Skip to content

Spaced Repetition for Language Learning: What the Research Actually Supports

Spaced repetition rests on two findings, not one. Here is what the research supports, what it does not, and how to use it without a scheduling algorithm.

Reign Creative Team8 min read

Key takeaways

  • Spaced repetition combines two separate findings — that reviews spread over time beat reviews packed together, and that recalling something is a stronger learning event than rereading it.
  • The specific intervals used by any given app are engineering conventions, not research results; the evidence supports spacing and retrieval, not one particular schedule.
  • Cramming produces higher confidence and worse retention, which is why the method that feels least effective in the moment is usually the one working.
  • Spaced repetition is a retention tool, not an acquisition tool — it keeps material you have understood, and does nothing for material you never understood.
  • You can get most of the benefit without any algorithm by reviewing on a widening schedule, always testing yourself before checking, and prioritising the items you keep getting wrong.

Everyone learning a language runs into the same wall. You study a set of words, you know them, and three weeks later they are gone — not fuzzy, gone. Then you relearn them, and they go again.

Spaced repetition is the standard answer to that problem, and it is a good one. It is also routinely oversold, wrapped in interval numbers presented as scientific findings when they are engineering choices, and applied to material it cannot help with.

This is what the research actually supports, what it does not, and how to get most of the benefit without any algorithm at all.

It is two findings, not one

"Spaced repetition" is usually treated as a single idea. It is two, and they are worth separating because they give different advice.

The spacing effect

Hermann Ebbinghaus, working on himself with nonsense syllables in the 1880s, established the basic shape of forgetting: retention of newly learned material drops steeply soon after learning and then more gradually. He also found that relearning was faster than learning, and that repetitions distributed over time were more efficient than the same repetitions packed together.

That second point is the spacing effect, and it has held up for well over a century of replication. Given a fixed amount of study time, spreading it across sessions produces better long-term retention than spending it all at once.

The 2013 Psychological Science in the Public Interest review of ten common learning techniques rated distributed practice as high utility — broadly applicable, well supported, and effective across ages, materials and subjects. It rated highlighting, rereading and summarisation considerably lower.

The testing effect

The second finding is less intuitive and does more work.

Retrieving information from memory is a stronger learning event than reviewing it. Trying to recall a word and succeeding strengthens it more than reading the word and its translation together, even though the second feels more productive. The Karpicke and Roediger work published in Science in 2008 made the point sharply: repeated retrieval practice produced far better long-term retention than repeated study, while learners' own predictions about which method was working were wrong.

The same 2013 review rated practice testing high utility alongside distributed practice. Between them, those two techniques were the only ones in the set to earn that rating.

This is why the format of a review matters as much as its timing. A flashcard you flip before genuinely attempting the answer has been converted from a test into a reread, and it loses most of its value. The pause where you sit there not knowing is not wasted time; it is the mechanism.

Why cramming feels better and works worse

Cramming produces a very specific illusion. Material seen five times in twenty minutes feels thoroughly known, because it is fluent and immediately available. Spaced review produces the opposite feeling — you keep failing on items you were sure you had learned.

Those feelings are almost exactly inverted from what is actually happening. Fluency in the moment is a poor predictor of retention later, and difficulty during practice often signals that useful work is being done. Researchers at UCLA's Bjork lab describe conditions of this kind as desirable difficulties: manipulations that slow learning down and depress performance during practice while improving retention and transfer afterwards.

The practical implication for a language learner is uncomfortable but simple. The study session where you feel competent is often the less useful one. The session where you keep drawing blanks and having to dig is usually the one paying off, provided you are still recovering the answers rather than failing outright.

What an interval schedule actually is

Here is where most explanations of spaced repetition overreach.

The research supports the principle: intervals should expand as an item becomes more secure, and a review is most valuable when recall is effortful but still successful. It does not hand down a table of days. When a system tells you to review at one day, then three, then seven, then twenty-one, those numbers are a design decision built on the principle — a reasonable one, but not a finding.

You should be sceptical of any product presenting its specific intervals as scientifically optimal. What matters, in order:

  1. That you review at all, more than once, after a gap.
  2. That the gaps widen rather than staying flat.
  3. That each review is a genuine attempt at recall before the answer appears.
  4. That items you keep failing come back sooner than items you keep getting right.

Everything past that is tuning, and the difference between a well-tuned schedule and a roughly sensible one is much smaller than the difference between reviewing and not reviewing.

Where spaced repetition does nothing for you

This is the part usually left out, and it is the reason a lot of people conclude the method does not work.

It is a retention tool, not an acquisition tool. It keeps material you have already understood. It does nothing for material you never understood in the first place. Drilling a grammar point you cannot parse produces a memorised string, not a usable structure — you will recall the card and remain unable to build the sentence.

It does not produce production ability on its own. Recognising a word on a card and reaching for it mid-sentence are different skills with different failure modes. Recognition is necessary and not sufficient. This gap is widest in exactly the languages where it hurts most: the Russian case endings covered in our guide to the Cyrillic alphabet can be recited from cards long before they can be chosen correctly while speaking, and the tone contrasts in Cantonese behave the same way.

It does not fix a backlog. The most common failure mode is not a bad schedule, it is a queue of several hundred overdue reviews that becomes psychologically impossible to face. The cure is preventative: add new material at a rate your future self can service, not at the rate your enthusiasm suggests today.

It is weakest on things that are not discrete items. Pronunciation, listening at speed, word boundaries in a script written without spaces, the reflex of choosing the right register — none of these decompose neatly into cards. They need exposure and production, not review.

Running it without an algorithm

You do not need a specialised system to get most of the benefit. You need three habits.

