreach efforts.
First Step
Phone: (734) 722-1772, (888) 453-5900
24-hour crisis intervention line: (734) 722-6800
Address:
4400 S. Venoy, Wayne, MI 48184
44567 Pinetree Drive, Plymouth, MI 48170
Website: http://www.firststep-mi.org/
Services provided: Sexual assault medical forensic examinations for women ages
12 and up. Crisis intervention for survivors of domestic violence and sexual
assault. Temporary emergency shelter for survivors; individual counseling or
group sessions available.
Turning Point
Phone: (586) 463-4430
Address: 158 S. Main St., Mt. Clemens, MI 48043
Website: www.turningpointmacomb.org/tp/about-turning-point
Services provided: Sexual assault medical forensic examinations for women
and children. Crisis intervention for survivors of domestic and sexual violence.
Temporary emergency shelter for survivors; individual counseling or group
sessions available.
4been
DetroitWayne
County
published by the
Department. Opinions or points of view expressed are those of the author(s)
Detroit Police Victims Assistance Program

Phone: (313) 833-1660
Address: 4707 Antoine, Suite M-167, Detroit, MI 48201
Website: https://www.detroitmi.gov/DepartmentsandAgencies/
PoliceDepartment/VictimsAssistance.aspx
Services provided: Free individual and group counseling services to sexual
assault and domestic violence survivors, counseling also available for people
whose family members were victims of homicide.
YWCA of Metropolitan Detroit-Sexual Assault and
Advocacy Services and Interim House
Crisis Line (313) 861-5300
Address: P.O. Box 21904, Detroit, MI 48221
Website: www.ywcadetroit.org
Services provided: Referrals, group counseling, individual counseling, 24-hour

hotline, and emergency shelter.
Detroit Receiving Hospital-Life Stress Center
Phone: (313) 745-4811
Address: 4201 St. Antoine Blvd, Detroit, MI 48201
Website: www.drhuhc.org
Services provided: Mental health services for psychological issues and life
stressors. Specializing in helping trauma victims with completing application
for crime-related medical care. They offer individual outpatient therapy,
including psychotherapy for crime victims.
Wayne County Prosecutors Office
Phone: (313) 224-5777 or (517) 224-5800
Address: 1441 St. Antoine, Detroit, MI 48226
Website: www.co.wayne.mi.us/prosecutor/index.htm
Services provided: The prosecutors office will work with victims to help
them locate information regarding their legal case. They will also advocate for
victims rights within the criminal justice system.
been published by the Department. Opinions or points of view expressedCommunity
are those ofResources
the author(s)5
DOMESTIC VIOLENCE SERVICES
National Domestic Violence Hotline 24 Hours/Day

Phone: (800)799-SAFE; (800) 799-7233; TTY (800) 787-3224
Website: www.ndvh.org
Services provided: Crisis intervention for victims and anyone calling on
behalf of a victim. Individuals are provided with safety planning strategies and
resources to agencies in their area.
Detroit Police Victims Assistance Program:
Domestic Violence (2nd Floor)
Phone: (313) 833-9813
Address: 1151 Taylor St. Bldg 6, Detroit, MI 48202
Website: www.detroitmi.gov/DepartmentsandAgencies/
PoliceDepartment/VictimsAssistance/tabid/1928/Default.aspx
Services provided: Free individual and group counseling services to sexual

assault and domestic violence survivors.
Personal Protection Order (PPO) Advocate
Phone: (313) 224-6292
Address: Coleman A. Young Municipal Building,
2 Woodward, Detroit, MI 48226
Website: www.co.wayne.mi.us/wcpo_divisions_dv.htm
Services provided:Free Personal Protection Orders (PPO). There may be a

cost to have the offender served with a copy of the PPO.
6been
DetroitWayne
County
published by the
Interim House Metropolitan YWCA

Crisis Line (313) 861-5300
Address: P.O. Box 21904, Detroit, MI 48221
Website: www.ywcadetroit.org
Services provided: Referrals, group counseling, individual counseling, 24-hour

hotline, and emergency shelter.
First Step
Phone: (888) 453-5900
Address: 4400 S. Venoy, Wayne, MI 48184
Phone: (734) 722-1772
44567 Pinetree Drive, Plymouth, MI 48170
Phone: (734) 416-1111
Website: www.firststep-mi.org
Services provided: SANE (sexual assault nurse examiner) free forensic sexual
exams for women ages 12 and up. Crisis intervention for survivors of domestic
violence and sexual assault. Temporary emergency shelter for survivors,
individual counseling or group sessions available.
Turning Point
Phone: (586) 463-4430
Address: 158 S. Main St., Mt. Clemens, MI 48043
Website: www.turningpointmacomb.org/tp/about-turning-point
Services provided: Free forensic sexual exams for women and children.
Crisis intervention for survivors of domestic and sexual violence. Temporary
emergency shelter for survivors, individual counseling or group sessions
available.
the author(s)7

Phone: (313) 745-4811
Website: www.drhuhc.org/
Services provided: Mental health services for victims of crime, victims of
trauma.
La Vida Program/CHASS Clinic
Phone: (313) 849-3104
Address: 5635 W. Fort St., Detroit, MI 48209
Website: www.chasscenter.org/laVida.html
Services provided: A domestic and intimate partner violence prevention,
support, and education program. Individual counseling and support groups
available for Spanish speaking women and youth.
SAFE (Sisters Acquiring Financial Empowerment)
Phone: (800) 757-4919 xt. 15
Address: 269 Walker Street, #721 Detroit, MI 48207
Website: www.newsafestart.org
Services provided: Domestic violence survivors are taught how to manage
and end the cycle of economic abuse. They also have employment-training
programs to help people find jobs.
8been
DetroitWayne
County
published by the
MENTAL HEALTH SERVICES
Wayne County 24 hour Crisis Line

Phone: (313)-224-7000
Services provided: Information referral line for shelter locations, emergency
care, utility assistance, mental health services, crisis counseling.
Detroit-Wayne County Mental Health &
Human Services-Access Center
Phone: (800)-241-4949
Services provided: Referral to enroll in mental health programs and 24-hour
crisis line for mental health.
Detroit-Wayne County Health & Human Services Crisis Center
Phone: (313) 745-3546
Address: 4201 St. Antoine, Detroit, MI 48201
Website: www.waynecounty.com/hhs_mh.htm
Services provided: Crisis center for individuals needing immediate assistance

receiving medication or mental health.
Black Family Development
Phone: (313) 758-0150
Address: 2995 E. Grand Boulevard, Detroit, MI 48202
Website: www.blackfamilydevelopment.org/index.html
Services provided: Outpatient mental health services and intensive outpatient
services for substance abuse. Provide referrals.
the author(s)9
Care First Community Health

Phone: (313) 846-5020
Address: 8097 Decatur Street, Detroit, MI 48228
Services provided: Outpatient facility for mental health, anger management
classes, parenting classes, domestic violence classes, andbehavioral treatment
for sex offenders
Catholic Social Services of Wayne County
Phone: (313) 883-2100
Address: 9851 Hamilton Ave., Detroit, MI 48202
Website: www.csswayne.org/index.html
Services provided: Mental health treatment and counseling, child welfare

programs, and education training.
Detroit Central City Community Mental Health, Inc.
Phone: (313) 831-3160
Address: 10 Peterboro St., Detroit, MI 48201
Website: www.dccmh.org
Services provided: Outpatient mental health treatment (referrals or walk-ins),

and substance abuse counseling.
Detroit Hispanic Development Corporation
Phone: (313) 967-4880; TTY 800-649-3777
Address: 1211 Trumbull St., Detroit, MI 48216
Website: www.DHDC1.org
Services provided: Mental health counseling for youth. Referrals for clothing,
employment training, and job placement. Substance abuse counseling for low
income or uninsured adults, free tattoo removal for gang members, English
classes, GED, peer support groups, free HIV testing (Must live in Detroit).
10
beenDetroitWayne
published by theCounty

