The multitrait-multimethod (MTMM) design of D. T. Campbell and D. W. Fiske (1959) is the most widely used paradigm for testing construct validity, but it continues to be plagued by problems associated with definition of terms, operationalizations of their guidelines, and analytic procedures used to test them. Using five diverse MTMM data sets, five current analytic approaches are demonstrated, compared, and contrasted. These include two manifest variables approaches: the original guidelines of Campbell and Fiske and the analysis of variance model. Also studied are the following latent variables approaches: (1) a taxonomy of confirmatory factor analysis (CFA) models; (2) the covariance component analysis (CCA) model; and (3) the composite direct product (CDP) model. Even though the five approaches use a common terminology (convergent validity, discriminant validity, and method effects), there is a fuzziness about what the concepts mean and how they are operationalized. This review and analysis recommends a common terminology and operationalization of terms based on CFA models. The use of four CFA models and the CDP model is also recommended, along with the original Campbell-Fiske guidelines. The strongest single model appears to be the CFA correlated uniqueness model. One figure and two tables present study findings, and a 52-item list of references is included. Two appendices provide an additional 10 pages of tables. (Author/SLD)