Inquiry and deliberation in judicial systems : The problem of jury size

Abstract:
We raise the question whether there is a rigorous argument favoring one jury system over another. We provide a Bayesian model of deliberating juries that allows for computer simulation for the purpose of studying the effect of jury size and required majority on the quality of jury decision making. We introduce the idea of jury value (J-value), a kind of epistemic value which takes into account the unique characteristics and asymmetries involved in jury voting. Our computer simulations indicate that requiring more than a > 50 % majority should be avoided. Moreover, while it is in principle always better to have a larger jury, given a > 50 % required majority, the value of having more than 12–15 jurors is likely to be negligible. Finally, we provide a formula for calculating the optimal jury size given the cost, economic or otherwise, of adding another juror.