Since canonical correlation analysis subsumes multiple regression as a special case, and since commonality analysis (a variance partitioning procedure) has proven useful in interpreting multiple regression results, the interpretation of canonical correlation results might also be enhanced by the use of commonality analysis. In this paper, a canonical correlation analysis using a data set of 64 observations, consisting of four predictor (independent) variables and four criterion (dependent) variables, is interpreted. Predictor variables were raw scores on four sections of the Myers-Briggs Type Indicator. Subjects were 64 school principals and assistant principals. A commonality analysis of the data is explained and illustrated, and results are interpreted in conjunction with the canonical correlation analyses. Commonality analysis offers several advantages for interpretation in that it: (1) indicates the degree to which predictors share variance with criterion variables; (2) indicates the extent of overlap of the variables; and (3) reinforces the recognition that canonical correlation is the most general linear model of parametric statistics. Four tables present analysis results, and a 20-item list of references is included. (SLD)