Principle
Keep the explanation close, but make it earn the space
A distant legend makes the reader solve two problems at once: understand the idea and remember which sentence belongs to which part. Bringing related words and visuals together often helps. But proximity is not a rule to pack every sentence into the picture. The evidence is strongest when both sources are needed for understanding, and it becomes much less tidy when the added text is redundant, the page is crowded, or the task makes the correspondence obvious already.
Before you use it
- Use when
- Readers keep looking between a diagram and a distant legend.
Keep this limit in mind
- Preserve accessible reading order and readable size. Crowding a diagram can create a different comprehension problem.
Current editorial reading
The reliable result is integration, not a magic distance
Schroeder and Cenkci combined 58 independent comparisons with 2,426 participants and found an overall advantage for integrated over spatially separated designs (g = 0.63). That is a meaningful average effect across the included learning studies. It supports the basic move behind this element: when a diagram and its explanation depend on each other, do not make the learner repeatedly search across the page to connect them. The result does not tell us that a label must sit a fixed number of pixels from its target, or that every close layout is automatically better.
The usual cognitive-load story is more uncertain than the design rule
The common explanation says that separation wastes working-memory capacity because the learner must search, match and mentally integrate two sources. A systematic review of 41 comparisons found little support for the prediction that integrated designs reliably lower measured cognitive load. Pouw and colleagues then varied spatial distance directly: greater distance changed integration time, eye movements or a secondary working-memory measure in different experiments, but their more realistic learning experiment found no learning-outcome difference. A better editorial interpretation is modest: proximity can make correspondence easier to establish, while the exact mechanism and the conditions that turn easier correspondence into better learning are still being worked out.
Near can become noise when the extra text is unnecessary
Rop and colleagues tested diagrams accompanied by text that repeated information learners could already infer from the diagram. In their second experiment, integrating that unnecessary text produced worse learning than keeping it separate, and eye tracking showed more attempts to integrate the unnecessary text with the diagram. This is an important boundary: closeness is useful when two sources need to be combined, not when one source merely restates the other. A 2024 experiment with 146 university students adds another caution: it found no significant effects of integrated versus split layout, presentation size, or their interaction on retention, comprehension or reported extraneous load. The broad effect is real enough to guide design, but individual layouts can produce null or reversed results.
Use proximity to remove a lookup, not to remove structure
For a screenshot, process map or technical diagram, first identify the exact correspondence the reader must make. Put a short label, value or explanation close to that component when doing so removes a real lookup. Keep longer reasoning in normal reading order and connect it with a clear callout, line, highlight or stable reference. Then test the page at the size people will actually use. If the text becomes tiny, overlaps the visual, or forces the reader to scan a dense cluster, move some explanation out again. The goal is not minimum physical distance; it is minimum avoidable matching work without sacrificing readability or accessibility.
Should every label sit inside the diagram?
No. The evidence compares designs that make related information easier or harder to integrate; it does not require text to be embedded inside the image. A nearby callout, aligned row or clearly connected block can preserve the correspondence while keeping text selectable, readable and compatible with a sensible reading order. If embedding the text makes the diagram crowded, the cure has started creating its own problem.
Is a numbered legend always a bad design?
No. A legend is costly when readers must repeatedly bounce between a number and a distant explanation to understand the main task. It can be perfectly reasonable for reference information that is rarely consulted, or when the mapping is obvious and stable. Before replacing it, watch the actual reading task: if people keep losing the referent, reduce the distance or strengthen the visual connection; if they do not, proximity may solve a problem that is not there.
What should I do on a small screen?
Do not preserve desktop proximity by shrinking everything. Keep readable text and a clear reading order first. On mobile, it may be better to pair one component with its explanation at a time, stack short callouts beside or below the relevant visual segment, or let the reader reveal detail without losing the target. The 2024 presentation-size study found no simple size rule, so treat the final mobile layout as something to test rather than as a deduction from the general contiguity effect.
Take it with you
Put related words and visuals close enough that the reader does not have to perform a needless lookup. Stop before proximity turns into crowding or redundancy. The evidence supports integrated designs on average, but newer and more diagnostic studies show that distance alone is not the mechanism and that unnecessary text can make integration actively unhelpful.
