Skip to content

Contiguity Principle (Spatial and Temporal)

Principle · All principles
Evidence · 30 claims (16 for, 14 mixed) · 24 studies (8 review, 6 causal, 5 quant-synthesis, 2 associational, 2 design, 1 theoretical), q1–q4 · 7 of 24 report an effect size · 25 claims rest on one study

Conditional relationship

This page owns one decision inside Multimedia Learning: where and when the words go, once a picture has a job. When an explanation's words and a picture refer to each other, so that neither can be understood alone (a diagram and its labels, a geometry figure and the proof steps, an animation and the narration describing it), and the learners are new to the material, placing each word or sentence next to the part of the picture it describes (spatial contiguity) and presenting the narration at the same moment as the matching animation (temporal contiguity) is expected to improve learning over designs that put them apart on the page or screen, or present one after the other. The split-attention effect is the same relationship from the learner's side: separated sources make the learner search, hold one source in mind and match it to the other, and learning suffers.

Claims in this wiki test the relationship directly, in meta-analyses, and report its boundaries:

  • Size and the two forms. Across 50 randomised between-group effects (2,375 students), integrating separated information gave d = 0.85 (Ginns 2006); a later meta-analysis of 58 comparisons (n = 2,426) found g = 0.63 for spatial integration (Schroeder & Cenkci 2018). Spatial (d = 0.72) and temporal (d = 0.78) effects did not differ significantly, which is not a test that they are equal.
  • Complexity of the material. The effect was d = 0.78 for high element-interactivity material and d = 0.28, with an interval crossing zero, for low (Ginns 2006). Across design principles generally, effects were g = 0.70 in complex and 0.20 in simple media (Noetel et al. 2022).
  • Pacing. Across design principles, effects were larger in system-paced media such as lectures (g = 0.41) than in learner-paced media such as websites (g = 0.27).
  • Expertise. Integrated text that novice electrical trainees needed became redundant for experienced trainees, who did better with the diagram alone (Kalyuga et al. 1998, as reported by two reviews). Noetel et al. found prior knowledge did not significantly moderate design-principle effects as a whole (p = 0.14). The two disagree in grain: one experiment programme on one principle, one moderator test across all principles. Plan for the reversal and check for it.
  • Medium and testing. Static (print) and dynamic (animation) presentations did not differ reliably in effect. Effects were larger under one-to-one testing (d = .89) than group testing (d = .65), and the group-tested effect is still large enough to matter in a classroom.
  • Mechanism. The usual explanation is extraneous load from search and matching. A systematic review of 41 comparisons found that cognitive-load measures mostly did not differ between integrated and separated designs (Schroeder & Cenkci 2020): the learning benefit stands, the load explanation is not established.

How this sits inside its neighbours: Multimedia Learning decides whether a picture is added and holds the umbrella model; this page decides placement and timing. Speaking the words instead of printing them (the modality principle) is the other way to remove the search a separated layout causes, and it has its own pacing limit. Cutting words that repeat what the picture already shows (redundancy and coherence) is the opposite move to integration, and which one fits depends on whether the words are needed at all. Cognitive Load Management and Cognitive Load Theory hold the wider model that integration is one lever of.

Default design, while the dose is untested

The meta-analyses establish the direction, its size and the conditions where it is larger; they do not set a dose or a schedule for withdrawing integration. Integration is a change to layout and timing, not added content, so it should not lengthen a lesson. The progression across sessions below is a labelled proposal.

  1. Audit the materials (before session 1, about 30 minutes for a lesson's materials). List every pair of sources that refer to each other: legends and figures, captions on another page, a worked example's text and its diagram, instructions in a manual for an interface on screen, narration that runs before or after the animation. For each, ask whether either source makes sense alone. Integrate only the pairs where neither does. Basis: contiguity across space and time [+M]; the "neither alone" test is the condition the integration studies used, and the rule for pairs that do stand alone follows the redundancy reversal below.
  2. Integrate in space. Put each label on, or touching, the part it names; put each sentence beside the region of the figure it explains; keep figure and text on the same page, slide or screen; replace numbered legends with labels in place. Basis: split attention degrades learning [+M]; Mayer's spatial tests, 5 of 5 [+M].
  3. Integrate in time. In video, narrated animation or a live lecture with slides, say each sentence while its part of the picture is on screen and pointed to; never explain first and show later, or show first and explain later. Basis: Mayer's temporal tests, 8 of 8 [+M]; spatial and temporal effects of similar size [~M].
  4. Where integration is impossible (lab equipment with a separate manual, two screens, a printed handout for a projected diagram), speak the explanation while learners look at the object, in system-paced settings; in self-paced settings keep the printed words but put them in the learner's line of sight beside the object. Basis: modality effect [+S] as the alternative, limited by pacing reversing it [~M].
  5. Sessions 1–2 of a topic (learners new to it). Use fully integrated materials. Judge on whether learners can explain how the parts relate, not on whether they found the page easy. Basis: integration effects for novices (step 2); load ratings and learning come apart [~S].
  6. From session 3, or once a learner can explain a figure unaided, check for reversal. Give the same kind of figure with no explanatory text and a changed-condition question ("what happens if this part fails?"). A learner who answers well moves to diagram-only or key-label versions of that topic's materials; a learner who does not stays with integrated versions. Re-check at each new topic, since expertise is topic by topic. Basis: integration reverses into redundancy for experienced trainees [~M]; the session count and the check are an untested proposal.
  7. Judge the change on a transfer question about the content (explain, predict, troubleshoot), and on a delayed test where the course has one. The effects above come mostly from short lessons tested soon after; a delayed horizon is not shown by them.

