Evidence-Based Teaching: Complete Guide to What Works

Why Most Teaching Advice Is Wrong

I’ve been in classrooms for over a decade. Earth science, Seoul National University graduate, ADHD diagnosis at 31. In that time I’ve watched schools adopt learning styles theory, adopt it hard, build entire professional development programs around it, then quietly drop it when the research didn’t hold up. The same thing happened with brain gym exercises. And with the idea that students learn better when they control the pace completely. Good intentions, zero evidence.

Related: cognitive biases guide

This is the problem with education: it runs on intuition dressed as insight. Something feels true — visual learners need diagrams, auditory learners need lectures — so it spreads. Teachers adopt it, parents demand it, administrators mandate it. Meanwhile the actual cognitive science sits in journals that nobody reads.

If you’re a knowledge worker who manages, trains, mentors, or teaches anyone, this matters to you directly. Because the same broken intuitions that run classrooms run corporate training, onboarding programs, and team skill-building. You are almost certainly doing some of it wrong — not because you’re careless, but because the right information is buried and the wrong information is loud.

Here’s what the evidence actually says.

The Techniques That Don’t Work (Even Though They Feel Like They Do)

Learning Styles

The idea that people are visual, auditory, or kinesthetic learners and should be taught accordingly has been studied extensively. The verdict is clear. A comprehensive review by Pashler et al. (2008) examined whether matching instruction to learning style produces better outcomes. It does not. The “meshing hypothesis” — that matching style to content helps — has not been supported by any methodologically sound study. Not one.

This doesn’t mean all people learn identically. It means the visual/auditory/kinesthetic taxonomy is not the useful variable. What matters is the nature of the content, not a fixed trait of the learner. Spatial information is better understood visually. Sequential processes are better explained step-by-step. That’s about the material, not the person.

Massed Practice (“Cramming”)

Studying everything at once feels efficient. You’re in the material, you’re building momentum, the information feels accessible. That accessibility is exactly the problem. When retrieval feels easy, your brain doesn’t work hard to consolidate it. The material is still in short-term working memory, not encoded into long-term storage. Three days later, it’s gone.

This has been replicated so many times it’s one of the most robust findings in cognitive psychology. Yet cramming remains the default strategy for most people, including professionals preparing for certifications, presentations, and client meetings.

Re-reading and Highlighting

Both feel productive. Neither works particularly well as a learning strategy. Re-reading creates familiarity, which the brain interprets as knowledge. Highlighting gives the sensation of selecting what matters without forcing you to actually retrieve or use it. Dunlosky et al. (2013) conducted a systematic review of ten common study techniques and rated both highlighting and re-reading as having low utility for durable learning.

The Techniques That Actually Work

Retrieval Practice

Testing yourself is not just a way to measure what you know. It is a way to build what you know. Every time you successfully retrieve information, you strengthen the neural pathway to that information. The act of retrieval — struggling to pull something from memory — does more for retention than any amount of re-exposure to the material.

Roediger and Karpicke (2006) showed that students who studied a passage once and then took repeated retrieval practice tests dramatically outperformed students who spent the same time re-studying. On a test one week later, the retrieval practice group scored around 80% while the re-studying group scored around 40%. Same content, same time investment, completely different outcomes.

For practical application: close your notes and write down everything you remember. Use flashcards with the answer hidden. Explain the concept to someone without looking at your materials. Answer practice questions before you feel ready. That discomfort of not-quite-knowing is where the learning happens.

Spaced Practice

Instead of one long session, spread your learning across multiple shorter sessions with gaps between them. The forgetting that happens between sessions is not a failure — it is the mechanism. When you return to material you’ve partially forgotten and retrieve it again, the memory becomes significantly more durable than if you’d never forgotten it in the first place.

The spacing effect is one of the oldest findings in memory research, dating back to Ebbinghaus in the 19th century. It holds across virtually every domain tested: languages, mathematics, medical knowledge, procedural skills. For knowledge workers, this translates directly: don’t do all your preparation for a presentation the night before. Review the material, then return to it two days later, then again a week out. Your fluency on the day will be substantially better.

Interleaving

Most people practice one type of problem until they’re good at it, then move to the next type. This is called blocked practice, and it produces fast initial gains that don’t transfer well. Interleaving — mixing different problem types within a single practice session — feels harder, produces slower immediate progress, but results in significantly better performance on tests that use different formats or apply knowledge in new contexts.

The reason is similar to spacing: when you know the next problem will be the same type as the last, your brain takes a shortcut and applies the same approach without really re-evaluating. When problem types are mixed, you have to identify what kind of problem you’re facing before solving it. That identification process strengthens both conceptual understanding and flexible application.

For teaching others: resist the urge to organize practice sessions by topic. Mix problem types. It will feel less satisfying in the moment and produce better results over time.

Elaborative Interrogation

This means asking “why” and “how” while learning rather than accepting facts at face value. When you encounter a claim — say, that spaced practice outperforms massed practice — you ask: why would that be true? What mechanism explains it? How does it connect to what I already know about memory? This process of generating explanations forces you to integrate new information with existing knowledge structures, which is exactly how expertise is built.

The practical version: after reading a section of material, close the source and write an explanation of it in your own words, including your best attempt at explaining why it works the way it does. Where your explanation breaks down reveals exactly where your understanding is incomplete.

How Expertise Actually Develops

Deliberate Practice Is Not Just Repetition

Ten thousand hours of work produces expertise only if the work is the right kind. Anders Ericsson’s research on expert performance established that what separates elite performers from experienced amateurs is not time spent practicing — it’s the quality and structure of that practice. Deliberate practice means operating at the edge of your current ability, receiving immediate feedback on errors, and focusing intensely on specific weaknesses rather than running through things you can already do comfortably.

