You are not worth the risk: The ethics of statistical discrimination in organisational selection of applicants

Abstract

Your job application is rejected unseen because you ticked a box admitting you smoke. The employer screened out applicants who ticked the 'smoker' box, because she had read empirical studies that suggest smokers, as a group, are a higher productivity risk than non-smokers. What distinctive ethical concerns inhere in the organisational practice of discriminating against applicants on the basis of group risk statistics? I argue that risk-focussed statistical discrimination is morally undesirable due to the lack of respect for applicants as unique autonomous agents. However, I argue further that the decision-making context affects the morality of this discrimination. Other things being equal, the morality of statistical discrimination varies depending on the purpose of the organisation, the level of detail in the discrimination, and whether the discrimination is transparent to applicants and includes some benefit for applicants. Because organisations may have good reason to use risk-focussed statistical discrimination when assessing applicants, I present some recommendations for decision-makers to mitigate the lack of respect for applicants as individual agents. Organisational decision-makers can focus on the extent to which the statistical data they use comprise i) factors that feature efforts and achievements of the applicant; ii) dynamic rather than static factors; and iii) data drawn from the applicant’s own history and actions over time.