Abstract
A substantial body of research examines time limits, abstinence, environmental modifications like grayscale, and clinical approaches for reducing social media use across a spectrum from mild dissatisfaction to severe compulsive use. Most controlled trials recruit mildly dissatisfied volunteers and run for only a few weeks, so findings are difficult to extend to severe or clinical cases and long-term durability remains almost entirely untested. The evidence identifies several plausible mechanisms and replicable short-term effects, but causality claims are tentative and effect sizes are often small or heterogeneous.
The verdict
The core question: Strategies and tools (including abstinence) are effective at reducing harmful social media use and improving well-being across varying levels of social media addiction
Contested, likely depends on context Score +0.04 26/100 confidence
⚡ The evidence splits into two camps rather than converging — this likely depends on context, not that we don't know.
Evidence weight →
+0.04
Hunt et al. 2018 'No More FOMO' study: limiting undergraduates to ten minutes per platform per day reduced loneliness and depressive symptoms over three weeks, with effects strongest among those who started most depressed Stance: +0.60 · Weight: 2.1 Other evidence · 1 primary source Low Click the bubble for sources
Facebook deactivation: modest well-being gain, persistent reduction Stance: +0.60 · Weight: 7.2 n≈580 · 1 primary source High Click the bubble for sources
Allcott 2020: deactivation reduced political polarisation Stance: +0.30 · Weight: 4.2 Randomized controlled trial (N not reported) · 1 primary source Medium Click the bubble for sources
Lambert 2022: one-week break improved well-being (demand bias caveat) Stance: +0.60 · Weight: 4.4 n≈154 · 1 primary source Medium Click the bubble for sources
Meta-analyses of digital detox interventions show inconsistent effects: some abstinence studies find null results or increased craving and boredom Stance: -0.60 · Weight: 8.0 n≈3,625 · 1 primary source High Click the bubble for sources
Vally and D'Souza 2019 UAE study found abstinence worsened well-being, plausibly because social media served genuine connection needs Stance: -0.60 · Weight: 2.1 Other evidence · 1 primary source Low Click the bubble for sources
Active social users lose more from abstinence than scrollers Stance: -0.30 · Weight: 5.0 n≈580 · 1 primary source High Click the bubble for sources
App timers: poor adherence, smaller effect than grayscale Stance: -0.60 · Weight: 4.3 n≈112 · 1 primary source Medium Click the bubble for sources
Orben and Przybylski: population screen-time correlations are tiny Stance: -0.30 · Weight: 2.0 n≈17,247 · 1 primary source Medium Click the bubble for sources
Short social-media abstinence shows no reliable effect on positive affect, negative affect, or life satisfaction, with substantial between-study heterogeneity. Stance: -0.60 · Weight: 8.1 n≈4,674 · 1 primary source High Click the bubble for sources
Facebook vs. Instagram deactivation: fundamentally different outcomes Stance: +1.00 · Weight: 4.2 Field Deployment (N not reported) · 1 primary source Medium Click the bubble for sources
Mobile internet block (n=467): dz=0.45 on well-being (volunteer sample) Stance: +0.60 · Weight: 7.1 n≈467 · 1 primary source High Click the bubble for sources
In the one sec mechanism decomposition, the ordering of contribution is: dismiss option > time delay > deliberation (≈ negligible). Stance: +0.30 · Weight: 2.2 n≈780 · 1 primary source Medium Click the bubble for sources
One sec: 37% fewer opening attempts (developer co-authorship caveat) Stance: +0.30 · Weight: 3.7 n≈780 · 1 primary source Medium Click the bubble for sources
Temporary incentives to cut social media use had persistent effects after removal, indicating social media is habit-forming and a short intervention can lower the habit stock. Stance: +0.60 · Weight: 4.2 Field Deployment (N not reported) · 1 primary source Medium Click the bubble for sources
Notification-disabling produces a specific paradox: it changes felt intentionality without changing behavior, because it only removes external cues while most checking is internally driven. Stance: -0.30 · Weight: 2.5 Quasi Experiment (N not reported) · 1 primary source Medium Click the bubble for sources