Phone: (313) 745-4811
Website: www.drhuhc.org
Services provided: Mental health services for victims of crime/trauma.
Detroit-Wayne County Community Mental
Health & Human Services Agency
Phone: (313) 833-3232
Address: 640 Temple, 8th fl Detroit, MI 48201
Services provided: Referrals to crisis hotline, mental health services, and
outpatient screening.
Family Service of Detroit and Wayne County
Phone: (313) 579-5989
Address: Peter Claver Building, 450 Eliot, Detroit, MI 48201
Address: 14200 Kercheval Detroit, MI 48215
Address: 13550 E. McNicholas, Detroit, MI
Website: http://www.fsiwc.org
Services provided: Mental health counseling, domestic violence programs,

employee assistance program, sliding scale fee.
Kids-TALK Childrens Advocacy Center
Phone: (313) 833-2970
Address: 40 East Ferry Street, Detroit, MI 48202
Website: http://www.guidance-center.org/kids-talk
Services provided: Comprehensive free treatment to suspected child victims
of sexual abuse, physical abuse, neglect or other forms of psychological trauma.
Services include forensic interviewing, advocacy and therapy as well as forensic
medical examinations through referrals to qualified pediatricians at Childrens
Hospital of Michigan.
Community
Resources
11
are those of
the author(s)
Northeast Guidance Center

Address:
12800 E. Warren Ave, Detroit, MI
20303 Kelly Road, Detroit, MI
2900 Conner Street, Detroit, MI
Website: www.neguidance.org
(313) 824-8000
(313) 245-7000
(313) 824-5623
Services provided: Mental health treatment.
NSO Tumaini Center

Phone: (313) 961-7990
Address: 3430 Third Ave., Detroit, MI 48201
Website: www.NSO-mi.org
Services provided: Adults homeless services including crisis intervention,

mental health services, healthcare, referral services, group and individual
counseling, emergency food, clothing, and shower facilities.
New Center Community Mental Health
Phone: (313) 961-3200
Address: 2051 W. Grand Blvd., Detroit, MI 48208
Website: www.newcentercmhs.org
Central DistrictDetroit
Southwestern DistrictDetroit
Northeastern DistrictDetroit
Western DistrictDetroit
Eastern DistrictDetroit
Northwestern DistrictDetroit
(313) 596-1300
(313) 596-5300
(313) 596-1100
(313) 596-1200
(313) 596-5900
(313) 596-5600
Services provided: Mental health treatment.
12
beenDetroitWayne
SUBSTANCE ABUSE SERVICES
HELP LINES
The National Clearinghouse for Alcohol and Drug Information
Phone: (800) 729-6686
Website: www.ncadi.samhsa.gov
Services provided: Treatment referrals, information about drugs and alcohol

abuse.
Alcoholics Anonymous Support Meetings
Detroit & Wayne County Office:
Phone: (313) 831-2555 10am-6pm or 24hr. hotline (313) 831-5550
Address: 4750 Woodward Ave #407, Detroit, MI 48291
Website: www.aa-semi.org
Services provided: Referrals for meetings, treatment centers, and assessments.

Narcotics Anonymous Meeting
Detroit Help line (248) 543-7200
Website: www.na.org
Services provided: Supportive meetings for addicts wanting a drug-free

lifestyle.
Community
Resources
13
are those of
the author(s)
INPATIENT & OUTPATIENT

CASS: Transitional Housing for Women and Children
(Moms Place)
Phone: (313) 883-2277
Address: 11850 Woodrow Wilson Street, Detroit MI, 48206
Website: www.casscommunity.org
Services provided: Residential program for homeless women and children,
substance abuse treatment program, case management, and housing placement
assistance.
Catholic Social Services of Wayne County
Phone: (313) 883-2100
Address: 9851 Hamilton, Detroit, MI 48202
Website: www.csswayne.org
Services provided: Outpatient substance abuse treatment and prevention,

foster care services, and adoption. Substance abuse funding is free for those on
parole, sliding scale fee for all other residents.
Detroit LIGHT House Program
Phone: (313) 832-1300
Address: 3800 Woodward Ave, Ste 400, Detroit, MI 48201
Website: www.pnlh.org
Services provided: Substance abuse inpatient and intensive outpatient services,

community funding for uninsured individuals (a referral is required).
New Life Home for Recovering Women
Phone: (313) 245-4357
Address: 17131 Gitre St, Detroit, MI 48205
Website: www.sacredheartcenter.com/index.aspx
Services provided: Residential long-term treatment (more than 30 days)
alcohol and drug abuse counseling; sliding fee scale.
14
beenDetroitWayne
Jefferson House- Capuchin Community Center

Phone: (313) 331-8900
Address: 8311 East Jefferson Avenue, Detroit, MI 48214
Website: www.mhweb.org/wayne/jeffhouse.htm
Services provided: Residential substance abuse treatment services for men.
Detroit Recovery Project
Address:
Comprehensive Outpatient Recovery Services Program, 18954 James
Couzens, Detroit, MI 48235; Phone: (313) 864-5306
Trent Recovery Home 355 West Grand Blvd., Detroit, MI 48207
Phone: (313) 579-5462
West Side Recovery Center, 1145 West Grand Blvd, Detroit, MI 48208
Phone: (313) 324-8900
Website: www.recovery4detroit.com
Services provided: Intensive outpatient, free intake assessments, one-on-one
counseling, group therapy, psycho-education relapse prevention, sober living
skills, family counseling, case management, dual diagnosis treatment, free
aftercare groups, driving under the influence classes, drug diversion class,
alcohol and drug screening, sliding scale payment options.
Operation Get Down Inc.
Phone: (313) 921-9422
Address: 10100 Harper St., Detroit, MI 48213
Website: www.operationgetdown.org
Services provided: Residential and outpatient substance abuse treatment,

uninsured, low-income residents receive state funding for services.
Community
Resources
15
are those of
the author(s)
Salvation Army Detroit Harbor Light (Acres of Hope)

Phone: (313) 361-6136
Address: 3737 Lawton Street, Detroit, MI 48208
Website: www.usc.salvationarmy.org/usc/www_usc_detroithl.nsf/vw-text-ind
Services provided: Residential substance abuse treatment, outpatient and
intensive outpatient substance abuse treatment, emergency shelter for single
men, substance abuse treatment program for women.
Self Help Addiction Rehabilitation and Education (SHARE)
Address:
1852 West Grand Blvd., Detroit, Mi 48208
(313) 894-8444
4216 McDougall, Detroit, MI 48207
(313) 923-6300
Website: www.sharinc.org
Services provided: Substance abuse residential and outpatient treatment for
men, and transitional housing, 30-90 day residential programs.
16
beenDetroitWayne
HEALTH & HUMAN SERVICES
United Way 211

Phone: (800) 552-1183
Address: 660 Woodward Avenue, Ste 300, Detroit, MI 48226
Website: http://211us.org
Services Provided: Help with food, housing, utilities, employment,
transportation, health care, counseling, and more.
MEDICAL & DENTAL SERVICES
American Indian Health & Family Services

