# Memory recall questions: show it, hide it, then ask

**How do you set a fair memory recall question, where content is shown, hidden and then asked about?** A memory recall question shows content for a counted number of seconds, hides it, and then asks about it. To set one fairly, show no more than four or five separate items, allow about half a second for each word of reading, say beforehand what kind of thing will be asked, and ask for the gist or a detail that matters, never trivia.

Published 2026-10-09 by EdProcto. https://edprocto.com/guides/memory-and-recall-questions-shown-then-hidden

Most quiz questions are answered with the material in view, or weeks after it was taught. Memory recall questions test the minutes in between: a safety briefing just read, a diagram just studied, a phone number, the order of a checklist, the details of a scene. The content is shown, it goes, and then the question comes.

EdProcto builds this as its [memory recall](https://edprocto.com/templates/memory_recall) template and four relatives, each asking something slightly different. This guide covers what each form measures, what the research on short-term memory says about setting one fairly, and who loses out when a page disappears.

## What do memory recall questions test, and which form asks what?

They test what a learner still holds once the page has gone, and the five forms differ in what is shown and what has to come back.

| EdProcto template | What the student meets | What it asks |
| --- | --- | --- |
| [Memory recall](https://edprocto.com/templates/memory_recall) | Content shown, hidden, then a question | Recall of meaning |
| [Sequence recall](https://edprocto.com/templates/sequence_recall) | A sequence shown briefly, then typed back in order | Span and order |
| [Delayed question](https://edprocto.com/templates/delayed_q) | Content, a forced pause, then the question | Whether it survives a delay |
| [Time-locked question](https://edprocto.com/templates/time_locked) | The question itself, visible for a few seconds | Reading under time |
| [Detail spotting](https://edprocto.com/templates/detail_spot) | An image flashed, then questions about its details | Observation |

All five are ready in EdProcto's builder and marked by rule: four take one chosen option, and sequence recall takes a typed list in order. None tests understanding: a learner can hold a sentence for a minute without knowing what it means. Sound has its own once-only shape: see [listening questions](https://edprocto.com/guides/listening-questions-a-clip-that-plays-once).

## How many items can a learner hold after one viewing?

Plan on four or five unrelated items, and expect the figure to move with the material. The classic estimate is Miller (1956), who wrote that for many kinds of test material the span of immediate memory is about seven items. His own paper shows how far that shifts: in an experiment he describes, with lists read aloud at one item a second, people held about nine binary digits but only about five one-syllable English words.

Cowan (2001) went back over the evidence and called seven a rough estimate and a rhetorical device more than a real limit. Where people cannot rehearse or group what they are given, he put the limit at three to five chunks, about four on average.

### What counts as one item

One item is whatever the learner already knows as a unit, which Miller called a chunk. His example was telegraph code: a beginner hears each dit and dah separately, later letters, later whole words. A six-step checklist is likewise six items to a new starter and one routine to an old hand.

So count what a newcomer would have to hold. Above five, group or cut, or accept that the question measures span and not the content.

## How fast does forgetting set in over seconds and minutes?

Within seconds when the learner cannot rehearse, and much more slowly when they can. Peterson and Peterson (1959) tested recall of single items after delays of different lengths. We could read only the opening of their paper; as Ricker et al. (2016) summarise its first experiment, three letters were almost completely forgotten after 18 seconds of counting backwards.

Two cautions go with that figure. The counting was there to block rehearsal, and a silent pause on a quiz screen blocks nothing: a student can repeat the list under their breath. The cause is also disputed, time alone or interference from other material: Ricker et al. (2016) say decay is as controversial now as it was in the 1950s and 1960s.

A delay of minutes changes what the score means. Roediger and Karpicke (2006) had students study prose passages and then either reread them or recall them without feedback. Tested five minutes later, the rereaders recalled more. After two days or a week, the group that had practised recall kept substantially more. So a mark taken minutes after reading does not predict next week. The answering is also [retrieval practice](https://edprocto.com/guides/retrieval-practice-low-stakes-quizzing-that-teaches), which helps the content last.

