The pretesting effect: a quiz before you teach

A pretest asks students about material they have not been taught, knowing most answers will be wrong. It looks like five wasted minutes. The pretesting effect is the finding that it is not: wrong guesses, followed by the right answers, leave those answers better remembered than study alone.

The effect is real, moderate and narrow: it helps the questions asked and almost nothing else.

EdProcto, which publishes this guide, runs a pretest as a short quiz made on its home page and shows, before you teach, which questions the class could already answer.

What is the pretesting effect, and how does a pretest differ from a pre-test?

The pretesting effect is better memory for an answer because the student tried to produce it before being taught it. Richland et al. (2009) named it. In five experiments students read an essay about vision; some answered questions on it first, and others had longer to read. Those questioned first did better afterwards, although the key ideas were in italics or bold for both groups, and in the fifth experiment the others saw the questions without answering them. The attempt matters, not only knowing what to look for.

Three quizzes share the name and differ in what the student is shown afterwards. A pretest to teach gives the answers the same day. A pre-test to measure shows nothing until its second sitting. Retrieval practice comes days after the teaching.

We read Geller et al. (2017), Latimier et al. (2019) and Pan and Rivers (2023) in full on 10 October 2026, the review by Pan and Carpenter (2023) in part, and every other study in abstract only, de Lima and Jaeger (2020) through an index.

What have experiments found about a quiz before teaching?

Laboratory experiments find a moderate gain on the questions asked. Classroom studies find less, and sometimes nothing.

The questions asked, and the rest of the lesson

St. Hilaire et al. (2024) pooled the experiments in a preregistered meta-analysis. Over 97 effects, material asked about beforehand was remembered better by about half a standard deviation (g = 0.54); over 91, the rest gained almost nothing (0.04). King-Shepard et al. (2025), in a multilevel meta-analysis, put the two at 0.66 and 0.01.

So on average a pretest neither helps nor harms the rest of the lesson. Single studies go both ways: Pan and Carpenter (2023) list video studies in which untested content gained, and Latimier et al. (2019) suspect that a reader who has seen the questions reads for those answers at the expense of the rest.

Lecture halls and classrooms

Three course studies gave three results.

• Carpenter et al. (2018) randomly assigned introductory psychology students to prequestions before each lesson or none. At the end of class the prequestion group remembered more, but only of what had been asked; a week later the prequestions had added nothing to the benefit of an end-of-class quiz.
• Geller et al. (2017) asked 62 chemical engineering students one multiple-choice question at the start of each of 14 lessons. At the end of class they did no better on it than on a fresh question, in a study with no comparison group.
• Soderstrom and Bjork (2023) gave a large research methods class a brief multiple-choice pretest straight before some lectures and not others. On the final exam, pretested and related material was answered better than control questions, by 8% to 9% in Pan and Carpenter's (2023) account.

Feedback, and how soon

The answers must follow, and how soon depends on the material. King-Shepard et al. (2025) found that prequestions with feedback taught more than prequestions alone. With word pairs, Grimaldi and Karpicke (2012) found in three experiments that a guess helped only when the words were related and the right answer came at once: a brief delay removed the gain. With trivia questions that mean something, Kornell (2014) found the gain survived a delay of up to 24 hours. In the three course studies no answer followed the pretest: the lesson had to supply it.

Children

Children learn from a wrong guess from about the age of five. Carneiro et al. (2018) had children guess a word related to one the experimenter said, then hear the right one: second- and third-graders and kindergarteners recalled more after guessing, and preschoolers did not. In two experiments with fourth- and fifth-graders reading an encyclopaedia text, de Lima and Jaeger (2020) found that questions before reading and questions after it both beat rereading a week later.

Confidence, and students' forecasts

Students do not notice the benefit. In the five experiments of Pan and Rivers (2023), 70% to 82% of the adults recalled guessed word pairs better than read ones in the first round, yet their forecasts gave guessing no advantage until they were shown their own scores for each method. Geller et al. (2017) found the opposite risk: at the end of class students were more confident on the repeated question than on the fresh one, and no more accurate.

Wrong answers can also discourage: in an online course described by Pan and Carpenter (2023), a study we did not read ourselves, learners given pretests were likelier to drop out.

Against a quiz afterwards

Head to head, the quiz afterwards has the better record. Latimier et al. (2019) had 285 adults study an online course on DNA with a quiz before each text, a quiz after it, or a second reading. A week later the quiz afterwards led (d = 0.74 over rereading) and the quiz before came second (d = 0.35). de Lima and Jaeger (2020) found the same order on recall in children.

One set of experiments found the reverse: see studying a textbook chapter by chapter. Pan and Carpenter (2023) conclude that a test before instruction can often improve learning, by an amount that varies with the procedure.

When is a pretest worth the minutes, and when is it not?

A pretest is worth five minutes when the lesson turns on a few facts with one answer each, the students know enough to guess, and the answers arrive the same day.

