Turning the Sheet Over: Retrieval Practice in Primary School

A correct fact in a pupil’s notebook does not tell us how it arrived there. It may have been copied from the sheet beside the notebook, or reconstructed while that sheet was out of sight. The finished sentence can look identical even though the learning activity was different. Retrieval practice makes this difference useful: asking for remembered information need not be reserved for the examination at the end of a course. It can become part of studying. That does not mean every lesson must become a high-pressure contest over marks.

In two classroom experiments published in 2013, Stuart J. Ritchie, Sergio Della Sala and Robert D. McIntosh investigated a particularly simple version of this idea. Children read short geographical factsheets. Some kept the material visible while making notes; others turned it over and worked from memory. A second comparison concerned whether the notes took the form of a mind map. The design separated the source of the information during note-taking from the arrangement of that information on the page, and also allowed the two techniques to be combined. [1]

What changed during the lesson

The first experiment involved 109 pupils aged eight to twelve at Towerbank Primary School in Edinburgh. Factsheets concerned Senegal, South Korea, Peru or Iran, countries chosen as likely to be unfamiliar to young pupils in the United Kingdom. Younger participating classes received eleven facts, while the older classes received sixteen. This was a deliberately manageable fact-learning task, not an attempt to teach an entire geography curriculum. The later questions could be tied to specific information supplied in the initial material.

Everyone began with five minutes of reading without writing. The next five minutes were for notes. In the retrieval condition, children turned their factsheet over and wrote on its blank reverse. In the other condition, they could still see the facts and wrote on a separate sheet. All pupils then read the original material again for three minutes and spent another five minutes making notes under their assigned condition. The session therefore lasted eighteen minutes for everyone. Children practising retrieval could see the source material for only eight of those minutes.

The equal session length matters. An advantage for the retrieval group could not simply be attributed to being given more scheduled study time. Nevertheless, an equal number of minutes does not establish an equal amount of mental engagement. One pupil may finish copying early, while another continues searching for missing information. The researchers considered such a possibility rather than treating the classroom clock as a complete account of what happened. The comparison is between activities carried out within that time, not merely between two durations of exposure.

A map was not the opposite of remembering

In the mind-mapping condition, children put the country’s name in the centre of the page and arranged facts along outward connections. In the other condition they could take notes however they wished, provided they did not use a mind map. The schools already used this uncomplicated form of mapping. The experiment was consequently not evaluating a lengthy training programme in a newly introduced representational technique. Mapping conditions were arranged between classes, whereas retrieval conditions were established within classes. The two interventions did not have identical levels of assignment.

A pupil could make a mind map from memory. Another could make one while looking at the factsheet. This is why the question was more informative than a simple choice between testing and drawing a map. If mapping helped regardless of access to the text, that would be one result; if it worked only in combination with retrieval, that would be another. An attractive or orderly page was not itself the outcome being scored. The central outcome was whether its owner could later supply the facts when asked.

Learning took place on Monday, followed by a written recall test on Friday. The experimenter read the questions aloud and allowed time for responses. Correct information did not lose credit merely because of a spelling error, and partly correct answers could receive half marks. The person scoring the tests did not know each child’s learning condition during scoring. These procedures helped keep the outcome focused on remembered content rather than penmanship, spelling accuracy or expectations about which method ought to win. They also make the test’s meaning more concrete than the loose description “school performance.”

A result that needed a second classroom

In the researchers’ preferred analysis of the first experiment, children who practised retrieval remembered more facts after four days. There was an important statistical qualification: the analysis took account of how much information children had recorded during the learning session. Removing that variable made the retrieval difference no longer statistically significant. Reporting only the favourable analysis would conceal a genuine sensitivity in the evidence. Moreover, the quantity of recorded facts was measured during learning, not in a separate assessment conducted before the intervention began.

The second experiment began with 209 pupils, again aged eight to twelve, at Bruntsfield Primary School in Edinburgh. Its task was made more challenging by increasing the number of facts at some school levels, reducing the session to sixteen minutes and extending the assessment schedule. Children were tested after one week and after five weeks. Twenty-three were unavailable at one or the other test, leaving 186 for the analysis of fact learning. The number enrolled at the beginning and the number contributing to that analysis should not be presented as interchangeable.

Here the retrieval advantage remained whether the additional variables were included or excluded. The second experiment therefore did more than supply another positive headline: it tested the earlier result under a different timetable, in another school and with a larger starting sample. Mind mapping did not show a similarly consistent advantage. In the first experiment it appeared beneficial in the non-retrieval condition, but that interaction did not recur in the second. This finding concerns the simple maps and factual recall studied here. It does not establish that every form of mapping, including elaborate representations of conceptual relationships, is educationally useless. [1]

The activity worked; its explanation remained open

The design cannot completely isolate why retrieval helped. Looking at the source again after trying to remember could alert a child to an omitted fact or help repair an error. The authors explicitly acknowledge that a direct effect of retrieval and an indirect benefit to subsequent studying were not fully separated. Rereading was part of the session for every group, rather than a technique the experiment prohibited. A sensible interpretation is not that books should remain shut throughout learning. It is that continuous access to the text is not the only productive way to use a study period.

The five-week test also requires care. These same facts had already been tested after one week, and that earlier test might itself have provided retrieval practice to both groups. The later result is not a measurement of memories left completely untouched for five weeks. Nor does recalling the requested facts automatically demonstrate skill in explaining a country’s political history, solving a novel geographical problem or comparing conflicting sources. Those outcomes would require their own assessments. Keeping the target skill specific helps a teacher decide where the technique belongs within a wider lesson.

One classroom implication is straightforward: after a short encounter with material, give pupils an opportunity to discover what they can bring back without it, then return to accurate information. This is a teaching suggestion drawn from the experiments, not a universally optimal minute-by-minute prescription or a guaranteed gain for each child. An incomplete response can also help a teacher locate the explanation that is still needed. The evidence-based contribution of these studies is that allocating some learning time to retrieval improved delayed recall of the studied facts in these primary-school samples. Turning over the page need not signal that learning has finished; it can start another part of the learning process.

Sources

[1] Ritchie, S. J., Della Sala, S., and McIntosh, R. D. (2013). Retrieval Practice, with or without Mind Mapping, Boosts Fact Learning in Primary School Children. PLOS ONE, 8(11), e78976. Primary research; methods, results of both experiments and general discussion directly examined. https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0078976

The source is published under a Creative Commons Attribution licence. This article is an independently written explanation, not a reproduction or translation of the source prose. Source access and factual review: 14 October 2026.