MTMM

MTMM stands for multitrait-multimethod matrix, a framework used in psychometrics to evaluate construct validity by examining how measures of multiple traits perform when assessed by multiple methods. Introduced by Campbell and Fiske in 1959, the MTMM approach aims to demonstrate convergent validity (that measures of the same trait with different methods relate to each other) and discriminant validity (that measures of different traits are distinct, especially across methods).

In a typical MTMM design, several traits are measured using several methods. For example, three traits might be assessed by three methods, producing a correlation matrix that includes all trait-method combinations. Key categories of correlations are: monotrait-heteromethod correlations, which compare the same trait across different methods and should be relatively high to support convergent validity; heterotrait-monomethod correlations, which compare different traits measured by the same method and should be relatively low to support discriminant validity; and heterotrait-heteromethod correlations, which compare different traits across different methods and should also be comparatively lower. Method effects can inflate correlations among measures that share the same method, potentially biasing interpretations of validity.

Analytically, MTMM can be approached through pattern interpretation of the correlation matrix or through more formal models, such as confirmatory factor analysis, which models trait factors and method factors separately to test convergent and discriminant validity. The MTMM framework remains a foundational tool for assessing whether measurements capture distinct underlying constructs rather than shared method variance. Limitations include sensitivity to sample size, design complexity, and the difficulty of interpreting imperfect or partially explained patterns.