Simple Mixed Design - Analysis of Variance Designs - page 388

Figure 13.32 Specifying the post hoc tests.

13.14 SAS OUTPUT FROM THE POST HOC ANALYSIS

The pairwise mean comparisons together with the least squares means
are shown in Figure 13.33. As we indicated in discussing the SPSS output,
this main effect is completely subsumed by the interaction effect and will
therefore not be described.

Least Squares Means

Standard
Error Effect

leadstyl
leadstyl
leadstyl

leadstyl Estimate DF t Value Pr > ltl

1
2
3

3.5000
6.6000
3.2000

0.3536
0.3536
0.3536

12
12
12

9.90
18.67
9.05

<.0001
< .0001

<.0001

These are the mean
differences for the
different levels of
leadstyl.

Differences of Least Squares Means

Standard
Error Effect

leadstyl
leadstyl
leadstyl

leadstyl
1
1
2

leadstyl
2
3
3

Estimate DF t Value Pr > ltl Adjustment Adj P

− 3.1000

0.3000
3.4000

0.5000
0.5000
0.5000

12
12
12

− 6.20

0.60
6.80

< .0001

0.5597
<.0001

Bonferroni
Bonferroni
Bonferroni

0.0001
1.0000
<.0001

This Table presents
both the
unadjusted p -value
and the Bonferroni
adjustment.

b

Figure 13.33 A portion of the output from the post hoc tests.