A three-week break from smartphones and social media was associated with longer self-reported sleep and lower stress in a small randomised trial of teenagers. But the gains had faded by a two-month follow-up, and wearable devices did not confirm a comparable improvement in sleep duration. The findings offer a measured perspective on a popular proposal: digital detoxes may help temporarily, but lasting benefits remain unproven.
A promising change that did not last
For three weeks, teenagers in a small UK randomised trial put away their smartphones and stopped using social media. Compared with peers continuing their usual habits, the abstinence groups reported longer sleep and some improvements in stress and mood. Yet the headline result comes with an important qualification: those improvements were not maintained at the two-month follow-up. The study therefore offers preliminary evidence that temporary disconnection can change some experiences in the short term, rather than proof that a brief digital detox produces lasting mental-health benefits.
Why the trial is different from a screen-time survey
Many investigations of teenagers’ digital habits ask participants how often they use their phones and compare those answers with their wellbeing. Such studies can be informative, but a correlation does not show which factor came first. Poor sleep might lead to more scrolling, scrolling might interfere with sleep, or another circumstance might influence both. In this pilot, researchers randomly assigned students to different conditions, making it possible to compare a deliberate change in access with business-as-usual use. Randomisation is a meaningful methodological advantage, although the small trial was designed to test feasibility rather than to provide a definitive estimate of health effects.
Three groups, including a basic-phone option
The researchers enrolled 82 students aged 13 to 18 from two schools in the United Kingdom. Twenty-six were assigned to abstain from smartphones and social media, another 26 to the same abstinence plan with access to a basic phone for calls and texts, and 30 to a control group that continued its usual smartphone and social media habits. This design addressed a practical concern: many teenagers rely on a phone to contact family members. The basic-phone group could retain essential communication without the features of a modern smartphone. Comparing the two abstinence approaches also offered a way to investigate whether preserving calls and texts reduced the burdens of disconnection.
What taking a break actually involved
Students in the abstinence groups had their smartphones stored in locked school offices for 21 days. They were instructed not to use social media on other devices, including computers, tablets and gaming consoles. Television viewing was permitted, while gaming itself was not prohibited. Those allocated to the basic-phone condition received a simple handset capable of calls and text messages. The control group continued as normal. This was a relatively demanding intervention, much stronger than simply silencing notifications at bedtime or limiting use of one social media application. It is important not to assume that a gentler restriction would produce exactly the same results.
Sleep duration improved in self-reports
At the end of the intervention, self-reported nightly sleep duration had increased by 49 minutes in the complete-abstinence group and 45 minutes in the basic-phone group. The corresponding increase among controls was 14 minutes. The differences in change between groups were statistically significant in the study’s exploratory models. These are changes from each group’s own starting point, not an assertion that all participants slept an identical amount. For a teenager who struggles to find sufficient sleep on school nights, an additional three-quarters of an hour could be meaningful. But the results are preliminary, and the two-month follow-up did not show the same advantage.
The wearable devices told a more cautious story
Participants also wore Garmin devices to estimate sleep duration and heart-rate variability. Unlike the self-reported sleep findings, the wearable-derived sleep measures did not show a clear change across groups or study weeks. Wearable data were available for smaller and varying subsets, making these estimates less precise. This discrepancy deserves attention because subjective and objective measures do not always capture exactly the same thing. A teenager may feel that sleep has improved even when a device does not register a substantial difference in duration. Neither measure is automatically worthless, but the disagreement prevents a simple claim that the detox objectively extended sleep for everyone.
Stress and depressive symptoms
Perceived stress scores fell by 6.63 points in the full-abstinence group and 6.44 points in the basic-phone group, compared with a decline of 2.29 points in controls. Depressive-symptom scores also declined, with a larger reduction in the complete-abstinence group than among controls. The study’s models identified some statistically significant differences in changes between groups. These outcomes suggest that disconnecting may temporarily ease aspects of psychological strain for some teenagers. However, other measures, including anxiety and negative mood, improved across multiple groups. The researchers cautioned that the study was not powered for firm causal conclusions about individual outcomes.
The follow-up changes the interpretation
The researchers assessed participants again approximately two months after the intervention. The short-term improvements attributed to abstinence were not maintained at that later assessment. This finding makes the distinction between an immediate response and a durable habit change central to the story. A 21-day restriction might temporarily alter bedtime routines or the social demands of constant connectivity, yet those benefits may diminish when normal patterns resume. The trial did not establish exactly why the gains faded. It is also possible that more gradual or supported changes would be easier to sustain, but that remains a hypothesis for future research rather than a result demonstrated here.
Working memory showed a limited signal
The study included computerised cognitive tasks, not just questionnaires. On a forward spatial working-memory task, accuracy rose by about 7% in the complete-abstinence group and 10% in the basic-phone group, compared with a decrease of roughly 1% among controls. A statistically significant interaction was identified for the basic-phone group on this measure. Other cognitive outcomes, including sustained-attention errors and a backward working-memory task, did not show the same pattern. It would therefore be misleading to say that giving up smartphones generally improves teenagers’ intelligence or concentration. The evidence points to a narrower, exploratory change in one task.
What teenagers said was difficult
The researchers also asked participants about their experiences. Among those responding to the post-intervention reflections, 39% described mixed feelings and 11% negative feelings, while half described the experience positively. Boredom was reported by 27%, practical reliance on smartphones by 48%, and fear of missing out or reduced communication by 36%. These responses show that an intervention can be scientifically interesting yet inconvenient in everyday life. Phones are used for navigation, timetables, friendships, family arrangements and access to information. Removing them entirely may impose costs that are easy to overlook when discussing digital wellbeing in abstract terms.
