How Learning Actually Works According to Science

Who created and checked this page

Correction history

Every accepted correction to this page is recorded with the exact change, so readers can see how the page improved over time.

1. 29 July 2026 · corrected by Emir Baycan

   Corrected two citations that misrepresented real research: a Doug Rohrer interleaving study was cited with the wrong journal and an invented reversal narrative not found in his real published work, and a real Roschelle, Feng, Murphy, and Mason 2016 study was misdescribed as being about Khan Academy mastery-based learning when it actually studied a different platform, ASSISTments, with a much smaller and differently measured effect.

   Before

   Article attributed a fabricated one-day-versus-one-month score reversal (89pct/72pct flipping to 38pct/61pct) to Rohrer in Psychological Science, and attributed a fabricated 52 percent improvement figure and Khan Academy mastery-learning framing to the real Roschelle et al. 2016 study.

   After

   Rohrer's interleaving finding corrected to the real, replicated result (72 percent versus 38 percent on a two-week delayed test, published in Psychonomic Bulletin and Review and Journal of Educational Psychology); the Roschelle et al. citation corrected to accurately describe the real ASSISTments randomized trial and its real effect size of 0.18 standard deviations.

   Why: Verified both named studies against their real published versions; found a real researcher's work altered with an invented narrative in one case, and a real study's subject, platform, and effect size all misrepresented in the other.

   View the full record →

Full contribution timeline

1. Correction 29 July 2026

   Emir Baycan: Corrected two citations that misrepresented real research: a Doug Rohrer interleaving study was cited with the wrong journal and an invented reversal narrative not found in his real published work, and a real Roschelle, Feng, Murphy, and Mason 2016 study was misdescribed as being about Khan Academy mastery-based learning when it actually studied a different platform, ASSISTments, with a much smaller and differently measured effect.

   Proof record →