How people say they feel about uncertainty does not match how they decide
Kamil Fuławka from our Computational Social Science group, together with colleagues Yannik Paul and Anya Pedersen from Kiel University, has published a new article in Computational Psychiatry. The study “Decomposing Intolerance of Uncertainty: No Association With Affective Decision Making in a Community Sample” addresses a central question in psychology: How do people respond to uncertainty, and what drives these responses?
Intolerance of uncertainty (IU) is considered a key factor across many psychological disorders, from anxiety to obsessive-compulsive symptoms. Yet the mechanisms behind it remain unclear. To explore this, the authors identified three possible pathways: overweighting negative outcomes, distorting probabilities of rare events, and discomfort with missing information. They then mapped these mechanisms onto well-known patterns of human decision-making described by Cumulative Prospect Theory.
To test this link, the team designed an experiment with 100 participants. Instead of abstract lotteries with money, they used medical scenarios in which people had to choose between two hypothetical painkillers, each with possible side effects and varying chances of success. In one condition, all information was provided upfront. In the other, participants had to discover probabilities by sampling outcomes themselves. This allowed the researchers to compare decision-making under risk with decision-making under uncertainty.
The results confirmed that participants displayed common biases: they overweighed side effects compared to benefits (loss aversion), and they distorted small probabilities. Choices also differed depending on whether information was described or experienced, echoing the well-known description–experience gap. However, none of these behavioral patterns correlated with how strongly individuals reported disliking uncertainty on questionnaires.
This finding is important because it suggests that self-reported intolerance of uncertainty and observable decision-making behavior may capture different aspects of the phenomenon. The study points to the need for more precise definitions of IU and for experimental designs that can create stronger affective contexts, possibly with real consequences for behavior.
