The hot stove effect in decision making under ambiguity
Abstract (Via Fujikawa, Takemi @ Uni Muenchen.de)
The ”hot stove effect” has been studied for repeated-play decision making under uncertainty (also referred to as experience-based decision making) in which the decision makers repeatedly face the Allais-type binary choice problems, and have to learn about the outcome distributions through sampling as the decision makers are not explicitly provided with prior information on the payoff structure. The previous studies have found mixed evidence: some studies have found that the hot stove effect is strong in repeated-play decision making under uncertainty, while other studies have found that the effect is weak. Thus, the evidence is inconsistent. This paper reports an experimental investigation of the hot stove effect in repeated-play decision making under ambiguity. The current experiment involves an ambiguity treatment in which (1) the participants perform two binary repeated-play choice problems, each involving 400-fold choice between a risky option and a riskless option; and (2) in each problem, there are two states of nature available: a favourable state and an unfavourable state, but only one of them obtains on any given trial. The realisation of the actual state is not disclosed to the participants, thus they would be expected to discover the actual state through sampling with immediate feedback. The current results suggest that the magnitude of the hot stove effect is significantly different between repeated-play decision making under uncertainty and repeated-play decision making under ambiguity. I shall show that the hot stove effect is attenuated in repeated-play decision making under ambiguity.
Introduction (Via Fuikawa, Takemi @ Uni Muenchen.de)
Ambiguity or “uncertainty about uncertainties” is a pervasive element of much real world decision making (Einhorn & Hogarth, 1986). An example of our real world decision making includes investment behaviour. Traders and investors often face the complexity of the situation and ambiguity in asset markets. Traders and investors do not often have perfect, prior information about asset markets, but ambiguous information about them. An example is concerned with investors who face a choice between a safe asset (e.g., cash) and a risky asset with an unknown return distribution. The distribution is unknown, but the investors know that the distribution obtains either a favourable distribution or an unfavourable distribution. If they consider the favourable distribution to be possible, then they may choose to invest in the risky asset. If, on the other hand, they consider the unfavourable distribution to be possible, then they consider not to invest in the risky asset (Easley & O’Hara, 2005).