Modeling the Subjective Perception of Uncertainty

The decision sciences have long striven to measure and model how people perceive uncertainty in specific choice situations. In the latter, people may be confronted with two types of uncertainty—epistemic and aleatory uncertainty—and in principle these may occur independently of each other. But how specific are people’s subjective perceptions thereof? And how are they related to perceptions of general uncertainty? We aimed to address these and related questions by reconciling different approaches to inducing and measuring perceptions of epistemic and aleatory uncertainty. Specifically, we conducted a study with a 2 (epistemic uncertainty: low vs. high) x 2 (aleatory uncertainty: low vs. high) x 2 (choice modality: described choice scenarios vs. incentivized monetary lotteries) mixed experimental design (N = 185). The key results suggest that people’s perceptions of epistemic and aleatory uncertainty are strongly correlated and that both types of uncertainty are associated with the perception of general uncertainty. Additionally, learning about choice options in decisions from experience was associated with reduced epistemic and aleatory uncertainty. Finally and at a more practical level, a novel and frugal measurement tool consisting of only two items achieved the same (and in some conditions an even better) performance as compared to the established but lengthier Epistemic Aleatory Rating Scale (Ülkümen et al., 2016). Overall, these findings provide insights on how to study epistemic and aleatory uncertainty with incentivized tasks and choice scenarios, as well as how to measure subjective perceptions of these types of uncertainty. As such the current study contributes to the conceptual debate on the structure of uncertainty perceptions.