A video is the easiest thing in a lesson to set and the easiest not to watch. It plays whether or not anyone attends, and a class that sat through a demonstration will tell you it was clear. Video quiz questions turn the watching into a task, and they separate the student who followed the procedure from the one who only had it on.
EdProcto builds the shape as its watch and answer template: a video, then questions with one option to choose, marked by rule. This guide covers the research, the length of the clip, what to ask, the replay rule and captions.
One limit first. Most of the studies below used recorded lectures and university students, so read their numbers as a direction and not a forecast.
What does research say about questions on a video?
Questions asked between short segments of a video hold attention and improve learning more than the same segments with the test kept for the end. The clearest trial is by Szpunar et al. (2013), who called the method interpolated testing.
Students watched a statistics lecture of 21 minutes in four parts, each about five and a half minutes long, with a break after every part. In the second of two experiments, 48 students were split three ways. One group was quizzed in each break. Another filled the breaks with arithmetic, and the last was shown the quiz questions with their answers. When an experimenter asked, at unannounced points, whether attention had drifted, the quizzed group said yes on 19% of occasions and the other two on 41% and 39%. The quizzed group then got 89% on questions about the fourth part, against 70% and 65%, wrote notes on more of the slides and reported less anxiety about the test at the end.
Seeing the answers without having to produce them changed nothing, so the effort of answering was what counted. And every student expected a test at the end, so a final quiz alone did not hold attention.
Questions handed out before the video help too. Lawson et al. (2006) had 127 psychology students watch a video in class. Those given eight guiding questions to answer in writing as they watched scored higher on the video's content in the test that followed.
Why watching alone teaches less than it feels
Because a clear demonstration feels understood whether or not it was. Crouch et al. (2004) compared ways of presenting classroom demonstrations: students who only watched understood the underlying idea no better than students who never saw the demonstration, and students who predicted the outcome first understood significantly more. Those were classroom demonstrations and not clips, and we read the abstract only. The lesson transfers: stop the clip before the result and ask what will happen.
How long should the clip be?
One to four minutes for one idea is our working rule. The research says short, and gives no figure for a quiz.
Guo et al. (2014) studied 6.9 million viewing sessions from four courses on the edX platform. Whatever a video's length, the median time spent watching it was at most six minutes, and students often got less than halfway through videos longer than nine minutes. Fewer students tried the problem that followed a longer video: 56% of sessions after the shortest group of videos, 31% after the longest.
The authors state the limits themselves: viewing time cannot show whether anyone was attending, and the result is not a finding about learning.
So cut. "Watch from 2:10 to 3:40" asks a student to operate a player and invites skipping. If the idea needs ten minutes, make three clips and ask after each, which is the arrangement Szpunar et al. (2013) tested.
Why can a video with narration and on-screen text overload a viewer?
Because the text and the moving picture compete for the eyes, while the narration reaches the ears. Mayer and Moreno (2003) set out the theory: people process pictures and words in separate channels, each channel holds little at one time, and overload comes when a lesson demands more than a channel has.
Brame (2016), applying that work to teaching videos, gives the plain case. An animation explained by a voice uses both channels. The same animation explained in printed text sends everything through the eyes, and learning suffers.
For a quiz clip, choose footage where the action is visible and the talk is spare. A clip that makes the viewer read a slide while watching a technique measures how well they divide their attention.
What should a question on a video ask?
Ask for what a viewer could only know by having watched: what happened, in what order, what was done wrong, what should happen next, and what they would do. EdProcto's own note on the template puts it in a line: ask about what can only be seen, not what the narration says.
| Ask | It shows | Shape |
|---|---|---|
| What happened? | The clip was watched | A choice after the clip: EdProcto's watch and answer |
| In what order? | The procedure was followed | Sequence the process |
| What was done wrong? | An error is recognised on sight | A choice, or a short written answer |
| What should happen next? | The idea is understood well enough to predict | A choice, with the clip cut before the result |
| What would you do? | Judgement | A short answer against a rubric |
| What does it show, in a sentence? | Observation and expression together | Caption it |
| Where, when or whose is it? | A trained eye or ear | Identify the source |
Never ask what colour the presenter's shirt was, unless observing is the skill: a witness statement, a patient's walk, a referee's call. Then say so beforehand, because viewers do not keep what nobody asked them to notice, as the guide to memory questions shows.
