I think that Dennis is referring to the fact that it can be the case
that the "simple main effects" of B may be in the same direction at all
levels of A (what Campbell and Stanley call a "monotonic interaction") but
the simple main effects of A be in different directions at different
levels of B (nonmonotonic interaction), even in a simple 2 x 2 factorial
design. Consider the following table of means for such a design:
B1 B2
A1 5 20
A2 10 15
At both levels of A, when B changes from level 1 to level 2, the mean
goes up, a "monotonic" interaction.
However, from the other perspective, this interaction is "nonmonotonic":
Changing A from level 1 to level 2 is associated with an increase in the
mean if we look only at the B1 column, but a decrease in the mean if we look
only at the B2 column.
Karl L. Wuensch, Professor, Graduate Faculty,
Director of Psychology/Social Work Computer Labs
Dept. of Psychology, East Carolina Univ.
Greenville, NC 27858-4353, phone 252-328-4102, fax 252-328-6283
Bitnet Address: PSWUENSC@ECUVM1
Internet Address: [EMAIL PROTECTED]
Web Address: http://core.ecu.edu/psyc/wuenschk/klw.htm
