People in one study demanded about a hundred dollars to give up Facebook for a month — and the ones who did came out the other side happier, with an extra hour a day and far less desire to go back, as if they'd been guarding the very thing that was quietly wearing them down

In the autumn of 2018 a few thousand Americans sat down at their laptops and were asked to name the smallest sum they would accept to switch off Facebook for four weeks. Not a hypothetical. A real offer with real money behind it, built so that inflating your number would cost you the payout. A lot of them said a hundred dollars.

Economists had seen that coming. Pilot data suggested a pile-up at exactly $100, so Hunt Allcott, Luca Braghieri, Sarah Eichmeyer and Matthew Gentzkow pitched their offer at $102, one nudge above the round number, then randomised the 61 per cent of their 2,743 recruits who came in under it into a group paid to deactivate and a group left alone. What came out of it, published in the American Economic Review, remains one of the largest and most carefully incentivized single experiments on what a month off social media does to a person.

I spent years pricing things for people who told me they were too expensive and ordered them anyway. What someone says a thing is worth and what they do about it are two separate measurements, and the gap between them is where the interesting behaviour hides.

What an hour a day bought

Deactivation freed up sixty minutes a day for the average person in the paid group. Self-reported time on other social media fell alongside it, since Facebook turns out to work as a complement to other screen use rather than a rival for it. Actual Twitter activity, tracked from real tweet counts rather than a survey answer, ticked up slightly, though not significantly. It went offline, to television watched alone, to dinners out, to friends and parents.

Wellbeing improved. A combined index of happiness, life satisfaction, depression and anxiety rose by 0.09 standard deviations, which the authors benchmark at roughly a quarter to two fifths of the effect of standard psychological interventions such as group training and one-to-one therapy. Daily text-message checks on mood pointed the same way without ever reaching significance.

Something was lost as well. News knowledge fell by 0.19 standard deviations, a bigger movement than the wellbeing gain, and self-reported time spent on news dropped by around fifteen per cent. The overall polarisation index fell too, by 0.16 standard deviations, driven mostly by reduced exposure to news that flattered people’s own side and a narrowing of views on policy issues.

Happier, calmer, less informed, less furious.

What people asked for the second time

“I realized how much time I was wasting,” one participant wrote afterwards. Everyone was asked a second time to name a price for another month away, and those who had actually served their four weeks revised their figure down by roughly fourteen per cent against the control group, which is a polite way of saying they had been quoting a price for something they had never costed.

They also cut back with no payment on the table. A month after the study closed, the treatment group reported about twelve fewer minutes a day on the Facebook mobile app, close to a quarter less than controls. Nine weeks out, five per cent of them still had their accounts switched off, against two and a half per cent of those who had never left.

Nobody was paying them by that point, and they still did not go back as hard.

How much weight one experiment can carry

Does that settle the question? Not even slightly.

A much larger follow-up, run before the 2020 election across tens of thousands of accounts and written up in a National Bureau of Economic Research working paper, found the same direction at a smaller size: six weeks off Facebook improved an index of happiness, depression and anxiety by 0.060 standard deviations, and six weeks off Instagram by 0.041. Facebook’s effect was concentrated in users over 35, Instagram’s in women under 25.

Widen the lens and the picture dissolves. A preregistered meta-analysis led by Laura Lemahieu, published in Scientific Reports, pooled ten abstinence experiments covering 4,674 adults and found no significant effect on positive affect, negative affect or life satisfaction. Their conclusion is admirably deflationary: quitting social media for a while does not appear to make you feel better, and does not make you feel worse either. Most of those trials ran a single week on small samples, several with compliance nobody could verify, and one in which fewer than 14 per cent of participants lasted the full seven days.

Allcott’s team turned the same scepticism on themselves. Inside their own baseline data, the raw correlation between heavy Facebook use and low wellbeing was about three times larger than the causal effect they went on to measure. Anyone quoting the correlation is quoting an inflated number, which covers most of what gets written about screens and misery.

Why the exit price is not the product price

Ask a thousand American university students what they would need to be paid to quit TikTok for four weeks and the average answer comes back at $55. Ask them instead what they would pay to have every TikTok user on campus quit at once, themselves included, and they will hand over $24. Leonardo Bursztyn, Benjamin Handel, Rafael Jiménez-Durán and Christopher Roth ran precisely that experiment, with the final version now out in the American Economic Review: 60 per cent of active TikTok users and 46 per cent of active Instagram users end up worse off for the platform existing at all, once you count what non-users lose by being left out.

Asked why they keep using something they would prefer did not exist, students named fear of missing out more than any other reason. Told that their classmates would be deactivating alongside them, they valued their own TikTok accounts about a third less. Bursztyn and his co-authors call this a product market trap, and they find the same shape in luxury handbags and annual iPhone releases, which suggests the mechanism has nothing to do with feeds and everything to do with standing in a room where everyone else is holding something.

The thing nobody is selling

Digital detox as a product assumes the problem is your relationship with an app. Four weeks of solitary abstinence buys an hour a day and a small lift in mood that may not survive contact with a well-powered meta-analysis, and the students Bursztyn surveyed were clear that this is not the arrangement they want. They want the coordinated version. There is no shop that sells it and no company with any reason to open one.

What sticks with me is that the whole thing had to be paid for. Four weeks costs nothing and requires no equipment, 80 per cent of the people who did it called it good for themselves afterwards, and it still took $102 a head to get them out the door. That figure has been reported ever since as the price of Facebook. It looks a lot more like the price of the door.