Fitting the design to a situation

Three facts change the decision most. Ask for them when the brief omits them: do the words and the picture need each other (would the diagram make sense without its text, and the text without the diagram)?; how much do these learners already know about this content and this kind of figure?; and who controls the pace, in what medium (print or learner-paced screens, system-paced video or a live lecture)?

If the brief says… Then change… Basis
Novice learners; a diagram with a separate legend, or text on another page or screen Put labels on the diagram parts and the text beside the region it explains, all on one page or screen Split attention degrades learning [+M] (g = 0.63, 58 comparisons)
The material has many interacting parts (a circuit, a body system, a proof) Integrate first here; this is where the effect is largest Integration helps most with complex material [+M] (d = 0.78 high vs 0.28 low)
The material is simple, with few interacting parts Integration is low priority; fix it if cheap, spend design time elsewhere Integration helps most with complex material [~M]; design principles matter less in simple media [~M]
Learners are advanced on this content (advanced), or already read this kind of diagram fluently Drop the explanatory text and give the diagram alone or with key labels; check with a changed-condition question Integration reverses into redundancy for experienced trainees [-M]; guidance that helps novices can become redundant [~S]; qualified by prior knowledge did not moderate design-principle effects overall [~M]
The text restates what the picture already shows, or each source makes sense alone Do not integrate; cut the redundant source, or keep it apart where learners can ignore it Redundant integrated notes retarded more knowledgeable learners [~M]
System-paced video or narrated animation Synchronise each sentence of narration with its part of the animation; no "explain, then show" Mayer's temporal tests, 8 of 8 [+M]; design principles matter more when system-paced [+M]
A live lecture with slides Put labels on the slide's figure; talk about each part while it is shown and pointed to, not before or after Mayer's spatial and temporal tests [+M], carried from computer and paper lessons to live lectures, which they did not test
Print: a textbook, worksheet or handout Labels on the figure; text on the same page as its figure; never "see Figure 3 overleaf" Static and dynamic presentations did not differ [~M]; Mayer's paper-based tests [+M]
Self-paced online course (online-self-paced), learners can pause and scroll Integrate anyway; expect a smaller gain than in system-paced media Design principles matter less when learner-paced [~M] (g = 0.27 vs 0.41)
Text and object cannot share a space (equipment and a manual, two monitors) System-paced: speak the explanation while learners look at the object. Self-paced: keep printed words, placed in the line of sight Modality effect [+S]; pacing reverses it [~M] (d = 0.93 system-paced, −0.14 self-paced)
VR, AR or a projected guide on a physical task Co-locate guidance with the work where the task is spatial (drawing, assembly); symbolic instructions can stay on a screen Co-located guidance matters most when instruction and task share a spatial frame [~W] (authors' interpretation of one study)
Phone-only, small screen Put labels on the figure, or reveal a label beside its part on tap; never put the explanation on a different screen from the figure untested proposal
Worked examples in mathematics or science (geometry, circuits) Write each step's explanation into the diagram where it applies, not in a block beside it Integrated worked examples restore their advantage [+M]
Children in primary school Use the same integrated layouts; check understanding with a short "what happens if" question Education level did not consistently moderate design-principle effects [~M]; the contiguity meta-analyses here are not broken down by age
Learners read in a second language, or have low literacy Integrate short explanatory notes into the text or figure; add captions to second-language video Integrated notes helped low language-competence learners [+M]; captioning second-language video [+M]
Learners are deaf or hard of hearing Keep printed words placed beside the picture, not narration; give pause control untested proposal; the temporal and modality evidence assumes learners hear the narration
Learners are blind or have low vision Keep the description of each part with that part: in a tactile diagram, braille labels on the part; in audio description, describe each part as it is presented untested proposal
A single session, or a fixed short time budget Integrate the existing materials; it moves words rather than adding them, so it costs no learner time untested proposal; the effects are on immediate tests of short lessons [+M]
One teacher with a large class, or group testing Integrate the shared materials once; the benefit holds in group-tested settings Effects larger one-to-one but still large under group testing [~M] (d = .89 vs .65)
The brief supplies its own materials, routine, rubric or assessment Keep them as given; re-lay out the existing figures and text and retime the narration, add no test the brief did not ask for, and judge with the brief's own assessment untested proposal
The material is private (patients' images, learners' own work) Annotate inside the same view, as a layer on the image, rather than in a separate document that must be shared alongside it untested proposal
A whole task kept whole (a real device, a full case) Integrate the guidance into the task environment (labels on the device or interface, prompts at the step) instead of a separate manual Integrated worked examples [+M]; extending it to whole tasks is untested
High stakes: a certification exam, clinical or safety training Do not rely on the effect alone to carry to delayed or on-the-job performance; test that performance directly untested; the effects here were measured soon after short lessons
Success is to be judged by learners' effort or load ratings Judge by an explanation or transfer question instead Load measures did not explain the integration benefit [~S]

Observation, state and explanation

These rows are proposals, untested with learners as diagnostics. Where a learner looked, how long they took and how hard they say it was are observations; "split attention" and "extraneous load" are explanations, and the second is not established by the measures in this wiki (Schroeder & Cenkci 2020).