Phone: (313) 846-3718
Address: 4880 Lawndale, Detroit, MI 48210
Website: www.aihfs.org
Services provided: Referrals for dental services forlow-income Native

Americans.
Childrens Center Clinic
Phone: (313) 833-2895
Address: 79 Alexandrine, Detroit, MI 48201
Website: www.thechildrenscenter.com
Services provided: Hospital dentistry. Comprehensive pediatric care, operation
room dentistry, sedation services, handicapped/special needs patients,
orthodontics, cleft palate/lip management, craniofacial anomalies.
Community
Resources
17
are those of
the author(s)
Community Health and Social Services

Phone: (313) 849-3920
Address: 5635 W. Fort St, Detroit, MI 48209
Website: www.chasscenter.org
Services provided: Primary and preventive medical care to all individuals
living in Detroit. Sliding scale fee, but also accepts Medicaid, private
insurance, and, MIChild.
Covenant Community Care Inc.
Phone: (313) 554-1095
Address: 559 West Grand Blvd., Detroit, MI 48216
Website: www.covenantcommunitycare.org
Services provided: On-site dental care (appointment needed). Accepts
Medicare and select Medicaid plans accepted.
Detroit Community Health Connection
Address:
(Eastside) 7900 Kercheval, Detroit, MI 48215
(Bruce Douglass) 6550 W. Warren, Detroit, MI 48210
(Nolan Center) 111 W. Seven Mile Rd, Detroit, MI 48203
(313) 921-5500
(313) 897-7700
(313) 369-2600
Website: http://www.dchcquality.org/index-6.html
Services provided: Comprehensive and emergency medical, OB/GYN, and
pediatrician services.
Detroit Community Health Connection
Woodward Corridor
Phone: (313) 832-6300
Address: 611 Martin Luther King Jr. Blvd., Detroit, MI 48201
Website: www.dchcquality.org/index-6.html
Services provided: Comprehensive and emergency medical, OB/GYN, and

pediatrician services.
18
beenDetroitWayne
Detroit Health Department-Herman Kiefer Family

Health Center, Pediatric Dental Clinic
Phone: (313) 876-4239
Address: 1151 Taylor, Detroit, MI 48202
Services provided: Cleanings, exams, X-rays, fillings, extractions. No root
canals or braces. Must be a Detroit resident up to 18 years old (if high-school
age must be enrolled in regular high school). Monday - Friday, 6 a.m. - 2:30
p.m. Appointments only.
Detroit Health Department-Herman Kiefer Family
Health Center, Adult Dental Clinic
Phone: (313) 876-4164
Address: 1151 Taylor, Detroit, MI 48202
Services provided: Cleanings, fillings, exams, X-rays, root canals, and oral
surgery. Typically two-week wait for new patients. Must reside in Detroit.
Monday and Tuesday 9 a.m. - 5 p.m. Appointments for preventive only.
Wednesday 7 a.m. walk-in for oral surgery only.
Detroit Hope Hospital
Phone: (313) 874-0100
Address: 801 Virginia Park, Detroit, MI 48202
Website: www.detroithope.com
Services provided: Affordable community-based hospital, including

emergency and dental care services. Free transportation is available.
Community
Resources
19
are those of
the author(s)
PREGNANCY & WOMENS HEALTH SERVICES
Pregnancy Aid
Phone: (313) 882-1000
Address: 17235 Mack Ave, Detroit, MI 48224
Website: www.pregnancyaid.com
Services provided: Pregnancy testing, counseling, baby clothes, and supplies.

Planned Parenthood-Detroit Health Center
Phone: (313) 831-7776
Address: 4229 Cass Ave., Detroit, MI 48201
Website: www.plannedparenthood.org/health-center/centerDetails.
asp?f=2890&a=90630&v=details
Services provided: Pregnancy testing, options counseling, emergency

contraception, STD screening, HIV testing, and medical exams.
Womens Center Michigan
Phone: (313) 526-3600
Address: 15650 East 8 Mile Rd., Detroit, MI 48205
Website: www.abortiononline.com
Services provided: OB/GYN medical services, abortion services, STD testing,

birth control, pap smears.
20
beenDetroitWayne
HIV TESTING SERVICES
AIDS Partnership Michigan: Information and Referral Line

Phone (313) 446-9800
Address: 2751 East Jefferson, Ste 301, Detroit, MI 48207
Website: www.aidspartnership.org
Services provided: Free HIV testing
Community Health Awareness Group

Phone: (313) 963-3434
Website: www.chagdetroit.org
Services provided: Free HIV testing andsupport groups
DISABILITY SERVICES
Disability Network Wayne County Detroit Center

Phone: (313)-923-1655
Address: 5555 Connors, Detroit, MI 48213
Services provided: Provide referrals for individuals with disability regarding
housing availability.
Community
Resources
21
are those of
the author(s)
LGBT SERVICES
Kick
Phone: (313) 285-9733
Address: 41 Burroughs Ste 109, Detroit, MI 48202
Website: http://e-kick.org
Services provided: LGBTQeducation and advocacy, mental health groups for
members of the LGBTQcommunity.
Ruth Ellis Center
Phone: (313) 252-1950
Address: 77 Victor Street, Highland Park, MI 48203
Website: www.ruthelliscenter.org
Services provided: Residential and drop-in programs for LGBTQ youth,

street outreach, and drop-in center for LGBTQ youth.
22
beenDetroitWayne
CLOTHING & FURNITURE ASSISTANCE
Capuchin Soup Kitchens: Emergency Goods,

Clothing, and Household Furnishings
Phone: (313) 925-0514
Address: 6333 Medbury St., Detroit, MI 48211
Website: www.cskdetroit.org/services_services.cfm
Services provided:Emergency food, clothing, household furnishings, and
substance abuse. Referrals required to receive furniture and appliances from
shelter.Must have at least one child under the age of 18. Must bring Social
Security card, photo identification and lease or rental application, proof of
income if available, and birth certificate if available.
St. Dominic Outreach Center
Phone: (313)831-6070
Address: 4835 Lincoln, Detroit, MI 48208
Website: www.stdominicoutreach.org/Pages/default.aspx
Services provided: Furniture, transportation cash, food pantry; must live in the
west side of Detroit, must provide proof of income, and residency if available.
Habitat for Humanity Detroit Restore
Phone: (313) 653-4890
Address: 17181 Mack Avenue, Detroit, MI 48224
Website: www.metrorestores.org/locations/detroiteast
Services provided: Gently used furniture at bargain prices.
Community
Resources
23
are those of
the author(s)
FOOD ASSISTANCE
All Saints Parrish

Phone: (313) 841-1428
Website: www.parishesonline.com/scripts/hostedsites/Org.asp?ID=15840
Services provided: Meals served Monday, Wednesday, & Thursday 11:00am12:30pm
Capuchin Soup Kitchens
Address: 4390 Conner, Detroit, MI 48215
(313) 822-8606
Mon-Sat Breakfast (8:30 9:30,) Lunch (11:00- 1:00), Dinner (4:00-6:00)
Address: 1264 Meldrum, Detroit, MI 48207
(313) 579-2100
Mon-Fri., Breakfast (8:30-9:30am), Lunch (11:00am-1:00pm)
Website: www.cskdetroit.org/services_services.cfm
Services provided: Provides hot meals to anyone.
CASS Community Social Services: Food Services
Phone: (313) 883-2277
Address: 11850 Woodrow Wilson St., Detroit, MI 48206
Website: www.casscommunity.org
Services provided: Homeless services, emergency food boxes, eviction
prevention, utility assistance, Saturday community lunch, free health clinic,
clothing, appliances and furniture for homeless, emergency homeless shelter
for women and children, rotating shelter for women and men, skill building
program, lunch daily Mon- Sun, 12:00-1:15.
24
beenDetroitWayne
Crossroads of Michigan
Address:
Main Office, 2424 W. Grand Blvd., Detroit, Mi 48208
Crossroads East Office, 21230 Moross, Detroit, MI 48215
(313) 831-2787
(313)822-4441
Website: www.crossroadsofmichigan.org/index.html
Services provided: Food pantry and Sunday afternoon soup kitchen12-3pm.
Focus Hope
Phone: (313) 494-4600
Address:
6353 W. Vernor Hwy, Detroit, MI 48209
1300 Oakman Blvd., Detroit, MI 48238
9151 Chalmers, Detroit, MI 48213
Website: www.saveourchildrencoalition.org/basic-needs/focus-hope.html
Services provided: Distributes supplemental food items such as infant
formula, milk, meat, and vegetables monthly. Eligibility: Pregnant women,
parents with children under 6 years old, or senior citizens. Must bring photo
identification and proof of income or local address. Hours: Monday-Thursday
9-5pm, Friday-9-12pm.
PantryNet.org
Website: www.pantrynet.org
Services provided: A website to help locate food pantries within your area.
St. Christine Christian Services

Phone: (313) 535-7272
Address: 15317 Dacosta, Detroit, MI 48233
Website: www.sccsdetroit.org/Contacts.htm
Services provided: Provides hot meals to anyone. Soup kitchen-Tues: 2:304:30pm, Saturday: 12-2pm. Food pantry-Tues: Seniors & Individuals with
disabilities-1:30-2:30pm, Able-bodied: 2:30pm-4:30pm
Community
Resources
25
are those of
the author(s)
St. Leos Church Soup Kitchen

Phone: (313) 897-6565
Address: 4860 15th Street, Detroit, MI 48208
Services provided: Provides hot meals from the soup kitchen M-Sat.
11:30am-12:30pm.
PeaceMakers International
Phone: (313) 923-5939
Address: 5322 Chene St., Detroit, MI 48211
Website: www.peacemakersinternational.org
Services provided: Hot lunch on Mon, Tues, Fri @ 12:00 1:00, Breakfast on
Wed @ 9:00- 12:00, Box Lunch on Tues (2:00) and Friday (11:00), have to
attend services in order to receive box lunch, clothing services on Tuesday at
2:30pm
St. Vincent de Paul Society: Detroit
Phone: (313) 393-2930 or (877) 788-4623
Address: 3000 Gratiot, Detroit, MI
Website: www.svdpdet.org/vec.cfm
Services provided: Emergency food assistance (10-2pm), thrift store

(weekdays 9-5pm) and furniture. Referrals needed. People in need of food
assistance can contact a participating Catholic church. This agency will
complete a referral form that can be taken to a community food depot. Call
for participating locations.
Wayne County Office Department of Human Services
Address: Go online for application
Website: https://www.mibridges.michigan.gov/access
Services provided: Application for MI Bridge card.
26
beenDetroitWayne
HOUSING ASSISTANCE
NETWORK
Community Housing Network
Phone: (866) 282-3119
Website: www.communityhousingnetwork.org/activek/content.
asp?catid=3&tid=1
Services provided: Help individuals who are homeless, have a disability, or

need help finding a secure housing.
Tumaini Neighborhood Services Organization
(NSO) Homeless Services
Phone: (313)832-3100
Address: 3430 Third Avenue, Detroit, MI 48201
Website: www.nso-mi.org/services.php
Services provided: Referral to affordable housing options for those individuals

with income or homeless.
Detroit Housing Commission
Phone: (313)877-8000
Address: 1301 East Jefferson, Detroit, MI 48207
Website: www.dhcmi.org
Services provided: Assist individuals with low-income housing for affordable

housing options.
Community
Resources
27
are those of
the author(s)
Detroit Non Profit Housing Corporation

Phone: (313)972-1111
Address: 8904 Woodward Ave., Detroit, MI 48202
Services provided: Monthly education classes on home buying. Helps
individuals through classes that prevent foreclosure.
TRANSITIONAL HOUSING
Bethlehem House
Phone: (313)923-6435
Address: 5603 Van Dyke, Detroit, MI 48213-2856
Services provided: Homeless shelter for females, money management,
life skills training, individual and group counseling, legal referrals, medical
referrals, laundry facilities, individual meal preparation, no children, working
fulltime, preferable over age 40
Lakewood Manor
Phone: (313) 821-0469
Address: 14200 Kercheval Street, Detroit, MI 48215-2848
Services provided: Transitional housing for mother and kids in Detroit area.
Salvation Army Denby Center for Children and Family Services
Phone: (313) 537-2130
Address: 20775 Pembroke Ave, Detroit, MI 48219-1345
Website: www.usc.salvationarmy.org/DenbyCenter
Services provided: Residential shelter for families, single females, and

veterans.
28
beenDetroitWayne
TEEN HOUSING
Alternative For Girls
Phone: (313) 361-4000
Address: 903 West Grand Blvd, Detroit, MI 48208
Website: www.alternativesforgirls.org
Services provided: Shelter, independent living program, prevention services

for girls 15-21 years of age.
Covenant House Michigan
Phone: (313) 463-2000
Address: 2959 Martin Luther King Jr. Blvd, Detroit, MI 48208-2475
Website: www.covenanthousemi.org
Services provided: Homeless shelter for youth between the ages of 18 and
22, men and women, no expectant mothers or children, GED programs,
alternative high school, career advancement program, free individual and
group counseling, identification necessary.
Detroit Rescue Mission Ministries: Teen Mom House I
Phone: (313) 993-6692
Address 3840 Fairview., Detroit, MI 48214
Website: www.drmm.org
Services provided: Shelter stay for 90 days, daycare services, GED assistance,
job search, support groups, group substance abuse treatment programs and
referrals. Serves teen mothers only.
Matrix Human Services: Off the Streets
Phone: (313) 873-0678
Address: 680 Virginia Park, Detroit, MI 48202
Website: www.matrixhs.org
Services provided: Twenty-one day temporary shelter, individual and group

counseling, prevention education for youth ages 12-17 years old.
Community
Resources
29
are those of
the author(s)
JOB TRAINING PROGRAMS
Crossroads of Michigan
Address:
Main Office, 2424 W. Grand Blvd., Detroit, MI 48208
Website: www.crossroadsofmichigan.org/index.html
(313) 831-2000
Services provided: Set up appointments for help creating and editing resumes.
Focus Hope
Phone: (313) 494-5500
Address: 1400 Oakman Blvd, Detroit, MI 48238
Website: www.focushope.edu
Services provided: Education and job training in technology, drop in
program for school age children, and food supplement program.
Urban League of Detroit
Phone: (313) 832-4600
Address: 208 Mack Ave, Detroit, MI 48201
Website: www.detroiturbanleague.org/urban_league_of_southeastern_
michigan_004.htm
Services provided: Provides screening, employability skills training, job

referrals and placement. Also, conducts an annual career/job fair.
Wayne County-City of Detroit Michigan Works! Service

Centers Detroits One Stop Service Center
Address: 707 West Milwaukee, 1st Floor, Detroit, MI 48202 (313) 873-7321
Address: 455 West Fort St. 1st floor, Detroit, MI 48226
(313) 962-9675
Address: Samaritan Center 5555 Conner St, Detroit, MI 48213
30
beenDetroitWayne
(313) 579-4925
Address: 9301 Michigan Avenue, Detroit, MI 48209
(313) 846-0814
Monday thru Friday 8:00am-5:00pm; Thursdays 8:00am-7:00pm

Services provided: Employment services such as workshops geared around
employability skills and help with resumes.
Community
Resources
31
are those of
the author(s)
WA Y N E C O U N T Y P R O S E C U T O R S O F F I C E
1 4 4 1 S T. A N T O I N E , R O O M 1 2 2 0
D E T R O I T, M I 4 8 2 2 6
313-224-7270
450
APPENDIX D: Data Collection Instruments
AppendixD1:
AppendixD2:
AppendixD3:
AppendixD4:
AppendixD5:
AppendixD6:
AppendixD7:
AppendixD8:
AppendixD9:
InterviewProtocolforDetroitStakeholders(FirstInterview)
InterviewProtocolforDetroitStakeholders(SecondInterview)
InterviewProtocolforNationalStakeholdersfromCriminalJustice/Forensic
ScienceandViolenceAgainstWomenSocialServiceOrganizations
QuestionsAskedofPublicOfficialsinComparableCities
PoliceReportCodingSheetsforVictim,Assailant,andCaseCharacteristics
ForensicTestingOutcomesCodingSheets
LawEnforcementTrackingSheetsforVictimNotifications
CommunityBasedAdvocatesTrackingSheetsforVictimNotifications
FocusGroupProtocolforDetroitStakeholders
451
APPENDIXD1:InterviewProtocolforDetroitStakeholders(FirstInterview)
DetroitSAKActionResearchProject
DetroitStakeholdersInterviewProtocol
FirstInterviewforLongitudinalInterviews;OnlyInterviewforCrossSectionalInterviews
INTRODUCTIONANDOVERVIEW
Thankyousomuchformeetingwithmetodayandfortakingtimeoutofyourbusyschedule.Asyouaremay
alreadyknow,oneoftheevaluationgoalsoftheDetroitSAKActionResearchProjectistointerviewthecore
membersofthisprojecttogainabetterunderstandingofwhyandhowsomanysexualassaultkitsremained
untestedinDetroit.Therefore,itisextremelyhelpfultohaveyoushareyourexpertiseregardingthisissue.
Specifically,Iwillbeaskingyoutoshareyourperceptionastothepossibleunderlyingcausesofthisproblem,as
wellashowyourorganizationhandlessexualassaultcasesandsexualassaultkits.
Thisinterviewshouldtakebetween12hourstocomplete.Ifitisokaywithyou,Iwouldliketorecordthis
interviewformyownuseincodingtheinterviewlateronitsgoingtobehardformetogeteverythingdownon
paper,sotherecordercanhelpmelateronfillinginanythingImighthavemissed.Theonlypeoplewhomight
listentothisrecordingwillbethetwoevaluatorsonthisproject.Whenwehavecompletedcodingtheinterview,
therecordingwillbedestroyed.
Everythingwediscusstodayisprivateyournamewillnotbeconnectedtoanythingyousay.Whatwediscuss
duringthisinterviewwillnotbesharedwithothercoremembersoftheDetroitSAKActionResearchProjector
withindividualsinyourorganizationorotherorganizations.Yournamewillnotbeplacedonthisintervieworthe
tape.Asweregoingthroughtheinterview,ifyouneedtotakeabreakorstop,justletmeknow.Ifthereareany
questionsthatyoudontwanttoanswer,justsayso,andIwillmoveontothenextsection.Youdonothaveto
answerallofthequestionsinthisinterview.
BeforewegetstartedIneedtogothroughtheprocedurestoobtainyourconsenttobeinterviewed(gothrough
procedurestoobtaininformedconsent).
Doyouhaveanyquestionsbeforewestart?
452
SECTION1:BACKGROUNDINFORMATION
Togetstarted,Iwouldliketoaskyousomequestionsaboutyourcurrentpositioninyouragency/organization.
1. Howlonghaveyouworkedin(lawenforcement,law,sexualassaultservices,crimelab)?
2. Whatareyourprimaryresponsibilitiesofthisposition?
3. Howlonghaveyoubeeninthisposition?
4. Towhatextenthaveyouworkedwithsexualassaultcasesand/orvictims?Whatisyourinvolvementin
sexualassaultcasesand/orsexualassaultvictims?
SECTION2:AGENCY/ORGANIZATIONROLEINPROCESSINGSAKs
Inthisnextsection,Iwouldliketoaskyouaboutyouragencysroleinprocessingsexualassaultkits.Iwantyouto
thinkaboutyourorganizationasawhole(policies,procedures),notabouttheindividualswhomakeupyour
organization.
1.
2.
3.
Thinkingofyourparticularorganization,whatpoliciesandproceduresareinplacetoprocesssexual
assaultcases?
Havethesepolicieschangedoverthepast10years?How?
Whatpoliciesandproceduresareinplacetoprocesssexualassaultkits?Whatarethestepstakento
processSAKs?
4.
Whichproceduresarethemosteffectiveinmakingcertainthatsexualassaultkitsareprocessed
appropriately?Whichproceduresarethemostineffective?
5. Isthereaperson/orgroupwhoisinchargeofoverseeingsuchpoliciesandprocedures?
6. Whatresources(i.e.,personnel,equipment,etc)areavailabletohelpmaketheseprocedureswork?
7. Arethereotherresources(thatarenotpartofcurrentprocedures)thatwouldbeusefulinmakingsexual
assaultkitsprocessmoreeffectively?
8. Canyouidentifyanygapsinthecurrentpolicies/proceduresthatmayhavecontributedtothelarge
quantitiesofuntestedsexualassaultkits?
9. Howdoesyourorganizationwork/collaboratewithotheragenciesinprocessingsexualassaultcases/kits?
Howwelldoyoudoyouthinkyourorganizationiscollaboratingwithotheragenciestomakethisprocess
work?
10. Whatarethenormsandexpectationsregardingtheimportanceofsexualassaultcasesinyour
organization?Howdoestheprocessingofsexualassaultcasescomparetoothercrimes?
453
SECTION3:INDIVIDUALSROLESRE:PROCESSINGSEXUALASSAULTCASES/KITS
Thisnextsectionisaboutyourperceptionofhowindividualswithinyourorganizationprocessandhandlesexual
assaultcasesandkits.Herewewillbediscussingindividualsskillsandabilitiestoperformtheirrespectiveroles
inprocessingSAKs.
1.
Howmanypeoplearetypicallyinvolvedinhandlingsexualassaultcasesinyourorganization?Dopeople
havespecificrolesthattheyadhereto?Whatarethey?
2. Whattypeoftrainingisinvolvedforthepersonnelinchargeofprocessingsexualassaultcases?
3. Dotheseindividualshavesupervisors?Doyouthinkthereisagapinhowtheseindividualswere/are
monitored/supervisedthatmayhavecontributedtountestedkits?
4. Wouldyousaytheseindividualsareknowledgeableaboutsexualassaultinparticular?(i.e.,theimpactof
sexualassaultonvictims?Currentstatsonactualsexualassaultincidents?Prevalence?Howsexual
assaultcasesareprosecuted?,etc.)
5. Whatwouldyousayisthecurrentsocialclimatesurroundingsexualassaultcases?Whataretheattitudes
andbeliefsthatindividualsinyourorganizationholdaboutsexualassaultvictims?Doyouthinkthese
beliefshavechangedovertime?Ifso,how?
6. Doyouthinkthattheattitudesheldbytheindividualsinyourorganizationcontributedtothewaysexual
assaultcases/kitswerehandled?Howso?
SECTION4:CONCLUDINGTHOUGHTS
Thankyouforsharingwithmeyourthoughtsandconcernsabouttheprocessingofsexualassaultcases/kitsin
yourorganization.Thisfinalsectionwillaskyoutoshareyouroverallfeelingsaboutthecurrentproblem.
1.
2.
3.
4.
Overall,whatwouldsayhascontributedthemosttotheproblemofuntestedsexualassaultkitsoverthe
lastseveraldecades?
Whatarethemajorstrengthsofyourorganizationinhandlingsexualassaultcases/kits?Whatarethe
majorweaknesses?
Whatarethemajorstrengthsoftheindividualswithinyourorganizationinhandlingsexualassault
cases/kits?Whatarethemajorweaknesses?
Whatdoyouforeseetobethemostdifficulttaskinrespondingtothisproblem?Bothwithinyour
organizationandacrossotherresponsibleagencies/organizations?
454
APPENDIXD2:InterviewProtocolforDetroitStakeholders(SecondInterview)
DetroitStakeholdersInterviewProtocol
Second/ThirdInterviewforLongitudinalInterviews
Thankyousomuchfortalkingwithmetodayandfortakingtimeoutofyourbusyschedule.DuringPhase1,oneof
theevaluationgoalsoftheDetroitSAKActionResearchProjectwastointerviewthecorememberstogainabetter
understandingofwhyandhowsomanysexualassaultkitsremaineduntestedinDetroit.
NowthatweareinPhase2,wearecontinuingqualitativeanalysesontheproblemaswellasgettingabetter
understandingofhowthisprojecthasimpactedtheworkofcore/collaborativepartners.Sinceyouhaveserveda
significantroleinthisprojectduringthepastyear,wewouldliketohearyourthoughts.Specifically,Iwillbe
askingyouwhatyourcurrentviewoftheproblemofuntestedkitsis,howparticipatinginthisprojecthas
impacted/changedyourviewsaboutsexualassault,howithasimpactedtheworkyoudooutsidetheproject,and
recommendationsfortheprojectaswemoveforward.
Thisinterviewshouldtakeabout30minutestocomplete.Ifitisokaywithyou,Iwouldliketorecordthisinterview
formyownuseincodingtheinterviewlateronitsgoingtobehardformetogeteverythingdownonpaper,so
therecordercanhelpmelateronfillinginanythingImighthavemissed.Theonlypeoplewhomightlistentothis
recordingwillbethetwoevaluatorsonthisproject.Whenwehavecompletedcodingtheinterview,therecording
willbedestroyed.
Everythingwediscusstodayisprivateyournamewillnotbeconnectedtoanythingyousay.Whatwediscuss
duringthisinterviewwillnotbesharedwithothercoremembersoftheDetroitSAKActionResearchProjector
withindividualsinyourorganizationorotherorganizations.Yournamewillnotbeplacedonthisintervieworthe
tape.Asweregoingthroughtheinterview,ifyouneedtotakeabreakorstop,justletmeknow.Ifthereareany
questionsthatyoudontwanttoanswer,justsayso,andIwillmoveontothenextsection.Youdonothaveto
answerallofthequestionsinthisinterview.
455
Iwouldliketostartoffbyaskingyouafewquestionsaboutyourroleinthisproject.
1. WhatisyourroleintheSAKActionResearchProject?Whatareyourresponsibilities?
SECTION2:THEPROBLEMOFUNTESTEDSAKS
NowIwouldliketotalkalittleabouttheproblemofuntestedkitsinDetroit.
2. Thinkingaboutthispastyearandwhatyoulearnedinthisproject,whatwouldyousayhascontributedto
thepileofuntested/unsubmittedSAKsinDetroit?
3. DoyouthinkthattheSAKprojectisbeginningtoaddressanyoftheseissues?Howso?
SECTION3:THEIMPACTOFNIJSAKACTIONRESEARCHPROJECT
Next,Iwouldliketotalkabouthowbeingpartofthisprojecthasimpactedyourworkandyouroverallthoughts
abouthowthisprojectisgoing.
4.
5.
6.
Howhasparticipatinginthisprojectimpactedyourworkoutsidetheproject?(probe:Hasyour
participationaffectedthewayyoudoworkregularly?Howso?)
Howhaveyoubeenmanaging/balancingtheworkyoudointheprojectandyourregularongoingwork?
Overall,howdoyouthinktheSAKprojectisgoing?Whatwerethebiggestchallenges?
7.
Howdoyouthinkthedifferentgroups/corepartnersareworkingtogether?Isthiswhatyouexpected
whenworkingwithamultidisciplinarycollaborativegroup?
SECTION4:CONCLUDINGTHOUGHTSANDRECOMMENDATIONS
Thankyou.Finally,Iwouldliketogiveyouthechancetoshareanyfinalthoughtsandproviderecommendationsto
thegroup.
8.
Knowingwhatyouknownow,aretherethingsyouthinkthegroupshouldhavedonedifferentlyto
respondtotheproblem?
9.
Doyouhaveanyrecommendationsforhowtheteamproceedswithtestingkitsandvictimnotification?
10. Asyouknow,whatisdoneinthisprojectwillhavenationalimpactacrossotherjurisdictionsgoing
throughthesameproblem.Doyouhaveanyadviceforhowotherscitiesrespondtotheproblemof
untestedkits?
456
APPENDIXD3:InterviewProtocolforNationalStakeholders
NationalStakeholdersInterviewProtocol
(FirstInterviewforLongitudinalInterviews;OnlyInterviewforCrossSectionalInterviews)
Thankyousomuchformeetingwithmetodayandfortakingtimeoutofyourbusyschedule.Asyoumayalready
know,theNationalInstituteofJusticefundedtwoActionResearchProjectsincitieswithlargenumbersof
untestedsexualassaultkits(SAKs).WearetheresearchersfromoneofthoseprojectstheDetroitsite.Aspartof
ourworkonthisproject,wewantedtoreachouttonationalstakeholderstocollecttheirthoughtsaboutthe
complexitiesofSAKtesting.Specifically,IwillbeaskingyoutoshareyourperceptionsregardingtheutilityofSAK
testingtotheworkofthecriminaljusticesystemandtovictimsandyourthoughtsaboutbestpracticesfor
victimnotification.
Thisinterviewshouldtakeapproximately3045minutestocomplete.Ifitisokaywithyou,Iwillbetakingnotesof
ourdiscussion.Everythingwediscusstodayisprivateyournamewillnotbeconnectedtoanythingyousay.What
wediscussduringthisinterviewwillnotbesharedwiththemembersoftheDetroitSAKActionResearchProjector
withindividualsinyourorganizationorotherorganizations.Yournamewillnotbeplacedonthisinterview.As
weregoingthroughtheinterview,ifyouneedtotakeabreakorstop,justletmeknow.Ifthereareanyquestions
thatyoudontwanttoanswer,justsayso,andIwillmoveontothenextsection.Youdonothavetoanswerallof
thequestionsinthisinterview.
BeforewegetstartedIneedtogothroughtheprocedurestoobtainyourconsenttobeinterviewed(gothrough
procedurestoobtaininformedconsent).
457
Togetstarted,Iwouldliketoaskyousomequestionsaboutyourcurrentpositioninyouragency/organization.
1. Howlonghaveyouworkedin(lawenforcement,law,sexualassaultservices,crimelab)?
2. Whatareyourprimaryresponsibilitiesofthisposition?
3. Howlonghaveyoubeeninthisposition?
4. Towhatextenthaveyouworkedwithsexualassaultcasesand/orvictims?Whatisyourinvolvementin
sexualassaultcasesand/orsexualassaultvictims?
SECTION2:PERCEPTIONSABOUTSAKTESTINGANDVICTIMNOTIFICATION
Inthisnextsection,IwouldliketoaskyouaboutyourthoughtsregardingtheutilityofSAKtestingandbest
practicesforvictimnotification
1.
2.
3.
4.
CanyoudescribehowSAKtestingcanbeusefultotheworkof...(lawenforcement,prosecutors,
forensicscientists)?
HowdoyouthinkSAKtestingcanbehelpfultovictims/survivors?
When/whydoyouthinkSAKsshouldbetested?
Whatareyourthoughtsabouttestingallvs.someunsubmittedSAKsinjurisdictionswithlargenumberof
unprocessedSAKs?
5.
6.
7.
Inyourprofessionalopinion/experience,how/whenshouldvictimsbenotifiedaboutSAKtesting?
Whatrecommendationsdoyouhaveforcreatingvictimnotificationprotocols?
Whatarekeyconfidentiality,privacy,andsafetyconcernsthatoughttobeattendedtowhencreating
victimnotificationprotocols?
SECTION3:CONCLUDINGTHOUGHTS
Thankyouforsharingwithmeyourthoughtsandconcernsabouttheprocessingofsexualassaultcases/kitsinyour
organization.Thisfinalsectionwillaskyoutoshareyouroverallfeelingsaboutthecurrentproblem.
1.
2.
Overall,whatwouldsayhascontributedthemosttotheproblemofuntestedsexualassaultkitsoverthe
lastseveraldecades?
WhatadvicedoyouhaveforjurisdictionsstrugglingwithlargenumbersofunsubmittedSAKs?
458
APPENDIXD4:QuestionsAskedofPublicOfficialsinComparableCities
QuestionsRegardingDNATestingandCODIS
1.
Howmanyforensiclabsserve________________(nameofcity)?
2.
Whatyeardidthelab(s)beginconductingDNASTRtesting?
3.
