The Multitrait Multimethod

The multitrait multimethod, an analytical tool introduced in 1959 by Campbell and Fiske in Psychological Bulletin, offers a useful way to test construct validity with measures of the same trait measured with different methods (i.e., convergent and discriminant validity). MTMM analysis relies on the assumption that correlations in a matrix are due to both traits and methods. High MTMM correlations correspond to measures of the same trait measured with the same method, while low MTMM correlations are associated with measurements of different traits that differ only in their measurement the multitrait multimethod method (i.e., heterotrait-monomethod correlations). MTMM is most commonly applied in conjunction with confirmatory factor analysis (CFA) through the specification of trait and method latent variables. However, most contemporary CFA-MTMM models restrict correlations between traits and methods to linear trait-method relationships.

In the MTMM framework, traits refer to the psychological characteristics or attributes being measured. These traits can include personality traits, cognitive abilities, emotional states, or any other constructs of interest. On the other hand, methods represent the tools, techniques, or instruments used to measure these traits. Examples of methods include self-report questionnaires, observational ratings, behavioral assessments, or physiological measures.

By employing the MTMM analysis, researchers can enhance the validity and reliability of their psychological assessments. The examination of trait convergence and method convergence allows for a comprehensive evaluation of measurement consistency and accuracy. This, in turn, ensures that the results obtained from psychological assessments are robust, replicable, and meaningful.

The MTMM analysis contributes to the field of psychological assessment by providing a systematic and rigorous framework for evaluating the validity and reliability of measurements. It helps researchers ensure that their assessments accurately capture the intended constructs and produce meaningful results.