Every number, and where it came from

Parry, D. A., Davidson, B. I., Sewall, C. J. R., Fisher, J. T., Mieczkowski, H., & Quintana, D. S. (2021). A systematic review and meta-analysis of discrepancies between logged and self-reported digital media use. Nature Human Behaviour, 5(11), 1535–1547.

SupportsSelf-reports correlate only moderately with device logs, r = 0.38 [0.33, 0.42], 106 effect sizes. Only 3 of 49 usable comparisons (6.1%) fell within 5% of logged use; 47% over-reported and 47% under-reported.

LimitationsPooled across ages, devices and media types; no Gen Z stratum.

Alexander, J. D., et al. (2024). Passively sensing smartphone use in teens with rates of use by sex and across operating systems. Scientific Reports, 14(1), 17982.

Supports185 min/day self-reported against 298 min/day passively logged, 1,415 adolescents, mean age 14, three weeks.

LimitationsOne cohort (ABCD Study). A single sample does not establish a general two-hour gap.

Pew Research Center. (2025, April 22). Teens, social media and mental health.

Supports45% of US teens say they spend too much time on social media (36% in 2022); 44% have cut back on social media and the same share on smartphone use.

Limitationsn = 1,391 US teens 13–17, fielded 18 Sept – 10 Oct 2024, MoE ±3.3.

Pew Research Center. (2026, April 15). Teens' experiences on TikTok, Instagram and Snapchat.

SupportsAbout 30% of teen TikTok users say they spend too much time on it; roughly six in ten say their time on each platform is about right.

LimitationsSurvey self-report. Most of this audience does not think it has a problem.

Asselin, G., Bilodeau, H., & Khalid, A. (2024, January 16). Digital well-being: The relationship between technology use, mental health and interpersonal relationships. Statistics Canada (Catalogue no. 22-20-0001).

Supports42% of Canadians aged 15–24 spend 20 or more hours a week online; 25% of Canadians overall.

LimitationsUnderlying data is the 2022 Canadian Internet Use Survey; measures general internet use, not social media.

Grüning, D. J., Riedel, F., & Lorenz-Spreen, P. (2023). Directing smartphone use through the self-nudge app one sec. PNAS, 120(8), e2213114120.

Supports57% fewer target-app openings after six weeks; self-reported 77 fewer minutes a day. In the controlled experiment (N = 500) the option to dismiss had the strongest effect, the delay also reduced consumption but added nothing on top of it, and the deliberation message was ineffective.

Limitations280 self-selected users, no control group in the field study. The app's developer is the second author.

Harkin, B., et al. (2016). Does monitoring goal progress promote goal attainment? A meta-analysis of the experimental evidence. Psychological Bulletin, 142(2), 198–229.

Supportsd+ = 0.40 [0.32, 0.48] for goal attainment across 138 randomised studies, N = 19,951; larger effects when progress was physically recorded or made public.

LimitationsJustifies daily check-ins. It does not justify penalising a missed day.

Zimmermann, L., & Sobolev, M. (2023). Digital strategies for screen time reduction: A randomized field experiment. Cyberpsychology, Behavior, and Social Networking, 26(1), 42–49.

SupportsDesign friction produced an immediate significant reduction; goal-setting a smaller gradual one; the self-monitoring control group did not reduce screen time.

LimitationsN = 112. One trial's control condition — it does not license 'all awareness interventions fail.'

Biedermann, D., Schneider, J., & Drachsler, H. (2021). Digital self-control interventions for distracting media multitasking – A systematic review. Journal of Computer Assisted Learning, 37(5), 1217–1231.

SupportsAcross 28 interventions, "especially interventions that relied purely on increasing the participants' awareness were barely effective."

LimitationsAuthors rate overall confidence low — small samples, short durations.

Thrul, J., Devkota, J., AlJuboori, D., Regan, T., Alomairah, S., & Vidal, C. (2025). Social media reduction or abstinence interventions are providing mental health benefits — Reanalysis of a published meta-analysis. Psychology of Popular Media, 14(2), 207–209.

SupportsInterventions under one week, Cohen's d = −0.175 (worse); one week or longer, d = +0.156 (better). Authors suggest around three weeks may be ideal and say explicitly that this needs confirmation.