Response Plausible explanations Observation that could change the interpretation
Looks back and forth between figure and text; slow; cannot say how the parts relate Separated sources force search and matching; the figure itself does not show the relation; the words are too hard to read Re-present one section integrated; if the explanation improves, integrate the rest. If not, check reading access and whether the figure shows the relation.
Explains well with integrated material but not from a plain diagram later Integration did its job and the learner now depends on it; the target requires reading unlabelled diagrams Decide whether the objective needs unaided reading; if so, fade labels across sessions and check with a plain diagram.
An experienced learner is slower or worse with integrated material The integrated text is redundant for them and hard to ignore Give the diagram alone with a changed-condition question; if they do as well or better, drop the text for them.
Misses the narration's point in a video Narration runs before or after the relevant visual; pace too fast; language or hearing access Check the timing of narration against the animation; offer pause control or captions placed by the figure.

Evidence and qualifications

Taken together: for learners new to complex material, words placed with and timed to the pictures they explain beat separated or successive presentation by a large average margin, in print and in animation, with spatial and temporal forms of similar size. The effect is smaller for simple material and for learner-paced media, and is reported to reverse for experienced learners, though one meta-meta-analytic moderator test did not find prior knowledge to matter across principles. No claim in this wiki tests a schedule for withdrawing integration, a delayed horizon, or learners with sensory impairments.

The evidence by question

Further evidence, not yet read against this model

Objective and learner-valued goal

The designer's objective is usually an explanation of how a system works, or the ability to use a diagram on a new problem. Ask what the learner needs: a technician who must read unlabelled schematics on the job needs integration early and its withdrawal later, while a learner who only needs to recognise parts may need neither. Ask also whether the figure will be available when the capability is used.

What would revise this model?

The expectation should weaken if new comparisons with novice learners on complex material, holding content and time equal, find no advantage for integrated or synchronised designs at a delayed test, or if the advantage disappears when separated designs give learners full control of pace. The expertise boundary needs a direct test: integrated against diagram-only versions with learners measured for prior knowledge of the specific content, on the same transfer test. If integration helps learners who report no difference in effort, that would be consistent with Schroeder & Cenkci (2020) and would revise the mechanism, not the design.

Claims

Examples

Key Sources

  • Ginns, P. (2006). Integrating information: A meta-analysis of the spatial contiguity and temporal contiguity effects. Learning and Instruction, 16(6), 511–525. doi:10.1016/j.learninstruc.2006.10.001
  • Schroeder, N. L., & Cenkci, A. T. (2018). Spatial contiguity and spatial split-attention effects in multimedia learning environments: A meta-analysis. Educational Psychology Review, 30(3), 679–701. doi:10.1007/s10648-018-9435-9
  • Schroeder, N. L., & Cenkci, A. T. (2020). Do measures of cognitive load explain the spatial split-attention principle in multimedia learning environments? A systematic review. Journal of Educational Psychology, 112(2), 254–270. doi:10.1037/edu0000372
  • 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
  • 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
  • Kalyuga, S., Ayres, P., Chandler, P., & Sweller, J. (2003). The expertise reversal effect. Educational Psychologist, 38(1), 23–31. doi:10.1207/s15326985ep3801_4
  • Kalyuga, S., Chandler, P., & Sweller, J. (1998). Levels of expertise and instructional design. Human Factors, 40(1), 1–17. doi:10.1518/001872098779480587
  • Atkinson, R. K., Derry, S. J., Renkl, A., & Wortham, D. (2000). Learning from examples: Instructional principles from the worked examples research. Review of Educational Research, 70(2), 181–214. doi:10.3102/00346543070002181
  • Ginns, P. (2005). Meta-analysis of the modality effect. Learning and Instruction, 15(4), 313–331. doi:10.1016/j.learninstruc.2005.07.001
Page metadata
Field Value
type principle
id contiguity
title Contiguity Principle (Spatial and Temporal)
description When words and a picture explain one thing and neither makes sense without the other, placing each word next to the part of the picture it describes, and presenting narration at the same time as the animation it describes, is expected to improve learning over separated or successive presentation for learners new to the material; meta-analyses test this (spatial and temporal effects of similar size), with larger effects for complex material and system-paced media and a reversal reported for experienced learners.
status review
canonical True
generated by claude/unspecified on 2026-10-14
trust unverified — no one has confirmed this page's content
sources 9 — ginns-2006, schroeder-cenkci-2018, schroeder-cenkci-2020, noetel-2022, mayer-2008, kalyuga-et-al-2003, kalyuga-et-al-1998, atkinson-et-al-2000, ginns-2005

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