Psychotherapy research has been running for decades and has produced a finding that most descriptions of therapy do not lead with.
The common factors finding
When different therapeutic approaches are compared directly for common conditions, the differences between them are frequently small.
Which has been replicated enough to have a name — the dodo bird verdict, after the character who declares that everyone has won and all must have prizes.
The interpretation is that factors common to all therapies account for a large share of the benefit.
Those factors are generally identified as the therapeutic relationship, the client's expectation of improvement, the provision of a coherent explanation for their difficulties, and the structure of regular attention.
Where it does not hold
The finding is frequently overstated, and it is worth being precise.
For specific conditions, specific approaches do outperform others.
Exposure-based approaches for anxiety disorders and obsessive-compulsive disorder have evidence that general supportive therapy does not match.
Specific trauma-focused approaches for post-traumatic stress have similar differentiation.
Which means the equivalence finding applies most strongly to general distress and depression, and less to conditions with a clear mechanism.
The alliance
The relationship between therapist and client is the most consistently replicated predictor of outcome across approaches.
It is measured as agreement on goals, agreement on tasks, and the emotional bond.
The causal direction is debated — a good alliance may cause improvement, or early improvement may cause a good alliance — and studies attempting to separate them suggest both operate.
Practically, this means the fit with a particular therapist matters, and it is a reasonable basis for changing therapists rather than a personal failing.
What cognitive behavioural approaches propose
The model is that thoughts, feelings and behaviour interact, and that identifying and testing unhelpful thought patterns changes the others.
The techniques are specific and teachable — monitoring thoughts, examining evidence, behavioural experiments, graded exposure.
Which makes it well suited to research, since it can be manualised and delivered consistently, and this is part of why it has the largest evidence base.
The criticism is that manualisation produces a technique-focused therapy that may miss what matters for a particular person.
What psychodynamic approaches propose
That current difficulties relate to patterns established earlier, frequently outside awareness, and that examining them in the context of the therapeutic relationship allows change.
Historically less researched, partly because the approach resists manualisation and involves longer treatment.
The evidence base has grown, and reviews find effects comparable to other approaches for several conditions, with some evidence of continued improvement after treatment ends.
Dose
Research on how much therapy is needed finds that a substantial proportion of improvement occurs early.
Which has driven service models offering shorter courses, on the reasoning that most people improve within them.
The counterargument is that averages conceal a group who need considerably more, and that short courses serve the responsive majority while failing the rest.
What predicts a poor outcome
Worth knowing, since these are partly addressable.
Poor alliance, which can prompt a change of therapist.
Lack of agreement on what the work is for.
And, in some studies, therapist factors that vary considerably between individual practitioners regardless of approach or training.
That last finding is uncomfortable and reasonably well established.
The necessary statement
I am not a clinician. This summarises published research on how therapy has been studied.
Anyone considering therapy should speak to their doctor or contact a regulated professional body, both of which will do considerably more than any article can.
Access and what it costs
The practical constraint that shapes what most people actually receive.
Waiting times in publicly funded systems are frequently long, and private provision is expensive enough to be inaccessible to many.
Which has driven expansion of lower-intensity options — guided self-help, digital programmes, group formats — that have evidence for milder presentations and are less established for severe ones.
Stepped care models, where people start with the least intensive option and escalate if it does not work, are the standard response and are efficient and slow for those who need more from the start.
Checking credentials
Worth a specific note because titles are inconsistently protected.
Some titles are legally protected and require registration with a statutory regulator. Others are not, and anybody may use them.
Which varies by country, and checking whether a practitioner appears on a statutory or accredited voluntary register is a five-minute exercise.
Registers also carry complaint mechanisms, which unregistered practice does not.
That is the single most useful practical check available and it is routinely skipped.
What to expect early on
Worth knowing since unmet expectations cause people to drop out.
Early sessions generally involve assessment and agreeing what the work is for, which can feel like nothing is happening.
Some approaches produce temporary worsening before improvement, particularly those involving exposure to avoided material, and this is generally explained in advance.
A therapist who cannot say what the approach is or how progress will be judged is a reasonable reason to ask more questions.