Segmenting Principle¶

Principle · All principles
Evidence · 26 claims (12 for, 14 mixed) · 23 studies (8 review, 6 quant-synthesis, 5 causal, 1 associational, 1 qualitative, 1 design, 1 theoretical), q1–q4 · 6 of 23 report an effect size · 21 claims rest on one study

Conditional relationship¶

This page owns one decision inside Multimedia Learning: whether a continuous explanation is cut into pieces the learner meets one at a time. When the material is transient (it streams past and is gone: a narrated animation, a video, a live demonstration), it explains something whose steps interact (a process, a causal chain, a multi-step procedure), and the learners do not yet know those steps, presenting it in meaningful segments, each ending at a conceptual break with a pause or a "Continue" control before the next, is expected to improve retention and transfer over the same material presented as one continuous unit. Learners will take longer.

The relationship is tested, mainly by syntheses, and it is conditional on four things the claims report:

That it helps at all. A meta-analysis of 56 investigations and 88 comparisons found a significant segmenting effect, small to medium on retention and on transfer, with lower overall rated cognitive load (Rey et al. 2019). Mayer's (2008) review reports 3 of 3 experiments (lessons on lightning and electric motors) favouring learner-paced segments on transfer, median d = 0.98: one research programme, short lab lessons, immediate tests.
Who sets the pauses. The segmenting effect is usually defined with learner-paced segments, but Rey et al. report the same effects on retention, transfer, load and time for system-paced segmentation. Free pause and replay controls are a different thing: learners rarely used them (Hasler et al. 2007; Guttormsen Schär & Zimmermann 2007), and learner control in general averaged g = 0.05 across 18 studies (Karich et al. 2014). Designed segments, not a pause button, carry the effect.
How complex the material is. In one experiment with primary-school pupils, learner-paced segments helped only on difficult, high-element-interactivity questions (Hasler et al. 2007); across reviews, multimedia design principles as a family mattered more for complex material (g = 0.70) than simple (g = 0.20) and more for system-paced media (g = 0.41) than learner-paced (g = 0.27) (Noetel et al. 2022). In a video study, easy tasks showed a floor with or without segments (Pimenova et al. 2026).
Prior knowledge, which runs both ways. Rey et al. found learners with high prior knowledge gained more on retention, against the expertise reversal prediction; Spanjers et al. (2011), one experiment with 75 secondary students, found segmented animated worked examples more efficient (equal transfer, lower effort) only at lower prior knowledge. Record prior knowledge and which outcome matters before deciding.

How it sits beside its neighbours: Multimedia Learning holds the general words-with-pictures model and the other design moves (placement, channel, cutting extras, pretraining, signalling); this page is its pacing-and-segmentation decision in depth. Chunking is the general idea of meaningful units; segmenting is its temporal form for transient media. Cognitive Load Management places segmenting among the moves that spread processing over a sequence. Reduced load is the explanatory hypothesis; the observations are test scores, time and rated effort.

Default design, while the dose is untested¶

The syntheses set the direction and the conditions; the claims in this wiki do not set a segment length, a number of segments or a schedule of fading. This default is a labelled proposal: the cut-at-breaks-with-a-pause structure follows the evidence, the timings are untested choices.

