Modeling MTMM Data from Additive and Multiplicative Covariance Structures: An Audit of Construct Validity Concordance

The purpose of this study was to determine the extent to which findings derived from four approaches to MTMM analyses were consistent in providing evidence of construct validity related to the measurement of four dimensions of perceived competence (social, academic, English, mathematics) across four maximally dissimilar rating methods (self, teacher, parent, peer). MTMM methodological approaches included that of Campbell-Fiske (1959), the general confirmatory factor analytic (CFA) model (Joreskog, 1969), the CFA (Correlated Uniqueness) model (Marsh, 1988), and the Composite Direct Product model (Browne, 1984). Procedures were applied to data from 158 grade 11 high school adolescents. Advantages, disadvantages, findings, and caveats related to each procedure are discussed.