Gabriel Go.Es. wrote:

Dear Doug,

In the case of a 3x2 design the post hoc contrast for evaluating the interaction effect would be as follows: D1Male D1Female D2Male D2Female D3Male D3Female

ANOVA to test the effect of diagnosis (D1, D2, D3) only for males [ 1 0 -1 0 0 0]/2 [ 1 0 0 0 -1 0]/2

Correct (don't divide by 2)

ANOVA to test the effect of diagnosis (D1, D2, D3) only for females [ 0 1 0 -1 0 0]/2 [ 0 1 0 0 0 -1]/2

Correct (don't divide by 2)

Is this correct?

Many thanks again. Gabriel