Meta-Analysis of Therapist Effects in Psychotherapy Outcome Studies

In a meta-analysis, we examined factors that could account for the differences in therapist efficacy evidenced in psychotherapy outcome studies. The factors investigated were: (1) the use of a treatment manual, (2) the average level of therapist experience, (3) the length of treatment, and (4) the type of treatment (cognitive/behavioral versus psychodynamic). Data were obtained from fifteen psychotherapy outcome studies that produced 27 separate treatment groups. For each treatment group, the amount of outcome variance due to differences between therapists was calculated and served as the dependent variable for the meta-analysis. Each separate treatment group was coded on the above four variables, and multiple regression analyses related the independent variables to the size of therapist effects. Results indicated that the use of a treatment manual and more experienced therapists were associated with small differences between therapists, whereas more inexperienced therapists and no treatment manual were associated with larger therapist effects. The findings are discussed in terms of the design and the analysis of psychotherapy outcome research.