Character, ‘Propensities’, and the (Mis)use of Statistics in Criminal Trials 1

Jones is on trial for a violent assault on his daughter. Proof of the following facts is available; each increases the statistical probability that Jones committed this assault: Jones belongs to an ethnic group among whom this kind of crime (of violence, of violence against family members, of violence against daughters) is significantly more prevalent than it is among the general population. Also, Jones has prior convictions for criminal offences; or for crimes of violence; or for assaults on his daughter. Should the prosecution be allowed to introduce such facts as evidence of Jones's propensity to commit the crime of which he is now accused? We engage with recent arguments in defence of the (in principle) admissibility of such evidence offered by Federico Picinali and by Mike Redmayne, and we suggest that the most persuasive arguments against its admissibility are grounded in political rather than in moral philosophy.