Why it works

The evidence behind Mosaic in under two minutes, with on-screen text.

0:00 / 1:50
Read the text of this video
  1. Which ways of studying actually work? The evidence, briefly.
  2. A major review rated ten study techniques (Dunlosky and colleagues, 2013). Only two were rated highly useful: practice testing and distributed practice. Elaborative interrogation, self-explanation and interleaved practice were rated moderately useful. Summarising, highlighting, the keyword mnemonic, imagery for text and rereading were rated low.
  3. Straight afterwards, rereading looks better: 5 minutes later, students who reread a passage recalled 81% of it, and students who tested themselves recalled 75%. A week later, testing wins: 56% for the tested group against 42% for rereading. In the same study, rereading made students feel more confident, not less (Roediger and Karpicke, 2006).
  4. It holds for doctors in training too. Forty paediatric and emergency medicine residents either took short tests or studied a review sheet after teaching. Six months later, the tested topics scored 39% and the studied topics 26% (Larsen and colleagues, 2009). In Mosaic: flashcards and Quick Test.
  5. Even a wrong guess helps, as long as you see the right answer straight after: it is remembered better later (Kornell and colleagues, 2009). In Mosaic, you answer first, then check, on every card and test question.
  6. Spread it out. The same study time, spread over weeks, is remembered for longer. The longer you need to remember something, the wider the gaps should be (Cepeda and colleagues, 2006, a review of 317 experiments, and 2008). In Mosaic, flashcards space your reviews for you.
  7. Explain why, not just what. Explaining an answer to yourself deepens understanding (Chi and colleagues, 1989; rated moderately useful by Dunlosky and colleagues, 2013). In Mosaic, AKT question writing asks you to explain every option.
  8. "The most important single factor influencing learning is what the learner already knows" (David Ausubel, 1968). New learning holds better when it's linked to what you already know. In Mosaic, the connection notebook links learning to your own patients.
  9. Plan one change, and say when. Planning when and where you'll act makes you more likely to follow through (Gollwitzer and Sheeran, 2006, a review of 94 studies). For example: if I see an older patient with type 2 diabetes on a sulfonylurea, then I'll ask about hypos and review their target. Family doctors who wrote down a planned change were more likely to make it (Wakefield and colleagues, 2003: 207 family doctors, judged by their prescribing). In Mosaic, the connection notebook asks for one thing you'll do this week.
  10. An honest caveat. Most of this evidence comes from laboratory and classroom studies. Studies with doctors are encouraging, but there are fewer of them. And there's less evidence that written reflection helps on its own. No app makes learning effortless. The effort is what makes it work.
  11. Mosaic: free learning tools for GP registrars, built on this evidence. mosaicgp.app. Not affiliated with or endorsed by the RCGP or NHS England.