Asst. Prof. Florence C. Navidad, RMT, RN, M.Ed.
1
Objec&ves
At
the
end
of
the
topic,
the
students
are
expected
to:
• Be
familiar
with
the
use
of
ANOVA
in
hypothesis
tes&ng;
• Differen&ate
the
three
types
of
varia&ons
that
are
considered
in
Anova
tests;
• Dis&nguish
between
One-‐way
and
Two-‐way
Anova
designs
• Perform
a
Post
Anova
test
aGer
each
hypothesis
tested
using
the
Tukey’s
HSD.
2
Different Forms of ANOVA
• One-‐way
ANOVA
(one-‐factorial
designs)
• One-‐Way
Between-‐Groups
(independent
groups)
ANOVA
• One-‐Way
Within-‐Subjects
(Repeated
Measures)
ANOVA
• Two
-‐way
ANOVA
(two-‐factorial
designs)
• Two-‐Way
Between-‐Groups
(Independent
groups)
ANOVA
• Post
–Hoc
Analysis
3
4
One-‐Way
or
One-‐Factor
Anova
[One-‐Way
Between-‐Groups
(Independent
groups)]
• Used
when
the
experiment
has
more
than
two
levels
of
the
independent
variable.
• results
stop
when
decision
is
the
acceptance
of
the
null
hypothesis.
• when
decision
is
the
rejec&on
of
null
hypothesis,
search
con&nues
which
pair
accounts
for
the
difference.
5
Steps
in
Conduc&ng
One-‐Way
Anova
1. Formulate the null and alternative hypothesis.
2. Choose a level of significance.
3. Choose the appropriate test statistic and
Compute the value of the statistical test
4. Establish the critical region
5. Make a decision
Decision Rule: Do not accept Ho if the test
statistic value lies inside the critical region,
otherwise accept Ho
6. State the appropriate conclusion.
6
ANOVA
TABLE
Source
of
Sum
of
dF Mean
Squares F
Sta2s2c
Varia2on Squares
Between
SSb k-‐1 MSb
=
SSb/k-‐1 MSb/MSw
Groups
Within
SSw N-‐k MSw
=
SSw/(N-‐k)
Groups
Total SSt N-‐1
7
Example
1
• Weight
gain
(oz)
of
20
baby
rats
fed
four
A
B
C
D
different
brands
of
1
7
9
8
cereal.
1
7
6
6
1
7
5
4
• Is
there
a
significant
1
7
3
1
difference
in
weight
1
7
2
1
gain
among
the
rats
that
were
fed
the
different
brands
of
cereal?
8
Steps
in
Hypothesis
Tes&ng
H0:
there
no
significant
difference
in
weight
gain
among
the
rats
that
were
fed
the
different
brands
of
cereal
H1:
there
is
a
significant
difference
in
weight
gain
among
the
rats
that
were
fed
the
different
brands
of
cereal
α
=
0.05
One-‐way
ANOVA:
SPSS
Cri&cal
region:
F
(3,16)
3.24
&
F-‐test:
7.353
P-‐value:
0.003
<
0.05
Decision:
do
not
cccept
null
hypothesis
and
accept
alterna&ve
hypothesis
Conclusion:
This
means
that
at
least
one
of
the
means
is
significantly
different
from
the
others.
9
POST
HOC
ANALYSIS
• When
the
null
hypothesis
is
rejected,
it
may
be
desirable
to
find
which
mean(s)
is
(are)
different,
and
at
what
ranking
order.
• Three
most
commonly
used
sta&s&cal
inference
procedures,
geared
at
doing
this,
are
presented:
– Fisher’s
least
significant
difference
(LSD)
method
– Tukey
HSD
test
(Tukey)
– Student-‐Newman-‐Keuls
(S-‐N-‐K)
10
Mul&ple
comparison
Example
1
Pair
P-‐value
decision
Results
A
–
B
0.002
p
<
α
Significant
A
–
C
0.033
p
<
α
Significant
A
–
D
0.139
p
>
α
Not
significant
B
–
C
0.442
p
>
α
Not
significant
B
–
D
0.139
p
>
α
Not
significant
C
–
D
0.868
p
>
α
Not
significant
11
EXAMPLE
2
A
medical
equipments
retailer
is
selling
4
brands
of
stethoscope.
A
B
C
D
7
9
2
4
The
owner
interested
if
there
is
3
8
3
5
a
significant
difference
in
the
5
8
4
7
average
sales
of
the
four
6
7
5
8
brands
of
stethoscope
for
one
9
6
6
3
week.
Perform
the
f-‐test
and
4
9
4
4
test
the
hypothesis
at
0.05
3
10
2
5
level
of
significance
that
the
average
sales
of
the
four
brands
of
stethoscope
are
equal.
12
Steps
in
Hypothesis
Tes&ng
H0
:
There
is
no
significant
difference
in
the
average
sales
of
the
four
brands
of
stethoscope.
H1
:
There
is
a
significant
difference
in
the
average
sales
of
the
four
brands
of
stethoscope.
