Evidential Proximity, Independence, and the evaluation of carcinogenicity

This paper analyses the methods of the International Agency for Research on Cancer (IARC) for evaluating the carcinogenicity of various agents. I identify two fundamental evidential principles that underpin these methods, which I call Evidential Proximity and Independence. I then show, by considering the 2018 evaluation of the carcinogenicity of styrene and styrene-7,8-oxide, that these principles have been implemented in a way that can lead to inconsistency. I suggest a way to resolve this problem: admit a general exception to Independence and treat the implementation of Evidential Proximity more flexibly where this exception applies. I show that this suggestion is compatible with the general principles laid down in the 2019 version of IARC's methods guide, its Preamble to the Monographs.