The therapeutic alliance
The most replicated process finding in the field, and the causal question it has not settled.
Bordin’s 1979 reformulation made the alliance pantheoretical: agreement on goals, agreement on tasks, and the affective bond. Every subsequent measure derives from it, and the alliance–outcome association has been found in hundreds of studies across every orientation, measure and rater.
The numbers
| Finding | Estimate |
|---|---|
| Alliance–outcome correlation | r ≈ 0.28 across 300+ studies |
| Patient-rated vs. therapist-rated | Patient ratings predict outcome more strongly |
| Early alliance | Predicts outcome even controlling for early symptom change |
| Rupture repair | r ≈ 0.29 with outcome |
| Therapist contribution | Therapists who form better alliances get better outcomes across their caseloads |
The causal problem
A correlation between alliance and outcome could mean the alliance causes improvement, or that improvement causes a better alliance, or that a third variable causes both. The strongest evidence for the causal direction comes from studies controlling for prior symptom change and from the therapist-level finding: therapists who form better alliances get better outcomes across their whole caseload, which is difficult to explain as a patient effect.
What to do with it
- 1Measure it, session by session, with a brief instrument.
- 2Solicit negative feedback explicitly; a neutral invitation produces a polite maximum score.
- 3Treat a small drop as a signal rather than as noise.
- 4When alliance and protocol conflict, repair first.
- 5Notice withdrawal ruptures, which feel like good sessions.
Sources
Flückiger, C., Del Re, A. C., Wampold, B. E., & Horvath, A. O. (2018). The alliance in adult psychotherapy: A meta-analytic synthesis. Psychotherapy, 55(4), 316–340
meta-analysisEubanks, C. F., Muran, J. C., & Safran, J. D. (2018). Alliance rupture repair: A meta-analysis. Psychotherapy, 55(4), 508–519
meta-analysis