α
=
0.05
One-‐way
ANOVA:
SPSS
Cri&cal
region:
F
(3,24)
3.01
&
F-‐test:
7.969
P-‐value:
0.001
<
0.05
Decision:
do
not
accept
null
hypothesis
and
accept
alterna&ve
hypothesis
Conclusion:
Therefore,
there
is
a
significant
difference
in
the
average
sales
of
the
4
brands
of
stethoscope.
13
Mul&ple
comparison
Example
2
Pair
P-‐value
decision
Results
A
–
B
0.025
p
<
α
Significant
A
–
C
0.350
P
>
α
not
Significant
A
–
D
0.999
p
>
α
Not
significant
B
–
C
0.000
P
<
α
significant
B
–
D
0.018
P
<
α
significant
C
–
D
0.432
p
>
α
Not
significant
14
Example
3
Three
brands
of
reducing
pills
were
tried
on
a
sample
of
8
female
adults.
The
data
is
reflected
on
the
table
below
in
terms
of
weight
loss
(lb).
Test
the
hypothesis
that
there
is
no
significant
difference
in
the
average
weight
loss
(in
lb)
among
the
three
groups
of
respondents
using
the
3
brands
of
reducing
pill
at
the
0.01
level.
Respondents
Brand
A
Brand
B
Brand
C
no.
1
4.5
3.2
3.0
2
4.1
3.0
2.8
3
3.6
3.8
3.2
4
5.3
3.9
3.6
5
4.8
4.2
3.5
6
2.7
3.1
3.5
7
4.3
4.0
2.9
8
3.8
3.3
3.6
15
One-‐Way
or
One-‐Factor
Anova
[One-‐Way
Within-‐Subjects
(Repeated
Measures)]
• Tes2ng
the
null
hypothesis
that,
on
the
average,
the
scores
of
subjects
remain
the
same
during
each
of
the
different
condi2ons
(the
baseline
condi2on
and
the
2,3,
or
4
different
condi2ons)
16
Example
1
• The
facilitator
of
a
stress
management
therapy
group
conducted
a
study
to
determine
the
most
effec2ve
relaxa2on
technique/s
for
stress
reduc2on.
The
10
members
of
his
stress
management
group
par2cipated
in
the
study.
The
heart
rate
of
each
par2cipant
was
monitored
during
each
of
five
condi2ons.
Each
par2cipant
experienced
all
five
condi2ons
during
the
same
session
to
control
for
varia2ons
in
the
amount
of
stress
experienced
from
day
to
day.
17
The
mean
heart
rate
(beats
per
minute)
for
each
subject
during
each
condi2on
is
as
follows:
Sub
Baseline
Medita& Comedy
Nature
Music
on
1
85
70
75
71
74
2
79
69
73
70
72
3
91
82
87
83
86
4
93
80
85
79
84
5
92
80
86
81
87
6
87
79
83
80
81
7
84
72
77
73
76
8
78
69
74
71
73
9
79
69
73
70
72
10
80
71
74
72
73
18
Steps
in
Hypothesis
Tes&ng
H0
:
There
is
no
significant
difference
in
the
mean
heart
rates
of
subjects
during
each
of
the
five
condi&ons.
H1
:
There
is
a
significant
difference
in
the
mean
heart
rates
of
subjects
during
each
of
the
five
condi&ons.
α
=
0.05
One-‐way
ANOVA:
SPSS
Cri&cal
region:
P-‐value:
0.000
<
0.05
NOTE:
Read
test
within-‐subjects
effect
for
the
analysis
or
Mul&variate
tests
Decision:
do
not
accept
null
hypothesis
and
accept
alterna&ve
hypothesis
Conclusion:
There
is
a
significant
difference
in
the
mean
heart
rates
of
subjects
during
each
of
the
five
condi&ons.
19
Example
2
• The
data
were
collected
in
a
study
inves&ga&ng
factors
associated
with
the
risk
of
developing
high
blood
pressure,
or
hypertension.
This
test
determines
if
there
is
evidence
that
the
diastolic
blood
pressure
changes
significantly
during
three
different
condi&ons:
res&ng,
doing
mental
arithme&c,
and
immersing
a
hand
in
a
cold
water.
Ho:
blood
pressure
does
not
change
during
the
stressors
20
Dbprest
Dbpma
Dbpcp
60.75
81.33
65.00
76.38
82.67
80.00
65.38
90.33
90.00
72.13
82.33
89.00
68.75
75.33
70.00
73.50
82.33
70.00
57.13
68.00
62.00
65.63
78.67
73.00
67.38
81.67
66.00
56.50
58.67
65.00
21
RANDOMIZED
BLOCKS
(TWO-‐WAY)
ANALYSIS
OF
VARIANCE
[Two-‐Way
Between-‐Groups
(Independent
Groups)
22
• Two
independent
variables
at
the
same
&me
• The
purpose
of
designing
a
randomized
block
experiment
is
to
reduce
the
within-‐treatments
varia4on.
• Determine
whether
there
is
significant
difference
in
the
mean
score
from
the
different
methods
of
treatments.
(columns)
• Determine
whether
there
is
a
difference
in
the
mean
scores
obtained
by
the
sample
(students).
