Does Meditation Actually Lower Your Stress Levels?

12 October 2026 · 7 min read

Does meditation actually lower stress?

Every meditation app on your phone makes this claim, and usually in the same breath as a photograph of a lake. So it is worth asking what has actually been measured, and how big it is.

Start with the cleanest evidence we have on how people feel.

In 2023 a collaboration of trialists did something unusual. Instead of pooling published summaries, they collected the raw data on every individual participant. The result was an individual participant data meta-analysis of 13 trials and 2,371 people across eight countries, median age 34, moderately distressed on average.

Compared with people who got nothing, those who completed a mindfulness programme reported less psychological distress one to six months later. The effect was a standardised mean difference of -0.32, with a confidence interval from -0.41 to -0.24. The authors call that small to moderate, and they rate their confidence in it as high.

Two things are worth holding onto. There was no detectable heterogeneity, rare in this field, meaning the trials broadly agreed. And the comparison was against passive controls, people on a waiting list, which is the generous one. These were in-person, teacher-led courses, not a phone.

So: a real effect, modest in size, measured against nothing.

What happens to the stress hormone?

This is where the claim gets interesting, because cortisol does not care how you feel about your week.

In September 2026 a team published a meta-analysis of 39 mindfulness-based stress reduction studies covering 2,059 participants, including only trials that controlled for the time of day the cortisol was collected. Cortisol swings across the day, so that restriction matters.

Compared with control conditions, MBSR did lower cortisol, with a between-group effect of g = -0.38 across 34 comparisons.

Then the detail that tells you more than the headline. Against inactive controls, the effect was -0.42. Against active controls, people who were given some other plausible thing to do, it was -0.29. And in the studies that had no control group at all, the before-and-after change was -0.06, which is nothing.

Read that sequence in order. The less the comparison group did, the better meditation looked. That is not a reason to dismiss the finding, because the effect survived in randomised trials and against active controls. It is a reason to distrust any number from an uncontrolled study, which is most of the numbers in app marketing.

The authors also report that significant heterogeneity remained unexplained. The trials do not agree on the size of this effect.

Why does the answer change depending on who you ask?

Because "cortisol went down" is not one measurement. It is a family of them, and they do not move together.

A 2018 review of nine workplace mindfulness studies found the programmes lowered high and low diurnal cortisol slopes, the shape of the hormone's daily curve. But they found no change in the cortisol awakening response, and no change in single morning or evening readings. Same people, same programmes. The answer depended entirely on which sample you drew and when.

Who is meditating matters just as much. A 2021 meta-analysis split trials by whether participants were at risk of elevated cortisol. Across 10 studies using blood samples, meditation produced a significant medium effect against control groups, but on closer inspection that effect was present only in at-risk samples, specifically patients with a somatic illness. Across 21 studies using saliva, the effect was small and not significant overall, with only a marginally significant effect for people living in stressful life situations.

The pattern the authors describe is that meditation does most for the people who need it most. Which is both encouraging and a caution: if you are a reasonably well person having a reasonably normal week, the physiological headroom for improvement is smaller.

The honest counterweight to all of this is a 2017 meta-analysis of 45 randomised trials that used active control groups only. Even held to that stricter comparison, meditation reduced cortisol, C-reactive protein, blood pressure, heart rate and triglycerides when all forms were analysed together. Focused attention practices, the kind a timed five-minute sit actually is, were the subtype that reduced cortisol.

So the physiology is not a myth. It is just smaller, noisier and more conditional than the lake photograph suggests.

Does any of this apply to five minutes on a phone?

Here is the gap. Almost all the cortisol evidence above comes from mindfulness-based stress reduction, an eight-week course with weekly group classes and long daily homework. Nobody in that literature tested five minutes a day.

What has been tested at something close to a realistic daily dose is how stressed people report feeling.

In 2025 a trial gave 138 university employees either the instruction to meditate for 10 minutes a day with an app, or a waiting list. Rather than asking once at the start and once at the end, the researchers pinged people five times a day for four days at a time, collecting 6,260 real-time reports of stress during ordinary life.

By week two, the meditators reported less subjective stress and less perseverative thinking. By week five they reported better coping. The effects held through week eight.

And the part that is easy to miss: over the same period, the control group got worse. Their reported stress and rumination rose and their sense of coping fell. Some of what the practice bought was not improvement. It was not sliding.

An earlier trial makes the adherence question concrete. A study of 238 employees at two UK companies offered 45 guided meditations of 10 to 20 minutes and asked for one a day for eight weeks. Average number completed: 17, with a range from 0 to 45. Barely a third of what was asked. The intervention group still reported significantly better well-being, less distress and less job strain than controls, with well-being and job strain still better 16 weeks in. The workday blood pressure change was only marginal.

Seventeen sessions over eight weeks is not a discipline. It is a normal human being with a job. It was enough to show up in the numbers.

Does it matter which app you use?

Probably less than the industry would like.

A 2025 Bayesian network meta-analysis in Nature Human Behaviour pooled 63 randomised trials and 20,454 participants on self-administered mobile stress interventions. Mindfulness meditation ranked among the top approaches, alongside structured stress management programmes and problem-solving therapy.

The finding to sit with is a negative one. There was no conclusive evidence that adding human support or more elaborate mobile technology improved outcomes. The authors are careful about the statistical power behind that analysis, so do not over-read it. But nothing here suggests the feature list is what helps.

So what is honest to say?

Three sentences, and they are the ones we stand behind.

Meditation reliably reduces how stressed people report feeling, by a small to moderate amount, and that holds up in the best-controlled evidence we have. It also moves the hormone, by less, mostly in people who are already strained or unwell, and only when you pick the right cortisol measure. And nobody has tested five minutes a day against cortisol, so anyone promising you a hormonal result from five minutes is guessing.

What five minutes has in its favour is the thing all this research quietly depends on. Every effect above required people to keep doing it, and the trials show they mostly did not. A dose you will actually take beats a dose you abandon in week two.

One thing to do today

Stop measuring this in hormones, and give it two weeks.

Two weeks is not arbitrary. It is where the app trial saw subjective stress and rumination separate from the control group. So sit today, at the same point in your day, for five minutes. Then pick a day a fortnight out and ask one honest question: in a bad moment this week, did I spend less time going round the same thought?

That is the outcome the research actually found. Not a calmer bloodstream. A little less time stuck in the loop, and a slightly better sense that you can handle the day. In this app that is one Session, the Daily Five, and then you close the phone and get on with it.

This article shares general information about meditation research. It is not medical advice. If you have concerns about your mental health, talk to a qualified professional.