Learner-constructed graphic organizers are not shown to outperform provided ones: the one direct test, with college readers, favoured provided organizers on transfer¶

Claim · All claims
Evidence · 2 studies · 1 causal r?, 1 quant-synthesis r? · q3–q4 · i2 medium
Currency · newest study 2018 · not yet checked against its sources · not reviewed by a person

The hypothesis that learners who construct their own graphic organizers (concept maps, diagrams, knowledge maps) learn more deeply than learners who study provided ones, because construction forces active selection, organization and integration, is not supported by the evidence recorded here. The one controlled head-to-head comparison (Stull & Mayer 2007, three experiments with college readers of a textbook passage) found provided organizers produced better transfer and no difference in retention; a meta-analysis (Schroeder et al. 2018) found a larger effect for constructing concept maps than for studying them, but each against its own comparison conditions, so it does not test the two against each other.

Subclaims¶

q4 i2 A meta-analysis of 142 effect sizes finds that creating concept maps (g = 0.72) shows a larger benefit over its comparison conditions than studying provided concept maps (g = 0.43), but this is a moderator contrast across different comparisons, not a head-to-head test of constructing versus studying. → Schroeder et al. 2018

q3 i2 In three direct experiments with readers of a passage, author-provided graphic organizers beat learner-constructed ones on transfer (d = 0.24 ns, 0.43, 0.84, with the gap growing as the construction task got simpler), with no significant differences on retention, which contradicts the claim for this setting. → Stull & Mayer 2007

Evidence¶

Schroeder et al. 2018¶

Schroeder, N. L., Nesbit, J. C., Anguiano, C. J., & Adesope, O. O. (2018). Studying and constructing concept maps: A meta-analysis. Educational Psychology Review, 30(2), 431–455. doi:10.1007/s10648-017-9403-9

q4 · random-effects meta-analysis · i2 · medium effect for constructing, g=0.72; studying g=0.43 · n=142 effect sizes (11,814 participants) · quant-synthesis · r?

A random-effects meta-analysis of 142 independent effect sizes from 11,814 participants that used concept and knowledge maps for learning. Overall, learning with maps had a moderate effect (g = 0.58). In a moderator analysis, constructing concept maps (g = 0.72) was associated with a larger benefit than studying provided maps (g = 0.43). Each subgroup was measured against its own comparison conditions, so the gap suggests an advantage for construction but does not directly compare the two in matched designs. Read from the publisher abstract only; the moderator statistics and the comparison conditions behind each subgroup were not checked.

Stull & Mayer 2007¶

Stull, A. T., & Mayer, R. E. (2007). Learning by doing versus learning by viewing: Three experimental comparisons of learner-generated versus author-provided graphic organizers. Journal of Educational Psychology, 99(4), 808–820. doi:10.1037/0022-0663.99.4.808

q3 · three peer-reviewed experiments · i2 · transfer favoured provided organizers, d=0.24 (ns), 0.43, 0.84 · n=3 experiments (participant counts not in the abstract) · causal · r?

In three experiments, learners read a passage that came either with author-provided graphic organizers or with space or templates for building their own, then took retention and transfer tests. The construction task went from most complex (27 organizers built in margin space) to least complex (10 templates). On transfer the provided-organizer group did better every time: d = 0.24 (not significant), 0.43 (p < .05) and 0.84 (p < .01). On retention there were no significant differences. The authors say the results are consistent with cognitive load theory, which posits that excessive activity can create extraneous cognitive load that disrupts generative processing; extraneous load itself is not reported as measured in the abstract. This directly contradicts the claim for studying a text passage.

Studies bearing on this claim¶

Does building a representation of the content teach more than studying one that is provided? This table gathers the studies behind this claim and the 5 narrower claims that bear on it: 7 distinct studies (3 quant-synthesis, 3 causal, 1 design), q2–q4, 4 of 7 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 direct evidence runs against the title. The only head-to-head test below, Stull & Mayer (2007), found provided organizers produced better transfer than learner-built ones across three experiments, with no difference on retention. The meta-analytic comparison favouring construction (Schroeder et al. 2018) sets each condition against its own controls, so it does not test the claim directly. The title now says so.

Generation as the mechanism. The proposed advantage of learner-constructed organizers rests on generation: deciding which concepts matter, how they relate, and how to represent them spatially engages selective organizing processes that studying a finished diagram can bypass. This aligns with the broader pattern that annotating improves learning and that activation improves learning — tasks requiring learners to produce structure outperform tasks requiring only recognition of structure. Dual coding may add a further benefit when learners translate verbal material into spatial form (dual coding theory).

Costs of construction. Building an organizer imposes its own load: learners must coordinate content analysis with layout decisions. Where working-memory resources are scarce, a well-designed provided organizer may function as an advance organizer that reduces working-memory load rather than adding to it. The comparison is therefore likely conditional on learner expertise and material difficulty, in line with cognitive load theory and the expertise reversal effect: provided organizers may help novices most, while construction benefits grow with developing skill.

Quality of the constructed product. A poorly constructed organizer can encode misconceptions or omit key relations; without feedback or a reference structure, construction may consolidate errors. Scaffolds such as partially completed maps or node/link constraints may capture much of the generation benefit while limiting these risks — a fading logic parallel to that used with worked examples.

