- Title: A Meta-Analysis of the Effectiveness and Boundary Conditions of Retrieval Practice, Distributed Practice, and Interleaved Practice in Authentic School Classrooms
- Authors: Sterre K. Ruitenburg et al.
- Access the original paper here
- Watch a video overview:
Paper summary
Retrieval practice, spaced practice and interleaving are widely recommended to schools, but most of the evidence behind them comes from labs and universities. This meta-analysis combined 53 studies, involving 7,303 students aged 6 to 18, carried out in real primary and secondary classrooms using curriculum content. All three strategies improved learning, with average effect sizes of 0.37 for retrieval practice, 0.20 for spacing and 0.40 for interleaving. These are smaller than lab-based reviews have found, but large by the standards of school interventions. The researchers also examined when each strategy worked. Retrieval practice helped most with factual knowledge. Spacing worked best when the gap between sessions was 10–100% of the time until the test. Interleaving helped when feedback was delayed or absent, but not when students got answers immediately.
If teachers remember one thing from this study, it should be…
Retrieval practice, spacing and interleaving all improved learning in real classrooms, but how you use them matters. Retrieval practice worked best for factual knowledge, spacing worked best when gaps suited the time until the test, and interleaving worked when students didn’t get instant answers.
*** PAPER DEEP DIVE ***
What are the key technical terms used in the paper?
- Meta-analysis: combining results from many studies.
- Effect size: how big a difference is (0.2 small, 0.5 medium).
- Retrieval practice: recalling information from memory rather than rereading it.
- Spaced (distributed) practice: spreading practice over sessions rather than cramming it into one.
- Interleaving: mixing problem types rather than practising one type in a block.
What are the characteristics of the participants in the study?
7,303 students aged 6 to 18 in regular primary and secondary schools, across 53 studies (71 experiments) published 2006–2024. Most studies came from the USA (16), the Netherlands (10) and Germany (6), with two from the UK. Studies used real curriculum content and compared each strategy against an active alternative.
What does this paper add to the current field of research?
Earlier meta-analyses found bigger effects, but mostly from labs and universities, and some found weaker results in schools. This is the first to pool only real primary and secondary classroom studies, with strict quality rules, and the first to test feedback timing as a factor in interleaving.
What are the key implications for teachers in the classroom?
- Keep using all three strategies, knowing the gains in real classrooms are smaller than lab studies suggest.By the standards of school interventions, the benefits were still large. Build them into your regular routines, for example a short low-stakes quiz at the start of lessons, homework that revisits topics from earlier in the year, and practice sets that mix different problem types.
- Use retrieval practice to secure the facts students need, and build understanding first. Retrieval practice reliably helped students remember factual knowledge such as vocabulary, definitions and key facts. It did not reliably help with understanding concepts, carrying out procedures or applying knowledge in new situations, although the authors stress this pattern is tentative. Their explanation is that recalling information strengthens memory of material students already understand, so the understanding has to be built beforehand. So when you teach a new idea, make sure students grasp it through explanation and tasks where they explain links or come up with their own examples, then use quizzing to keep the key facts that underpin it secure.
- Match your practice questions to how students will be tested. Retrieval practice worked better when the practice questions and the final test used the same format, for example short-answer practice before a short-answer test. If students will need to write a definition from memory in the exam, quizzing them only with multiple-choice questions may not prepare them well. When you plan a quiz, look at the end assessment first and mirror its question style.
- Space practice out, and set the gap according to when students need the knowledge. Spreading practice across sessions clearly beat doing it all in one sitting (an effect size of 0.39), though spreading it even more thinly gave no clear extra benefit. The gaps that worked best were between a tenth of, and the whole of, the time until the test. So if an end-of-unit test is four weeks away, leave somewhere between three days and four weeks between practice sessions on a topic, instead of practising it on consecutive days. The authors offer this range tentatively, but it is a useful starting point for planning homework and starter quizzes.
- When you mix problem types, hold the answers back until the end. This is the finding most likely to surprise teachers. Interleaving improved learning when students got no feedback (0.25) or feedback after the practice session (0.82, though based on only five results), but it made no difference when students saw the answer after each question. The authors suggest that instant answers remove the need to work out which method each problem calls for, which is the skill interleaving is meant to build. Mixed practice also feels harder, so seeing lots of wrong answers straight away may put students off trying. Set a mixed problem set, go through the answers together at the end of the lesson or the start of the next one, and consider switching off instant marking on online platforms for these tasks.
Why might teachers exercise caution before applying these findings in their classroom?
These are averages across very varied studies. Many findings about when each strategy works rest on few studies and overlap, so the authors call their practical suggestions tentative. The interleaving evidence is thin (ten studies, mostly maths), and only two of the 53 studies came from the UK.
What is a single quote that summarises the key findings from the paper?
“Observed moderator patterns suggest that the testing, spacing, and interleaving effects only occur reliably if their underlying cognitive processes are elicited during practice.”
