The weekly quiz at school: low-stakes and cumulative

A weekly quiz earns its ten minutes when it does two things an end-of-unit test cannot. It asks about last month as well as this week, and nothing rides on the score. Without the first, each topic is learned for one Friday. Without the second, the quiz is a small exam.

This guide is for a teacher of pupils aged 11 to 16; a lecturer wants the university version. EdProcto, which publishes it, is a quiz maker for teachers, and its Free plan covers a weekly quiz for one class.

Sources were read on 10 October 2026: Bailey et al. (2021) and the Association for Psychological Science (2022) in full, Harlen and Deakin Crick (2002) in its published summary, and the other studies as abstracts ("abstract read").

What does cumulative mean in a weekly quiz?

Cumulative means that every quiz asks about earlier weeks as well as this one, in a fixed share. For ten questions that is six, two and two.

The split is our working rule, not a research finding.

A returning question is the old idea in new words: a new number, a new example, or the question turned round, so that pupils recall the idea and not the sentence.

One week's ten questions

An invented Year 8 maths class in week 7, with percentages taught this week.

• New. Find 15% of 80 (12). Increase 60 by 25% (75). Write 0.35 as a percentage (35%). Write 18 out of 24 as a percentage (75%). Which is larger, 30% of 50 or 50% of 30 (neither: each is 15)? A coat costs £40 after 20% off: what was the full price (£50)?
• Last month. Share £35 in the ratio 2:5 (£10 and £25). Work out 2/3 + 1/4 (11/12).
• Last term. Work out −7 − (−3) (−4). Find the area of a triangle with base 10 cm and height 6 cm (30 cm²).

What did research find on cumulative testing and interleaving?

Cumulative quizzes beat quizzes on the latest material alone in two of the three direct comparisons we found. In schools the firmer evidence is for the two things a returning question does: it spaces practice and it mixes topics.

Cumulative against non-cumulative testing

Two of the three come from universities and the third from an online course whose abstract gives no level. Three searches of ERIC on the day of writing found no such comparison in a secondary classroom.

• Weekly vocabulary quizzes. Nakata et al. (2021; abstract read) assigned 72 Japanese university students at random to two kinds of weekly quiz while they learned 80 English words, ten a class: on the previous class's ten alone, or on earlier words as well. By the authors' estimate the cumulative quizzes were about twice as effective on the receptive final test and three times on the productive one.
• Weekly exams, half and half. Gayman et al. (2023; abstract read) gave one section of an online course weekly exams on that week alone, and another weekly exams with half the questions from earlier weeks. On a shared final the second section averaged 4.91% higher and earned more As and Bs.
• Where it made no difference. Bailey et al. (2021) compared two sections of a university biology course, each sitting seven tests of 12 questions. In one, about half of every test came from earlier units. The shared final exam showed no difference, and the authors suggest cumulative tests may add nothing there when testing is already frequent. On a retention test five months later, taken by 62 of more than 300 students, cumulative tests had raised scores by about 15% for the lowest reasoners and lowered them by about 15% for the highest. The authors treat that with caution.

Spacing and interleaving in schools

• Spaced practice in Year 7. Emeny et al. (2021; abstract read) ran two classroom experiments with pupils aged 11 and 12, who did their maths practice in one session or over three sessions a week apart. A month later the spaced pupils scored higher both times. They had also predicted their scores fairly well, while the others were far too confident.
• Interleaving in weekly science quizzes. Sana and Yan (2022; abstract read) gave 155 pupils in grades 9 to 12 a weekly science quiz for four weeks, on half of each week's concepts, with the questions grouped by concept or mixed. A month after the last quiz the pupils scored 47% on concepts never quizzed, 54% on those quizzed in blocks and 63% on those quizzed mixed. The Association for Psychological Science (2022) adds that the mixed quizzes scored lower on the day, 55% against 59%, and took ten to twelve minutes of class a week.

Our guide to interleaving has the trials in school mathematics. No study here tested six, two and two across a term. They support its parts: bring a topic back after a gap, mix it with others, and expect lower scores on the day.

How do you build ten questions in fifteen minutes?

