Spaced repetition in corporate training: what the research says

What is spaced repetition?

Spaced repetition means bringing material back several times, with gaps between the sessions, instead of covering it once. It is usually paired with retrieval practice: each time the material returns, the learner tries to recall it, for example by answering a question, rather than reading it again. Research treats these as two separate effects that work well together.

The idea is old. In his 1885 book on memory, Hermann Ebbinghaus reported experiments on himself, learning series of nonsense syllables and measuring the effort needed to relearn them. In one comparison, 38 repetitions spread over the three previous days worked as well as 68 repetitions made the day before. He concluded that, for any considerable number of repetitions, spreading them over time is “decidedly more advantageous” than massing them at once.

What does the research say about spacing?

Spacing works, but the best gap depends on how long you need to remember. A 2006 meta-analysis in Psychological Bulletin by Cepeda and colleagues pooled 839 assessments from 317 experiments in 184 articles. It found that the gap between study sessions that gave the best retention grew as the delay before the final test grew.

A large follow-up study made this concrete. In 2008 the same group taught more than 1,350 people a set of obscure facts online, gave them a review after a gap of up to 3.5 months, and tested them up to a year later. Lengthening the gap first helped, then slowly hurt. The best gap was about 20 to 40% of the test delay when the test came a week later, and about 5 to 10% when it came a year later. For something needed in a year, that works out to a review several weeks after first learning, not a refresher the next day.

Why does answering questions beat rereading?

Recalling something strengthens memory more than reading it again, at least over days rather than minutes. This is called the testing effect, or retrieval practice. In a 2006 study in Psychological Science, Roediger and Karpicke had university students study short prose passages, then either restudy them or write down what they could recall, with no feedback.

Rereading looked better at first: after five minutes, students who restudied recalled 81% against 75% for those who had been tested. Two days later the order flipped, 68% against 54%, and after a week it was 56% against 42%. In a second experiment, students who studied once and took three recall tests remembered 61% after a week, against 40% for those who studied four times. Yet the students who only reread were the most confident they would remember.

A 2013 review of ten common study techniques by Dunlosky and colleagues, in Psychological Science in the Public Interest, gave its highest utility rating to just two: practice testing and distributed practice. Feedback matters too. Butler and Roediger found in 2008 that showing the right answer after a multiple-choice question, either straight away or later, raised correct answers on a later test and reduced the wrong options that learners repeated as if they were true.

How strong is the evidence for workplace training?

Weaker than many vendors suggest. Most spacing research uses students, word lists or trivia, and short delays. In the 2006 meta-analysis, 80% of the comparisons tested memory less than a day later, and only 4% waited more than a month. Studies of employees learning job knowledge over several months are rare.

An older meta-analysis in the Journal of Applied Psychology, by Donovan and Radosevich in 1999, looked at task performance across 63 studies. Spaced practice beat massed practice on average, with a mean effect size of 0.46, but the benefit was smaller for more complex tasks, and less rigorous studies reported larger effects. Spacing in that work mostly meant rest breaks within practice, not reviews weeks apart.

Classroom evidence is growing. A 2025 meta-analysis by Mawson and Kang found 22 classroom reports with 31 effect sizes and a moderate benefit for distributed over massed practice (d = 0.54). It covered schools and universities, though, and none of the studies tested memory more than 42 days later.

One of the better workplace examples comes from medicine. In a randomized trial published in 2010, Kerfoot and colleagues sent 95 primary care clinicians at eight US Veterans Affairs medical centers short case questions by email, with immediate feedback, in four cycles over 36 weeks. A control group received nothing. In the 72 weeks after the program, 7.8% of the prostate cancer screening tests the spaced group ordered were inappropriate, against 13.1% in the control group: a 40% relative reduction. It is one profession and one topic, so read it as encouraging, not as proof for your setting.

Is there a best review schedule?

No single schedule is proven best. The research agrees that the right gap depends on how long the knowledge must last, and that one short gap is rarely enough for knowledge people need months later. It does not settle the exact intervals, or whether each gap should be longer than the last.

Expanding schedules, where each gap is longer than the one before, are common in apps and training tools. The 2006 meta-analysis found they did better than fixed gaps on average, but called the conclusion tentative because studies disagreed: one researcher found expanding gaps better, another found fixed gaps better, and a third found no difference. Treat any schedule, including ours below, as a sensible design choice rather than a proven optimum.

How do you design spaced review for staff training?

Keep each review short, make learners recall rather than reread, show the right answer afterwards, and spread reviews across the period in which people need the knowledge. Then check whether it worked with a test taken weeks later, not straight after the course.

• Write short retrieval questions that each test one thing: a fact, a step, or a decision in a short scenario
• Ask learners to produce or choose the answer before they see it; rereading a summary is not retrieval
• Give feedback every time: the right answer and a line on why, especially after multiple-choice questions
• Spread reviews over the weeks or months the knowledge must last, with longer gaps when it is needed further ahead
• Keep each session to a few minutes so it fits a working day
• Focus on what people must recall without looking it up, such as safety steps, product facts and policy rules
• Measure retention some weeks after the course, and if you can, compare a group that reviews with one that does not

How does Classbase handle spaced review?

Classbase brings questions back in a short daily review. Questions a learner answers in a course, in lessons or in checks, go into review. A right answer moves a question to a longer gap, and a wrong one sends it back to the start. The gaps are 1, 3, 7, 16 and 35 days.

Those numbers are our design choice, not a research-proven optimum. We chose a simple box system (Leitner boxes) over more complex algorithms because a learner can understand it, and a few minutes a day doesn’t need more. By default the review asks up to five questions a day, about three minutes, shows the right answer and why after each one, and has no points, streaks or timers. Each answer also counts toward the learner’s mastery of the skills it is tagged with, and that score fades if a skill goes unpractised. Review can be turned off for the whole organization or for one course.

Whether it improves retention for your people is worth measuring, and we’re happy to plan that with you in a demo.