Sit down to revise for almost any subject and you face the same choice: should you test yourself with multiple-choice questions, or with free-recall — open-ended prompts where you have to produce the answer from a blank page? The intuitive answer is that free recall must be harder, and therefore better. The intuition is half right. Free recall is generally harder, and difficulty matters, but the picture from a half-century of testing-effect research is genuinely more interesting than “harder is better.” Some carefully designed multiple-choice formats outperform free recall on long-term retention; some don’t. This article walks through what the research actually shows, and how to set up self-tests that turn quizzing into learning rather than just measurement.

The testing effect, briefly

Before getting to the format comparison: the underlying finding the whole conversation rests on. In experiments published in Psychological Science in 2006, Henry Roediger and Jeffrey Karpicke had college students study short prose passages and then either take free-recall tests on them or restudy them the same number of times. On a final test five minutes later, restudying did slightly better. On final tests two days and a week later the order reversed, and the students who had been tested remembered substantially more — even though restudying had made students more confident. Retrieval, not rereading, was what laid down durable memory.

The effect is one of the most heavily replicated findings in the cognitive psychology of learning. Christopher Rowland’s 2014 meta-analysis in Psychological Bulletin, pooling comparisons of testing and restudying across many studies, found a reliable overall benefit of testing. The size of the effect varies from study to study; its direction is consistent. Practising recall outperforms practising re-exposure.

The follow-on question is the one this article addresses: what kind of test?

What multiple choice gets you

The objection to multiple-choice testing is well-known. You can hit the right answer without producing it from memory — recognition, the argument goes, doesn’t exercise the same retrieval pathway as recall, and so doesn’t strengthen memory in the same way. Worse, exposure to wrong answers (the “distractors”) might actively interfere with later recall by planting incorrect associations.

The research turns out to be more nuanced. Marsh, Roediger, Bjork and Bjork (2007) found that taking a multiple-choice test did produce a testing benefit on a later test — and also that students sometimes carried a wrong option they had picked into their later answers. Kang, McDermott and Roediger (2007) compared multiple-choice and short-answer practice tests directly. Without feedback, multiple choice came out ahead, largely because students got more of those questions right the first time. With feedback, short-answer practice produced the best later recall. The “wrong-answer interference” worry is real, and how much it matters depends heavily on feedback.

Butler and Roediger (2008), in Memory & Cognition, tested that directly. Students studied passages and then took a multiple-choice test with immediate feedback, delayed feedback, or none. Compared with no feedback, both kinds of feedback increased correct answers on a later recall test and reduced how often students repeated the wrong options. If you take a multiple-choice quiz and never find out which answers were right, the risk of learning a wrong answer is at its highest. If you see the correct answer afterwards, most of that risk goes away and the testing benefit dominates.

That feedback finding is one of the most practically useful things in the entire literature, and it’s almost always the difference between self-testing that works and self-testing that doesn’t.

What free recall gets you

Free recall — given a prompt, produce the answer with no scaffolding — is the harder test format, and “desirable difficulty” (a term coined by Robert Bjork in the 1990s) is one of the better-supported principles in modern learning research. Tasks that feel harder during study, within limits, tend to produce stronger long-term retention.

The mechanism is reasonably well understood. Free recall requires you to construct the retrieval pathway yourself, without the safety net of recognising the answer in a list. Each successful retrieval strengthens that pathway. Each unsuccessful retrieval — even when it ends in “I don’t know” — produces some learning when feedback follows (the so-called “pretesting” effect, replicated by Kornell and colleagues across multiple studies in the late 2000s and early 2010s).

The size of the free-recall advantage over multiple choice grows with the retention interval — the gap between study and test. On an immediate test, the two formats are often comparable. At a one-week interval, free recall typically pulls ahead. At a one-month interval, the gap widens further. If you’re studying for the day after tomorrow, format matters less. If you want the material to be available six months from now, the format gap becomes substantial.

When multiple choice wins

The most surprising and consistent finding in this area: well-designed multiple-choice questions can produce free-recall-comparable retention if they meet a few criteria.

The distractors are plausible. A four-option question with three obviously-wrong distractors is functionally a free-recall question (you knew the answer without seeing the options). A four-option question with three plausible distractors forces real discrimination and exercises a genuinely retrieval-adjacent process.

Feedback is given. As Butler and Roediger found, feedback is doing a lot of the work: seeing the correct answer after a multiple-choice question removes most of the risk of learning a wrong option, whether the feedback comes straight away or a little later.

The test happens more than once, spaced. A single multiple-choice test on Monday is a much weaker intervention than three multiple-choice tests on Monday, Wednesday, and the following Monday. The spacing effect (Cepeda et al., 2008) compounds with the testing effect.

