John C. Hershey

Outcome bias in decision evaluation.

In 5 studies, undergraduate subjects were given descriptions and outcomes of decisions made by others under conditions of uncertainty. Decisions concerned either medical matters or monetary gambles. Subjects rated the quality of thinking of the decisions, the competence of the decision maker, or their willingness to let the decision maker decide on their behalf. Subjects understood that they had all relevant information available to the decision maker. Subjects rated the thinking as better, rated the decision maker as more competent, or indicated greater willingness to yield the decision when the outcome was favorable than when it was unfavorable. In monetary gambles, subjects rated the thinking as better when the outcome of the option not chosen turned out poorly than when it turned out well. Although subjects who were asked felt that they should not consider outcomes in making these evaluations, they did so. This effect of outcome knowledge on evaluation may be explained partly in terms of its effect on the salience of arguments for each side of the choice. Implications for the theory of rationality and for practical situations are discussed.

Sources of Bias in Assessment Procedures for Utility Functions

Utility functions are an important component of normative decision analysis, in that they characterize the nature of peoples risk-taking attitudes. In this paper we examine various factors that make it difficult to speak of the utility function for a given person. Similarly we show that it is questionable to pool risk-propensity data across studies (for descriptive purposes) that differ in the elicitation methods employed. The following five sources of bias or indeterminacy are hypothesized and demonstrated. First, certainty equivalence methods generally yield greater risk-seeking than probability equivalence methods. Second, the probability and outcome levels used in reference lotteries induce systematic bias. Third, combining gain and loss domains yields different utility measures than separate examinations of the two domains. Fourth, whether a risk is assumed or transferred away exerts a significant influence on peoples preferences in ways counter to expected utility theory. Finally, context or framing differences strongly affect choice in a nonnormative manner. The above five factors are first discussed as essential choices to be made by the decision scientist in constructing Von Neumann-Morgenstern utility functions. Next, each is examined separately in view of existing literature, and demonstrated via experiments. The emerging picture is that basic preferences under uncertainty exhibit serious incompatibilities with traditional expected utility theory. An important implication of this paper is to commence development of a systematic theory of utility encoding which incorporates the many information processing effects that influence peoples expressed risk preferences.

Risk Taking and Problem Context in the Domain of Losses: An Expected Utility Analysis

This paper presents an experimental investigation of risk taking in the domain of losses. The results are partly compatible with expected utility theory, assuming an inflection point in the utility function over losses. However, overweighting of low probabilities and underweighting of high ones was observed, which runs counter to the expected utility model. Additionally, a strong context effect was observed in which choices presented in an insurance context were judged with greater risk aversion than mathematically identical choices presented as standard gambles. Both the normative and descriptive implications for expected utility theory are discussed. This article examines two related issues regarding decision-making under conditions of objective or known risk. The first one concerns the nature of peoples risk preferences for losses, and its compatibility with expected utility theory; the second concerns the extent to which such risk preferences are influenced by problem context.

Cognitive Processes and the Decisions of Some Parents to Forego Pertussis Vaccination for Their Children

Public health analyses suggest that, in spite of the possibility that pertussis vaccine may cause rare cases of neurological injury, catastrophic risks to individual children are lower if they are vaccinated. A number of parents, however, choose not to vaccinate their children. The purpose of this study was to investigate the decision processes of some parents who choose to vaccinate and some parents who choose not to do so. Surveys were mailed to 500 randomly selected subscribers of Mothering magazine. Two hundred and ninety-four completed questionnaires were returned (59%). In addition to well-recognized factors in vaccination decisions, perceived dangers of the vaccine, and of the disease and susceptibility to the disease, several cognitive processes not previously considered in vaccination decision studies were found to be important predictors in this population of parents: perceived ability to control childrens susceptibility to the disease and the outcome of the disease; ambiguity or doubts about the reliability of vaccine information; a preference for errors of omission over errors of commission; and recognition that if many other children are vaccinated, the risk to unvaccinated children may be lowered. Although perhaps most cases of undervaccination for pertussis reflect more general problems of health care access, some parents choose to forego vaccination for their children for other reasons. Traditional risk-benefit arguments alone will be unlikely to persuade these parents to reassess their decisions. Efforts to increase childhood vaccination must incorporate an understanding of the cognitive processes that help drive these decisions.