Start from last week's quiz, write six questions, reword four, and use only shapes a rule can mark. The fifteen minutes are our estimate, not a measured figure.

1. Copy last week's quiz (1 minute). In the EdProcto builder, "Duplicate this quiz" makes a draft with the same questions, settings and kind, a new code and no submissions. Rename it.
2. Clear it (1 minute). "Remove" the questions that have done their work. Keep any that most of the class missed: reworded, they return this week.
3. Write six new questions (7 minutes). Three that ask for a fact or a term, two that use it, and one aimed at a mistake you saw in the books. Multiple choice and fill in the blank are marked by rule. Or paste the week's notes into "Write with Procto" and read every draft against them.
4. Reword four old ones (4 minutes). Two from the last month, two from last term.
5. Label and order (2 minutes). Type the unit under "Topic (for the results by topic)" and set "Difficulty" to "Easy", "Medium" or "Hard". Press "Move up" and "Move down" until no two neighbours share a unit.

Keep a one-line note of each week's most-missed questions as you read the results. Step 4 is then copying from a list.

How do you keep a weekly quiz low-stakes?

Decide before the first quiz what its score will never be used for, tell the class, and keep to it. A quiz is low-stakes when a bad week costs a pupil nothing that lasts.

How results are said

Say results about questions, not about pupils. Three sentences cover most weeks:

• To the class: "Most of us missed question 7. It comes back next week."
• To one pupil: "You missed both of these three weeks ago, and today you got them."
• To a parent: "The weekly quiz is practice. It is not on the report."

The evidence is about marks in general. Koenka et al. (2021; abstract read) pooled studies from primary and secondary schools: pupils given grades achieved more than pupils given no feedback but were less motivated, and against pupils given comments they came out worse on both. Harlen and Deakin Crick (2002), reviewing 19 studies of summative tests, advise: "Teachers should avoid comparisons between students based on test results." So no scores read out and no chart on the wall.

What does a term's timetable look like?

Twelve quizzes, one a week in the same slot, with more returning questions before each break. The table is for a twelve-week term with a break after week 6. "Older" means last term, last year or, late in a term, its first weeks.

Over the term that is 120 questions, 62 new and 58 returning.

What do you do about absence, and the pupil who always scores low?

Treat both the same way: the next quiz is the remedy, so nobody sits an old quiz in class and nobody is measured against the room. The rules below are our judgement.

Absence

A missed quiz is not made up in lesson time. Its most-missed questions return within the month, which is what cumulative is for.

• One week missed. Nothing to do. An online quiz can stay open until the next lesson for practice at home.
• Several weeks missed. The pupil restarts with this week's quiz. Say that the returning questions will feel new, and why.

For a test that does count, see retakes and make-ups.

The pupil who always scores low

First look at where the misses fall. On the new questions, the week's teaching did not land. On the returning ones, it landed and faded, which more practice mends.

• Compare the pupil with their own last month, never with the class. Harlen and Deakin Crick (2002) report that after national tests came into England's schools, low-achieving pupils had lower self-esteem than high achievers, a link absent before.
• Keep the returning questions. In Bailey et al. (2021) the lowest reasoners retained more after cumulative tests, though they were adults and few.
• Lower the barrier, not the content. Ask an idea as multiple choice the first time it returns and as a typed answer the next. How often and how hard has the success rates to aim for.
• Offer a second go at home. Duplicate the quiz and set the copy's kind to "Study quiz": "Each answer shown at once, misses retried."

What do you stop doing when the routine starts to slip?

Stop the extras in a fixed order, and keep the slot and the returning questions to the last. A routine is slipping when the quiz takes more than fifteen minutes to make or ten to run.

1. Stop writing ten fresh questions. Four are old ones reworded.
2. Stop marking by hand. Every question goes in a shape a rule marks.
3. Stop going through all ten answers. Pupils see each answer at hand-in. Teach the two the class missed most.
4. Stop recording scores, if any have crept into the markbook.
5. Stop at six questions before you skip a week: three new, two from the last month, one older.

Do not save time by cutting the returning questions: a quiz of new questions alone is the non-cumulative quiz of the comparisons above. The step from no quiz to a weekly one is the large one, as frequent low-stakes quizzes reports, so protect the week before the length.

