Reading focus
Syllabus: How Do We Judge and Decide?
Reconstruct the three measures and follow their different predictions across task difficulty and feedback. Notice which information is available about self and others.
Why you're reading it. Make trust assessable through specified constructs and observations.
A question to consider. A forecaster says they are better than their peers and gives very narrow prediction ranges. What separate data would you collect to assess those two claims?
Why this reading
Someone can overestimate their score while underestimating their position relative to other people. Moore and Healy separate measures that the same everyday word often conceals, then propose and test an explanation for their different patterns. This makes construct definition part of explaining the evidence.
Before you read
Overestimation means predicting too high an absolute performance. Overplacement means placing oneself too high relative to others. Overprecision means expressing more certainty than the evidence warrants. Calibration concerns the agreement between expressed confidence and observed outcomes across appropriate cases; confidence alone does not measure accuracy.
Read for
- Reconstruct all three measures from the quiz design and its elicited probability distributions. Identify where familiar confidence measures confound them.
- Follow predictions for easy and difficult tasks and for beliefs before and after feedback. Connect these predictions to unequal information about self and others.
- Separate the theory's explanation of task-dependent patterns from questions it does not explain, including motivation and individual differences.
Consider afterwards
- A student predicts 7 correct answers and expects others to score 8; both actually score 6. Which judgments are too high or too low, and what additional information is needed to evaluate overprecision?
Editorial context
The study combines a theoretical reconciliation with an incentivized trivia experiment involving 82 university participants. It supports distinguishing the three constructs rather than treating overconfidence as a single stable personal defect. The account explicitly leaves motivational and individual-difference effects unresolved. Generalizing to professional forecasts requires suitable measures of outcomes, uncertainty, and information—not simply asking whether an expert sounds confident.