Science communication is a deeply social enterprise. Scientific arguments, theories and data are not presented in a vacuum, but are delivered to the public by particular individuals with unique characteristics. Some of these characteristics involve ostensibly unrelated (im)moral or (un)ethical past actions. In theory, this information should not influence the audience's evaluation of the veracity of the individual's scientific claims. For example, knowledge that a scientist has engaged in a sexual indiscretion should not affect our assessment of his or her claims in an unrelated field (e.g. physics). In this paper, we present results from a series of controlled experiments demonstrating that this is not the case. All participants read about a debated scientific theory written by an expert in his field. In the main experiment, they were randomly assigned to one of three conditions: (1) morally good expert; (2) morally bad expert; (3) morally neutral expert (control condition). We hypothesized that academic training and experience should counteract these types of biases, and therefore tested both undergraduate students and professional scientists. Using both surveys and behavioral measures, we analysed the participants' judgement of the expert and his claims. We found that the audience's pre-existing knowledge of this person's (im)moral reputation significantly influenced their judgement of his work. Although this pattern of a distorted perception of the arguments was more pronounced among the undergraduate students, the trend was similar for the professional scientists. In a follow-up experiment, we refined the nature of the transgression to examine whether academic misdeeds (e.g., data fabrication) fared differently to moral offenses in the audience's scientific evaluations. We will discuss the implications of our ongoing research in this area for science communication. In particular, we ask how much-and what kind-of an expert's biographical detail should be communicated?
A copy of the full paper has not yet been submitted.