# Knowledge check questions for training: three per module

**How do I check understanding during a training workshop?** Knowledge check questions for training work best as three questions at the end of each workshop module: one recall, one application to the delegates' own work and one asking which statement is wrong. Delegates answer on phones with no score shown. Read the counts in two minutes, then move on, repeat one point or park it for a follow-up quiz a week later.

Published 2026-10-10 by EdProcto. https://edprocto.com/guides/knowledge-checks-inside-a-workshop-three-questions-per-module

The usual way to check understanding in a workshop is to ask "Any questions?" and read the silence. Silence shows who will speak in front of colleagues, not who can use the idea on Monday.

Knowledge check questions for training do better. Three questions close each module, delegates answer on their phones, and two minutes later you know what to repeat before the room moves on. It is formative assessment in corporate training at its plainest: no score, one decision.

EdProcto, which publishes this guide, runs such a check from a link that opens on any phone; its page [for trainers](https://edprocto.com/solutions/trainers) shows the rest. Our working rules are labelled: they are not research findings.

## What is a knowledge check in a workshop for?

A knowledge check in a workshop is for one decision, made on the spot: move on, repeat one point, or park the point for later.

It reports to nobody. The test that reports to a client is another instrument: see [before and after tests](https://edprocto.com/guides/before-and-after-tests-proving-a-training-course-worked). If no answer could change your next ten minutes, drop the check.

## What does research say about short tests between modules?

Short tests between the parts of a long session helped later learning in most experiments, not all, and nearly every participant was a student.

### Tests between the segments of a lecture

Szpunar et al. (2013) cut a 21-minute recorded lecture into four segments. Undergraduates given six questions after each scored higher on a closing test and reported mind wandering at 19% of the moments they were asked, against 39% and 41% in the comparison groups. That is interpolated testing.

Later results are mixed. Jing et al. (2016) again found more learning than from restudying, yet in their first experiment minds wandered as often with tests as without. Welhaf et al. (2022), with 195 undergraduates, found a small drop in off-task thought and no gain on the later test.

All three used video lectures and undergraduates. We found no trial of a live workshop for working adults: take the direction, leave the sizes.

### Retrieval practice with adults and professionals

Answering from memory is called retrieval practice. Much of its evidence is from classrooms: Yang et al. (2021) pooled 222 studies and 48,478 students and found that quizzing raised achievement by about half a standard deviation. Three studies sit closer to a training room.

- **Older adults.** In Meyer and Logan (2013), adults aged 55 to 65 gained as much from being tested as university students did.
- **Doctors in training.** Larsen et al. (2009) randomised 40 residents after a teaching session. Six months on, topics tested three times with feedback were recalled better than topics restudied: 39% against 26%.
- **Employees.** Kim et al. (2019) analysed 10,514 people answering repeated questions in workplace training: the best gap between repeats grew with the time the knowledge had to last. It did not test quizzing against no quizzing.

## Does a check on the day prove the training will stick?

No. A check on the day shows what the room can do at that minute, a poor guide to what it will do in a month.

Soderstrom and Bjork (2015) conclude from their review that performance during training is often an unreliable index of lasting learning, and that learners mistake one for the other. Larsen's residents scored 39% after six months even on the topics tested three times. Yang et al. (2012), reviewing 11 studies of healthcare staff after advanced life support courses, concluded that knowledge and skills decay within six months to a year.

A check says what to repeat today. It is never evidence for the client that the training worked.

## What can three questions tell you about a room of twelve?

Three questions can tell you which point most of the room missed and which wrong idea several delegates share. They cannot grade a person or rank the room.

The arithmetic is ours.

- **Read counts, not percentages.** One delegate is 8.3 points; "58% right" is seven people.
- **One person's score means little.** Someone guessing blind at three four-option questions gets two or more right 16% of the time.
- **The room's count means more.** Twelve blind guessers average three right on such a question; seven or more would turn up about once in 70 tries.