Facts, yes. Hausman and Rhodes (2018) compared factual and conceptual pretest questions across three experiments. Factual pretests improved later answers to the same questions. Conceptual pretests did not improve conceptual learning, and wrong conceptual answers were likelier to be repeated at the final test.

Skills, unproven. We read no study of a pretest on a skill such as writing a paragraph. Geller et al. (2017) suggest prequestions suit a brief stretch of fairly simple material better than a 50-minute lesson on harder ideas.

No footing, no gain. A guess teaches when it lands near the answer: Grimaldi and Karpicke's unrelated word pairs gained nothing, and Overoye et al. (2021), in two experiments on invented topics, found a larger pretesting effect after a short paragraph of context. For a class with no footing, give two sentences of context, then ask.

One quiz only. If there is time for one quiz, make it the one after teaching, as the school experiment in our retrieval practice guide found.

How do you write pretest questions?

Write questions a student could reason towards, on the points the lesson will state plainly.

1. Point at what the lesson will stress. The gain falls on what is asked, so ask the lesson's few most important facts.
2. Make each answerable in principle. A thoughtful student should be able to offer a sensible guess.
3. One answer, which the lesson says aloud. A guess the lesson never corrects is an error left standing.
4. About five questions in five minutes, as a working rule.
5. Multiple choice with plausible wrong options, or one word. Both are marked at hand-in.
6. Keep the questions for the end of the lesson and next week's review.

How do you run a pretest so that being wrong costs nothing?

Say that wrong answers are expected, show no score, record no grade, and give the answers the same day.

"I have not taught this yet, so I expect wrong answers. Guess anyway: a guess makes the answer easier to remember when we reach it. Nothing here counts."

• No score on the screen, no mark in the book. A low-stakes pretest that prints 20% is no longer low stakes.
• Read it by question, never by name.
• Flag each answer as it comes: "That was question 4."
• Ask the five again as the lesson closes. Learners believed in the benefit only after seeing their own scores (Pan & Rivers, 2023).
• Not every lesson. Once or twice a week is our working rule.

Which pretest fits your purpose?

The right quiz depends on what follows it; the table is our recommendation.

A worked example: one lesson and a five-question pretest

A Year 8 lesson on the rock cycle opens with five questions, answered on phones as the class comes in.

1. Which type of rock forms when molten rock cools and hardens? (Igneous)
2. What is molten rock called while it is still underground? (Magma)
3. Limestone is which type of rock? (Sedimentary)
4. Which rock does limestone become under great heat and pressure? (Marble)
5. In which type of rock are fossils most often found? (Sedimentary)

Suppose 28 students hand in and the shares right, in invented figures, are 21%, 75%, 36%, 7% and 43%.

• Before teaching. Question 2 is known, so the recap on magma goes. Question 4 is the gap, so marble gets a sample on the desk.
• In the last three minutes. The same five again, read as an exit ticket.
• Next week. The five return, reworded, in the review quiz.
• Everything else. Weathering and erosion were not on the pretest, gained nothing from it, and are taught as carefully as ever.

How does EdProcto run a pretest?

EdProcto is the best tool for a pretest because it marks the answers by rule at hand-in: each question's share right is in front of you before you teach, and since the gain falls only on what is asked, that reading decides the lesson.

1. Write the five. The quiz maker on the home page drafts questions from your lesson notes for you to edit. Type each answer's reason in "Explanation shown to the student afterwards".
2. Choose the kind. The builder's selector explains its three: "Exam: handed in once, marked at the end. Study quiz: each answer shown at once, misses retried. Quiz night: played live, with a scoreboard." Before a lesson, choose "Exam". For reading done alone, choose "Study quiz": the student presses "Check" after each answer and reads "Right." or "Not this time.", with "Answer:" and the right one after a miss, and your explanation either way. A study quiz is practice: its start screen gives "Attempts" as "No limit".
3. Set what is shown. The "Practice" preset removes the time limit and the security and puts the three "Feedback" settings at "At once". Set "Show the correct answer", "Show the explanation" and "Show the score" to "Never" (the other choices are "After submitting" and, for the score, "When released"): a student who hands in reads "This quiz does not show a score." In a study quiz, "Check" still shows each answer. Switch off "An account is needed to take it" and the start screen asks only for "Your name".
4. Read it before you teach. On every plan the results show, under "Each question", the share who "Got it", with the cautions, headed "Read this first" below Educator: "Fewer than five submissions: the numbers below describe these students, not the class or the teaching yet." On Educator a question opens to each option's "Chose it", and the analysis adds sentences such as "Most of the class missed Q4. Check the key first, then whether this was taught." On a pretest that is the expected reading: it marks what to stress.
5. Give the answers. After the lesson, set the first two to "After submitting". Students open the link again on the same device, press "See your result" and read "Answer:" and "Explanation:" under each question, still with no score.

