Document Type


Publication Date



This paper proposes and tests an empirical model of impartiality, inspired by Adam Smith (1759), that is based on the moral judgments of informed third parties (or spectators). The model predicts that spectatorship produces properties widely considered desirable in both the normative and descriptive literature of philosophy and the social sciences, namely, unbiasedness and consensus. This informs a vignette experiment that elicits moral judgments about real world policy issues while varying the information conditions (relevant and irrelevant information) and roles (spectator and stakeholder) of respondents across treatments. The results indicate that spectator views are unbiased, and that relevant information reduces stakeholder bias to insignificance, whereas irrelevant information reduces bias but does not eliminate it. Relevant information promotes a kind of consensus among both spectators and stakeholders. I argue that this model can inform descriptive and prescriptive political analysis and that it complements empirical work on deliberation and public opinion.

Included in

Economics Commons