Learning Science Evidence Map: What the Research Actually Shows

This is an evidence map for the core findings of learning science - a structured, sourced reference to what the research actually shows about how people learn durably. It is built for a specific use: when you want to know not just that a technique works, but how strong the evidence is, where it comes from, and what it does not claim.

Every entry links the claim to a named, peer-reviewed source you can verify yourself. Where the evidence is strong, we say so. Where it is mixed, narrow, or contested, we say that too. This page is maintained as a reference, not a one-time roundup.

How to read this map

Each technique below is rated on two axes that are easy to conflate but should be kept separate:

• Evidence strength - how well-replicated and methodologically sound the supporting research is. "Strong" means multiple independent replications, often synthesized in meta-analyses; "moderate" means real support with meaningful caveats or narrower conditions; "contested/limited" means the picture is genuinely unsettled.
• Effect type - what the technique actually improves. Many learning techniques help long-term retention and transfer while feeling worse in the moment than easier strategies. This gap between performance-during-study and durable-learning is itself one of the most robust findings in the field, and it is why learner intuition is an unreliable guide.

A note on effect sizes: where we cite them, they are drawn from the named meta-analysis and should be read as that synthesis reported them, not as universal constants. Effect sizes vary with population, material, retention interval, and study design. Treat them as directional evidence of strength, not precise predictions.

The evidence map

Technique	What it is	Evidence strength	What the evidence supports	Key source
Retrieval practice (testing effect)	Attempting to recall information from memory, rather than re-reading it.	Strong	Substantially better long-term retention than restudying the same material; one of the highest-utility techniques reviewed across the literature.	Roediger & Karpicke (2006); Dunlosky et al. (2013)
Spaced practice (distributed practice)	Distributing study across time rather than massing it into one session.	Strong	Reliably better long-term retention than massed practice across content types and ages; the optimal gap grows with the target retention interval.	Cepeda et al. (2006)
Interleaving	Mixing different problem types or topics within a session rather than blocking them.	Moderate-Strong	Better discrimination learning and transfer, especially in maths and category learning; feels harder during practice (a desirable difficulty).	Rohrer & Taylor (2007); Kornell & Bjork (2008)
Elaborative interrogation	Asking and answering "why is this true?" for facts being learned.	Moderate	Improves retention of factual material; benefits depend on prior knowledge and are clearest for discrete facts.	Dunlosky et al. (2013)
Self-explanation	Explaining to yourself how new information relates to what you know, or the steps of a worked example.	Moderate	Improves learning and transfer across tasks; effortful and not always done well without prompting.	Dunlosky et al. (2013)
Worked examples (early learning)	Studying fully worked solutions before attempting problems, when a domain is new.	Moderate-Strong	Reduces extraneous cognitive load and accelerates early skill acquisition; the advantage fades and can reverse as expertise grows (the expertise-reversal effect).	Sweller (1988)
Highlighting / re-reading	Marking text or reading it again - the most common student strategies.	Strong evidence of LOW utility	Popular and low-effort, but consistently among the least effective techniques for durable learning; produces fluency that is easily mistaken for mastery.	Dunlosky et al. (2013)
Matching instruction to "learning styles"	Teaching to a learner's preferred modality (visual/auditory/kinaesthetic).	Contested - not supported	The meshing hypothesis (matching modality improves outcomes) has repeatedly failed to find support in controlled tests, despite widespread belief.	Pashler et al. (2008)

The search transparency gap: an original measurement

The table above reflects the scientific record. It does not tell you what a typical reader actually encounters when they search for these study techniques. So we measured that directly: for each technique, we pulled the top organic web search results and checked, page by page, whether the result disclosed any evidence-strength caveat relevant to the specific claim, a lack of controlled support, a contested status, a boundary condition, or (for the two low-utility techniques) an explicit statement that the popular belief is not well supported.

This is original data. It is not copied from any source; it is our own one-time measurement, dated below, and the methodology is disclosed so it can be checked or repeated.

In one line: the clearest myth on this page, matching instruction to "learning styles", is also the least disclosed. Only about a quarter of top search results mention that the idea has repeatedly failed controlled tests, while the two techniques we rate as popular-but-low-utility (highlighting and self-explanation coverage) are disclosed almost everywhere.

