Evidence for Using Concept Maps #

Every substantive claim on the Using Concept Maps page is checked against current research. Here is each claim, how well today’s evidence supports it, and the sources. The full, de-duplicated source list lives on the references page.

Supported · strong evidence — The most valuable way to use a concept map for memory is to build (or redraw) it from memory and then check it against the source, because the retrieval attempt itself is what produces durable learning.

Karpicke & Blunt (2011, Science) directly tested concept mapping against retrieval practice and found that practising retrieval (free recall of the material) produced substantially more learning on later tests than elaborative concept mapping done with the source available. Reframing mapping as a retrieval task – close the source, reproduce the map, then check – aligns the technique with the strongest, most replicated finding in the study-strategy literature (the testing effect). The page’s emphasis on retrieving from memory before checking is exactly the manipulation the evidence supports.

Sources: Karpicke, J. D., & Blunt, J. R. (2011), Retrieval practice produces more learning than elaborative studying with concept mapping. Science, 331(6018), 772-775 – https://doi.org/10.1126/science.1199327 · Roediger, H. L., & Karpicke, J. D. (2006), Test-enhanced learning: Taking memory tests improves long-term retention. Psychological Science, 17(3), 249-255 – https://doi.org/10.1111/j.1467-9280.2006.01693.x · full reference ›

Supported · strong evidence — Retrieval practice (self-testing) is one of the highest-utility study strategies, so turning a map into a self-test by blanking out nodes or link phrases puts it to work as a tested-memory tool rather than a rereading exercise.

Dunlosky et al.’s (2013) review rated practice testing as one of only two techniques with ‘high’ utility across learners, materials and settings; the redraw-and-check and skeleton-self-test forms the page describes are practice testing applied to map content. The claim that self-testing on the map’s structure beats passively rereading it is well supported.

Sources: Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham, D. T. (2013), Improving students’ learning with effective learning techniques. Psychological Science in the Public Interest, 14(1), 4-58 – https://doi.org/10.1177/1529100612453266 · full reference ›

Supported · moderate evidence — A concept map is a delivery vehicle for proven learning methods (a tool for elaboration and organisation), not a memory engine in its own right; constructing one beats studying a ready-made one, but its edge over other active, effortful methods is modest.

Nesbit & Adesope’s (2006) meta-analysis found concept-map use associated with modest-to-moderate gains in retention and transfer, with effect sizes varying widely by how the map was used and the comparison condition – larger against passive controls (reading text, attending lectures) and smaller against other active strategies. This supports treating mapping as one good format among several rather than a uniquely powerful method, and is consistent with Dunlosky et al.’s (2013) lower rating for imagery/organisational techniques relative to retrieval practice. The honest ‘modest edge’ framing matches the evidence.

Sources: Nesbit, J. C., & Adesope, O. O. (2006), Learning with concept and knowledge maps: A meta-analysis. Review of Educational Research, 76(3), 413-448 – https://doi.org/10.3102/00346543076003413 · Dunlosky, J., et al. (2013), Improving students’ learning with effective learning techniques. Psychological Science in the Public Interest, 14(1), 4-58 – https://doi.org/10.1177/1529100612453266 · full reference ›

Supported · moderate evidence — Choose the map format to fit the material and the task in front of you (a branching argument, a process, a set of causes), trying a few formats and keeping the one that carries the content with least strain.

Matching the representation to the structure of the material is consistent with the moderator findings in the concept-map meta-analyses: benefits depend on how and where maps are used rather than on the format being intrinsically superior. This is a sensible task-analysis recommendation rather than a strong empirical law, hence moderate strength. Crucially it is matching to the TASK, not to a supposed learner ‘style’ – see the styles row.

Refuted · strong evidence — Map choice should fit the material and task, NOT a learner’s supposed ’learning style’ – matching instruction to a preferred style does not improve learning.