Answer your question with the video off: if general knowledge is enough, it does not test the clip. Then answer it from the spoken words alone: if they are enough, it is a listening question.
Captions and sources as answers
A caption question asks for one sentence that says what the clip shows. List in the rubric what a full caption needs: what was done, what changed, under what condition. The guide to short answers and rubrics shows how.
An identify question asks for the period, the technique or the maker: which stroke a swimmer is using, which decade a newsreel comes from. Pick a typical example over a famous exception, and ask how the student can tell, so that the reasoning earns the marks.
Should students be able to replay, pause or go back?
Allow what the real task allows, and tell students the rule before the clip starts.
| Choice | The mark then shows |
|---|---|
| Replay allowed | Understanding with the material to hand, as it is at work |
| One play | What was caught at first sight: an incident witnessed, a decision in real time |
| Pausing | Close reading of a frame: a detail question becomes an image question |
| Going back to change an answer | The student's last word, including anything a later question gave away |
A single play of a long clip tests memory more than understanding, so keep one-play clips short. A prediction question survives any number of replays, because its answer is not in the clip.
In EdProcto the replay rule is a switch on the question, labelled "Plays once, no replay". With it on, the student gets the picture with one button under it, "Play the clip", and a line that fills: the video runs a single time from start to end, with no pause and no going back. Going back is a setting of the quiz: "Allow going back to change an answer".
A worked example: one clip, four questions
The clip is invented for this guide: ninety seconds, no narration. A teacher puts a ball of steel wool in a dish on a balance, which reads 48.0 g, holds the wool in a flame until all of it has glowed, lets it cool in the dish and sets the dish back on the balance. The clip ends before the display is shown.
1. What happened. What did the teacher do to the steel wool between the two weighings? A. Heated it in a flame until it glowed. B. Soaked it in water. C. Pulled it into two pieces.
Key: A. This shows only that the clip was watched.
2. In what order. Put these in order: leave the wool to cool; hold the wool in the flame; return the dish to the balance; note the first reading.
Key: reading, flame, cooling, balance.
3. What should happen next. Compared with 48.0 g, the second reading will be: A. higher. B. the same. C. lower.
Key: A. Iron combines with oxygen from the air, and the oxide left in the dish holds the mass of both. C is the belief that burning destroys matter, carried over from wood, whose products leave as gas. B treats an open dish as though it were sealed.
4. Why. In one or two sentences, explain your prediction.
Rubric: one mark for saying that the iron reacts with oxygen from the air, and one for saying that the solid left behind includes that oxygen.
Anyone who watched can answer the first two. The last two need understanding, and no replay helps.
Who needs captions, transcripts or audio description?
Students who are deaf or hard of hearing need captions, students who are blind or partially sighted need audio description or a text version, and many hearing students do better with captions as well. The web's accessibility standard requires captions on prerecorded video with sound at Level A, its lowest, together with either audio description or a full text alternative; Level AA requires audio description (World Wide Web Consortium, 2024). Audio description is narration added to tell a viewer the visual details the soundtrack leaves out.
Gernsbacher (2015) reviewed more than 100 studies and concluded that captions improve comprehension of, attention to and memory for a video, most of all for deaf and hard-of-hearing viewers, second-language viewers and people learning to read. That sits awkwardly beside the finding that printed text can crowd the eyes, and nothing we read settles it. Captions a viewer can switch off are the practical answer.
When a transcript changes the test
A transcript changes the test when the answer is in the words: a question about what was said becomes a reading question. If the answer can only be seen, captions and transcripts give nothing away.
Audio description is the hard case, because describing the action to a blind student may state the answer. The standard lets a test that would be invalid as text be identified in text and no more, but its audio description criteria carry no such exception. Set that student a different question on the same aim, and put your institution's policy and the law first.
EdProcto does not add captions to a video. Have the captions made where the video is kept, before you attach it.
What rights do you need to use a video in a quiz?
You need the right to use the video in the way you are using it; this is general caution and not legal advice. A video you filmed yourself is the simplest case, provided the people in it agreed to be shown. For anyone else's video, read its licence or terms: some allow copying and cutting, and some allow showing only where the owner published it. Where that is the licence, link to the video and do not download, trim or re-upload it. Rules for teaching differ by country, so ask whoever handles copyright at your institution before a clip goes into a test that counts.