Technique or claim	Results checked	Disclose the evidence caveat
Learning styles (meshing hypothesis)	7	28.6% (2 of 7)
Spaced practice	10	40% (4 of 10)
Worked examples	8	50% (4 of 8)
Interleaving	8	62.5% (5 of 8)
Elaborative interrogation	7	71.4% (5 of 7)
Retrieval practice	10	80% (8 of 10)
Highlighting / re-reading	8	100% (8 of 8)
Self-explanation	5	100% (5 of 5)

The pattern here runs differently from a simple "popular claims get less scrutiny" story. Highlighting and re-reading are extremely popular study habits, and yet every usable result we checked disclosed that the research rates them low-utility; the low-utility finding itself has become common knowledge, so the disclosure IS the popular narrative now. Learning styles is the opposite case: it is arguably the single most debunked idea in education research, cited in dozens of studies as a myth, but the pages that rank for the term overwhelmingly describe how to identify and teach to a "style" rather than noting the idea does not hold up.

A reader who searches "learning styles" today is more likely to get a quiz than a correction.

Two findings that frame everything above

1. Desirable difficulties

Several of the most effective techniques - retrieval, spacing, interleaving - share a counterintuitive property: they make studying feel harder and slower in the moment, yet produce stronger long-term retention. Conditions that speed up apparent learning during practice often weaken it long term, and vice versa.

This is Robert Bjork's concept of "desirable difficulties," and it explains why learners systematically prefer less effective strategies: the easy ones feel like they are working.

2. The performance-learning distinction

Performance is what you can do during or immediately after study, under supportive conditions. Learning is what remains after a delay and generalises to new contexts. They frequently dissociate.

Because learners and instructors usually observe performance, not learning, intuitive judgments about "what's working" are unreliable - which is the practical reason this evidence map exists.

Methodology and scope notes

• What counts as evidence here: peer-reviewed experimental research and meta-analyses/syntheses. Where a row cites a synthesis (e.g. Dunlosky et al., 2013), the rating reflects that synthesis's assessment of utility, not a single study.
• Effect sizes are conditional. They depend on population, material type, retention interval, and design. This map reports evidence strength and direction, not guaranteed magnitudes for any individual.
• Lab vs. classroom. Much of the strongest evidence comes from controlled studies; classroom replications generally support the same directions but with smaller and more variable effects. Treat real-world gains as smaller than lab maxima.
• This is not advice for a specific person or condition. It summarises general findings about typical learners; it is not individualised educational, clinical, or accessibility guidance.
• Maintenance: this page is updated when the underlying evidence base changes materially, not on a fixed schedule. A visible "last reviewed" date reflects a real review.
• Search transparency gap methodology: for each technique, we searched its plain common name and checked the top 10 organic (non-ad, non-video) results, real general-audience articles, encyclopedia entries, or reference pages. Pages that could not be fetched or parsed were excluded from the denominator rather than counted as non-disclosing, so sample sizes actually checked ranged from 5 to 10 usable pages per technique. A single disclosing sentence anywhere on the page counted as disclosure, a low bar, so the percentages likely understate rather than overstate the real transparency gap. This is one search snapshot, not a tracked index; a different query phrasing would return a different set of pages. Measured August 2026; we plan to re-run this measurement periodically.

Sources checked: the raw audit

For transparency, every page we checked for the search transparency gap measurement above is listed below, one collapsible section per row, with our verdict and the exact disclosing sentence where one was found. This is the full working, not a summary of it.

Learning styles (meshing hypothesis): 2 of 7 disclosed

• en.wikipedia.org: disclosed. "Studies contradict the widespread "meshing hypothesis" that a student will learn best if taught in a method deemed appropriate for the student's learning style"
• apa.org: excluded, page unreadable
• success.uark.edu: not disclosed
• snhu.edu: disclosed. "I hear people say, 'I'm a hands-on learner,' which in their mind equates to a belief that they can't learn any other way."
• td.org: excluded, page unreadable
• serc.carleton.edu: not disclosed
• ebsco.com: not disclosed
• vark-learn.com: not disclosed
• ncbi.nlm.nih.gov: not disclosed
• nifa.usda.gov: excluded, page unreadable

Spaced practice: 4 of 10 disclosed

• en.wikipedia.org: disclosed. "Some papers find expanding intervals to be beneficial for recall. Other meta-analyses tend to conclude that both methods yield similar results."
• en.wikipedia.org: disclosed. "Shaughnessy (1977) found that the spacing effect is not robust for items presented twice after a 24-hour delay in testing."
• en.wikipedia.org: disclosed. "Not much attention has been given to the study of the spacing effect in long-term retention tests."
• oxfordlearning.com: not disclosed
• training.safetyculture.com: not disclosed
• maestrolearning.com: not disclosed
• thirdspacelearning.com: not disclosed
• lifehack.org: not disclosed
• justinmath.com: disclosed. "The current literature on mathematical models for determining optimal repetition spacing is limited to the setting of independent flashcard-like tasks."
• thedecisionlab.com: not disclosed