The styles-matching (‘meshing’) hypothesis is refuted: Pashler, McDaniel, Rohrer & Bjork (2008) found that the studies meeting the experimental criteria needed to test it provided no support, and several found the opposite of what the hypothesis predicts; this remains the consensus in 2026. The page is rated ‘refuted’ on the styles claim because it correctly tells readers to match the map to the task and material rather than to a personal style – it teaches the evidence-based position, so the page is concordant while the styles-matching idea it rejects is the refuted one.

Sources: Pashler, H., McDaniel, M., Rohrer, D., & Bjork, R. (2008), Learning styles: Concepts and evidence. Psychological Science in the Public Interest, 9(3), 105-119 – https://doi.org/10.1111/j.1539-6053.2009.01038.x · full reference ›

Supported · moderate evidence — Concept mapping as a teaching tool rests on Ausubel’s principle that meaningful learning happens when new knowledge is tied to relevant concepts the learner already has – ’the most important single factor influencing learning is what the learner already knows’.

Novak built IHMC concept mapping on Ausubel’s assimilation theory, and the foundational claim – that integrating new material with relevant prior knowledge produces meaningful, durable learning rather than rote learning – is well aligned with modern accounts of schema construction, elaboration and the generative/prior-knowledge effects. It is a foundational theoretical principle (1968) rather than a single quantified result, so moderate strength; the modern mechanism (prior knowledge as the scaffold for encoding) remains consensus.

Sources: Ausubel, D. P. (1968), Educational Psychology: A Cognitive View. Holt, Rinehart & Winston · Fiorella, L., & Mayer, R. E. (2016), Eight ways to promote generative learning. Educational Psychology Review, 28(4), 717-741 – https://doi.org/10.1007/s10648-015-9348-9 · full reference ›

Supported · moderate evidence — A dense, fully-linked expert map can overload a novice because cognitive load goes into reading the diagram rather than learning the content, so beginners should start from skeleton or fill-in maps and move toward free construction as their knowledge grows (scaffold then fade).

This is a direct application of the expertise-reversal effect (Kalyuga, Ayres, Chandler & Sweller, 2003): instructional support that helps novices becomes redundant and can impair more knowledgeable learners, and detailed representations that aid experts can overload novices. The prescription – high guidance (skeleton/fill-in maps) early, faded to free construction later – follows from the effect and from worked-example/guidance-fading research. It is well-grounded in cognitive load theory; rated moderate because the optimal fading point is task- and learner-dependent rather than fixed.

Sources: Kalyuga, S., Ayres, P., Chandler, P., & Sweller, J. (2003), The expertise reversal effect. Educational Psychologist, 38(1), 23-31 – https://doi.org/10.1207/S15326985EP3801_4 · full reference ›

Supported · moderate evidence — Drawing a map from memory with the source closed is a low-stakes way to surface what a learner can actually reproduce (versus what merely feels familiar), and pairing it with a prediction of how much they expect to recall turns the gap between expected and actual into a calibration signal.

Two well-supported ideas combine here. First, retrieval (drawing from memory) both measures and builds learning – it does double duty (Karpicke & Blunt 2011; Roediger & Karpicke 2006). Second, rereading and viewing produce fluency that inflates confidence (the illusion of competence / overconfidence from familiarity), while a retrieval attempt gives a more honest read of what is actually stored, and predict-then-test exposes the calibration gap. The metacognitive-calibration claim is supported in the broad literature; rated moderate because calibration improvements from a single map-from-memory exercise are smaller and more variable than the core retrieval benefit.

Sources: Karpicke, J. D., & Blunt, J. R. (2011), Retrieval practice produces more learning than elaborative studying with concept mapping. Science, 331(6018), 772-775 – https://doi.org/10.1126/science.1199327 · Roediger, H. L., & Karpicke, J. D. (2006), Test-enhanced learning. Psychological Science, 17(3), 249-255 – https://doi.org/10.1111/j.1467-9280.2006.01693.x · full reference ›