Two Sample TTests

1
Two Sample T-Tests

• Data from Wuensch, K. L., Castellow, W. A.,
  Moore, C. H. (1991)

2
Random Sample of 10 Cases

• Wuensch, K. L., Castellow, W. A., Moore, C. H.
  (1991). Effects of defendant attractiveness and
  type of crime on juridic judgment. Journal of
  Social Behavior and Personality, 6, 713-724.
• Compare rated happiness of physically attractive
  defendant to that of not physically attractive
  defendant.

3
Matched Pairs
The data were actually independent samples. I
created matched pairs for pedagogical purposes.
4
The Hypotheses

• H0 µA µB (µA - µB) 0 µDiff 0
• H1 µA ? µB (µA - µB) ? 0 µDiff ? 0
• Evaluate with the .05 criterion of statistical
  significance.

5
Compute Correlated t
6
Obtain p

• df n - 1 4
• COMPUTE p2CDF.T(-5.099,4).EXECUTE.
• p .007
• The rated happiness of the physically attractive
  defendants was significantly greater than that of
  the physically unattractive defendants.

7
Obtain Confidence Interval
8
Estimate Cohens d
9
SPSS Output
10
(No Transcript)
11
Independent Samples
12
Standard Error of Difference Between Means
13
T
14
p

• COMPUTE P2CDF.T(-2.213,8).EXECUTE.
• df N 2 8 p .0578
• The effect of the defendants physical
  attractiveness on rated happiness fell short of
  statistical significance.
• Note The independent samples analysis had less
  power that the correlated samples analysis.

15
Confidence Interval
16
Estimated Cohens d

• Same as with correlated samples, but now we can
  compute a confidence interval for the
  standardized effect size.

17
SPSS Output