12 evidence reviews

Check the research

Popular claims about phones, attention and policy, checked against the primary studies behind them rather than the summary that usually circulates. Where the evidence is mixed, these say so instead of rounding it into a confident answer. Seehow we research for the sourcing standard behind that.

Is Phone Addiction Real? What the Research Actually Sayschecked oct 2026Phone addiction is not a recognised diagnosis, and how common it looks depends almost entirely on the definition: about 27 per cent in screening studies, about 1 per cent once real harm is required. Where the line sits, and when to see someone.7 min read Break Up With Your Phone: Does Catherine Price's Plan Work?checked sep 2026How to Break Up With Your Phone sells a 30-day plan, and no trial has ever tested it. Graded day by day against the research, some weeks are well supported and the first one is the control condition.7 min read How to Reduce Screen Time: 6 Ways, Ranked by the Evidencechecked sep 2026Six ways to reduce screen time, ranked by the strength of the evidence behind each one rather than by how often they get repeated. One of the most-quoted tips does not survive its own meta-analysis.9 min read Is Video Game Addiction Real?checked sep 2026The WHO recognises gaming disorder. The American Psychiatric Association still does not. What separates a real video game addiction from a teenager who plays a lot.12 min read Do Blue Light Glasses Actually Work?checked sep 2026Blue light glasses block far less blue light than the marketing implies, and seventeen randomised trials covering 619 people have not shown they relieve eye strain.10 min read Does a Digital Detox Actually Work?checked sep 2026Two preregistered randomised trials found real improvements in mood, sleep and stress during a detox, and both found screen time rebounding to near baseline once it ended. What that makes a detox good for.4 min read How Much of Teen Mental Health Is Actually About Screens?checked sep 2026What the evidence links social media to in teen mental health, and how weakly. Across 355,358 adolescents the association with wellbeing accounts for at most 0.4% of the variation, and a 2025 causal analysis ranks peer relationships far above screens.4 min read Do Apple Screen Time and Digital Wellbeing Actually Work?checked sep 2026Tracking your screen time reliably changes what you know about it and rarely changes the number. What the field studies and the one randomised trial actually found, and which tools moved the figure.5 min read Do Workplace Phone Bans Actually Work?checked sep 2026The evidence is genuinely mixed: the one real workplace field experiment found bans raise effort but not output, the one meeting-perception study found a written ban barely beats having no policy, and the largest evidence base, from schools, disagrees with itself.8 min read Do School Phone Bans Actually Work?checked sep 2026Do school phone bans work? The best US evidence found a real test-score gain and a racially uneven suspension spike; the best UK evidence found no benefit to wellbeing, grades or behaviour at all. What each one actually measured.8 min read Does the Right to Disconnect Actually Work?checked sep 2026Most "right to disconnect" laws require your employer to negotiate a policy, not to leave you alone, and the one country that has studied its own law found 97% of employees saw no observable change. The court case everyone cites as proof of enforcement, read directly, turns out not to be about the right to disconnect at all.8 min read Phone Addiction Statistics 2026: What the Research Showschecked sep 2026The phone addiction statistics quoted online are routinely wrong, out of date or misread. Here are the figures that survive checking, each one linked to its source, and the popular ones that do not.8 min read

Definitions of the terms these reviews use are in the glossary. Looking for a tool instead of a claim to check? Head to the comparisons.

What “checked” means here

It means someone opened the study a claim is attributed to and read what it reports, rather than repeating the summary that circulates. That sounds like a low bar. It is not one this subject usually clears, and the reviews here exist because of what turns up when you try.

Four failure modes come up repeatedly:

The citation that does not contain the claim. The most-quoted number in this field, that it takes about 23 minutes to recover from an interruption, traces to a 2006 interview rather than to a paper, and the study usually credited for it reports something close to the opposite.
The finding that decayed. Interventions are often reported at their first-week effect. Where a study followed people longer, the effect usually shrank, and a headline that quotes the peak is describing the best week rather than the result.
Correlation described in causal language. Most of the evidence linking screens to wellbeing is cross-sectional. Distressed people may use their phones more, which fits the same data as the reverse.
The sample that cannot carry the claim. A result from 143 undergraduates over three weeks is worth knowing and is not a fact about the general population, so sample sizes are recorded alongside each source.

Questions worth asking of any number you meet

Whether or not it appears on this site: does the article link to a study rather than to another article? Does the study measure what the headline says it measures? Is a correlation being described with causal words? And what was the sample? A number that fails those four questions probably should not change your behaviour, and a claim that never links to a study is usually quoting someone who was quoting someone.