Mapping risk preferences: psychometric and empirical properties across measurement approaches and domains

Risk preference, a fundamental building block of economic theory, plays a significant role in understanding individual differences in risk-taking across the behavioural sciences. Despite employing various measurement approaches and domains for its assessment, inconsistencies persist regarding these operationalizations’ reliability and the inherent characteristics of risk preference.
This dissertation consists of three manuscripts that extensively investigate the psychometric properties (i.e., convergent validity and temporal stability) of risk preference and is supplemented by a fourth manuscript that delves into its empirical properties (i.e., selection and socialization effects). The aim is to enhance our comprehension of the individual differences in risk preference and to quantify its development over time.
In our analysis of psychometric properties, we employed meta-analytic methods to thoroughly analyse and synthesise data, with the goal of estimating and comparing the convergent validity and temporal stability (i.e., rank-order stability and mean-level change) of risk preference across various measures. Our meta-analyses provided several significant insights. Firstly, we discovered a notable lack of convergence among different measurement approaches and domains. Secondly, we identified substantial differences in rank-order stability across these measures. Thirdly, our results challenge existing theories by demonstrating an absence of significant age-related differences in revealed risk preferences across various task types. However, we consistently found a decline in stated risk preference with age across diverse risk domains, genders, and populations.
In our exploration of empirical properties, we applied the propensity score matching (PSM) method to control for confounding variables, meticulously examining the intricate relationship between self-reported risk preferences and life events in family and career domains. Our detailed analysis indicated that selection effects were more pronounced than socialization effects, suggesting that individuals’ self-reported risk preferences exhibited greater predictive ability on life events while being less influenced
by those events.
In conclusion, this dissertation conducts an in-depth examination of both the psychometric and empirical properties of risk preferences, aiming to elucidate whether, how, and why risk preferences vary among individuals. It reveals the complex nature of risk preferences across various measurement approaches and domains, providing invaluable insights into individual differences in risk preferences for researchers, professionals, and policymakers.