Free takes thirty students a quiz, Tutor 100, and Educator 500 with the full analysis: see pricing and the pages for teachers and for tutors. A pretest in EdProcto records what a class guessed before a lesson, never a grade.

Quick questions

Is a pretest the same as a pre-test for measuring progress?

No. A pretest to teach gives the answers the same day. A pre-test to measure shows nothing until the post-test: see the measuring pair.

Should a pretest count towards a grade?

No. The answers are expected to be wrong, and the benefit comes from the attempt. In EdProcto, "Show the score" at "Never" keeps the score off the student's screen.

Should students see the answers straight after a pretest?

For meaningful questions the lesson can be the answer: Kornell (2014) found the gain survived a 24-hour delay. For paired facts such as vocabulary, show it at once. Quiz feedback covers timing after teaching.

Does pretesting work with young children?

From about five, on the evidence so far: Carneiro et al. (2018) found a benefit for kindergarten and early primary children and none for preschoolers.

References

• Carneiro, P., Lapa, A., & Finn, B. (2018). The effect of unsuccessful retrieval on children's subsequent learning. Journal of Experimental Child Psychology, 166, 400–420. https://doi.org/10.1016/j.jecp.2017.09.010
• Carpenter, S. K., Rahman, S., & Perkins, K. (2018). The effects of prequestions on classroom learning. Journal of Experimental Psychology: Applied, 24(1), 34–42. https://doi.org/10.1037/xap0000145
• de Lima, N. K., & Jaeger, A. (2020). The effects of prequestions versus postquestions on memory retention in children. Journal of Applied Research in Memory and Cognition, 9(4), 555–563. https://doi.org/10.1016/j.jarmac.2020.08.005
• Geller, J., Carpenter, S. K., Lamm, M. H., Rahman, S., Armstrong, P. I., & Coffman, C. R. (2017). Prequestions do not enhance the benefits of retrieval in a STEM classroom. Cognitive Research: Principles and Implications, 2(1), Article 42. https://doi.org/10.1186/s41235-017-0078-z
• Grimaldi, P. J., & Karpicke, J. D. (2012). When and why do retrieval attempts enhance subsequent encoding? Memory & Cognition, 40(4), 505–513. https://doi.org/10.3758/s13421-011-0174-0
• Hausman, H., & Rhodes, M. G. (2018). When pretesting fails to enhance learning concepts from reading texts. Journal of Experimental Psychology: Applied, 24(3), 331–346. https://doi.org/10.1037/xap0000160
• King-Shepard, Q. W., Walker, J., Nokes-Malach, T. J., Carpenter, S. K., & Fraundorf, S. H. (2025). The effect of prequestions on learning: A multilevel meta-analysis. Educational Psychology Review, 37(4), Article 115. https://doi.org/10.1007/s10648-025-10075-7
• Kornell, N. (2014). Attempting to answer a meaningful question enhances subsequent learning even when feedback is delayed. Journal of Experimental Psychology: Learning, Memory, and Cognition, 40(1), 106–114. https://doi.org/10.1037/a0033699
• Latimier, A., Riegert, A., Peyre, H., Ly, S. T., Casati, R., & Ramus, F. (2019). Does pre-testing promote better retention than post-testing? npj Science of Learning, 4, Article 15. https://doi.org/10.1038/s41539-019-0053-1
• Overoye, A. L., James, K. K., & Storm, B. C. (2021). A little can go a long way: Giving learners some context can enhance the benefits of pretesting. Memory, 29(9), 1206–1215. https://doi.org/10.1080/09658211.2021.1974048
• Pan, S. C., & Carpenter, S. K. (2023). Prequestioning and pretesting effects: A review of empirical research, theoretical perspectives, and implications for educational practice. Educational Psychology Review, 35(4), Article 97. https://doi.org/10.1007/s10648-023-09814-5
• Pan, S. C., & Rivers, M. L. (2023). Metacognitive awareness of the pretesting effect improves with self-regulation support. Memory & Cognition, 51(6), 1461–1480. https://doi.org/10.3758/s13421-022-01392-1
• Richland, L. E., Kornell, N., & Kao, L. S. (2009). The pretesting effect: Do unsuccessful retrieval attempts enhance learning? Journal of Experimental Psychology: Applied, 15(3), 243–257. https://doi.org/10.1037/a0016496
• Soderstrom, N. C., & Bjork, E. L. (2023). Pretesting enhances learning in the classroom. Educational Psychology Review, 35(3), Article 88. https://doi.org/10.1007/s10648-023-09805-6
• St. Hilaire, K. J., Chan, J. C. K., & Ahn, D. (2024). Guessing as a learning intervention: A meta-analytic review of the prequestion effect. Psychonomic Bulletin & Review, 31(2), 411–441. https://doi.org/10.3758/s13423-023-02353-8