The “learning retention pyramid” is one of the most widely shared diagrams in education, the one claiming people remember 10% of what they read but 90% of what they teach to others. It’s a popular way to argue for active learning. It’s also, as far as anyone can tell, made up.
That doesn’t mean the underlying idea is wrong. Active learning really does outperform passive learning, and the real research behind that claim is stronger than the pyramid ever was. Here’s what the diagram actually gets right, what it invents, and what to use instead.
What Is the “Learning Retention Pyramid”?
The pyramid ranks learning methods from least to most effective, usually attaching a specific retention percentage to each: lecture (5%), reading (10%), audiovisual (20%), demonstration (30%), group discussion (50%), practice by doing (75%), and teaching others (90%). It’s almost always credited to the National Training Laboratories (NTL) Institute in Bethel, Maine.
The diagram is genuinely useful as a rough intuition: doing tends to beat watching, and watching tends to beat listening. The problem is the specific numbers, and the research supposedly behind them.
Why the Original Percentages Are a Myth
NTL Institute itself has acknowledged that the original study behind these percentages was never published, and the underlying data no longer exists. In a detailed historical analysis, Letrud (2012) traced the pyramid back to two separate, unrelated sources: Edgar Dale’s 1946 “Cone of Experience,” which never included any retention numbers at all, and an old, uncredited retention chart that had circulated in various forms since the early 1900s. Neither source supports the specific percentages now attached to NTL’s name.
A later analysis in medical education, Masters (2020), found the same myth still being taught in medical schools nearly a decade after Letrud’s rebuttal, describing it as a case study in how a convenient, good-sounding statistic can outlive the evidence for it. Neither analysis disputes that active learning works. What they dispute is that anyone ever measured it at 90% versus 10%.
What the Real Research Says About Active vs. Passive Learning
Set the fake percentages aside, and the actual evidence for active learning is substantial, just less quotable.
A large-scale meta-analysis of 225 studies across STEM courses, Freeman et al. (2014), found that students in traditionally lectured courses were 1.5 times more likely to fail than students in active learning courses, and exam scores improved by roughly half a letter grade under active learning conditions.
A separate study, Deslauriers et al. (2019), found something particularly relevant for parents and teachers: students in active learning classrooms actually learned more, but felt like they had learned less than students in passive lecture classrooms. In other words, active learning can feel harder in the moment precisely because it’s working, which is worth knowing before judging a method by how smooth it feels.
Closer to the age range most parents and educators are planning for, a 2023 meta-analysis focused specifically on K-12 settings, Tutal and Yazar (2023), reviewed 398 studies on academic achievement and 88 on learning retention. Both showed large, statistically significant effects in favor of active learning over teacher-led instruction, with retention showing an even larger effect than achievement scores.
Practical Takeaways for Parents and Educators
The research points to a few concrete habits worth building, without needing a pyramid or a specific percentage to justify them:
- Favor doing over watching. A child solving a problem retains more than one watching someone else solve it.
- Let a child explain it back. Asking a child to explain a concept in their own words is one of the more consistent, well-supported active learning techniques.
- Don’t judge a method by how easy it feels. Per Deslauriers et al. (2019), the learning that feels most comfortable isn’t always the learning that sticks.
- Keep it short and repeated. The K-12 research base is built on classroom-scale interventions, not marathon study sessions, so frequent short practice tends to outperform occasional long ones.
How Magrid Applies Evidence-Based Active Learning
Magrid’s visual, language-free method is built around active engagement by design, not passive presentation. Every activity asks a child to observe, decide, and act, tracing a shape, comparing quantities, completing a pattern, rather than watching a demonstration or listening to an explanation. That structure lines up directly with what Freeman et al. (2014) and Tutal and Yazar (2023) found: doing consistently outperforms watching and listening for both learning and retention.
Sessions are also deliberately short, around 10 to 15 minutes, which matches the kind of frequent, focused practice the K-12 research base is built on rather than long, occasional study blocks. Built on more than a decade of cognitive science research at the University of Luxembourg, Magrid’s approach reflects the same active-learning principles the research actually supports, without leaning on a pyramid that doesn’t hold up.
