Issues of measurement and mechanism in meta-analyses: comment on Westen and Morrison (2001).

The authors suggest that D. Westen and K. Morrison's (2001) meta-analysis of treatment is critically limited in the consideration of measurement and mechanisms of therapeutic change. The measures included in the analysis fail to represent a comprehensive coverage of the domains within which change is expected. Moreover, they do not measure the theoretically derived constructs currently conceived as being central to each disorder. Further, the particular meta-analytical approach taken prohibits evaluation of the treatment components responsible for change. The authors reviewed the most recent data on comorbidity as an issue of treatment efficacy and generalizability, proffer an interpretation for the difference in outcome results across the 3 diagnostic groups, and discuss internally valid methodologies for the bridging from research to clinical practice.