Experience and rationality under risk: re-examining the impact of sampling experience

Abstract

A recent strand of the literature on decision-making under uncertainty has pointed to an intriguing behavioral gap between decisions made from description and decisions made from experience. This study reinvestigates this description-experience gap to understand the impact that sampling experience has on decisions under risk. Our study adopts a complete sampling paradigm to address the lack of control over experienced probabilities by requiring complete sampling without replacement. We also address the roles of utilities and ambiguity, which are central in most current decision models in economics. Thus, our experiment identifies the deviations from expected utility due to over-(or under-) weighting of probabilities. Our results confirm the existence of the behavioral gap, but they provide no evidence for the underweighting of small probabilities within the complete sampling treatment. We find that sampling experience attenuates rather than reverses the inverse S-shaped probability weighting under risk.

An inquiry into the nature and causes of the Description - Experience gap

“…For example, while DE gap studies that estimate prospect theory weighting functions have generally reported inverse S-shaped probability weighting curves in "Description," there is considerable heterogeneity in the shapes of curves elicited in "Experience" conditions: Abdellaoui et al (2011b), and Kemel and Travers (2016) reported inverse-S shaped weighting; Ungemach et al (2009) reported S-shaped weighting; while Hau et al (2008) found linear weighting. More recently, Kopsacheilis (2018) put forward the "Relative Underweighting Hypothesis," according to which people overweight rare events in "Experience" but less so than in "Description," a hypothesis that was later corroborated by findings in Aydogan and Gao (2020). This hypothesis is accommodated by an inverse S-shaped probability weighting function in "Experience" that is closer to the diagonal for probabilities closer to 0 or 1, when compared to "Description.…”

“…But, there is evidence that DE gaps can also arise, albeit typically weaker, in the absence of sampling bias, from studies that control for it by engineering 'Experience' treatments to ensure that experienced and objective probabilities coincide (e.g. Hau et al, 2010;Ungemach et al, 2009;Barron & Ursino, 2013;Aydogan & Gao, 2020). DE gaps observed in such setups require an explanation that goes beyond biased information, prompting consideration of preferences, cognitive processes or both.…”

“…Classic references in psychology include Barron and Erev (2003), Hertwig et al (2004), and Weber et al (2004). Examples of more recent interest among economists include Abdellaoui et al (2011b), Kopsacheilis (2018), and Aydogan and Gao (2020). Wulff et al (2018) provides a meta-analytic review.…”

“…However, using multiple measurement approaches is an ambition shared by other recent papers. One example is Aydogan and Gao (2020), which prefaces an "individual data" analysis of probability-weighting functions with an "aggregate data" analysis based on choice frequencies. Yet, their aggregate analysis is not entirely model-free, as it uses a model to control for utility curvature.…”

The Making-Of: Innovation. Understanding and Designing the Environment for Non-R&D Innovation

“…In some studies, these information asymmetries appear to fully explain choice differences (e.g. Aydogan and Gao 2020;Fox and Hadar 2006;Rakow et al 2008), in others, sampling bias only attenuates the gap while differences of varying degrees still exist (e.g. Camilleri and Newell 2011b;Frey et al 2015;Hau et al 2010;Ungemach et al 2009).…”