Return to 440 problems & solutions page
Question #81
This is another example of extreme interaction effect, using extreme interaction.....
The simple question about whether the drugs have the same effect (regardless of gender) would show almost no difference.
Similarly, the question about whether the gender produce the same effect (regardless of the drug being used) would show no effect at all.
So if we consider JUST the two MAIN effects SEPARATELY, both null hypotheses would NOT be rejected and no effects would be noticed. However, the effects ARE powerful but with an opposite effect within each level of the other variable, and they were being cancelled out by the interaction. The interaction can be seen to be very strong in the ANOVA table below It also demonstrates, that IF interaction is present, the effects of the two independent variables have to be described separately within the levels of the OTHER indep var.