Dashboards and tracking-only tools produce null reduction Stance: -0.30 · Weight: 3.9 n≈121 · 1 primary source Medium Click the bubble for sources
Reduced use and improved well-being are separable claims Stance: -0.30 · Weight: 8.2 n≈5,544 · 1 primary source High Click the bubble for sources
Continuous passive environmental modification during use (grayscale mode) produces ~20 minutes/day reduction in session duration and is superior to active threshold-based app timers in immediate effect, while passive tracking-only (dashboards) produces null objective reduction; grayscale reduces engagement depth but not unlock frequency (session initiation). Stance: +0.60 · Weight: 3.8 n≈84 · 1 primary source Medium Click the bubble for sources
Real-time session-contingent interruption (notifications triggered when Instagram session exceeds threshold) shows null effect on usage reduction and mixed user receptivity (only 6 of 21 participants found helpful; 2 found annoying), and is mechanistically distinct from continuous passive numeric ambient display. Stance: -0.60 · Weight: 3.8 n≈21 · 1 primary source Medium Click the bubble for sources
Passive tracking-only control (dashboards, weekly digests, or real-time passive monitoring displays without environmental modification or friction) produces null objective usage reduction in controlled tests, consistent with hypothesis that such tools must bridge an intention-behavior gap and require user self-regulation at future execution moments. Stance: -0.60 · Weight: 5.5 n≈233 · 2 primary sources Medium Click the bubble for sources
Adjacent-domain literature (industrial safety, energy conservation, clinical alert fatigue) predicts that static continuous non-contingent ambient displays lose effectiveness over weeks to months through perceptual habituation; this attenuation is slower (weeks-to-months) for behavioral habituation alone and affects abstract metrics (screen time, kilowatt-hours) more than sensory-coupled feedback (water flow temperature/pressure). Stance: -0.30 · Weight: 1.8 Expert opinion & Other evidence · 3 primary sources Low Click the bubble for sources
Grayscale reduces engagement depth (time per session) but not initiation frequency (unlock count), suggesting the operative mechanism is reward-signal reduction (muting dopaminergic salience of colorful badges/content) rather than reducing the compulsive checking initiation pattern driven by boredom/habit. Stance: +0.30 · Weight: 2.5 n≈84 · 1 primary source Medium Click the bubble for sources
Whittaker meTime: elapsed-time display reduced social and email use Stance: +0.60 · Weight: 4.2 Field Deployment (N not reported) · 1 primary source Medium Click the bubble for sources
A preregistered RCT of complete category-level mobile internet blocking (Freedom app, n=467, 2 weeks), representing the strongest form of substitution-channel closure, found intent-to-treat pre-post intervention effects of dz=0.45, demonstrating that full channel closure produces meaningful improvements in mental health, well-being, and sustained attention outcomes. Stance: +0.60 · Weight: 4.7 n≈467 · 1 primary source Medium Click the bubble for sources
NO — refuted 0 YES — supported
Stance on the premise →
Hunt et al. 2018 'No More FOMO' study: limiting undergraduates to ten minutes per platform per day reduced loneliness and depressive symptoms over three weeks, with effects strongest among those who started most depressed
Stance +0.60 Weight 2.1 Low
Facebook deactivation: modest well-being gain, persistent reduction
Stance +0.60 Weight 7.2 High n≈580
Allcott 2020: deactivation reduced political polarisation
Stance +0.30 Weight 4.2 Medium
Lambert 2022: one-week break improved well-being (demand bias caveat)
Stance +0.60 Weight 4.4 Medium n≈154
Meta-analyses of digital detox interventions show inconsistent effects: some abstinence studies find null results or increased craving and boredom
Stance -0.60 Weight 8.0 High n≈3,625
Vally and D'Souza 2019 UAE study found abstinence worsened well-being, plausibly because social media served genuine connection needs
Stance -0.60 Weight 2.1 Low
No linked source citation available for this finding.
