Journal of Economic Behavior and Organization 21 (1993) 91-98. North-Holland Attitude towards risk An empirical demonstration of context dependence Mir Anjum AltaP University of Karachi, Karachi, Pakistan Received June 1989, final version received February 1992 Following the breakdown of the consensus that marked the theory of choice under uncertainty there is an increasing realization of the importance of context-specific considerations in explaining risk preferences. This paper provides experimental evidence that one such consider- ation, the relative level of success with which participants enter a game, has a significant influence on attitudes towards risk. 1. Intr~u~tion This paper explores the extent to which risk preferences are context dependent. March (1988, p. 5) has remarked that ‘[R]isk aversion is generally assumed in the formal literature on decision theory, where it is typically defined as a concavity in the utility for monetary (or other) return. . . . Most studies indicate, however, that risk preference is not fixed, but depends on the context of a choice’. In his survey of the field, Machina (1987, p. 121) observed that the consensus that surrounded the theory of choice under uncertainty has broken down. While no unified model has emerged to integrate various challenges to the expected utility model of preferences over random prospects, there is general agreement that the theory has to incorporate more context-specific considerations. In this paper I use a natural experiment to study one such factor suggested by March (1988, p. 20), the level of success at which different participants enter a game, which can be taken as a proxy for their historical experience relevant to the gaming situation. The experimental evidence supports the claim that such context-dependent considerations have a marked impact on the risk prefer- ences of the participants. Correspondence to: Dr. Mir Anjum Altaf, Visiting Associate Professor, Department of Environmental Sciences and Engineering, University of North Carolina at Chapel Hill, Chapel Hill, NC 275994060, USA. *I would like to thank an anonymous referee for comments that led to significant improvement in the paper and Dr. Richard Day for encouraging suggestions. I would also like to acknowledge the help of Shanzi Ke in the statistical analysis. 0167-2681/93~~06.~ 0 199%Elsevier Science Publishers B.V. All rights reserved