Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.584737
Title: Investigating the influence of outcome utility on estimates of probability
Author: Harris, Adam J. L.
Awarding Body: Cardiff University
Current Institution: Cardiff University
Date of Award: 2009
Availability of Full Text:
Access through EThOS:
Access through Institution:
Abstract:
This thesis is concerned with the rationality of human probability estimates, specifically the potential influence of outcome utility on estimates of probability. Intuitively, and normatively, the desirability of an outcome should not make that outcome seem more or less likely to occur. Chapter 1 provides a background to the subsequent empirical work by addressing some general issues surrounding a probabilistic approach to human reasoning. The major questions addressed are whether people represent uncertainty quantitatively and their competence at doing so, considered with reference to the status of extant biases in the literature on human probability judgment. Chapter 2 presents seven studies investigating the effect of negative utility on estimates of probability, using a minimal paradigm in which there is a visually defined, objective, probability. When there was an indication that future human action could influence the likelihood of an outcome (the outcome was, in some way, controllable), negative outcomes were rated as more likely to occur than neutral outcomes.. This moderating effect of control can be given a decision-theoretic explanation in terms of loss function asymmetry (e.g., Weber, 1994). Consequently, these effects can be understood as rational reactions to the recognition of the uncertainty of human cognition. Chapter 3 investigates the effects of positive utility using the same visual representation of probability. Across four studies, no effect of positive outcome utility was observed, a result consistent with the asymmetric loss function explanation proposed for the findings in Chapter 2. Chapter 4 presents a statistical-based critique of the 'unrealistic optimism' phenomenon (e.g., Weinstein, 1980). Additionally, two empirical studies failed to find any evidence that the 'unrealistic optimism' phenomenon is more than just a statistical artifact. The results from all three experimental chapters provide support for the contention that people's probability estimates are not systematically biased by utility considerations.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.584737  DOI: Not available
Share: