Psychological therapy has been studied extensively, and the evidence supports it. It is also frequently oversold, and an honest account of what it does and does not achieve is more useful than either enthusiasm or dismissal.
What the research shows
Meta-analyses consistently find that psychological therapy outperforms no treatment and waiting-list controls for depression, anxiety disorders, PTSD and OCD.
Effect sizes are generally described as moderate, comparable to many accepted medical treatments.
For several conditions, therapy performs comparably to medication in the short term and shows better maintenance of gains after treatment ends, which is a meaningful advantage.
What moderate means
It means therapy helps a substantial proportion of people meaningfully, and does not help everyone, and rarely eliminates a problem entirely.
Typical findings are that around half of people receiving treatment for depression or anxiety improve substantially, with a further group improving somewhat and a minority not benefiting.
Anyone promising transformation is overstating what the evidence supports.
The common factors debate
One of the longest-running arguments in the field concerns whether specific techniques matter or whether the shared elements do the work.
Comparisons between established therapies frequently find smaller differences than expected, which has been called the dodo bird verdict.
The common factors, including the therapeutic relationship, expectation of improvement and having a coherent explanation for your difficulties, account for a substantial share of the benefit.
The qualification is important. For specific conditions with specific treatments, technique matters considerably. Exposure and response prevention outperforms general counselling for OCD by a wide margin, and trauma-focused therapy outperforms supportive therapy for PTSD.
The therapeutic alliance
The relationship between therapist and client is among the more robust predictors of outcome across approaches.
This includes agreement on goals, agreement on tasks, and the quality of the bond. It is measurable, and stronger alliances predict better outcomes.
Importantly, the alliance early in treatment predicts later improvement rather than simply reflecting it.
Therapist variation
Outcomes vary considerably between therapists, more than they vary between approaches.
Some therapists consistently achieve better results, and this is not fully explained by experience or training.
This is a reason to change therapist if it is not working rather than to conclude that therapy does not work.
Where it can do harm
Rarely discussed and worth stating. Studies find a minority of people deteriorate during therapy.
Contributing factors include poor fit, a therapist working outside their competence, approaches that are inappropriate for the problem, and the therapy destabilising someone without adequate support.
Boundary violations by therapists cause serious harm and are the reason regulation and complaint procedures matter.
The publication problem
Effect sizes in this field have been affected by publication bias, where studies finding no effect are less likely to be published.
Analyses correcting for this generally find smaller effects than the original literature suggested, though still positive.
This is a reason for realistic expectations rather than for dismissing the evidence.
What improves your odds
Attending consistently and completing the course rather than stopping when things improve.
Doing the between-session work where the approach involves it, which is one of the stronger predictors in CBT.
Raising problems with the therapist rather than disengaging quietly.
Matching the approach to the condition where a specific treatment exists.
Why comparisons with medication are difficult
Trials comparing therapy with medication face a structural problem: participants cannot be blinded to which one they are receiving, whereas drug trials use placebos.
This affects how the two literatures compare, and it means confident claims that one substantially outperforms the other should be treated cautiously.
For moderate to severe depression, guidance generally supports offering both, and combining them frequently outperforms either alone.
The realistic expectation
Therapy is a treatment with reasonable evidence, moderate effects and meaningful benefit for a substantial proportion of people.
It is not a guarantee, and if it does not help, that is information about the fit rather than about you.
This is general information rather than medical advice.