Active social users lose more from abstinence than scrollers
Stance -0.30 Weight 5.0 High n≈580
App timers: poor adherence, smaller effect than grayscale
Stance -0.60 Weight 4.3 Medium n≈112
Orben and Przybylski: population screen-time correlations are tiny
Stance -0.30 Weight 2.0 Medium n≈17,247
Short social-media abstinence shows no reliable effect on positive affect, negative affect, or life satisfaction, with substantial between-study heterogeneity.
Stance -0.60 Weight 8.1 High n≈4,674
Facebook vs. Instagram deactivation: fundamentally different outcomes
Stance +1.00 Weight 4.2 Medium
Mobile internet block (n=467): dz=0.45 on well-being (volunteer sample)
Stance +0.60 Weight 7.1 High n≈467
In the one sec mechanism decomposition, the ordering of contribution is: dismiss option > time delay > deliberation (≈ negligible).
Stance +0.30 Weight 2.2 Medium n≈780
One sec: 37% fewer opening attempts (developer co-authorship caveat)
Stance +0.30 Weight 3.7 Medium n≈780
Temporary incentives to cut social media use had persistent effects after removal, indicating social media is habit-forming and a short intervention can lower the habit stock.
Stance +0.60 Weight 4.2 Medium
Notification-disabling produces a specific paradox: it changes felt intentionality without changing behavior, because it only removes external cues while most checking is internally driven.
Stance -0.30 Weight 2.5 Medium
No linked source citation available for this finding.
Dashboards and tracking-only tools produce null reduction
Stance -0.30 Weight 3.9 Medium n≈121
Reduced use and improved well-being are separable claims
Stance -0.30 Weight 8.2 High n≈5,544
Continuous passive environmental modification during use (grayscale mode) produces ~20 minutes/day reduction in session duration and is superior to active threshold-based app timers in immediate effect, while passive tracking-only (dashboards) produces null objective reduction; grayscale reduces engagement depth but not unlock frequency (session initiation).
Stance +0.60 Weight 3.8 Medium n≈84
Real-time session-contingent interruption (notifications triggered when Instagram session exceeds threshold) shows null effect on usage reduction and mixed user receptivity (only 6 of 21 participants found helpful; 2 found annoying), and is mechanistically distinct from continuous passive numeric ambient display.
Stance -0.60 Weight 3.8 Medium n≈21
No linked source citation available for this finding.
Passive tracking-only control (dashboards, weekly digests, or real-time passive monitoring displays without environmental modification or friction) produces null objective usage reduction in controlled tests, consistent with hypothesis that such tools must bridge an intention-behavior gap and require user self-regulation at future execution moments.
Stance -0.60 Weight 5.5 Medium n≈233
Adjacent-domain literature (industrial safety, energy conservation, clinical alert fatigue) predicts that static continuous non-contingent ambient displays lose effectiveness over weeks to months through perceptual habituation; this attenuation is slower (weeks-to-months) for behavioral habituation alone and affects abstract metrics (screen time, kilowatt-hours) more than sensory-coupled feedback (water flow temperature/pressure).
Stance -0.30 Weight 1.8 Low
No linked source citation available for this finding.
Grayscale reduces engagement depth (time per session) but not initiation frequency (unlock count), suggesting the operative mechanism is reward-signal reduction (muting dopaminergic salience of colorful badges/content) rather than reducing the compulsive checking initiation pattern driven by boredom/habit.
Stance +0.30 Weight 2.5 Medium n≈84
Whittaker meTime: elapsed-time display reduced social and email use
Stance +0.60 Weight 4.2 Medium
A preregistered RCT of complete category-level mobile internet blocking (Freedom app, n=467, 2 weeks), representing the strongest form of substitution-channel closure, found intent-to-treat pre-post intervention effects of dz=0.45, demonstrating that full channel closure produces meaningful improvements in mental health, well-being, and sustained attention outcomes.
