How to Stop Forgetting Vocabulary: Research and Real Data

Forgetting vocabulary is usually a method problem, not a memory problem. Cover the answer and recall it before you check, review the next day, then let the gaps between reviews grow. Done that way, 8 to 9 in 10 words are still remembered a week to a month later. That is what we see in Tomaru's anonymized learning data, and it matches more than a century of memory research.

• Recall beats rereading. Students who reread a text four times remembered 40% of it a week later. Students who read it once and tested themselves three times remembered 61% (Roediger & Karpicke, 2006).
• The next-day review saves the most words. Review a brand-new word the next day and about 6 in 10 stick. Wait a week and it's under half.
• Spread it out. Across 317 experiments, studying in separate sessions beat studying in one block, 47.3% to 36.7% (Cepeda et al., 2006).

Written by the Tomaru team, the people who build the Tomaru vocabulary app. Source for Tomaru figures: Tomaru anonymized learning data (2026).

The answer in one table

Why we forget new words

Memory loss is fastest right after learning, then slows down. Hermann Ebbinghaus showed this in 1885 by memorizing lists of nonsense syllables and measuring how much time he saved when he relearned them later. In 2015, researchers at the University of Amsterdam repeated his experiment and got almost the same curve: after 31 days, relearning a list saved only about 20% of the original learning time (Murre & Dros, 2015, *PLOS ONE*).

The famous "you forget 67% in a day" is a misreading. Ebbinghaus measured savings, meaning how much faster relearning was, not the share of words still remembered. His savings were 58% after 20 minutes, 34% after one day and 21% after 31 days. He used nonsense syllables, so read the curve as a shape, not as exact numbers for real words.

The curve describes something learned once and then left alone. That is how most people study vocabulary: read the list, copy it out, move on to the next page. Losing a word you saw only once doesn't mean your memory is bad. It means the method stopped too early. The rest of this guide is about the days after you first see a word.

For a longer explanation, see the forgetting curve.

Why rereading and copying don't stick

Rereading feels productive because the words look familiar, but familiarity is not recall. In a study of 120 college students, people who reread a passage scored higher five minutes later (81% vs 75%), but after a week, those who had tested themselves remembered more (56% vs 42%). In a second experiment, students who read a text four times remembered 40% a week later, while students who read it once and tested themselves three times remembered 61% (Roediger & Karpicke, 2006, *Psychological Science*).

Rereading right away barely helps. Across four experiments with textbook chapters and magazine articles, reading the same material a second time immediately did not meaningfully improve test scores, for strong and weak readers alike (Callender & McDaniel, 2009, *Contemporary Educational Psychology*).

Psychologists rate rereading and highlighting as low-utility study methods. A review of 10 common techniques rated only two as high utility: practice testing and distributed practice. Rereading, highlighting, summarizing, keyword mnemonics and imagery were rated low utility, and the authors suggest replacing rereading with self-testing (Dunlosky et al., 2013, *Psychological Science in the Public Interest*).

Copying a word ten times has the same weakness. Like rereading, it is done with the answer in front of you, so you never practice pulling the word out of memory. Writing a word is fine. Writing it from memory, then checking, is much better.

What works: recall first, then space it out

Recall the answer before you look at it. In an experiment with 40 Swahili–English word pairs, students who kept testing themselves on words even after getting them right remembered about 80% a week later. Students who stopped testing a word once they got it right remembered only 33–36%, and extra rereading did not close the gap (Karpicke & Roediger, 2008, *Science*). It works the same way for Japanese 慣れる, Korean 습도 or Chinese 参加: look at the word, say the meaning, then check.

When you get a word wrong, see the answer right away. In a study of people learning Luganda–English words, seeing the correct answer after a mistake made a large difference to what they remembered a week later; feedback after correct answers made almost no difference (Pashler et al., 2005, *Journal of Experimental Psychology: Learning, Memory, and Cognition*).

Give yourself a few seconds to try. English speakers learning Spanish words from pictures remembered more when they got a short pause to recall the word before seeing it, both right away and a week later (Barcroft, 2007, *Language Learning*).

Spread the same study over several sessions. A review of 317 experiments found that people who studied in separate sessions averaged 47.3% on later tests, compared with 36.7% for the same amount of study done in one block; only 12 of 271 comparisons showed no benefit or the reverse (Cepeda et al., 2006, *Psychological Bulletin*).

It holds for language learning too. A meta-analysis of 48 second-language experiments found the benefit of spacing to be medium to large, and when the test came later, longer gaps did better than short ones (Kim & Webb, 2022, *Language Learning*).

For more on the method itself, see spaced repetition for vocabulary and the testing effect.

How far apart should reviews be?

The right gap depends on how long you need to remember. In a study of 1,354 people, the best gap between two study sessions was 1 day when the test was a week away, about 11 days when it was 35 days away, and about 21 days when it was 70 days or a year away. Choosing a good gap improved final recall by about 64% compared with reviewing on the same day (Cepeda et al., 2008, *Psychological Science*). Put simply: the further away your exam, the longer the gaps can be.

