The alliance in child and adolescent psychotherapy is widely recognized as an important factor in therapy. Studies on the alliance have increasingly focused on assessment of the alliance as a dyadic construct, measuring both client and therapist alliance ratings. However, cross-informant reports of the alliance in child psychotherapy have not yet been subjected to meta-analysis. Therefore, the present meta-analysis aims to increase knowledge on the degree of convergence and divergence between child, parent, therapist, and observer alliance ratings in child and adolescent psychotherapy.

A series of three-level meta-analyses of 78 studies was conducted to investigate differences and associations between child, parent, therapist, and observer alliance ratings in child and adolescent psychotherapy.

Findings indicated that children and parents in general rated the alliance more positively than their therapists (d = 0.35, d = 0.72, respectively), and that child-therapist and parent-therapist alliance ratings were moderately correlated (r = .32, r = .23, respectively). Associations between child and therapist ratings and observer ratings were moderate to large (r = .43, r = .53, respectively).

It can be concluded that children and parents generally report more positively on the alliance compared to their therapists, which is consistent with research on the alliance in adult populations. The small to moderate associations between alliance ratings indicate that individuals to some extent have a shared perspective on their alliance, and that the various perspectives on alliance should be acknowledged when dealing with children and parents in therapy. Implications for future research are discussed.