A strategy is outlined for evaluating the Myers-Briggs Type Indicator, in terms of its theory, the claims made for it, and its virtues which are independent of the theory. The importance of compiling norms, of determining reliability, and of determining the effect of faking are stressed. The relative merits of ways of reporting results are described: the type category approach, the score approach, and the compound type-plus-score approach. The importance of cross-validation has been taken for granted. The importance of determining situational effects is also discussed; spurious inferences may result when the same situational factors affect both Indicator scores and the criterion in the same way.