LimitationsEffect sizes are Cohen's d, not Hedges' g. The pooled effect across all 20 studies was d = 0.081, non-significant. The under-one-week stratum contains only 4 studies. This is a three-page reanalysis of someone else's dataset.

Lemahieu, L., Vander Zwalmen, Y., Mennes, M., Koster, E. H. W., Vanden Abeele, M., & Poels, K. (2025). The effects of social media abstinence on affective well-being and life satisfaction: A systematic review and meta-analysis. Scientific Reports, 15(1).

SupportsPre-registered; null on positive affect, negative affect and life satisfaction. Break duration was not a significant moderator.

Limitations10 studies, N = 4,674, 38 effect sizes. This disagrees with Thrul et al. (2025) on whether length matters. Both are listed; they disagree.

Burnell, K., Meter, D. J., Andrade, F. C., Slocum, A. N., & George, M. J. (2025). The effects of social media restriction: Meta-analytic evidence from randomized controlled trials. SSM – Mental Health, 7, 100459.

SupportsRestricting social media improved subjective wellbeing, mean Hedges' ḡ = 0.17, 95% CI [0.08, 0.27], across 32 articles, 91 effect sizes, 5,544 individuals, mean age 23.38.

LimitationsThe authors call the estimates "small in magnitude, suggesting only weak support for the effectiveness of restricting social media use."

Allcott, H., Braghieri, L., Eichmeyer, S., & Gentzkow, M. (2020). The welfare effects of social media. American Economic Review, 110(3), 629–676.

SupportsN = 1,661 paid to deactivate Facebook for four weeks. Wellbeing index +0.09 SD; still 12 min/day lower a month after the experiment ended.

LimitationsAlso news knowledge −0.19 SD. The largest and best-identified trial here found a smaller effect than the meta-analytic average.

Pieh, C., Dale, R., Jesser, A., Probst, T., Plener, P. L., & Humer, E. (2025). Smartphone screen time reduction improves mental health: A randomized controlled trial. BMC Medicine, 23(1), 107.

SupportsN = 111 students, mean age 22.7, three-week reduction to ≤2 h/day. Partial η² = .109 depression, .085 stress, .053 wellbeing, .048 sleep quality.

Limitations"Screen time increased rapidly after the intervention and at follow-up the values were once again approaching the initial level."

Hunt, M. G., Marx, R., Lipson, C., & Young, J. (2018). No more FOMO: Limiting social media decreases loneliness and depression. Journal of Social and Clinical Psychology, 37(10), 751–768.

Supports143 undergraduates, three weeks at 10 min per platform per day. Loneliness and depression improved relative to control.

LimitationsAnxiety and FOMO improved in both arms. Single-site convenience sample, unblinded. The repeated "30 minutes a day" is 10 × 3 platforms, not an empirical threshold.

Lally, P., van Jaarsveld, C. H. M., Potts, H. W. W., & Wardle, J. (2010). How are habits formed: Modelling habit formation in the real world. European Journal of Social Psychology, 40(6), 998–1009.

SupportsRange 18–254 days to 95% of the automaticity asymptote. Missing a single day did not materially affect habit formation — the basis for our grace day.

Limitations96 recruited, 82 usable, model fitted for 62, good fit for only 39. Measures building a habit, not breaking one.

Singh, B., Murphy, A., Maher, C., & Smith, A. E. (2024). Time to form a habit: A systematic review and meta-analysis of health behaviour habit formation and its determinants. Healthcare, 12(23), 2488.

SupportsMedians 59–66 days, means 106–154 days, individual range 4–335 days across 20 studies, N = 2,601. Self-selected habits showed greater strength.

Limitations11 of 20 studies at high risk of bias. Habit formation, not habit breaking.

Buyalskaya, A., Ho, H., Milkman, K. L., Li, X., Duckworth, A. L., & Camerer, C. (2023). What can machine learning teach us about habit formation? Evidence from exercise and hygiene. PNAS, 120(17), e2216115120.

Supports"Contrary to the popular belief in a 'magic number' of days," gym habits take months and handwashing weeks. Complexity drives the timeline, not a fixed count.

