Meditation is the practice of deliberately training attention, usually by resting it on something simple like the breath and returning to it each time the mind wanders. The evidence for it is real but more modest than the popular framing suggests: the best available analysis found moderate evidence of improvement in anxiety, small effects overall, and no sign that meditation outperforms other active treatments such as exercise or therapy. It is worth doing. It is not a cure for anything.
What follows is what the research actually found, what a structured program involves, how to start without joining anything, and the honest caveats about safety and about who it does not suit.
What does the research actually show?
The most useful reference point is a meta-analysis published in JAMA Internal Medicine and available via PubMed Central. Its headline finding was that "Mindfulness meditation programs had moderate evidence to improve anxiety [ES 0.38 (CI 0.12 to 0.64) at 8 weeks; ES 0.22 (0.02 to 0.43) at 3–6 months]".
An effect size of 0.38 is small to moderate, and the drop to 0.22 at three to six months suggests the benefit fades as practice does. The same paper is equally clear about what it did not find:
- "We found either low evidence of no effect or insufficient evidence of any effect of meditation programs on positive mood, attention, substance use, eating, sleep, and weight."
- "We found no evidence that meditation programs were better than any active treatment (drugs, exercise, other behavioral therapies)."
That second point deserves emphasis, because it is routinely omitted. Meditation performed comparably to other active approaches, not better than them. If you would rather run than sit, the evidence does not tell you that sitting is the superior choice.
How reliable is the evidence base?
Less than the volume of research suggests. The National Center for Complementary and Integrative Health states directly that "much of the research on these topics has been preliminary or not scientifically rigorous", and explains why drawing conclusions is hard: "Because the studies examined many different types of meditation and mindfulness practices, and the effects of those practices are hard to measure, results from the studies have been difficult to analyze and may have been interpreted too optimistically."
"Interpreted too optimistically" is a federal research body describing the field's own coverage of itself. It is a reasonable lens for any confident claim you encounter about what meditation does to the brain.
What is mindfulness-based stress reduction?
MBSR is the standardised program most of the good research studies, which is why it comes up so often. According to a peer-reviewed analysis of the program, "Participants engage in structured weekly meetings, each lasting ~2 to 2.5 h, over a period of 8 weeks", and "Additionally, a 5-h silent retreat is conducted during the sixth week." The content "combines elements of mindfulness meditation, Hatha yoga and body awareness practices to discover stress triggers and unhelpful automatic stress reactions".
| Element | What it involves |
|---|---|
| Duration | 8 weeks |
| Weekly session | Roughly 2 to 2.5 hours, in a group |
| Retreat | A 5-hour silent retreat during week six |
| Practices | Mindfulness meditation, Hatha yoga, body awareness |
| Between sessions | Daily home practice |
This is worth knowing because it sets expectations. The results in the literature come from a substantial commitment, not from five minutes with an app. That does not make shorter practice worthless, but it does mean you should not expect study-sized effects from a fraction of the study-sized dose.
Is meditation safe?
Mostly, with a caveat that is often skipped. NCCIH notes that meditation and mindfulness practices "usually are considered to have few risks", but immediately qualifies it: "However, few studies have examined these practices for potentially harmful effects, so it isn't possible to make definite statements about safety."
Where it has been looked at, the picture is not blank. NCCIH reports that in one review, "about 8 percent of participants had a negative effect from practicing meditation, which is similar to the percentage reported for psychological therapies. The most commonly reported negative effects were anxiety and depression."
The comparison is the useful part. Eight percent is not alarming, and it is roughly what psychological therapies produce. But it does mean meditation is an active intervention rather than an inert one, and that a bad reaction to it is a known phenomenon rather than evidence you are doing it wrong. If sitting with your own thoughts consistently makes you feel worse, that is a reason to stop and speak to someone, not to persist harder.
How do you actually start?
You need no equipment, no app, and no belief system. The practice is repetitive by design.
- Sit somewhere you can stay still for a few minutes. Upright and comfortable beats any particular posture.
- Pick one anchor. The feeling of the breath at the nose or the belly is the usual choice, because it is always available.
- Rest attention there. Not forcefully. Just notice the sensation.
- Expect to drift. Wandering is not failure, it is the thing you are practising with.
- Return without commentary. Noticing you drifted and coming back is the entire exercise, repeated.
- Stop before you want to. Ending while it still feels manageable is what makes tomorrow likelier.
Five to ten minutes daily is a sane starting point. Our piece on what meditation does to a busy brain goes further into what is happening while you sit.
Why is it so hard at the start?
Because the discomfort is the material. Most people arrive expecting a calm mind and find a noisy one, then conclude they are bad at it. Noticing the noise is what changes, not the noise itself. The frequent complaint of "I cannot stop thinking" describes the normal condition of the exercise rather than a failure to perform it.
How long before you notice anything?
The trial evidence measured outcomes at eight weeks, so it is reasonable to give it that long before deciding. Bear in mind that the measured effects were modest even then, and that they diminished at three to six months in the meta-analysis. Meditation looks less like a treatment with a completion date and more like a practice whose benefits track how much of it you are currently doing.
Who should be cautious?
- Anyone with a psychiatric condition, particularly where anxiety or depression is prominent, given that these were the most commonly reported negative effects.
- Anyone with a history of trauma, for whom extended silent attention can be difficult; a trauma-informed teacher is worth seeking out.
- Anyone treating meditation as a replacement for care they already need. The meta-analysis found no evidence it outperforms active treatments, so it is not a substitute for one.
If you want the calming effect without the sitting, breathwork produces measurable results faster, and movement performed comparably in the same body of research.