How does EdProcto run a video question?
EdProcto is the best tool for this job when the questions must be about your own clip, because each question carries its own video address. A quiz can then be a run of short clips with a question after each, the arrangement that held attention in the trial above.
Choose watch and answer among the 111 templates, paste the video's address into the question's "Video address" field, then write the prompt and the options. The address has to lead to the video file itself: a link to a page on a video-sharing site will not play inside the question, so sit the quiz as a student before the class does. EdProcto stores the address with the question and marks the chosen option by rule, with the key kept on the server. A question on what can only be seen also gives a chatbot little to work from, which is a hurdle and not a guarantee.
In caption it and identify the source the medium can be an image, audio or video, and the written answer is marked by you or, when you ask, by Procto AI against your rubric.
For a first draft, paste the transcript or your notes on the clip into the quiz maker on the home page. Procto AI writes questions about what the text says, so edit them towards what is seen.
Free covers the core templates, with marking by rule and a result for every question. Tutor opens all 111 templates and adds grading by Procto AI against your rubric. Educator adds the full analysis, and audio and video as sources for Procto AI's writing, transcribed in your browser. The pricing page has the detail.
What are the common mistakes?
Most come from treating the video as the lesson and the questions as an afterthought.
- One long video with every question at the end.
- A question that general knowledge answers with the video off.
- A later question that gives away an earlier one.
- No captions, or a transcript that hands over the answer.
- An address that opens for you and not for them. Sit the quiz as a student, on a phone, before the class does.
Quick questions
How long should a video be for a quiz question?
One to four minutes for one idea. Guo et al. (2014) found that the median time spent watching was at most six minutes whatever the video's length.
Should the questions come during the video or at the end?
Between short parts. Szpunar et al. (2013) found that students questioned after each segment of a lecture mind-wandered less and learned more than students tested once. In EdProcto, give each question in watch and answer its own short clip.
Should students be allowed to replay the video?
Yes where the real task allows a second look, and no where it gives one look only. Say which before the clip starts.
Do I need captions on a video in a quiz?
Yes, unless the video has no sound that matters. The accessibility standard requires them at Level A (World Wide Web Consortium, 2024), and Gernsbacher (2015) concluded that they help hearing viewers as well.
Can Procto AI write questions from a video?
On the Educator plan, from the video itself: drop it into the Procto AI pane in the builder, and it is transcribed in your browser before Procto AI writes. On any plan, paste the transcript or your notes into the quiz maker on the home page and edit what it drafts.
References
- Brame, C. J. (2016). Effective educational videos: Principles and guidelines for maximizing student learning from video content. CBE Life Sciences Education, 15(4), Article es6. https://doi.org/10.1187/cbe.16-03-0125
- Crouch, C. H., Fagen, A. P., Callan, J. P., & Mazur, E. (2004). Classroom demonstrations: Learning tools or entertainment? American Journal of Physics, 72(6), 835–838. https://doi.org/10.1119/1.1707018
- Gernsbacher, M. A. (2015). Video captions benefit everyone. Policy Insights from the Behavioral and Brain Sciences, 2(1), 195–202. https://doi.org/10.1177/2372732215602130
- Guo, P. J., Kim, J., & Rubin, R. (2014). How video production affects student engagement: An empirical study of MOOC videos. In Proceedings of the First ACM Conference on Learning @ Scale (pp. 41–50). Association for Computing Machinery. https://doi.org/10.1145/2556325.2566239
- Lawson, T. J., Bodle, J. H., Houlette, M. A., & Haubner, R. R. (2006). Guiding questions enhance student learning from educational videos. Teaching of Psychology, 33(1), 31–33. https://doi.org/10.1207/s15328023top3301_7
- Mayer, R. E., & Moreno, R. (2003). Nine ways to reduce cognitive load in multimedia learning. Educational Psychologist, 38(1), 43–52. https://doi.org/10.1207/S15326985EP3801_6
- 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
- World Wide Web Consortium. (2024). Web Content Accessibility Guidelines (WCAG) 2.2. https://www.w3.org/TR/WCAG22/