Title: Bias in asking for a second expert’s opinion

(joint work with Barbara Osimani, Michal Sikorski and Zhitao Zhang)

Abstract: In science, policy, and everyday life it is often advantageous to ask for a second expert’s opinion. Intuitively, it seems beneficial to provide the second expert with some background information, including the fact that there was an earlier assessment and its result. In this paper, we use causal models to show that providing such contextual information can lead to bias and double-counting. We argue that the bias involved is more severe than it might look at first glance since it can even arise if one is not telling the second expert the first expert’s opinion and cannot easily be counterbalanced. We use a simplified example from radiology for illustration, but the problem we describe generalises to any field in which the interpretation of data by experts is relevant.

The seminars are held online at 14:30 (Rome time) on the Microsoft Teams platform, here.