Teenagers overestimate their own sleep by 28 minutes — and not evenly

Almost everything the public knows about how much a population sleeps comes from asking people. Surveys are cheap, scale nationally, and produce the round numbers that end up in headlines. A study published in Sleep Health on 6 November measured how far those answers drift from what a wrist accelerometer records — and found the drift is not random [s1].

The comparison

The study used the age-15 wave of the Future of Families and Child Wellbeing Study. Participants (N=634; mean age 15.4 years; 51% Black, 30% Hispanic, 18% White) first self-reported typical weekday and weekend bedtimes and waketimes, then wore a wrist actigraph for one week [s1].

Running the self-report first and the measurement second is the right order for this question: it avoids the possibility that wearing a device changes what people say about their own habits.

After adjustment, self-reported sleep duration overestimated actigraphy-assessed duration by 27.87 minutes (95% CI: 34.96, 20.78) [s1].

Agreement between the three duration categories — short (under 8 hours), recommended (8–10 hours), and long (over 10 hours) — was 73%, but the weighted kappa statistic was 0.12 (95% CI 0.07 to 0.18) [s1]. That combination is the finding most likely to be misread. Seventy-three per cent sounds like reasonable agreement; a kappa of 0.12 says that almost all of it is what chance alone would produce, because most adolescents fall into one dominant category regardless of method.

The mean discrepancy was larger on weekends (74.3 ± 144.7 minutes) than on weekdays (27.5 ± 92.7 minutes) [s1].

The gap is not evenly distributed

Within-group analyses found that self-reported duration overestimated actigraphy-assessed duration for Black adolescents and for those whose primary caregivers had a high school education or less [s1].

This is the part with consequences beyond methodology. If self-report inflates sleep more in some groups than others, then survey-based estimates will understate insufficient sleep specifically in the groups already carrying more of it — and the measured disparity will look smaller than the real one. The authors frame the conclusion in exactly those terms: reliance on self-report may misclassify sleep duration and underestimate insufficient sleep, particularly in marginalised groups [s1].

What the measured national picture looks like

Two weeks later the same journal published an actigraphy-based description of sleep across a nationally representative US sample, which gives a sense of what the objective data show when survey answers are set aside [s2].

That analysis used 13,656 US residents aged 3 to 80 in the National Health and Nutrition Examination Survey (2011–2014), each wearing an ActiGraph GT3X+ on the wrist for up to seven days, with sleep metrics derived using GGIR version 3.0.0 and analyses accounting for the complex sampling design [s2].

Sleep duration was shorter with greater age [s2]. Sleep onset, midpoint, and waketime were latest among those aged 10 to 30, and social jetlag followed a similar distribution [s2]. Wake after sleep onset was highest among children and lower with greater age [s2]. Females generally slept more than males [s2].

The socioeconomic and racial patterns match the direction of the measurement-bias finding. Adults with a low poverty-income ratio tended to have worse sleep than adults with a higher ratio [s2]. Compared with non-Hispanic White adults, Hispanic, non-Hispanic Asian, and non-Hispanic Black adults had shorter sleep duration and higher wake after sleep onset, and non-Hispanic Black adults had the highest social jetlag [s2].

Limits

Actigraphy is not polysomnography. It infers sleep from movement, tends to overestimate sleep in people who lie still while awake, and cannot stage sleep. Treating it as ground truth overstates what it does; it is a better measure than self-report for duration and timing, not a perfect one.

The measurement-bias study is a single cohort at a single age wave, with 634 participants [s1], and its racial and socioeconomic composition is not nationally representative. The NHANES actigraphy data are from 2011–2014 [s2], now more than a decade old.

Why it matters

Sleep guidance, school start time debates, and workplace policy all rest on population estimates of how much people sleep. If those estimates come from surveys, they are systematically optimistic by roughly half an hour in adolescents [s1], and more optimistic in some groups than others. That is a correction worth carrying into any figure quoted from a self-report survey.

This article is informational and does not constitute medical advice.

Sources

• [s1] Measurement bias in adolescent sleep duration: Comparison of self-reported and actigraphy-assessed sleep duration among adolescents in the Future of Families and Child-Wellbeing Study. Sleep Health, 6 November 2025. https://doi.org/10.1016/j.sleh.2025.10.003
• [s2] Monitoring sleep duration, timing, and continuity among US youth and adults in NHANES using actigraphy. Sleep Health, 20 November 2025. https://doi.org/10.1016/j.sleh.2025.10.008