Transcript Chapter 11
Chapter 11
Hypothesis Tests: Two
Related Samples
Overview
Learning
objectives
Vocabulary lesson again
Introduce t test for related samples
Advantages and disadvantages
An example
Review questions
2
Learning Objectives
Difference
between independent-measures
& related-samples experimental design
Difference between repeated-measures &
matched-subjects experimental design
Compute t test for dependent groups
Advantages and disadvantages
Measures of effect size
3
Vocabulary
Related-samples
t statistic Repeated-
measures design
Matched-samples design
Difference scores (estimated standard error
of D-bar)
Individual differences
Carry-over effects
4
Related-samples t statistic
Two
forms
• Repeated-measures design
• Matched-samples design
Use difference scores between two
measurement points rather than means
5
Repeated-measures
The
same participants give us data on two
measures (e. g. Before and After treatment)
• Aggressive responses before video and
aggressive responses after
Accounts for the fact that if someone is high
on one measure probably high on other.
6
Matched-samples
Individuals
in one group are matched to
individuals in a second sample
• Matching based on variables thought to
be relevant to the study
• Not always perfect match
Also called matched pairs or pairwise t test
7
Difference Scores
Calculate
difference between first and
second score (between individual scores or
matched pairs)
• e. g. Difference = Before – After
• D = X2-X1
Base subsequent analysis on difference
scores
8
The Formulas
D D
t
, df n 1
sD
sD
2
s
s
n
n
9
Hypothesis Testing
Null
states that
• The population of difference scores has a
mean of zero
• No systematic or consistent difference
between the conditions
H 0 : D 0
Alternative states that
H1 : D 0
• There is a real difference
10
Advantages of Related Samples
Eliminate
subject-to-subject variability
• Makes the test more powerful
Control for extraneous variables
Need fewer subjects
11
Disadvantages of Related
Samples
Order
effects
Carry-over effects
Subjects no longer naïve
Change may just be a function of time
Sometimes not logically possible
12
An Example
Therapy
for rape victims
• Foa, Rothbaum, Riggs, & Murdock
(1991)
A group (n=9) received Supportive
Counseling
Measured post-traumatic stress disorder
symptoms before and after therapy
13
Step 1
Null:
there is no difference in symptoms in
individuals after treatment
Alternative: there is a difference in
symptoms
α=.05, two tailed
H 0 : D 0
H1 : D 0
14
Step 2
With
a sample of 9
• df = n-1 = 9-1 = 8
• Critical value = +2.306
Sketch
15
The Data: Therapy for PTSD
Before
Mean
St. Dev.
21
24
21
26
32
27
21
25
18
23.84
4.20
After
15
15
17
20
17
20
8
19
10
15.67
4.24
Diff.
6
9
4
6
15
7
13
6
8
8.17
3.60
16
Eye test of Results
The
Supportive Counseling group
decreased number of symptoms
Was this enough of a change to be
significant?
Before and After scores are not
independent; use related-samples t test
17
Step 3
Compute t test for related samples
D D D D 8.22 0 8.22
t
6.85
sD
3.6
1.2
sD
n
9
df = n - 1 = 9 - 1 = 8
18
Step 4
critical value with 8 df, α=.05, twotailed = +2.306
We calculated t = 6.85
Since 6.85 > 2.306, reject H0
Conclude that the mean number of
symptoms after therapy was less than mean
number before therapy.
Supportive counseling seems to work.
The
19
SPSS
Next
slide shows SPSS Printout
• Similar printout from other software
• Results match ours
20
Paired Samples Statistics
Mean
Pair
1
Std.
Deviation
N
Std. Error
Mean
POST
15.6667
9
4.2426
1.4142
PRE
23.8889
9
4.1966
1.3989
Paired Samples Correlations
N
Pair 1
POST & PRE
Correlation
9
.637
Sig.
.065
Paired Samples Test
Paired Differences
POST - PRE
Mean
Std.
Deviation
Std. Error
Mean
-8.2222
3.5978
1.1993
95% Confidence
Interval of the
Difference
Lower
Upper
t
-10.99
-5.46
-6.86
df
Sig.
(2-tailed)
8
.000
21
Magnitude of difference by
computing effect size
Two
methods for
computing effect size
Cohen’s
r2
d
D
Cohen' s _ d
s
2
t
r2 2
t df
22
Review Questions
Why
do we say that the two sets of measures
are not independent?
What are other names for “related samples?”
How do we calculate difference scores?
• What happens if we subtract before from
after instead of after from before?
23
Cont.
Review Questions--cont.
do we usually test H0: D = 0?
Why do we have 8 df in our sample when
we actually have 18 observations?
What are the advantages and disadvantages
of related samples?
Why
24