The Effect-Size Map

"Meditation works" is a sentence with a missing column: works for what, compared with what, and how big? This page builds the scorecard outcome by outcome — anxiety, depression, pain, blood pressure — using the actual meta-analytic numbers, including the ones the marketing leaves out.

A One-Minute Course in Effect Sizes

Every claim on this page is expressed as Cohen's d, the field's standard yardstick: 0.2 is small, 0.5 moderate, 0.8 large. Two rules keep the yardstick honest. First, the comparator rule: an effect measured against a waitlist (people who get nothing) is almost always larger than the same effect measured against an active control (people who also get attention, structure, and hope). Second, the publication rule: small trials with dramatic results get published; small trials with null results often don't, which quietly inflates pooled estimates. The meta-analyses below differ in how much they correct for both — which is why two papers can both be right about the same practice and print different numbers.

The parent topic (Meditation & Mindfulness) gave the overview; this page gives the columns: outcome, comparator, number, verdict. The pattern holds when the comparison is done per-outcome rather than per-study, and it is one of the more replicated findings in the field.

Anxiety: The Strongest Card

If the evidence base were a deck, anxiety would be the card worth keeping. The most cited meta-analysis of mindfulness-based treatments for anxiety disorders reports an effect size of roughly 0.8 against waitlists and roughly 0.35 against active treatments (Vøllestad, Clinical Psychology Review, 2012). Both numbers matter: the first tells you the intervention does something real; the second tells you the something is ordinary-sized once you compare it with any credible alternative.

• 🎯 In the trials: an 8-week MBSR program reduced anxiety more than a matched stress-management education course in adults with generalized anxiety disorder (Hoge, Journal of Clinical Psychiatry, 2013).
• 💊 Against medication: in a 2023 randomized trial, MBSR was statistically non-inferior to escitalopram for anxiety symptoms over 8 weeks (Hoge, JAMA Psychiatry, 2023) — different side-effect profiles, comparable anxiety relief.
• 🧾 The honest read: meditation is a reasonable first tool for sub-clinical and mild anxiety, and a defensible option for clinical anxiety — not because it outperforms therapy, but because it performs in the same neighborhood.

Depression: Real, and Smaller Than the Hype

The landmark JAMA Internal Medicine meta-analysis put the mindfulness effect on depressive symptoms at about 0.30 (Goyal, 2014) — small, consistent, and real. Broader meta-analyses land around 0.5 before a comparison group is involved, and around 0.3 once one is (Khoury, Clinical Psychology Review, 2013). Where the depression evidence genuinely earns its reputation is relapse prevention: for people with three or more past episodes, mindfulness-based cognitive therapy (MBCT) reduced relapse risk by roughly 43% against usual care (Piet & Hougaard, Clinical Psychology Review, 2011), and in a randomized non-inferiority trial relapse rates over two years were similar to staying on maintenance antidepressants — 44% versus 47% (Kuyken, Lancet, 2015).

The pattern worth internalizing: meditation treats acute depression modestly and prevents its return well. That asymmetry matters when you decide where a practice belongs in your own toolkit.

Pain: A Different Mechanism

Chronic pain is where meditation does something mechanistically interesting: it barely touches the pain signal and still helps. The same JAMA meta-analysis put the effect at about 0.33 (Goyal, 2014); a dedicated systematic review of mindfulness for chronic pain found small, consistent improvements in pain, depression, and quality of life (Hilton, Annals of Behavioral Medicine, 2017), and a meta-analytic review of acceptance-based approaches found small effects on pain itself with somewhat larger effects on the distress pain causes (Veehof, Pain, 2016). The consistent story across studies: people report the pain matters less and interferes less, even when its intensity barely moves. That is attention training doing what attention training does — changing the relationship with a signal rather than deleting it.

Blood Pressure: The Weakest Card

The physiology scorecard is where the humility belongs. A meta-analysis of transcendental meditation trials reported reductions of roughly 4 mmHg systolic and 2 mmHg diastolic (Bai, American Journal of Hypertension, 2015) — real, modest, and concentrated in one practice family, with substantial between-study disagreement. A broader review of mindfulness and physiological stress markers found reductions in cortisol, blood pressure, heart rate, and inflammatory markers, but noted the effects were inconsistent and often small (Pascoe, Journal of Psychiatric Research, 2017). The landmark JAMA meta-analysis went further: across programs, effects on stress physiology were weak to inconclusive (Goyal, 2014). For context on what a few mmHg do and don't buy, the Cortisol 101 topic owns the biology; the short version is that meditation is a poor substitute for the levers that actually move blood pressure.

Questions, Answered Briefly

• Which number should I believe? The active-controlled one. It is smaller, less flattering, and closer to what you will actually experience.
• Does app-based meditation get the same results? Online and app-delivered programs show small effects versus waitlists and near-zero versus active comparators (Spijkerman, Clinical Psychology Review, 2016) — directionally the same, smaller in size.
• Is meditation better than therapy? Meta-analyses comparing mindfulness with established psychological treatments find them broadly comparable, not superior (Goldberg, Clinical Psychology Review, 2018). Choose on fit and access.
• How much practice do these effects assume? The trials used 20–45 minutes daily for 8 weeks. The dose question — how little you can get away with — is the next subtopic.

Related Topics