(rows)
• Major
difference:
We
now
have
3
separate
tests,
and
three
separate
F
values
23
Randomized
Blocks
Block
all
the
observa&ons
with
some
commonality
across
treatments
Treatment
4
Treatment
3
Treatment
2
Treatment
1
Block3
Block2
Block
1
24
Anova
Table
for
the
Randomized
Complete
Block
Design
Source
of
Sum
of
dF Mean
Squares F
Sta&s&c
Varia&on Squares
Treatments
SStr k-‐1 SStr/k-‐1 MS(SStr)/
(column)
MS(SSr)
Blocks
(rows) SSb n-‐1 SSb/n
-‐
1
Residual
SSr (k
–
1)
SSr/(k
–
1)(n
–
1)
(error)
(n
–
1)
Total SSt Kn
-‐
1
25
Three
dis&nct
hypothesis
tests
• Test
for
the
main
effects
– The
mean
difference
between
levels
of
the
first
factor.
– The
mean
difference
between
level
of
the
second
factor
• Test
for
the
interac&on
– Any
other
mean
differences
that
may
result
from
the
unique
combina&on
of
the
two
factors.
26
EXAMPLE
1
Forty-‐five
sta&s&cs
students
were
randomly
assigned
to
one
of
three
instructors
and
to
one
of
the
three
methods
of
teaching.
Achievement
was
measured
on
a
test
administered
at
the
end
of
the
term.
Use
the
two-‐way
ANOVA
at
0.05
level
of
significance
to
test
the
following
hypothesis.
27
TWO-‐FACTOR
ANOVA
with
Significant
Interac&on
TEACHER
FACTOR
A
B
C
Method
of
Teaching
1
40
50
40
41
50
41
40
48
40
39
48
38
38
45
38
Method
of
Teaching
2
40
45
50
41
42
46
39
42
43
38
41
43
38
40
42
Method
of
Teaching
3
40
40
40
43
45
41
41
44
41
39
44
39
38
43
38
28
Problem
• Is
there
a
significant
difference
in
the
performance
of
students
under
the
three
different
teachers?
• Is
there
a
significant
difference
in
the
performance
of
students
under
the
three
different
methods
of
teaching?
• Is
there
a
significant
difference
in
the
interac&on
of
the
three
groups
of
students
under
three
different
methods
of
teaching
and
instructors?
29
Hypothesis
1. H0
:
There
is
no
significant
difference
in
the
performance
of
the
three
groups
of
students
under
three
different
instructors.
2.
H0
:
There
is
no
significant
difference
in
the
performance
of
the
three
groups
of
students
under
three
different
methods
of
teaching.
3.
H0
:
There
is
no
significant
difference
in
the
interac&on
of
the
three
groups
of
students
under
three
different
methods
of
teaching
and
instructors.
30
Descrip&ve
A
B
C
total
Method
1
39.6
28.2
39.4
42.4
Method
2
39.2
42
44.8
42
Method
3
40.2
43.2
39.8
41.07
total
39.67
44.47
41.3
31
Conclusion
1. Computed
F-‐value
(column).
Accept
null
hypothesis.
Therefore,
there
is
no
significant
differences
in
the
performance
of
the
three
groups
of
students
under
three
different
instructors.
2. Computed
F-‐value
(row).
Accept
null
hypothesis.
Therefore,
there
is
no
significant
differences
in
the
performance
of
the
students
under
the
three
different
methods
of
teaching
.
3. Computed
F-‐value
(column-‐row).
Reject
null
hypothesis.
Therefore,
there
is
a
significant
difference
in
the
interac&on
of
the
three
groups
of
students
under
three
different
methods
of
teaching
and
instructors.
32
Example
2
A
male
psychology
student
decides
to
conduct
a
study
to
determine
the
effec&veness
of
various
approaches
to
mee&ng
someone.
The
three
approaches
he
wants
to
inves&gate
are
as
follows:
causal
conversa&on
approach,
humor
approach,
and
pick-‐up
line
approach.
He
also
decided
to
include
aurac&veness
as
a
2nd
independent
variable.
The
measure
of
success
for
each
approach
is
the
length
of
&me
in
minutes.
The
results
are
as
follows
using
0.05
level
of
confidence.
33
Subject
aurac&ve approach
&me
nss
1
1
1
43
2
1
1
35
3
1
1
52
4
1
1
61
5
1
2
45
6
1
2
52
7
2
2
30
8
2
2
45
9
2
3
1
10
2
3
2
11
2
3
3
12
2
3
4
34
Hypothesis
1. Ho:
there
is
no
significant
difference
in
the
main
effect
of
aurac&veness.
2. H0:
there
is
no
significant
difference
in
the
main
effect
approach
type;
3. H0:
there
is
no
significant
difference
in
the
interac&on
of
aurac&veness
and
approach
type
35
Conclusion
1.
there
is
no
significant
difference
in
the
main
effect
of
aurac&veness.
2.
there
is
a
significant
difference
in
the
main
effect
approach
type;
3.
there
is
a
significant
difference
in
the
interac&on
of
aurac&veness
and
approach
type
36
The
end
“It is possible to fail in many ways…
While to succeed is possible only in one
way.”
- Aristotle
37
