Multilevel Models
An Introduction and FAQs
Note that some of these materials were written more recently than others. We believe that the older materials, marked with this symbol - 
, are still valuable and that the majority of what they contain is still true, but note that for example they may refer to an earlier version of MLwiN or other software, and they may not reflect the most up-to-date methodological developments.
- What do multilevel models do and why should I use them?
- What are the data structures that multilevel models can handle?
- What type of model can be fitted?
- What size of sample is best?
-
Introduction to Multilevel Modelling and MLwiN by Wen-Jung Peng - Various video presentations (eg,' What is it and why you should do it', 'The sorts of analyses you can do and the results you can get', etc.
- Introduction to multilevel modelling - workshop materials
Do multilevel models ever give different results? by Kelvyn Jones It is sometimes said that the use of multilevel models over OLS regression makes no substantive difference to interpretation and represents something of a fuss over nothing. This short paper demonstrates with a simple example that this is not always the case.- What substantive issues have been tackled by multilevel modelling?
- What software is available for fitting the models?
- Partitioning variation across levels
- What is the intra cluster correlation?
- Differential weightings
- Sandwich estimators for standard errors
Other terms used for Multilevel Modelling
- Bayesian hierarchical models
- mixed models
- hierarchical linear models
- hierarchical modelling
- random effects models
- random coefficient models
- subject specific models
- variance component models
- variance heterogeneity