Marianne Welmers

Chapter 6 138 but was found to negatively impact alliances with adolescents (DiGiuseppe et al., 1996). Ultimately, playing an active role in a conjoint treatment process with adults, for example by providing feedback, is important as well as challenging for children and adolescents. It should therefore be carefully guided by the therapist both verbally and non-verbally (Escudero & Friedlander, 2017; Moore & Bruna-Sue, 2011; O’Reilly, 2008; Strickland-Clark et al., 2000). The importance of understanding and attending to children’s and adolescents’ experiences of the alliance is underlined by results of analyses in Chapter 5, indicating that youth’s self-reports of the alliance were strongly associated with more positive treatment outcomes in terms of youth problems, in contrast with parents’ self-reports of the alliance. Although the number of included youth with self-reports on the alliance was too small to draw reliable conclusions, this finding is in line with some previous family treatment studies (Friedlander et al., 2012; Johnson et al., 2002; Keeley et al., 2011; Rienecke et al., 2016) and with meta-analytic research in adolescent treatment (Murphy & Hutton, 2018; Roest et al., 2021b), both underlining the importance of youth’s self-reported alliance in predicting treatment outcome. In the meta-analysis (Chapter 2) there was no significant difference in the effect on treatment outcome between alliances with parents and alliances with youth (therapist-, self- and, observer-reports collapsed). We did, however, find a substantial difference between parents and youth with regard to changes in the alliance over time, indicating that improvement of alliances with youth over the course of treatment affected outcomes more strongly than improvement of alliances with parents. This underlines the importance of a process-oriented approach both for practice and scientific research: building alliances with youth in family treatment, likely because of the specific challenges outlined before, may take considerable time. Balancing Alliance Differences: Associations with Outcome Moving beyond the observation that alliances differ notably between family members, the present dissertation also aimed at studying the presumed effect of these alliance differences – referred to as unbalanced or split alliances – with two prominent findings. First, the meta-analysis (Chapter 2) indicated no effect of unbalanced or split alliances on treatment outcome. However, only five studies reported on the effect of unbalanced or split alliances, and these studies were highly heterogeneous in their definitions and measures of split alliances and treatment outcomes. A more recent meta-analysis in couple and family therapy identified seven studies on unbalanced or split alliances and outcome, and found a significant moderate effect size ( r = .316), indicating that a greater imbalance in family members’ alliances leads to less favorable treatment outcome (Friedlander et al., 2018).