Skip to content

Elaboration Theory

Pattern · All patterns
Evidence · 12 claims (3 for, 9 mixed) · 14 studies (4 causal, 4 quant-synthesis, 3 review, 2 theoretical, 1 associational), q2–q4 · 2 of 14 report an effect size · 9 claims rest on one study

Description and scope

A reusable course- or unit-level sequencing policy for a body of content whose ideas are organised (concepts, principles, procedures that depend on each other): begin with an epitome, a simplified but complete version of the whole that carries its few most general and inclusive ideas; then elaborate it level by level toward detail, special cases and complexity; and at the end of each level reconnect the new detail to the whole by summary and synthesis, before using the growing structure on tasks. It instantiates the Sequencing principle and applies Clear Structure: the epitome and the syntheses make the content's organisation explicit, and that page's model (signals and organizers help learners without an organisation of their own, qualified by structural match, prior knowledge and outcome) is the relationship this pattern relies on at each step. Chunking owns the narrower question of how large one unit is and where its boundaries fall; this page owns the order in which units of increasing detail arrive and how they are tied back to the whole.

4C/ID sits beside this pattern rather than inside it. Both move from simple to complex, but 4C/ID sequences whole tasks in task classes for a complex skill, with support fading inside each class; Elaboration Theory sequences content from a general overview to detail. Where the objective is whole-task performance (complex-skill), use 4C/ID and treat its supportive information as the place an elaboration sequence may live.

No claim in this wiki tests the elaboration sequence itself (epitome first, levels of elaboration, summarizers and synthesizers) against another order. The nearest evidence is one report of three small experiments on how a hierarchical organisation of the same physics knowledge, with general ideas above detail, compares with a single-level one. The study configurations below are evidence about that; the sequence and the response-dependent policy are untested design proposals, kept from the earlier page where it gave usable guidance and labelled as such.

Inputs: establish the design brief

Field Record or elicit
Learner and starting observation Before any instruction, ask the learner to sketch or list how they think the domain is organised (the main ideas and how they relate) and to place two or three specific details in it. Record what they place correctly, what they leave out, and whether they hold an organisation already. Prior knowledge here is novice, intermediate or advanced for this content's organisation, not for the wider field. Preserve unknowns.
Content and its organisation The candidate most general ideas (the epitome), the kind of organisation (conceptual: kinds and parts; procedural: steps and decisions; theoretical: causes and principles), which details depend on which, and the element interactivity of each level (low or high). These are design judgements; record their basis.
Context and activities Setting (classroom, online-self-paced, online-instructor-led, workplace-clinical), duration (single-session, days-weeks, term-plus), whether a teacher is present, whether a persistent map or outline can stay visible, language of instruction and access needs.
Learner-valued goal Ask what the learner wants to be able to do with the content and why. Record divergence from the designer's objective; it changes which ideas belong at the top of the epitome.
Designer objective The knowledge type (concept, principle, procedure; verbal-association alone is a poor fit), and the tasks learners will later face: tasks that use the general ideas (explaining, deriving, problem solving) or tasks that rest on single details. The study below found the organisation helped the tasks that matched it.
Outcome A representative task scored by an agreed rubric, kept apart from a recall of details; immediate-performance, delayed-retention and near-transfer or far-transfer as separate outcomes, with the horizon stated.

A request to "use Elaboration Theory" or "sequence from simple to complex" does not supply these inputs. Ask which tasks the learners will face before choosing the epitome's top level, and ask what the learners already hold before deciding how long the epitome needs to be.

Sequence and conditional policy

Every step is an untested proposal from the earlier page unless a claim is named.

