Using and Interpreting Logistic Regression: A Guide for Teachers and Students

Despite the frequent use of logistic regression in the social sciences, considerable confusion exists about its use and interpretation. This confusion is attributed to a lack of adequate teaching materials and to unfamiliarity with logistic regression by many statistics instructors. The purpose of this paper is to define and illustrate basic concepts of dichotomous logistic regression (DLR) and to present strategies for teaching these concepts. Strategies include 1) using analogies between ordinary least squares (OLS) regression and logistic regression, 2) illustrating concepts with contingency tables, 3) focusing initially on the bivariate case for ease of understanding, and 4) linking logistic regression concepts to interpretation of statistics in computer outputs. After discussing several examples of logistic regression, we present and illustrate statistics for evaluating the goodness of fit and predictive efficacy of a DLR model.