Section 14.9 Two Within-Subjects Variables and One Between-Subjects Variable 491(^6) For those who want to see the calculations, the corresponding pages from the previous edition can be found at
http://www.uvm.edu/~dhowell/method7/More_Stuff/Table14-11Analysis.html.
to perform slightly better (less suppression during Phase II) than did the other group, espe-
cially on the first cycle or two.
Group A-Awas also treated in the same way as Group A-Bexcept that both Phases were
carried out in the same box—Box A. Because there were no differences in the test boxes to
serve as cues (i.e., animals had no way to distinguish the no-shock from the Shock Phases),
this group would be expected to show the most suppression during the No-shock Phases.
Bouton and Swartzentruber predicted that overall there would be a main effect due to
Phase (i.e., a difference between shock and No-shock Phases), a main effect due to Groups
(A-Band L-A-Bshowing less suppression than A-A), and a main effect due to Cycles (ani-
mals tested in Box Bwould learn over time that it was a safe location). They also predicted
that each of the interactions would be significant. (One reason I chose to use this example,
even though it is difficult to describe concisely, is that it is one of those rare studies in
which all effects are predicted to be significant and meaningful.)
The data and analysis of variance for this study are presented in Table 14.11. The analy-
sis has not been elaborated in detail because it mainly involves steps that you already know
how to do. The results are presented graphically in Figure 14.4 for convenience, and for the
most part they are clear-cut and in the predicted direction. Keep in mind that for these data a
lower score represents more suppression—that is, the animals are responding more slowly.
Rather than present literally three pages of tables and calculations, which few people
would have the patience to work through, I have chosen to carry out the analysis using
SPSS.^6 The data would be entered just as they appear in Table 14.11, with a column for
Groups on the left. You would selectAnalyze, General Linear Model, Repeated Mea-
suresfrom the drop-down menus and specify that there were two repeated measures
(Cycles with 4 levels and Phases with 2 levels). Then click on Defineand specify the vari-
ables that are associated with each of the cells and the variable(s) that define the Between-
Subject Factor(s). This dialogue box follows, where C1P1 – C4P2 would be moved to
the Within-Subjects Variables box and Group would be moved to the Between-Subjects
Factor(s) box.
55
50
L A B
45
40
35
30
25
20
15
10
Suppression ratio
Shock
Cycle 1 2 3 4
A A
A B
Figure 14.4 Conditioned suppression data