  1. Choose the epitome from the target tasks. Pick the few most general ideas that the tasks learners will later face depend on, and state them as a simplified whole, not a table of contents. As a starting size, three to five top-level ideas (untested proposal; the figure is the chunking [~W] claim's estimate of what novel material can hold in memory tasks, not a tested epitome size). Which ideas go at the top changes which tasks the organisation helps (organisation–task match [~M]). If two kinds of task matter, consider two views of the same content rather than one epitome that favours one.
  2. Elicit, then present the epitome. Use the elicitation in the brief (sketch the organisation, place details). If the learner already explains the whole correctly, shorten the epitome to a check and move to the level they need. If they cannot follow the epitome at all because its ideas interact too much, pre-teach the one or two elements it depends on first (isolated parts first, for novices [~M]), then return to it.
  3. Elaborate one level at a time. Each level adds detail, conditions or special cases to one part of the epitome and says which part it elaborates. A level should be small enough to finish in one session or one study block, with a visible map showing where it sits. Use analogies to familiar content where learners lack the base for the new idea (untested here).
  4. Synthesize at the end of each level, with the learner doing it. Ask the learner to mark the new detail on the overview, arrange items into the structure, or explain how the level changes the whole, rather than only reading or writing a summary (tasks that make learners explicate the organisation [+M]). A teacher-written summary at the end of each level is the minimum.
  5. Apply, then reobserve. After each level or two, give a task that uses the structure built so far. Check detail placed low in the structure separately, since in the study below it was recalled least (recall by hierarchy level [~M]). Recheck at a stated delay (the study used two weeks).
Observation Discriminating probe Proposed action, with its uncertainty
Restates the epitome but cannot place a new detail in it Give the detail and ask which top-level idea it elaborates and why; compare with a detail from the previous level. If both fail, the overview is held as words, not as a structure: add a synthesis task where the learner arranges items into the map. If only the new detail fails, the level may be too large or its link to the whole unstated. Neither reading is validated.
Top-level ideas recalled, lower-level detail lost after a delay Compare free recall with recall cued by the top-level idea. If cueing restores the detail, add retrieval of details from their parent idea; if not, re-teach the level. Loss of low-level detail is the expected direction in the study below, not by itself a sign of failure.
Cannot follow the epitome itself Present one of its ideas on its own; ask what it means and how it relates to one other idea. If isolated ideas are understood, simplify the epitome or add an analogy; if not, pre-teach the elements first (tested for novices on interacting elements in one experiment; untested as an entry to an elaboration sequence).
Solves one kind of task but not another Give matched tasks of each kind; ask which general idea each draws on. The organisation may favour one kind of task. Add a second organisation or synthesis that puts the other kind's ideas on top, provisionally.
Already holds a full organisation (draws the map correctly before instruction) Ask for an explanation of a relation two levels down. Skip or shorten the epitome and start at the level where explanation fails; a full overview for such a learner is redundant and may cost time. The exact point is not established.

Fitting the design to a situation

The three facts that most change the decision:

  • Whether learners already hold an organisation for this content. A learner without one needs the epitome and explicit syntheses; a learner who has one needs a quick check and the level they lack. If unstated, ask: can they sketch how the main ideas relate and place a detail, unaided, before instruction?
  • Which tasks they will face. The organisation helped the tasks that matched the ideas at its top and showed no detected benefit on tasks resting on single details. If unstated, ask: what will they be asked to do with this content, explain and solve, or recall specific details?
  • How much the ideas interact, against the learner's prior knowledge. If the epitome's ideas must be held together and the learner knows none of them, an overview first may not be followable. If unstated, ask: can one idea be understood without the others?
If the brief says… Then change… Basis
Learners are novice and the content is high in element interactivity Keep the epitome very short (two or three ideas); pre-teach the one or two elements it depends on in isolation, then present the epitome; make each level small and end it with a learner-built synthesis. Isolated parts first, for novices [~M] (one experiment; sample and horizon not reported on the claim page)
Learners are intermediate or advanced for this content Replace the epitome with the elicitation sketch; start at the first level they cannot explain; drop syntheses they can already produce. untested proposal; advance organizers [~M] found no special benefit for low-knowledge learners, so neither direction of an expertise effect is established for overviews
Learners have weaker prior attainment in the subject Do not rely on the hierarchy alone: add learner-built syntheses at every level, practice tasks at each level, and a check on the top-level ideas before moving down. Hierarchical organisation did not help low-ability students [~M] (college physics volunteers; untested in other groups)
Learners are child or adolescent Make the epitome concrete (a picture, a story or a model of the whole), keep levels to one lesson each, and have the teacher lead the synthesis aloud before asking learners to arrange items. untested proposal; the study tested college and advanced high-school physics students, untested with children
The goal is verbal-association (facts, vocabulary, names) Do not build an elaboration sequence; use a plain grouping and retrieval practice instead. No detected effect on tasks resting on single details [~M] (a non-significant difference, not equivalence)
The goal is a complex-skill (whole-task performance) Sequence whole tasks with 4C/ID and use an elaboration sequence only for the supportive information around them. untested proposal
Assessment is high-stakes and uses one kind of task (e.g. deductive problems, or historical and contextual questions) Put the ideas that task draws on at the top of the epitome and at the top of each synthesis; assess a matched task, not a recall of the outline. Organisation–task match [~M] (20 high-school physics students, two-week test)
online-self-paced, no teacher present Keep the epitome as a persistent map on every page, with the current level highlighted; make each synthesis an activity the system checks (arrange items into the map, mark the new detail), not a passage to read. Learners explicating the organisation [+M] for the synthesis activity (same 20-student high-school experiment; untested online); the persistent map is an untested proposal
Time is one session (single-session) rather than a unit or course Present the epitome and one level of elaboration only, end with one synthesis task and one application task; leave further levels for later sessions rather than compressing them. untested proposal
Some learners study in a second language, or need accessible formats State the epitome in short sentences with a labelled diagram; keep the same terms for the same ideas at every level; offer the synthesis as an arrange-or-label task rather than a written summary. untested proposal

Five rows rest on a claim (each tested in one small study, or one experiment, in a population the row names); the other five, including the persistent map in the self-paced row, are untested proposals.

Choosing configurations from evidence

Six of the claims below come from one report (Eylon & Reif, 1979, three experiments with physics students), so they are one research programme, not replications. None has been checked against its sources, and none tests an elaboration sequence across a course.

  • Hierarchical against single-level organisation. A hierarchical organisation improved complex-task performance [+M]: 36 paid volunteers from an introductory college physics course, blocked by ability and randomly assigned to a hierarchical organisation of the knowledge or a single-level organisation presented once or twice; the hierarchical group scored higher on composite complex recall and problem-solving tasks (F(2,27) = 10.84). The horizon is not reported on the claim page, nor the order in which the levels were presented, so it tests the organisation an epitome-plus-elaboration aims at, not the sequence. In the same experiment, tasks resting only on single detailed steps [~M] showed no significant difference (equivalence not tested), and low-ability students showed no significant benefit [~M] while medium-ability students did; in a third experiment (eight volunteers) higher-ability students recalled more overall.
  • What sits low is recalled least. Information at higher levels was recalled better than information at lower levels [~M]: in free-recall summaries two weeks later, recall fell from the highest to the lowest level (one group's figures: 50%, 50%, 27% and about 5% at the lowest level). It supports placing what matters most near the top and checking detail separately; it does not show that detail is learned worse than it would be without the hierarchy.
  • Match the top of the organisation to the tasks, and have learners work it. Organisation helps the tasks that match it [~M]: 20 students in an advanced high-school physics class, blocked by ability and randomly assigned to an organisation with deductive or with historical information at the higher levels, tested two weeks later; each group did better on the tasks matching its organisation, with no effect size printed. In the same experiment, acquisition tasks in which learners marked organisational charts and arranged items into a chart [+M] outperformed writing a structured summary of the same well-organised materials for the same time (F(1,12) = 5.96). This is the nearest test of an active synthesizer; it compares two learner activities, not synthesis against none.
  • When the whole is too much at first. Isolated parts before the integrated whole [~M]: in one experiment on tasks high in element interactivity, novices did better when first given isolated parts before the integrated whole (sample, outcome and horizon not reported on the claim page; not yet checked against its sources). It limits the policy of always starting with a whole epitome; it was not a test of an elaboration sequence.
  • Overviews before instruction, in general. Advance organizers [~M]: two older meta-analyses found small average benefits (a mean effect of 0.21 across 110 effect sizes in one, unweighted and without significance tests), larger for high-ability than low-ability learners, and no special benefit for low-ability or low-knowledge learners in the other; a narrative review argues they help only in appropriate situations with suitable outcome measures. An epitome is a richer object than an organizer, so this is a neighbouring finding: it warns against assuming that an overview helps weaker learners most. One of its three entries has been checked against its source.