Prospect theory's reflection hypothesis: A critical examination

Abstract In a recent new theory of decision making under risk (called prospect theory), Kahneman and Tversky suggest that peoples preferences among lotteries on the loss side are a mirror image of those on the gain side. This article examines the reflection hypothesis at both an across-subject and within-subject level. Since significant reflection at either level does not necessarily imply significant reflection at the other, upper and lower bounds are computed for each. Four types of within-subject reflectivity are distinguished and analyzed. Relationships among these types are examined theoretically as well as empirically. Experimental results from three separate studies are presented, which seriously question the generality of prospect theorys reflection hypothesis.

The Role of Physicians’ Recommendations in Medical Treatment Decisions

Background . A shift away from the medical paternalism of the past has occurred, and today, the law and ethics advocate that physicians share decision-making responsibility with their patients. It is unclear, however, what the appropriate role of physicians’ recommendations ought to be in this new shared decision-making paradigm. One way to approach this question is to assess the influence of physicians’ recommendations. Objective . In this study, the authors examine the influence of physicians’ recommendations on hypothetical treatment decisions. Do physicians’ recommendations influence treatment decisions in scenarios where the decision that maximizes health is obvious and apparent to subjects? Do recommendations pull subjects away from the treatment choice that they otherwise prefer (based on their decision when unaware of the physicians’ recommendation)? Design . An experimental web questionnaire presented hypothetical medical treatment scenarios in which the treatment choice that maximized health was obvious. Across scenarios, the authors varied physicians’ recommendations in 3 ways:(1) physicians’ recommendations supporting what maximized health, (2) physicians’ recommendations that went against what maximized health, and (3) no physicians’ recommendation. The participants were 102 volunteers. Results . Hypothetical treatment decisions were significantly influenced by physicians’ recommendations (P < 0.0001), and physicians’ recommendations against the decision that maximized health pulled subjects away from the treatment decision that they made when no recommendation was given (P <0.0001). Conclusion . Physicians’ recommendations can lead people to make decisions that go against what is best and against what they would otherwise prefer. Physicians must take care in making recommendations and should incorporate patient preferences into their recommendations.

Determinants of Priority for Risk Reduction: The Role of Worry

One hundred twenty-two members (experts) of the Society for Risk Analysis completed a mailed questionnaire and 150 nonexperts completed a similar questionnaire on the World Wide Web. Questions asked included those about priorities on personal and government action for risk reduction, badness of the risk, number of people affected, worry, and probabilities for self and others. Individual differences in mean desire for action were largely explained in terms of worry. Worry, in turn, was largely affected by probability judgments, which were lower for experts than for nonexperts. Differences across risks in the desire for action, within each subject, were also determined largely by worry and probability. Belief in expert knowledge about the risk increased worry and the priority for risk reduction. A second study involving 91 nonexperts (42 interviewed and 49 on the Web) replicated the main findings for nonexperts from the first study. Interviews also probed the determinants of worry, attitudes toward government versus personal control, and protective behaviors.

