Skip to content

Memory science

Retrieval Practice and Spacing for Vocabulary: New Studies

Two 2025 studies test spacing and retrieval practice on second-language words. See what they found, what they cannot prove and how to use both.

By iranimij Published 6 min read

Two 2025 studies of second-language vocabulary support the same advice: test yourself on words, check the answer, and review them again after a growing gap. One replicated a classic spacing experiment online and found spaced study beat massed study. The other found that a plan combining retrieval practice and spacing produced substantially better learning and retention than typical study.

This post explains what each study did, what it can and cannot tell you, and how to turn it into a daily routine.

What are retrieval practice and spacing?

Retrieval practice means trying to recall something from memory; spacing means spreading those attempts over time. Both are among the best-supported study methods in memory research, and both feel harder than rereading, which is part of why they work.

Take the Spanish word ventana (window). Looking at a card that shows ventana = window is restudy. Covering the answer, trying to remember what ventana means and then checking is retrieval. Doing that on Monday, then again on Thursday, then the following week is spacing.

If you want the background, see spaced repetition vs cramming for a century of evidence on timing and how to remember vocabulary for nine techniques, including testing yourself. The new studies add recent second-language data to that picture.

What did Rogers, Nakata and Chiu (2025) find about spacing online?

Spaced study beat massed study again, this time in an online replication. John Rogers, Tatsuya Nakata and Ming Chiu published a conceptual replication of Experiment 1 from Cepeda and colleagues (2009) in Studies in Second Language Acquisition in 2025 (Rogers et al., 2025). A conceptual replication repeats the design of an earlier study with a new group of people or setting, to check whether the result holds up.

Cepeda and colleagues (2009) had tested second-language vocabulary learning with two study sessions followed by a final test 10 days later. The replication kept that structure and ran it online. The main result matched the original: the spaced conditions significantly outperformed the massed condition.

One difference matters for learners. The online participants scored 10 to 20 percent lower than the participants in the original study in every condition. The authors point to the challenges of running experiments on the web, such as less controlled conditions. The benefit of spacing survived the noisier setting, which is reassuring for anyone studying on a phone on the bus.

What did Karatas and colleagues (2025) test?

They compared typical study with a plan that combined retrieval practice, feedback and spaced, expanding gaps. The study by Nur Basak Karatas, Michael Ullman and colleagues appeared in Language Teaching Research in 2025 (Karatas et al., 2025). It followed 48 Turkish university students who were learning English as a second language. They trained on 64 English words over 17 days.

One group followed a typical pattern, restudying the words on the first and last days. The other followed an optimized regimen of retrieval practice with feedback, spread over the period with gaps that expanded. The authors report that the optimized regimen substantially improved both learning and retention of the words, and that the advantage was still there more than a week after training ended.

The contrast is easy to picture. The typical group met each word twice, at the start and at the end. The optimized group met each word several times in between, each time trying to recall it first and then seeing the answer. That is more work per session, but the work is concentrated on the moments when memory is just starting to fade.

An expanding schedule means the gap between reviews gets longer each time you get a word right: for example 1 day, then 3, then 7. It is the idea behind box systems like the one in our guide to the Leitner system.

Where does Radicard fit in?

If you want retrieval practice and spacing without managing a schedule, an app can do the bookkeeping. I built Radicard for this reason. You photograph any text, such as a street sign, a café menu or a book page, and Google’s Gemini AI picks out 5 to 15 vocabulary words matched to your CEFR level, with translations.

The words go into an on-device library and are reviewed with the Leitner spaced-repetition system: seven boxes with review gaps of 1, 2, 4, 8, 16, 31 and 120 days. Everything except scanning a new photo works offline. The free plan includes 5 scans per day, and you can read more about how it works or open the web app in any browser. A paper box and a pen give you the same mechanism, so the method matters more than the tool.

What do these studies not show?

They do not show which ingredient does the work, and neither gives you a magic schedule. Being honest about the limits makes the findings more useful, not less.

  • The Karatas study tested a package. By the authors’ own account, the design did not test spacing alone, retrieval practice alone, expanding versus equal gaps, or retrieval with and without feedback. You cannot tell from it which part produced the gain.
  • It is one group of university students. The sample was 48 people learning English in one setting. Results for adult learners of Japanese or Arabic on their own may differ.
  • The replication used two sessions. The Cepeda design tests the idea of spacing, not a full long-term review plan.
  • Lab words are not your words. Both studies used word lists chosen by researchers. Words from a menu or a conversation you care about may stick differently.

For the underlying idea that the best gap depends on how long you want to remember, see the review by Cepeda and colleagues (2006) and the later Cepeda et al. (2009) study the replication was based on.

How do you combine retrieval practice and spacing in your own study?

Quiz yourself on each word, check the answer straight away, and move words you got right to a longer gap. Here is a plan that mirrors what the studies tested without needing special software.

  1. Learn a small batch. Ten new words is plenty for one sitting. Read each with its translation and one example sentence.
  2. Retrieve before you look. Cover the answer and try to say the meaning, or the word from the meaning, out loud.
  3. Check and correct. Feedback matters. If you were wrong, read the correct answer and say it once.
  4. Space the reviews. Right answers go to a longer gap, for example 1, 2, 4 and 8 days. Wrong answers go back to tomorrow.
  5. Stop the session while it still feels slightly hard. Easy repetition feels good but teaches less.

What should you do next?

Pick one change you can make today.

  1. Turn one list of words you are learning into cards where you must recall the answer, not just read it.
  2. Set up two or three review gaps, such as tomorrow, in four days and in two weeks.
  3. Write the date of the next review on each batch of cards, or let an app do it.
  4. Keep your sessions short, ten to fifteen minutes, and do them most days.
  5. After two weeks, check how many of your old words you can still recall without a prompt. That number is your own evidence.

Frequently asked questions

Is retrieval practice better than rereading for vocabulary?

Retrieval practice means trying to recall a word before you look at the answer. In the 2025 study by Karatas and colleagues, learners who combined retrieval practice with spacing did substantially better than learners who mostly restudied. Because the two techniques were combined, the study cannot show how much of the gain came from retrieval alone.

What is the difference between spacing and retrieval practice?

Spacing is about timing: you spread reviews of a word over days or weeks instead of one sitting. Retrieval practice is about method: you try to recall the word from memory instead of just looking at it again. Flashcard systems such as the Leitner box use both at once.

Does spaced repetition still work if you study online and on your own?

Evidence says yes. Rogers, Nakata and Chiu (2025) replicated a classic spacing experiment with online participants and again found that spaced study beat massed study. Their participants scored lower overall than in the original lab study, so expect noisier conditions to cost some accuracy, not the benefit itself.

How long should I wait between reviews?

The studies here do not give one number that suits everyone. A practical rule from earlier research is that the right gap grows with how long you want to remember the word. The Leitner system is one simple version, with gaps that grow from 1 day to 120 days across seven boxes.