Empirical Investigation of Belief Updating in Decisions from Experience

In this work, I argue that case-based sequences of decisions is a fruitful empirical field for studying dynamics of ambiguity attitudes, which, in turn, can be informa- tive of the nature of the concept itself. By decomposing ambiguity attitudes into aversion to it (pessimism) and perception of it (likelihood insensitivity), I show how those can be explored if a decision maker experiences statistical learning - sampling from iid variables, distributions of which are unknown to her. To connect ambi- guity attitudes with case-based reasoning and to legitimate quantitative analysis of concepts, I suggest a brief theoretic extension of several other models of choice. This happens to be a bridge from Bayesian definition of ambiguity attitude to a frequentist notion. Further, to include the case of a complete ignorance about both probabilities and outcomes as a starting point, I’m extending the framework to an infinite sample of outcomes, while most papers base their work on finite spaces. Hence, I’m also touching upon the question of belief formation. My theoretical extension can serve further empirical work in decisions under ambiguity. Due to direct statistical integration of observed information into estimation of its effects, findings overall can be applied to study dynamic equilibriums in financial markets and in other highly ambiguous environments like health and technology studies.