Heuristics Can Produce Surprisingly Rational Probability Estimates: Comment on Costello and Watts (2014)
2016 (English)In: Psychological review, ISSN 0033-295X, E-ISSN 1939-1471, Vol. 123, no 1, 103-111 p.Article in journal, Editorial material (Refereed) Published
Costello and Watts (2014) present a model assuming that people’s knowledge of probabilities adheres toprobability theory, but that their probability judgments are perturbed by a random noise in the retrievalfrom memory. Predictions for the relationships between probability judgments for constituent events andtheir disjunctions and conjunctions, as well as for sums of such judgments were derived from probabilitytheory. Costello and Watts (2014) report behavioral data showing that subjective probability judgmentsaccord with these predictions. Based on the finding that subjective probability judgments followprobability theory, Costello and Watts (2014) conclude that the results imply that people’s probabilityjudgments embody the rules of probability theory and thereby refute theories of heuristic processing.Here, we demonstrate the invalidity of this conclusion by showing that all of the tested predictions followstraightforwardly from an account assuming heuristic probability integration (Nilsson, Winman, Juslin,& Hansson, 2009). We end with a discussion of a number of previous findings that harmonize very poorlywith the predictions by the model suggested by Costello and Watts (2014).
Place, publisher, year, edition, pages
2016. Vol. 123, no 1, 103-111 p.
probability judgment, rationality, conjunction error, configural weighted average, random variation
Social Sciences Psychology
Research subject Psychology
IdentifiersURN: urn:nbn:se:uu:diva-270761DOI: 10.1037/a0039249ISI: 000367327000008PubMedID: 26709414OAI: oai:DiVA.org:uu-270761DiVA: diva2:890468
FunderSwedish Research Council, 2012–1212