Every learning principle in our library on one page: how strong the research is, the question it answers, and
what it means for your course. It grows as each new entry is published.
9 principles · last reviewed 10 November 2026
How we rate the evidence
Strong
Consistent results across many studies and reviews, including with real learners in real courses.
Moderate
Consistent experimental support, but mostly from labs, one research programme or one kind of learner.
Emerging
Early, small or mixed studies. Worth trying in a course, but don't expect a guaranteed effect.
What keeps learners motivated in a self-paced online course, and what does the research say helps?
In practice: Help learners feel capable first with early wins and clear feedback, then add real choices and contact with others; don't lean on rewards.
Motivation and engagementReviewed 10 November 2026
Do learners really remember 10% of what they read and 90% of what they teach?
In practice: Don't quote the pyramid's percentages or rank formats by them; build retrieval practice, active tasks and explanation into each module because the evidence for those is good.
Should a course match its lessons to each learner's learning style?
In practice: Don't sort learners by style or build parallel versions of lessons; choose each format for the content, and offer captions and transcripts for access.
Claims you'll still see in training decks and course-creator advice, and what the research actually says.
“People learn best when teaching matches their learning style: visual, auditory or kinaesthetic.”
Not supported
A major review found virtually no evidence that matching teaching to a preferred style improves learning, and several well-designed studies that contradict it. Its authors concluded there is no adequate evidence base for using learning-styles assessments in general teaching. Giving every learner words and a relevant picture together (dual coding) is a different idea, and better supported.
“We remember 10% of what we read, 20% of what we hear, and 90% of what we teach others.”
Not supported
The percentages have no known empirical source. Tracing the model's history, Letrud and Hernes write that nobody knows how the learning pyramids originated or whether any evidence supported them. The model is often credited to Edgar Dale's Cone of Experience, which ranked teaching media by how abstract they are, not by how much people remember.
“Course videos should never be longer than six minutes.”
Overstated
The figure comes from a study of 6.9 million viewing sessions on edX, which found median viewing time topped out at about six minutes whatever the video's length. It measured how long people kept watching, not what they learned, and its authors warned against drawing conclusions about learning from it. Short segments are a sensible default, not a proven rule.
Ratings are our judgement of the research each entry cites, and every entry links to its sources. We re-check each
entry at least once a year. Spotted something the evidence has moved on? Tell us.