It is not uncommon to hear practitioners say they choose not to work with couples explaining that they find this therapy daunting and difficult. Forty years of research including meta-analytic studies spanning decades allows us to confidently assert that couples therapy works. However, while research has also attempted to identify the mechanisms for this effectiveness, the how and why it works remain unclear. Perhaps this explains why otherwise confident and competent therapists avoid this domain.
How Can This Be?
With over 48 couple therapy process studies conducted over the last 40 years how can this be so? Anderson et al (2025) addressed this question using Doss’s (2004) framework to systematically review and categorize all aspects of the change process reported in quantitative couple therapy research. The model suggests that processes of change reflect the interaction between therapist interventions and client response and include four components. The first is therapist actions, the active ingredients of change and include interventions, directives, homework tasks, and theoretical techniques. These aim to trigger client change processes in behaviours, reactions and experiences both within the session and beyond. This recursive loop between client and practitioner produces change mechanisms, alterations in client’s skills and capacities that appear outside the practitioner’s direct influence. The final step is achieving the goal of therapy.
What Did They Find?
In applying this framework and detailed reporting of available studies the researchers came to six conclusions. Firstly, none of these studies explored the entire change process. A second observation was that while some studies did report promising results none of these have been directly replicated. Of the studies which explored hypothesized associations few reached significance and authors were unable to explain why some associations were supported while others were not. A fourth concern was that studies often reported on a single session, therapist, client or without reference to the broader context of change. A fifth observation was that participants in the studies were generally ‘white, middle-class, married, college‐educated, heterosexual individuals experiencing mild to moderate levels of relationship distress with few psychological comorbidities. Finally, the authors note methodological concerns with the research including small sample size, low statistical power, and a reliance on a limited set of coding schemes.
What Does this mean for Practitioners?
Clearly, greater collaboration between researchers and clinicians is needed with a focus on understanding theorized mechanisms of change within models and common principles of change across models. In addition, an exploration of the entire change process is required, to identify practical approaches that practitioners can ethically and easily apply.
While it is tempting to feel disheartened, the authors conclusion is encouraging. ‘The field, in essence, is a 40‐year‐old toddler—full of potential but still taking uncertain steps into the future. However, the limited and fragmented findings we do have show promise and offer fertile ground for future exploration.’ Hopefully, greater clarity and direction will allow more practitioners to venture into this challenging terrain.
©Copyright Bower Place™ Pty. Ltd. 2026
