Spaced Repetition: Does It Work? What the Research Found, and Who It Does Not Suit

By AI Anyone · 2024-06-11 · Updated Oct 21, 2026 · 9 min read · Learning Tips

Spaced repetition has some of the strongest evidence in learning research, for long-term recall. Here is what Ebbinghaus, Murre and Dros, Cepeda and others actually found, how large the benefit is, and who it does not suit.

Spaced repetition works for its main job, which is keeping what you learned available weeks and months later. Across hundreds of memory experiments, studying the same material in sessions spread over time has beaten studying it all at once, and reviews of the research rate it among the most useful study techniques we have. It is weaker at helping you understand something new, it cannot rescue a test that is tomorrow morning, and it asks for a daily habit that some people will not keep.

That is the short answer. The rest of this article is the evidence behind it: what the forgetting curve shows, which studies tested spacing and what they actually found, how big the benefit is, and who should use something else.

Plenty of people are asking. Spaced repetition draws about 5,400 US searches a month, yet the direct question "does spaced repetition work" registers only about 20 (DataForSEO, October 2026). Most searchers want to know what it is; far fewer stop to ask whether it delivers. Both deserve a straight answer.

What is spaced repetition?

Spaced repetition is reviewing material at increasing intervals, with each review timed for when you are starting to forget, rather than reviewing it many times in one sitting. It rests on an older and broader finding called the spacing effect: the same amount of study produces better long-term memory when it is split across separate sessions.

Two terms often get blurred:

  • The spacing effect is the research finding. Spread-out study beats massed study (cramming) for later recall.
  • Spaced repetition is a method built on it. You break material into items, usually flash cards or questions, and review each one on its own schedule, with gaps that grow when you get it right and shrink when you get it wrong.

The top Google result for "spaced repetition" this month is its Wikipedia entry (DataForSEO, October 2026), which is a fair sign that most searchers are still at the definition stage.

Nearly every version also tests you at each review. You try to recall the answer before seeing it. That makes spaced repetition a combination of spacing and retrieval practice, and much of its power comes from the pairing.

The forgetting curve behind spaced repetition

About 20 people a month in the US search "spaced repetition curve" (DataForSEO, October 2026), and the curve they mean is the forgetting curve. Memory for new material falls off quickly at first and then more slowly, and the classic evidence for this comes from one man testing himself. Ebbinghaus (1885) memorized lists of nonsense syllables, then measured how much less time he needed to relearn each list after delays ranging from minutes to about a month. The savings dropped steeply over the first day and leveled off after that. The shape he drew is now called the forgetting curve.

A single subject in the 1880s is thin evidence on its own, which is why it matters that Murre and Dros (2015) repeated the experiment with a modern participant following Ebbinghaus's method closely, and got a curve very similar to his. Read their paper.

What the curve suggests for study is simple. A review that happens soon after learning catches the memory before most of it is lost. Each review also seems to slow the next decline, so the following review can wait longer. Spaced repetition schedules (1 day, then 3, then 7, and so on) are an attempt to ride that pattern.

One caution: the curve was measured with meaningless syllables. Material you understand and connect to other ideas usually fades more slowly. The shape holds; the exact speed depends on what you are learning.

Does spaced repetition work? What the studies found

Yes, for long-term recall the evidence is strong and consistent. These are the studies most worth knowing, with what each one tested.

Cepeda et al. (2006), a review of distributed practice. Published in Psychological Bulletin, this review brought together several hundred experiments on remembering words and facts. Spreading practice across sessions produced better recall than massing it, across many kinds of verbal material. Read the review.

Cepeda et al. (2008), the best gap depends on the test. More than a thousand people learned obscure facts online, waited between study sessions for gaps of different lengths, and were tested after delays of up to about a year. The gap that produced the best recall grew as the test delay grew. A short gap suits a test next week; a longer gap suits knowledge you want next year. Read the study.

Rohrer and Taylor (2007), maths problems. Spacing is not limited to vocabulary lists. In one of their experiments, students who split the same set of practice problems across two sessions did better on a test a week later than students who did all the problems at once. Read the study.

Dunlosky et al. (2013), a rating of ten study techniques. This large review judged common techniques on how well the evidence supports them across learners, materials and settings. Distributed practice was one of only two techniques rated high utility; the other was practice testing. Rereading, highlighting and summarizing were rated low. Read the review.

Kang (2016), from lab to classroom. Written for educators and policy makers, this review concluded that spaced repetition improves long-term retention efficiently and recommended building it into teaching. Read the review.

How the main techniques compare

TechniqueRating in Dunlosky et al. (2013)What it is for
Distributed practice (spacing)HighKeeping material for weeks and months
Practice testingHighMaking memories easier to recall
Interleaved practiceModerateTelling similar problem types apart
Elaborative interrogationModerateUnderstanding why facts are true
Self-explanationModerateUnderstanding steps and reasoning
RereadingLowFamiliarity, not recall
HighlightingLowVery little on its own

The two high ratings are the two halves of a spaced repetition session: spacing decides when, testing decides how.