Stance +0.60 Weight 4.7 Medium n≈467
High ≥3 consistent independent studies, or one strong-design study (meta-analysis, systematic review, RCT) with no conflicting results and no funding concerns.
Medium A moderate-design study (cohort, case-control), or fewer than 3 independent studies, or a strong-design study downgraded by a conflict of interest or a single funder.
Low No independent primary source found in the evidence bank, only weak-design evidence (cross-sectional, case report, preprint, expert opinion, community anecdote, news coverage), conflicting effect directions between studies, or a material conflict of interest.
Surprising findings
What people assume Passive awareness tools like screen-time dashboards help people use their phones less once they see how much time they spend.
What the evidence shows Tracking-only controls, which function like commercial dashboards, produce null objective usage reduction in controlled trials.
A three-arm RCT by Zimmermann and Sobolev compared tracking-only, grayscale, and active timers. The tracking-only condition (resembling Google Digital Wellbeing) produced no significant reduction in device-logged screen time, while grayscale produced an immediate significant reduction. Information about use does not automatically translate into behaviour change without an environmental modification or decision-point intervention at the moment of use.
What people assume The more severely addicted someone is, the less likely a simple reduction strategy will help them.
What the evidence shows The most distressed users sometimes gain the most from simple time-limit interventions.
Hunt et al. 2018 found that well-being effects of limiting use to ten minutes per platform were strongest among participants who started most depressed. Digital health intervention research similarly finds that severely burdened individuals need not be excluded from internet-based stress management. This suggests that the gradient of need does not always predict poorer response to accessible behavioural tools, at least in the short term.
What people assume Quitting social media entirely for a week or two reliably improves your mood and reduces anxiety.
What the evidence shows A meta-analysis of 10 randomised trials found no significant pooled effect of abstinence on positive affect, negative affect, or life satisfaction.
While individual studies by Tromholt, Allcott, and Lambert reported positive effects, pooling 4,674 participants across 10 RCTs yields null results on affective well-being and life satisfaction. The positive individual findings are likely inflated by demand characteristics: 91% of participants suspected researchers were trying to show social media is harmful, biasing self-reported outcomes. The overall picture is profound heterogeneity, not consistent benefit.
What people assume Simply turning off notifications is an effective way to reduce how much time you spend on social media.
What the evidence shows Disabling notifications increases the felt sense of intentionality but produces no measurable objective reduction in usage.
Roughly 90% of social media interactions are user-initiated rather than notification-triggered, so removing external prompts cannot reach the internally cued checking driven by boredom or habit. Controlled data show that notification disabling changes subjective perceptions of control without altering device-logged behaviour, making it a poor intervention for actual usage reduction.
What people assume If you deactivate Instagram and Facebook, the effects should be similar since both are major social platforms.
What the evidence shows Facebook deactivation persistently reduces total app use by roughly 9 minutes per day while Instagram deactivation results in full substitution with no net reduction.
An NBER study found that most freed time from Facebook deactivation is substituted to other apps yet total usage still falls, whereas all freed time from Instagram deactivation is simply redirected to other apps with no net reduction. The operative mechanism appears to be platform-specific habit-stock disruption tied to Facebook's unique social-graph integration rather than simple channel closure, meaning lessons from Facebook deactivation studies do not transfer to Instagram.
Notes on interpreting findings in this space
Beware of the following when reading this research
Nearly all controlled trials recruit self-selected, mildly dissatisfied volunteers, so findings may not generalise to severe or compulsive users who represent the most clinically important group.
Most studies run only 2 to 6 weeks on compensated samples, making it impossible to know whether effects persist, decay, or reverse over months.
Self-reported screen time correlates poorly with device logs (r around 0.38), and the heaviest and most distressed users are the least accurate reporters, undermining studies that rely on self-report outcomes.
Demand characteristics are substantial: 91% of participants in abstinence RCTs suspected researchers were trying to show social media is harmful, inflating self-reported well-being benefits in the direction of hypothesised effects.
