Evidence for Mistakes #

Every substantive claim on the Mistakes page is checked against current research. Here is each claim, how well today’s evidence supports it, and the sources. The full, de-duplicated source list lives on the references page.

Mixed · moderate evidence — Mistakes are a key part of learning and you can progress faster by making the most of your own mistakes.

Errors can accelerate learning, but conditionally rather than automatically. Metcalfe’s review concludes that committing an error and then receiving corrective feedback typically produces better retention than studying the correct answer alone — but the benefit depends on the corrective feedback being given and attended to. The page’s original ‘mistakes are a key part of learning / you progress faster’ is honest once reframed as ‘a corrected, reflected-on error helps; an uncorrected one does not’.

Sources: Metcalfe (2017), Learning from errors, Annual Review of Psychology — https://doi.org/10.1146/annurev-psych-010416-044022 · full reference ›

Supported · moderate evidence — A mistake on its own teaches very little; what does the teaching is the corrective feedback and reflection that follow it.

Central finding of the errorful-learning literature reviewed by Metcalfe: the memory benefit of generating an error comes from the error–correction sequence, not the error in isolation. Errors made without feedback are not beneficial and can entrench the wrong response. Consistent with the broader feedback literature.

Sources: Metcalfe (2017), Learning from errors, Annual Review of Psychology — https://doi.org/10.1146/annurev-psych-010416-044022 · Kornell, Hays & Bjork (2009), Unsuccessful retrieval attempts enhance subsequent learning, JEP:LMC — https://doi.org/10.1037/a0015729 · full reference ›

Supported · moderate evidence — People often learn less from being told they got something wrong than from being told they got it right, even when the failure feedback carries equally useful information, because the error stings the ego and they tune out.

Eskreis-Winkler & Fishbach’s experiments in Psychological Science directly demonstrate reduced learning from failure relative to informationally-equivalent success feedback, attributed to ego threat and disengagement; lowering the self-relevance of the failure restored learning. Magnitude depends on stakes and ego involvement.

Sources: Eskreis-Winkler & Fishbach (2019), Not learning from failure—the greatest failure of all, Psychological Science — https://doi.org/10.1177/0956797619881133 · full reference ›

Supported · moderate evidence — Deliberately reflecting on what went wrong — a short structured ‘what happened, why, what I’ll do differently’ — improves subsequent performance compared with simply getting more practice time.

Di Stefano, Gino, Pisano & Staats found that adding brief structured reflection to practice produced significantly better performance than equivalent additional practice; reflective-practice and after-action-review evidence broadly supports articulating lessons, with effects dependent on the quality of reflection. Single-team field/lab studies, so moderate rather than strong.

Sources: Di Stefano, Gino, Pisano & Staats (2016), Making experience count: The role of reflection in individual learning, Harvard Business School working paper — https://www.hbs.edu/faculty/Pages/item.aspx?num=43592 · full reference ›

Supported · strong evidence — Framing the lesson at the attempt (’that approach was too fast’) rather than at the self (‘I’m hopeless at this’) matters, because feedback aimed at the person rather than the task tends to be ineffective or even harmful for learning.

Kluger & DeNisi’s meta-analysis of 607 effect sizes found feedback that cues attention to the self is among the least effective, with roughly a third of feedback interventions reducing performance; Hattie & Timperley concur that self-level feedback is the weakest. Reframing an error as task-level information is well supported.

Sources: Kluger & DeNisi (1996), The effects of feedback interventions on performance, Psychological Bulletin — https://doi.org/10.1037/0033-2909.119.2.254 · Hattie & Timperley (2007), The power of feedback, Review of Educational Research — https://doi.org/10.3102/003465430298487 · full reference ›

Mixed · weak evidence — You can learn from other people’s mistakes (a shortcut), but vicarious learning is usually weaker than first-hand learning, so a borrowed lesson is a head start you still need to practise.

Observational/vicarious learning is well established (Bandura’s social learning theory) and organisational research shows people and firms do learn from others’ failures, but transfer from observed to enacted skill is typically partial and weaker than enactive mastery, and vicarious error-learning depends on attention, similarity and the observer drawing the right inference. The page’s framing of others’ mistakes as a ‘shortcut you still need to practise’ is the honest qualification of the original unconditioned ‘shortcut’ claim.

Sources: Bandura (1986), Social Foundations of Thought and Action: A Social Cognitive Theory, Prentice-Hall · KC, Staats & Gino (2013), Learning from my success and from others’ failure: Evidence from minimally invasive cardiac surgery, Management Science — https://doi.org/10.1287/mnsc.1120.1657 · full reference ›