Death Studies

ISSN: 0748-1187 (Print) 1091-7683 (Online) Journal homepage: http://www.tandfonline.com/loi/udst20

Suicide contagion, gender, and suicide attempts

among adolescents

Muhammed Yıldız, Ugur Orak, Mark H. Walker & Ozgur Solakoglu

To cite this article: Muhammed Yıldız, Ugur Orak, Mark H. Walker & Ozgur Solakoglu (2018):
Suicide contagion, gender, and suicide attempts among adolescents, Death Studies, DOI:
10.1080/07481187.2018.1478914

To link to this article: https://doi.org/10.1080/07481187.2018.1478914

Published online: 19 Jun 2018.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

http://www.tandfonline.com/action/journalInformation?journalCode=udst20
DEATH STUDIES
https://doi.org/10.1080/07481187.2018.1478914

Suicide contagion, gender, and suicide attempts among adolescents

Muhammed Yıldıza, Ugur Oraka, Mark H. Walkera, and Ozgur Solakoglub
a
Sociology Department, Louisiana State University, Baton Rouge, LA, USA; bSociology Department, Turkish Military Academy,
Ankara, Turkey

ABSTRACT
Exposure to suicidal behavior (ESB) through social networks is often associated with an
increased risk of suicidality, but empirical research—largely limited to the Western popula-
tions—has produced mixed results. Using survey data, we examined (a) the association
between ESB and suicide attempt, and (b) the exacerbating role of ESB on gender differen-
ces in suicide attempt risk among Turkish adolescents (N ¼ 2035). ESB was significantly asso-
ciated with increased risk of suicide attempt, and it contributed to the higher attempt risk
among girls. Suicidal behaviors spread through social networks via ESB, and suicide conta-
gion exacerbates the existing gender differences.

Suicide is among the leading causes of death world- factors that researchers disagree on. One factor that
wide, claiming around one million premature deaths has sustained debate is exposure to suicidal behaviors
annually (World Health Organization, 2014). Although (ESB) of others on one’s own suicidality, otherwise
suicide rates remained stable for most subpopulations known as suicide contagion (Randall, Nickel, &
over the last 45 years, statistics suggest an increase in Colman, 2015). Recent studies examining the effect of
suicide among youth (Fleischmann & De Leo, 2014). ESB adapted social learning theory of deviance (Akers
Suicide attempts are the strongest precursors of subse- & Jennings, 2009) to the case of suicide and suggested
quent completed suicides (Nock et al., 2008), and they that the more exposure an individual has to suicidal
are usually more prevalent among the younger than behaviors, such as suicidal plans, attempts, or comple-
older populations (Evans, Hawton, Rodham, Psychol, tions, the more acceptable and feasible that behavior
& Deeks, 2005). Moreover, there are large gender dif- may become (Mueller & Abrutyn, 2015). The exposure
ferences in suicide attempts among youth, with girls effect is thought to be salient especially for adoles-
reporting at least 2 to 4 times more attempt than boys cents, because their self-image is still forming
in most countries (Canetto, 2008). These trends were (Crosnoe & Johnson 2011), they are particularly
similarly reflected in statistics on Turkey, where the socially conscious, and they are more susceptible to
youth suicide rates have been continually increasing others’ influence (Giordano, 2003). Thus, researchers
since 1990 (Coskun, Zoroglu, & Ghaziuddin, 2012). focus on social contagion of suicidal behaviors among
The highest suicide attempt rate was in the 15–19 age adolescents (Crepeau-Hobson & Leech, 2014).
group, where the mean annual rate of suicide attempt Much of the supporting evidence for the suicide con-
per 100,000 was 255 for girls and 75 for boys tagion hypothesis comes from studies of (a) the associ-
(Devrimci-Ozguven & Sayil, 2003). ation between suicides in mass media platforms and
Determining the factors that lead adolescents to suicide rates (Gould, Jamieson, & Romer, 2003), (b) the
suicidal behavior is important in developing appropri- association between celebrity suicides and suicide rates
ate intervention methods. To this end, researchers (Stack, 2005), and (c) the geographic and temporal clus-
identified various demographic (Brent, Baugher, tering of suicides (Baller & Richardson, 2002). Overall,
Bridge, Chen, & Chiappetta, 1999), psychological these studies indicate a possible link between exposure
(Brent, 1995; Thompson, Mazza, Herting, Randell, & and suicidal behavior. Relatively few studies have exam-
Eggert, 2005), and social (Wray, Colen, & Pescosolido, ined the effects of ESB through social networks, and the
2011) precursors. Nevertheless, there are still some existing research is contradictory. Some found an

CONTACT Muhammed Yıldız [email protected], [email protected] Department of Sociology, Louisiana State University, 126 Stubbs
Hall, Baton Rouge, LA 70803, USA.
ß 2018 Taylor & Francis
2 M. YILDIZ ET AL.

increased risk of suicidality among exposed adolescents Third, past research on suicide contagion through
(Abrutyn & Mueller, 2014; Bearman & Moody, 2004; social networks occurred primarily in the West, mostly
Mueller & Abrutyn, 2015), whereas others found no the United States (Abrutyn & Mueller, 2014; Bearman
relation (Harris & Bettiol, 2017). Some researchers even & Moody, 2004; Harris & Bettiol, 2017; Mueller &
argued that ESB of known others may enable an indi- Abrutyn, 2015). Less is known about the effects of
vidual to see the negative consequences of suicide, and ESB in Middle Eastern countries such as Turkey
hence prevent any tendencies towards suicidal behavior where suicide research fail to go beyond simple
(Mercy et al., 2001). descriptive studies of samples of completed and
Apart from the inconsistent findings, there are also attempted suicides (Devrimci-Ozguven & Sayil, 2003).
considerable gaps in what we know about the effects Compared to the United States and other Western
of ESB through social networks. First, previous studies countries, Turkey is a racially, culturally, and reli-
have ignored the potential exacerbating role of ESB on giously homogeneous country where 90–99% of the
gender differences in suicide attempt risk. Researchers people identify as Turk and Muslim (Turkish Ministry
often attributed girls’ higher risk of suicide to a of Religious Affairs, 2014). According to differential
greater likelihood of internalizing behavior (e.g., identification theory (Glaser, 1956), the transfer of
depression, negative attributional style, etc.) or differ- behaviors is more likely to occur in homogeneous
ential socialization that presents suicidality as a femin- societies like Turkey, because similarity facilitates imi-
ine way of coping with stressors (Canetto & tation by providing a base for identification with
Sakinofsky, 1998; Wichstrom & Rossow, 2002). others. Therefore, the effects of ESB might be more
Suicide contagion through social networks may aggra- salient for Turkish adolescents than for adolescents
vate this existing gender differences. During adoles- growing up in culturally and demographically diverse
cence, boys and girls socialize primarily within same- societies such as United States.
sex contexts (Maccoby, 1998). This gender difference Thus, we expected that ESB would be significantly
is especially true for Muslim societies where culture and positively associated with suicide attempts among
discourages interaction with opposite sex individuals, Turkish adolescents, even after controlling for risk and
especially in education and leisure (Ibrahim, 1982). protective factors associated with adolescent suicidal
Furthermore, girls’ same-sex friendships are often behavior (Hypothesis 1). In addition, we expected that
defined by greater intimacy and more disclosure of ESB would partially mediate the effects of female gen-
personal information than do boys (Benenson &
der on suicide attempt (Hypothesis 2).
Christakos, 2003). Therefore, girls have a higher
chance of experiencing ESB within their social net-
works, and this increased risk of exposure may partly Method
mediate the gender-suicide attempt link. Participants
Second, most previous studies controlled for only
basic demographic variables and often failed to control The survey data used in this study was conducted in
for important risk and protective factors associated Bagcilar, a district of Istanbul. Researchers randomly
with suicidal behaviors of adolescents (Harris & Bettiol, sampled all high schools in Bagcilar district, selecting
2017; Mercy et al., 2001). According to assortative relat- 2447 students for participation. We excluded 412 stu-
ing hypothesis, those at high risk of suicidal behavior dents with missing values. The final sample (n ¼ 2035)
are more likely to be friends with other high risk indi- included 1025 (50.36%) girls and 1010 (49.64%) boys.
viduals, and thus, the observed effect of exposure may The mean age was 14.35 years (SD ¼ 0.67), with partic-
simply be explained by unobserved, preexisting similar- ipants ranging in age from 13 to 17. Most students
ities between suicidal person and the exposed individ- were born in Turkey (96%) and lived with both their
ual (Joiner, 2003; Randall, Nickel, & Colman, 2015). parents (92%). Many (n ¼ 938, 46.09%) had been
Adolescent health problems such as depression exposed to at least one suicidal behavior at some point
(Thompson et al., 2005), smoking (Innamorati et al. in their lives, and 200 (9.83%) reported a suicide
2011), and alcohol use (Galaif, Sussman, Newcomb, & attempt. See Table 1 for descriptive statistics.
Locke, 2007) are well-known risk factors of suicidal
behaviors, and adolescents who have these problems
Measures
tend to cluster together (Van Ryzin & Dishion, 2014).
It is important to control for these psychological factors Suicide attempt is coded “1” if the respondent
to account for preexisting similarities between the sui- answered “yes” to any of two questions about having
cidal person and the exposed adolescent. attempted suicide (Have you ever attempted suicide?
DEATH STUDIES 3

Table 1. Descriptive statistics (n ¼ 2035). Symptom Checklist Questionnaire (Derogatis,

Mean Standard Min. Max. Lipman, & Covi, 1973) was 15 statements (e.g., I was
or (%) deviation value value
sad or had little interest in doing things) about the pre-
Suicide attempt 9.83%
Female 50.36% vious week. Responses ranged from 0 ¼ Almost never
ESB 1.10 1.47 0 5 to 3 ¼ Often. We used means and these questions were
Age 14.35 0.67 13 17
Parental SES 2.84 0.96 0 4
internally consistent (a ¼ 0.93).
Living with both parents 92.29% Negative life events (Zhang, Jia, Zhang, Wang, &
Parental support 1.97 0.63 0 3 Liu, 2015) was 19 experiences such as serious accident,
Parental regulation 2.13 0.59 0 3
Support from friends 2.26 0.62 0 3 severe illness, or death of a parent or sibling, in the
School attachment 3.10 0.95 0 4 past 12 months. Due to the small percentage of
Religious attendance 1.81 1.08 0 3
Neighborhood change 14.25% respondents with multiple negative life experiences, we
School change 12.92% coded negative life events as 0 ¼ None, 1 ¼ One, and
Depression 1.07 0.77 0 3
Negative life events 0.53 0.77 0 2 2 ¼ 2 or more.
Smoking 13.27% There were also yes/no questions about family
Alcohol use 8.94%
structure (living with both parents), neighborhood and
school change (moving to a different neighborhood or
and Did you attempt suicide during the current school school in the last 12 months), smoking (at least one
year?) and “0” otherwise. cigarette use), and alcohol use (at least one alcoholic
The ESB scale (Bernburg, Thorlindson, & drink during the last 30 days). Finally, there was one
Sigfusdottir, 2009) was five yes/no questions about 4-point question about religious attendance.
anyone telling you about considering suicide, or an
acquaintance or friend attempting or dying by suicide. Procedure
In the present sample, these items were internally con-
sistent (a ¼ 0.77). 2008 Youth in Europe survey (YIE; Kristjansson
The Family Affluence Scale (Boyce, Torsheim, Sigfusson, Sigfusdottir, & Allegrante, 2013) was con-
Currie, & Zambon, 2006) was four questions about ducted by European Cities Against Drugs (ECAD), a
family wealth that focus on being financially poor, and non-profit organization. Istanbul, as a member city of
unable to afford a car, necessities, and extracurricular ECAD, participated in the 2008 YIE, and the second-
activities. Response categories range from 0 ¼ Almost ary survey data used in this study comes from this
never to 4 ¼ Almost always. With reverse scoring, the project (see ECAD’s website for more detail: http://
totals were internally consistent (a ¼ 0.80) and ranged www.ecad.net). The survey questions, designed by the
Icelandic Center for Social Research and Analysis,
from 0 (lowest SES) to 4 (highest SES).
covered social problems among youth, such as health
Parental and friend support was five questions each
problems, delinquency, and suicidal behavior.
on how easy or difficult it was to receive: caring and
Teachers and research assistants administered ques-
warmth, conversations about private issues, advice
tionnaires in class. Students answered anonymously
about school, advice about other issues, and assistance
(Sigfusdottir, Kristjansson, & Agnew (2012) for a
for things (parents a ¼ 0.73, friends a ¼ 0.80). Scored
full discussion).
on a 4-point scale, higher values indicated more
intense care and support.
Parental regulation was six items about parental rules Results
and monitoring. Rules involved home, outside, and cur- In Table 2, we provide a series of regression models
few. Monitoring involved friends, associates, and even- for the main aims of the study, which are an examin-
ing time places. Responses ranged from 0 ¼ Applies very ation of the impact of ESB on suicide attempt
poorly to me to 3 ¼ Applies very well to me. We used (Hypothesis 1), and an assessment of the mediating
mean scores that ranged from 0 (minimum parental effects of ESB on the gender-suicide attempt associ-
regulation) to 3 (maximum parental regulation), and ation (Hypothesis 2). Our test of mediation follows
these items were internally consistent (a ¼ 0.74). conditional requirements where: (a) female gender sig-
School attachment was four 5-point Likert items on nificantly affects suicide attempt in the absence of
how adolescents felt about their school and teachers. ESB, (b) ESB has a significant effect on suicide
Items included feeling bad at school, wanting to quit attempt, (c) the effect of female gender on suicide
or change school, and getting along badly with teach- attempt shrinks upon addition of ESB to the model,
ers (a ¼ 0.75). Responses were reverse coded so that and (d) female gender significantly affects ESB (Baron
higher values indicated greater attachment. & Kenney, 1986). Following these steps, we first
4 M. YILDIZ ET AL.

Table 2. Suicide attempt and ESB regressed on varia- Model 2 shows the results when ESB was added. As
bles (n ¼ 2035). predicted, the results indicate that ESB (OR ¼ 1.95,
Suicide Attempta ESBb p < 0.001) had a statistically significant and positive
Model 1 Model 2 Model 3 effect on adolescents’ suicide attempts, even after con-
Coefficient OR Coefficient OR Coefficient IRR trolling for various risk and protective factors associ-
Female 0.53 1.70 0.27 1.31 0.39 1.48 ated with adolescents’ suicidal behavior. The effect
Control variables was quite robust; thus, for every unit increase in the
Age 0.19 1.21 0.06 1.06 0.17 1.19
Parental SES 0.01 0.99 .03 0.97 0.01 1.01 ESB scale, the odds of suicide attempt increased by
Living with both parents 0.39 0.68 .27 0.76 0.22 0.81 around 95 percent. This finding not only provided
Parental support 0.22 0.80 .17 0.84 0.03 0.98
Parental regulation 0.17 0.85 .33 0.72 0.08 1.08 support for the Hypothesis 1, but also fulfilled the
Support from friends 0.26 0.77 .32 0.73 0.01 1.01 second condition (b) of the mediation analysis, which
School attachment 0.27 0.76 .24 0.79 0.05 0.96
Religious attendance 0.01 0.99 0.09 1.09 0.08 0.93 required ESB to have a significant effect on suicide
Neighborhood change 0.37 1.44 0.06 1.06 0.28 1.32 attempt. Moreover, the results in Model 2 show that
School change 0.61 1.84 0.35 1.43 0.24 1.27
Depression 0.69 1.99 0.53 1.70 0.32 1.38
controlling for ESB dramatically shrank the coefficient
Negative life events 0.47 1.60 0.30 1.35 0.25 1.29 of gender, reducing the size of the gender coefficient
Smoking 0.65 1.92 0.42 1.53 0.28 1.32 by nearly half (1–0.27/0.53 ¼ 0.49), and made the asso-
Alcohol use 0.01 1.01 0.17 0.84 0.24 1.28
Mediator Variable ciation between gender and suicide attempt insignifi-
ESB – – 0.67 1.95 – – cant. Because the original focal relationship between
Model Statistics
Nagelkerke R-squared 0.21 0.35 0.14 gender and suicide attempt was eliminated with the
LR Chi-Square 216.72 371.56 295.84 inclusion of ESB, the third condition (c) of the medi-
p < 0.001, p < 0.01, p < 0.05.
a
ation analysis was also achieved.
Logistic regression.
b
Negative binomial regression. Finally, Model 3 shows the incidence rate ratios
(IRR) when ESB was regressed for gender and the
control variables to test the final condition (d) of the
regress suicide attempt on gender and control varia-
mediation, which required gender to be a significant
bles together to test condition (a) (Model 1). Next, we
predictor of ESB. As predicted, the female gender had
add ESB into the model (Model 2). In addition to pro-
a significant effect on ESB (IRR ¼ 1.48, p < 0.001)
viding evidence in support of necessary conditions (b)
which meant that the female adolescents in our sam-
and (c), Model 2 will also directly test Hypothesis 1.
ple were exposed to more suicidal behaviors than
Finally, we regress ESB on gender and control varia-
males within their social networks. Among the control
bles for condition (d) (Model 3). Hierarchical multiple
variables, higher age (IRR ¼ 1.19, p < 0.001), neighbor-
logistic regression was used in models predicting sui-
hood change (IRR ¼ 1.32, p < 0.01), school change
cide attempt (Model 1 and 2). Negative binomial
(IRR ¼ 1.27, p < 0.01), depression (IRR ¼ 1.38,
regression was used for predicting ESB (Model 3),
p < 0.001), negative life events (IRR ¼ 1.29, p < 0.001),
which was an over-dispersed, count variable with a smoking (IRR ¼ 1.32, p < 0.01), and alcohol use
larger variance than its mean. (IRR ¼ 1.28, p < 0.05) were all significantly associated
Model 1 shows the results when suicide attempt with higher levels of ESB. Conversely, higher religious
was regressed for the gender and control variables attendance (IRR ¼ 0.93, p < 0.01) was significantly
together. As expected, being female is significantly associated with lower levels of ESB.
associated with higher risk of suicide attempt
(OR ¼ 1.70, p < 0.01), net of other control variables.
This finding fulfills the first condition (a) of the medi- Discussion
ation, which requires gender to be a significant pre- Present results indicate that among this large group of
dictor of suicide attempts. Among the control adolescents in Turkey, a predominantly Muslim coun-
variables, school change (OR ¼ 1.84, p < 0.01), depres- try, ESB through social networks was significantly
sion (OR ¼ 1.99; p < .001), experiencing negative life associated with attempting suicide. We also found
events (OR ¼ 1.60, p < 0.001), and smoking strong support for our mediation hypothesis, suggest-
(OR ¼ 1.92, p < 0.01) were all significantly associated ing that the higher suicide attempt risk among females
with higher odds of suicide attempt, whereas high lev- could partly be attributed to higher levels of ESB
els of support from friends (OR ¼ .77, p < 0.05) and among females.
school attachment (OR ¼ .76, p < 0.01) were signifi- The significant relationship between ESB and sui-
cantly associated with a lower risk of suicide attempt cide attempt found in this study is consistent with
among adolescents. those few previous studies in the U.S. that reported an
DEATH STUDIES 5

increased risk of suicidality among adolescents longitudinal analyses are needed to truly assess causal
exposed to family or friend suicidal behavior (Abrutyn effects, particularly in the link between ESB and sui-
& Mueller, 2014; Bearman & Moody, 2004) and pro- cide attempt. Similarly, we cannot strictly dismiss the
vides further support for the suicide contagion potential for network selection processes due to the
through social networks. In addition, it adds to the lit- cross-sectional design of this study. Even though we
erature by providing empirical evidence from a non- controlled for individual’s levels of social integration,
Western society. The fact that the effect of ESB per- regulation, demographic characteristics, depression,
sisted even after controlling for various demographic, stress-exposure, and behavioral risk factors, we cannot
psychological, and social risk factors of suicide pro- completely count out the possibility that assortative
vided some confidence that our results are not due to relating is occurring on some factor that is not
the assortative relating hypothesis and other alterna- accounted for in our models. Another limitation of
tive explanations. this study is the fact that we were unable to include
Our study also extends the literature by assessing some demographic variables that are generally
the mediating effect of ESB on the relationship included in studies conducted in United States.
between gender and suicide attempt risk. Previous Specifically, we could not control for self-reported
studies have mostly focused on theories of differential race/ethnicity and religion of the adolescents, as these
risk factors or gendered socialization to explain the questions were not included in the Turkey part of the
higher suicide attempt rates among females (Canetto survey. We view this as a minor limitation due to the
& Sakinofsky, 1998; Wichstrom & Rossow, 2002). Our homogeneous nature of the Turkish context; nonethe-
study demonstrated that in addition to these factors, less, since understanding the impact of ESB on sui-
suicide contagion also serves as a spreading mechan- cidal behaviors for minority subgroups is important in
ism that enlarges the gender differences in suicide its own right, future research should examine the role
attempt risk. Because adolescents usually socialize in that race/ethnicity and religion play in this process.
same-sex contexts and female adolescents exhibit A final limitation of this study is related to the
more suicidal behavior than do males (due to differen- measurement of ESB. In contrast to previous studies
tial risk factors or differential socialization), they also that used dichotomous measures only for exposure to
have a higher risk of being exposed to suicidal behav- family members’ or friends’ suicidal behavior
ior in their social networks. The resulting increased (Abrutyn & Mueller, 2014; Bearman & Moody, 2004;
levels of ESB among female adolescents exacerbates Mueller & Abrutyn, 2015), our measure of ESB was
the existing gender differences even more. created by adding up different types of exposures ado-
Clinicians, researchers, and policy makers should lescents experienced. Although this type of measure-
ensure that appropriate identification methods and ment might be more comprehensive than a simple
prevention activities are developed to reduce the sui- dichotomous one, it is limited because it does not dif-
cides that could result from ESB. Specifically, a social ferentiate the exposure effects based on closeness to
system should be established that enables immediate the victim or fatality of the outcome. Further research
intervention when a suicide or suicide attempt occurs is needed on this issue since who exhibited which sui-
within a family, school, or community. Previous cidal behavior might significantly influence the conta-
research on this issue has indicated that appropriate gion effect.
crisis intervention methods such as first talk-throughs Despite these limitations, a rigorous methodology
and psychological debriefings by a trained mental was employed to explore the associations among ESB,
health professional can be effective in inhibiting the gender, and suicide attempt among adolescents in an
spread of suicidal behavior among adolescents overlooked society. To the best of our knowledge, this
(Poijula, Wahlberg, & Dyregrov, 2001). Crisis inter- was the first study that examined suicide contagion
vention strategies should take robust gender differen- among adolescents in a Muslim country, and the first
ces into consideration as well, as this may help study ever that examined the mediating role of ESB
determine gender-specific strategies for prevent- on gender-suicide attempt association. Our findings
ing suicides. indicate that suicidal behavior can spread through
Although our findings provided important insights, social networks via ESB, and females’ greater risk of
this study had some limitations that must be consid- experiencing ESB can contribute significantly to the
ered. First of all, due to the cross-sectional nature of existing gender differences in suicide attempts among
the data, causal predictions were limited. Although the Turkish adolescents. Studies of psychological, behav-
theoretical approach and past research guiding these ioral, and social risk factors associated with adolescent
analyses improve our ability to infer directionality, suicidality are undoubtedly important, but researchers
6 M. YILDIZ ET AL.

should also keep in mind network contagion processes Crepeau-Hobson, M. F., & Leech, N. L. (2014). The impact
in suicidal behavior while developing prevent- of exposure to peer suicidal self-directed violence on
youth suicidal behavior: a critical review of the literature.
ive strategies.
Suicide and Life-Threatening Behavior, 44, 58–77.
doi:10.1111/sltb.12055
References Crosnoe, R., & Johnson, M. K. (2011). Research on adoles-
cence in the twenty-first century. Annual Review of
Abrutyn, S., & Mueller, A. S. (2014). Are suicidal behaviors Sociology, 37, 439–460. doi:10.1146/annurev-soc-081309-
contagious in adolescence? Using longitudinal data to 150008
examine suicide suggestion. American Sociological Review, Derogatis, L. R., Lipman, R. S., & Covi, L. (1973). SCL-90:
79, 211–227. doi:10.1177/0003122413519445 an outpatient psychiatric rating scale-preliminary report.
Akers, R. L., & Jennings, W. G. (2009). The social learning Psychopharmacology Bulletin, 9, 13–28.
theory of crime and deviance. In Handbook on crime and Devrimci-Ozguven, H., & Sayıl, I. (2003). Suicide attempts
deviance (pp. 103–120). New York, NY: Springer. in Turkey: results of the WHO—EURO multicentre study
Baller, R. D., & Richardson, K. K. (2002). Social integration, on suicidal behaviour. The Canadian Journal of
imitation, and the geographic patterning of suicide. Psychiatry, 48, 324–329. doi:10.1177/070674370304800508
American Sociological Review, 67, 873–888. doi:10.2307/ Evans, E., Hawton, K., Rodham, K., Psychol, C., & Deeks, J.
3088974 (2005). The prevalence of suicidal phenomena in adoles-
Baron, R. M., & Kenny, D. A. (1986). The moderator–me- cents: a systematic review of population-based studies.
diator variable distinction in social psychological research: Suicide and Life-Threatening Behavior, 35, 239–250.
conceptual, strategic, and statistical considerations. doi:10.1521/suli.2005.35.3.239
Journal of Personality and Social Psychology, 51, 1173. Fleischmann, A., & De Leo, D. (2014). The World Health
doi:10.1037/0022-3514.51.6.1173 Organization's report on suicide: a fundamental step in
Bearman, P. S., & Moody, J. (2004). Suicide and friendships worldwide suicide prevention. Crisis, 35, 289–291.
among American adolescents. American Journal of Public doi:10.1027/0227-5910/a000293
Health, 94, 89–95. doi:10.2105/AJPH.94.1.89 Galaif, E. R., Sussman, S., Newcomb, M. D., & Locke, T. F.
Benenson, J. F., & Christakos, A. (2003). The greater fragil- (2007). Suicidality, depression, and alcohol use among
ity of females' versus males' closest same-sex friendships. adolescents: a review of empirical findings. International
Child Development, 74, 1123–1129. doi:10.1111/1467- Journal of Adolescent Medicine and Health, 19, 36.
Giordano, P. C. (2003). Relationships in adolescence.
8624.00596
Bernburg, J. G., Thorlindsson, T., & Sigfusdottir, I. D. Annual Review of Sociology, 29, 257–281. doi:10.1146/
annurev.soc.29.010202.100047
(2009). The spreading of suicidal behavior: The context-
Glaser, D. (1956). Criminality theories and behavioral
ual effect of community household poverty on adolescent
images. American Journal of Sociology, 61, 433–444.
suicidal behavior and the mediating role of suicide sug-
doi:10.1086/221802
gestion. Social Science and Medicine, 68, 380–389.
Gould, M., Jamieson, P., & Romer, D. (2003). Media conta-
doi:10.1016/j.socscimed.2008.10.020
gion and suicide among the young. American Behavioral
Boyce, W., Torsheim, T., Currie, C., & Zambon, A. (2006).
Scientist, 46, 1269–1284. doi:10.1177/0002764202250670
The family affluence scale as a measure of national Harris, K. M., & Bettiol, S. (2017). Exposure to suicidal
wealth: validation of an adolescent self-report measure. behaviors: A common suicide risk factor or a personal
Social Indicators Research, 78, 473–487. doi:10.1007/ negative life event? The International Journal of Social
s11205-005-1607-6 Psychiatry, 63, 70–77. doi:10.1177/0020764016682361
Brent, D. A. (1995). Risk factors for adolescent suicide and Ibrahim, H. (1982). Leisure and Islam. Leisure Studies, 1,
suicidal behavior: mental and substance abuse disorders, 197–210. doi:10.1080/02614368200390161
family environmental factors, and life stress. Suicide and Innamorati, M., Leo, D. D., Rihmer, Z., Serafini, G.,
Life-Threatening Behavior, 25, 52–63. doi:10.1111/j.1943- Brugnoli, R., Lester, D., … Girardi, P. (2011). Tobacco
278X.1995.tb00490.x smoking and suicidal ideation in school-aged children
Brent, D. A., Baugher, M., Bridge, J., Chen, T., & 12–15 years old: Impact of cultural differences. Journal of
Chiappetta, L. (1999). Age- and sex-related risk factors Addictive Diseases, 30, 359–367. doi:10.1080/
for adolescent suicide. Journal of the American Academy 10550887.2011.609802
of Child and Adolescent Psychiatry, 38, 1497–1505. Joiner, T. E. (2003). Contagion of suicidal symptoms as a
doi:10.1097/00004583-199912000-00010 function of assortative relating and shared relationship
Canetto, S. S. (2008). Women and suicidal behavior: a cul- stress in college roommates. Journal of Adolescence, 26,
tural analysis. The American Journal of Orthopsychiatry, 495–504. doi:10.1016/S0140-1971(02)00133-1
78, 259. doi:10.1037/a0013973 Kristjansson, A. L., Sigfusson, J., Sigfusdottir, I. D., &
Canetto, S. S., & Sakinofsky, I. (1998). The gender paradox Allegrante, J. P. (2013). Data collection procedures for
in suicide. Suicide and Life-Threatening Behavior, 28, school-based surveys among adolescents: The Youth in
1–23. doi:10.1111/j.1943-278X.1998.tb00622.x Europe Study. Journal of School Health, 83, 662–667.
Coskun, M., Zoroglu, S., & Ghaziuddin, N. (2012). Suicide doi:10.1111/josh.12079
rates among Turkish and American youth: A cross-cul- Maccoby, E. E. (1998). The two sexes: Growing up apart,
tural comparison. Archives of Suicide Research, 16, 59–72. coming together (Vol. 4). Cambridge, MA: Harvard
doi:10.1080/13811118.2012.640612 University Press.
DEATH STUDIES 7

Mercy, J. A., Kresnow, M-J., O'Carroll, P. W., Lee, R. K., Threatening Behavior, 35, 121–133. doi:10.1521/
Powell, K. E., Potter, L. B., … Bayer, T. L. (2001). Is sui- suli.35.2.121.62877
cide contagious? A study of the relation between exposure Thompson, E. A., Mazza, J. J., Herting, J. R., Randell, B. P.,
to the suicidal behavior of others and nearly lethal suicide & Eggert, L. L. (2005). The mediating roles of anxiety,
attempts. American Journal of Epidemiology, 154, depression, and hopelessness on adolescent suicidal
120–127. doi:10.1093/aje/154.2.120 behaviors. Suicide and Life-Threatening Behavior, 35,
Mueller, A. S., & Abrutyn, S. (2015). Suicidal disclosures 14–34. doi:10.1521/suli.35.1.14.59266
among friends: using social network data to understand Turkish Ministry of Religious Affairs (2014). Türkiye’de dini
suicide contagion. Journal of Health and Social Behavior, hayat araştırması. Ankara: Diyanet İşleri Başkanlığı
56, 131–148. doi:10.1177/0022146514568793 Yayınları.
Nock, M. K., Borges, G., Bromet, E. J., Alonso, J., Van Ryzin, M. J., & Dishion, T. J. (2014). Adolescent
Angermeyer, M., Beautrais, A., … Williams, D. (2008). deviant peer clustering as an amplifying mechanism
underlying the progression from early substance use
Cross-national prevalence and risk factors for suicidal
to late adolescent dependence. Journal of Child
ideation, plans and attempts. The British Journal of
Psychology and Psychiatry, 55, 1153–1161. doi:10.1111/
Psychiatry: The Journal of Mental Science, 192, 98–105.
jcpp.12211
doi:10.1192/bjp.bp.107.040113
World Health Organization (2014). Global status report on
Poijula, S., Wahlberg, K. E., & Dyregrov, A. (2001).
alcohol and health-2014. Geneva, Switzerland: World
Adolescent suicide and suicide contagion in three second- Health Organization.
ary schools. International Journal of Emergency Mental Wichstrøm, L., & Rossow, I. (2002). Explaining the gender
Health, 3, 163–170. difference in self-reported suicide attempts: a nationally
Randall, J. R., Nickel, N. C., & Colman, I. (2015). Contagion representative study of Norwegian adolescents. Suicide
from peer suicidal behavior in a representative sample of and Life-Threatening Behavior, 32, 101–116. doi:10.1521/
American adolescents. Journal of Affective Disorders, 186, suli.32.2.101.24407
219–225. doi:10.1016/j.jad.2015.07.001 Wray, M., Colen, C., & Pescosolido, B. (2011). The soci-
Sigfusdottir, I. D., Kristjansson, A. L., & Agnew, R. (2012). ology of suicide. Annual Review of Sociology, 37, 505–528.
A comparative analysis of general strain theory. Journal doi:10.1146/annurev-soc-081309-150058
of Criminal Justice, 40, 117–127. doi:10.1016/ Zhang, W. C., Jia, C. X., Zhang, J. Y., Wang, L. L., & Liu,
j.jcrimjus.2012.01.001 X. C. (2015). Negative life events and attempted suicide
Stack, S. (2005). Suicide in the media: a quantitative review in rural China. PLoS One, 10, e0116634. doi:10.1371/
of studies based on nonfictional stories. Suicide and Life- journal.pone.0116634