SAS software can fit many different kinds of regression models. In fact a common question on the SAS Support Communities is "how do I fit a <*name*> regression model in SAS?"
And within that category,
the most frequent questions involve how to fit various logistic regression models in SAS.

There are many types of logistic regression models:

- The traditional logistic model has a binary (or binomial) response variable.
- Well-known extensions of the logistic model include ordinal regression (for an ordinal response) and multinomial regression (for a discrete but unordered response).
- Special models handle situations such as repeated measures (longitudinal data) or random effects.
- Other models include conditional logistic regression, survey logistic regression, Bayesian logistic regression, and fractional logistic regression (which gets the Coolest Name Award).

If you know the procedure that you want to use, you can read the procedure documentation and follow the examples to perform the analyses. In practice, however, you might know the reverse information: You know the analysis that you want to perform, but you do not know which SAS procedure implements that regression model.

I was therefore pleased to discover a short SAS Knowledge Base article titled "Types of logistic (or logit) models that can be fit using SAS." This article provides a "reverse look-up": Given the type of logistic regression model that you want to fit, it provides the name of the SAS procedures that support that analysis!

*Logistic models that can be fit with SAS: A reverse look-up resource. #SAStip*

Click To Tweet

Use this resource the next time you need to know which SAS procedure can conduct a certain variant of logistic regression. Bookmark it, print it out, tattoo it on your forearm, or come to my blog and type "HOW TO FIT A LOGISTIC MODEL" into the search box. But don't forget it! This resource is a treasure for the SAS statistical programmer.

The post How to fit a variety of logistic regression models in SAS appeared first on The DO Loop.