LimitationsObservational, two specific behaviours.

Reed, P., Fowkes, T., & Khela, M. (2023). Reduction in social media usage produces improvements in physical health and wellbeing: An RCT. Journal of Technology in Behavioral Science.

SupportsThe reduce-only arm cut 37 min/day. The reduce-plus-assigned-activity arm increased use by 25 min/day, with 53% getting worse; the authors attribute it to resistance to being instructed.

Limitationsn = 17 per arm. One study — but the only head-to-head test.

Six, S. G., Byrne, K. A., Tibbett, T. P., & Pericot-Valverde, I. (2021). Examining the effectiveness of gamification in mental health apps for depression: Systematic review and meta-analysis. JMIR Mental Health, 8(11), e32199.

SupportsAcross 38 studies, n = 8,110, the number of gamification elements predicted neither outcomes (β = −0.03, p = .38) nor adherence (β = −1.93, p = .40).

LimitationsNo study isolates streaks as a mechanism. Our streak is a design choice.

Dekker, C. A., Baumgartner, S. E., Sumter, S. R., & Ohme, J. (2025). Beyond the buzz: Investigating the effects of a notification-disabling intervention on smartphone behavior and digital well-being. Media Psychology, 28(1), 162–188.

SupportsPre-registered RCT, N = 205, one week, objectively logged. No effect on checking frequency or screen time; increased fear of missing out.

LimitationsOne week. This is why the app does not tell you to turn off notifications.

Laird, B., Hook, J. N., & Van Tongeren, D. R. (2025). Spirituality in Clinical Practice, 12(4), 502–513.

SupportsRandomised, N = 192, three arms. No significant between-condition differences.

LimitationsThis is the causal evidence on faith-based wellness apps, and it is null. The earlier uncontrolled trial (Laird et al., 2024, n = 77) performed no subgroup comparison by race or faith.

Lai, N. M., et al. (2025). The effectiveness of school-based interventions for reducing screen time — A systematic review and meta-analysis. Child and Adolescent Mental Health, 30(3), 223–237.

SupportsScreen time: 27 studies, n = 19,751, SMD −0.10 [−0.14, −0.06], I² = 85%. Physical activity: 21 studies, n = 14,944, SMD +0.10 [0.02, 0.19].

LimitationsGRADE low certainty. The screen-time effect rests on the sub-analysis sample, not the 39-trial headline.

Contardo Ayala, A. M., et al. (2024). Effectiveness of interventions on sedentary behaviours and physical activity in secondary schools. Sports Medicine – Open.

SupportsNull pooled effects for sedentary time and physical activity at all intensities in secondary schools specifically.

LimitationsSecondary schools only.

Bourke, M., Maddren, C. I., Sippel, F., & Thomas, G. (2026). Within-person association between daily screen use and sleep in youth: A systematic review and meta-analysis. JAMA Pediatrics, 180(5), 500–509.

Supports25 studies, 4,562 participants, ages 3–25. Later sleep onset r = 0.079 [0.010, 0.149]. No significant within-person association with sleep duration, latency, efficiency or subjective quality. Use after bedtime was the significant moderator of subjective quality (r = −0.092, p = .007).

LimitationsThe defensible claim is about use after bedtime, not screen time in general.

Maguire, J., Persson, E., & Tinghög, G. (2025). Opportunity cost neglect: A meta-analysis. Journal of the Economic Science Association.

Supportsd = 0.22 [0.15, 0.27] across 39 experiments, N = 14,005; falls to d = 0.16 excluding the original studies; publication bias detected.

Limitations21 of 39 experiments were consumer choice, and there were insufficient studies in the time-use domain for robust conclusions. Our results page is an extrapolation and does not claim otherwise.

US Department of State, Foreign Service Institute. Foreign language training (archived page, 2017–2021).

Supports"Category I Languages: 24-30 weeks (600-750 class hours)" to reach Speaking-3 / Reading-3 — FSI's "Professional Working Proficiency."

LimitationsThe current State Department page gives 552–690 class hours against a slightly different criterion. Those are classroom hours with a teacher; on your own it takes longer.