This paper provides a general overview of the mixed model, a powerful tool for analyzing correlated data. Numerous books and other sources exist that cover the mixed model comprehensively. However, we aimed to provide a relatively concise introduction to the mixed model and describe the primary motivations behind its use. Recent developments of various aspects of this topic are discussed, including estimation and inference, model selection, diagnostics, missing data, and power and sample size. We focus on describing the mixed model as it is used for modeling normal outcome data linearly, but we also discuss its use in other situations, such as with discrete outcome data. We point out various software packages with the capability of fitting mixed models, and most importantly, we highlight many important articles and books for those who wish to pursue this topic further. © 2008 Elsevier B.V. All rights reserved.
