Department Seminars

General Information

When: Thursdays from 4:00–4:50 p.m.
Where: 1170 TMCB

Simon Sheather, Chair
Texas A&M University
Department of Statistics

2007-11-05

Diagnostics for linear mixed models

As the name suggests, a linear mixed model is one which consists of linear fixed effects,
random effects and an error structure. Linear mixed models are widely used in practice.
In this talk we will examine the following issues surrounding linear mixed models:

• In what circumstances least squares (LS) residuals provide direct information on
  how the model for the fixed effects is misspecified.
• How, and why, to transform a linear model with a general error structure into a
  model with an iid error structure
• The different types of residuals in mixed models and their uses
• SAS diagnostics for fixed effects and covariance parameters in mixed models

Real data sets will be used to illustrate the points being made.