This paper examines the empirical validity of stochastic dominance rules and the mean-variance framework by analysing data generated through an experiment on individual investment decisions under uncertainty. The analyses indicated that none of the two approaches provided adequate explanation for the observed pattern of choice. An alternate framework, based on preference for skewness, in addition to mean and variance, was examined. This framework provided a significantly better explanation compared to the two parameter framework. The preference for skewness was significant at higher levels of borrowing and at all levels of wealth.