Pyc, Agarwal and Roediger’s 2014 chapter for teachers, in the American Psychological Association’s Applying Science of Learning in Education, summarises how to use test-enhanced learning in the classroom. The practical point that runs through this whole literature is that the bigger gap tends to be between students who test themselves at all and students who don’t, rather than between one test format and another.

A practical guide to designing self-tests that work

If you’re putting together your own quizzes — whether for revision, language learning, or knowledge maintenance — the consolidated guidance from this research breaks down to a small number of rules.

  1. Test yourself rather than rereading. This is the headline finding and it’s worth restating. The lowest-effort study habit that visibly fails the research is rereading. The highest-effort habit that consistently wins is testing.

  2. Use a mix of formats. Free recall for the material you most want to retain long-term; multiple choice for breadth coverage and warm-up. A typical weekly revision routine might be 20 minutes of free-recall practice on the most important topics, followed by 20 minutes of multiple-choice coverage across a broader set.

  3. Get feedback every time. Without feedback, multiple-choice testing risks installing wrong answers; with feedback, the risk largely vanishes. Spend the extra moment to read the correct answer and the explanation before moving on.

  4. Space the tests. A 2008 study by Cepeda, Vul, Rohrer, Wixted and Pashler found that the best gap between study sessions depends on how long you want the material to last: the longer you need to remember it, the longer the gaps should be. A practical version is to come back after a day, then a few days, then a week or two. Cramming is the opposite of this, and a poor way to keep anything beyond the next test.

  5. Make distractors plausible. If you’re writing your own multiple-choice questions, the test of a good distractor is whether someone who half-knows the topic could reasonably pick it. Implausible distractors give you false confidence without the underlying learning.

  6. Test before you’ve finished studying. Self-testing on material you only partly know — including answering “I don’t know” — produces gains even before feedback. This is the pretesting effect, and it’s one of the better-replicated findings of the last fifteen years.

  7. Don’t conflate “feels easy” with “is learning.” Bjork’s desirable-difficulty principle is uncomfortable to apply. Rereading feels productive because the material flows past you smoothly. Testing yourself feels punishing because you’re constantly hitting the limits of what you actually know. The discomfort is the point. The smoother study session is usually the less effective one.

What this means for using a quiz site

If you’re a casual user of an online quiz platform — playing rounds for entertainment, with the side benefit of learning — none of this is critical. You will still pick up information, particularly over time, simply by playing.

If you’re using one as a deliberate study tool, the implications are more pointed. The most effective use pattern is:

  • Play a round on a topic you want to know better.
  • Read the feedback for each question — particularly the wrong ones.
  • Return to the same topic later in the week and play another round.
  • Test yourself across the topic from memory, without the quiz, every few weeks.

That last step is the one most quiz players skip, and it’s the one that converts “answering a quiz question” into “knowing the material.” Free recall, even brief and informal — “what do I remember about the geography of South America” — is the strongest single learning intervention available to you, and you don’t need any tool to do it.

For a longer reading on the underlying habits, our evidence-based study guide and improving general knowledge pieces sit alongside this one. If you want to put the principles to immediate use, pick a category from the main quiz list and play a round; the feedback on each question is built specifically for the testing-effect mechanism this article describes.

Sources and further reading

  • Roediger, H. L., & Karpicke, J. D. (2006). Test-enhanced learning: taking memory tests improves long-term retention. Psychological Science, 17(3), 249–255.
  • Rowland, C. A. (2014). The effect of testing versus restudy on retention: a meta-analytic review of the testing effect. Psychological Bulletin, 140(6), 1432–1463.
  • Butler, A. C., & Roediger, H. L. (2008). Feedback enhances the positive effects and reduces the negative effects of multiple-choice testing. Memory & Cognition, 36(3), 604–616.
  • Kang, S. H. K., McDermott, K. B., & Roediger, H. L. (2007). Test format and corrective feedback modify the effect of testing on long-term retention. European Journal of Cognitive Psychology, 19(4–5), 528–558.
  • Marsh, E. J., Roediger, H. L., Bjork, R. A., & Bjork, E. L. (2007). The memorial consequences of multiple-choice testing. Psychonomic Bulletin & Review, 14(2), 194–199.
  • Kornell, N., Hays, M. J., & Bjork, R. A. (2009). Unsuccessful retrieval attempts enhance subsequent learning. Journal of Experimental Psychology: Learning, Memory, and Cognition, 35(4), 989–998.
  • 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.
  • Pyc, M. A., Agarwal, P. K., & Roediger, H. L. (2014). Test-enhanced learning. In V. A. Benassi, C. E. Overson, & C. M. Hakala (Eds.), Applying Science of Learning in Education. American Psychological Association.
  • Bjork, R. A. (1994). Memory and metamemory considerations in the training of human beings. In J. Metcalfe & A. Shimamura (Eds.), Metacognition: Knowing about knowing. MIT Press.