Skip to content

Activation

Principle · All principles
Evidence · 25 claims (10 for, 13 mixed, 2 against) · 29 studies (11 causal, 8 review, 6 quant-synthesis, 2 qualitative, 1 associational, 1 theoretical), q1–q4 · 7 of 29 report an effect size · 17 claims rest on one study

Conditional relationship

For a learner meeting new material that connects to something they already know, a short activity that makes the relevant prior knowledge visible before instruction (a prediction, an explanation of a familiar case, a pre-question, an organizer that relates the new content to the known) may improve comprehension or later retention of that material, compared with starting the instruction directly. The relationship is conditional on at least four things: that the learner has relevant and accurate knowledge to activate (expertise matters, but the useful observation is the task-specific starting response, not the label); that the activity surfaces knowledge that the following instruction actually uses; what kind of knowledge is activated (topic knowledge, metacognitive knowledge, a misconception); and the outcome and horizon (immediate-performance, conceptual-understanding, delayed-retention). Where the activated knowledge is inaccurate, the expected benefit depends on the instruction confronting it, not on the activation itself.

No claim in this wiki tests the relationship as a whole against a no-activation comparison across these conditions. The recorded evidence is split: synthesis evidence on advance organizers and experiments on pre-questions point one way, and two experiments of topic-knowledge activation found little or no effect. The Merrill activation pattern is the matching pattern page (not yet in this format); prior knowledge activation and advance organizers are the elements that carry it.

Fitting the design to a situation

Ask for three facts when the brief omits them: do learners hold relevant and accurate knowledge of the topic, or likely misconceptions?, will the instruction that follows use what is activated?, and is the outcome recall of specific content or broader understanding, and at what delay?

If the brief says… Then change… Basis
Learners know little about the topic Do not open with a free brainstorm of what they know; ask a few questions on the coming content instead Prior-knowledge activation had little effect on retention [-M]; Pretesting enhances learning [+M]
The material is expository text or video to be studied Put a few pre-questions before the reading or between video segments, and expect the gain mainly on the questioned content Pretesting enhances memory for tested text [+M]; Pretesting enhances memory for tested video [+M]
The outcome covers content the pre-questions did not ask about Do not count on pre-questions for it; benefits for untested content were inconsistent across experiments Effects on untested materials are inconsistent [~M]
A multi-week course with a high-stakes exam Give short ungraded pretests at the start of units on the content the exam will cover Classroom pretesting improves exam performance [+M]
A voluntary online course (MOOC) where persistence matters Keep any pretest short and optional, and tell learners it is not graded; pretests reduced persistence in one MOOC Pretesting can harm motivation [-M]
The content is vocabulary or paired items Give the correct answer immediately after each pre-question; delayed feedback removed the benefit for paired items Feedback timing moderates pretesting [~M]
Learners are likely to hold misconceptions Surface the belief, then confront it directly (state the misbelief, then the fact and why) rather than presenting the correct account alone Misconceptions interfere with new learning [~M]; Refutational text enhances recall [+M]
The designer can write an organizer relating the new content to what learners know Use one, and expect a small effect, larger for higher-ability learners Advance organizers improve learning [+M]
Primary-school pupils will read a text for comprehension Activate what they know about how to read (strategies, monitoring) rather than what they know about the topic Activation improves learning [~M]
Learners resist answering questions on material they have not studied Tell them why it helps; learners rank pretesting as the least effective strategy even when it helps them Learners underrate pretesting [~M]

Observation, state and explanation

What a designer observes is a response to an activation prompt under stated conditions: what the learner wrote or said, how much of it bears on the coming content, how confidently it was offered, and how long it took. "Has relevant prior knowledge", "holds a misconception" and "is now primed" are inferred states. A blank or thin response does not show that nothing is known, and a fluent one does not show that the knowledge is accurate or that the learner will connect it to the new material. The rows below are proposals, untested with learners.

Response Plausible explanations Observation that could change the interpretation
Little or nothing relevant produced No relevant knowledge; knowledge present but the cue does not reach it; unwillingness to guess in front of others Re-ask with a concrete familiar case or a recognition item instead of open recall, privately; record whether the response changes.
Fluent, confident but inaccurate idea A coherent misconception; a correct idea in the wrong context; a guess stated with assurance Ask for the reasoning behind it and a prediction for a discrepant case; check whether the same idea recurs on a differently worded item.
Accurate ideas, but not connected to the new task Knowledge available but not seen as relevant; prompt aimed at topic knowledge rather than the relation the lesson needs After instruction, ask the learner to say how the activated idea bears on the new problem; compare with a probe that names the connection.
Better later recall after a pre-question, even when it was answered wrongly The attempt directed attention to the answer during study; the question itself flagged what matters; extra exposure to the item Compare items that were attempted with items only read, and items pre-questioned with items not; keep all exposures in the record.
No later difference after activation Activation added nothing beyond what instruction supplied; too little relevant knowledge was surfaced to matter; outcome measured only immediately or on a measure the activation does not touch Record how many relevant ideas were produced, the outcome instrument and the delay; a null on one outcome does not establish equivalence on another.

These observations can shift confidence between explanations; none identifies a cause on its own, and the probe itself is a further activation that can change the state.