EdProcto says as much on every plan: until five people have handed in, its results page reads "Fewer than five submissions: the numbers below describe these students, not the class or the teaching yet."

## How do you write the three knowledge check questions?

Write one recall question, one application question set in the delegates' own work, and one that asks which statement is wrong. The mix is our working rule.

- **Recall.** The module's central rule, asked plainly. It shows whether the room has the idea at all.
- **Application.** A case from the delegates' own job. It shows whether the idea survives real work.
- **Which is wrong.** Four statements, one carrying the mistake you expect. It shows which wrong idea is shared.

Recall does not prove application: a delegate can recite the rule and still not use it on their own letter. Our guide to [questions above recall](https://edprocto.com/guides/how-to-write-questions-above-recall-application-analysis-evaluation) shows how to ask for the use itself.

The third question has a critic. Chiavaroli (2017) notes that most guides discourage negatively worded questions: strong candidates answer as if the question were positive, and highlighting the negative word does not prevent it. He writes about high-stakes exams; here a misread costs a false alarm, not a mark. Our rule: say aloud that question three hunts for the error, and if its misses scatter, suspect the wording first.

1. **Draft from the module.** Paste its notes into the quiz maker on EdProcto's [home page](https://edprocto.com/) and Procto AI drafts [multiple choice](https://edprocto.com/templates/mcq) questions from that text, yours to edit.
2. **Rewrite question two by hand** around a real document from the client.
3. **Build each wrong option from a mistake** you have heard before; see [writing distractors](https://edprocto.com/guides/how-to-write-distractors-that-work).

## How do you read the answers in two minutes?

Read three counts and one wrong option, decide, and say the decision aloud.

1. **Wait for the room.** Once the first answers are in, EdProcto's results page lists who is still "In the quiz now" and how far each has got.
2. **Read "Got it"** for each question: the share who answered it right.
3. **Open the weakest question.** "Got it" names it on every plan; in the full analysis, on the Educator plan, each option shows how many chose it.
4. **Decide**, and tell the room what their answers changed.

| What you see | Decision |
| --- | --- |
| Ten or more of twelve right on every question | Move on |
| Three or more on the same wrong option | Repeat one point: show the option, say where it fails, give a fresh example |
| Scattered misses, or a gap too big for five minutes | Park it for the follow-up quiz |

The thresholds are our working rule for twelve; no study we read tests a cut-off. One exception: if half the room misses the recall question, the module did not land, so teach its core again another way.

Read on your own laptop, not the projector: lower down, the page lists each name with a score.

## How do you test adults without making them feel examined?

Keep the check unscored, keep names out of it where you can, and say why you are asking.

The evidence is from students. In two experiments by Hinze and Rapp (2014), quizzes taken under high-stakes pressure gave no advantage over rereading on a later test; low-stakes quizzes did. Szpunar's tested students also reported less anxiety about their final test.

Anonymity is our working rule. No student in those studies had a manager two chairs away; for an adult at work the stakes are social, so remove the audience and say so: "Three questions. No score, nothing goes to your manager. They tell me what to go over again."

In EdProcto:

- **"Show the score" at "Never".** After handing in, the delegate reads "This quiz does not show a score." The right answers still appear.
- **"An account is needed to take it" off.** The start screen asks only for "Your name".
- **A nickname for anonymity.** A typed name is still a name: the check is anonymous only if you tell delegates to type a nickname, each a different one, because a second person typing a name already handed in is turned away. Or set "Names shown to" at "No one": no name is asked, and you see each delegate by a number.

## A worked example: one day, four modules, twelve delegates

Twelve people from a housing association's customer team take a one-day workshop on plain writing in four modules. The figures are invented to show the method. This is the check on module two, active verbs.

