Correlated Samples, SPSS

Matched Pairs t Tests
Source of the Data
Cognitive-behavioral intervention to
reduce African American adolescents' risk
for HIV infection
Group EC was given a standard 2-hour
educational program
Group BST went through an 8 week
behavioral skills training program.
The Dependent Variables
• The log-transformed frequency of condomprotected sexual intercourse
• Measured before the intervention, at
posttest (8 weeks later), 6 weeks after
posttest, and 12 weeks after posttest.
A Plot of the Means
Download the Data
Bring the data into SPSS
We want to do two matched pairs t tests
One for the EC group
One for the BST group
Split the File by Group
Data, Split File
Conduct the Test
• Analyze, Compare Means, Paired
Samples T Test, Pretest versus fu6
Obtain Estimates of Cohen’s d
• Download the .sav and .sps files
D-EqN-- Estimate Cohen's d, equal n data
(including correlated samples)
d-EqN-Doc -- instructions
d-EqN-Dat -- necessary data file
d-EqN-Run -- syntax file
Enter Statistics Into .sav File
Run the Syntax File
Look Back at .sav File
Summary Statement, BST
• Adolescents in the behavioral skills
training group (n = 20) engaged in
significantly more condom-protected
intercourse six weeks after the training (M
= 18.80, s = 11.04) than prior to the
training (M = 13.45, s = 13.77), t(19) =
2.35, p = .030, d = .43.
Summary Statement, EC
• Adolescents in the educational control
group (n = 20) engaged in significantly
less condom-protected intercourse at the
six week followup (M = 8.60, s = 11.09)
than at pretest (M = 19.8, s = 21.87), t(19)
= 2.40, p = .027, d = .65.

similar documents