Simple main effects analysis is 'a priori', post hoc is 'a posteriori'. In the latter case you correct for the many tests, e.g. with Bonferroni correction. In the first case you do not correct for the many tests.

Simantini Ghosh, a simple main effect in regression analysis refers to the coefficient of Y (e.g. wage) on X1 (e.g. educational level) at a specific value (e.g. female) of X2 (i.e. gender) making up the interaction term (X1*X2).

Hello Simantini,

Yes, you could apply a traditional, pairwise, post hoc comparison approach to the cells involved in a significant interaction. The advantage of going with simple effects, though, is that it stays with comparisons that are likely to be consonant with your original research question: (e.g., are there differences across levels of factor A with respect to outcomes on the DV?).

It's your choice, and your obligation to explain your analytic scheme to your audience.

Good luck with your work.

Simple main effects analysis is 'a priori', post hoc is 'a posteriori'. In the latter case you correct for the many tests, e.g. with Bonferroni correction. In the first case you do not correct for the many tests.

If you did predict any simple effect, use the (a priori) simple main effects analysis, if you did not know in advance where there could be significant effects, use the post hoc approach. Correct the p-values with e.g. the Holm correction (Bonferroni is too conservative).