What a meditation app is actually worth
Pooled across forty-five trials, the apps beat doing nothing. Set against anything else that takes the same twenty minutes, the advantage disappears.
There is one app in this shop, and it would be dishonest to sell it without writing this. The evidence for meditation apps is real, it is consistent, and it is much smaller than the category's marketing implies.
What the pooling shows
Linardon and colleagues pooled forty-five randomised controlled trials of mindfulness apps in Clinical Psychology Review in 2023. Against control conditions, the apps improved:
- Depression, at a Hedges g of 0.24.
- Anxiety, at a Hedges g of 0.28.
Those are small effects. In this measure 0.2 is small, 0.5 is moderate and 0.8 is large. They are also positive, replicated across dozens of trials, and achieved by something that costs a few pounds a month and asks nothing of a health service. Small and real is not nothing.
The comparison that changes the answer
Most of those trials compare the app against a waiting list. A person in a waiting list group gets nothing at all, and is aware of getting nothing at all, for the length of the study.
A smaller number of trials compared the app against an active therapeutic comparison, meaning some other thing that also took time and attention. In those comparisons the effects were not statistically significant: a Hedges g of -0.15 for depression and 0.10 for anxiety. The authors note the number of such studies was low, so this is a weak finding rather than a refutation.
But the direction is the point. What survives comparison with nothing does not obviously survive comparison with something.
Nobody has tested this particular app
The trials in that literature mostly test Headspace and Calm. There is no independent randomised controlled trial of Waking Up, the app in our shop. There is one pre-registered evaluation published outside peer review, and that is all.
So the honest statement about the product we link to is this: the category it belongs to has small, replicated effects against doing nothing, and this specific member of the category has not been tested at all. We cannot transfer the category's result onto it and we are not going to pretend otherwise.
The other thing the trials keep finding
Attrition and engagement are the chronic problem in this field. People download the app, use it for a while and stop, and the trials lose them. That is not a flaw in the research so much as a finding about the thing being researched: the practice works if you do it, and an app does not by itself make somebody do it.
What we do with this
We keep the app in the shop, because a thirty-day course that teaches attention as a skill is a reasonable purchase and the instruction is good. We describe it as instruction rather than treatment, and we do not attach a number to it.
And we say the cheaper thing out loud. The Monday Service is free, it is an hour a month, and it has never been tested either. If you want to find out whether sitting still does anything for you, the version that costs nothing will answer that question as well as the version that costs money.
