Proceedings of the Annual Meeting of the Cognitive Science Society

One of the most persistent findings in causal cognition is that individuals frequently deviate from normative Bayesian reasoning, most notably through Markov violations, where they incorporate conditionally independent information into their judgments. Recent research has sought rational explanations for these deviations, ranging from structural uncertainty to memory sampling limitations. In this study, we evaluated four rational models using a novel experimental paradigm that introduced structural uncertainty into a causal chain by manipulating the functional properties of the causal links. We found that participants systematically violated the Markov assumption when causal links shared the same mechanism. Computational modeling identified the Bayesian Uncertainty Model (BUM) as the most plausible explanation for these results, with sampling-based methods like the Mutation Sampler appearing as close contenders. We discuss these findings in light of the need to rethink the definition of rationality in Bayesian causal reasoning.