Do Nonpatients Underestimate the Quality of Life Associated with Chronic Health Conditions because of a Focusing Illusion

Background. A number of studies show that the general public often estimates that the quality of life (QOL) associated with various health conditions is worse than patients say it is. These studies raise the possibility that people overestimate the impact that unfamiliar health conditions will have on their quality of life. One possible reason people overestimate this is because they are susceptible to a focusing illusion—when asked to imagine themselves in unfamiliar circumstances, people overestimate the emotional impact of those features of their life that would change. Methods. The authors surveyed members of the general public to test the hypothesis that their QOL ratings of hypothetical health conditions would be higher (indicating a better quality of life) after thinking about how the health condition would affect a broad range of life domains. Across 3 experiments, the authors varied the health conditions people were asked to consider (either paraplegia, below-the-knee amputation, or partial blindness), the life domains they were asked to consider, the response mode with which they evaluated how each health condition would affect each life domain, whether subjects rated the health condition before and after considering life domains or only after, and whether subjects rated their own current quality of life first. Results. Across 3 experiments, using 10 different questionnaire versions, only 1 instance was found in which subjects’ ratings were significantly higher after thinking about the effect of the health condition on life domains than before, and the magnitude of this increase was small. Conclusion. It could not be established that a focusing illusion contributes significantly to the discrepancy in QOL ratings of patients and nonpatients. Further research should explore other factors that could contribute to the discrepancy or other ways of testing for the influence of a focusing illusion.

Life insurance and breast cancer risk assessment: adverse selection, genetic testing decisions, and discrimination.

Life insurance industry access to genetic information is controversial. Consumer groups argue that access will increase discrimination in life insurance premiums and discourage individuals from undergoing genetic testing that may provide health benefits. Conversely, life insurers argue that without access to risk information available to individuals, they face substantial financial risk from adverse selection. Given this controversy, we conducted a retrospective cohort study to evaluate the impact of breast cancer risk information on life insurance purchasing, the impact of concerns about life insurance discrimination on use of BRCA1/2 testing, and the incidence of life insurance discrimination following participation in breast cancer risk assessment and BRCA1/2 testing. Study participants were 636 women who participated in genetic counseling and/or genetic testing at a University based clinic offering breast cancer risk assessment, genetic counseling, and BRCA1/2 testing between January 1995 and May 2000. Twenty‐seven women (4%) had increased and six (1%) had decreased their life insurance since participation in breast cancer risk assessment. The decision to increase life insurance coverage was associated with predicted breast cancer risk (adjusted OR 1.03 for each 1% absolute increase in risk, 95% CI 1.01–1.10) and being found to carry a mutation in BRCA1/2 (OR 5.10, 95% CI 1.90–13.66). Concern about life insurance discrimination was inversely associated with the decision to undergo BRCA1/2 testing (RR 0.67, 95% CI 0.52–0.85). No respondent reported having life insurance denied or canceled. In this cohort of women, these results indicate that information about increased breast cancer risk is associated with increase in life insurance purchasing, raising the possibility of adverse selection. Although fear of insurance discrimination is associated with the decision not to undergo BRCA1/2 testing, there was no evidence of actual insurance discrimination from BRCA1/2 testing.

Heuristics and biases in diagnostic reasoning: II. Congruence, information, and certainty☆

Six experiments were carried out to examine possible heuristics and biases in the evaluation of yes-or-no questions for the purpose of hypothesis testing. In some experiments, the prior probability of the hypotheses and the conditional probabilities of the answers given each hypothesis were elicited from the subjects; in other experiments, they were provided. We found the following biases (systematic departures from a normative model), and interviews and justifications suggested that each was the result of a corresponding heuristic: Congruence bias. Subjects overvalued questions that have a high probability of a positive result given the most likely hypothesis. This bias was apparently reduced when alternative hypotheses or probabilities of negative results are explicitly stated. Information bias. Subjects evaluated questions as worth asking even when there is no answer that can change the hypothesis that will be accepted as a basis for action. Certainty bias. Subjects overvalued questions that have the potential to establish, or rule out, one or more hypotheses with 100% probability. These heuristics are explained in terms of the idea that people fail to consider certain arguments against the use of questions that seem initially worth asking, specifically, that a question may not distinguish likely hypotheses or that no answer can change the hypothesis accepted as a basis for action.