I'm trying to run a two-way RM ANOVA. I have two groups that received different treatments over the course of 6 days; n=10 in each group so 20 subjects total. I have a significant interaction effect. When I run the post-hoc tests I'm a little bit confused by the degrees of freedom used in the calculation; for timepoint * group each session has a df of 18. I thought that in the post-hoc test the pooled error term is used and therefore the dfs is (n-1)(a-1)? Any guidance would be very apprecaited! I'm new to statistics. post_hoc = pg.pairwise_tests( data=long_df, dv='score', within='timepoint', between='group', subject='subject', padjust='bonf', ) Continue reading...
