đź“–Deliberate Practice and Performance in Music, Games, Sports, Education, and Professions: a Meta-Analysis

Macnamara, Brooke N. and Hambrick, David Z. and Oswald, Frederick L.
  • Critique of deliberate practice
    • Ignored decades of psychological theorizing (meh)
    • Deliberate practice might be correlated with success because we stop practicing what we’re not good at. (p.2)
    • Deliberate practice and talent are not mutually exclusive
    • The amount of deliberate practice for master-level chess players varies greatly (from ~3,000 to 23,000 hours)
    • (I haven’t done much deliberate practice with programming. perhaps besides contests)
  • Deliberate practice is important but not the sole factor (it only explains about one third)
  • Goals
    • Investigate correlation between deliberate practice and performance
    • Find factors that influence deliberate practice relevance
      • domain: music, games, sports, education, profession
      • predictability of environment (how much environment changes unpredictably under you while you perform the task)
        • (programming seems to be highly predictable environment)
      • methodological moderators: how deliberate practice was assessed in the original paper
  • 88 studies analyzed
  • Data: https://osf.io/rhfsk ./files/macnamara2014-delib-pract-perfor-music-games/data.zip
  • Professions are the least-studied with only 7 studies and 321 people
    • Seems to only include non-solitary deliberate practice
  • Results:
    • In most studies, the correlation between deliberate practice and performance is positive. Only 10 (out of 157) showed negative correlation, and 8 of them were statistically insignificant (p < .05)
    • Variability explained by deliberate practice:
      • per domain
        • games 26%
        • music 21%
        • sports 18% (19% if team performance is excluded)
        • education 4%
          • primarily included students, where performance = grade
        • professions 1% (and not statistically significant)
      • predictability of task environment
        • high 24%
        • intermediate 12%
        • low 4%
      • by methodology
        • retrospective interview 20%
        • retrospective questionnaire 12%
        • log method 5%
          • (in Ericsson…1993 (original study) there was little correlation between good and best groups in terms of current practice per day—but best groups had more accumulative practice)
      • how assessed
  • theory
    • low correlation in professions might be explained by lack of well-defined deliberate practices for professions
    • Starting age negatively predicts performance in chess (even after statistically controlling for deliberate practice)
    • General intelligence might explain some performance results
    • Another predictor might be working memory capacity
  • Q:
    • the study seems to cram in different methodologies. According to Ericsson…1993, only accumulated practice matters, so log method is off the table (?)

See also: