Reading focus
Syllabus: How Do We Judge and Decide?
For each proposed heuristic, identify the question being answered, the cue being used, and the evidence that the cue can mislead.
Why you're reading it. Move from noticing errors to evaluating explanations that predict particular patterns.
A question to consider. Would finding that a vivid example changes a frequency estimate establish availability as the mechanism? Specify a comparison with an informative but less vivid example.
Why this reading
A difficult probability question may be answered through a more accessible cue, such as resemblance or ease of recall. This paper organizes experimental examples around proposed judgment strategies and their characteristic errors. It gives readers a way to ask how a judgment was produced, beyond whether it was correct.
Before you read
A heuristic is a simplifying strategy. A bias is a systematic departure from a specified standard, not any answer a researcher dislikes. A base rate is the frequency of an event in the relevant population. Representativeness concerns resemblance; availability concerns ease of bringing examples to mind; anchoring concerns the influence of a starting value.
Read for
- For each example, state the target judgment, the proposed substitute cue, and the standard used to assess the answer.
- Distinguish an observed answer pattern from evidence that identifies the mental process responsible for it.
- Look for circumstances in which the cue is informative and circumstances in which it can mislead. Consider what task interpretation the explanation assumes.
Consider afterwards
- After hearing a vivid account of a delayed train, someone raises their estimate of how often trains are late. What data and comparison would help distinguish a sensible update from reliance on ease of recall?
Editorial context
The article is an influential theoretical synthesis with demonstrations, not a complete taxonomy of errors or proof that people are generally irrational. A named heuristic is an explanatory proposal that needs more precise predictions. Later work on information formats and environmental structure qualifies broad claims about error. The scientific task is to identify the procedure, its inputs, and its conditions of success—not to diagnose a person from a single response.