All articles
5 min readlearning / training

86% versus 66% on the last segment: quiz as you teach

I run full-day trainings, and the last hour is the one I lose. Everyone is still in the room, laptops open, nodding. Ask them a week later what happened after 4pm and the answers thin out. Karl Szpunar, Novall Khan and Daniel Schacter found where the repair goes: in the gaps between segments.

The quizzed group held the last segment.

Szpunar and his colleagues split a 21-minute statistics lecture into four equal segments. One group answered short recall questions after each segment. A second group restudied the same material. A third watched the lecture straight through. Random probes during the lecture asked people whether their attention had drifted. The quizzed group reported drifting on 19% of probes, against 39% for the review group and 41% for the uninterrupted group 1.

The scores separated where attention usually gives out. On the fourth segment the quizzed group scored 86%, against 72% with no interruptions and 66% after review 1. Across the full cumulative test they averaged 90%, ahead of both other groups. They also wrote more notes during the lecture and reported less anxiety about the test to come 1.

Jing, Szpunar and Schacter stretched the design to a 40-minute public health lecture cut into eight 5-minute segments, with free recall after each one. The quizzed group learned more than a group that reviewed its slides for the same amount of time, and it tied the material together better both inside segments and across the whole lecture 2.

The classroom literature is larger. Chunliang Yang, Liang Luo, Miguel Vadillo, Rongjun Yu and David Shanks pooled 222 studies covering 48,478 students and found a medium benefit of quizzing on achievement, g = 0.499 3.

The pause does nothing without the recall.

Zohara Assadipour, Dahwi Ahn and Jason Chan tested what the interruption has to contain. In two experiments, 249 people in a lab and 246 online, the interpolated task was either recall of the segment people had watched, recall of an autobiographical memory, or retrieval of general knowledge. Only recall of the lecture material improved learning of the segment that followed 4.

Chan, Ahn, Szpunar, Assadipour and Hymnjyot Gill then ran 703 university and community college students through a 20-minute lecture in four 5-minute segments, with either quiz questions or review slides between them, under three levels of onscreen distraction: nothing, memes, or TikTok videos. Quizzes improved learning for both student populations 5. The distraction produced a result the authors flag as unexpected: with TikTok playing, the review group caught up to the quizzed group 5.

The attention claim is the shaky half.

Welhaf, Miyake, Kane and their co-authors ran 195 undergraduates through a narrated statistics lecture with the same manipulation. Interpolated testing lowered task-unrelated thoughts compared with restudying, but the effect was small and the Bayes factors gave it no support. They found no evidence of a learning advantage either, and concluded that interpolated testing may do little for attention 6. Jing and colleagues also measured no difference in overall mind wandering between their groups, though wandering hurt the review group's final score more than the quizzed group's 2.

So the 19% against 39% from the original experiment has not held up as the mechanism. The learning effect survives in the larger samples, including 703 students across two kinds of institution 5 and 222 classroom studies 3. I build the session around retrieval and treat a quieter room as a bonus.

The protocol.

  1. Cut the session into 5 to 10 minute teaching blocks. The studies that worked used 5-minute segments 25 and segments of a little over 5 minutes 1. A 3-hour module comes out at 18 to 36 blocks.
  2. End each block with 2 to 4 recall questions on that block alone. Closed notes, 60 to 90 seconds, free recall or short answer. In the studies the questions covered only the segment people had watched 12.
  3. Ask about the content you taught. Recalling anything else, including general knowledge, left learning of the next segment unchanged 4. A stretch break is not a substitute.
  4. Run the quiz after the very first block. Being tested once raises what people expect of the blocks that follow 5, and the tested group kept taking notes through the lecture 1.
  5. Show the answers right after and record no scores. The quizzed group in the original experiment ended up less anxious about the final test than the group that had never been tested 1.
  6. Give the final block its own quiz. On the fourth segment the quizzed group beat the restudy group by 20 points 1.
  7. Keep the quizzes in remote sessions. With 703 students and a TikTok feed running on screen, the quizzed groups still learned more than the baseline conditions 5.

For my next full-day training the three slots I care about are minute 5, minute 10, and the first block after lunch.

Sources.

  1. Szpunar, K. K., Khan, N. Y., & Schacter, D. L. (2013). Interpolated memory tests reduce mind wandering and improve learning of online lectures. Proceedings of the National Academy of Sciences, 110(16), 6313-6317. doi.org/10.1073/pnas.1221764110
  2. Jing, H. G., Szpunar, K. K., & Schacter, D. L. (2016). Interpolated testing influences focused attention and improves integration of information during a video-recorded lecture. Journal of Experimental Psychology: Applied, 22(3), 305-318. doi.org/10.1037/xap0000087
  3. Yang, C., Luo, L., Vadillo, M. A., Yu, R., & Shanks, D. R. (2021). Testing (quizzing) boosts classroom learning: A systematic and meta-analytic review. Psychological Bulletin, 147(4), 399-435. doi.org/10.1037/bul0000309
  4. Assadipour, Z., Ahn, D., & Chan, J. C. K. (2025). Interpolated retrieval of relevant material, not irrelevant material, enhances new learning of a video lecture in-person and online. Behavioral Sciences, 15(5), 668. doi.org/10.3390/bs15050668
  5. Chan, J. C. K., Ahn, D., Szpunar, K. K., Assadipour, Z., & Gill, H. (2025). In-lecture quizzes improve online learning for university and community college students. Communications Psychology, 3, 54. doi.org/10.1038/s44271-025-00234-5
  6. Welhaf, M. S., Phillips, B. A., Smeekens, B. A., Miyake, A., & Kane, M. J. (2022). Interpolated testing and content pretesting as interventions to reduce task-unrelated thoughts during a video lecture. Cognitive Research: Principles and Implications, 7, 26. doi.org/10.1186/s41235-022-00372-y

Keep reading