Is there an allegiance bias on the efficacy of systemic therapy? A three-level meta- regression analysis

Allegiance bias affects the reproducibility of psychotherapy outcomes, but its impact in systemic therapy (ST) remains largely unexplored.

Two meta-analyses were conducted: Study 1 with 42 RCTs (364 effect sizes) for children and adolescents, as well as Study 2 with 30 RCTs (161 effect sizes) for adults. Allegiance was measured using a comprehensive rating scale across researcher, therapist, trainer, and supervisor levels. Furthermore, risk of bias (RoB) was assessed using the adapted tool for psychotherapy. Three-level meta-regressions were performed, controlling for dosage.

No significant moderating effect of the overall allegiance score was found in either study, Study 1: F(2, 361) = .88, p = .42, k = 364, s = 42; Study 2: F(2, 158) = 2.35, p = .10, k = 161, s = 30 . Researcher allegiance predicted higher treatment efficacy in Study 1 (β = .14, p = .02) but not in Study 2 (β = .02, p = .82). A significant moderating effect of RoB emerged in Study 2 (F(8, 151) = 3.02, p = .004, k = 161, s = 30), indicating that higher RoB was associated with larger effects favoring ST. This pattern was absent in Study 1 (F(8,355) = 1.04, p = .41, k = 364, s = 42).

The findings underscore the role of researcher allegiance investigating ST for children and adolescents and RoB in ST for adults, highlighting the need for transparent methods and replication efforts.