Check the condition. Is the material transient (it plays rather than sits on a page)? Does it explain interacting steps? Are the learners new to them? If the material is a static page the learner already reads at their own pace, or the content is simple, segmenting adds little (design principles matter more for complex and system-paced media [~M]).
Cut at conceptual breaks, one step per segment. End each segment where one step of the process, one stage of the causal chain or one sub-procedure ends, never mid-step. As a reference dose, Spanjers et al. cut each animated probability example into 5–7 segments with 2-second pauses (segmentation benefits shrink with expertise [~M]); Mayer's lessons used a "Continue" button after each segment (segmenting principle [+M]). Proposal: segments of about one to three minutes of narration for an explanatory video, and no single video longer than about nine minutes, since median viewing fell below half the video's length past nine minutes in MOOC logs (MOOC engagement declines with video length [+W], associational).
Put a deliberate break between segments. In a learner-paced lesson, a "Continue" control the learner must press; in a system-paced one (a lecture, a video played to a class), a pause of a few seconds or a short spoken transition. Either works (effects confirmed for system-paced segmentation [+M]). Do not substitute a free pause button for designed breaks (flow control was rarely used [~M]).
Budget more time. Segmented lessons take longer (segmentation increases learning time [~M]; no magnitude is printed on the claim). Plan the session's minutes with that in mind rather than squeezing the segmented version into the continuous version's slot.
Optionally, one short prompt at each break (name the step just shown, or predict the next). Untested proposal: no claim in this wiki tests segment-plus-question against segments alone, and the prompt is a learning activity, not an added test; leave it out where the brief does not want one.
Across sessions (untested proposal). Sessions 1–2 on a new topic: designer-set segments, one step each, with a continue control. Middle sessions, once a learner explains a segment's step correctly and answers a changed-condition question about it: longer segments (two or three steps each) and chapter markers the learner may use. Late sessions, once they explain and predict without the segments: a continuous presentation with chapter markers, kept for reference. The fading follows Spanjers et al.'s efficiency result; keep segments longer where the outcome is retention, which Rey et al. found high prior knowledge learners still gained on.
When to move a learner on (untested proposal): after a segment, when they can say what happened in that step and why, not when the segment has played.
Judge it on retention and transfer, as the brief's own assessment measures them, and at a delay where the brief has one: the recorded effects are mostly from immediate tests after short lessons.

Fitting the design to a situation¶

Three facts change the decision most. Ask for them when the brief omits them: is the material transient and complex for these learners (a video or animation explaining interacting steps, or a static page, or simple content)?; who controls the pace (system-paced lecture or broadcast, or learner-paced on a device)?; and how much do the learners already know about this content, and is the outcome retention, transfer or speed?

If the brief says…	Then change…	Basis
A narrated animation or explanatory video of a process, for learners new to it (`novice`)	Cut it at each step into segments with a continue control	Segmenting improves multimedia learning [+M]; segmenting principle [+M] (3 of 3 experiments, median d = 0.98)
Learners already know much of the content, and the goal is retention	Keep the segments; prior knowledge did not remove the retention benefit	High prior knowledge learners benefited more on retention [~M]
Learners already know much of the content and want to move fast; the goal is efficient transfer	Offer longer segments or the continuous version with chapter markers, and check with a changed-condition question	Segmentation benefits shrink with expertise [~M] (efficiency only, one study)
The content is simple, with few interacting steps, or the task is easy	Do not segment for its own sake; spend the time elsewhere	Design principles matter more for complex media [~M] (g = 0.70 vs 0.20, multimedia principles as a family); easy tasks showed a floor [~M]
System-paced delivery: a live lecture, a video played to the class, a broadcast	Build the pauses in: stop at each conceptual break for a few seconds or a short transition	Effects confirmed for system-paced segmentation [+M]
Self-paced online, learners have pause and replay controls	Still cut into designed segments with continue points; do not rely on learners pausing at the right places	Flow control was rarely used and did not track learning [~M]; learner pacing helped on complex material, learner control in general near zero [~M]
Children (primary school)	Segment the animation and keep the pacing control simple (one continue button)	Learner pacing helped primary pupils on difficult questions [~M] (one experiment)
A recorded lecture of 20 minutes or more	Re-cut it into several shorter videos at topic breaks, each a coherent step	Students said long recorded lectures impeded focus [+W]; engagement declines with video length [+W]
Slides or print the learner reads at their own pace	Segmenting in this sense matters less; use one idea per slide or page and headings, and spend design effort elsewhere	untested; the claims here are on transient media, and in one study learners with static pictures navigated far more than with animation (animations appear to pacify learners [~M])
VR, a simulation or an interactive demonstration	Break the walkthrough into stages with a stop between them before the learner acts	untested proposal; the claims here are on video and animation
Phone-only learners	Short segments, each a separate screen or clip with a large continue control	untested proposal
A single session only	Segment within the session and budget extra minutes for the segmented version	Segmentation increases learning time [~M]
A fixed short time budget (e.g. ten minutes a day)	Cut the scope of the session, not the segments: fewer steps per session, each segmented, so the minutes add up	untested proposal, from segmentation increases learning time [~M]
One teacher for many learners	Teacher pauses the shared video at the same breaks for everyone; use the break for a whole-class prompt rather than individual pacing	Effects confirmed for system-paced segmentation [+M]; the prompt is an untested proposal
Learners with ADHD or other attention difficulties	Segment; expect a benefit at least as large as for other learners	Segmentation benefit was larger for the ADHD group [~M] (interaction not significant); segmented videos reduced hesitations with ADHD [+M]
Low literacy or second-language learners, and the video shows spoken plus written text	Segment, and expect the spoken–written duplication to help less once segmented	Spoken–written advantage is smaller for segmented presentations [~M] (g = 0.15 segmented vs 0.53 nonsegmented)
Learners who are deaf or hard of hearing	Caption each segment; short segments make re-watching one step easier	Students said long recorded lectures impeded focus [+W] (one student with a hearing disability valued shorter videos for re-watching)
Learners who are blind or have low vision	Narrate what each segment's animation shows; segment boundaries still fall at the narrated steps	untested proposal
The brief supplies its own video, routine, rubric or assessment	Keep them as given: segment the brief's own material at its breaks, run its assessment as written, and add no test it did not ask for	untested (authoring rule)
The material is private (patient data, internal procedures)	Segmenting needs no logging of viewing; do not add pause or view analytics the brief did not ask for	untested proposal
The goal is a whole task that must stay whole (a full procedure performed end to end)	Segment the demonstration, not the learner's practice: after the segmented walkthrough, the learner performs the whole task	untested proposal
High stakes (certification, safety)	Use segmenting, but do not rely on the recorded effects to carry to a delayed or on-the-job test; check at the horizon that matters	untested; the recorded effects are mostly immediate tests

