How can I assess whether one mean (from a treatment group) is significantly different from another when running a GLM in JMP 13? I know how to do that in the other model platforms (REML, SLSs), but I can't seem to find that option when performing GLMs. For example, I have a GLM model where the response variable is a logit-transformed proportion value (Y) and all effects are categorical (treatment groups) that include main and interactive effects.
Y ~ A, B, C, A*B, A*C, B*C
How can I figure out if the mean from one group (say, between A’s in treatment C)? Can I use the profiler to see if, when I change the values, the mean lies outside of the error bars? Thank you!
