Skip to content

Knowledge Organization

Principle · All principles
Evidence · 28 claims (17 for, 11 mixed) · 31 studies (10 causal, 8 review, 6 quant-synthesis, 3 theoretical, 2 design, 1 associational, 1 qualitative), q1–q4 · 5 of 31 report an effect size · 23 claims rest on one study

Conditional relationship

This page owns the organization a learner ends up holding for a body of knowledge: which ideas sit at the top, how the rest hangs from them, and what learners do to make that organization their own rather than merely see it. For a learner who holds a topic's ideas as separate items (a novice on this material: they can recall facts but cannot say how they connect), and who will later need those ideas for tasks that cut across them (recalling the topic as a whole, explaining, solving problems that need several parts at once; concept and principle targets, conceptual-understanding, near-transfer), teaching the same content with an explicit organization (a few central ideas with the rest nested beneath them, chosen to serve the tasks to come) and having learners work with it (sort items into it, mark it, explain its links, rebuild it from memory) is expected to improve performance on those cross-cutting tasks compared with the same content taught as a single-level list. It is not expected to improve tasks that need only one detail at a time.

The relationship is conditional on four things:

  • Match between organization and task. An organization helps the tasks it was built for. Ideas placed high in it are recalled and used better; ideas placed low are recalled worst. There is no organization that is good in general.
  • The learner's starting knowledge and ability. In the one direct test, the lowest-ability learners showed no benefit; learners who already hold an organization of their own may find a given one redundant.
  • What the learner does with the organization. Tasks that make the learner place items into it and mark its structure beat a task that only asks for a well-organized summary of the same material.
  • The comparison. Against reading a list, organization helps; against an equally long retrieval activity, building a map while looking at the text loses.

What is tested, and what is not. One report (Eylon & Reif, 1979: three small experiments with college physics students and an advanced high-school physics class) tests the core contrast directly, the same content taught with a hierarchical or a single-level organization, and its claim pages carry the match, ability, detail-task and acquisition-task qualifications below. It is one study in one domain, with samples of 8 to 36, and the claim pages print F tests but no effect sizes. Two concept-mapping meta-analyses test one visual way for learners to build an organization. No claim tests other kinds of organization (causal chains, sequences, comparison matrices) as organization, a course-long framework revisited unit by unit, or how to fade a given organization across sessions. The default design below keeps the earlier page's guidance as labelled proposals around that tested core.

How the converted neighbours sit beside this page: Clear Structure holds the relationship that making the material's organization visible (signals, previews, matching graphics) helps learners who lack one; Graphic Organizers holds the visual form (which diagram, who builds it); Chunking holds the size of units and where their boundaries fall; Activation holds connecting new material to what learners already know before instruction; Self-Explanation and Retrieval Practice hold the two activities this page asks learners to do with an organization. This page asks a prior question: which organization should the learner end up holding, given the tasks they will face, and how do they come to hold it?

Default design

The earlier page's guidance, made concrete. The core contrast (organized against single-level) has one direct test; each step says what stands behind it.