Observation, state and explanation¶

These rows are proposals, untested with learners as diagnostics.

Response	Plausible explanations	Observation that could change the interpretation
Can recite each segment but cannot explain how the steps connect	Breaks fell mid-step or fragmented the causal chain; segments too short; the test asked for recall	Re-cut one stretch at the step boundaries and ask for a "what happens if this step fails" explanation.
Presses "Continue" immediately every time	Material already known; the break carries no task; disengagement	Ask a changed-condition question at one break; if answered, lengthen segments (as for higher prior knowledge).
Does no better with segments than without	Content too simple; learner already knows it; segments arbitrary rather than conceptual; test too easy	Check element interactivity and prior knowledge; give a harder transfer item before deciding.
Runs out of time in the segmented version	Segmenting lengthens study; the session was planned at continuous-video length	Compare the minutes actually used with the plan; reduce scope per session rather than removing breaks.

Evidence and qualifications¶

Segmenting improves multimedia learning [+M]: Rey et al.'s (2019) meta-analysis of 56 investigations and 88 comparisons; significant segmenting effect, small to medium on retention and transfer, with no pooled value printed in what was read. The claim page says the effects held "specifically" for system-paced segmentation; the quoted abstract says they are confirmed for system-paced segmentation as well, which is how this page reads it.
Segmenting principle: learner-paced segments of narrated animation improve transfer [+M]: Mayer's (2008) review, 3 of 3 experiments, median d = 0.98 on transfer; one research programme, short computer lessons.
Effects confirmed for system-paced segmentation [+M] and segmentation reduces overall cognitive load [+M]: subgroup and outcome findings of the same meta-analysis.
Segmentation increases learning time [~M]: the cost, same meta-analysis, no magnitude printed.
High prior knowledge learners benefited more on retention [~M] against segmentation benefits shrink with expertise [~M]: the meta-analytic moderator is on retention; the experiment (75 Dutch secondary students, animated probability worked examples, 5–7 segments with 2-second pauses) found no interaction on transfer scores, only on effort and efficiency. They are not a direct contradiction, but neither licenses dropping segments for experts on its own.
Learner pacing helped on complex material; learner control in general near zero [~M]: Hasler et al. (2007), primary pupils, day-and-night animation, segmented and stop/play versions beat system-paced continuous only on difficult questions, and the stop/play buttons were rarely used; Karich et al. (2014), 18 studies, g = 0.05.
Design principle effects are moderated by pacing and element interactivity [~M]: Noetel et al. (2022), multimedia principles as a family, not segmenting alone.
Segmentation benefit was larger for the ADHD group [~M] and segmented videos reduced hesitations for participants with ADHD [+M]: Pimenova et al. (2026), a within-subjects study of segmented versus non-segmented programming videos, 27 participants of whom 17 had ADHD, on medium and hard tasks; errors fell with segmentation for both groups (d = 0.85 ADHD, 0.66 control), the group difference was not significant, and medication status did not interact with segmentation [~M]. An arXiv preprint.
Spoken–written advantage is smaller for segmented presentations [~M]: Adesope & Nesbit (2012), a moderator in a redundancy meta-analysis.