Most professional practice is not deliberate in this sense. A teacher who’s been teaching for twenty years but has never gotten systematic feedback on specific weak points and systematically worked to address them is not building expertise — they’re performing an established routine. Competence plateaus. Deliberate practice doesn’t.

The Role of Mental Models

Experts don’t just know more facts than novices. They organize knowledge differently. An expert chess player doesn’t see individual pieces — they see board configurations, patterns, strategic implications. An experienced surgeon doesn’t consciously process every instrument or movement — they perceive the surgical field as a structured whole with meaningful landmarks.

This chunking — organizing individual elements into meaningful patterns — is what allows experts to work faster, make fewer errors, and transfer skills to new situations. The educational implication is significant: teaching isolated facts is far less valuable than teaching the patterns and structures that connect facts into coherent systems. Schema first, detail second.

For knowledge workers building skill in a domain: seek out the underlying frameworks. What are the 5-7 core patterns that experts in this field recognize? Learning to perceive those patterns is more valuable than accumulating additional facts.

Teaching Other Adults Specifically

Adults Need Relevance Established First

Children will often learn material because an authority figure says it matters. Adults require a more compelling answer to “why does this apply to my situation right now?” This is not resistance — it’s a cognitive efficiency mechanism. Adult working memory is largely allocated to real ongoing problems. Information that doesn’t connect to those problems doesn’t get prioritized for encoding.

The practical implication: never lead with content. Lead with the problem the content solves. Not “today we’re going to learn about retrieval practice” but “you probably spend a lot of time preparing for things and feel underprepared anyway — here’s why that happens and what actually fixes it.” Problem first, mechanism second, technique third.

Worked Examples and Fading

When teaching a new skill, worked examples — where the expert solution is shown step-by-step — are more effective than problem-solving for novices. This seems counterintuitive; shouldn’t learners build understanding by struggling through problems? For novices, the struggle produces cognitive overload rather than productive learning because they don’t yet have the schemas to make sense of what they’re doing wrong.

The key is fading: as competence builds, progressively remove support. Start with a fully worked example. Then provide a partially worked example where the learner completes the final steps. Then provide the problem with hints. Then remove hints. This gradual transition from guided to independent performance is more effective than either extreme — complete guidance or immediate independent practice — for most learners in most domains.

Feedback Timing and Specificity

Feedback should be specific, timely, and actionable. “Good job” produces nothing. “Your explanation of the mechanism was clear, but you didn’t address what happens when the variable changes sign” gives the learner exactly what to work on. Feedback also needs to arrive close enough to the performance that the learner can connect it to specific decisions they made — delayed feedback on performance people can’t remember is largely useless.

One counterintuitive finding: immediate feedback during practice can actually reduce long-term retention compared to slightly delayed feedback. When feedback is instant, learners rely on it rather than developing their own error-detection. A short delay forces them to evaluate their own performance first, which itself is a valuable metacognitive skill.

What This Means for Your Practice Right Now

If you train people — whether you’re a manager running onboarding, a team lead upskilling your team, or a teacher in any formal sense — the gap between what works and what most organizations do is enormous. Most training is a single dense session, delivered to a passive audience, organized by topic, followed by no systematic retrieval practice. The retention rate from that format is somewhere between dismal and negligible.

The alternative doesn’t require more time. It requires different structure: shorter initial instruction, retrieval practice built into the session (not saved for a quiz at the end), spaced follow-up over subsequent days or weeks, mixed practice rather than blocked topics, and feedback that is specific enough to be actionable.

For your own learning, the principle is the same. Identify what you’re trying to learn. Design retrieval practice for it. Space your practice sessions. Mix topics rather than blocking them. And when something feels too easy, that’s usually a signal that you’re not learning — you’re performing something already consolidated, which feels good and does very little.

The research on this is not ambiguous. Dunlosky et al. (2013) evaluated ten common learning techniques across five criteria: generalizability across subjects, learner characteristics, materials, and study conditions. Retrieval practice and spaced practice received the highest utility ratings. The techniques most people default to — highlighting, re-reading, massed practice — received the lowest. The gap between evidence and common practice in education is one of the most consistent findings in educational psychology.

Knowing this doesn’t automatically change behavior. But it does give you the right target. The question isn’t whether you’re working hard at learning something. The question is whether the structure of your practice is the kind that actually builds durable, transferable knowledge. Usually, it can be redesigned in ways that take the same amount of time and produce significantly better results. That redesign starts with retrieval, not review.

Last updated: 2026-03-31

References

• Every Learner Everywhere (2023). Six Examples of Evidence-Based Teaching Practices and a Resource Library with Many More. Transform Learning.Link
• Gebhardt, M., et al. (2025). Evidence-based development of inclusive schools. International Journal of Inclusive Education.Link
• Teacher Created Materials. Evidence-Based Research Library. Teacher Created Materials.Link
• Knogler, M., et al. (2025). Pre-service teachers’ knowledge of evidence-based classroom management practices in physical education. Frontiers in Education.Link
• Evidence Based Education. Resources – The Great Teaching Toolkit. Evidence Based Education.Link
• Hoare, E., Thomas, K., & Ofei-Ferri, S. (2025). Evidence-based practices in school settings for student wellbeing. Australian Government Department of Education.Link

Related Reading

• DCA Strategy for Beginners [2026]
• Basic Car Maintenance Everyone Should Know: Beginner Guide [2026]
• How to Find the North Star: Navigation 101