Social media addiction is not a recognised clinical diagnosis (unlike gaming disorder in ICD-11), and the scales used to define severe cases show weaker validity against objective behaviour than even basic duration measures.
Key findings — confidence & importance
High confidence + high importance
High confidence + medium importance
Medium confidence + high importance
Medium confidence + medium importance
Low / contested confidence
Observation about the evidence base
Well-being and mental health outcomes from abstinence
Facebook deactivation: modest well-being gain, persistent reduction
Mobile internet block (n=467): dz=0.45 on well-being (volunteer sample)
Volunteer compensated sample; effect-size comparisons to antidepressants are rhetorically strong but methodologically loose.
Lambert 2022: one-week break improved well-being (demand bias caveat)
Facebook vs. Instagram deactivation: fundamentally different outcomes
Heterogeneous effects and population-level correlations
Active social users lose more from abstinence than scrollers
Reduced use and improved well-being are separable claims
Orben and Przybylski: population screen-time correlations are tiny
Political polarization and evidence gaps in severe cases
Allcott 2020: deactivation reduced political polarisation
CBT-IA for severe cases: assumed, not demonstrated for social media
Friction-based and awareness-based intervention techniques
App timers: poor adherence, smaller effect than grayscale
One sec: 37% fewer opening attempts (developer co-authorship caveat)
The strongest one sec evidence is co-authored with the app developer. Replication by unaffiliated teams is thin.
Dashboards and tracking-only tools produce null reduction
Whittaker meTime: elapsed-time display reduced social and email use
Grayscale cuts session time ~20 min/day (4 replications)
What people think — researchers, practitioners & communities
Sceptical of mainstream narrative
Cautionary / warning of harm
Nuanced / conditional
Methodological concern
"One user reported that after a year of abstinence, offline information now came filtered through people with shared values, indicating a genuine net gain in connection quality for users with dense offline social networks."
Community user report
"A systematic review found no significant effects of social media abstinence interventions on positive affect, negative affect, or life satisfaction across ten studies and 4,674 participants."
Lemahieu and Zwalmen et al. (2025)
"Uninstall-reinstall cycles cluster around mood dips, loneliness, and boredom; reinstallation risk is highest at moments of lowest mood, not uniformly distributed over time."
Community forum users
"Problematic social media use rarely travels alone; clinicians treat use and underlying mood disorder simultaneously rather than hierarchically, with depression making the feed tempting and the feed deepening the depression."
Expert practitioner consensus
"Reporting by social media platforms on friction interventions deployed to date lacks transparency related to the testing and implementation process, making it difficult for researchers to study different countermeasures."
Jahn (2023)
"Self-reported media use correlates only moderately with logged measurements, and measures of problematic media use show an even weaker association with usage logs."
Parry et al. (2021)
"The most fundamental limitation in the literature is a lack of long-term follow-up data; no intervention was longer than three weeks and only three studies measured depression later than post-intervention."
Burnell (2025)
What remains unknown
Hard to study
Do the well-being gains observed in individual abstinence RCTs like Lambert et al. 2022 persist after correcting for demand characteristics statistically or through pre-registered designs explicitly addressing researcher-expectancy effects?
Untested
What is the post-intervention persistence timeline for abstinence and reduction effects beyond the few available 2-4 week follow-ups? Do effects decay, stabilize, or grow over 3-6 months?
Untested
For severe problematic social media use specifically, is an addiction-framed (CBT-IA style) treatment approach or a comorbidity-integrated approach more effective, and should they be used sequentially or simultaneously?
Untested
What is the persistence of grayscale's approximately 20 min/day effect beyond 3 weeks? Does the effect attenuate through perceptual habituation as adjacent-domain literature predicts, and if so, at what timescale?
Hard to study
For within-person longitudinal analyses using EMA as the exposure measure, does the validity deficit (r approximately 0.30 for daily fluctuation) mean most published ESM/EMA studies of social media and daily mood outcomes are operating on near-random measurement noise?