Do not combine these into a predicted effect for the pattern. They differ in learners (college and high-school physics students, novices on complex tasks, mixed groups in old meta-analyses), comparator and outcome, and none measured a course built as an elaboration sequence against one built another way.

The evidence by question

Each question below links a general claim that gathers the studies behind it into one table, one row per distinct study, with each study's design and effect as recorded and no averaged verdict.

Further evidence, not yet read against this model

Claims this page cited before it was rewritten as a conditional model. They are evidence about the relationship, but none has yet been re-read against the model above; each says how far its own sources have been checked.

Illustrative design instance and observation record

Illustration only, invented, not a tested course. A regional water authority wants volunteer river monitors (adults, adult-workplace, mixed prior knowledge, some studying in a second language) to understand how their catchment works; the volunteers say they want to be able to explain to neighbours why the river floods. The course is online-self-paced over days-weeks, with no teacher. The table changed the design in four places. Because the volunteers' valued task is explaining causes, the epitome's top level is causal (rain, land, channel, people) rather than the authority's list of monitoring procedures (task-match row). The epitome is a labelled diagram kept on every page, and each of four levels ends with an arrange-the-items synthesis the system checks (self-paced and language rows). Volunteers who draw the catchment correctly in the opening sketch start at the second level (expertise row). Species names, a verbal-association goal, are taught as a separate retrieval set, not as a level (facts row). The outcome is a delayed explanation of a flood scenario scored by a short rubric, kept apart from a quiz on details.

Record: opening sketch and placed details → epitome and level given → synthesis response → candidate interpretations and confidence basis → distinguishing probe → next level or repeat, and why → next observation at the stated horizon. Preserve the learner's stated purpose and any change in it. An agent may summarise this record, but must not invent missing inputs or assign numerical state probabilities without a calibrated model.

Elements and limits

Epitome, simple-to-complex sequencing, advance organizers, conceptual overviews, concept maps, summarization and synthesis, analogies, activation, part-task practice, practice, reflection and spiral curriculum. The pattern also applies metaphors and analogies, creating visual representations, guided practice, worked examples, scaffolding and cognitive load theory.

This pattern is scoped to sequencing organised content over a unit or course; it is a poor fit for isolated facts, and a forced epitome can distort a domain with no single organising idea. Its design cost is real: choosing the epitome and the levels is a content analysis that needs someone who knows the domain. The evidence above tests a hierarchical organisation of knowledge in small physics studies, not the sequence, and the conditional branches remain to be tested with learners.

Claims

Key Sources

  • Reigeluth, C. M. (1979). In search of a better way to organize instruction: The elaboration theory. Journal of Instructional Development, 2(3), 8-15. doi:10.1007/bf02984374
  • Reigeluth, C. M., & Stein, F. S. (1983). The elaboration theory of instruction. In C. Reigeluth (Ed.), Instructional-design theories and models (pp. 335-381). Erlbaum.
Page metadata
Field Value
type pattern
id elaboration-theory
aliases reigeluths-elaboration-theory
title Elaboration Theory
description A reusable course-level sequencing policy: an epitome of the most general ideas, elaborated level by level toward detail and tied back to the whole by learner-built synthesis; one small study of hierarchical knowledge organisation bears on it, and no claim tests the sequence itself.
status review
grain_size course
author Charles Reigeluth
canonical True
generated by claude/unspecified on 2026-10-02
trust unverified — no one has confirmed this page's content
sources 1 — reigeluth-1979

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