www.cmm.bristol.ac.uk/forum

Indeed! Thx!

It is an assumption that there must be homogeneity.
Is it true that for multi level analysis not having homogeneity is not so bad (as compared to OLS)?
I think it's just as bad, but perhaps with the automatic shrinkage factor, it will be OK anyway?

I work with the tutorial database.
I want to do a 2 level model.
I have done following steps:
- open worksheet
- toggle cat
- sort data
- model equations
- y: 2 levels (school = 2, student = 1)
- add cons (also 2 levels)
- estimates
- start


But my variation on level 2 stays at 0. Can anyone help ...