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 well-defined subsets of a population, and random effects, which apply to specific experimental or observational units in the study. The workshop will introduce mixed-effects models and the lme4 package for fitting, analyzing and displaying linear mixed-effects models, generalized linear mixed models and nonlinear mixed models with scalar or vector-valued random effects in nested, crossed or partially crossed configurations. We will use recently developed capabilities in lme4 that allow for hypothesis testing on and interval estimation of the model parameters using profiled likelihood.
The section on fitting linear mixed models includes multilevel and hierarchical linear models as a special case. The methods in lme4 are particularly effective on very large data sets (millions of observations on hundreds of thousands of subjects) and models in which the random effects are not nested.