Much of the literature on statistical evidence and legal decision-making has focused on what is peculiar, epistemically speaking, with statistical evidence. In common with a minority of others, however, I focus rather on what is peculiar, morally speaking, with (some kinds of) statistical evidence. Others have proposed there are moral costs in appealing to disrespectful generalisations about people.

I consider a different and independent kind of moral cost, however, to do with arbitrariness and thus unfairness in the supporting evidence grounding a case against an individual. Whether this ought to be mitigated is another matter, since on my account (unlike a recent proposal of Di Bello and O’Neil 2020), unfair evidence is ubiquitous, if not always so extreme as in the striking cases of statistical evidence.