Tutorial: Mixed models in R using the lme4 package

Douglas Bates, Department of Statistics, University of Wisconsin - Madison

Abstract

Mixed-effects models or, more simply, mixed models are statistical models that incorporate both fixed-effects parameters, which apply to an entire population or to certain well-defined subsets of a population, and random effects, which apply to specific experimental or observational units in the study. Certain types of mixed models, called hierarchical linear models (HLMs) or multilevel models, have been widely used in the social sciences but these are only a subset of the possible linear mixed models (LMMs) for data on a continuous scale or generalized linear mixed models (GLMMs) for binary data or count data. Recent developments in computational methods for mixed models, embodied in the lme4 package for R, allow for fitting these and more general forms of mixed models, including models with crossed random effects such as random effects for both subject and stimulus. The workshop will introduce mixed-effects models and the lme4 package for fitting, analyzing and displaying such models.

Outline

The following topics will be covered.

1) Organizing and displaying data collected on multiple experimental or observational units.
2) Models with simple, scalar random effects
3) Models for longitudinal data
4) Generalized linear mixed models (GLMMs)
5) Theory and computational methods
6) Nonlinear mixed models (NLMMs)
7) Item response models in a mixed model framework

Potential attendees

The workshop is aimed as users of R wishing to fit and analyze mixed-effects models.

Elementary knowledge of statistical concepts at the level of a ﬁrst course in biostatistics is assumed. This tutorial is intended to appeal to public health and medical researchers involved in genetic investigations, as well as biologists, statisticians and computer scientists with interests in bioinformatic tools. Topics will extend coverage in UseR!2008 tutorial.