Hi,

I'm running a linear mixed model with two categorical fixed effects, two continuous effect and one random nested effect. Inside the model I am running a multiple comparisons test for the interaction between the categorical fixed effects. When I get the results for the estimated means from the comparison the degrees of freedom are an order of magnitude higher for one term than the other three. What is especially weird is that the DF for the interaction term that seems off is higher than the number of observations in that group. The random effect in the model is individual and in the case of this interaction term there is only one individual in the model. Can anyone explain how Multiple comparisons calculates DF and if there is something wrong with my model?

Thanks

Here is the means table, the interaction between GT*CST only has 23 observations, the others have over 100.