Information Technology Reference
In-Depth Information
Figure 12.27
Specifying the effects for which we wish to obtain least square means.
12.13 SAS OUTPUT FROM THE SIMPLE EFFECTS ANALYSIS
The results of the simple effects tests are shown in Figure 12.31 and are
read in a manner analogous to that described for the two-way interaction
discussed in Section 11.14. Many of the comparisons are statistically
significant, despite the large number of paired comparisons that were
performed (and hence the substantial Bonferroni correction used by
SAS
Enterprise Guide
). Generally, these comparisons yield similar conclusions
to those we drew from the SPSS analysis.
Consider, for example, the second comparison in the table since this
isthecomparisonweillustratedintheSPSSoutputforthesimpleeffects
tests. The first three coded columns represent one condition and the last
three coded columns represent the other condition. We are thus comparing
the mean for
powcolor1, granules0,boxcolor1
(the white powder with no
granules in the aqua box with a mean of 2.7) with
powcolor 1
,
granules 1
,
boxcolor 1
(the white powder with granules in the aqua box with a mean
of 6.5). The
t
value for that comparison is
−
9.71, which is statistically
significant (
p
<.
0001). We may therefore conclude that the participants
