When do we believe in experts? : the power of the unorthodox view

Alison, Laurence and Almond, Louise and Christiansen, Paul and Waring, Sara and Power, Nicola and Villejoubert, Gaëlle (2012) When do we believe in experts? : the power of the unorthodox view. Behavioral Sciences and the Law, 30 (6). pp. 729-748. ISSN 0735-3936

Abstract

This paper examines the extent to which orthodoxy (degree of typicality) and congruence (degree of similarity with own opinion) mediate the influence of expert advice on decision makers’ judgments. Overall, 227 members of the public and 60 police officers completed an online questionnaire involving an investigation into a child sex offence. Participants were asked to first (i) formulate their own “profile” of a likely offender then (ii) estimate the guilt of two presented suspect descriptions (orthodox vs. unorthodox), and, following the presentation of an “expert’s” profile that matched either the orthodox or the unortho- dox suspect, (iii) re-evaluate their guilt judgments of the two suspects based on this new advice. Finally, (iv) the perceived similarity (congruence) between the participants’ own and the expert profile was assessed. Results revealed two key findings. First, expert profiles that matched a suspect’s description elevated perceptions of guilt in that suspect, whilst also, simultaneously, very significantly decreasing the perception of guilt of the alternative suspect. This suggests a powerful rejection and downward revision of the other suspect. Second, perceived similarity of the profile (to one’s own profile) was only a significant factor in increasing guilt judgments when assigning guilt to the unorthodox (as opposed to orthodox) suspect. Comparisons of lay judgments with those of police officers revealed few significant differences in effects. The finding that advice is most influential when unorthodox and incongruent suggests that decision makers are more likely to reevaluate judgments when expert advice contributes novel information that contradicts their beliefs. The practical implications of these findings are discussed for profilers, police, and decision research in general.