What a decisive test would look like. A strong comparison would hold content and time constant across a construction condition and a provided-organizer condition, measure both immediate and delayed outcomes, and include a no-organizer control to isolate whether provided organizers help, harm, or merely fall short of construction. Outcome measures matter: construction should show its largest advantage on transfer and relational-knowledge items, and may show none on simple recall, where a provided organizer's chunking support could suffice.

Open questions. Two entries are recorded: one set of three direct experiments (against the hypothesis, for college readers of a text) and one meta-analysis whose comparison is indirect. Missing are direct comparisons in other settings and with younger learners. Key moderators to establish empirically: learner expertise, domain (well-structured vs. ill-structured content), whether provided organizers are complete or partial, and whether outcome measures reward recall or transfer. Until further controlled comparisons are added, the advantage of construction should be treated as a hypothesis the one direct test contradicts.

Merged from “Learner Generated Organizers Beat Provided Organizers” (learner-generated-organizers-beat-provided-organizers): Mechanism. Generation requires learners to select main ideas, decide on relational structure, and map content onto that structure — activities consistent with generative processing and with the benefits described in Activation improves learning and Learner highlighting of text gives a small memory benefit but no reliable comprehension benefit, and evidence on other forms of annotation is not yet recorded here. A provided organizer, by contrast, can be skimmed passively, and may impose its own structure rather than the learner's. Self-constructed organizers also serve as a form of retrieval and elaboration: deciding what belongs at the top of an outline or which nodes connect in a concept map forces learners to reorganize material in their own words rather than recognize someone else's.

Boundary conditions. The advantage of generation is likely moderated by learner expertise and by the quality of the generated organizer. Novices may lack the knowledge to construct a useful structure, in which case a well-designed provided organizer (see Advance organizers improve learning) may outperform a poor self-generated one — an instance of the expertise reversal pattern described in Expertise reversal effect. Generation also costs time and working-memory resources, so under high cognitive load the benefit may shrink or reverse; scaffolds such as partially completed concept maps or provided node lists may capture much of the benefit while limiting the cost. A further caveat: a poorly constructed organizer can entrench misconceptions, since the learner's erroneous relational structure goes uncorrected unless reviewed.

Open questions. The direct evidence runs against the title. The only head-to-head test recorded above, Stull & Mayer (2007), found provided organizers produced better transfer than learner-built ones across three experiments, with no difference on retention. The meta-analysis favouring construction (Schroeder et al. 2018) sets each condition against its own controls, so it does not test the claim directly. The page was retitled on 2026-10-05 to say so. Key comparisons to document include same-time-on-task controls (generation takes longer, so unequal-time comparisons overstate the advantage) and measures of quality of the generated organizer as a mediator of learning gains.

Related Claims¶

• Advance organizers improve learning. — provided organizers are the comparison condition; helps define when giving structure beats asking learners to build it
• Annotating improves learning. — a related generation task in which learners actively mark up material rather than receive it pre-processed
• Chunking reduces working memory load. — organizers work partly by chunking related concepts; relevant to the load cost of construction
• Cognitive overload degrades learning. — construction demands can overload learners for whom a provided organizer would have helped
• Cognitive load theory. — frames the trade-off between the generative benefit and the extraneous load of construction
• Expertise reversal effect. — predicts the advantage of construction grows while the value of provided organizers shrinks as expertise develops
• Studying worked examples improves novices' algebra performance on problems with the same structure as the examples — analogous provided-structure condition; its fading logic suggests how to scaffold organizer construction
• Advance organizers improve learning — provided organizers are the comparison condition this claim is measured against
• Learner highlighting of text gives a small memory benefit but no reliable comprehension benefit, and evidence on other forms of annotation is not yet recorded here — another learner-generated processing activity with similar generative mechanisms
• Chunking reduces working memory load — organizers work by imposing structure that chunks content
• Cognitive overload degrades learning — generation costs resources that overloaded learners may not have
• Concept mapping improves learning over most comparison conditions by a moderate average amount, but not over time-matched retrieval practice — related
• Adding relevant pictures or graphics to text improves learning from it, with a small pooled effect on reading comprehension, though whether dual coding explains the benefit is disputed — related
• Graphic Organizers Support Novice Comprehension — related
• Generative Learning Improves Comprehension — related
• Productive Failure Improves Conceptual Learning — related
• A map risks becoming a new metanarrative unless it is continually remapped and readers actively construct their own maps — related
• Self-constructing a complete concept map produces better delayed recall than completing, sorting a map, or writing a summary, with the largest effect in the I+C structure — a narrower finding that bears on this claim
• Merely adding an expert example concept map does not add to the value of students creating their own concept maps for this population — related
• 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 — related
• Active-processing focus-support techniques such as testing, self-explaining, concept mapping, and drawing show positive effects on learning — related
• Self-generated concept mapping benefits learning from grade 4 onwards with similar effect sizes across grades 4–8, 9–12, and college — related
• Sequencing generated drawing before viewing a provided visual improves transfer, suggesting provided visuals work best as feedback — related