A First Step toward a Unified Theory of Richly Parameterized Linear ModelsUsing mixed linear models to analyze data often leads to results that are mysterious, inconvenient, or wrong. Further compounding the problem, statisticians lack a cohesive resource to acquire a systematic, theory-based understanding of models with random effects.Richly Param

James S. Hodges

Using mixed linear models to analyze data often leads to results that are mysterious, inconvenient, or wrong. Further compounding the problem, statisticians lack a cohesive resource to acquire a systematic, theory-based understanding of models with random effects. This book takes a first step in developing a full theory of richly parameterized models, which would allow statisticians to better understand their analysis results. It examines what is known and unknown about mixed linear models and identifies research opportunities. Supporting materials are available on the author's website.

Mixed Linear Models: Syntax, Theory, and Methods: An Opinionated Survey of Methods for Mixed Linear Models. Two More Tools: Alternative Formulation, Measures of Complexity. Richly Parameterized Models as Mixed Linear Models: Penalized Splines as Mixed Linear Models. Additive Models and Models with Interactions. Spatial Models as Mixed Linear Models. Time-Series Models as Mixed Linear Models. Two Other Syntaxes for Richly Parameterized Models. From Linear Models to Richly Parameterized Models: Mean Structure: Adapting Diagnostics from Linear Models. Puzzles from Analyzing Real Datasets. A Random Effect Competing with a Fixed Effect. Differential Shrinkage. Competition between Random Effects. Random Effects Old and New. Beyond Linear Models: Variance Structure: Mysterious, Inconvenient, or Wrong Results from Real Datasets. Re-Expressing the Restricted Likelihood: Two-Variance Models. Exploring the Restricted Likelihood for Two-Variance Models. Extending the Re-Expressed Restricted Likelihood. Zero Variance Estimates. Multiple Maxima in the Restricted Likelihood and Posterior.