Instructional guidance that helps novices can become redundant or counterproductive as expertise grows.¶

Claim · All claims
Evidence · 5 studies · 3 review r2, 1 causal r?, 1 theoretical r2 · q1–q3

The expertise reversal effect describes a recurring pattern in instructional research: scaffolds, explanations, and worked examples that reduce cognitive load for novices lose value once learners have already internalized the underlying schema.

Subclaims¶

q2 i? Highly guided instruction tends to benefit novices more than experts. → Kalyuga et al. 2003

q2 i? As prior knowledge increases, redundant explanations and worked examples can depress performance relative to less guided conditions. → Kalyuga et al. 2003

q2 i? A synthesis of cognitive-load studies across split-attention/redundancy, modality, worked-example, isolated-elements and imagination-effect paradigms finds that guidance formats which help novices (integrated text, worked examples, auditory narration, isolated elements, direct study) repeatedly lose their advantage — and in several paradigms reverse to a disadvantage — once learners have acquired domain-specific schemas. → Kalyuga et al. 2003

q3 i? In three experiments teaching electrical trainees to read circuit diagrams, novice trainees learned better from diagrams with integrated explanatory text, but as trainees gained experience the best format shifted to a diagram with the text eliminated, and the most experienced trainees performed significantly better with text removed than with it present. → Kalyuga et al. 1998

q1 i? A conceptual paper on serious games restates, citing Kalyuga et al. (2003), that additional text is redundant for experts and has negative rather than positive effects, and argues that instructional design should be tailored to learners' level of experience; it presents no new data. → Obikwelu 2013

q2 i? Instructional methods that benefit novices can become unnecessary or detrimental for advanced learners, as a full or partial reversal. → Slava Kalyuga 2010

q2 i? A special case of ATI in which an instructional parameter increases learning gains in novices but decreases them in experts, highlighting the need for dynamic conceptualization of aptitudes. → Leonard Tetzlaff 2021

Evidence¶

Primary evidence link: https://doi.org/10.1207/S15326985EP3801_4

Kalyuga et al. 2003¶

Kalyuga, S., Ayres, P., Chandler, P., & Sweller, J. (2003). The expertise reversal effect. Educational Psychologist, 38(1), 23–31. doi:10.1207/s15326985ep3801_4

q2 · narrative review of experimental cognitive-load studies · i? · no pooled effect size reported (individual experiments described as producing statistically significant reversals, but no d/F values given in the text) · n=N/A (review; ~10 primary studies cited) · review · r2

A narrative review by the group that coined the term, synthesizing its own and others' experiments across five separate cognitive-load paradigms — split-attention/redundancy (including the Kalyuga, Chandler, & Sweller (1998) study below), modality/redundancy, worked examples, isolated interacting elements, and the imagination effect. In each paradigm the same instructional design that most helped inexperienced learners either lost its advantage or actively hurt more experienced learners once they held relevant schemas, because processing the extra guidance became redundant cross-referencing that consumes working memory rather than useful support. The review frames this as a consequence of cognitive load theory: guidance that substitutes for a missing schema helps a novice, but the same guidance duplicates what an expert's automated schema already provides.

Kalyuga et al. 1998¶

Kalyuga, S., Chandler, P., & Sweller, J. (1998). Levels of expertise and instructional design. Human Factors, 40(1), 1–17. doi:10.1518/001872098779480587

q3 · peer-reviewed experiment (three studies) · i? · no effect size reported in the read text (differences described as "significantly better," no d/F/p values available in the abstract) · n=N/A (abstract only; per-experiment n not stated in the text read) · causal · r?

Three experiments with trainees learning to read electrical/circuit diagrams compared a diagram with text physically integrated onto it against a diagram-only format with the (redundant) text eliminated. For less experienced trainees, the integrated diagram-plus-text format worked best — a diagram alone was not intelligible to them. As trainees' domain experience increased across the three experiments, the best-performing format shifted, and the most experienced group performed better with the diagram-only format than with the integrated text present, replicating the split-attention benefit for novices with a redundancy-driven reversal for experts. Only the abstract was available for this entry (see provenance); the numeric effect sizes are not asserted here because they were not present in the text read.

Obikwelu 2013¶

Obikwelu, C, Read, J, and Sim, G. (2013). Children’s Problem-Solving in Serious Games: The “Fine-Tuning System (FTS)” Elaborated. The Electronic Journal of e-Learning, 11(1). No Crossref record was found for this article (title search, 2026-10-05); the link formerly given was the journal's homepage, https://www.ejel.org, which is not a citation of it.

q1 · conceptual paper, no new data · i? · the article prints no effect size for this finding · theoretical · r2

Section 6 of this conceptual paper reports the expertise reversal effect as attributed to Kalyuga et al. (2003): "Providing additional text is redundant for experts and will have negative rather than positive effects". The article also cites work proposing that eliminating redundant material reduces cognitive load for experienced learners. No new data are presented.

"Providing additional text is redundant for experts and will have negative rather than positive effects, thus demonstrating the expertise reversal effect."

Slava Kalyuga 2010¶

Slava Kalyuga, Alexander Renkl. (2010). Expertise reversal effect and its instructional implications: introduction to the special issue. Instructional Science. https://doi.org/10.1007/s11251-009-9102-0

q2 · i? · review · r2

The authors define the expertise reversal effect, noting it has been replicated in many studies with a large range of instructional materials and participants, either as a full reversal (a disordinal interaction) or more often as a partial reversal with a significant interaction.

"The expertise reversal is a reversal in the relative effectiveness of instructional methods as levels of learner knowledge in a domain change."

Leonard Tetzlaff 2021¶

Leonard Tetzlaff, Florian Schmiedek & Garvin Brod. (2021). Developing Personalized Education: A Dynamic Framework. Educational Psychology Review. https://doi.org/10.1007/s10648-020-09570-w

q2 · i? · review · r2

The review describes the expertise reversal effect (Kalyuga 2007) as evidence that instructional parameters effective at the start of learning can actively impede learning as expertise grows; even reading a short text can alter the effectiveness of subsequent instruction (Rey and Fischer 2013).

"The expertise reversal effect is present if a certain instructional parameter leads to increased learning gain in novices, but decreased learning gains in experts."

Studies bearing on this claim¶

Do the cues still help learners who already know the material, or do they add load? This table gathers the studies behind this claim and the 38 narrower claims that bear on it: 30 distinct studies (11 causal, 9 review, 5 theoretical, 5 quant-synthesis), q1–q4, 4 of 30 with an effect size coded. One row per study, codes as each entry records them; no effect is averaged and no verdict is drawn. Built by scripts/synthesize_claim.py from eval/synthesis/; a row that looks wrong is fixed on the claim it comes from.

Discussion¶

The claim does not mean experts need no support. It means support should change form: less completion of routine steps, more emphasis on challenge, transfer, and selective feedback.

Merged from “Expertise Reversal Guidance Hurts Experts” (expertise-reversal-guidance-hurts-experts): The expertise reversal effect is the dynamic counterpart to guidance benefits for novices. Under cognitive load theory, guidance reduces the search burden that overwhelms novices; but once learners have automated domain schemas, that same guidance duplicates information they can generate internally, imposing extraneous load and interfering with schema construction. The effect is therefore inherently conditional: the same instructional treatment produces opposite effects at different levels of expertise, and no fixed level of guidance is optimal across a learning trajectory.

Practical consequences follow. Guidance should be faded as competence grows rather than held constant — see Worked examples can become redundant or counterproductive for advanced learners.. Adaptive designs that tailor guidance to measured learner expertise outperform one-size-fits-all sequences. Assessment of prior knowledge is thus a prerequisite for applying any guidance-heavy pattern such as worked examples or scaffolding, and supports adaptive difficulty designs that adjust support to current competence.

Boundary conditions matter. The effect is documented primarily for guidance that duplicates what experts can already derive — full worked steps, redundant explanations, high-support scaffolds. Guidance that adds genuinely new information (e.g., feedback on errors, novel problem constraints) does not automatically reverse; the mechanism is redundancy, not the mere presence of support. Expertise is also domain-specific: a learner expert in one topic of a course may still be a novice in the next, so reversal must be assessed per topic, not per student.

Open questions include how finely expertise must be measured (domain-specific vs. general ability) and how quickly guidance should be faded within a single lesson versus across a curriculum. Three sources are recorded above: a narrative review by the group that named the effect (Kalyuga et al. 2003), one of its three-experiment primary studies (Kalyuga et al. 1998, circuit diagrams, read from the abstract), and a conceptual paper restating the review (Obikwelu 2013). None prints an effect size in the text read, and no meta-analysis is recorded.

Merged from “Multimedia Principles Benefit Novices” (multimedia-principles-benefit-novices): Why novices benefit more. The theoretical rationale comes from Cognitive Load Theory: novices lack the prior-knowledge schemas that let experienced learners compensate for poorly designed materials. Principles such as integrating words and images, signaling essential content, and segmenting animation reduce extraneous processing precisely for learners whose working memory is most easily overwhelmed — see Cognitive overload degrades learning and Chunking reduces working memory load.

Expertise reversal as boundary condition. The same design features that help novices can hinder advanced learners, who must process redundant explanations or unnecessary graphics they no longer need — the expertise reversal effect. This means "apply all multimedia principles everywhere" is the wrong takeaway; designers should match design density to learner expertise and fade support as competence grows.

Scope caveat. The claim is a moderator claim (principles × prior knowledge), not a claim that any single principle works in all conditions. Effects depend on the principle in question, the medium, and the domain; some principles, such as removing irrelevant material, are well supported in their own right — see Coherence principle: irrelevant material hurts learning.

Design implications. For novice-facing materials, apply the principles at full strength: pair narration with images rather than on-screen text, segment continuous media into learner-paced parts, use pre-training to introduce key terms before the main explanation, and strip decorative content. As learners gain expertise, progressively relax these constraints — allow learners to skip redundant narration, permit denser integrated materials, and shift toward problem-centered formats such as worked examples faded into independent practice.

Open questions. Evidence entries are still needed to quantify how large the novice advantage is for each individual principle, and whether the moderator pattern holds for newer media (interactive simulations, video) as it does for classic text-and-graphics materials.

Learner Variables¶

• Prior Knowledge — moderator: an instructional effect differs with it

Related Claims¶

• Supplementing simulation-based inquiry with learner guidance has a modest effect on learning outcomes (d = 0.49) and inquiry skills (d = 0.41), independent of guidance type (review attribution to D'Angelo et al., 2014) — the general claim this finding bears on, with the studies behind it

• In one experiment with 256 undergraduates, computer-based lesson versions that included practice produced significantly higher posttest scores than versions without practice — the general claim this finding bears on, with the studies behind it

• Multimedia Principle Improves Learning — the general claim this finding bears on, with the studies behind it

• Contingent scaffolding improves learning more than fixed or absent support. — the general claim this finding bears on, with the studies behind it

• Chunking reduces working memory load by grouping information into fewer, more meaningful units.

