The evidence question
Can a Word Become Easier to Remember Simply Because You Were Tested on It?
A vocabulary quiz can do more than measure memory. Research on retrieval practice shows when trying to recall a word can strengthen the memory itself.
A question worth asking.
08 sources, open to exploreIn this article 13 sections
A quiz usually feels like the thing that comes after learning.
First you study.
Then someone tests whether the learning worked.
But memory research has repeatedly found something stranger:
the test can become part of the learning.
Trying to retrieve a word can change how well you remember it later, even when the test adds no new definition, example sentence, or explanation.
So when a flashcard asks you for a word you studied yesterday, it may be doing two jobs at once.
It is measuring the memory.
And it may be modifying it.
Yes. Successfully retrieving a word can make that word more likely to be retrieved again later.
This is usually called the testing effect or retrieval-practice effect. Across memory research, tests that require retrieval often produce better delayed retention than spending the same learning phase simply restudying material.
Foreign-language vocabulary is one of the classic places where the effect has been demonstrated.
But “testing improves memory” needs qualifications. Retrieval has to be possible often enough to succeed, mistakes need a route to correction, repeated retrieval has diminishing time-efficiency, and the exact kind of test influences what is learned.
The useful principle is not:
quiz yourself constantly.
It is:
after a memory has been established, make it retrieve information instead of only showing it the information again.
A test is not just a measurement
Suppose you study:
scarce → not enough; limited
Five minutes later you have two options.
Restudy
scarce → not enough; limited
Retrieval
scarce → ?
Both expose you to the cue scarce.
Only one requires your memory to reconstruct the target information before it is shown.
That reconstruction is important.
A test therefore differs from rereading even when both activities concern exactly the same word.
The surprising claim of retrieval-practice research is that successfully bringing information back can alter later accessibility more than another exposure to the same information.
Foreign-language vocabulary produced one of the clearest demonstrations
Karpicke and Roediger's 2008 experiments used foreign-language vocabulary pairs and manipulated what happened after learners first recalled an item correctly.
Some words continued to be studied.
Some continued to be tested.
Some were dropped from study or testing after the first correct response.
On a delayed test one week later, repeated studying after an item had already been recalled produced essentially no additional benefit, while repeated retrieval produced a large retention advantage.
That result is memorable because the extra study trials felt like additional learning opportunities.
The extra tests were the events that mattered most for delayed recall.
The effect is much bigger than one vocabulary experiment
Retrieval practice is one of the most extensively studied learning phenomena in cognitive psychology.
A major meta-analysis by Rowland compared testing with restudying across the experimental literature and found an overall retention advantage for testing. The benefit depended on conditions: tests requiring recall tended to produce larger benefits than recognition tests, and the relationship between the practice test and the eventual final test mattered.
That is useful for language learners because “test yourself” can describe very different activities.
Recognizing scarce among four definitions is not identical to producing its meaning with no alternatives visible.
Producing scarce from a meaning cue is another retrieval problem again.
All can be legitimate tests.
They do not necessarily strengthen exactly the same retrieval route.
Why a flashcard can become easier after you struggle with it
Imagine this sequence:
Day 1: scarce → “limited”
Day 3: scarce → ... → “limited”
Day 8: scarce → “limited”
The Day 3 attempt feels worse than simply rereading the pair.
There is a pause.
You are uncertain.
The answer takes effort to reconstruct.
Yet the important comparison is not how fluent the practice feels.
It is what happens when the cue returns later.
Retrieval practice repeatedly forces the cue to reach the target without the target already being visible. That makes it much closer to the future problem you care about: encountering scarce and getting its meaning from memory.
This is also why recognition and recall are not the same vocabulary skill.
A practice method can become extremely easy because the answer is always present.
Ease during study is not proof of accessibility later.
In L2 vocabulary, more retrieval can help — but time changes the calculation
Tatsuya Nakata tested 98 Japanese learners studying English–Japanese word pairs with different numbers of retrievals within a learning session.
When practice amount was compared directly, five and seven retrievals produced higher later scores than one and three retrievals.
But there was a catch.
More retrievals also consume more time. When learning efficiency was evaluated with time-on-task controlled, a single retrieval produced the greatest gain per unit of time.
This is a useful correction to simplistic advice.
If five successful recalls are better than one, it does not follow that you should spend five times as long testing every word today.
Learning happens under a time budget.
80 words across nine weeks: cumulative testing changed retention
Retrieval practice is not limited to tightly controlled laboratory sessions.
Nakata, Tada, McLean and Kim studied 72 Japanese university students learning 80 low-frequency English words across nine weekly classes.
