Concept
Overconfidence (Above-Average Effect)
self-perception cognitive-bias confidence
Most people rate themselves as above average on nearly every desirable trait — intelligence, honesty, driving, kindness, cooking — which is a statistical impossibility once aggregated across a population, yet holds up consistently across studies, cultures, and even among people with strong incentive to be accurate. Will Storr frames this as a feature, not a bug: a psychologically healthy brain distorts self-perception to make a person feel somewhat heroic, since a genuinely accurate, unflattering self-view is closer to how depression actually presents.
A poll of 100 Nudge listeners found average self-ratings of 7/10 for intelligence, 8/10 for honesty, and 6/10 for cooking ability — all comfortably above the "5 = average" midpoint the question defined (see Nudge Listener Self-Rating Overconfidence Survey). The pattern holds even among people with every incentive to be self-critical: prison inmates rate themselves above average for compassion and honesty; both partners in a couple independently claim to do more household chores than the other (a logical impossibility); and long-term couples confidently but wrongly predict how their partner would answer questions about their own sexual history, with confidence and long relationship tenure increasing rather than reducing the error rate (see Couples Overconfidence Studies).
Experts aren't exempt. Psychologists Howard Raffia and Mark Alpert asked participants to give a numeric range for uncertain quantities (e.g. total US egg production, Panama Canal toll revenue) wide enough that they'd be wrong no more than 2% of the time. Participants were wrong roughly 40% of the time — even with total freedom to pick an arbitrarily wide range, they chose ranges far too narrow (see Raffia Alpert Range-Estimation Overconfidence Study). Economists forecasting oil prices five years out are shown to be no more accurate than zookeepers doing the same — but deliver their forecasts with far more unwarranted confidence. A commonly-cited (though hard-to-verify) figure holds that 84% of French men rate themselves as above-average lovers.
Consequence: in roles where seniority isn't tightly tied to measurable performance, this bias helps explain why unqualified or mediocre people reach senior positions — not through dishonesty, but because they genuinely, sincerely believe in their own above-average competence, same as everyone else does.
Group consensus as a corrective, and a Kahneman quote (Episode 105). David Halpern, in Inside the Nudge Unit, cites research finding a group's combined judgment about whether a project will finish on schedule is far more accurate than any individual's, since aggregating independent views filters out any one person's overconfidence — and reports Daniel Kahneman saying that if he could eliminate one bias with a magic wand, he'd pick overconfidence. Overconfidence itself is found to grow more pronounced with seniority and success (see Mindgym CEO Confirmation-Bias Performance Study).
Further survey evidence and a real-stakes illustration (Episode 278). Van Hoorens' 1993 review found students rate themselves above-average on honesty, persistence, and originality, and below-average on hostility and vanity. Donald Trump's repeated public claims that "nobody knows more than me" about a wide range of subjects are cited as a real-world illustration. A related puzzle is used to show overconfidence in unfamiliar domains: told lily pads on a lake double in size every day and take 48 days to fully cover it, most people answer "day 24" for when the lake is half-covered — the correct answer is day 47, since the pads double on the final day.
Discussed in
- Episode 105 — Why -doing everything- at work won-t work (group-consensus corrective; Kahneman quote)
- Episode 125 — 125-why-stupid-people-think-they-re-smart
- Episode 278 — 278-real-world-examples-of-cognitive-biases (Van Hoorens review; Trump illustration; lily-pad puzzle)
Related
- Nudge Listener Self-Rating Overconfidence Survey
- Couples Overconfidence Studies
- Raffia Alpert Range-Estimation Overconfidence Study
- Groupish Belief Formation (Confirmation Bias and Status)
- Will Storr
- The Science of Storytelling
- Dunning-Kruger Effect
- Illusion of Validity
- Mindgym CEO Confirmation-Bias Performance Study
- Inside the Nudge Unit