  1. Start from the tasks, then choose the top level. Write down three to five tasks learners must eventually do with the topic. Put at the top of the organization the ideas those tasks turn on: governing principles if learners must derive or solve, a timeline if they must trace how ideas developed, causes if they must explain outcomes, categories if they must classify. If two kinds of task matter, build two short organizations of the same content rather than one that serves neither. Basis: in advanced high-school physics, students given principles at the top did better on deductive tasks and students given history at the top did better on historical tasks, two weeks later (A hierarchical organization facilitates tasks only when matched to the task domain [+M]).
  2. Put what matters most high, and keep the top level small. Three to five top-level ideas, two or three levels deep per view; split a larger topic into an overview and one view per branch. Recall falls with depth: in the same report, recall two weeks later was about 50% for the top two levels, 27% for the third and about 5% for the lowest (Information at higher levels of an acquired hierarchical organization is recalled better than information at lower levels [+M]). The counts are an untested proposal; the reason for a small top level is the limit on what learners can hold at once (Chunking reduces working memory load by grouping information into fewer, more meaningful units. [+M], whose recent estimate is three to five chunks for novel material, a memory finding, not a test of lesson design).
  3. Show the whole organization before the detail (5 minutes). Open the topic with the top level and how the parts relate, then teach each part as a branch of it, saying where it sits. Basis: college physics students taught the same content with a hierarchical organization did better on complex recall and problem-solving tasks than those taught it at a single level, presented once or twice (A hierarchical internal knowledge organization facilitates performance on complex recall and problem-solving tasks more than a single-level organization of the same knowledge [+M]). Showing it first, rather than at the end, is an untested proposal; for organizers given before instruction, see Activation.
  4. Have learners place items into the organization, not just summarize it (10–15 minutes a session). Give a chart of the organization and a set of items (facts, examples, terms) to sort into it, then have learners mark which level each belongs to and label the links. Basis: with study time equal, students whose tasks included marking organizational charts and arranging items into a chart outperformed students who wrote a structured summary (Strong acquisition tasks explicating the organization produce better internalization than weak tasks asking only for a structured summary [+M]; one small study). The durations are untested proposals.
  5. Ask for the links in words. For two or three links per session, ask "why does this sit under that?" or "what would change here if that changed?". Basis: prompting learners to explain improved understanding and problem solving on immediate tests by a moderate average amount (Prompting learners to self-explain improves understanding and problem solving on immediate tests by a moderate average amount, with little evidence yet on delayed or classroom gains. [+S]); explaining links in an organization is a carried application, not what the studies tested.
  6. Then have learners rebuild it from memory (5 minutes), and check. Close the materials; learners sketch the top two levels and as much below as they can, then compare with the reference and fix gaps. Basis: building concept maps without the texts in view gave better short-answer performance a week later than building them while studying the texts (The same quizzing or concept-mapping activity produces more learning when done without viewing the material, as retrieval, than while viewing it [+M], reported second-hand); building a map with the text open lost to time-matched retrieval practice (Concept mapping improves learning over most comparison conditions by a moderate average amount, but not over time-matched retrieval practice [~S]).
  7. Return to the same frame as detail is added. Each later session opens on the organization, places the new material in it, and closes with a short rebuild. Untested; from the earlier page's "sequenced unit frameworks".
  8. Test with both kinds of task, and expect the gain only on one. Check learning with at least one task that needs several parts together and one that needs a single detail. Basis: the hierarchical organization made no significant difference to tasks relying only on individual detailed steps (Hierarchical versus single-level organization does not significantly affect performance on local tasks relying only on individual detailed steps [~M]; equivalence was not tested). If detail matters on the assessment, keep detail practice as well.
  9. Plan the arc across sessions. A proposed default, untested as a whole: sessions 1–2, the organization is given complete, learners sort items into it and explain two links; sessions 3–4, a partial frame (top level given, lower levels and links left blank), learners place new material and label links, then rebuild from memory; later sessions, learners rebuild from memory at the start, extend it with new material, and reorganize it once for a different task (for example, from a timeline to a cause-and-effect organization of the same events). Move a learner on when they can rebuild the top two levels from memory and explain two of three links they are asked about; give learners who already do so on the first day the partial or blank frame at once (Instructional guidance that helps novices can become redundant or counterproductive as expertise grows. [~M]).
  10. Watch the weakest learners. In the direct test, the lowest-ability third gained nothing from the hierarchical organization while the middle third did (Ability moderates the hierarchical-organization advantage: low-ability subjects showed no significant benefit, while medium-ability subjects did [~M]). For learners who cannot yet use the organization, slow the sort-and-rebuild cycle, use fewer items per sort, and check the top level alone before adding levels; these adjustments are untested.

When a condition is not met: if the tasks learners will face are unknown, ask before choosing a top level (step 1), because an organization built for the wrong task gives no benefit on the right one; if the content really is a list of independent items (vocabulary, isolated facts), use retrieval practice on the items and do not impose a hierarchy; if learners already state the organization and its links unaided, skip steps 3–4 and use steps 6 and 8 only.