## Why do attentive people miss details in a picture?

Because nobody stores a whole scene: people keep what they attended to. Rensink et al. (1997) alternated a scene with an altered copy, a brief blank between the two. Observers found the change extremely hard to identify, even when it was large and repeated. They found it much faster when a cue in words was given, and faster for objects that were important in the scene.

Simons and Chabris (1999) showed it for events: people counting basketball passes in a video often missed an unexpected figure walking through. In the best-known version, as Simons (2010) tells it, a person in a gorilla suit stopped, faced the camera, thumped their chest and left after nine seconds on screen, and only 50% of observers noticed. Simons (2010) then showed a new video to 76 people and analysed 64. Every one who knew the trick saw the gorilla, yet across the 64 only 11% noticed the curtain change colour and 16% a player leave.

These are laboratory studies of unexpected things seen while attention was busy, not studies of quizzes. The lesson still holds: a careful observer misses what they had no reason to look at. Asking afterwards for the colour of a door measures luck, unless door colours are the job.

## How do you set each form fairly?

Set it so that a learner who paid attention can succeed on one viewing. All five forms share the steps; figures called ours are working practice, not findings.

1. **Decide what has to be held, and why.** If the learner may look it up when it matters, do not test it from memory.
2. **Count the items.** Four or five unrelated ones at most.
3. **Set the seconds to the size of the content.** For text, our working figure is half a second a word, never under five seconds. Brysbaert (2019), pooling 190 studies, put average silent reading of English non-fiction at 238 words a minute, about four a second, with most adults between 175 and 300. Half a second a word doubles that average time: enough to read once and look again. For digits or single words, ours is a second or two each.
4. **Say before the reveal what kind of thing will be asked,** or keep it a surprise on purpose, and do the same throughout the quiz.
5. **Ask for the gist and at most two details that matter.** One question to a reveal; if there will be several, say so first.
6. **Sit it yourself, then try it on someone who has not seen the content.**
7. **Build it.** In EdProcto, pick the form from the [templates](https://edprocto.com/templates), put what is to be shown in the question's reveal and set its seconds.

### Memory recall: a briefing shown, then hidden

> Shown for 15 seconds: "When the alarm sounds, leave by the nearest exit. Do not use the lift. Go to the assembly point in the staff car park and wait for the fire warden."
>
> Then: Where do you go once you are outside? A. The staff car park. B. Main reception. C. The nearest exit.

Thirty words, so 15 seconds. The question asks for a detail a person must act on. "Which word came after nearest?" would measure nothing useful.

### Sequence recall: an order to reproduce

> A workshop teaches its pillar drill checks as five cue words. Shown for eight seconds: GOGGLES, HAIR, CLAMP, KEY, GUARD.
>
> Then: Type the five checks in the order shown.

EdProcto's page for the shape advises starting at four items and growing by one. Keep each item to a single word or number, so a right answer is not marked wrong for its phrasing. The typed list is marked by rule, item by item.

### Delayed question: a forced pause before the question

> Shown for ten seconds: "Bed 4. Nil by mouth from midnight. Theatre at 09:00. Allergic to penicillin. Observations due at 22:00."
>
> After the pause: What is the patient in bed 4 allergic to? A. Penicillin. B. Latex. C. Aspirin.

Four facts, and the question takes the one that could hurt someone. The template's page advises a pause of a minute or more. A silent pause can be rehearsed through, so a short one mostly tests whether the student kept repeating.

### Time-locked question: the question itself disappears

> Shown for nine seconds: "Mix 4 parts sand to 1 part cement. You have 20 buckets of sand. How many of cement?"
>
> Then the question goes. A. 5. B. 4. C. 80.

