7.2 — What It Actually Does: the Honest Evidence

Meditation research has a quality problem, and you should know its shape before reading any claim about it, including the ones on this page.

Many studies are small. Many have no proper control group — and comparing meditation against doing nothing tells you nothing, because expecting to improve produces improvement (1.3). Many rely on self-report from participants who know which group they are in and have usually been told what to expect. And publication bias is strong, because a study showing meditation does nothing is harder to publish and less interesting to fund.

A large review commissioned by a United States government agency in 2014 assessed nearly 19,000 records, found 47 trials with reasonable methods, and concluded that mindfulness programmes showed moderate evidence of small to moderate improvements in anxiety, depression and pain, and little or no evidence of improvement in mood, attention, substance use, eating, sleep or weight compared with active control conditions.

That is a real result. It is far smaller than the popular presentation, and it is worth holding both halves: meditation does something measurable, and it is not transformative on the timescale that is usually implied.

What the evidence supports

Sorted by strength.

Anxiety, depression and pain: small to moderate effects, comparable to other active treatments in some comparisons. This is the strongest area. Structured eight-week programmes — Mindfulness-Based Stress Reduction, and Mindfulness-Based Cognitive Therapy — have the best evidence, because they are manualised, taught by trained instructors, and have been tested against real comparison conditions.

Preventing depression relapse: genuinely good evidence. Mindfulness-Based Cognitive Therapy was designed for this specific purpose and is recommended in clinical guidelines in several countries for people with recurrent depression. In trials it performs comparably to maintenance antidepressants for relapse prevention in some populations. This is the single strongest claim in the field, and it is notably narrower than "meditation is good for you".

Emotional reactivity: modest reductions, with sustained practice. Both self-reported and, in some studies, physiological.

Sustained attention and mind-wandering: modest improvements. Reasonably consistent. This is the mechanism most relevant to the rest of this volume.

Stress markers: mixed but broadly positive. Some studies find reduced cortisol and blood pressure; the effects are small and not universally replicated.

What the evidence does not support

Dramatic brain changes from short practice. Much-publicised imaging studies reported structural differences in long-term meditators, and some found changes after eight weeks. Several of these have failed to replicate, including a well-designed larger study that found no structural brain changes after an eight-week course. Treat all "meditation rewires your brain in eight weeks" claims as unsupported.

Improved general cognitive ability. Same transfer problem as brain-training (1.9).

Meditation as a treatment for serious mental illness on its own. It is an adjunct, and for some conditions it is contraindicated.

Large effects from apps used occasionally. Most of the evidence comes from eight-week structured programmes with a teacher and roughly forty-five minutes of daily practice. An app used for ten minutes three times a week is a different intervention and should not be expected to produce the same results.

And the effect on prosocial behaviour — that meditation makes people kinder — is genuinely mixed. Some studies find it, a substantial meta-analysis found that effects largely disappeared when the control conditions were properly designed and when the study author was also the intervention teacher.

The dose question

How much practice, for how long, produces the effects above?

The clinical trials use roughly forty-five minutes a day for eight weeks, with a weekly class. That is the dose behind most of the positive findings, and almost nobody outside a trial does it.

Below that, the evidence thins quickly. There is some support for shorter daily practice producing smaller benefits, and it is much weaker.

The honest position: ten minutes a day is worth doing, will produce something, and should not be expected to produce the trial results. And ten minutes a day sustained for two years is a different thing from ten minutes a day for three weeks, which is the pattern most people actually manage.

Where it can do harm

Under-reported, and it matters.

Adverse effects occur in a meaningful minority. Reviews of meditation-related adverse events find rates that vary widely by study and population, but the phenomenon is real and not rare. Reported effects include increased anxiety, depersonalisation — the sense of being detached from yourself or watching yourself from outside — intrusive images, and re-experiencing of traumatic memories.

The highest-risk situations:

Trauma history. Sitting still with closed eyes and attention turned inward removes the external distraction that has been keeping traumatic material at bay. Standard mindfulness instructions can precipitate flashbacks, and this is common enough that trauma-sensitive adaptations exist for exactly this reason (7.9).

Long intensive retreats. Ten days of silence with many hours of daily practice is a substantially stronger intervention than a daily sit, and adverse reactions are more common there.

Current psychosis or a history of it. Intensive practice can destabilise and is generally not recommended without clinical supervision.

Severe depression. Turning attention inward during a severe episode can amplify rumination rather than interrupting it. Mindfulness-Based Cognitive Therapy is designed for people in remission, between episodes, not for people who are currently severely depressed.

None of this means avoid it. It means the framing of meditation as universally safe and appropriate for everyone is wrong, and 7.9 covers what to watch for and what to do.

Which practice for which purpose

Since the practices differ, matching them matters.

For catching your own state early — the capacity underlying Parts 4, 5 and 6: breath focus and noting (7.3, 7.5).

For physical tension and interoception: body scan (7.4).

For a hostile or self-critical inner life: loving-kindness (7.6). This one has its own literature and its effects are somewhat distinct from the attention practices.

For sleep onset: body scan or progressive muscle relaxation (7.8).

For acute stress right now: breathing with extended exhale (6.2), which is a physiological intervention rather than a meditation and works in ninety seconds.

And for preventing depression relapse, specifically: the structured eight-week MBCT programme with a trained teacher, because that is the intervention that was tested.

The honest summary

Meditation is a real, modest, slow-acting intervention that improves the capacity to notice what your mind is doing, with knock-on effects on anxiety, low mood and reactivity. It is not a cure for anything, the brain-change claims are overstated, the industry around it oversells consistently, and a minority of people are made worse by it — particularly those with trauma histories doing intensive practice without support.

Worth doing, at a small daily dose, sustained over years, for the specific capacity it trains. That is the claim this Part makes, and the following pages give you the actual practices.

What to do with this

Set your expectations to the evidence. Ten minutes a day will not transform you in a month. It will, over a year, make you slightly better at catching your own state, which compounds because everything else in this volume depends on that.

Choose the practice that matches what you want, from the list above, rather than defaulting to whatever an app opens with.

If you have a trauma history, read 7.9 before you start, and consider starting with eyes open, shorter sessions, and an external anchor like sound rather than an internal one.

And judge it by an outcome outside the sessions, not inside them. Whether a session felt calm is not the measure. The measure is whether, three months in, you catch yourself getting irritated slightly earlier than you used to. That is what is being trained, and it is the only thing worth checking.

Next: 7.3 is the first practice, with a full script you can follow tonight — the breath, and the four variations of it that do different jobs.