Combining imprecise or conflicting probability judgments: A choice-based study

Sources of uncertainty appear to affect attitude towards ambiguity. For instance, when two advisors agree on a range of probabilities and, when they disagree - one advisor predicting the upper bound of the range while the other predicts the lower bound of the range †decision-makers might have different beliefs about the risk although in both case the mean probability is the same. This study draws upon prospect theory and ambiguity research to empirically test how sources of uncertainty affect decision-makers†beliefs. It contrasts revealed beliefs †the precise probability leading to the same choice as an uncertain probability forecast †with judged beliefs †decision-makers†best estimate of the probability of the risk. It also equips beliefs with two properties - pessimism and likelihood sensitivity†to allow them to vary as a function of the source of uncertainty. Two experiments compare beliefs across sources of uncertainty and across elicitation methods (judged vs. revealed beliefs). Findings support the predictions that the source of ambiguity matters in particular for low and high probability events and that revealed and judged beliefs differ.