Risky Choice 2

The literature suggests that profit-harming trading behaviors are driven either by beliefs or preferences, while the relative contribution of each is unclear. We address this question by using computational modelling to determine whether belief formation or preferences better explain these decision patterns. We analyzed a dataset of 192 participants who completed an investment task over 150 rounds. We observed that individuals tended to hold the asset more frequently than shorting it, were not invested in a number of rounds, and reacted slowly to price changes. We defined risk-neutral Bayesian updating as our benchmark model and investigated the improvement of models along beliefs and preferences, based on a selection of well-established mechanisms. We tested all models that included combinations of both dimensions by changing belief formation mechanisms through reinforcement learning and differential updating for gains and losses, and by incorporating preferences such as risk preferences and loss aversion. Model comparisons show that modelling belief mechanisms was much more important than preferences. Reinforcement learning and differential learning in gains and losses showed strong effects. Risk and loss preferences, and even their combinations with different types of beliefs, led to only small refinements. Inspired by the importance of learning, we tested whether people learn not only by reward, but also by repetition, thereby building habitual preferences. Implementing habits produced strong effects across all mechanisms used to form beliefs. The good fit of the augmented RL model together with habitual tendencies to our data demonstrates the crucial role of learning processes in financial decisions.

This is an in-person presentation on July 21, 2024 (11:40 ~ 12:00 CEST).

The present study examines the effect of social distance on choice behavior through the lens of a probabilistic modeling framework. In two identical experiments, conducted three weeks apart, participants made incentive-compatible choices between lotteries in three different social distance conditions: self, friend, and stranger. We conduct a layered, within-subjects analysis that considers four properties of preferential choice. These properties vary in their granularity. At the coarsest level, we test whether choices are consistent with transitive underlying preferences. At a finer level of granularity, we evaluate whether each participant is best described as having fixed preferences with random errors or probabilistic preferences with error-free choices. In the latter case, we further distinguish three different bounds on response error rates. At the finest level, we identify the specific transitive preference ranking of the choice options that best describes a person's choices. At each level of the analysis, we find that the stability between the self and friend conditions exceeds that between the self and stranger conditions. Stability increases with the coarseness of the analysis: Nearly all people had transitive preferences regardless of the social distance condition, but very few had the same preference ranking in every social condition. This pattern of results replicated across the two experiments. Overall, while it matters whether one makes a choice on behalf of a friend versus for a stranger, the differences are most apparent when analyzing the data at a high level of granularity.

This is an in-person presentation on July 21, 2024 (12:00 ~ 12:20 CEST).

Many important decisions are based on experience. Recently, the experimental paradigm of sampling a fixed number of outcomes, and afterwards making one consequential decision has received much attention in the decision-making literature. A puzzling phenomenon in this literature was that participants systematically chose the higher-variance option (e.g., Ludvig & Spetch, 2011; Tsetsos et al., 2014), that way contradicting classical work in decision-making concluding that people are risk averse for symmetrical outcome distributions. Here, we examined the robustness of this phenomenon in three experiments (sample size at least 176 per experiment). We varied whether single valuations came directly before choice task or were separated from each other; whether sequences were presented simultaneously or sequentially; and whether participants chose/valued two or four sequences per block. In all experiments, participants were risk seeking or risk neutral in the choice tasks, but risk averse in single independent valuations of the same outcome sequences (certainty equivalents). With computational modeling we show that the effect of choosing high-variance options can be explained through the comparison process between the presented outcomes in choices. In contrast, in single valuations, a noisy compressed mental number line explains the data best. We conclude that risk taking behavior must not necessarily indicate risk seeking or convex mental number lines, but rather can be explained through the process of outcome comparison during information sampling. Therefore, to understand risk taking across different answer modalities and contexts, it is crucial to explicitly model information processing.

This is an in-person presentation on July 21, 2024 (12:20 ~ 12:40 CEST).