Identifying moral antecedents of decision-making in discrete choice models

Abstract

Discrete Choice Models are valuable tools for quantitative decision-making analysis: they allow analysts to draw behavioural conclusions from data, better understand and predict choices, and evaluate policies. However, up until recently, they had a blind spot for morality. Moral values often play an essential role in decision-making; fairness or loyalty can deter people from following self-interest. Moral motivations can also prompt decision-makers to change their minds when contemplating a dilemma or hide their preferences when they want to avoid judgement. These notions are not aligned with crucial behavioural assumptions traditional Discrete Choice Models are based on, such as stable preferences echoing through choices or decision-makers maximizing their utility. This thesis aims to develop and test new Discrete Choice Models that help identify morality in a mathematically rigorous framework, thus increasing the behavioural realism of Discrete Choice Models in moral decision-making. To do this, it uses two approaches.