Analysis of a Binary Outcome Variable Using the FREQ and the LOGISTIC Procedures

ABSTRACT A common application in the health care industry is to investigate the association between a response variable interest that has dichotomous outcome such as having or not having a certain disease with one or more variables. This type of study can be analyzed by using the FREQ procedure if there are only one or two explanatory categorical variables. A more general approach to study a binary outcome variable would be building a logistic regression model by using the LOGISTIC procedure which can handle one or more categorical or continuous independent variables. In this talk, in addition to reviewing both PROC FREQ and PROC LOGISTIC, other model building issues including detecting confounding variables and identifying effect modifiers will also be addressed.