Longitudinal Data Analysis with Discrete and Continuous Responses
This course is for scientists and analysts who want to analyze observational data collected over time. It is not for SAS users who have collected data in a complicated experimental design; they should take the Mixed Models Analyses Using the SAS System course instead.
Learn how to
- create individual and group profile plots and sample variograms
- use PROC MIXED to fit a general linear mixed model and a random coefficient model
- plot information criteria for models with selected covariance structures
- generate diagnostic plots in PROC MIXED
- fit a binary or ordinal GEE model in PROC GENMOD
- compute GEE deletion diagnostic statistics in PROC GENMOD
- fit a generalized linear mixed model in PROC GLIMMIX.
Who should attend
Epidemiologists, social scientists, physical scientists, and business analysts
Duration: 3 days
Before attending this course, you should be able to
- execute SAS programs and create SAS data sets. You can gain this experience by completing the SAS Programming 1: Essentials course.
- fit models using the GLM and REG procedures in SAS/STAT software. You can gain this experience by completing the Statistics II: ANOVA and Regression course.
Longitudinal Data Analysis Concepts
- understanding the merits and analytical problems associated with longitudinal data analysis
- examining the differences between PROC GLM and PROC MIXED
Exploratory Data Analysis
- graphing individual and group profiles
- identifying cross-sectional and longitudinal patterns
General Linear Mixed Model
- understanding the concepts behind the linear mixed model
- examining the different covariance structures available in PROC MIXED
- fitting a general linear mixed model in PROC MIXED
Evaluating Covariance Structures
- creating a sample variogram that illustrates the error components in your model
- plotting information criteria for models with selected covariance structures
Model Development, Interpretation, and Assessment
- learning the model building strategies in PROC MIXED
- creating interaction plots
- specifying heterogeneity in the covariance structure
- computing predictions using EBLUPs
- fitting a random coefficient model in PROC MIXED
- generating diagnostic plots in PROC MIXED using ODS Graphics
Longitudinal Models with Binary Responses
- examining the correlation structures available in PROC GENMOD
- fitting a binary GEE model in PROC GENMOD
Longitudinal Models with Ordinal Responses
- examining cumulative logit plots
- fitting an ordinal GEE model in PROC GENMOD
GEE Model Diagnostics
- computing GEE deletion diagnostic statistics in PROC GENMOD
Generalized Linear Mixed Models
- fitting a longitudinal model with a discrete response using the GLIMMIX procedure
Students have unlimited 24/7 access to the interactive multimedia course on the Web for one year.
The course requires only a standard Web browser and a free Flash plug-in. The course includes data
that can be downloaded, printable lesson summaries, a Quick Reference Guide, practices and solutions,
step-by-step demonstrations, and a Certificate of Completion.
This course addresses SAS/GRAPH, SAS/STAT.
You should
- use Internet Explorer 6.0 or later or Firefox 1.5 or later, with JavaScript and cookies enabled
- have the latest version of Macromedia Flash Player installed
- have base SAS software visible on the same machine that you are taking the training on so that you can practice your new skills in your own SAS environment.
Back to SAS Global Forum Training