How adults learn: practice, testing and spacing

Testing as a way of learning

Roediger and Karpicke (2006) asked students to study prose passages and then either take one or three free-recall tests without feedback, or restudy the passage the same number of times. A final retention test followed after 5 minutes, 2 days or 1 week. After 5 minutes, repeated studying produced better recall than repeated testing. On the delayed tests, however, earlier testing produced substantially greater retention, even though restudying had made students more confident in their memory. The authors concluded that testing is a powerful means of improving learning, not just assessing it.^[1]

Karpicke and Roediger (2008) taught students foreign-language vocabulary under four conditions. Once an item had been recalled correctly, it was either restudied without further testing, tested without further study, or dropped from both. Repeated studying after success had no effect on delayed recall, while repeated testing had a large positive effect. Students' predictions of their own performance were uncorrelated with how they actually did. These experiments used foreign-language vocabulary, so they show the mechanism clearly but say little about complex job tasks.^[2]

What the later syntheses add

Later work pooled many experiments. Rowland (2014) meta-analysed studies comparing testing with restudy and reported that initial recall tests produced larger testing benefits than recognition tests, which the paper reads as support for the role of effortful processing. Adesope and colleagues (2017) found practice tests more beneficial for learning than restudying and all other comparison conditions, and noted that the size of the effect depended on features of the tests, the participants, the outcomes and the study methods. The summary abstracts do not give a single headline figure, so none is quoted here.^[3][4]

Dunlosky and colleagues (2013) reviewed ten learning techniques and rated how widely each one generalises across learning conditions, learner characteristics, materials and outcome measures. Practice testing and distributed practice received the highest utility ratings because they benefit learners of different ages and abilities. Rereading and highlighting were rated low, because they do not consistently boost performance, and the authors suggested practice testing in place of rereading. Elaborative interrogation, self-explanation and interleaved practice were rated moderate because the evidence for them was still limited.^[5]

Spacing practice over time

Cepeda and colleagues (2006) synthesised 839 assessments of distributed practice drawn from 317 experiments in 184 articles. They compared massed study with spaced study, compared shorter and longer gaps, and looked at gaps that expand over time. Their analyses suggest that the gap between study sessions and the length of time before the final test operate jointly. The gap that produced the best retention became longer as the interval before the final test became longer.^[6]

In practice this means there is no single ideal spacing. A gap that suits a test next week may not suit knowledge that must last for months. The review concerned verbal recall tasks, so its evidence is about remembering words and facts rather than complex skills. The later monograph by Dunlosky and colleagues still rated distributed practice as high utility because it benefits learners of different ages and abilities across many criterion tasks, which is one reason spacing sits beside testing in most practical summaries of the research.^[6][5]

Worked examples and feedback

A meta-analysis by Barbieri and colleagues (2023) screened 8,033 abstracts and kept 43 articles, reporting 55 studies and 181 effect sizes. The average effect of worked examples on mathematics performance was medium, with g = 0.48. Studies using correct examples alone showed larger effects than those using incorrect examples alone or in combination with correct ones. Adding self-explanation prompts to examples moderated the effect negatively. The outcomes were mathematics performance, from elementary grades to postsecondary settings, so the finding should not be assumed to hold for every subject.^[7]

Hattie and Timperley (2007) described feedback as one of the major influences on learning but stressed that the type of feedback and the way it is given can be differentially effective. Wisniewski, Zierer and Hattie (2020) later meta-analysed 435 studies with 994 effects and more than 61,000 participants and found a medium average effect (d = 0.48). They also found significant heterogeneity, meaning feedback is not a single consistent treatment. Impact depended on the information the feedback carried, and was higher for cognitive and motor skills than for motivational and behavioural outcomes.^[8][9]

Limits and cautions

Several cautions follow from the papers themselves. The first testing study found that restudying beat testing when the final test came after only five minutes, so the benefit appears when retention is measured later. Students in the same study felt more confident after restudying even though they remembered less, which means how learners feel about a method is not a reliable guide to how well it works. Effect sizes also differ across test formats, populations and outcomes, so a single number should be read as an average.^[1][2][4]

Most of the studies cited here used university students, school-age learners or laboratory tasks. The abstracts do not report trials with employees on job-related training, so applying the results to workplace learning is an inference rather than an established finding. The Dunlosky review also notes that some promising techniques had not yet been adequately evaluated in educational settings. A sensible reading is that the principles are well supported for recall and comprehension, and that their effect on on-the-job performance needs separate evidence.^[5][7]

Questions