Worked examples: 4 of 8 disclosed

• en.wikipedia.org: disclosed. "As learners gain expertise in the subject area of interest, worked examples lose their effectiveness due to the expertise reversal effect."
• structural-learning.com: disclosed. "Helpful guidance for beginners can become annoying for experts."
• tandfonline.com: excluded, page unreadable
• innerdrive.co.uk: not disclosed
• psychologistworld.com: not disclosed
• eduaide.ai: excluded, page unreadable
• newsletter.jamieleeclark.com: disclosed. "In fact, worked examples are most powerful early in learning, when content is new and schemas are still forming."
• education.nsw.gov.au: disclosed. "While fully guided instruction is very effective for teaching students new material, it becomes less effective as students become more expert at a particular skill."
• tomneedhamteach.wordpress.com: not disclosed
• instructionaldesign.org: not disclosed

Interleaving: 5 of 8 disclosed

• effectiviology.com: disclosed. "in some cases, it's no better than blocked practice, and can even be worse"
• psychology.ucsd.edu: disclosed. "Examples of materials that interleaved practice can benefit include similar types of math problems, easily confused grammatical tenses, and similar classes of visual stimuli."
• arc.duke.edu: not disclosed
• structural-learning.com: disclosed. "the benefits may reverse if topics are too dissimilar, particularly with vocabulary learning"
• edweek.org: disclosed. "Neither massed practice nor interleaving alone results in deep mastery. It takes both."
• ctl.wustl.edu: not disclosed
• mlpp.pressbooks.pub: excluded, page unreadable
• cei.umn.edu: excluded, page unreadable
• coursera.org: not disclosed
• scientificamerican.com: disclosed. "Otherwise, interleaving can sometimes be more confusing than helpful."

Elaborative interrogation: 5 of 7 disclosed

• en.wikipedia.org: not disclosed
• vaia.com: excluded, page unreadable
• litfl.com: excluded, page unreadable
• learningscientists.org: not disclosed
• structural-learning.com: disclosed. "If students do not yet know enough to generate a plausible explanation, they will produce incorrect elaborations."
• arc.duke.edu: excluded, page unreadable
• learningscientists.org: disclosed. "Elaboration has been shown to help students who are more familiar with the topic, while those who are less familiar don't benefit as much."
• uwlax.edu: disclosed. "Some research indicates that students need background knowledge of a topic in order to produce good explanations."
• smowl.net: disclosed. "One of the main mistakes often made when implementing this technique in the classroom is that students do not have prior knowledge of a topic."
• frontiersin.org: disclosed. "Elaborative interrogation are especially beneficial for learners that already possess a relatively high amount of prior knowledge, but less for learners with rather low prior knowledge."

Retrieval practice: 8 of 10 disclosed

• en.wikipedia.org: disclosed. "Some researchers have applied aspects of cognitive load theory to suggest the testing effect may disappear with increasing task difficulty due to increased element interactivity."
• ctl.wustl.edu: disclosed. "One possible disadvantage to open-book settings is that they may have the unintended effect of reducing student time spent studying if the students know that the future activity will be open book."
• structural-learning.com: disclosed. "Feedback significantly changes how well retrieval practice works"
• studyingmachine.com: not disclosed
• mike-taylor.org: not disclosed
• ncbi.nlm.nih.gov: disclosed. "generalization from the laboratory context to educational environments may not be entirely warranted"
• ncbi.nlm.nih.gov: disclosed. "the testing effect is more robust when test-trials involve more effortful retrieval (e.g., recall vs. recognition)"
• ncbi.nlm.nih.gov: disclosed. "The backward testing effect is often quite small or even absent when the retention interval between study and test is relatively short."
• ncbi.nlm.nih.gov: disclosed. "A 2021 meta-analysis found virtually no advantage (Hedges' g = 0.095) of retrieval practice over various forms of elaborative learning strategies."
• frontiersin.org: disclosed. "virtually no advantage of retrieval practice over various forms of elaborative learning strategies"

Highlighting / re-reading: 8 of 8 disclosed

• psychologicalscience.org: disclosed. "Rereading and highlighting, seem to provide minimal benefits to their learning and performance."
• ideas.time.com: excluded, page unreadable
• edutopia.org: disclosed. "Most studies have shown no benefit to highlighting (as it is typically used) over and above the benefit of simply reading."
• learningcenter.unc.edu: disclosed. "Highlighting tends not to be a very effective reading strategy."
• kent.edu: disclosed. "Rereading and highlighting, seem to provide minimal benefits to their learning and performance."
• pubmed.ncbi.nlm.nih.gov: excluded, page unreadable
• facultyfocus.com: disclosed. "Research on study strategies rated rereading as having low utility."
• lawschooltoolbox.com: disclosed. "Both highlighting text and rereading text surprisingly received a Low Utility rating."
• innerdrive.co.uk: disclosed. "Compared to other learning strategies, such as Retrieval Practice, research has shown that highlighting to improve learning and memory tends to fare quite poorly"
• learningscientists.org: disclosed. "Never use highlighting as the ONLY means of studying."