Both groups received vocabulary quizzes. The key difference was whether quizzes tested only the newest material or also brought older vocabulary back.
The cumulative-testing group performed substantially better on later vocabulary tests; the authors reported cumulative tests could be roughly twice as effective for receptive knowledge and three times as effective for productive knowledge in that study.
The practical mechanism is important: cumulative quizzes repeatedly create retrieval opportunities for material that would otherwise disappear from active practice.
Testing and spacing solve different problems
Retrieval practice is often bundled together with spaced repetition.
They are related, but they are not the same variable.
Retrieval asks:
What does the learner have to do when the word appears?
Spacing asks:
When does the word appear again?
You can retrieve repeatedly in one ten-minute session.
You can also space repeated rereading without ever requiring retrieval.
A strong vocabulary schedule often combines both: reconstruct the memory, then make it survive again after time has passed.
This distinction also explains why the testing effect article is different from Should You Check the Answer Immediately After Testing Yourself — or Wait?. That article asks what should happen after the attempt. Here, the important event is the attempt itself.
Myth: a harder quiz automatically creates stronger memory
Myth: If retrieval is good, the hardest possible test must be best.
Reality: A retrieval attempt that almost never succeeds can become a poor learning environment.
Retrieval practice needs an existing memory to retrieve. For a completely unknown word, initial encoding still matters. If the learner repeatedly produces nothing — or repeatedly produces the wrong answer without useful correction — difficulty alone does not create the desired memory.
Research on foreign-language vocabulary also shows that retrieval practice can depend on how many retrieval opportunities learners receive and whether feedback is available. In experiments where retrieval was very limited, retrieval alone has not always beaten restudy; combining retrieval with effective encoding or corrective information can matter.
For feedback specifically, see the separate VocaNum article on when to check the answer after testing yourself.
“The testing effect” is not one perfectly uniform phenomenon
There is a newer methodological complication worth knowing.
A 2026 meta-analysis of free-recall research argues that studies commonly grouped under “the testing effect” can mix at least two consequences of testing:
- a direct testing effect, where testing improves later memory for the material that was tested;
- a forward testing effect, where taking a test changes learning of material encountered afterward.
The authors found that many prior studies did not cleanly isolate those two contributions.
That does not erase the large literature showing that retrieval can support retention. It does mean that a single label can hide multiple mechanisms and experimental designs.
For a vocabulary learner, the safest practical claim remains narrower:
requiring previously learned words to be retrieved is often more valuable for delayed accessibility than repeatedly exposing yourself to answers that are already visible.
What should you do with a word after you first learn it?
Do not replace learning with testing.
Create something worth retrieving first.
Understand the core meaning. Hear or read a useful example. Make sure the form–meaning relationship is not a guess.
Then remove the answer.
Try to reconstruct it.
If you fail, get the correct information rather than rehearsing an error. If you succeed, do not assume the word is permanently learned. Let time pass and ask memory to solve the problem again.
The useful cycle is:
encode → retrieve → correct → wait → retrieve again
The test is not the finish line.
It is one of the training events.
A quiz can change the memory it appears to measure
That is the counterintuitive idea behind retrieval practice.
If you only reread scarce = limited, the answer is continuously supplied by the environment.
When you see scarce = ?, the environment stops doing part of the work.
Memory has to produce something.
Across general memory research and foreign-language vocabulary studies, repeatedly creating that retrieval event can improve later retention. But retrieval is not magic difficulty: learners still need initial encoding, useful feedback when necessary, sensible spacing, and a practice schedule that respects limited time.
So the next time a vocabulary test feels like an interruption to learning, remember:
if it makes you genuinely retrieve what you previously learned, the test may be part of the learning itself.
Sources and further reading
- Karpicke & Roediger (2008) — The Critical Importance of Retrieval for Learning
- Karpicke & Roediger (2007) — Repeated retrieval during learning is the key to long-term retention
- Rowland (2014) — The effect of testing versus restudy on retention: A meta-analytic review of the testing effect
- Nakata (2017) — Does Repeated Practice Make Perfect? The Effects of Within-Session Repeated Retrieval on Second Language Vocabulary Learning
- Nakata et al. (2021) — Effects of Distributed Retrieval Practice Over a Semester
- Fritz et al. (2007) — Comparing and combining retrieval practice and the keyword mnemonic for foreign vocabulary learning
- Huang et al. (2024) — The facilitative effect of the keyword mnemonic on L2 vocabulary retrieval practice
- Gereau & Mulligan (2026) — Research on “the testing effect” routinely conflates direct and forward testing effects