AN INTRODUCTION TO MIXED LINEAR MODELS

IS THE LINEAR MODELS PROCEDURE GLM OBSOLETE?

When fitting linear statistical models, both SAS and SPSS (and other statistical programs) include a procedure called GLM (General Linear Models). It has been one of the most important analysis routines of these systems for nearly three decades.

Examples of ANOVA-like computations with a relatively new SAS and SPSS procedure called MIXED (often in comparison with GLM) will be considered in the following sections:

• 1. ANOVA Ideal Conditions
• 2. Introducing PROC MIXED
• 3. Data Structure for PROC MIXED
• 4. Why Switch to PROC MIXED?
• 5. Syntax for MIXED
• 6a. Fixed Effects ANOVA
• 6a. Contrasts and Estimates in ANOVA
• 7a. Computing Means and Variances from Clustered Data
• 7b. Repeated Measures ANOVA
• 7c. Multivariate Analysis
• 8. Understanding the Correlation Structure
• 9. Random Effects in ANOVA Models
• 10. The Unequal Variance ANOVA Model
• 11. PROC MIXED and the Output Delivery System
• 14. PROC MIXED as a Study Planning Tool