3.6. Goodness-of-Fit Statistics

The LOGISTIC and GENMOD procedures produce many different statistics that can help you evaluate your models. We’ve already examined some of the default statistics, notably the global chi-square tests produced by LOGISTIC (Section 2.7) and the deviance chi-square produced by GENMOD (Section 2.8). In this section, we’ll look at some of the optional statistics.

It’s essential to keep in mind that there are many ways of approaching model evaluation, and statistics that appear to be similar may actually answer quite different questions. For example, LOGISTIC’s global chi-square addresses the question “Is this model better than nothing?” A significant chi-square signals a “yes” answer, suggesting that the model is acceptable. ...

Get Logistic Regression Using SAS®: Theory and Application now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.