Evidence desk
What does the research actually let us say?
“Phones are harmful” is too broad to guide a decision. Reducing leisure use at home, using a screen in bed, and restricting phones at school are different questions. These three studies illustrate why the setting, activity, and outcome matter.
Published by Digital Dopamine Dispenser · October 2, 2026 · Editorial approach
How this comparison was selected
This is a focused reading guide, not a systematic review or a complete survey of the latest research. We selected three primary studies addressing decisions covered by this site: a family agreement, an evening routine, and school restrictions. The trial summary uses the published article; the sleep and school summaries use the authors’ indexed abstracts. Follow each source to inspect its scope.
Each card separates the study result from our interpretation. The studies did not test this website, its tools, or its classroom materials. This page has not received independent clinical review.
Randomized trial · Denmark · 2024
Can reducing leisure screen use help?
Who and how long: 89 families, including 181 children and adolescents; a two-week intervention.
What was compared: Families were assigned to reduce leisure screen use or continue as usual. The reduction group was asked to limit leisure screens to three hours per week per person and hand over smartphones and tablets. This was a substantial family-wide change, not a test of muting one app.
What was found: The reduction group improved on a behavioral difficulties questionnaire compared with the control group. The estimated difference in change was −1.67 points, with a 95% confidence interval from −2.68 to −0.67.
What it cannot establish: Two weeks cannot establish lasting effects. A questionnaire outcome is not a diagnosis. Random assignment strengthens the causal interpretation for this intervention, but does not establish which part of the family-wide change produced the benefit.
Our interpretation: this supports taking displacement of other activities seriously. It does not make the study’s strict temporary limit a universal household prescription or validate our agreement worksheet.
Read the source: Schmidt-Persson and colleagues, JAMA Network Open (2024)
Repeated observations · New Zealand · 2024
Does it matter when screens are used?
Who and how long: 79 young people aged 11–14.9; four nonconsecutive nights.
What was compared: Researchers recorded evening screen activity with cameras and measured sleep using wrist-worn accelerometers. This captured activity in bed separately from activity earlier in the evening.
What was found: Screen use in the two hours before bed was not associated with most sleep measures that night. Use once in bed was associated with less sleep, particularly interactive activity and multitasking.
What it cannot establish: The study observed behavior rather than assigning a screen restriction. It cannot prove that removing a device would cause a particular sleep improvement. A small sample and a few nights also limit how widely the results should be generalized.
Our interpretation: ask about the last activity in bed, not just the daily total. Do not use this result to promise a specific number of extra minutes of sleep or to claim that all earlier evening use is harmless.
Read the source: Brosnan and colleagues, JAMA Pediatrics (2024)
Cross-sectional comparison · England · 2025
Are school restrictions enough on their own?
Who and how long: 1,227 students aged 12–15 across 30 secondary schools.
What was compared: SMART Schools compared students at 20 schools with restrictive policies and 10 with permissive policies. Schools were not randomly assigned a policy.
What was found: Restrictive policies were associated with less phone and social media time at school. Researchers found no evidence of a difference in mental wellbeing or overall weekday and weekend use between policy groups.
What it cannot establish: Comparing existing schools cannot isolate the policy from every other difference between them. No evidence of a difference is not proof that all possible policies have zero effect.
Our interpretation: separate the goal of a manageable lesson from a claim about overall wellbeing. A classroom rule may address a classroom problem while leaving evening use unchanged. Our lesson teaches design awareness; this study does not test that lesson.
Read the source: SMART Schools, The Lancet Regional Health – Europe (2025)
Why these results are not contradictory
The family trial changed leisure screen use throughout a household and measured short-term questionnaire outcomes. The school comparison examined existing policies and wellbeing across different schools. The bedtime study examined particular activities on particular nights. A benefit in one setting and no detected difference in another do not cancel each other out.
Before sharing a headline, finish this sentence: “In this group, this kind of use or change was linked to this measured outcome over this period.” If those details disappear, a narrow result can become a claim the research did not test.
A worked decision: the late group chat
Imagine a student who finishes homework online, then keeps answering friends after getting into bed. A daily screen total combines both activities. The bedtime study suggests a more useful question: what continues after getting into bed? It does not tell us whether this particular student’s tiredness is caused by messages.
A practical next step is to discuss a final check, preserve a reliable urgent-contact route, and make optional alerts quieter. Then ask whether the plan was workable and whether anything important was missed. This is a household observation, not an experiment that can establish a medical cause. Persistent sleep difficulties deserve individual support.
Build and test a notification planThree questions to bring to any new study
- Was behavior assigned or observed? Randomization helps separate the intervention from pre-existing differences; observation can show patterns without settling causation.
- What exactly changed? A phone setting, a school rule, and a whole-family restriction are not interchangeable interventions.
- What was measured? Fewer minutes, better questionnaire scores, and longer sleep are distinct outcomes. An improvement in one is not automatic evidence of improvement in all three.
Found an error or a study that changes this interpretation? Send the passage and source. Corrections should change the explanation, not merely refresh its date.