Decades of psychotherapy outcome research — including meta-analyses by Norcross, Lambert, Horvath, and colleagues — consistently show that the therapeutic alliance is one of the strongest predictors of treatment success. The finding has been replicated across hundreds of studies, dozens of treatment modalities, and a wide range of clinical populations and settings.
What is less examined — and what the field has conspicuously avoided acting on — is what this finding means for how practices should be designed.
The Horvath et al. (2011) meta-analysis found an effect size of r = 0.275 — a moderate but highly reliable relationship between alliance and outcome, accounting for approximately 7.5% of outcome variance. A separate meta-analytic review found an overall correlation of .278, explaining approximately 7.7% of variation in clinical outcomes — consistent with the broader literature. Specific treatment models, by comparison, account for less than 1% of outcome variance. The alliance accounts for roughly seven times more.
Norcross and Lambert’s “Psychotherapy Relationships That Work” identifies the alliance as a demonstrably effective relationship element and reports that the psychotherapy relationship makes substantial and consistent contributions to outcome independent of treatment type.
The intake problem
Most practices assign clients to clinicians based on availability and specialty match. These are reasonable criteria. They are not criteria that optimize for alliance formation.
The research on alliance formation suggests that early session alliance ratings — typically measured around session three — are highly predictive of eventual outcomes. Research on early alliance and outcome consistently demonstrates that alliance problems identified early are addressable, but only if they are identified. Practices that do not have a system for measuring alliance early are allowing a significant proportion of their caseload to proceed toward poor outcomes that were detectable and potentially addressable.
If you knew that roughly 7% of your outcome variance was determined by something you could measure and intervene on in the first three sessions, would you measure it? Most practices do not.
— James Calloway, Business & Therapy
What measurement looks like in practice
Several validated brief measures of therapeutic alliance exist and can be administered in under two minutes at the end of a session. The Session Rating Scale is the most widely used in practice settings. It produces a score that flags alliance concerns before they become dropout events.
The practices that use these measures report that they identify alliance problems they would not otherwise have caught, and that early intervention — a direct conversation with the client about what is and is not working — frequently resolves those problems before they lead to dropout or poor outcomes.
The matching question
The deeper implication of the alliance research is one the field is not yet ready to act on at scale: if alliance quality matters more than technique, then matching clients to clinicians based on interpersonal fit — rather than diagnosis or modality — should improve outcomes. Research on early alliance and client-therapist fit supports the idea that early alliance formation is influenced by factors beyond diagnosis and presenting problem.
This is operationally complex. It requires information about clinician interpersonal style that most credentialing and referral systems do not collect. It requires a willingness to rematch clients who are not forming strong alliances. And it requires a practice culture in which poor alliance formation is treated as clinical information rather than clinician failure.
What we still do not know
The alliance research is robust on the question of whether alliance matters. It is considerably less developed on the question of how alliance quality can be systematically improved at the practice level. The intervention research in this area is promising but limited. This is a gap the field needs to close.
Do you have additional information about therapeutic alliance research or alliance measurement tools in practice? We update our articles and research regularly. Contact our editorial team with corrections, updates, or sources.