What you'll learn:
- Understand what Generalized Linear Mixed Models (GLMMs) are
- See how they can be used to fit most other statistical models (T-tests, ANOVA, Regression, etc.)
- Learn how to construct, run, and interpret GLMMs in SAS using PROC GLIMMIX
- Test side-by-side comparisons of statistical models with PROC GLIMMIX
This course will teach you how to use the generalized linear mixed model framework to replicate (and go beyond) the classical statistical tests such as t-test, ANOVA, and regression in SAS.
In the 1st lecture, I start by asking why use generalized linear mixed models (GLMMs)?
Next, I explain how to get and use a free version of SAS, SAS Studio.
The 2nd lecture goes over the details of GLMMs. This includes definitions, model examples, parts, assumptions, and caveats. The section ends with a quick example in SAS.
Lecture #3 covers how to run GLMMs in SAS using PROC GLIMMIX. It starts with an overview of the procedure, then delves into the core statements, other statements, and common statement options.
Lecture #4 focuses on the output table and datasets produced by PROC GLIMMIX. The first half covers table interpretation.
The second half of the course wrangles data outputs to graph bar plots and scatterplots.
Lectures #5-10 go through side-by-side examples of standard statistical tests compared to generalized linear mixed models, using the same datasets.
Model specification and output interpretation is considered for each.
The tests covered are:
Chi-Square Test of Independence
Two-Way T-test
One-Way T-test
Paired T-test
One-Way ANOVA
Two-Way ANOVA
Repeated-Measures ANOVA
ANCOVA
Simple Linear Regression
Multiple Linear Regression
Logistic Regression (Categorical)
Logistic Regression (Numerical)
Poisson Regression
Linear Mixed Models
Generalized Linear Mixed Models
In the 11th and final lecture, I explain what tests you can’t do with a GLMM format, provide next steps to take your learning, and list useful references.
By the end of the course, you will be confidently using GLMMs as a new paradigm for statistical analysis.