Whether gaps grow or stay fixed matters less than people think. A meta-analysis of 29 studies found no meaningful difference between expanding gaps (1, 3, 7 days...) and equal gaps (Latimier, Peyre & Ramus, 2021, *Educational Psychology Review*). A vocabulary experiment with Japanese university students found the same: whether you space at all matters far more than the exact pattern (Nakata, 2015, *Studies in Second Language Acquisition*).

The practical rule: see a new word again the next day. If you remember it, let the next gap be longer. If you forget it, bring it back soon.

What is the 2-7-30 rule, and does it work?

The 2-7-30 rule is a rule of thumb, not a research finding. It usually means reviewing something 2 days, 7 days and 30 days after you first learn it. No specific study points to those three numbers.

What it gets right: it spaces reviews out, and the gaps get longer. Both ideas have strong support (Cepeda et al., 2006; Kim & Webb, 2022).

Where it falls short for vocabulary:

• The first review comes late. For brand-new words, the next day is the safest first review: about 6 in 10 stick when you review them the next day, and under half when you wait a week.
• Every word gets the same schedule. Some words you'll never forget after one look; others slip again and again. A fixed calendar spends the same time on both.
• It stops at 30 days. If your exam is three months away, Cepeda's results suggest one more review, with a longer gap.

So: 2-7-30 is far better than cramming, and fine for a handful of facts. For hundreds of words, let each word's next gap depend on whether you remembered it.

What we see in Tomaru's data

We build Tomaru, so we can see how words hold up when people study on a schedule. Two things stand out.

On a schedule, most words stay. When a word comes back a week or a month after the last review, 8 to 9 in 10 are still remembered.

The first review decides a lot. For words people didn't know at all the first time, a review the next day keeps about 6 in 10. When that first review slips to a week later, it's under half. If you only change one habit, make it the next-day review.

How many new words a day?

Pick a number you can keep for months. Every new word brings reviews with it over the following weeks, so a number that feels easy on day one gets heavier by week three. Ten to twenty new words a day works for most people.

How long common word lists take at 20 new words per study day:

For word counts by level, see JLPT vocabulary by level, TOPIK vocabulary by level, EPS-TOPIK vocabulary and New HSK 3.0 vocabulary by level.

How to stop forgetting vocabulary: 8 steps

1. Choose one word list that matches your goal, such as an exam level or a frequency list, and stick with it instead of mixing books and apps.
2. Hide the answer. Look at the word and say the meaning out loud, or look at the meaning and produce the word, before you check.
3. Mark honestly. If you hesitated a long time or got it wrong, count it as forgotten and look at the answer right away.
4. Review the next day. This is the step that protects the most words.
5. Let the gaps grow for words you remember: a few days, then one to two weeks, then a month. Bring forgotten words back the next day.
6. Do reviews before new words. On a busy day, spend your five minutes on reviews and skip new words.
7. Keep new words small and steady: 10 to 20 a day, every day, beats 100 on Sunday.
8. Count back from your exam date. Divide the list by the study days you have, and stop adding new words one to two weeks before the exam so the end is review only.

Doing this in Tomaru

We wrote this guide and we build Tomaru, so here is how it handles the steps above. Tomaru has ready-made word lists for English speakers learning Japanese (JLPT N5–N1, Irodori A1/A2 for JFT-Basic), Korean (TOPIK I and II, EPS-TOPIK), Chinese (HSK 1–9, following the 2026 syllabus) and Spanish (Core 1000 and 2500). Set an exam date on any list and it works out how many new words a day you need and roughly how many minutes that takes. Each answer button shows when the word will come back, and the day after you start, it tells you how many of yesterday's words you still remember. The core vocabulary features are free forever, with no ads. You can browse every list on the word lists page.

FAQ

How many times do you need to review a word to remember it?

Fewer than most people think, if the timing is right. Recalling a word several times in one sitting helps, but once study time is taken into account, spreading those recalls across different days is the better deal (Nakata, 2017, *Studies in Second Language Acquisition*). In practice: the next day, a few days later, one to two weeks later, then a month. Words you remember drop out of the rotation quickly; words you forget come back sooner.

How much vocabulary do you forget after a day, a week or a month?

Without review, forgetting is fastest in the first day, then slows down (Ebbinghaus, 1885; Murre & Dros, 2015). With scheduled reviews, it looks very different: in Tomaru's anonymized learning data, 8 to 9 in 10 words are still remembered when they come back a week or a month later.

What is the 2-7-30 rule for memory?

It is a rule of thumb for reviewing 2, 7 and 30 days after learning. No specific study backs those numbers. Spacing and growing gaps are well supported, but for new vocabulary a first review on the next day is safer, and gaps work best when they depend on whether you remembered each word.

