Abstract
This note reassesses the basic result in Mukhopadhaya (2003) that, when jurors may acquire costly signals about a defendant's guilt, with a larger jury size the probability of reaching a correct verdict may in fact fall, contrary to the Condorcet Jury Theorem. We show that if the jurors coordinate on any one of a number of (equally plausible) asymmetric equilibria other than the symmetric equilibrium considered by Mukhopadhaya, the probability of accuracy reaches a maximum for a particular jury size and remains unchanged with larger juries, thus mitigating Mukhopadhaya''s result somewhat. However, the case for limiting the jury size – a recommendation by Mukhopadhaya – gains additional grounds if one shifts the focus from maximizing the probability of reaching a correct verdict to the maximization of the overall social surplus, measured by the expected benefits of jury decisions less the expected costs of acquiring signals.