Why do I get different t-values for paired samples t-test compared to post-hoc RM-ANOVA?

roundcircle

Using JASP, I ran a RM-ANOVA and performed post-hoc tests giving me t-values and corrected p-values (bonferroni). It also gives me a Cohen's d (uncorrected) for which I wanted to get 95% CIs. I did not see this option in the post-hoc tests, so I thought I would just do paired samples t-tests equivalent to my post-hoc tests and get the Chohen's d 95% CIs from this test. However, t-values and thus Cohen's d are different when testing with the paired samples t-test compared to post-hoc tests. I assumed that post-hoc with bonferroni only corrected p-values, but this is apparently not the case? Why do I get different t-values? And is there a way to get Chohen's d 95% CIs of my post hoc tests?

Thanks in advance!

JohnnyB

Hi @roundcircle ,

This is probably due to the post hoc tests using the pooled error term, whereas the paired t-tests use only the error term of the two groups under consideration. If you untick this option in the RM ANOVA post hoc tests, the values should correspond to the paired t-test. I would advise to use the pooled error term though, since this corresponds to the actual ANOVA model (which assumes equal variances).

Cheers

Johnny

roundcircle

Great, thanks @JohnnyB! Would you have any idea how then to get the 95% CIs for Chohen's d?