• Worked examples can become redundant or counterproductive for advanced learners.
• Fading support promotes the transfer of responsibility from instructor to learner.
• Minimal guidance is less effective for novices than explicit instruction — related
• Expertise reversal effect — the theoretical account in cognitive load theory
• Cognitive overload degrades learning. — the load mechanism by which redundant guidance harms experts
• Presenting words as spoken narration rather than on-screen text alongside graphics improves learning — related
• Split Attention Effect Degrades Learning — related
• Redundancy Effect Impairs Learning — related
• On-screen text that duplicates the narration of a diagram can hurt learning, but spoken–written redundancy is not harmful on average and helps over spoken-only presentation in some conditions — related
• Expert performers counteract automaticity by building increasingly complex mental representations, unlike everyday skill learning — related
• Redundant text in the diagram did not affect posttest accuracy or difficulty ratings — related
• Current serious games give all learners the same unregulated scaffolding, which the authors argue contradicts the notion of scaffolding — a narrower finding that bears on this claim
• The FTS proposes that fading distinguishes expert from non-expert peers via a game-progress threshold, and peer-tutoring collaboration is expected to improve non-expert game progress and knowledge gained — related
• Expertise reversal: cognitive load effects for novices decrease, disappear, and can reverse as expertise increases — a broader claim this one bears on
• The advantage of guidance begins to recede only when learners have sufficiently high prior knowledge to provide internal guidance — possibly the same claim (merge candidate)
• Cognitive load effects including integration benefits may reverse for more experienced learners, whose schemas make integrated additional information redundant — a narrower finding that bears on this claim
• Coherence principle: irrelevant material hurts learning — one of the core multimedia principles whose benefit is strongest for novices
• Cognitive Load Theory — the theoretical framework underlying the multimedia principles
• In one author's classroom media experiments, intuitive learners tended to score above sensing learners, though not in every format comparison — related
• Four of six activity groups are rated better in multimedia format, but quizzes and picture descriptions are rated equally across formats — related
• A practice article reports that the studies it cites found better post-test outcomes from words with relevant pictures than from words alone, especially with spoken words — related
• Learners with high prior knowledge benefit more from segmenting in retention performance than learners with no or low prior knowledge — related
• Design principle effects are larger in system-paced and in complex media, and learner prior knowledge was not a significant moderator (p = 0.14) — reports the opposite
• Split-attention worked examples lose their advantage over conventional practice, while integrated formats restore it — a narrower finding that bears on this claim
• Experiment 5 showed superior performance for learners who had previously studied isolated elements followed by interacting elements instruction, supporting expertise as the key factor in the technique's benefits — related
• Segmentation of dynamic visualizations shows an expertise reversal effect: students with lower prior knowledge benefit, while students with higher prior knowledge learn equally well with or without segmentation — a narrower finding that bears on this claim
• Adaptive fading of worked examples based on current skill level produced better learning outcomes than fixed fading or tutored problem solving — related
• Adapting instruction to individual learners requires more advanced measurement of expertise and cognitive load — related
• Complex learning is lengthy, requiring instruction to account for learners' motivational states and expertise development — a broader claim this one bears on
• Within cognitive load theory, the element interactivity and expertise reversal effects are usually not treated as closely related effects — related
• Iconic visual scaffolds in chemistry visualizations helped low-prior-knowledge high-school students but not higher-prior-knowledge ones, with the effect eliminated for younger middle-school children — a narrower finding that bears on this claim
• Embedded explanatory notes improved Shakespearean text comprehension for novice readers but reversed for experts — a narrower finding that bears on this claim
• The positive effect of integrated illustrations on transfer is stronger for students who lack experience in the text's domain (meteorology) than for high-experience students — a narrower finding that bears on this claim
• Phase 4 low-guidance procedure was truncated for children who already demonstrated unconfounded designs — a narrower finding that bears on this claim
• Expertise reversal: high-guidance worked examples become ineffective or detrimental once learners have developed cognitive schemata — possibly the same claim (merge candidate)
• The imagination effect (cognitive rehearsal of example procedures) enhances learning only for learners with prior knowledge — a narrower finding that bears on this claim
• The expertise reversal effect reflects the strength of experts' knowledge base resisting redundant external information, and a motivation-related form is predicted — a narrower finding that bears on this claim
• The worked example effect holds for the schema-acquisition goal, but other instructional goals may be better served by other means — related
