Is it possible in MLWin to see what the Standard Deviation / Variance is of 1 explanatory variable once it has been layered/grouped? I am trying to tackle what I believe is a "repeated measures" structure but would rather remove the layer if of little significance. As an example, say I have a dataset of student marks for 3 tests they perform at various times in one year, I could get the overall SD of the marks or (probably more accurately) the SD from a layered structure i.e., layer 1 = mark, layer2 = student. I am not trying to create a regression/predictive model, rather just trying to see what the difference are for the SD of an "overall" mark versus the SD of a "layered" mark.
I am a complete "newbie" in this type of modelling so would appreciate any help. If MLWin is incapable of this, perhaps somebody has a formula(s) / can help redirect me to relevant research so I can figure out how to calculate this myself?