Evidence and qualifications

  • Activation improves learning [~M]: the claim page's own entries qualify its title. An advance-organizer meta-analysis (135 studies) reports a facilitative effect on learning and retention, effect size not in the abstract read; a randomized 2 × 2 experiment with 88 primary-school students found that activating topic knowledge did not improve text comprehension once prior knowledge was controlled, while activating metacognitive knowledge did; and an experiment with 165 secondary students found no difference in conceptual knowledge between activating a high or low share of a concept's components, with no no-activation control. Partly checked: 2 of 3 entries pass, the rest could not be confirmed (abstract).
  • Prior-knowledge activation instructions before study had little effect on retention [-M]: students talked for 1.5 minutes about topic-relevant or unrelated headings before studying science facts; the only effect was a free-recall interaction favouring the unrelated activation for belief-consistent facts, which the authors say to read with caution because students generated very few ideas matching the content (about one per topic). This bears on open learner-generated activation when little relevant knowledge exists. Not yet checked against its sources.
  • Advance organizers improve learning [+M]: two meta-analyses (135 and 112 studies) associate designer-authored organizers with better learning and retention; one reports a small mean effect (0.21 on learning, from a conference version of the analysis), larger for high-ability (0.23) than low-ability learners (0.13), and the other found no special benefit for low-ability or low-knowledge learners. Effect sizes were unweighted and mostly on immediate tests; the evidence predates modern designs. It does not predict that organizers help those who know least. Partly checked: 1 of 3 entries pass, the rest could not be confirmed (abstract).
  • Pretesting enhances learning [+M]: in laboratory experiments with undergraduates, attempting questions on not-yet-read material, then reading it, beat equivalent extra study time on a later test of the same material (Experiment 1: 75% against 56%, d = 1.1), even for items answered wrongly at pretest; reading the questions without attempting them helped less. The exact delay to the later test is not reported on the claim page, and the materials were expository text, trivia and word pairs. It supports a question-first form of activation, not open recall of what one knows. Not yet checked against its sources.
  • Misconceptions interfere with new learning [~M]: a meta-analysis and a review of conceptual-change interventions in science and reading report that standard exposition rarely dislodges misconceptions, and that interventions producing conceptual conflict, or refutation text, were more effective. This is the limiting condition: surfacing an inaccurate idea is expected to help only if the instruction then confronts it. Both entries were read as abstracts; no pooled effect is recorded. Not settled: the abstract available could not confirm the entries.
  • Prior knowledge determines new learning [~M]: a meta-analysis of 8,776 effect sizes finds prior knowledge strongly correlated with later knowledge (r = .534) but almost unrelated to normalized learning gains (r = −.059), with a prediction interval from −.688 to .621. It is correlational and does not test activation; it qualifies the assumption that prior knowledge reliably helps new learning, which the activation rationale rests on. Not yet checked against its sources.

Keep each study's learners, prompt, duration, comparator, outcome and delay when transporting these results. The six claims differ in design, comparator and outcome; do not rank them by effect label.

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 (eval/synthesis/activation.json, scripts/synthesize_claim.py). The count is of distinct studies across the claim and the narrower claims that bear on it.

Further evidence, not yet read against this model

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

Objective and learner-valued goal

The designer's objective is a capability on the new material, with a representative instrument, a criterion and a horizon; activation is a means to it, and its own output (a list of ideas, a concept map) is not the outcome. Ask separately what the learner wants from the topic, and treat agreement or divergence as an observation.

For example, adults starting a unit on household electricity may want to know whether a wiring job at home is safe, while the designer's objective is explaining current and resistance on a written test. An opening that asks what they already believe about fuses can serve both, but only if the later instruction returns to those beliefs, and a written test alone cannot show whether a confidently held misconception about safety has changed.

What would revise this model?

The expectation should weaken if comparable learners, prompts and aligned outcomes show no benefit of activation over starting instruction directly, especially on delayed outcomes; two recorded experiments already point that way for open topic activation. It should be narrowed if the benefit appears only for question-first forms (pre-questions, organizers) and not for learner-generated recall. If activating a misconception and then refuting it does no better than refutation alone, the claim that activation is needed to surface misconceptions loses its support. If the number of relevant ideas a learner produces predicts later learning only because both reflect prior knowledge, the activation adds nothing a pretest would not reveal. Do not protect the model by relabelling every null as "the wrong knowledge was activated".

The present evidence does not establish an optimal activation format or duration, a benefit for learners with little relevant knowledge, or durability beyond the tests the studies used.

Claims

Examples

Illustrative

Pre-Reading Activities — Before a dense article or textbook section, learners generate predictions, define likely key terms, or answer a short anticipatory question set. The value comes from making prior ideas visible so the reading can confirm, refine, or contradict them.

Prior Knowledge Activation — In a science lesson, learners first explain what they think causes seasons or why objects float, then compare those ideas to a demonstration or text. This turns activation into a diagnostic starting point rather than a ritual opener.

Background Knowledge Activation — A routine in which an instructor uses a quick write, brainstorm, or concept map to surface what learners already know before introducing new material. Strongest when the teacher refers back to the activation during the lesson rather than abandoning it after the opener.

Key Sources

Page metadata
Field Value
type principle
id activation
aliases prior-knowledge-growth-point-new-knowledge
title Activation
description For learners whose relevant prior knowledge is accurate and is then used by the instruction that follows, an activity that makes it visible before instruction may improve comprehension or retention, while activating topic knowledge alone has often shown little effect and activated misconceptions need confronting.
status review
moved_from pattern/merrills-first-principles-activation
canonical True
generated by claude/unspecified on 2026-10-02
trust unverified — no one has confirmed this page's content
sources 3 — bransford-2000, qiong-jia-2010, merrill-2002

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