Most likely explanation
Passive environmental modification (grayscale) and platform deactivation produce the most reliable short-term reductions, but no strategy shows durable benefit across all users and all outcomes.
Low-moderate confidenceThe most consistent finding is that altering the environment continuously and passively during use (grayscale, platform deactivation) outperforms strategies that rely on users exercising willpower at the moment of use (timers, dashboards). Grayscale is replicated across four independent studies at roughly 20 minutes per day reduction in session duration, while passive tracking tools show null effects in direct comparison. For mild dissatisfaction, time limits are comparably effective to abstinence without its social costs; for severe compulsive use, no controlled trial has demonstrated a clearly superior approach and clinical best practice favours treating underlying mood disorders simultaneously rather than targeting social media use in isolation.
Main caveats: All replication chains run only 2-6 weeks on volunteer samples; adherence decay, long-term durability, and generalisation to severe clinical users remain essentially untested, meaning current effect sizes are likely upper bounds on real-world population impact.
Best practical tips from the research
Use grayscale mode to reduce session length
Switching your phone display to grayscale has been replicated across four controlled studies and reduces daily screen time by roughly 20 minutes by muting the visual reward signals that sustain scrolling, though it will not reduce how often you pick up your phone.
Strongest evidence Set per-platform time limits rather than quitting cold turkey
For people who are mildly unhappy with their usage, a ten-minute-per-platform daily cap produces small but meaningful reductions in loneliness and depressive symptoms without requiring full abstinence, which shows no reliable pooled benefit.
Moderate evidence Do not rely on your phone's built-in screen-time dashboard
Passive tracking tools such as Apple Screen Time or Google Digital Wellbeing produce no measurable objective reduction in use in controlled trials; actual environmental modification or friction is required for behavioural change.
Strongest evidence Assess social function before advising blanket reduction
For people whose social media use serves genuine connective needs (LGBTQ+ youth without local community, isolated older adults, people with stigmatising conditions), reducing use without offering alternatives may increase loneliness rather than improve well-being.
Moderate evidence Treat severe cases alongside underlying mood disorders
Clinicians broadly favour treating compulsive social media use alongside comorbid depression or anxiety simultaneously rather than treating either as strictly primary, because the relationship between mood and use is transactional.
Expert consensus Plan for reinstallation during emotional low points
Community evidence and the intervention literature both show that app deletions fail most often during mood dips, loneliness, and boredom; pre-commitment strategies or environmental scaffolding are more robust than relying on willpower at the moment of craving.
Caveat - real-world pattern About the author
Research methodology: AI analysis and synthesis across more sources than a traditional manual review allows, with human editorial direction and review. Intended for directional understanding rather than a formal meta-analysis — read primary sources before making important decisions based on these findings.
Core sources
2026 Digital Behavioral Addictions in Adolescents and Psychotherapeutic Interventions: A Systematic Review Adolescent Research Review · Özge Canoğulları
2025 Blocking mobile internet on smartphones improves sustained attention, mental health, and subjective well-being PNAS Nexus · Castelo, Kushlev, Ward, Esterman, and Reiner
2025 Examining measurement discrepancies in adolescent screen media activity with insights from the ABCD study npj Mental Health Research · Yihong Zhao
2025 The Effect of Deactivating Facebook and Instagram on Users' Emotional State (NBER Working Paper) NBER · Hunt Allcott, Matthew Gentzkow
2025 The effects of social media abstinence on affective well-being and life satisfaction: a systematic review and meta-analysis Scientific Reports · Lemahieu, Zwalmen et al.
