ABSTRACT
In this chapter, the authors introduce to logistic regression, which can also be used when the outcome is categorical and that allows model prediction. They discuss methods that can be used to enter predictors in logistic regression models. To understand how the logistic regression equation operates, there are three primary computational concepts that must be understood: probability, odds, and the logit. Maximum likelihood estimation performed by statistical software usually begins the estimation process with all regression coefficients equal to the most conservative estimate. Stepwise logistic regression is a data-driven model building technique where the computer algorithms drive variable entry rather than theory. Simulation research suggests that logistic regression is best used with large samples. Noncollinearity is applicable to logistic regression models with multiple predictors just as it was in multiple regression. Outliers and influential cases are problematic in logistic regression analysis just as with ordinary least squares regression.
