Assessment and Knowledge Checks Inside Training Video

Why a check belongs inside each module, what a good question tests, and how the results tell you about the content rather than the learner.

Training video without assessment produces a record of attendance rather than of comprehension, and for most organisations the distinction is the reason the training exists. A short check inside each module converts viewing into evidence, and it does something more useful: it tells the organisation which parts of the content are not working.

The placement that works is inside each module rather than at the end of the course. A check immediately after the material is fresh tests whether that segment communicated, and the learner receives correction while the context is still active. A single assessment at the end tests recall across everything and identifies nothing specific when it is failed.

There is also a learning effect beyond measurement. Retrieval, being asked to produce an answer rather than recognise one, strengthens retention, which is why a check is a teaching device as well as a test. Cheng et al. (2023), examining the benefits of asking students to make an instructional video of a multimedia lesson, clarified the learning by teaching hypothesis, and the same underlying principle supports asking a learner to articulate rather than only absorb.

A good question tests the behaviour rather than the words. Asking which sequence is correct, or what should be done when a specific condition occurs, tests whether the procedure was understood. Asking what the module said about a topic tests whether the narration was heard. The second is easier to write and measures nothing that matters.

Scenario questions outperform factual ones for procedural content. Presenting a situation and asking what the operator should do next requires the learner to apply the material rather than recall it, and it surfaces the misunderstandings that a factual question conceals. These take longer to write and are the difference between an assessment and a formality.

The cognitive design should respect the same constraints as the content. Beege and Ploetzner (2025), studying learning from interactive video, examined how navigation and cognitive load influence what learners take from video material, and Ludwig et al. (2026) found that instructional design and cognitive load affect knowledge acquisition and problem solving. A check with eight questions after a five minute module inverts the ratio and turns the learning into an examination.

Feedback on a wrong answer should teach rather than score. Returning the learner to the relevant moment in the video, with an explanation of why the answer was incorrect, closes the loop. Marking an answer wrong and moving on wastes the most valuable moment in the module, which is the point where the learner has just discovered they misunderstood something.

The results are more useful as data about the content than as data about the learners, and almost nobody uses them that way. A question that most people fail is usually identifying a segment that explained badly rather than a workforce that is inattentive. Reviewing the failure rates by question, after the first cohort, produces a specific list of which modules to revise.

For compliance purposes the record has to be defensible, which shapes the implementation rather than the content. Who took it, when, what they scored, whether they retook it, and whether the version they saw is identifiable. Versioning matters here: when a module is updated, the record should show which version a given learner completed, because that is the question an auditor will ask.

Language and accessibility apply to the assessment as much as to the video, and are frequently overlooked because the questions are text. Zheng et al. (2022) found that adding subtitles to audio visual material assists comprehension, and the equivalent obligation for a check is that the question is comprehensible to a learner whose first language is not the one it was written in. A learner failing on the phrasing of a question rather than on the content is a defect in the assessment.

References

Cheng, M., Wang, F., & Mayer, R. E. (2023). Benefits of asking students to make an instructional video of a multimedia lesson: Clarifying the learning-by-teaching hypothesis. Journal of Computer Assisted Learning, 39(5), 1636–1651. https://doi.org/10.1111/jcal.12823

Beege, M., & Ploetzner, R. (2025). Learning from interactive video: The influence of self-explanations, navigation, and cognitive load. Instructional Science, 53(1), 99–119. https://doi.org/10.1007/s11251-024-09693-5

Ludwig, S., Rausch, A., & Taub, M. (2026). Effects of instructional design, instructional preferences, and cognitive load on problem solving and knowledge acquisition in a computer-based office simulation. Learning and Instruction, 101, Article 102255. https://doi.org/10.1016/j.learninstruc.2025.102255

Zheng, Y., Ye, X., & Hsiao, J. H. (2022). Does adding video and subtitles to an audio lesson facilitate its comprehension? Learning and Instruction, 77, Article 101542. https://doi.org/10.1016/j.learninstruc.2021.101542