ResearchPsychological flexibility

Does spaced repetition actually work?

Last reviewed 17 August 2026

The verdict

Yes, and by a large enough margin to change how you learn anything. Across 22 classroom studies and more than 3,000 learners, spacing beat cramming at d = 0.54. Weekly intervals were the most consistent, fewer sessions worked better than many, and gains grew as the delay before testing grew. Studies of procedural skills are less consistent than studies of facts.

Strength of evidence

  1. traditional
  2. emerging
  3. informed
  4. established

Supported by controlled trials or meta-analysis, replicated by more than one research group. How we grade

What is spaced repetition?

Spacing means spreading your study of something across several sessions with gaps between them, instead of doing it all at once. The alternative, cramming, is called massed practice. Spaced repetition adds a rule for when to revisit material, usually widening the gap each time you get it right.

The comparison researchers make is straightforward. Hold total study time constant, split it differently, and see who remembers more later. That last word does a lot of work, because the advantage shows up at delay rather than immediately.

How much better is it?

Substantially. A 2025 meta-analytic review of classroom research found distributed practice beat massed practice at d = 0.54 across 22 studies, 31 effect sizes, and more than 3,000 learners. That is roughly half a standard deviation, which is large for an education intervention that costs nothing.

Mawson and Kang published this in Behavioral Sciences, and the classroom setting matters. Much of the spacing literature comes from laboratories with word lists, and this review deliberately looked at applied research in real teaching.

Heterogeneity was very high at 92 per cent, so the studies disagree with each other considerably and the authors could only describe moderators rather than test them statistically. They also note some studies confused spacing with interleaving, particularly in mathematics.

What intervals should you use?

Weekly gaps produced the most consistent positive effects in this review. Longer delays before the final test showed larger benefits, though no study ran past 42 days, so the long-horizon question stays open.

One finding runs against the instinct to do more. Fewer re-exposures, three or under, produced stronger benefits, with diminishing returns as sessions multiplied. Reviewing something six times in a week is closer to cramming than to spacing.

Effects were larger in secondary school than primary, and larger for language learning, though mathematics was consistently positive too. University-level studies were scarce.

Why does cramming feel better?

Because it works, briefly. Massed study produces strong immediate recall, which feels like learning while you are doing it. The material fades faster afterwards, and the exam is usually far enough away for that to matter.

Spacing feels worse in the moment. You return to material you have partly forgotten, retrieval is effortful, and effort reads as failure. That difficulty is the mechanism, and the discomfort is the price of the durability.

How do you actually use it?

Split any study you were going to do in one sitting across three sessions about a week apart. That single change captures most of the effect and requires no software.

Make each session retrieval rather than review. Close the book and try to produce the answer before checking, since the spacing effect and the testing effect compound and rereading captures neither.

Resist adding sessions. Three well-spaced passes beat six crowded ones on this evidence, which makes spacing cheaper than the method it replaces.

For physical or procedural skills, be more careful. Some studies found spacing less robust, or absent, when the thing being practised is a procedure rather than a fact.

Questions people ask

How long should the gaps be?
Weekly intervals were most consistent in the classroom review. Longer gaps helped when the test was further away, which is the case that matters for anything you want to keep.
Do flashcard apps work?
They automate the scheduling, which is the useful part. The evidence supports spacing and retrieval, and the app is one convenient delivery method for both.
Is more repetition always better?
No. Three or fewer re-exposures produced stronger benefits than more, with diminishing returns beyond that.
Does it work for skills as well as facts?
Less reliably. Some studies of procedural learning found weak or absent effects, so the strongest support is for factual and language material.

Sources

Every linked source below was opened and checked on the review date. Where a record could not be opened, it is named in the text with no link.

  1. Mawson RD, Kang SHK (2025). The Distributed Practice Effect on Classroom Learning: A Meta-Analytic Review of Applied Research. Behavioral Sciences.

    Twenty-two studies, 31 effect sizes, more than 3,000 learners, d = 0.54 favouring spacing. Very high heterogeneity at 92 per cent limited formal moderator analysis, and no study exceeded a 42 day retention interval.

More research

All briefs

Put this into practice with people doing the same

Inside HumanOS, people compare approaches with the evidence and work useful practices into ordinary days.

Join HumanOS free