Fitting the design to a situation

The three facts that most change the decision:

  • What the learner will have to do with the knowledge. The organization should be built for those tasks; one built for another task can help nothing or help the wrong task. If unstated, ask: name three things learners must do with this topic afterwards, unaided, and when.
  • How the learner holds the topic now. A learner with isolated facts needs a given organization; one with an organization of their own may need it challenged or extended, not replaced; one who cannot yet use any organization may not benefit. If unstated, ask the learner to sort ten items from the topic into groups and say why, before teaching.
  • Whether the content has relational structure at all. Causes, hierarchies, principles and developments have it; a list of independent items does not, and imposing one adds work without benefit. If unstated, ask: do the tasks need learners to connect items, or only to recall each one?
If the brief says… Then change… Basis
Learners are novice on this topic, holding unconnected facts Give the organization complete at first, with sorting tasks; move to partial frames only after they rebuild the top level from memory A hierarchical internal knowledge organization facilitates performance on complex recall and problem-solving tasks more than a single-level organization of the same knowledge [+M] (college physics, one report)
Learners are advanced, or already state the organization unaided Skip the given organization; ask them to rebuild it from memory, then reorganize it for a new task or critique where it fails Instructional guidance that helps novices can become redundant or counterproductive as expertise grows. [~M] (review of cognitive-load studies; carried to organizations, which it did not test)
Learners are low-attaining, or struggled with the topic before Fewer items per sort, top level first and checked alone, more sort-and-rebuild cycles before adding levels; check whether they use the organization at all Ability moderates the hierarchical-organization advantage: low-ability subjects showed no significant benefit, while medium-ability subjects did [~M] (college physics; the adjustments are untested)
Learners are child (primary age) Use story-like or picture-card organizations of two levels, sorted physically; for narrative content, organize around goal, events and outcome untested for organization design; narrative recall is better when stories follow a familiar story structure (Narrative memory is superior when content follows stereotypical story grammar structure, and goal information is critical to comprehension [+W], second-hand review, narratives only)
Goal is solving problems that use several principles (principle, near-transfer) Put the principles at the top and examples and procedures beneath them; test with problems that need two branches at once A hierarchical organization facilitates tasks only when matched to the task domain [+M] (advanced high-school physics, two weeks)
Goal is recall of individual facts or carrying out single steps Do not build a hierarchy for its own sake; practise the items with retrieval and feedback; use organization only where items need connecting Hierarchical versus single-level organization does not significantly affect performance on local tasks relying only on individual detailed steps [~M] (null result; equivalence not tested)
Goal is telling similar categories or cases apart (concept) Organize as a comparison: place two or three cases side by side and have learners list similarities and differences before the principle is named Comparing cases side by side improves learning and transfer by a moderate average amount, though in one algebra study it raised procedural knowledge and flexibility but not conceptual knowledge [+S] (tests comparison, not organization as such)
online-self-paced, no teacher present Use drag-and-drop sorting into the organization with automatic checking, then a closed-book rebuild the learner compares with a reference map; do not drop the rebuild to save time untested; the closed-book rebuild rests on The same quizzing or concept-mapping activity produces more learning when done without viewing the material, as retrieval, than while viewing it [+M] (second-hand, settings not stated)
Mixed modes: some learners remote, some in the room, or phone-only Keep the organization as an indented outline that fits a phone screen, not a wide diagram; remote and in-room learners sort into the same shared document; read the top level aloud for those who cannot see the board untested proposal
classroom, one teacher and a large group One shared organization on the board; pairs sort printed cards into it; the teacher checks only top-level placement across the room, then everyone rebuilds individually from memory untested at class scale; pairs building maps together is reported to beat individuals (An ICAP theory article reports, second-hand and without describing the design, a study in which concept mapping with a peer enhanced learning more than concept mapping alone [+W], one study reported second-hand)
Time is one session only Show the organization, one sort task, one explained link, one closed-book rebuild at the end; record what the rebuild shows and claim nothing about retention untested proposal
Facilitator time is scarce (volunteers, a busy shift, a tutor checking in briefly) Give a finished organization rather than having learners build maps from scratch, which is slow; use the tutor's minutes only to look at a 3-minute closed-book sketch and ask about one link untested; mapping is slow and its advantage over time-matched retrieval is not shown (Concept mapping improves learning over most comparison conditions by a moderate average amount, but not over time-matched retrieval practice [~S])
Learners have low confidence or low literacy Use pictures or objects on cards for items, let learners explain links aloud or by pointing, and keep labels to a few words; sort before writing untested proposal
Adults in a work setting Build the top level from the decisions or tasks of the job (for example, "what to check first, what to escalate"), using the learners' own cases as the items to sort untested; the task-match finding is carried from physics students to workplace tasks (A hierarchical organization facilitates tasks only when matched to the task domain [~M])
Learners must produce an explanation, talk or product for a real audience Have them build their own organization of the topic first, then select from it what this audience needs; the audience sees the explanation, not the working map untested proposal
Stakes are high, and the exam mixes detail items with integrative ones Keep separate retrieval practice on the details; judge readiness on both kinds of item, never on the organization exercise alone Hierarchical versus single-level organization does not significantly affect performance on local tasks relying only on individual detailed steps [~M]
The brief names its own routine (a note-taking template, a KWL chart, a unit map the school uses) Keep it; change only two things: tie its top level to the tasks learners will face, and add one closed-book rebuild of it per session untested; the rebuild rests on The same quizzing or concept-mapping activity produces more learning when done without viewing the material, as retrieval, than while viewing it [+M]

