Degree Date



Doctor of Philosophy (PhD)


Social Work and Social Research


A large body of evidence points to the therapeutic alliance as one of the major and universal factors contributing to outcomes in psychotherapy (Luborsky, 1976; Orlinsky, Ronnestad, & Willutzki, 2004; Wampold, 2001). Young adults are at heightened risk for mental health problems, yet they are less likely than other adults to receive psychotherapy (Office of Applied Studies, 2008), and there is little research on the impact of the alliance on psychotherapeutic outcomes for young adult clients. This study aimed to synthesize available evidence on the alliance and outcomes for 18 to 34 year olds, using current research synthesis techniques. Eligible studies included young adult clients involved in bona fide psychotherapies (Wampold, 1997) delivered in individual, face-to-face interactions with mental health professionals or trainees. Eligible studies used longitudinal cohort designs, with at least one measure of the therapeutic alliance followed in time by at least one outcome measure. Eligible alliance and outcome measures demonstrated evidence of reliability and validity. Outcomes included measures of mood, psychiatric disorders, global functioning, specific outcomes, and treatment participation or termination status. Data in included studies could be obtained from direct observations, interviews, self-reports, and/or clinical records. A rigorous search for eligible studies included multiple sources of published and unpublished literature. Twenty-five reports from 21 independent published studies were deemed eligible, after screening over 18,000 potentially eligible studies. Data were extracted using a standardized procedure with inter-rater reliability checks. Conceptually distinct outcomes and alliance reports from different sources were analyzed separately. A

tabular analysis of six commonly reported outcomes suggested that results of the primary studies were only partially reported. This sparse reporting pattern was a barrier to comprehensive analysis. The meta-analysis consisted of three forest plots. Two of these plots included studies with overlapping samples, which could not be synthesized independently. It was possible to synthesize results across studies in one forest plot which included partial correlations of client-rated alliance and the Outcome Questionnaire 45 from three studies; results indicated that more positive alliances were related to a decrease in psychiatric symptoms. This systematic review highlights the poor quality of data reporting in studies of the alliance and outcomes of psychotherapy for young adults. Sparse and uneven reporting of results in the primary studies severely restricted my attempts to conduct comparative analysis and synthesis. High levels of outcome reporting bias limited many analyses. This underscores the importance of increased adherence to reporting standards and the potential value of replicating basic therapeutic alliance and outcome studies using current methodologies and reporting standards to strengthen the evidence about the role of the alliance in positive changes in individual psychotherapy.