Aims: The presence of conflicts of interest in psychotherapy, with the consequent errors and/or falsifications, has not been adequately considered in psychotherapy. In this work, a much cited study by Jonathan Shedler of mega-analyses of efficacy studies of various treatments in psychiatry, both psychotherapeutic and pharmacological, is taken into consideration for in-depth examination, in which two Authors declared conflicts of interest: General psychotherapy, CBT, Psychodynamic Psychotherapy (PP), Antidepressant Psychopharmacotherapy (PTAD).
Method: Examination of the method and results of the studies considered in J. Shedler’s mega-analysis in the light of statistical methodology. We then proceeded to examine: a) the method used for merging the results, b) the correctness of the calculations, c) the congruence between results and conclusions.
Results: Numerous meta-analyses included in this study do not have the sufficient number to consider reliable the calculation of the Effect Size (ES or Degree of Effect), one in the PTAD group, six in the PP group. Furthermore, meta-analysis studies are included together with some mega-analyses, and a meta-analysis study of PP does not have control groups, therefore it is not comparable to the others. Therefore, only one of the meta-analyses has characteristics that can lend themselves to verifying the efficacy of the PP, that of Abbas & coll. (2006), but many sub-analyses of this study do not have the sufficient number to consider the calculation of the ES reliable, therefore they should be excluded. Finally, the final result (ES = .77) is erroneously reported in the conclusions as ES = .97.
Discussion: In view of the above, the study presents numerous errors, which can sometimes be interpreted as manipulation or falsification of data, and represents a real example of abuse of the meta-analysis method.