For example, in an experiment the output data (y axis) becomes more dispersed as some input variable (x axis) is changed. So if there seems to be a relationship between variance (or standard deviation) and the input... is variance an acceptable predictor of the input?
I am new using R and I need a model to fit both the mean and the SD of the model (similar to the Double hierarchical generalized linear models). Is it possible to perform using the MCMCglmm package? I already have a model for the mean, but I couldn't manage to module the SD.