A basic phone did not solve every problem
Providing a basic handset was intended to preserve communication while reducing the temptation to scroll. Yet the qualitative findings suggest that this substitute did not eliminate the wider difficulties associated with giving up a smartphone. Modern devices serve as multi-purpose tools, and replacing calls and texts does not replace every useful function. The study raises a practical design question for schools and families: is a complete ban necessary, or could more targeted changes reduce problematic use while retaining essential features? The researchers suggested that restricting internet access or targeting bedtime habits may be worth testing, but did not establish which approach would work best.
Participation was encouraging, recruitment was limited
Of 82 enrolled students, 71 completed the intervention, representing attrition of approximately 13.4%. This retention suggests that a demanding 21-day school-based intervention is possible to deliver. However, recruitment was a much bigger challenge. Only 113 of 1,845 eligible students initially expressed interest, a rate of 6.1%, and the final sample was smaller still. Volunteers may have differed from teenagers who were unwilling or unable to participate. They may have been more motivated to reduce phone use or had family circumstances that made abstinence easier. These differences limit the generalisability of the findings.
Why this was a pilot rather than a definitive test
A pilot trial is designed primarily to determine whether researchers can recruit participants, deliver an intervention, retain people and collect usable data. The researchers explicitly stated that the study was not statistically powered to detect reliable differences across all the health and cognitive outcomes measured. Although some exploratory comparisons reached conventional significance thresholds, a small study testing multiple outcomes can generate unstable estimates. The appropriate conclusion is that the design appears feasible and certain results are promising enough to investigate further. A larger preregistered trial would be needed to determine whether the patterns are reproducible and clinically important.
The challenge of measuring adherence
Participants were asked to avoid social media on all devices, but compliance was assessed using questionnaires rather than continuous objective monitoring. Among respondents to a post-intervention question, 66% reported avoiding social media for the full 21 days. This suggests that not everyone followed the intervention perfectly. In real life, even a carefully organised school programme cannot guarantee that students never access an account through another device. Future trials could combine consent-based app-use records with transparent privacy safeguards to measure adherence more accurately. Such monitoring would need to be proportionate, particularly because the participants are minors.
What the study cannot tell us
The experiment does not show that smartphones cause depression, that every social media platform is harmful, or that removing devices is the best policy for all schools. Participants came from two UK schools, and their circumstances may not resemble those of teenagers elsewhere. The study did not restrict gaming, even though gaming may also influence sleep. Researchers were not blinded to group assignment, and many outcomes depended on self-reports. The trial also cannot isolate which part of the intervention mattered most: less social comparison, fewer late-night notifications, reduced online stimulation or changes in social routines. Those mechanisms require separate investigation.
What this means for parents and schools
The most useful practical message is not that families should confiscate every smartphone for three weeks. Rather, the results suggest that access patterns may be modifiable and that some young people experience short-term benefits when constant connectivity is interrupted. Families might discuss when a phone is genuinely necessary, which functions interfere with sleep, and how to make changes without cutting off important communication. Schools considering restrictions should examine students’ practical needs and avoid assuming that one policy will benefit everyone equally. The study offers questions to test and discuss, not a universal prescription.
Why the policy response is noteworthy
The study reported that 83% of respondents to a policy question supported some form of government action on young people’s smartphone or social media use. That is an interesting expression of concern from adolescents themselves. It is not a representative referendum on national legislation: the responses came from a small, self-selected trial population, and the question encompassed different forms of action, including education and regulation. Support for doing something should not be confused with agreement on what exactly should be done. Policymakers need evidence about effectiveness, unintended consequences and the rights of young people before treating a ban as the obvious solution.
The bigger research question
Digital-health debates often ask whether screen time is harmful as though every minute and every activity were equivalent. This study points toward a more informative approach: test specific changes, measure sleep and wellbeing in several ways, and examine whether any benefits last. A late-night messaging habit may differ from daytime educational use or an essential call home. Research that separates activities and times of day could produce more useful guidance than treating all smartphone use as a single exposure. Larger trials with better adherence measures and longer follow-up would help establish whether targeted interventions offer lasting improvements without the disruption of complete abstinence.
The bottom line
In this pilot randomised trial, teenagers who gave up smartphones and social media for 21 days reported more sleep and some short-term improvements in stress and mood than peers continuing usual use. The benefits were not maintained two months later, and objective sleep measurements did not show a matching clear improvement. The intervention was feasible for those who enrolled, but recruiting a broad range of teenagers was difficult and complete abstinence posed practical challenges. The findings justify further rigorous research, not sweeping claims that a three-week digital detox is a proven long-term mental-health treatment.
Related Research Today coverage
For further context, read our coverage of minimal-phone experiments and a year-long study of social media and happiness. These examined different interventions and populations.
Source Information
Study title: Innovations in Practice: A pilot randomised trial on smartphone and social media abstinence – effects on sleep quality and psychological wellbeing in adolescents
Authors: Emma C. Sullivan, Scott A. Cairney and Lisa-Marie Henderson
Journal: Child and Adolescent Mental Health
Publication date: 9 August 2026
Study design: Preregistered three-arm pilot randomised controlled trial with 21-day intervention and two-month follow-up
Sample or scope: 82 adolescents aged 13–18 recruited from two UK schools; 26 complete abstinence, 26 abstinence with basic phone, 30 controls
Key finding: Abstinence groups reported 45–49 minutes more nightly sleep after the intervention compared with a 14-minute increase in controls, but improvements were not sustained at two months
DOI: 10.1111/camh.70118