Thirteen of the seventeen rows cite a claim, and four (mixed modes, one session, low literacy, a real audience) are untested proposals with no claim. Only the rows on novice learners, low-attaining learners, problem-solving goals, detail goals and high stakes rest on the one direct test of organization, and that test used physics students; the adults row carries it to work settings. The rest carry a neighbouring finding (comparison of cases, concept mapping, retrieval, story structure, expertise reversal) into this setting and say so, and several of those are untested as moves.

Observation, state and explanation

What can be observed: how a learner sorts items into groups, the order and grouping of their free recall, a sketch they draw from memory, which links they can explain, and performance on tasks that need one part or several. "Has a well-organized understanding" is an inferred state. A neat map drawn with the materials open shows what the learner can copy, not what they hold.

Response Plausible explanations Observation that could change the interpretation
Recalls many details but cannot say how they relate No organization taught; one taught but never used by the learner; the test rewards details only Ask for a sketch of the main ideas from memory; give one task that needs two parts together and compare.
Rebuilds the organization correctly but fails problems that need it The organization learned is not the one the task needs (mismatch); learned as a picture without its meaning; the problem needs a procedure not yet practised Ask which branch the problem belongs to and why; give a problem matched to the organization taught and one matched to another.
Sorts items correctly with the chart in view, fails to rebuild it from memory Copying rather than holding; too many items or levels at once Shrink to the top level only and rebuild again; compare with a rebuild after a day.
Learner with prior knowledge resists or ignores the given organization They hold a different, possibly valid organization; the given one is redundant for them; it conflicts with a misconception Ask them to draw their own first, then compare the two on a task each should serve.
Recall of lower-level detail is poor while the top level is solid Detail was placed low and is expected to be recalled worst; detail was never practised on its own Separate detail-recall check; move a detail that matters higher, or practise it by retrieval.

These rows are proposals, untested with learners. A sort or rebuild is itself practice and changes what it measures, so record each one, and keep open the explanations a single probe does not rule out.

Evidence and qualifications

The direct evidence is one report, Eylon & Reif (1979), read through five claim pages that draw on different experiments in it. All five are small, all are physics, and none prints an effect size (the i codes on these pages rest on F tests or percentages, not on a printed effect size).

Do not rank these by effect label: physics students on hierarchies, pooled concept-map studies and retrieval experiments compare different things. None measures retention beyond a week or two, and none tests a course-long organization.

