Individual differences in belief updating across tasks variants

Belief updating is fundamental to adaptive behavior. This process of revising expectations in light of new evidence has been operationalized in different tasks putatively targeting the same core construct. However, this assumption of construct validity has not been explicitly tested. Here, we take an initial step to address this question by testing the generalization of individual differences across two close variants of an established belief updating paradigm, the helicopter task (Fromm et al., 2023) and the predator task (Satti et al., 2025). In the helicopter task, participants predict the next outcome by positioning a bucket at the bottom of the screen, with the cover story that they are catching money dropped from a helicopter hidden behind clouds. In the predator task, participants defend themselves against a predator by placing a fire at a given angle along a circle around the center of the screen whenever a warning tone signals the predator's approach. Thus, both task variants require estimation of spatial outcomes under uncertainty, yet they differ in features that are putatively superficial and thus irrelevant to the core belief updating construct, allowing us to probe cross-task stability in belief-updating mechanisms. The computational models applied to both tasks capture multiple forms of uncertainty, including relative uncertainty (RU), which reflects imprecision in the current belief relative to environmental noise, and change-point probability (CPP), which tracks the likelihood that a change in the underlying generative process has occurred and modulates how strongly new outcomes influence belief updating. Our primary hypothesis is therefore that participant-level descriptive and model-based parameters will correlate across task variants. By focusing first on this minimal pair of task variants, we aim to determine whether reliable, domain-general belief updating signatures emerge before scaling up to a broader task battery in later studies.