Delete when the app is the whole problem and you can afford to lose what's inside it. Limit when your friendships, your work or your family live there. The best-evidenced version is a fixed break with an end date rather than a permanent goodbye, because the trial that measured one found the effect lasted well past the four weeks.
The honest comparison
| | Deleting or deactivating | Limiting | |---|---|---| | What it's good at | Breaking the reach-for-it reflex fast. The strongest measured effect on later use. | Keeping the contacts, the group chats and the work you actually need. | | What it costs you | Real access. One trial measured a drop in factual news knowledge alongside the benefits. | Slower. You stay in daily contact with the thing you're trying to use less. | | How it usually fails | You reinstall inside a week and feel worse than before you tried. | The limit quietly loosens until it isn't a limit. | | Best for | People whose use is mostly passive scrolling with nobody on the other end. | People whose social life, income or family contact runs through the app. | | Make it work by | Setting an end date up front, and changing the condition that lets it back on. | Handing the passcode to someone else so the loosening needs a conversation. |
What a four-week deactivation actually did
The cleanest evidence for a time-boxed break comes from a randomised experiment that paid people to deactivate Facebook for the four weeks before the 2018 US midterm election. Deactivating ["reduced online activity, while increasing offline activities such as watching TV alone and socializing with family and friends"](https://www.aeaweb.org/articles?id=10.1257/aer.20190658). It ["increased subjective well-being"](https://www.aeaweb.org/articles?id=10.1257/aer.20190658). And the finding that matters most if you're deciding today, it ["caused a large persistent reduction in post-experiment Facebook use"](https://www.aeaweb.org/articles?id=10.1257/aer.20190658).
That last one is the argument for a break over a rule. The effect outlasted the four weeks. People came back and used it less, having recalibrated what it was worth to them.
There was a cost, and it belongs in this paragraph rather than a footnote. Deactivation ["reduced both factual news knowledge and political polarization"](https://www.aeaweb.org/articles?id=10.1257/aer.20190658). You will know less about what happened. For some people that's the point. For others, particularly anyone whose community organises through those platforms, it's a genuine loss worth planning around.
And the trial where everyone got better
Here's the counterweight, and you should have it before you build your hopes too high.
A 14-day social media abstinence experiment randomised people into an abstinence group and a control group. Screen time and body image dissatisfaction improved for the group that stopped. But depression and anxiety symptoms, problematic smartphone use, fear of missing out and loneliness ["showed a decrease during the overall intervention time but no difference between the investigated groups could be observed"](https://pubmed.ncbi.nlm.nih.gov/38481298/).
Everybody improved. Including the people who kept scrolling.
A [meta-analysis of 10 detox studies](https://pubmed.ncbi.nlm.nih.gov/39280291/) lands in a similar place: a significant effect on depression (standardised mean difference -0.29), and nothing significant for life satisfaction, stress or overall wellbeing.
Take the break anyway. Just be precise about what you're buying. Two weeks off gets your hours back and, in that study, a better relationship with how you looked. It did not beat simply being in a study where somebody paid attention to how you felt. If your mood is the frightening part, the app is unlikely to be the whole answer, and that's worth taking to a GP rather than to your settings screen.
The reinstall loop
This is the part almost nobody writes about, and it's the most common thing people are actually living.
You delete it. You feel briefly excellent. Twenty-four hours later you're logging back in, and the account is exactly where you left it because deactivation was designed to be reversible. So now you're back and ashamed, which is a worse position than the one you started from.
Nothing has gone wrong with you. Four things went wrong with the setup.
The first is that deleting an app removes the icon. It leaves the account, the password in your keychain, and the version in your phone's browser that logs you straight in. The door is still there. You took the handle off.
The second is that you deleted at a peak. Almost every deletion happens at 1am after a bad session, which is the moment you have the least information about what you'll want on Thursday afternoon when you're bored at work. Decisions made at the worst moment tend to be reversed at the next ordinary one.
The third is the gap. Two hours a day vanished and nothing moved in. Boredom is a real physical state rather than a failure of character, and an empty evening is the most reliable reason people reinstall.
The fourth is that the moment itself never changed. Same chair, same time, same restless half hour, and the app was the only thing that ever went there.
So change the condition rather than trying harder:
- Get the password out of your own hands. A long random one, generated, not memorised, held by someone you trust or locked behind a manager you have to deliberately open.
- Delete the account properly, or don't pretend you did. Most platforms have a real deletion path with a waiting period. Deactivation is a pause, and it's fine to choose a pause, as long as you call it one.
- Block it at the device level, where the home screen can't help you. [Shield](/shield) is the part of Renovyn that keeps the block in place when you're the one arguing against it. Screen time limits with a passcode somebody else holds work the same way.
- Decide in advance what the two hours become, and make it something with a start time. "Read more" fails. "Gym at 7 on Tuesday and Thursday" survives.
If you're reading this having already reinstalled, that's an ordinary event. It isn't a verdict on you. [What to do after you reinstall](/guides/social_media/you-reinstalled-the-app-what-now) picks up exactly there.
If your friendships live in the app
Some people cannot delete, and telling them to is useless advice. Your group chat is there. Your family in another country is there. Your freelance work comes through the direct messages.
The move there is to keep the channel and kill the feed. Turn off every notification except direct messages. Use it in a browser, where the experience is deliberately worse. Unfollow every account you don't know personally, which usually empties the feed within an hour and leaves the people intact. Give it one slot in the day and let the rest be quiet.
That's limiting done honestly, with a different aim from deleting: keeping what the platform is genuinely good for, removing the part that eats evenings.
Still unsure which side of this you're on? [Work out whether your use is a problem](/guides/social_media/is-my-social-media-use-a-problem) first.
Common questions
**Did deleting Instagram change people's lives?** The measured version of that question is the Facebook deactivation experiment, where four weeks off raised subjective wellbeing, moved time to offline activity, and cut later use well after the study ended. It also cut how much people knew about the news. So: a real effect, a modest one, with a trade-off attached.
**Why do I reactivate within 24 hours every time?** Because deactivation is built to be undone, your password is still saved, and nothing has filled the hours you freed up. The loop breaks when you change the conditions rather than your resolve. Move the password out of reach and put something with a fixed start time into the gap.
**How do I stop the addiction after I've already deleted the accounts?** Deleting removes the access. The reaching stays for a while. Expect a couple of weeks where your hand goes for the phone out of habit and finds nothing, which is uncomfortable and passes. [Getting your attention back](/guides/social_media/getting-your-attention-back) covers what that stretch feels like and what shortens it.
**How long should the break be?** The experiment with the most persistent effect used four weeks. The 14-day trial found gains in screen time and body image, though not more than the control group on mood. Pick a length you'll finish, put the end date in the calendar, and treat finishing as the goal rather than never going back.
Sources
- [The Welfare Effects of Social Media, American Economic Review, 2020](https://www.aeaweb.org/articles?id=10.1257/aer.20190658). The four-week Facebook deactivation, the increase in wellbeing and offline activity, the reduction in news knowledge and polarization, and the persistent drop in later use.
- [Effects of a 14-day social media abstinence on mental health and well-being, BMC Psychology, 2024](https://pubmed.ncbi.nlm.nih.gov/38481298/). Screen time and body image improving in the abstinence group, and mood, problematic use and loneliness improving in both groups with no difference between them.
- [Impacts of digital social media detox for mental health, Narra J, 2024](https://pubmed.ncbi.nlm.nih.gov/39280291/). Ten studies, a significant effect on depression, and no significant effect on life satisfaction, stress or overall mental wellbeing.
