Imprecise Data Sets as a Source of Ambiguity: A Model and Experimental Evidence

In many circumstances, evaluations are based on empirical data. However, some observations may be imprecise, meaning that it is not entirely clear what occurred in them. We address the question of how beliefs are formed in these situations. The individual in our model is essentially a “frequentist.” He first makes a subjective judgment about the occurrence of the event for each imprecise observation. This may be any number between zero and one. He then evaluates the event by its “subjective” frequency of occurrence. Our model connects the method of processing imprecise observations with the individual's attitude toward ambiguity. An individual who in imprecise observations puts low (high) weight on the possibility that an event occurred is ambiguity averse (loving). An experiment supports the main assertions of the model: with precise data, subjects behave as if there were no ambiguity, whereas with imprecise data subjects turn out to be ambiguity averse. This paper was accepted by Brad Barber, Teck Ho, and Terrance Odean, special issue editors.

• 저널Management Science · 58(1) · 188–202
• 토픽Bayesian Modeling and Causal Inference · Artificial Intelligence
• DOI10.1287/mnsc.1110.1463
• 저자Ayala Arad, Gabrielle Gayer