Regression in the service of the superego: the do's and don'ts of stepwise multiple regression

Stepwise multiple regression is a very powerful but often misused technique. It can be used to find a set of independent variables which can predict some outcome. However, there are problems when the results of a stepwise solution are used to try to explain or understand the dependent variable. This paper discusses the different types of stepwise regressions, some of the legitimate and illegitimate uses of this technique, some of the difficulties encountered when trying to interpret the results and other solutions to the problems posed by using them.