Risky weighting in discrete choice

•We investigate a broad class of cumulative risky weighting functions.•We propose a unified cumulative risky weighting function.•We develop a multivariate method for choice analysis with risky prospects.•The new method accounts for decision-makers' individual-specific risk perception.

This paper presents a new approach to discrete choice analysis for risky prospects. Conventional discrete choice analysis focuses on riskless prospects and does not deal with the scenario where the alternatives that the decision-makers choose from are associated with risk. In this paper, we investigate decision-makers' risk perception and choice behaviour in choice experiments when they are facing several risky prospects. We propose a broad class of cumulative risky weighting functions, upon which a unified cumulative risky weighting function is developed. We show that this unified cumulative risky weighting function includes several existing cumulative risky weighting functions as special cases. We then develop a multivariate method for choice analysis with risky prospects to account for decision-makers' individual-specific risk perception and the impact of various factors on the value function respectively. We illustrate the developed method using an empirical study on road tolling in Australia.