Does More Sleep Actually Improve College Grades?

Last updated August 14, 2026

Start a student at a 3.6 GPA. Five nights a week of bad sleep — the ordinary state of a college semester — and that becomes a 3.5. Seven nights, and it’s a 3.46. That comes from an analysis of 55,322 undergraduates, where each additional night of poor sleep in a typical week was associated with about 0.02 lower GPA. The 3.6 starting point is ours, chosen to make the size of it concrete; the per-night figure is theirs.

That is a correlation, though, and for years it was all anyone had. What changed this year is that a randomized experiment tested the other direction: improve sleep on purpose, and see whether grades move.

They did. Researchers randomly assigned 1,149 university students to a sleep intervention or a control group, tracked their sleep with Fitbits, and then linked the results to official academic transcripts. Students in the treatment group slept about 19 more minutes per weeknight and earned semester grades 0.075 to 0.088 grade points higher. Because students were assigned at random, the difference is attributable to the intervention itself and not to the sort of student who already sleeps well. It is the first field experiment to demonstrate that improving sleep improves academic performance.

Why this study is different from the ones before it

The link between sleep and grades has been documented for years. Larger surveys have consistently found that students who sleep less earn lower grades, and a 2023 study tracking first-year students found that nightly sleep duration predicted GPA.

The problem with all of it was direction. Students who sleep well may differ from students who don’t in a dozen other ways — organization, workload, health, finances, whether they hold a job. Observational data cannot separate the effect of sleep from the effect of being the kind of person who gets it.

A randomized experiment can. Published in the Journal of Political Economy in January 2026 — DOI 10.1086/740221 — this study reports a field experiment run at the University of Pittsburgh across seven semesters between 2019 and 2022. The University of Pittsburgh announced the findings publicly in August 2026.

How the experiment worked

All 1,149 participants wore a Fitbit and installed a custom smartphone app. After a baseline period, they were randomly assigned to a control group or to a treatment.

The main treatment group received three things every weeknight for four weeks: a personalized bedtime reminder set about an hour earlier than their usual bedtime, morning feedback on whether they had hit a seven-hour sleep goal, and an immediate $4.75 payment when they did. Payment arrived by Venmo the same day. The control group wore the Fitbit and did nothing else.

Grades came from the University Registrar — actual transcripts, not self-reports.

What the students looked like at baseline

Before anything changed, the participants were sleeping like ordinary college students:

• An average of 6.6 hours per weeknight
• Seven or more hours on only 43% of weeknights
• Less than six hours on 29% of weeknights
• About half had an average bedtime after 1 a.m.; a quarter after 2 a.m.
• Only 4% hit seven hours on all five weeknights. One in five hit it on none.

Two further findings from the baseline data are worth sitting with. Students overestimated their own sleep: about two-thirds reported typically sleeping at least seven hours, while the Fitbit data showed only about a third managed it on three or more nights. And they wanted more sleep than they were getting — 94% reported an optimal sleep duration longer than their actual one.

So this was not a population that needed persuading. It was a population that was wrong about its own behavior and unable to change it on its own.

That distinction matters if you are reading this as a parent. These students were not indifferent — 94% of them said they wanted more sleep than they were getting. Wanting it and getting it turned out to be different problems, and only one of them is solved by being told sleep is important.

What changed

During the four-week intervention, the share of weeknights with at least seven hours of sleep rose by 11.5 percentage points — a 26% increase over baseline. Average weeknight sleep rose by an estimated 19 minutes.

Then the grades. Across all classes, students in the treatment group earned 0.075 more grade points that semester (p = 0.045). Restricted to lecture courses, which made up 80% of enrollments, the estimate was 0.088 (p = 0.035). Both are equivalent to about a tenth of a standard deviation.

Where the effect showed up

The gains were not evenly distributed, and the pattern is informative.

First-term freshmen showed substantially larger effects than other students. The authors suggest habits are more malleable before routines have set.

STEM courses drove most of the result: 0.132 grade points, against 0.013 in non-STEM courses. Two explanations are offered — sleep may matter more for quantitative work, and non-STEM grading left less room to move, with an average GPA of 3.69 compared with 3.27 in STEM.

Midday classes, those starting between 10 a.m. and 2 p.m., showed the largest effects. Notably not early-morning classes, which rules out the simplest explanation that better-rested students merely showed up to their 8 a.m.

What students gave up to sleep more

This is the part most relevant to a student weighing the trade-off.

Total study time did not change. Sleep did not come out of studying. What it came out of was screen time — internet, video and games, excluding screens used for studying — which fell by about 11.5 minutes a day, concentrated around bedtime. The size of the decrease was close to the size of the sleep increase.

The researchers also observed study time shifting from evening and night hours toward the morning. They are careful to say this pattern was not precisely estimated, so it should be read as suggestive rather than established. But the direction is consistent with the rest of the picture: the same hours of work, done earlier.

Did the change stick?

Partly, and the detail matters.

After the payments stopped, weeknight sleep stayed about eight minutes above the control group for the following four weeks — roughly half the effect. Sleep regularity also improved and persisted, with less week-to-week variation in bedtime and wake time.

But average bedtime returned to baseline. The researchers had designed the intervention around cue-based habit formation, expecting the bedtime reminder to trigger the behavior automatically once the reward was removed. That is not what happened. Students who kept receiving reminders did no better than students who stopped.

