ANALYSIS OF MULTITRAIT-MULTIRATER PERFORMANCE-APPRAISAL DATA - COMPOSITE DIRECT-PRODUCT METHOD VERSUS CONFIRMATORY FACTOR-ANALYSIS

This study addresses the manner in which trait and rater variance combine in multitrait-multirater (MTMR) performance appraisal data. The Confirmatory Factor Analytic (CFA) model assumes trait and rater variance combine additively, whereas the Composite Direct Product (CDP) model assumes a multiplicative relationship. Implications of these models are explicated for MTMR data, and empirical differences are examined using four data sets. Results indicated that the fit of the CDP model was superior to that of the CFA model in all four data sets. Discussion centered on strengths and weaknesses of the CDP and CFA models, assumptions regarding trait/rater relationships and conflict between conceptual simplicity and realistic representation of relationships. Guidelines for applying the CDP method were provided.