Can I memorize 1,000 words in a day?

You can look at 1,000 words in a day, but you won't keep most of them. Every word needs a few reviews over the following weeks, so 1,000 new words today means thousands of reviews later. In one experiment, recalling a word more times in a single session helped, but once study time was counted, spreading practice across days was the better use of time (Nakata, 2017). Ten to twenty new words a day is a pace you can keep.

Is writing words down a good way to memorize them?

Writing a word from memory, then checking, is a form of self-testing and works well. Copying a word many times while looking at it works like rereading: it feels like effort but does little for recall a week later.

Is spaced repetition better than cramming?

Yes, for anything you need to remember for more than a few days. Across 317 experiments, spaced study averaged 47.3% on later tests vs 36.7% for studying the same material in one block (Cepeda et al., 2006). Cramming can get you through a test tomorrow morning, but most of it is gone soon after.

Should review gaps keep getting longer?

Gaps should get longer as a word becomes familiar, but research finds little difference between strictly expanding gaps and equal gaps (Latimier, Peyre & Ramus, 2021). Having gaps, and making them long enough for your exam date, matters more than the exact pattern.

Does this work for Japanese kanji, Korean or Chinese characters?

Yes. Many of the experiments above used foreign-language words (Swahili, Luganda, Spanish and English as a second language), and a meta-analysis of 48 second-language experiments found medium-to-large benefits from spacing (Kim & Webb, 2022). For Japanese and Chinese, learn the reading and meaning together. For Korean, listen to each word, because spelling and pronunciation often differ.

I already forgot words I learned months ago. Do I have to start over?

No. Relearning is faster than learning the first time; that is exactly what Ebbinghaus measured. Go back through the same list, test yourself instead of rereading, and review again the next day. The words come back much faster than they took the first time.

References

• Barcroft, J. (2007). Effects of opportunities for word retrieval during second language vocabulary learning. *Language Learning, 57*(1), 35–56. https://doi.org/10.1111/j.1467-9922.2007.00398.x
• Callender, A. A., & McDaniel, M. A. (2009). The limited benefits of rereading educational texts. *Contemporary Educational Psychology, 34*(1), 30–41. https://doi.org/10.1016/j.cedpsych.2008.07.001
• Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. *Psychological Bulletin, 132*(3), 354–380. https://doi.org/10.1037/0033-2909.132.3.354
• Cepeda, N. J., Vul, E., Rohrer, D., Wixted, J. T., & Pashler, H. (2008). Spacing effects in learning: A temporal ridgeline of optimal retention. *Psychological Science, 19*(11), 1095–1102. https://doi.org/10.1111/j.1467-9280.2008.02209.x
• Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham, D. T. (2013). Improving students' learning with effective learning techniques. *Psychological Science in the Public Interest, 14*(1), 4–58. https://doi.org/10.1177/1529100612453266
• Ebbinghaus, H. (1885). *Über das Gedächtnis.* Leipzig: Duncker & Humblot. English reprint: *Annals of Neurosciences* (2013). https://doi.org/10.5214/ans.0972.7531.200408
• Karpicke, J. D., & Roediger, H. L., III. (2008). The critical importance of retrieval for learning. *Science, 319*(5865), 966–968. https://doi.org/10.1126/science.1152408
• Kim, S. K., & Webb, S. (2022). The effects of spaced practice on second language learning: A meta-analysis. *Language Learning, 72*(1), 269–319. https://doi.org/10.1111/lang.12479
• Latimier, A., Peyre, H., & Ramus, F. (2021). A meta-analytic review of the benefit of spacing out retrieval practice episodes on retention. *Educational Psychology Review, 33*(3), 959–987. https://doi.org/10.1007/s10648-020-09572-8
• Murre, J. M. J., & Dros, J. (2015). Replication and analysis of Ebbinghaus' forgetting curve. *PLOS ONE, 10*(7), e0120644. https://doi.org/10.1371/journal.pone.0120644
• Nakata, T. (2015). Effects of expanding and equal spacing on second language vocabulary learning. *Studies in Second Language Acquisition, 37*(4), 677–711. https://doi.org/10.1017/S0272263114000825
• Nakata, T. (2017). Does repeated practice make perfect? *Studies in Second Language Acquisition, 39*(4), 653–679. https://doi.org/10.1017/S0272263116000280
• Pashler, H., Cepeda, N. J., Wixted, J. T., & Rohrer, D. (2005). When does feedback facilitate learning of words? *Journal of Experimental Psychology: Learning, Memory, and Cognition, 31*(1), 3–8. https://doi.org/10.1037/0278-7393.31.1.3
• Roediger, H. L., III, & Karpicke, J. D. (2006). Test-enhanced learning: Taking memory tests improves long-term retention. *Psychological Science, 17*(3), 249–255. https://doi.org/10.1111/j.1467-9280.2006.01693.x
• Tomaru anonymized learning data (2026).