The multitrait-multimethod matrix is a versatile tool for structuring the design and analysis of studies in many areas of psychology, including the aging of psychological processes. The basic goal for which the multitrait-multimethod matrix was developed is the establishment of the construct validity of measures. With newer forms of analyzing multitrait-multimethod data, particularly those using structural equation modeling, it is possible to specify models and employ model comparisons for an even broader goal: the more adequate generalization of conclusions across the facets included in the study. Topics addressed in this paper include the design of multitrait-multimethod studies, the specification of an array of structural models for such data, model comparisons that allow the estimation of the degree of convergent validity, discriminant validity, and method variance exhibited by a set of measures, and the extension of structural models to more than a single group, which allows the testing of questions regarding means, as well as the traditional tests of differences in variance and covariance. The procedures discussed are then explicated by describing the representation of data from a hypothetical four-group study of memory employing a form of design that is common in research on aging. Strengths and weaknesses of the structural modeling approach to the analysis of multitrait-multimethod data are discussed.