20 years of emotions and risky choices in the lab: A meta-analysis

This paper is a meta-analysis of experimental studies dealing with the impact of incidental emotions on risky choices, so as to explain traditional heterogeneity of outcomes in the literature. After devising a standard search strategy and filtering out studies that do not comply with a list of eligibility criteria, we include 24 articles from which 109 observations are drawn at the treatment level. At this point, we code a set of moderator variables representing experimental protocols and adopt Hedges’s g as comparable metric of effect size. Subgroup analysis and meta-regressions find causal impact of both sadness and fear on risk aversion, albeit to a small extent, as well as highly contrasting patterns depending on the nature of incentives offered in the experiments. The use of monetary incentives turns out to reduce data variability and affects information processing by making subjects more susceptible to emotions. When studies provide real stakes, our results also show that individualism moderates the relationship between emotions and risk attitude by increasing risk propensity. We discuss possible interpretations of our findings.