Discriminant function analysis (DFA) is a statistical technique used to determine which variables are best at predicting group membership. It creates linear combinations of predictor variables called discriminant functions that discriminate between categories of a dependent variable. DFA is similar to regression and ANOVA. It works by maximizing between-group differences and minimizing within-group differences to classify cases into groups based on predictor variables. The assumptions of DFA include normally distributed predictors and equal variance-covariance matrices within groups.