Sources and evidence (5)
What we checked
A random-effects meta-analysis of 58 independent comparisons with 2,426 participants found an overall learning advantage for integrated over spatially separated designs of g = 0.63.
This supports the general spatial-contiguity principle across the included learning studies. It does not identify one optimal distance or prove that every integrated layout will outperform every separated layout.
Spatial Contiguity and Spatial Split-Attention Effects in Multimedia Learning Environments: a Meta-Analysis · Abstract; overall random-effects meta-analysis
A systematic review of 41 comparisons found that cognitive-load measures largely did not support the claim that integrated designs work by reliably lowering extraneous cognitive load.
The learning advantage can still exist even if the standard cognitive-load explanation is incomplete. The review proposed facilitation of integrative processing as an alternative account.
Do Measures of Cognitive Load Explain the Spatial Split-Attention Principle in Multimedia Learning Environments? A Systematic Review · Abstract; analysis of 41 comparisons
Across three experiments, increasing spatial distance changed some integration-process measures, but the more ecologically valid learning experiment did not show a learning-outcome difference between the spatial conditions.
This suggests that spatial distance by itself is neither the only nor a sufficient condition for a learning effect.
The cognitive basis for the split-attention effect · Abstract; Experiments 1–3
When explanatory text was unnecessary because the diagram already conveyed the information, one experiment found worse learning for integrated text than separated text and more integration attempts in the integrated condition.
This is a direct boundary on the practical rule: integration can be counterproductive when it makes learners process redundant material they did not need.
The effect of layout and pacing on learning from diagrams with unnecessary text · Summary; Experiment 2; General Discussion
A 2024 experiment with 146 university students found no significant effects of integrated versus split format, presentation size, or their interaction on retention, comprehension, or reported extraneous cognitive load.
The null result does not erase the meta-analytic average, but it reinforces that the effect has boundary conditions and should not be converted into a universal pixel-level layout rule.
Does presentation size of instructional materials influence the split-attention effect? · Abstract; Results
Editorial sources (5)
- Spatial Contiguity and Spatial Split-Attention Effects in Multimedia Learning Environments: a Meta-Analysis — Noah L. Schroeder, Ada T. Cenkci
- Do Measures of Cognitive Load Explain the Spatial Split-Attention Principle in Multimedia Learning Environments? A Systematic Review — Noah L. Schroeder, Ada T. Cenkci
- The cognitive basis for the split-attention effect — Wim Pouw, Gertjan Rop, Bjorn de Koning, Fred Paas
- The effect of layout and pacing on learning from diagrams with unnecessary text — Gertjan Rop, Anne Schüler, Peter P. J. L. Verkoeijen, Katharina Scheiter, Tamara Van Gog
- Does presentation size of instructional materials influence the split-attention effect? — Shirong Zhang, Bjorn B. de Koning, Fred Paas
A small scene. Your next move.
Try this idea.
Make a move, see what changes, then bring it to your own situation.
Save and practise this idea
My learning
Read. Recall. Repeat.
Save this idea and return to it on any device.
Keep exploring
Connected ideas
You are readingKeep the explanation close, but make it earn the space
Ideas that link here
These ideas link to this one. Check their conditions and limits before use.
Useful with this ideaDraw the process while you explain it
Original edition evidence and sources
These records belong to the original released edition. The current editorial reading has its own evidence above; neither set silently replaces the other.
The meta-analysis found an overall learning benefit for integrated text-and-diagram designs across its included comparisons.
It does not imply that visual crowding or tiny text is acceptable.
Spatial Contiguity and Spatial Split-Attention Effects in Multimedia Learning Environments: a Meta-Analysis · Abstract, overall comparison.
Original edition, downloads and provenance
The reading above is editorial version 1.0.0, bound to this exact source edition. The original data and saved downloads retain that edition and the separate editorial layer. They may contain earlier wording; use the current reading above when interpreting it.
Version 0.1.0
No review details are supplied for this edition.
Read the original edition and overlay as JSON · Release details