First-order stochastic dominance violations, risk preferences and framing effects

Abstract

This paper experimentally investigates whether decision-makers violations of the first-order stochastic dominance (FSD) criterion of rational choice, as documented in previous studies, are correlated with their risk preferences. As argued in the literature, FSD violations require framing effects to affect an individual’s decisions: the frame of the lotteries in an individual’s choice set must prevent the individual from the inferiority of choosing stochastically dominated lotteries. This paper moves a step ahead by investigating whether (consistent or erratic) risk-averse or risk-seeking attitudes towards uncertainty make individuals more or less receptive to the framing effects which can induce FSD violations. Our main finding is that consistent (and, to a minor extent, “erratic”) risk-seeking individuals are statistically significantly more prone to violate the FSD criterion in their choices than individuals exhibiting different risk attitudes (i.e., consistent risk-averse and systematically inconsistent risk-taking individuals). We finally control for the role of the framing effects in our results. Consistently with the previous literature, we find that the framing effects play a crucial role. As we re-frame lotteries in such a way as to make the FSD superior prospects more easily distinguishable, any participants hardly violate the FSD criterion, irrespective of their risk preferences, gender, and exerted cognitive abilities.

Key words: first-order stochastic dominance, framing effects, risk preferences

Please email me for a copy of the paper.

Features

• Experimental Design: The study employs a controlled experimental design to elicit risk preferences and FSD violations under two different framing conditions.
• Software: o-Tree for experimental design and data collection, and R/Python for data analysis.
• Participants: 140 UK-resided participants from diverse backgrounds recruited through Prolific.