The Ellsberg 'Problem' and Implicit Assumptions under Ambiguity

Abstract

Empirical research has revealed that people try to avoid ambiguity in the Ellsberg problem and make choices inconsistent with the predictions of Expected Utility Theory. We hypothesized that people might be forming implicit assumptions to deal with the ambiguity resulting from the incomplete information in the problem, and that some assumptions might lead them to deviate from normative predictions. We embedded the Ellsberg problem in various scenarios that made one source of ambiguity (i.e., the implied distribution of the unknown number of the colored balls) explicit. Results of an experiment showed that more people chose consistently (and hence rationally) when the scenario encouraged them to think that the probability distribution of the number of balls was normal. The results give insight into the implicit assumptions that might lead to choices congruent with normative models.

Overcoming Ambiguity Aversion Through Experience

“…Note that the average number of samples taken in the EP group was significantly higher than that of the ND group—a pattern that indicates that the participants in the EP group needed to sample more from the ambiguous box in order to gain insight about the underlying distributions than those in the ND group. This finding supports the idea that the implication of the EP distribution was relatively more difficult for participants to intuit (Güney & Newell, ). A similar relationship between the “informativeness” and the level of ambiguity aversion is observed for the (known) FOPs (Becker & Brownson, ; Budescu et al, ; Curley & Yates, ; Kramer & Budescu, )—more precisely defined FOPs lead to less ambiguity aversion.…”

“…This is mainly because ND implies that the FOP is more likely to be around 0.5, whereas EP, for example, indicates that FOP could be anything between 0 and 1 with an equal probability. Therefore, participants might be more confident in predicting/estimating the potential likelihood of winning in the former and could demonstrate less aversion (Güney & Newell, ).…”

The Description–Experience Gap in Risky and Ambiguous Gambles

“…However, Arló‐Costa and Helzner (, ) have recently shown that people seem to behave as ambiguity‐averse even in non‐comparative cases when urns A and B are not presented simultaneously. One reason for ambiguity aversion in the non‐comparative cases could be that people form implicit assumptions to deal with the ambiguity resulting from the unknown information about urn B, and these assumptions might lead them to behave as ambiguity‐averse (Guney & Newell, ).…”

Ambiguity attitudes in qualitative contexts: The role of prior beliefs

“…Although this explanation may not apply so much to the classic Ellsberg situation -in which people can choose which colour to bet on -it could be applicable more broadly, particularly in real-world situations. After all, ambiguous and risky choices are only equal if people follow the principle of indifference (Marquis de Laplace, 1902) and presume that all possible outcomes are equally likely (or that the probability distribution underlying the ambiguity is otherwise symmetrical around the 'risky' probability; Guney & Newell, 2011;Güney & Newell, 2015). Under this assumption, a risky decision in which the two outcomes are assigned 50/50 probability is equivalent to an ambiguous two-choice scenario in which all options are equally likely.…”