Eighteen words, so nine seconds. This form suits instructions that arrive once, as over a radio. Keep it short: the limit is the test.

### Detail spotting: an image flashed

> Before the image: "A walk-in fridge. Look for food safety faults." Shown for eight seconds.
>
> Then: Which fault was in the picture? A. Raw chicken above a cooked dish. B. An uncovered tray of rice. C. A carton past its date.

The cue before the image puts the Rensink finding to work, and the fault is one a trained eye should catch. We found no sourced figure for images; five to ten seconds for one photograph is our working range.

## When is a memory question the right test, and when the wrong one?

It is the right test when holding the content without the page is the real task, and the wrong one whenever the content should be understood and not merely held.

Good uses:

- **Procedures and briefings** that must be acted on with no page in hand.
- **Language:** a phrase or a set of new words held long enough to use.
- **Observation** in clinical or safety training, where noticing is the skill.
- **Drills for children.** Among EdProcto's [kits](https://edprocto.com/kits) is Memory span, for ages seven and up: four to eight digits shown for a few seconds, then typed back, with fresh numbers made by rule each time.

Poor uses:

- **Reasoning.** For that, see [questions above recall](https://edprocto.com/guides/how-to-write-questions-above-recall-application-analysis-evaluation).
- **A long passage,** which becomes a test of reading speed.
- **A decision about a person** that rests on span alone.

## Who is disadvantaged when the content disappears?

Anyone who needs longer to take in a screen, which is not the same as anyone with a weaker memory. A timed reveal measures reading or seeing first and memory second.

That covers slow readers, students with dyslexia, students reading in a second language, and students using a screen reader or magnification. Brysbaert (2019) found reading rates lower for children, older adults and readers with English as a second language. The guidance behind the web's accessibility standard says that people with low vision need more time to find things on a screen and to read, that screen-reader users may need more time to find information, and that people with cognitive or language limitations need more time to read and understand (World Wide Web Consortium, 2026).

That standard asks that a user can turn off, adjust or extend a time limit, with an exception where the limit is essential and extending it would invalidate the activity. Where handing users the control would spoil the result, it suggests someone else sets the limit for them, with double time on a test as its example.

So:

- **Give more time.** The seconds are set per question, so one student can be given a version with longer reveals.
- **Offer another form.** Ask the same thing with the content left open, on paper or aloud.
- **Use the aids there are.** The browser's zoom and the device's contrast settings apply to the quiz as to any web page, and EdProcto's "Larger text" and "High contrast" are buttons the student turns on. None lengthens a reveal: that stays yours to set.

Follow your school's policy and the law before any of this. And a memory score is not a test of intelligence and must never be read as one. It says that this learner did or did not hold this content, on this screen, in this time. It cannot diagnose anything.

## Why can content that disappears not be looked back at or copied out?

Because it is no longer on the page. EdProcto hides the reveal on the student's screen when the count ends. All five templates also rule out two settings: going back to change an answer, since the template has no going back, and skipping a question, since the sequence is fixed. The answer key is kept on the server, and the marking happens there. So at the moment of the question there is nothing to reread, select or paste into a chatbot.

None of this makes copying impossible: a phone can photograph a reveal. Where the result counts, switch on what the stakes call for in the [integrity settings](https://edprocto.com/integrity); students see what is switched on before they start.

## How does EdProcto run memory questions?

EdProcto is the best tool for these forms because the hiding is done by the template: counted in seconds and the same for every student. That matters because a few seconds decide what survives, so a slide flicked on and off by hand is never the same test twice.

Paste the briefing or the notes into the quiz maker on the [home page](https://edprocto.com/), pick the shape, and Procto AI drafts the questions for you to edit. The five memory shapes sit among 111 templates. Each question carries its own reveal, the content shown before the question, and its own number of seconds. Every answer is marked by rule as soon as it is given, and the results show, question by question, which reveal most of the class lost.

