This paper discusses the principles underlying discriminant analysis and constructs a simulated data set to illustrate its methods. Discriminant analysis is a multivariate technique for identifying the best combination of variables to maximally discriminate between groups. Discriminant functions are established on existing groups and used to predict the classification of subjects with unknown group membership. The simulated study sought to discriminate high, medium, and low job achievement groups comprised of 10 employees each. Five employee variables were created varying in their degree of correlation with the independent variable. A discriminant analysis was performed on this ideal data set with the CANDISC feature of the Statistical Analysis System (SAS). The rest of the paper treats in detail the statistical output of this analysis and serves as a guide to interpreting it. The discussion of results for the simulated data includes comparison of the univariate and multivariate statistics, use of the correlation matrix to detect multicollinearity, and application of the Wilks' Lambda test. The paper concluded with the graphing of centroids by discriminant functions as a means of visually identifying the different subgroups. (Author/LPG)