Whendidthelab(s)becomeaccreditedforCODIS?(andbywhatmeans?)
4.
HowmanyDNAscientists(benchscientists)didthelabhaveperyearforDNAtesting(focusyears:
19902009)?(i.e.,numberoffulltimescientistsavailableforcaseworkperyear)
QuestionsRegardingProsecutionResources
1.
Howmanytrialattorneysperyear(19902009,ifavailable)
2.
Doestheprosecutorsofficehaveaspecializedprosecutionunitforsexualassault?Ifso,describe;ifso,
yearinwhichitwasestablished.
QuestionsRegardingMedical/SANEResources
1.
Does____________(nameofcity)haveaSANE/SAFEprogram?Ifso,yearitwasestablished.
2.
WhatistheaveragenumberofSANE/SAFEmedicalprovidersintheprogram(frombeginningofprogram
to2009)?
QuestionsRegardingVictimAdvocacy
1.
Doesthe_________(nameofcity)policedepartmenthaveavictimadvocacyprogram?Ifso,describe;if
so,howmanystaffareemployed(vs.volunteer).
1.
Doesthe_________(nameofcity)haveanonprofitsexualassaultvictimadvocacyprogram/rapecrisis
center?Ifso,howmanystaffareemployed(vs.volunteer).
459
AppendixD5:PoliceReportCodingSheetsforVictim,Assailant,andCaseCharacteristics
Dateofassault
Dateofexam
Howlongagoassaultoccurred
(inyears,fromDecember31,
2013,asreferencedate)
TimeBetweenOffenseand
Exam(Days)
VictimGender
VictimRace
VictimAge
AssailantGender
AssailantRace
AssailantAge
Openentryfield
888=missing(becauseentirefilemissing)
999=missing(becauseinfonotinfile)
Openentryfield
Openentryfield
Openentryfield
0=Female
1=Male
0=AfricanAmerican
1=ArabAmerican/Chaldean
2=AsianAmerican/PacificIslander
2=Caucasian
3Hispanic/Latina
4=Multiracial
Openentryfield
0=Female
1=Male
0=AfricanAmerican
1=ArabAmerican/Chaldean
2=AsianAmerican/PacificIslander
3=Caucasian
4=Hispanic/Latina
5=Multiracial
Openentryfield
(referencevariable)
(referencevariable)
CASECHARACTERISTIC
CASECHARACTERISTIC
VICTIMCHARACTERISTIC
ASSAILANTCHARACTERISTIC
460
VictimOffenderRelationship
AssaultInvolvedMultiple
Perpetrators
Alcoholand/orDrugsInvolved
intheAssault(eitherforthe
victimortheassailant)
WeaponUsedintheAssault
PhysicalForceUsedinthe
Assault
1=Stranger
2=BySight/Nickname
3=Friend/Associate/FamilyMember
4=Current/PastIntimatePartner
0=No
1=Yes
0=No
1=Yes
0=No
1=Yes
0=No
1=Yes
CASECHARACTERISTIC
CASECHARACTERISTIC
CASECHARACTERISTIC
CASECHARACTERISTIC
CASECHARACTERISTIC
461
AppendixD6:ForensicTestingOutcomesCodingSheets
SAKcontainedbiologicalevidence
SAKhadsample(s)thatpassedserologyscreeningand
proceededtoDNAtesting
DNAProfileEnteredintoCODIS
CODISHit(ofanykind)
CODISHitOffenderHit
(offenderidentified)
IfOffenderHit,wasCODISqualifyingoffenseasexual
assault
CODISHitForensicHit
(hittoanothercase,offenderunidentifiedinboth)
IfForensicHit,wasCODISqualifyingcrimescene
evidencefromasexualassault
CODISHitOffenderForensicHit
(hittoanothercaseandoffenderidentified)
IfOffenderForensicHit,wasCODISqualifyingoffense
and/orcrimesceneevidencefromasexualassault
IfCODISHit(ofanykind),casetocaseassociationto
anotherNIJSAK?
0=no
1=yes
0=no
1=yes
0=no
1=yes
0=no
1=yes
0=no
1=yes
0=no
1=yes
0=no
1=yes
0=no
1=yes
0=no
1=yes
0=no
1=yes
0=no
1=yes
Ifyes,provideIDnumberofwhichcase(s)hittoeachother
462
AppendixD7:LawEnforcementTrackingSheetsforVictimNotifications
LOCATETracking(trackingallactionstakentolocatethevictim)
ALLINFOCOLLECTEDFOREACHACTIONTAKENUNTILVICTIMLOCATED.
DateofAction
Openentryfield
TimeofAction
Openentryfield
NameofInvestigatorCompletingAction
Dropdownfield(optionsincludenamesof
potentialinvestigators)
TypeofAction
Dropdownfield(optionsincludephone,email,
databasesearch,inperson,letter)
NameofDatabaseSearched(ifactionisdatabasesearch)
Dropdownfield(optionsincludeCLEMIS,CRIM,
CRISNET,LEIN,LEXUS,ODYSSEY,OTHER,TLO)
RecipientofAction(ifactionisnotdatabasesearch,recipientofemail,
letter,phonecall,orinpersoncontact)
Openentryfield
OutcomeType
Dropdownfield(optionsincludelead,negative)
OutcomeNarrative
Openentryfield
NextStep
Openentryfield
NOTIFICATIONTRACKING(trackingallVNinteractionswiththevictim)
ALLINFOCOLLECTEDFOREACHVNINTERACTIONWITHTHEVICTIM
DateofAction
Openentryfield
TimeofAction
Openentryfield
NameofInvestigatorNotifyingVictim
Dropdownfield(optionsincludenamesof
potentialinvestigators)
Location
Openentryfield
OthersPresentforVictimNotification
Openentryfield
NarrativeofInteraction
Openentryfield
Whatinformationwasprovidedtothevictimandinwhatform?
Openentryfield
Howdidthevictimreactphysically?
Openentryfield
Whatwasthevictimsemotionaldemeanor?Diditchangeoverthecourse
oftheinteraction?
Openentryfield
Whatquestionsdidthevictimhaveandweretheyanswered?
Openentryfield
Whatarethenextstepsforthecase?
Openentryfield
Howdidthenotifierfeelaftertheinteraction?
Openentryfield
Whatdidand/ordidnotwork,andwhatdoyourecommendforfuture
notifications?
Openentryfield
463
AppendixD8:CommunityBasedAdvocateTrackingSheetsforVictimNotifications
NOTIFICATIONTRACKINGAdvocateDatabase(trackingadvocateandinvestigatorinteractionwiththevictim)
ALLINFOCOLLECTEDFORVNMEETINGWITHADVOCATE,INVESTIGATORS,ANDVICTIM
Whatinformationwasprovidedtothevictimandinwhatform?
Openentryfield
Howdidthevictimreactphysically?
Openentryfield
Whatwasthevictimsemotionaldemeanor?Diditchangeovertime?
Openentryfield
Whatquestionsdidthevictimhaveandweretheyanswered?
Openentryfield
Howdidtheadvocatefeelaftertheinteraction?
Openentryfield
Wouldyoumakeanyrecommendationstothenotifier/investigatoron
whattheycoulddodifferently?
Openentryfield
AdditionalNotes
Openentryfield
464
APPENDIXD9:FocusGroupProtocolforDetroitStakeholders
DetroitStakeholdersEndofProjectFocusGroups
Priortothismeeting,everyonereceivedaDRAFTofouranalysesregardingkeylessonslearnedthroughoutthis
project.Today,wewillbediscussingTable(INSERTNAMEcensus,testing,victimnotification,overallproject).
Wehavetheremaining(90or30)minutesintodaysmeetingtodiscussthesefindingsandanyotherinformation
youdliketoprovideregardinglessonslearned.
Iappreciatethatthisgroupisusedtoworkingtogether,talkingandsharingideasregularly,evenarguing
respectfullywitheachother.Today,forthisconversation,Iwantustobemindfulthateveryoneneedsachanceto
speakandbeheard.Pleasemonitoryourownparticipationinthediscussionsothateveryonehasachancetotalk
andeveryonefeelscomfortabletalking.PleasenotethatImaybeaskingyou(verballyorthroughgestures)to
holdbackonyourfeedbacktogiveothersachancetotalkorasking/encouragingyoutoweighin,ifyouhavenot
donesoalready.Aswithallotherprojectactivities,everythingwediscusstodayisprivateyournamewillnotbe
connectedtoanythingyousay.
QUESTIONS
1.
Letsbeginwithoverallreactions.Whatdoyouthinkarethemostimportantlessonslearnedinthe
projectwithrespectto(TOPIC)(conductingthecensus,developingtestingplan,creatingvictim
notificationprotocols,overall)
2.
Lookingback,whatdoyouthinkwedidrightwithrespectto(TOPIC)
3.
Lookingback,whatdoyouseearethemostimportantcoulda,woulda,shouldasregarding(TOPIC)
whatshouldwehavedonedifferentlyandwhy?
4.
Letsrevieweachsectionofthetablenowforeachpart,letsopenitupfor
questions/comments/challengestothematerialpresented.