Taken together: for transient, complex material and learners new to it, designed segments with a break between them improve retention and transfer by a small-to-medium amount and cost time, whether the learner or the system sets the pauses; free pause controls alone do not do the same job; the benefit is smaller for simple content; how it changes with prior knowledge depends on the outcome. No claim in this wiki sets a segment length or tests segmenting at a delayed horizon.

The evidence by question¶

Each question links a general claim that gathers the studies behind it and the narrower claims bearing on it (eval/synthesis/segmenting.json, scripts/synthesize_claim.py), one row per distinct study, with no averaged verdict.

Does breaking a continuous presentation into segments improve learning over presenting it whole? Segmenting Improves Multimedia Learning [+M]
Who should set the pauses, and do pause controls do the same job as designed segments? Letting learners pace instruction helped on complex material in one experiment, while learner control in general showed almost no effect across 18 studies [~M]
For which learners does segmenting help more or less (prior knowledge, attention)? Learners with high prior knowledge benefit more from segmenting in retention performance [~M]
What does segmenting cost, and where does it add little? Segmentation increases learning time [~M]

Further evidence, not yet read against this model¶

Segmenting multimedia instruction into learner-paced segments improves retention and transfer performance with small-to-medium effects [+M]: the same Rey et al. (2019) finding as the claim cited above; a fold candidate.
Cognitive Load Management [+M]: carries the Rey et al. meta-analysis beside a worked-examples meta-analysis.
Multimedia principles benefit novices [~M]: the expertise reversal argument for multimedia supports in general, with no segmenting study.
Expertise reversal: cognitive load effects for novices decrease, disappear, and can reverse [~M]: the general prediction Rey et al.'s retention moderator runs against.
Animations appear to pacify learners [~M]: navigation counts by format in one small study.
Students pause the video after a few sequences to consult a segment menu [~W]: a classroom design report; learners judged workload by the number of segments.
Guo et al.: short videos, talking heads and Khan-style drawing were more engaging [+W]: engagement, as reported in a review, not learning.
Pretraining principle [+M]: the companion move that front-loads component names before a segmented or continuous lesson.

Objective and learner-valued goal¶

The designer's objective is usually understanding of how a process works, retained and usable on a new case. Ask separately what the learner wants: a learner who wants to find one step quickly (how to reset a device) values chapter markers and a short segment for that step, not a full sequence; a learner revising for an exam values being able to replay one hard step. Neither is served by a single long video, and neither shows by watching to the end that they understood it.

What would revise this model?¶

The expectation should weaken if comparisons with comparable learners and material, holding content equal, find no retention or transfer advantage for segments placed at conceptual breaks, or if the advantage seen on immediate tests disappears at a delay. The prior-knowledge qualification should be revised once a study crosses prior knowledge with both retention and transfer outcomes in one design. A study that varies segment length, or that compares segments with and without a prompt at each break, would replace the untested dose in the default design.

