ABSTRACT
This chapter contains an overview of new methodological approaches in the area of multitraitmultimethod (MTMM) modeling and shows how psychotherapy research can profit from a multimethod research and assessment program. The implications of multimethod approaches for test validation, multimethod measurement, indication and evaluation in psychotherapy research are discussed. To interpret the results of evaluation studies appropriately, the valid measurement of change processes is a necessary prerequisite. Construct validation deals with the question of whether tests and questionnaires measure the intended constructs of interest. Classical MTMM models are models for cross-sectional data. They can be applied to analyze the convergent and discriminant validity of state measures or trait measures. In particular, longitudinal MTMM models offer new ways to validate change measures and to validate indication rules that are based on the detection of differential change. Modern psychometric approaches such as latent statetrait (LST) models allow decomposing an observed score into a trait component that is stable over time.