How does EdProcto run a weekly quiz, and which plan does it need?

EdProcto is the best tool for a cumulative quiz, because each week's quiz starts from the one before: one press copies its questions, settings and kind, and a rule marks every objective answer. Of the 120 questions in the term above, 58 are old ones reworded, and all can be in shapes a rule marks.

The kind. In the builder, set "Kind of quiz" to "Exam" for the class's sitting. The line beside the selector then reads "Handed in once, marked at the end." The word names the kind, not the stakes.

The settings. The "Practice" preset reads "No time limit, going back allowed, answers shown at once, no security." In an exam, "At once" is the hand-in. Practice also switches shuffling off, so every pupil meets the questions in the order you built.

The results. Every plan has the results question by question. Results by topic are part of the full analysis, on the Educator plan.

The plan. Free makes ten quizzes a month, for thirty students a quiz.

On Free each full class needs its own copy, because a quiz takes thirty students. "Duplicate this quiz" makes it, and every copy is one of the month's ten, as is a quiz later deleted. On Tutor, three classes of thirty can share one quiz and one code. Procto AI writes 20 questions a month on Free and 50 on Tutor; six new questions a week come to 24 to 30.

Quick questions

Should a weekly quiz count towards a grade or a report?

Not for pupils aged 11 to 16. Record who sat it and which questions the class missed, and keep the scores off reports.

Is interleaving the same as a cumulative quiz?

No. Cumulative is which weeks the questions come from, and interleaving is the order they sit in. In Sana and Yan (2022), mixed weekly quizzes led to 63% a month later, against 54% for grouped ones.

What do I do when a pupil misses the weekly quiz?

Nothing in lesson time. The questions most of the class missed return within the month.

Which EdProcto plan does a weekly quiz for several classes need?

Two classes fit Free's ten quizzes a month with a copy each. Three or more need Tutor, which has no limit on quizzes: start at EdProcto.

References

• Association for Psychological Science. (2022, June 2). Mix it up: Testing students on unrelated concepts can help jump-start learning. APS Observer. https://www.psychologicalscience.org/observer/mix-it-up-testing-students
• Bailey, E. G., Greenall, R. F., Tullis, M. M., & Williams, K. R. (2021). The retention benefits of cumulative versus non-cumulative midterms in introductory biology may depend on students' reasoning skills. PLOS ONE, 16(4), Article e0250143. https://doi.org/10.1371/journal.pone.0250143
• Emeny, W. G., Hartwig, M. K., & Rohrer, D. (2021). Spaced mathematics practice improves test scores and reduces overconfidence. Applied Cognitive Psychology, 35(4), 1082–1089. https://doi.org/10.1002/acp.3814
• Gayman, C. M., Jimenez, S. T., Hammock, S., Taylor, S., & Rocheleau, J. M. (2023). The effects of cumulative and noncumulative exams within the context of interteaching. Journal of Behavioral Education, 32(2), 261–276. https://doi.org/10.1007/s10864-021-09451-4
• Harlen, W., & Deakin Crick, R. (2002). A systematic review of the impact of summative assessment and tests on students' motivation for learning. EPPI Centre, Institute of Education, University of London. https://eppi.ioe.ac.uk/cms/publications/systematic-reviews/summative-assessment-and-motivation
• Koenka, A. C., Linnenbrink-Garcia, L., Moshontz, H., Atkinson, K. M., Sanchez, C. E., & Cooper, H. (2021). A meta-analysis on the impact of grades and comments on academic motivation and achievement: A case for written feedback. Educational Psychology, 41(7), 922–947. https://doi.org/10.1080/01443410.2019.1659939
• Nakata, T., Tada, S., McLean, S., & Kim, Y. A. (2021). Effects of distributed retrieval practice over a semester: Cumulative tests as a way to facilitate second language vocabulary learning. TESOL Quarterly, 55(1), 248–270. https://doi.org/10.1002/tesq.596
• Sana, F., & Yan, V. X. (2022). Interleaving retrieval practice promotes science learning. Psychological Science, 33(5), 782–788. https://doi.org/10.1177/09567976211057507