Further Statistical Methods - Lecture 8
Multilevel Analysis
Multilevel analysis is concerned with observations in a nested structure, for example on individuals within classes within schools within geographical region, corresponding to four levels of nesting.
T. A. B. Snijders and R. J. Bosker (1999). Multilevel Analysis. Sage Publications, London. is an excellent source. The material for this lecture is contained in chapters 1, 2 and 4.
An alternative way of analysing multilevel data is through Bayesian hiearchical models. These are well described as directed graphical models and analysed using WinBUGS. The WinBUGS manual is a good place to read about this. For this particular lecture, look at the examples: Rats, Dyes, and Schools.
Overheads
Links
Last updated: Tuesday, 20 February 2007 09:48. Steffen L. Lauritzen