Hi. Being a student of statistical research methods, I’m looking for some good study material to help me understand covariance structures and their use in multilevel modeling. Any suggestions?
what covariance is important? When covariance in multilevel analysis is relevant to you, you can check an MLA book such as the 2nd edition by Snijder and Boskers.
That is my question exactly! I have worked through the chapter in DSUS bij Andy Field on multilevel modelling. It had a short piece on the different kinds of covariance structures to be used. It isn’t clear to me what shoud co-vary with what and so on and how it influences the outcome of the analysis.
Thanks for the suggestion anyway! I’ll look it up.
Covariance structures are involved in multilevel modeling precisely because you model the data-generating process on multiple levels. But Snijder/Boskers should help (or, more advanced and for Stata users http://stata-press.com/books/multilevel-longitudinal-modeling-stata/)
So, if I understand you correctly, the question is basically about: how does the covariance between X and Y on one level compare to the covariance between the same variables on a different level?