• Revised visuals reduce cognitive processing of math-relevant visuals for low prior knowledge students but increase it for middle prior knowledge students — a narrower finding that bears on this claim
• Revised visuals reduce cognitive processing of lesson text for low prior knowledge students but increase it for middle prior knowledge students — a narrower finding that bears on this claim
• The worked example effect is explained by reduced element interactivity relative to problem solving, and reverses with expertise — related
• The effectiveness of instructional techniques and procedures interacts with levels of learner prior knowledge, a pattern known as the expertise reversal effect — a broader claim this one bears on
• The review reports that worked examples benefit inexperienced learners but become redundant with increased experience, when problem solving proves superior — a narrower finding that bears on this claim
• Revised visuals reduce cognitive load while reading lesson text for low prior knowledge students but increase it for middle prior knowledge students — a narrower finding that bears on this claim
• Empirical patterns of the expertise reversal effect fit those found in Aptitude Treatment Interaction studies initiated in the mid-60s — related
• Experts have reduced access to memory traces of their own cognitive processes due to automatization, and large domain knowledge can disadvantage performance on certain tasks — related
• Definitions, checkpoints, and staged pacing improved conceptual clarity for novice programmers — a narrower finding that bears on this claim
• Collaboration facilitates learning only under specific knowledge and experience conditions; with advanced learners it can be detrimental via the expertise reversal effect — a narrower finding that bears on this claim
• The prior knowledge of the intended audience determines a text's information density, rate of new information, and accompanying background — related
• The review reports that the modality effect reverses with expertise: audio-visual presentations helped inexperienced learners, but after further training a diagram-only condition showed a substantial advantage — a narrower finding that bears on this claim
• Element interactivity and expertise reversal effects: cognitive load effects diminish and may reverse as expertise increases because known elements chunk into single elements — a broader claim this one bears on
• Design guidelines beneficial for novice learners can be ineffective or even detrimental when applied to experts (expertise reversal) — possibly the same claim (merge candidate)
• Redundancy effect: when one source of information is intelligible in isolation, physically integrating it with another source fails to help and eliminating the redundant source facilitates learning — a narrower finding that bears on this claim
• Experts introduce procedure and planning explanations over time, typically after an initial step or action is completed — related
• Low variability materials enhanced learning regardless of guidance; expertise reversal found for variability in Experiment 2 — a narrower finding that bears on this claim
• Expertise reversal effect: instructional aids effective for low-knowledge learners hinder high-knowledge learners across all three layers — possibly the same claim (merge candidate)
• Expertise reversal effect: learners with developed fluency do not benefit from unnecessarily explicit instruction and benefit more from open problem-solving approaches — possibly the same claim (merge candidate)
• Moderate levels of interactivity tend to be most beneficial for learning, with benefits depending on task integration and assessment format — related
• Scaffolding effectiveness depends on knowledge base: high-guidance scaffolding suits less advanced learners, high-self-regulation scaffolding suits advanced learners — a narrower finding that bears on this claim
• Imagining worked examples benefits learners with adequate prior knowledge but harms performance for low-knowledge learners — a narrower finding that bears on this claim
• The expertise reversal effect relies on changes in relative levels of element interactivity due to changes in relative levels of expertise — related
• The review reports that the isolated-elements mixed instructional method benefits novices but the benefit disappears for experienced learners, and that the imagination effect reverses with expertise — a narrower finding that bears on this claim
• Cognitive Load Management — related
• Example-first conditions outperformed problem-first conditions on an immediate circuit test — related
• Instructional method and prior knowledge interact on conceptual understanding of statistics (expertise reversal effect) — a narrower finding that bears on this claim
• Element interactivity moderates the effectiveness of cognitive offloading, with greater benefit for high element interactivity materials — related
• Feedback benefits are stronger for children with low prior knowledge, resembling an expertise-reversal effect — a narrower finding that bears on this claim
• The Perform First Hypothesis proposes direct sensorimotor action early in expertise and observing, imagining, or abstract representations later — related
• On the immediate test, novices showed a worked example effect (high guidance superior) while more knowledgeable learners showed a generation effect (low guidance superior) — a narrower finding that bears on this claim
• The thesis hypothesises that with increasing expertise the worked example effect may be eliminated or reversed while the generation effect remains robust — a narrower finding that bears on this claim