Practice Tests Beat Re-Reading: What 50 Classroom Experiments Show
Pulling an answer out of memory teaches more than reading it again. A classic experiment, a review of 50 classroom studies, what they leave open, and how to use them tomorrow.

In this post
Ask a class how they revise and most will say the same thing: they read their notes again, maybe with a highlighter. It feels productive, because the material looks more familiar each time. Researchers who study memory have spent decades testing a different habit, called retrieval practice: closing the notes and trying to recall the answers. This post walks through one experiment that made the idea famous, a review of fifty experiments run in real schools and universities, the questions that are still open, and a routine you can use next lesson.
The experiment that started the conversation
In 2006 Henry Roediger and Jeffrey Karpicke gave students short prose passages to learn. Some students read a passage and then read it again. Others read it once and then wrote down everything they could remember, with no feedback on what they got right. A final test came 5 minutes, 2 days or a week later.
Five minutes after studying, the re-readers did better. That is the result most students would predict, and it is why re-reading feels like the safer choice. But on the tests two days and a week later the picture flipped: the students who had practised recalling remembered substantially more. Repeated studying also left students more confident that they would remember, which is exactly the wrong lesson to take away.
Two details make this study useful for teachers. First, nobody was told which of their answers were right, so the benefit came from the act of recalling itself. Second, the trap is built into how learning feels. A method that wins after five minutes and loses after a week will always look better in the moment.
From the lab to fifty real classrooms
Laboratory findings do not always survive a noisy classroom, so in 2021 Pooja Agarwal, Ludmila Nunes and Janell Blunt went looking for evidence from real lessons only. They screened nearly two thousand abstracts and kept fifty experiments in which students practised retrieval individually and in person, as part of their normal course. Studies run in a lab, online or in groups were left out on purpose, to keep the picture clean.
| What they looked at | What they found |
|---|---|
| Size of the evidence | 50 experiments, 49 effect sizes, 5,374 students |
| How big the benefit was | 57% of the effects were medium or large |
| Results against practice | Only 3 of the 49 favoured lessons without it |
| Who took part | 20 experiments in schools, 20 at university, 10 in medical schools |
| When the final test came | From one day later to the end of the semester |
The benefit held across subjects, ages and test delays. The largest effects came from middle-school classrooms, though the authors point out that all of those studies were run in a single school district. University studies showed the smallest effects, but they also had the largest samples.
A second, wider review backs this up. Olusola Adesope, Dominic Trevisan and Narayankripa Sundararajan pooled experiments from the lab and the classroom and concluded that practice tests helped learning more than restudying and more than every other comparison they examined, with the size of the gain depending on how the tests were given and to whom.
Three details that change the result
- Show the answer afterwards. Most of the classroom experiments gave immediate feedback. The ten that gave none produced mostly small or very small effects. Recalling helps on its own, as Roediger and Karpicke showed, but in a real class the correct answer straight after is what turns a guess into learning.
- Match the kind of question to the exam. When students practised with multiple choice and were then tested with short written answers, the effects were smaller. If the exam asks students to write, some of the practice should too.
- Keep it low-stakes. The authors note that these classroom activities were rarely even called tests. They were quizzes, warm-ups and recall prompts, and the point was practice, not a mark.
What the evidence does not tell us yet
Every review has edges, and these are worth knowing before you quote the research to a colleague.
Almost all of the classroom studies come from a narrow slice of the world. Only 6% of the experiments were run outside what researchers call WEIRD countries (Western, educated, industrialised, rich and democratic), so we know far less about classrooms elsewhere.
How much practice is enough is still unclear. Many studies mixed up how often students practised, when they did it and whether the same questions came back, so the review cannot say whether a weekly quiz beats a daily one.
Transfer is the open question. Remembering the answer to a question you have practised is well supported; applying the idea to a new kind of problem is harder to show, and the authors call for more research on exactly that.
A ten-minute routine for tomorrow
Retrieval practice in a normal lesson
- Open with four or five questions on last week’s topic, not today’s, so that students have to dig for the answers rather than repeat what they heard ten minutes ago.
- Everyone answers alone, with books and notes closed.
- Reveal the correct answer straight after each question, with a sentence on why.
- If the exam needs written answers, make one or two of the questions typed rather than multiple choice.
- Note the question most of the class missed, and bring it back in a week.
It is worth telling students why you do this. The Roediger and Karpicke result is easy to explain: re-reading makes material feel familiar, and familiar feels like learned. A short demonstration helps. Give the class a list to learn, have half of them re-read and half of them test themselves, and check a few days later. Students who see the difference in their own scores are more likely to change how they revise at home.
If you already start lessons with a quiz, you are most of the way there. The research suggests three small changes: make sure the right answers are shown, reach back further than yesterday, and let some answers be written rather than chosen. For a related habit, asking questions before a topic is taught, see pre-testing.
Sources
- Roediger, H. L. & Karpicke, J. D. (2006). Test-enhanced learning: Taking memory tests improves long-term retention. Psychological Science, 17(3), 249–255.
- Agarwal, P. K., Nunes, L. D. & Blunt, J. R. (2021). Retrieval practice consistently benefits student learning: A systematic review of applied research in schools and classrooms. Educational Psychology Review, 33(4), 1409–1453.
- Adesope, O. O., Trevisan, D. A. & Sundararajan, N. (2017). Rethinking the use of tests: A meta-analysis of practice testing. Review of Educational Research, 87(3), 659–701.
Try it in Giga Quizzes
Free, in the browser, and nothing for your students to install. These guides show each step on the real screens.
New to Giga Quizzes?
Free live quizzes in the browser. Students join from any phone, with no app.
Make your first quiz →