The evidence by question

Further evidence, not yet read against this model

Objective and learner-valued goal

The designer's objective is usually unaided performance on integrative tasks (explaining, solving problems across parts of a topic, recalling it as a whole) at a stated delay. Ask separately what the learner wants from the topic: to pass a test of facts, to make sense of something that has confused them, to use it in a job. A learner who values passing a detail-heavy test may see the organization exercise as a detour, and on detail items they may be right; a learner who wants to understand may welcome it. Record both, and if they diverge, show the learner one task they care about that the organization helps with.

For example, an adult apprentice electrician may want to find faults quickly on site, while the course objective is recall of the regulations by section. An organization built around the fault-finding decisions (what to check first, what each reading rules out) serves the learner's goal; one built around the regulation chapters serves the exam. Building both from the same content, and saying which task each serves, keeps the two from being confused.

What would revise this model?

The expectation should weaken if further comparisons of organized against single-level teaching of the same content, with learners outside physics and with aligned integrative tasks, show no advantage, or if an advantage at two weeks disappears at a longer delay. If closed-book retrieval of the content's items, without any taught organization, matches organized teaching on integrative tasks, the organization adds nothing beyond retrieval and this page should defer to Retrieval Practice. If low-attaining learners consistently gain nothing, the model should say who it is for. If studies of causal, sequence or comparison organizations behave unlike hierarchies, the top-level and depth advice is hierarchy-specific. Do not protect the model by calling every failure "the wrong organization" after the fact; name the matching task before teaching.

Assisted sorting, unaided rebuilding, integrative-task performance, detail recall and use in the learner's own setting are separate claims. The present evidence establishes no number of levels, no session dose and no criterion for moving on.

Claims

  • Cognitive Load Management — knowledge organization is often one concrete route to load reduction
  • Creating Visual Representations — visual structures can make conceptual relations visible
  • Metacognition — organized knowledge is easier for learners to inspect, monitor, and revise
  • Graphic Organizers — the visual form an organization can take, and who builds it
  • Clear Structure — making the material's own organization visible in presentation
  • Chunking — the size of units and where their boundaries fall
  • Activation — connecting new material to what learners already know, including advance organizers
  • Retrieval Practice — rebuilding an organization from memory is a retrieval activity
  • Self-Explanation — explaining the links in an organization
  • Sequencing — the order in which branches of an organization are taught

Examples

Illustrative

Concept Mapping — Learners build visual maps that make relationships among concepts explicit and revisable.

Graphic Organizers — Compare-contrast charts, timelines, hierarchies, and causal maps give learners a structure for complex information.

Sequenced unit frameworks — An instructor repeatedly returns to the same conceptual map as new detail is added, helping learners keep the big picture intact.

Key Sources

  • Novak, J. D. (2010). Learning, creating, and using knowledge (2nd ed.). Routledge. doi:10.4324/9780203862001
  • Ausubel, D. P. (1968). Educational psychology: A cognitive view. Holt, Rinehart and Winston.
  • Eylon, B.-S., & Reif, F. (1979). Effects of internal knowledge organization on task performance. https://eric.ed.gov/?id=ED171804
  • Armbruster, B. B., & Anderson, T. H. (1982). Structures for Explanations in History Textbooks or So What If Governor Stanford Missed the Spike and Hit the Rail? (Technical Report No. 252). https://eric.ed.gov/?id=ED218595
Page metadata
Field Value
type principle
id knowledge-organization
aliases make-frame-slots-explicit-history-explanations
title Knowledge Organization
description For a learner who holds a topic as unconnected facts and must later use it for tasks that cut across it, teaching the content with an explicit organization built for those tasks, and having learners sort into, explain and rebuild that organization from memory, is expected to improve integrative tasks but not single-detail tasks; one small 1979 physics report tests the core contrast, and no claim tests other kinds of organization or a course-long framework.
status review
canonical True
generated by claude/unspecified on 2026-10-05
trust unverified — no one has confirmed this page's content
sources 1 — armbruster-1982

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