Does Blocking the Internet on Your Phone Actually Help?

In a study published in early 2025, 467 people in the US and Canada agreed to block all internet on their phones for two weeks. Calls and texts still worked. Laptops still worked. Just nothing online on the phone.

119 of them kept the block on for at least 10 of the 14 days.

That’s the number this post is about. But first, the study, because it’s better than most.

A good study, for once

The paper, by Noah Castelo and four colleagues, appeared in PNAS Nexus in February 2025. Most research on phones and mental health is correlation: heavy users are more anxious, but anxious people may simply reach for their phones more. This one was an experiment.

Participants were iPhone users recruited online, average age 32. They were split at random. One group blocked straight away for two weeks. The other carried on as normal for two weeks, then blocked. That gives a clean comparison for the first fortnight. The block was an app called Freedom, in a locked mode, and the app logged whether it was on each day. So compliance was recorded, not self-reported. Attention was measured with a five-minute test, not a questionnaire. The analysis plan was filed in advance.

The results. The waiting group didn’t change on anything. The blocking group improved on all three measures: mental health, wellbeing and sustained attention. Their screen time, taken from screenshots, fell from 5 hours 14 minutes a day to 2 hours 41.

The line that travelled is the weakest one

The paper says the mental-health effect “was larger than the meta-analytic effect of antidepressants.” Headlines ran with it: This Simple Phone Hack Can Improve Mental Health As Much as Antidepressants.

It doesn’t hold up as a comparison:

• The phone number isn’t depression. It’s a combined score covering depression, anxiety, anger, social anxiety and personality functioning.
• It’s a before-and-after change. The antidepressant figure the authors cite, from Kirsch and colleagues (2008), is drug minus placebo. In those trials the placebo groups got more than 80% of the improvement the drug groups did. The 0.32 that’s left is what remains after that’s subtracted. A raw before-and-after change will almost always look bigger.
• These weren’t patients. They were volunteers who wanted to use their phones less.

The authors also say the attention gain matched about ten years of age-related decline. That’s their comparison, drawn from an earlier study; treat it as theirs. Attention was the smallest of the three effects, and just under six in ten people improved on it.

The compliance number is three numbers

“A quarter complied” is true, but it hides three different things:

Step	People
Agreed to block	467
Downloaded the blocking app	272
Set it up and finished all the surveys	249
Kept the block on for 10 of 14 days	119

So about four in ten never downloaded the blocker at all. The paper doesn’t say why. Of the people who did set it up and stayed, just under half made it. And the bar was 10 days out of 14, not the “full two weeks offline” some coverage reported.

The people who didn’t comply improved too. That’s either the most useful finding in the paper, or the most worrying.

The people who didn’t do it got better anyway

Here’s the part that matters most. The headline results come from everyone who finished the surveys (313 people), whether they kept the block on or not. That includes about 60 who never switched it on at all.

Looking at only the 119 who complied, the effects were larger, but not by much. On mental health, the effect size went from 0.56 to 0.68. On attention, from 0.23 to 0.26. The authors say even those who didn’t fully comply saw “significant, though more modest, improvements.” They don’t give a separate number for them.

There are two ways to read this.

Reading one: a little is enough. This is the authors’ own view: “simply reducing mobile internet use may be sufficient.” There’s good support for it. The blocking group’s screen time halved, including people who didn’t stick to it. And the waiting group, who were just as keen to cut down, managed only 14 minutes less a day on their own (5 hours 36 to 5 hours 22). Wanting to use your phone less barely moved anything. A leaky block did. The improvements in wellbeing also tracked real changes in what people did: more time with people in person, more exercise, more time outdoors, more sleep.

Reading two: some of it wasn’t the block. These volunteers signed up for a study about smartphones and wellbeing, and 83% said they wanted to cut down. There was no “active” control: no group doing something that looks helpful but shouldn’t be. A waiting list controls for time passing, not for hope. People who were told their turn was coming weren’t expecting to feel better yet. The authors acknowledge that placebo and demand effects “could have contributed.”

The attention test is the best answer to reading two. It’s much harder to shift by feeling optimistic than a questionnaire is, and it did improve. But it was the smallest change, and none of the things the authors measured (time outdoors, sleep, socialising) explained it.

A useful comparison, used carefully

Allcott and colleagues (2020) paid people US$102 to deactivate Facebook for four weeks around the 2018 US midterms, and kept checking. Accounts were off at about 90% of checks. Wellbeing rose by 0.09 standard deviations.

That’s much smaller, but don’t divide one number by the other. It’s one app rather than all phone internet, and it’s a comparison between groups rather than each person against themselves. What it does show is that when you pay people and check, compliance is high.

Who could do it?

The question in the title is the one the paper can’t answer. It doesn’t report how the 119 differed from the rest. It does report that the people who finished all the surveys started out with better mental health and attention than those who dropped out. If the people who could keep the block on were also the ones who needed it least, the study measures how much it helps people who were going to be fine. Nobody knows yet.

After the block came off, the first group’s wellbeing and mental health stayed above where they started for at least two weeks. Their screen time rose again, to 4 hours 25, but stayed well below the 5 hours 14 they began with.

---

The study is good, and its most quoted line is its weakest. The real finding isn’t that three-quarters couldn’t do it. It’s that you didn’t have to do all of it to get much of the benefit. That’s either good news about how little it takes, or a sign some of the benefit was never the block. This study can’t tell which.

---

Jurisdiction: participants were iPhone users in the US and Canada, recruited online, and most already wanted to cut down. The Facebook study’s payment is in US dollars. Nothing here is from New Zealand.

Sourcing note: two weeks of blocking and two weeks of follow-up. The mental-health measure is a symptom questionnaire, not a diagnosis. The paper reports no separate result for the people who didn’t keep the block on, and doesn’t say how they differed from those who did.

---

Sources: Castelo, Kushlev, Ward, Esterman & Reiner, “Blocking mobile internet on smartphones improves sustained attention, mental health, and subjective well-being,” PNAS Nexus, 2025 · Kirsch et al., “Initial severity and antidepressant benefits: a meta-analysis of data submitted to the Food and Drug Administration,” PLoS Medicine, 2008 · Allcott, Braghieri, Eichmeyer & Gentzkow, “The Welfare Effects of Social Media,” American Economic Review, 2020.