한국어

Sleep or Study Before an Exam? What the Research Says

At a glance
  • Sleep, in most cases. Students who traded sleep for extra study had more trouble in class and on tests the next day.
  • In one experiment, an hour spent napping still showed a benefit a week later; an extra hour of cramming no longer did.
  • Grades tracked sleep over the weeks the material was taught, not sleep on the night before the test.
  • Cramming feels effective. 90% of learners remembered more with spacing, yet 72% believed massing had worked better.
  • Tonight: set a bedtime, spend the time left on closed-book recall, then sleep. Next time, spread the review out.

In most cases, sleep. Students who cut sleep to study more than usual had more trouble understanding class and doing tests the next day. In one experiment, an hour spent napping still helped memory a week later, while an extra hour of cramming no longer did. And grades tracked how students slept during the weeks the material was taught, not on the night before the exam.

Does pulling an all-nighter help you study?

It gives you more hours, but less to work with the next day. Days after students traded sleep for study went worse.

Gillen-O'Neel and colleagues (2013) had 535 high school students keep a daily diary for 14 days in 9th, 10th and 12th grade: how long they studied each night, how long they slept, and what went wrong at school the next day. Whether a student usually studied a lot or a little, on days after they cut sleep to study more than usual, they had more trouble understanding what was taught in class and were more likely to struggle on an assignment or test. Because students sacrificed sleep for study more often in later grades, the cost grew over time.

Staying up also weakens the next day's learning. Yoo and colleagues (2007) had people either sleep normally or stay awake all night, then learn new material the next day. After a sleepless night, activity in the hippocampus, the brain area that forms new memories, dropped markedly during learning, and people remembered less later. The authors concluded that sleep before learning prepares the brain to form new memories the next day.

So the morning review before the test, and making sense of an unfamiliar question on the paper, both happen with a brain that missed its night.

An hour of cramming or an hour of napping: which lasts?

The nap. The cramming benefit lasted until that evening and then faded.

Cousins and colleagues (2019) had 84 university students in Singapore learn facts about twelve species of ants and crabs over an afternoon, in four 40-minute blocks. In the middle, each group spent one hour differently: one napped (about 60 minutes on average), one studied the same material for another hour, and one watched a movie. Everyone was tested 30 minutes after learning and again a week later.

The cramming benefit was gone after a week; the nap benefit remained

  • 1-hour nap
  • 1 more hour of cramming
00.30.60.91.21-hour nap · After 30 minutes · 0.950.951 more hour of cramming · After 30 minutes · 0.60.6After 30 minutes1-hour nap · After a week · 0.560.561 more hour of cramming · After a week · 0.270.27After a weekOnly the nap still madea significant difference
Effect size (Cohen's d) against the group that watched a movie for the same hour, for memories rated "certain" (correct minus incorrect). The 0.27 for cramming after a week was not statistically significant.Source: Cousins et al. (2019), Results
See the numbers
1-hour nap1 more hour of cramming
After 30 minutes0.950.6
After a week0.560.27

After 30 minutes, both the nappers and the crammers remembered more than the movie group. That is no surprise for the crammers, who had an extra hour of study. A week later, the gap between cramming and the break had shrunk to nothing significant, while the gap between the nap and the break remained. The nappers also reported feeling less sleepy before the afternoon's final study blocks.

One caveat. Compared directly, the nap and cram groups did not differ significantly after a week. So the accurate reading is not "napping beats cramming." It is "studying an hour less and sleeping instead cost nothing, and only the sleep still showed a benefit a week later."

Does sleep the night before an exam decide your grade?

No. Grades tracked sleep during the weeks the material was taught, not on the night before.

Okano and colleagues (2019) gave Fitbits to 88 students in an MIT chemistry course for a full semester and compared their sleep with eight quizzes and three midterms. How long or how well students slept the night before a test was not related to their score on it. Average sleep over the month before each midterm, and the week before some quizzes, was.

