My repeted factor represents an effect of an antipsychotic drug on some quantitative variable measured in 5 time points. When that is the only factor in the analysis I have significant differences between time points. I have included gender, age, drug doze and years of treatment in the model. Now I don't have the effect of repeated factor (in the within subjects box in SPSS output). So it could be that those other factors explained some variance so the effect of the repeted factor is no longer significant. But there is a problem. I still have significant post hoc tests (Sidak) for the repeated factor. So in the second model the differences are significant and not significant?Any idea how this can happen and how to interpret that?