Claims¶

Modality Principle
Redundancy Principle
Signaling Principle
Contiguity Principle (Spatial and Temporal)
Coherence Principle
Pre-training Principle
Multimedia Learning: the umbrella model this decision sits inside
Chunking: meaningful units in general; segmenting is its form for transient media
Cognitive Load Management: spreading processing demand over a sequence
Cognitive Load Theory: the explanatory account of why transient, complex material overloads novices
Worked Examples: Spanjers et al.'s material was segmented animated worked examples
Sequencing: the order of segments and of topics across sessions

Examples¶

Multimedia Learning pattern: the reusable policy that chooses segmentation alongside placement and channel.
Segmenting and Structured/Segmented Instruction: the strategies that carry this principle into an activity.
Interactive Video Quizzing: a prompt at each break (step 5, untested here).
Self-Paced Learning: the element through which learners control pace.
Segment dynamic visualizations with pauses between meaningful segments for novice learners, choosing segment length deliberately

Key Sources¶

Rey, G. D., Beege, M., Nebel, S., Wirzberger, M., Schmitt, T. H., & Schneider, S. (2019). A Meta-analysis of the Segmenting Effect. Educational Psychology Review, 31(2), 389–419. doi:10.1007/s10648-018-9456-4
Mayer, R. E. (2008). Applying the Science of Learning: Evidence-Based Principles for the Design of Multimedia Instruction. American Psychologist. doi:10.1037/0003-066x.63.8.760
Spanjers, I. A. E., Wouters, P., van Gog, T., & van Merriënboer, J. J. G. (2011). An expertise reversal effect of segmentation in learning from animated worked-out examples. Computers in Human Behavior, 27(1), 46–52. doi:10.1016/j.chb.2010.05.011
Hasler, B. S., Kersten, B., & Sweller, J. (2007). Learner control, cognitive load and instructional animation. Applied Cognitive Psychology, 21(6), 713–729. doi:10.1002/acp.1345
Karich, A. C., Burns, M. K., & Maki, K. E. (2014). Updated Meta-Analysis of Learner Control Within Educational Technology. Review of Educational Research, 84(3), 392–410. doi:10.3102/0034654314526064
Noetel, M., Griffith, S., Delaney, O., Harris, N. R., Sanders, T., Parker, P., del Pozo Cruz, B., & Lonsdale, C. (2022). Multimedia design for learning: An overview of reviews with meta-meta-analysis. Review of Educational Research. doi:10.3102/00346543211052329
Adesope, O. O., & Nesbit, J. C. (2012). Verbal Redundancy in Multimedia Learning Environments: A Meta-Analysis. Journal of Educational Psychology. doi:10.1037/a0026147
Pimenova, V., Lee, C., Bhakdibhumi, B., Chu, S., & Begel, A. (2026). Leveling the Playing Field: Temporal Video Segmentation for Individuals with ADHD in Computing Education. arXiv preprint. arXiv:2607.24612
Guttormsen Schär, S., & Zimmermann, P. G. (2007). Investigating Means to Reduce Cognitive Load from Animations: Applying Differentiated Measures of Knowledge Representation. Journal of Research on Technology in Education, 40(1), 64–78. ERIC EJ826063

Page metadata

Field	Value
`type`	`principle`
`id`	`segmenting`
`title`	`Segmenting Principle`
`description`	When a lesson presents complex material as a continuous, transient stream (a narrated animation, a video, a live demonstration) to learners for whom its steps are new, presenting it in meaningful segments with a pause or a continue control between them is expected to improve retention and transfer over one continuous presentation, at the cost of more learning time; meta-analyses test this, and it holds when the system sets the pauses, matters more for complex than for simple material, and its relation to prior knowledge is mixed.
`status`	`review`
`canonical`	`True`
`generated`	by `claude/unspecified` on `2026-10-14`
`trust`	unverified — no one has confirmed this page's content
`sources`	9 — `rey-2019`, `mayer-2008`, `spanjers-2011`, `hasler-2007`, `karich-2014`, `noetel-2022`, `adesope-nesbit-2012`, `pimenova-2026`, `guttormsen-schar-2007`

Frontmatter is the page's machine-readable record — see the Schema & Guide.