2025 The effects of social media restriction: Meta-analytic evidence from randomized controlled trials ScienceDirect · Kaitlyn Burnell
2024 Agreement between self-reported and objectively measured smartphone use among adolescents and adults ScienceDirect (Computers in Human Behavior Reports) · Kimberly M. Molaib
2024 Are active and passive social media use related to mental health, wellbeing, and social support outcomes? A meta-analysis of 141 studies Journal of Computer-Mediated Communication, Oxford Academic · Statistical detail not reported
2024 Is life brighter when your phone is not? The efficacy of a grayscale smartphone intervention addressing digital well-being SAGE Journal of Digital Health · Dekker, Baumgartner
2023 Directing smartphone use through the self-nudge app one sec PNAS · Statistical detail not reported
2022 A Nudge-Based Intervention to Reduce Problematic Smartphone Use: Randomised Controlled Trial International Journal of Mental Health and Addiction, Springer · Jay A. Olson
2022 Digital Addiction American Economic Review · Hunt Allcott
2022 Digital detox: An effective solution in the smartphone era? A systematic literature review Mobile Media & Communication · Theda Radtke
2022 The associations of active and passive social media use with well-being: A critical scoping review New Media & Society / SAGE · Patti M. Valkenburg, Irene I. van Driel, Ine Beyens
2021 A systematic review and meta-analysis of discrepancies between logged and self-reported digital media use Nature Human Behaviour · Douglas A. Parry
2021 The accuracy and validity of self-reported social media use measures among adolescents ScienceDirect · Verbeij, Pouwels, Beyens, Valkenburg
2020 Digital Nudges for Screen Time Reduction: A Randomized Control Trial with Performance and Wellbeing Outcomes OSF PsyArXiv · Zimmermann, Sobolev
2020 The Welfare Effects of Social Media American Economic Review · Hunt Allcott, Luca Braghieri, Sarah Eichmeyer, Matthew Gentzkow
2019 Screens, Teens, and Psychological Well-Being: Evidence From Three Time-Use-Diary Studies Psychological Science (SAGE) · Amy Orben, Andrew K. Przybylski
2018 No More FOMO: Limiting Social Media Decreases Loneliness and Depression Journal of Social and Clinical Psychology · Melissa G. Hunt
2017 Do Social Network Sites Enhance or Undermine Subjective Well-Being? A Critical Review Social Issues and Policy Review · Philippe Verduyn
2016 'Don't Waste My Time': Use of Time Information Improves Focus (meTime ambient display study) ACM CHI · Whittaker, Kalnikaité, Hollis, Guydish
2016 The relationship between addictive use of social media and video games and symptoms of psychiatric disorders: A large-scale cross-sectional study PubMed / Journal of Behavioral Addictions · Cecilie Schou Andreassen, Billieux et al.
2016 Validation of the DSM-5 internet gaming disorder framework for clinical diagnosis of problematic social media usage PMC / PubMed Central · Regina J. J. M. van den Eijnden
Evidence landscape
Systematic review (13) High confidence evidence
Meta-analysis (8) High confidence evidence
Randomized controlled trial (5) High confidence evidence
Randomized controlled trial (17) Moderate confidence evidence
Cohort study (9) Moderate confidence evidence
Quasi Experiment (7) Moderate confidence evidence
Field Deployment (6) Moderate confidence evidence
Expert opinion (29) Low-moderate confidence evidence
Other evidence (22) Low-moderate confidence evidence
Cross-sectional study (11) Low-moderate confidence evidence
Qualitative (3) Low-moderate confidence evidence
News coverage (2) Low-moderate confidence evidence
Cohort study (1) Low-moderate confidence evidence
Community reports (3) Low confidence evidence
LowLow-moderateModerateHigh
Evidence quality / confidence →
136 sources across the full evidence base.
Low confidence Individual case reports, personal anecdotes, testimonials, personal quotes, social media posts.
Low-moderate confidence Case-control studies, cross-sectional studies, small or poorly controlled studies, mechanistic or laboratory evidence extrapolated to real-world outcomes, individual expert opinion.
Moderate confidence Individual randomized controlled trials, prospective cohort studies, large observational studies, natural or quasi-experimental studies, systematic reviews with substantial heterogeneity, expert consensus.
High confidence High-quality systematic reviews and meta-analyses; well-designed, adequately powered randomized controlled trials; strong evidence syntheses or guidelines built on systematic evidence.
Try Unscroll