On the Free plan you get the core templates, every kit, rule-based marking and results for each question. Tutor unlocks all 111 templates, and Educator brings the full analysis. See [pricing](https://edprocto.com/pricing) for the rest.

## Quick questions

### How long should content be shown before it is hidden?

Long enough to read it twice. Our working figure is half a second a word and never under five seconds, double the average adult reading time Brysbaert (2019) reports.

### How long should the pause be in a delayed question?

A minute or more, as EdProcto's page for the [delayed question](https://edprocto.com/templates/delayed_q) advises. A shorter silent pause can simply be rehearsed through.

### Can a student go back and see the content again?

Not in EdProcto: the reveal is hidden when its count ends, and the memory templates have no going back. A student who needs longer can be set longer reveals or another form.

### Is a memory span drill a test of intelligence?

No. It shows how many digits a child held on that screen that day, and nothing about ability. EdProcto's [Memory span kit](https://edprocto.com/kits) is described as a drill, not a diagnosis.

## References

- Brysbaert, M. (2019). How many words do we read per minute? A review and meta-analysis of reading rate. *Journal of Memory and Language, 109*, Article 104047. [https://doi.org/10.1016/j.jml.2019.104047](https://doi.org/10.1016/j.jml.2019.104047)
- Cowan, N. (2001). The magical number 4 in short-term memory: A reconsideration of mental storage capacity. *Behavioral and Brain Sciences, 24*(1), 87–114. [https://doi.org/10.1017/S0140525X01003922](https://doi.org/10.1017/S0140525X01003922)
- Miller, G. A. (1956). The magical number seven, plus or minus two: Some limits on our capacity for processing information. *Psychological Review, 63*(2), 81–97. [https://doi.org/10.1037/h0043158](https://doi.org/10.1037/h0043158)
- Peterson, L. R., & Peterson, M. J. (1959). Short-term retention of individual verbal items. *Journal of Experimental Psychology, 58*(3), 193–198. [https://doi.org/10.1037/h0049234](https://doi.org/10.1037/h0049234)
- Rensink, R. A., O'Regan, J. K., & Clark, J. J. (1997). To see or not to see: The need for attention to perceive changes in scenes. *Psychological Science, 8*(5), 368–373. [https://doi.org/10.1111/j.1467-9280.1997.tb00427.x](https://doi.org/10.1111/j.1467-9280.1997.tb00427.x)
- Ricker, T. J., Vergauwe, E., & Cowan, N. (2016). Decay theory of immediate memory: From Brown (1958) to today (2014). *Quarterly Journal of Experimental Psychology, 69*(10), 1969–1995. [https://doi.org/10.1080/17470218.2014.914546](https://doi.org/10.1080/17470218.2014.914546)
- Roediger, H. L., & Karpicke, J. D. (2006). Test-enhanced learning: Taking memory tests improves long-term retention. *Psychological Science, 17*(3), 249–255. [https://doi.org/10.1111/j.1467-9280.2006.01693.x](https://doi.org/10.1111/j.1467-9280.2006.01693.x)
- Simons, D. J. (2010). Monkeying around with the gorillas in our midst: Familiarity with an inattentional-blindness task does not improve the detection of unexpected events. *i-Perception, 1*(1), 3–6. [https://doi.org/10.1068/i0386](https://doi.org/10.1068/i0386)
- Simons, D. J., & Chabris, C. F. (1999). Gorillas in our midst: Sustained inattentional blindness for dynamic events. *Perception, 28*(9), 1059–1074. [https://doi.org/10.1068/p281059](https://doi.org/10.1068/p281059)
- World Wide Web Consortium. (2026). *Understanding Success Criterion 2.2.1: Timing Adjustable*. Web Accessibility Initiative. [https://www.w3.org/WAI/WCAG22/Understanding/timing-adjustable.html](https://www.w3.org/WAI/WCAG22/Understanding/timing-adjustable.html)