Does therapy actually work?

A: Yes — and this isn't a hopeful yes, it's one of the better-documented findings in health care. Meta-analyses across hundreds of randomized trials put psychotherapy's average effect size around 0.7–0.8 for depression and anxiety — comparable to or better than many common medical treatments, and with benefits that persist after treatment ends, which medication alone often doesn't show.¹ The honest caveats: fit matters more than method (the therapeutic relationship predicts outcomes better than the brand of therapy — which is why switching → beats quitting), it takes weeks, not days (about half of clients improve meaningfully within 8–20 sessions¹), and the claim is "therapy works," not "any therapist will work for you." Roughly 5–10% of people don't improve or feel worse with a given therapist — the signal to change therapists, not to conclude you're untreatable.

Worked example: moderate depression, weekly CBT. Typical arc: weeks 1–3 feel like expensive talking, weeks 4–8 the sleep and the mornings shift, week 12 the outside-world numbers (missed workdays, canceled plans) move. That lagging-then-compounding curve is the normal shape, not a slow start.

Do this: pick by fit and availability (the 5-day method →), give it 4–6 sessions before judging, and name it in the room if it isn't working — "I'm not sure this is helping" is a clinically useful sentence, not rudeness.

Source: 1. Cuijpers et al. (2023), World Psychiatry — meta-analytic overview of psychotherapy efficacy; Wampold & Imel (2015), The Great Psychotherapy Debate; APA resolution on psychotherapy effectiveness — apa.org.