1. Chapter 11: Bivariate Statistics and Statistical Inference
“Figures don’t lie, but liars figure.”
Chapter 11: Bivariate Statistics and Statistical Inference
Key Concepts: Testing Group Differences 1 1

2. Testing Mean Differences
Independent sample t-test
Do two independent samples belong to the same population?
E.g., Is there a difference in satisfaction between in-person and telephone interviews?
Conclude, NO, those interviewed in-person have equal satisfaction as those interviewed by phone.
t(340) 1.338, p<.18 2 2

3. t-tests (con’t.) Paired sample t-test
Are dependent samples different over time?
E.g., are computer skills different for foster youth from pretest to posttest?
Pre
Post
Yes, skills have improved. Sample error is unlikely. 3
t(36) , p=.00 3

4. t-tests (con’t.) One sample t-test
Is a sample mean different from a population mean?
E.g., Does this group differ in IQ scores from national norms ? (Test value = 100)
t(9) 1.847, p=.10
We conclude: No Difference in IQ. 4 4

5. ANOVA Analysis of Variance
Do two or more means differ? Usually used with 3 or more means.
Do attitudes about leadership style differ based on agency role?
Administrators, Supervisors, Direct Service Workers
We conclude: YES, the groups differ.
F (3,294) 4.051, p=.01 5 5

6. Post hoc Analysis But exactly which groups differ?
There are several post hoc analysis tests.
Direct Service workers differ (lower on the scale) from Supervisors and Administrators but not “Others”. There are no other significant differences. 6 6