Logistic Regression¶

A generalized linear model (GLM) is a generalization of ordinary linear regression to the case where the response variable’s error is best modeled by a non-normal distribution.

Stat-Ease 360 currently supports logistic regression for ungrouped binary response data consisting of two numeric values. A response consisting of pass/fail data coded to 1s (pass) and 0s (fail) respectively is an example of a candidate for logistic regression analysis.

The logistic regression analysis option for this type of GLM may be selected on the Transform tab. The level that codes to 1, the success level, may also be chosen on the Transform tab. This is most useful when your data consists of two values different from 0 and 1 or when you wish to reverse the coding order.