Their reading is that participants developed their own sustainable routines rather than responding to an external cue. Which is a more useful result than the one they were looking for.

How this compares with other ways of raising grades

The authors benchmark their result against programs designed to improve college performance directly. A recent meta-analysis found that performance-based financial incentives — paying students for grades — raise GPA by an average of 0.041 grade points at about $460 per student per semester.

The sleep intervention produced roughly twice the effect at about $112, including the cost of the Fitbit.

That comparison is about institutional policy rather than anything an individual student can act on. But it is a striking way to frame the underlying point: the cheapest available lever on academic performance turned out not to be an academic intervention at all.

What this doesn’t show

Three honest limits.

The effect is small. A tenth of a standard deviation is real but modest — roughly a quarter of the GPA gap between students with above- and below-median high school grades.

It faded. Grades were still higher the following semester, though that estimate did not survive correction for multiple testing. Two semesters later there was no effect at all.

It is one university. Participants were recruited through a research lab at a single institution, and women and Asian students were over-represented relative to the national college population. The authors themselves call for replication.

How this fits with what was already known

The most useful thing about this study is that its estimate converges with work done a completely different way.

An analysis of 55,322 undergraduates — observational, not experimental — found that each additional night of poor sleep in a typical week was associated with roughly 0.02 lower GPA. Studies of school start times, using natural variation rather than randomization, find that shifting start times by an hour increases sleep by 6 to 36 minutes and improves academic performance by 0.06 to 0.16 standard deviations.

This experiment lands at 19 minutes of extra sleep and 0.10 to 0.11 standard deviations. Three different methods, three different sets of assumptions, and estimates that sit on top of each other. That convergence is a stronger reason to believe the finding than any single study, including this one.

Frequently asked questions

Does more sleep actually cause better grades, or are they just correlated?
Until recently the evidence was correlational. A field experiment published in the Journal of Political Economy in January 2026 changed that. Researchers randomly assigned 1,149 University of Pittsburgh students to a sleep intervention or a control group, then linked the results to official academic transcripts. Students in the treatment group earned semester grades 0.075 to 0.088 grade points higher. Because assignment was random, the difference can be attributed to the intervention rather than to the kind of student who chooses to sleep more.

How much extra sleep did students actually get?
About 19 minutes per weeknight during the four-week intervention. The share of nights with at least seven hours of sleep rose by 11.5 percentage points, a 26 percent increase over the baseline rate of 43 percent.

Nineteen minutes sounds like very little. Why would that matter?
It is a small change, and the resulting grade effect is also small: 0.075 to 0.088 grade points, or about a tenth of a standard deviation. What makes it notable is that it is measurable at all from a modest change, sustained over four weeks, in a real semester rather than a laboratory.

Did students study less in order to sleep more?
No. Total study time did not change. The researchers observed a shift of study time from evening and night hours toward the morning, though they note this pattern was not precisely estimated. The clearest trade-off was screen time, which fell by about 11.5 minutes per day, concentrated around bedtime.

Which students benefited most?
Effects were largest among first-term freshmen, in STEM courses, and in classes held between 10 a.m. and 2 p.m. In STEM courses the estimated gain was 0.132 grade points. In non-STEM courses it was 0.013 — close to zero, though the authors note non-STEM grades were already high enough to leave little room to improve.

Did the sleep improvement last after the payments stopped?
Partly. Weeknight sleep remained about eight minutes above the control group for the following four weeks, roughly half the effect seen during the intervention. Sleep regularity also improved and persisted. Average bedtime, however, returned to baseline, which suggests students established sustainable routines rather than the automatic habit the researchers had designed for.

Do students know they are not sleeping enough?
They underestimate the shortfall. About two-thirds of participants reported sleeping at least seven hours on a typical weeknight, while Fitbit data showed only about a third managed that on three or more nights. At the same time, 94 percent said their optimal sleep would be longer than what they were actually getting.

Does this mean paying students to sleep is the answer?
The payment was the research tool used to change sleep, not the point of the finding. What the experiment establishes is that when sleep improved, grades improved. The authors note the intervention was unusually cost-effective compared with programs that pay students for grades directly, but the practical implication for an individual student is about sleep rather than about incentives.

Related reading:

• How Does Sleep Affect College Grades and Friendships? — the observational side of the same question
• How Does Lack of Sleep Make College Stress Worse? — the mechanism behind the effect

Sources:

• Giuntella, O., Saccardo, S. & Sadoff, S. (2026). “Sleep: Educational Impact and Habit Formation.” Journal of Political Economy. DOI: 10.1086/740221. Earlier version: NBER Working Paper 32550 (2024). Pre-registered at AEA RCT Registry, AEARCTR-0003235.
• Hartmann, M.E. & Prichard, J.R. (2018). “Calculating the contribution of sleep problems to undergraduates’ academic success.” Sleep Health, 4(5), 463–471.
• Creswell, J.D., et al. (2023). “Nightly sleep duration predicts grade point average in the first year of college.” PNAS, 120(8), e2209123120.
• Carrell, S.E., Maghakian, T. & West, J.E. (2011). “A’s from Zzzz’s? The causal effect of school start time on the academic achievement of adolescents.” American Economic Journal: Economic Policy, 3(3), 62–81.
• Lintner, T. (2024). “Effects of performance-based financial incentives on higher education students: A meta-analysis using causal evidence.” Educational Research Review.