Mixed Models Analyses Using SAS

Classroom duration: 3.0 days Live Web duration: 6 half-day sessions CEU: 1.8

This course teaches you how to analyze linear mixed models using PROC MIXED. A brief introduction to analyzing generalized linear mixed models using PROC GLIMMIX is also included.

Learn how to

analyze data (including binary data) with random effects
fit random coefficient models and hierarchical linear models
analyze repeated measures data
obtain and interpret the best linear unbiased predictions
perform residual and influence diagnostic anlaysis
deal with convergence issues.

Who should attend

Statisticians, experienced data analysts, and researchers with sound statistical knowledge

Prerequisites

Before attending this course, you should

know how to create and manage SAS data sets
have experience performing analysis of variance using the GLM procedure of SAS/STAT software
have completed and mastered the Statistics II: ANOVA and Regression course or completed a graduate-level course on general linear models.

Exposure to matrix algebra will enhance your understanding of the material. Some experience manipulating SAS data sets and producing graphs using SAS/GRAPH software is also recommended.

Course Contents

Introduction to Mixed Models

identifying fixed and random effects
describing linear mixed model equations and assumptions
fitting a linear mixed model for a randomized complete block design using the MIXED procedure
writing CONTRAST and ESTIMATE statements to perform custom hypothesis tests

Examples of Mixed Models in Some Designed Experiments

fitting a linear mixed model for two-way mixed models
fitting a linear mixed model for nested mixed models
fitting a linear mixed model for split-plot designs
fitting a linear mixed model for crossover designs

Examples of Mixed Models with Covariates

fitting analysis of covariance models with random effects
performing random coefficient regression analysis
conducting hierarchical linear modeling

Best Linear Unbiased Prediction

explaining BLUPs and EBLUPs
producing parameter estimates associated with the fixed effects and random effects
explaining the difference between LSMEANS and EBLUPs
computing LSMEANS and EBLUPs using the MIXED procedure

Repeated Measures Analysis

discussing issues on repeated measures analysis, including modeling covariance structure
analyzing repeated measures data using the four-step process with the MIXED procedure

Mixed Models Residual Diagnostics and Troubleshooting

performing residual and influence diagnostics for linear mixed models
troubleshooting convergence problems

Additional Information about Linear Mixed Models (Self-Study)

discussing issues associated with unbalanced data, data with empty cells, estimation and inference of variance parameters, and different denominator degrees of freedom estimation methods

Introduction to Generalized Linear Mixed Models and Nonlinear Mixed Models

discussing the situations where generalized linear mixed models and nonlinear mixed models analysis are needed
performing the analysis for generalized linear mixed models using the GLIMMIX procedure

Software Addressed

This course addresses the following software product(s): SAS/STAT.

Classroom Course Materials

Students receive a hardcopy of the course notes and, in some courses, can choose to take home a copy of the course data.

Live Web Course Materials

Students attend Live Web classes using a Web browser and a telephone and interact with their instructor and fellow classmates in real time. Each student receives an e-mail with instructions on how to join the class three business days before the class begins. The instructions e-mail includes a link to download the course materials, including the exercise files. Students need to download and print the course materials prior to class.

System Requirements

For Live Web, you must

review and follow the general system requirements.
complete the course exercises through our virtual lab. The virtual lab allows you to access the software used in class over the Internet, so that you do not need this software on your local machine.
run this test to connect to a virtual lab session. If firewall problems prevent you from connecting to the virtual lab, then you will need the following software installed and configured in your environment to participate in the course exercises:
- Base SAS, SAS/STAT, and SAS/GRAPH software, Release 9.2 or 9.1.3 on a Windows operating system

U.S. Schedule

18MAR2009 SAS Global Forum-DC,

Check for additional and updated schedule information online at support.sas.com/courses/aglm92.html.

Registration

To register for this course in the US, call 800-333-7660 or visit support.sas.com/training. This course is also available for on-site training, or you can create a custom course by combining material from several courses. For more details, contact SAS Education in Cary, NC at 919-531-7321 or send e-mail to training@sas.com.