> **1.** A sentence is active when its subject:
> A. does the action  B. receives the action  C. is left out
>
> **2.** From your own arrears letter: "Your payment plan will be reviewed in March." Which rewrite is active?
> A. We will review your payment plan in March.
> B. Your payment plan will be reviewed by us in March.
> C. Your payment plan is being reviewed in March.
>
> **3.** Three of these are sound. Which ONE is wrong?
> A. The passive can leave out who did it.
> B. The passive is useful when the doer is unknown.
> C. A passive sentence is a grammatical mistake.
> D. Adding "by us" to a passive sentence does not make it active.

The answers are A, A and C. Eleven of twelve got question one. Seven got question two, and four chose B. Eight got question three, and three chose D.

Questions two and three say one thing twice: four people believe that naming the doer makes a sentence active, and three of them said so again by calling statement D wrong. The decision is to repeat one point. After lunch the trainer shows option B, asks where the doer sits, and has the room rewrite two lines from their own letters.

Module four's check, at 15:45, showed a gap too wide for the ten minutes left. It was parked for the follow-up quiz.

| Time | What happens |
| --- | --- |
| 09:30 | Welcome; why the checks carry no score |
| 09:40 | Module 1 |
| 10:45 | Check 1 on [EdProcto](https://edprocto.com/), then the break |
| 11:05 | Repeat slot, then module 2 |
| 12:15 | Check 2, then lunch |
| 13:05 | Repeat slot, then module 3 |
| 14:15 | Check 3, then the break |
| 14:35 | Repeat slot, then module 4 |
| 15:45 | Check 4, repeats and the parked list |
| 16:00 | Action plans |
| 16:15 | Close |

Each check takes five minutes, three to answer and two to read, and sits just before a break: the decision is made over coffee and the repeat opens the next session. The four checks use 20 of the day's 330 working minutes, about 6%; with the repeat slots, 45 minutes, about 14%.

## What should the follow-up quiz a week later ask?

The follow-up quiz should ask fresh versions of the day's questions, the repeated and parked points first.

"A week" is our working rule. Cepeda et al. (2008) taught facts to more than 1,350 people and varied the gap before one review. The best gap was about 20 to 40% of the delay for a test a week away and 5 to 10% for a test a year away, which is 18 to 37 days. For knowledge that must last a year, the study points nearer a month. We say a week because the parked points are owed an answer; a second quiz at a month is better still. See [a spaced repetition schedule](https://edprocto.com/guides/re-tests-that-come-due-how-spacing-is-scheduled) for the gaps.

Write new questions: the day's answers were shown. In two-day training sessions, open day two with three quiz questions on day one.

## How does EdProcto run a knowledge check in a workshop?

EdProcto is the best tool for a check between modules when the questions must come from that morning's material and the answers must be read before the break ends. Procto AI drafts from the module's own notes, choice questions are marked by rule at hand-in, and the results fill while the quiz is open.

1. **Make each check its own quiz.** Leave the kind at "Exam", which means only handed in once and marked at the end. "Study quiz" shows each answer at once, and "Quiz night" puts a scoreboard of names on the host's screen.
2. **Change three settings.** "Time limit" to "None", "Show the score" to "Never", "An account is needed to take it" off. Leave security off: [integrity](https://edprocto.com/integrity) is for sittings that count.
3. **Press "Publish & share"** as the module ends and show the link, which ends in the code.
4. **Press "Close the quiz"** when you have read the results.

For the day's last check, the [exit ticket](https://edprocto.com/templates/exit) template, ready in the builder, adds an unmarked reflection; it rules out the "Show the score" setting.

A four-module day fits the Free plan: four checks and a follow-up are five of its ten quizzes a month, twelve delegates are inside its thirty a quiz, and the twenty questions a month that Procto AI writes cover the four checks with eight to spare. Tutor has no limit on quizzes and takes a hundred people a quiz; Educator adds certificates and the full analysis. Educator also puts your logo and colour on the delegate's screen. See [pricing](https://edprocto.com/pricing).

