Are the parts as good as the whole? A meta-analysis of component treatment studies

J Consult Clin Psychol. 2013 Aug;81(4):722-36. doi: 10.1037/a0033004. Epub 2013 May 20.


Objective: Component studies compare standard treatments to treatments with added components or dismantled components. A previous meta-analysis (Ahn & Wampold, 2001) failed to find any differences in outcome between treatments with more components and those with fewer components, leading the authors to conclude that common factors and not specific ingredients account for therapeutic change.

Method: The current random effects model meta-analysis of psychotherapy component studies conducted between 1980 and 2010 included more than 3 times as many studies as Ahn and Wampold's (2001) meta-analysis (k = 66). Unlike the previous meta-analysis, this study conducted separate meta-analyses for additive and dismantling studies and also examined treatment outcomes at follow-up.

Results: For the dismantling studies, there were no significant differences between the full treatments and the dismantled treatments. For the additive studies, the treatment with the added component yielded a small, but significant, effect at completion (d = 0.14) and a slightly larger effect at follow-up (d = 0.28), but only for the specific problems that were targeted for treatment. Despite the diversity of populations studied, problems treated, and treatments examined, there was little heterogeneity among the results of these studies.

Conclusion: These findings suggest that added specific ingredients may contribute modestly to treatment outcomes.

Publication types

  • Meta-Analysis

MeSH terms

  • Humans
  • Models, Psychological*
  • Psychotherapy / methods
  • Psychotherapy / standards
  • Psychotherapy / statistics & numerical data
  • Research Design / standards
  • Research Design / statistics & numerical data
  • Treatment Outcome*