Sleep in the weeks before a test tracked scores; the night before did not

00.10.20.30.40.5Correlation between sleep duration and score (r)NightbeforeNight before · 0.01~0.260.01~0.26quizzes: none significantWeekbeforeWeek before · 0.3~0.40.3~0.43 of 8 quizzes significantMonthbeforeMonth before · 0.25~0.340.25~0.34all 3 midterms significant
Range of correlations (r) between sleep duration and test scores. The night before quizzes was never significant; the week before was significant for 3 of 8 quizzes; the month before was significant for all 3 midterms. These are correlations, so they do not show that sleep caused the scores.Source: Okano et al. (2019), Results
See the numbers
FromTo
Night before0.010.26
Week before0.30.4
Month before0.250.34

Sleep on the night before each midterm was not related to the score either (all r below 0.20). The authors suggest that sleeping well while the tested material is being taught may matter more than the night before. Together, sleep duration, quality and consistency accounted for about 24% of the differences in semester scores.

When students usually slept mattered too. Splitting them at the median bedtime (1:47 a.m.) and median wake-up time (9:12 a.m.) gives this.

Semester averageEarlier than medianLater than median
Bedtime77.25%70.68%
Wake-up time78.28%69.63%

These are mean overall semester scores, quizzes and exams combined, from the Results of Okano et al. (2019).

This was an observational study, not an experiment. Students who sleep well might also be the ones who plan well. Still, it points the same way as the experiments above: sleep is not something to catch up on the night before, but something to keep while you learn.

Why does cramming feel like it works?

Right after a cram session the material is vivid, so it feels known. That feeling did not predict memory days later.

Kornell (2009) had participants learn GRE-style word pairs with web-based flashcards. With the same number of looks, he compared studying cards massed together with spacing them out in one large stack, and also with cramming, meaning massing study on the last day before the test. Across three experiments, spacing beat massing.

In one line: 90% of participants remembered more when they spaced their study. Yet after the first session, 72% believed massing had worked better.

That is why cramming seems to pay off. With the test the next morning, you can hold on for one day. But if you need the subject again next term, what you crammed is the first to go. When to review instead is covered in Spaced repetition schedule: when should you review?

The exam is tomorrow. What should you do tonight?

Spend the time you have left on closed-book recall instead of rereading, and go to bed at a set time. Then plan the next exam differently.

  1. Pick your bedtime first. The time between now and then is tonight's study time. The studies above found that cutting sleep made the next day worse.
  2. Close the book and recall. Look only at the section headings, write what you know on a blank page, then open the book and fill in only the gaps. With little time, finding what you do not know comes first. The method is in Is note-taking a waste of time?
  3. Let go of what you cannot cover. Making a few likely topics solid to recall lasts better than skimming everything once more.
  4. Recall again for 10 minutes in the morning. Cover last night's page and write it out again. Whatever is missing is what to look at before the test.
  5. Spread it out next time. If you recall a little each week while the material is taught, there is less left for the night before.

If you have an hour free during the day, sleeping instead of resting is an option too. In the experiment above, the nap group studied an hour less and lost nothing.

In Aloom

What cramming cannot give you is recall spread over time. In Aloom, the answers you keep from asking AI come back as questions, as many each day as you choose (1, 3 or 5). You recall a little every day instead of all at once, and each time you do, the node's ink grows darker. If you ask and keep answers while a course is running, the night before the exam you already have things you have recalled several times. See how Aloom works

FAQ

Is it better to pull an all-nighter or sleep for 3 hours?

The studies here did not compare 3 hours with none. They do agree on direction: the more sleep you cut, the harder understanding and new learning got the next day. If you must stay up late, set a bedtime and get some sleep rather than none.

Is cramming effective at all?

For a test the next day, somewhat. In the experiment above, the group that studied an extra hour did remember more after 30 minutes. A week later that advantage was no longer significant. For anything you need to keep, spacing lasts much better.

What if I can't sleep the night before an exam?

In one semester-long study, sleep on the night before a test was not related to the score. A restless night does not doom the exam. If you slept well in the weeks before, that is already working for you.

References