The research base II: what actually works
Understand common factors, therapist effects, alliance, and the limits of the specificity claim.
What you should be able to do afterwards
- 1State the dodo bird verdict precisely, including what it does not claim
- 2Describe the alliance–outcome relationship and the causal problem with it
- 3Explain therapist effects and why they are larger than treatment effects
- 4Describe deterioration rates and why clinicians fail to detect them
The reading
The uncomfortable findings, stated plainly. Bona fide psychotherapies produce broadly similar outcomes for most common presentations. The alliance predicts outcome across every orientation at around r = 0.28. Who delivers the treatment accounts for more variance than which treatment is delivered. Between five and ten per cent of adults deteriorate in psychotherapy, and clinicians identify only a minority of those cases without data.
The contextual model
Wampold’s account: therapy works through three pathways — the real relationship; the creation of expectations through a plausible explanation and treatment; and the enactment of health-promoting actions. On this view specific techniques matter mainly because they provide a coherent rationale and a ritual, not because of their particular content.
The medical model reply
The counter-position, held by serious researchers, is that this reads the aggregate over the particular: it is true on average and false in the specific cases where mechanism-targeted treatment demonstrably outperforms. The dispute is genuine, live, and unlikely to be settled by another meta-analysis.
What belongs to this unit
Questions to take away
For a seminar, a supervisor, or an argument with yourself.
- 1If therapist effects exceed treatment effects, what should training look like?
- 2The alliance predicts outcome. Does it cause it? What evidence would settle that?
- 3You discover you are in the lower half of your service’s outcome distribution. What do you do?