Hi Everyone, my outcome is a number between 0 and 1. I have a fair number of 0s and 1s. I would like to fit a model using 5 binary predictors. I have 6 repeated measurements from each respondent, each with a different set of covariates. I've done a logit transform on my outcome. I've subtracted and added 1E-16 to the 1's and 0's, respectively, to make this work. I want to implement this in MCMCglmm. I just wanted to check: a) I can use a gaussian family because I have already done the transform b) Is R=list(V=1,fix=1) appropriate in this case c) when I make predictions, I have to add 1 to linear predictor? Thanks Mark [[alternative HTML version deleted]]