Subscribe Bookmark RSS Feed

Degrees of freedom in post-hoc t-tests


Community Member


Feb 23, 2015

Hi, I have a question regarding the post-hoc t-tests in JMP. I run a repeated measures ANOVA with one random and two fixed factors. The two-way interaction between the fixed factors is included. (Y=random(A) + A + B + A*B). Following a significant A*B effect, I use the 'effect details' option to do post-hoc t-tests, Bonferroni-corrected for the number of comparisons. But. When I chose 'detailed comparison' in the t-test, to check the t-value and degrees of freedom, I get degrees of freedom that are a lot higher than i would expect, and that also vary from one Y variable to another, EVEN when I have the same number of independent datapoints for each Y. Has anyone run into this and knows how the dfs are calculated for these tests? Help much appreciated!