Proceedings of the Annual Meeting of the Cognitive Science Society

When reasoning about evidence under conditions ofuncertainty, one important consideration for accurate updatingis the presence (and influence) of dependencies. For instance,if considering whether a patient has a disease, the value oftwo doctors’ diagnoses indicating the presence of the diseasemay carry more value if such diagnoses were conductedindependently, rather than if, all else being equal, one doctorhas seen the other’s diagnosis before making their own. In thepresent paper, we demonstrate that lay reasoners prefer toavoid dependencies when considering evidential support.However, we additionally illustrate two cases in whichdependencies may carry evidential advantage: namely, wheninformation is partial or contradictory. Lay reasonerserroneously remain averse to dependencies even in suchcases, reflecting the difficulties inherent to considerations ofdependence.