Detroit Sexual Assault Kit Action Research Projectort

Uploaded by

Copyright:

Available Formats

Detroit Sexual Assault Kit Action Research Projectort

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Detroit Sexual Assault Kit Action Research Projectort

Uploaded by

Copyright:

Available Formats

The author(s) shown below used Federal funds provided by the U.S.

Department of Justice and prepared the following final report:

The Detroit Sexual Assault Kit (SAK) Action

Opinions or points of view expressed are those

Rebecca Campbell, Ph.D.

Giannina Fehler-Cabral, Ph.D.

Steven J. Pierce, Ph.D.

Deborah Bybee, Ph.D.

Jessica Shaw, Ph.D.

CHAPTER 2: The Scope of the Problem

CHAPTER 3: Underlying Reasons

CHAPTER 4: Testing Kits

CHAPTER 5: Notifying Victims

APPENDIX C: Victim Notification Resources

APPENDIX D: Data Collection Instruments

CHAPTER 2: The Scope of the Problem

CHAPTER 3: Underlying Reasons

CHAPTER 4: Testing Kits

CHAPTER 5: Notifying Victims

CHAPTER 2: The Scope of the Problem

CHAPTER 3: Underlying Reasons

CHAPTER 4: Testing Kits

CHAPTER 5: Notifying Victims

CHAPTER 3: Underlying Reasons

Unsubmitted Sexual Assault Kits (SAKs): A Growing National Problem

An Action Research Approach to the Problem of Unsubmitted SAKs

One Citys Shocking Discovery: The Problem of Unsubmitted SAKs in Detroit

The Collaborative Team

The Collaborative Partnership Process

A Special Note About Identity, Confidentiality, and Privacy

About This Report

CHAPTER 2: The Scope of the Problem

The Discovery of the Kits and Initiating a Census

CBS News.com, March 4th, 2010

The Process of Conducting the Census in Detroit

The Results of the Census: How Many Unsubmitted SAKs in Detroit

Summary & Conclusions: The Scope of the Problem

CHAPTER 3: Underlying Reasons

Ecological Systems Theory: Understanding Interdependent Organizations

Historical Context: Organizational Resources & Practices 1980 - 2009

History of DNA Testing and CODIS Access 1980 2009

0.242 0.409 <.001

Police Department Resources & Policies 1980-2009

Crime Laboratory Resources 1980-2009

Prosecutors Office Resources 1980-2009

Medical System Resources 1980-2009

0.242 0.409 <.001

Victim Advocacy Resources 1980-2009

The Impact and Legacy of Chronic Resource Depletion

Underlying Factors: Front-Line Practices & Inter-Organizational Communications

Police Department Front-Line Decision Making & Practices

The Police and the Medical System: Minimal Inter-Organizational Communication

The Police and Victim Advocacy Organizations: Intra- & Inter-Organizational

Summary & Conclusions: Why So Many Unsubmitted SAKs in Detroit

CHAPTER 4: Testing Kits

The Process of Developing a SAK Testing Plan

Implementing the Detroit SAK Testing Plan

Defining the Sampling Criteria for SAK Selection

Selecting SAKs for the Testing Groups

Challenges Implementing the Testing Plan and Coordinating Post-Testing Activities