Categorical Data Analysis Using SAS has been a favorite among SAS Books readers for quite some time, both as a useful discussion of categorical data analysis techniques as well as an invaluable aid in applying these methods with SAS. Maura Stokes is lead author of the recently published third edition. I asked her for three reasons why owners of previous editions might consider updating:
- One big difference between the second and third editions is that the examples have been updated to the most recent release of SAS software. The second edition is over ten years old, and the software has grown considerably in capabilities. In some cases, that means that we rewrote entire chapters to take advantage of new features, such as the STRATA statement in the LOGISTIC procedure, which provides you with conditional logistic regression. SAS/STAT 12.1 was used to run the examples, but only a few features come from that release. I’d say 98% of the examples work fine with SAS/STAT 9.3 (and the great majority work with SAS/STAT 9.2). The few features specific to SAS/STAT 12.1 are spelled out in the Preface.
- Another big difference is that ODS Graphics have been incorporated throughout the book. You’ll see effect plots, odds ratio plots, prediction plots, diagnostic plots, agreement plots, and mosaic plots produced and interpreted. The incorporation of ODS Graphics into SAS/STAT software has been a huge advancement in recent years, and they enrich this book. See the mosaic plot below.
- And finally, this edition reflects the methods used in modern categorical data analysis, including more recent methods developed and/or possible with statistical computing advances. It now includes discussions of exact Poisson regression, incidence density ratios and their confidence intervals, modern confidence intervals for the difference of proportions, the use of the QIC statistic in GEE analysis, and the Firth penalized likelihood approach for logistic regression, for example.
Clearly the third edition offers invaluable insights and updates. "We didn’t just rerun examples and add additional material to the book," Maura added. "We went through it chapter by chapter, consolidating some material and rewriting other material, with the aim of staying true to the original scope of the project, which was to provide a solid treatment of the fundamentals of categorical data analysis illustrated with SAS."