## Quick questions

### How many knowledge check questions should each module have?

Three: one recall, one application, one asking which statement is wrong. The number is our working rule.

### Should a knowledge check in training be scored?

No. Hinze and Rapp (2014) found that quizzes taken under pressure lost their advantage over rereading. Show the answers, not a score.

### Can delegates answer on their phones without an account?

Yes. With "An account is needed to take it" off, EdProcto asks only for a name. The page [for trainers](https://edprocto.com/solutions/trainers) shows the path from module to code.

### Does a good result on the day mean the training worked?

No. It means the room followed the module. Ask again a week later.

## References

- Cepeda, N. J., Vul, E., Rohrer, D., Wixted, J. T., & Pashler, H. (2008). Spacing effects in learning: A temporal ridgeline of optimal retention. *Psychological Science, 19*(11), 1095–1102. [https://doi.org/10.1111/j.1467-9280.2008.02209.x](https://doi.org/10.1111/j.1467-9280.2008.02209.x)
- Chiavaroli, N. (2017). Negatively-worded multiple choice questions: An avoidable threat to validity. *Practical Assessment, Research, and Evaluation, 22*, Article 3. [https://doi.org/10.7275/ca7y-mm27](https://doi.org/10.7275/ca7y-mm27)
- Hinze, S. R., & Rapp, D. N. (2014). Retrieval (sometimes) enhances learning: Performance pressure reduces the benefits of retrieval practice. *Applied Cognitive Psychology, 28*(4), 597–606. [https://doi.org/10.1002/acp.3032](https://doi.org/10.1002/acp.3032)
- 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. [https://doi.org/10.1037/xap0000087](https://doi.org/10.1037/xap0000087)
- Kim, A. S. N., Wong-Kee-You, A. M. B., Wiseheart, M., & Rosenbaum, R. S. (2019). The spacing effect stands up to big data. *Behavior Research Methods, 51*(4), 1485–1497. [https://doi.org/10.3758/s13428-018-1184-7](https://doi.org/10.3758/s13428-018-1184-7)
- Larsen, D. P., Butler, A. C., & Roediger, H. L., III. (2009). Repeated testing improves long-term retention relative to repeated study: A randomised controlled trial. *Medical Education, 43*(12), 1174–1181. [https://doi.org/10.1111/j.1365-2923.2009.03518.x](https://doi.org/10.1111/j.1365-2923.2009.03518.x)
- Meyer, A. N. D., & Logan, J. M. (2013). Taking the testing effect beyond the college freshman: Benefits for lifelong learning. *Psychology and Aging, 28*(1), 142–147. [https://doi.org/10.1037/a0030890](https://doi.org/10.1037/a0030890)
- Soderstrom, N. C., & Bjork, R. A. (2015). Learning versus performance: An integrative review. *Perspectives on Psychological Science, 10*(2), 176–199. [https://doi.org/10.1177/1745691615569000](https://doi.org/10.1177/1745691615569000)
- 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. [https://doi.org/10.1073/pnas.1221764110](https://doi.org/10.1073/pnas.1221764110)
- Welhaf, M. S., Phillips, N. E., 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*, Article 26. [https://doi.org/10.1186/s41235-022-00372-y](https://doi.org/10.1186/s41235-022-00372-y)
- 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. [https://doi.org/10.1037/bul0000309](https://doi.org/10.1037/bul0000309)
- Yang, C.-W., Yen, Z.-S., McGowan, J. E., Chen, H. C., Chiang, W.-C., Mancini, M. E., Soar, J., Lai, M.-S., & Ma, M. H.-M. (2012). A systematic review of retention of adult advanced life support knowledge and skills in healthcare providers. *Resuscitation, 83*(9), 1055–1060. [https://doi.org/10.1016/j.resuscitation.2012.02.027](https://doi.org/10.1016/j.resuscitation.2012.02.027)