The contribution of individual variables to overall multivariate significance in a multivariate analysis of variance (MANOVA) is investigated using a combination of canonical discriminant analysis and Roy-Bose simultaneous confidence intervals. Difficulties with this procedure are discussed, and its advantages are illustrated using examples based on the following four data sets with different characteristics: (1) one-way MANOVA with three dependent variables and five subjects in each treatment; (2) 3 treatments, 4 dependent variables, and 12 subjects in each treatment; (3) one-way MANOVA with eight treatment levels and two dependent variables; and (4) 3 treatment groups and 3 variables with 125, 90, and 25 subjects, respectively. The proposed process--Roy/STP (moderate)--identifies treatment mean contrasts and dependent variables worthy of further consideration. Important parts of the process discussed can be gleaned from standard packaged statistical routines, but the step of using the Roy-Bose confidence intervals is left to the researcher. A program to complete the procedure is available in SAS and Fortran. Sixteen references are listed. An appendix illustrates the confidence interval procedure. (SLD)