Self-explanation: 5 of 5 disclosed

• learning.northeastern.edu: disclosed. "After-the-fact explanation also has an impact, but it is not as effective as explaining in the moment."
• psychology.ucsd.edu: disclosed. "Other methods may also be helpful (however there is less available research and/or only specific types of information may benefit)."
• lib.sfu.ca: excluded, page unreadable
• innerdrive.co.uk: disclosed. "Other research has suggested that self-explanation is more beneficial for subjects such as science and maths, as it is particularly suited to subjects that require reasoning or problem solving."
• biointeractive.org: excluded, page unreadable
• takinglearningseriously.com: disclosed. "Self-explaining will not be very effective if you have very little knowledge of the material."
• activerecalling.com: excluded, page unreadable
• kaporfoundation.org: disclosed. "Can be less effective if students do not have any prior knowledge to make connections with new material."
• ablconnect.harvard.edu: excluded, page unreadable

Sources

All sources are peer-reviewed and linked to their DOI or canonical record. These are the primary references behind the ratings above.

• Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham, D. T. (2013). Improving students' learning with effective techniques: Promising directions from cognitive and educational psychology. Psychological Science in the Public Interest, 14(1), 4-58. DOI: 10.1177/1529100612453266
• Roediger, H. L., & Karpicke, J. D. (2006). Test-enhanced learning: Taking memory tests improves long-term retention. Psychological Science, 17(3), 249-255. DOI: 10.1111/j.1467-9280.2006.01693.x
• Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin, 132(3), 354-380. DOI: 10.1037/0033-2909.132.3.354
• Kornell, N., & Bjork, R. A. (2008). Learning concepts and categories: Is spacing the "enemy of induction"? Psychological Science, 19(6), 585-592. DOI: 10.1111/j.1467-9280.2008.02127.x
• Rohrer, D., & Taylor, K. (2007). The shuffling of mathematics problems improves learning. Instructional Science, 35, 481-498. DOI: 10.1007/s11251-007-9015-8
• Sweller, J. (1988). Cognitive load during problem solving: Effects on learning. Cognitive Science, 12(2), 257-285. DOI: 10.1207/s15516709cog1202_4
• Pashler, H., McDaniel, M., Rohrer, D., & Bjork, R. (2008). Learning styles: Concepts and evidence. Psychological Science in the Public Interest, 9(3), 105-119. DOI: 10.1111/j.1539-6053.2009.01038.x
• Bjork, R. A., & Bjork, E. L. (1992). A new theory of disuse and an old theory of stimulus fluctuation. In A. Healy, S. Kosslyn, & R. Shiffrin (Eds.), From Learning Processes to Cognitive Processes (Vol. 2, pp. 35-67). Erlbaum.

Related reading on When Notes Fly

• How Learning Works: A Research Guide - the full pillar this map supports
• Spaced Repetition Explained
• The Testing Effect
• Deliberate Practice Explained
• What Is Cognitive Load

Frequently Asked Questions

What is a learning science evidence map?

It is a structured reference that rates each major learning technique by how strong the supporting research is and what it actually improves, with every claim linked to a named peer-reviewed source. It separates evidence strength from effect type, and is honest about where findings are contested.

Which learning techniques have the strongest evidence?

Retrieval practice (the testing effect) and spaced practice have the strongest, most replicated evidence for durable long-term retention. Interleaving and worked examples (early in learning) have moderate-to-strong support. Highlighting and re-reading are popular but have strong evidence of LOW effectiveness.

Do learning styles work?

The ‘meshing hypothesis’ - that matching instruction to a learner’s preferred modality (visual/auditory/kinaesthetic) improves outcomes - has repeatedly failed to find support in controlled experiments, despite being widely believed. The evidence does not support tailoring teaching to learning styles.

Why do effective study techniques feel harder?

Several of the most effective techniques (retrieval, spacing, interleaving) create ‘desirable difficulties’ - they feel harder and slower during study but produce stronger long-term retention. Easy strategies create fluency that learners mistake for mastery, which is why intuition is an unreliable guide to what works.