I ran an factorial ANOVA in SPSS (1 dependent, 2 independent variables). One of the factors was significantly related to the dependent variable and there was a significant interaction with the other variable. Surprisingly most post-hoc tests (except least significant difference) were not significant for the variable that had p