Spaced repetition theory: why gaps help memory

Nobody has one agreed explanation for the spacing effect, and the honest version is that several ideas probably contribute. The common ones:

  • Retrieval effort. After a gap, recalling the answer takes work, and effortful recall seems to strengthen memory more than easy recall. When you review five minutes after studying, the answer is still in mind and the review does little.
  • Varied context. Sessions on different days happen in different moods, places and frames of mind, which may give a memory more routes back to it.
  • Attention. The second pass in a massed session gets less attention because the material feels familiar. A spaced review feels newer, so you process it more fully.

None of these needs to be settled for you to use the method. They do explain a practical point: a review that feels too easy is probably too early.

How big is the benefit, honestly?

There is no single number for how much spaced repetition helps, and you should be wary of anyone who quotes one. The size of the effect depends on at least three things.

  1. The gap compared with the test delay. Per Cepeda et al. (2008), a gap that is right for a one week test is too short for a one year test.
  2. The material. Most experiments use facts, word pairs and vocabulary. Evidence for problem solving exists (the maths study above) but is thinner, and evidence for physical skills and complex reasoning is thinner still.
  3. What you do at each review. Spacing out rereading helps less than spacing out self-testing, because the testing is doing a lot of the work.

It also does not feel like it is working. Kornell and Bjork (2008) found that people learning painters' styles did better when the paintings were mixed and spread out, yet most of them believed the blocked, back to back version had taught them as much or more. Read the study. That gap between feeling and result is why so many students go back to cramming: cramming feels like learning while you do it.

Who spaced repetition does not suit

Spaced repetition is not the right first tool in every case. Be skeptical of it here:

  • You do not understand the material yet. Scheduling flash cards for an idea you cannot explain turns it into memorized words. Learn it first (worked examples, explanations, asking why), then space the review.
  • Your test is tomorrow. There is no room left for gaps. Test yourself on the most likely questions tonight and sleep, then build a schedule for the next course.
  • You are learning a hands-on skill. Spacing practice sessions is sensible, but most of the evidence is about memory, not movement. Treat the card-and-app version as a support for the facts behind a skill, not the skill itself.
  • You will not keep a daily habit. A schedule that collapses after ten days into a backlog of hundreds of reviews does more harm to motivation than good to memory. A weekly review session, spaced across the term, still uses the effect.
  • The course already spaces itself. Cumulative subjects such as maths and languages revisit old material constantly. Your extra reviews can be lighter there.

What a spaced repetition algorithm adds

A spaced repetition algorithm is the rule an app uses to set each item's next review date from your past answers. About 90 people a month in the US search "spaced repetition algorithm" (DataForSEO, October 2026), often while comparing apps.

The research above supports spacing and testing in general. It does not tell you which app's formula is best, and the difference between reasonable schedules is smaller than the difference between reviewing and not reviewing. A paper box with five sections and fixed gaps of 1, 3, 7, 14 and 30 days uses the same principle. Choose whatever you will open every day.

For more on how the timing works in practice, see our guide to the science of distributed practice, and for the related idea of mixing topics in one session, interleaving as a learning strategy.

Try this: test the spacing effect on yourself

Run a small two week experiment instead of taking the research on faith.

  1. Pick 40 items you need to learn anyway: vocabulary in a language, anatomy terms, formulas, dates for a history course.
  2. Split them at random into two lists of 20.
  3. List A, massed. On day 1, study it for 30 minutes in one sitting, testing yourself until you get every item right.
  4. List B, spaced. Study it for 10 minutes on day 1, 10 minutes on day 3 and 10 minutes on day 7, testing yourself each time. Same total time.
  5. Day 14, test both. No looking. Count how many of each list you get right.

Write down which list you expected to win before you test. Most people expect the massed list, because it felt solid on day 1.

If you want help generating the items, on AI Anyone Courses you can type a topic and it builds a short course with lessons and quick checks; pull your 40 items from those. For the day 14 test, call a character in your browser and ask it to quiz you on both lists in random order, one item at a time. After the call the transcript lands in your chat, so you can count your score against it. Or browse characters to quiz you and pick a teacher whose subject matches your list.

The bottom line

Spaced repetition has some of the best evidence in learning research behind it, as long as you know what that evidence covers: long-term recall of material you already understand, strongest when each review is a self-test. It will feel slower than cramming and work better. Use it for what you need to keep, and use something else to understand things in the first place.

Spaced RepetitionLearning ScienceMemoryStudy Techniques

Keep going with this topic

Two ways to go deeper right now. Chat is free with an account, and a full course takes a few minutes to build.

Chat free with a science and tech mentor Generate a course on this
More from the AI Anyone blog →
© 2026 AI Anyone. All rights reserved.