Always attempt before revealing. Whatever the tool — physical cards, an app, a notebook covered with your hand — say or write the answer before checking. If you skip this, you have a reading exercise.

Widen the gaps on a schedule you can hold. A workable pattern for a new item is same day, next day, later in the week, the following week, then the following month. Those numbers are a convention rather than a finding, and any similar expanding sequence will do the job. Consistency matters more than precision.

Sort ruthlessly by what you get wrong. Time spent on items you already know is close to wasted. Time spent on items you fail repeatedly is where the returns are. If you are tracking nothing else, track your failures.

Using the review modes in our courses

Worth being precise about what the apps do, because the honest version is less impressive than the marketing version and more useful.

Learn Cantonese for Beginners has seven review modes — Daily Review, Speed Run Review, Mistake Rescue, Listening Blitz, Sentence Builder, Match Blitz and Dictation Mode — all of which draw on material you have already covered, plus a missed-questions list that collects what you keep getting wrong. Learn Russian and Learn Cambodian, Speak Khmer keep saved words and missed questions in the same way, and the Khmer course adds flashcards you build and shuffle yourself.

What that gives you is the raw material for spaced review: a record of what you have met and a record of what is not sticking. What it does not give you is a per-item scheduler computing an optimal next review date, and we are not going to claim it does.

So the schedule is yours. In practice that means running Daily Review as the habit, using Mistake Rescue as the priority queue, and deliberately revisiting an old unit after a gap rather than only ever working forward. The mode variety helps for a reason worth naming: reviewing the same word through recognition, through dictation and through sentence building tests different retrieval routes to the same item, which is closer to how you will actually need it than a single card format ever gets.

One design detail matters here more than it looks. There is no hearts or lives system in these courses — nothing stops you after a mistake. A review method that depends on you failing frequently is incompatible with a system that punishes failure, and any app that makes errors expensive is quietly teaching you to avoid the difficulty that makes review work.

A weekly shape that survives contact with real life

Rough, adjustable, and built around the reality that some days you will do almost nothing:

  • Most days: one short review session first, then new material only if there is time left. Reviews before new content, always. When the two compete, reviews win — new material you never review is wasted effort, while reviews without new material still hold the line.
  • Two or three days a week: production practice that is not card-shaped. Speaking, dictation, building your own sentences. This is where recognition gets converted into use.
  • Once a week: deliberately open something from a few weeks back. This is the review your schedule will otherwise never prompt, and it is the one that shows you what has actually stuck.
  • Bad days: do the review session and nothing else. The habit is the asset.

What to expect

Honest framing, since this is where a lot of study advice starts making promises.

Spaced repetition will meaningfully reduce how much vocabulary you lose. It will make the second and third encounters with a word cheaper. It will show you clearly which items are weak, which is information you cannot get any other way.

It will not turn recognition into speech by itself, it will not teach you anything you have not first understood, and it will not compress the time a language takes — a point our guide to the Thai script makes in a different guise, where knowing every letter still leaves the harder half of the reading job undone.

What it does is stop you paying for the same words twice. Over months, that is a much larger effect than it sounds. If you want to see how the method applies to a specific language's particular sticking points, our language learning apps each have a beginner's guide that names them, and the Cantonese guide is the most detailed on the recognition-versus-production gap.

Good to know

Frequently asked questions

What is spaced repetition?
Spaced repetition means reviewing material at intervals that widen over time rather than reviewing it repeatedly in one sitting. The underlying finding, known as the spacing effect, is that the same total study time produces better long-term retention when it is distributed across days than when it is packed into one session. Most implementations add a second ingredient, retrieval practice, by making you recall the answer before showing it.
Is spaced repetition better than cramming?
For retention beyond a few days, distributed practice consistently outperforms massed practice at equal total study time. Cramming does work for an exam tomorrow, which is why it survives as a strategy. The problem for language learners is that they need the material next month and next year, not on Tuesday, so the case for spacing is much stronger for vocabulary than it is for a one-off test.
What is the ideal spaced repetition interval?
There is no single answer, and any app that presents its numbers as scientifically derived is overstating the case. The research supports the general principle that intervals should widen as an item becomes more secure, and that a review is most valuable when recall is effortful but still successful. The exact day counts in any given system are design decisions built on that principle rather than findings in their own right.
Do Reign Creative apps use a spaced repetition algorithm?
No, and it would be misleading to say otherwise. Learn Cantonese for Beginners has review modes that draw on material you have already covered, plus a missed-questions list that collects what you keep getting wrong, and other courses keep saved words and missed questions in the same way. That gives you the raw materials for spaced review, but the schedule is yours to set rather than something the app computes per item.
How many words should I review per day?
Fewer than you think, and consistently rather than in bursts. The constraint is not how many new items you can meet in one sitting but how many reviews your future self will actually complete, because a backlog of overdue reviews is the single most common reason people abandon a review habit. A daily amount you will still do on a bad day beats an ambitious one you will do for nine days and then drop.

Sources

  1. Memory: A Contribution to Experimental Psychology (1885) (opens in a new tab)Classics in the History of Psychology, York University — accessed
  2. Improving Students' Learning With Effective Learning Techniques: Promising Directions From Cognitive and Educational Psychology (Dunlosky et al., Psychol Sci Public Interest 2013;14(1):4–58) (opens in a new tab)PubMed, U.S. National Library of Medicine — accessed
  3. The Critical Importance of Retrieval for Learning (Karpicke & Roediger, Science 2008;319(5865):966–8) (opens in a new tab)PubMed, U.S. National Library of Medicine — accessed
  4. Bjork Learning and Forgetting Lab (opens in a new tab)University of California, Los Angeles — accessed

Apps mentioned in this article

Keep reading