Abstract

The Condorcet Jury Theorem, together with a large and growing literature of ancillary results, suggests two conclusions. First, large committees outperform small committees, other things equal. Second, heterogeneous committees can, under the right circumstances, outperform homogeneous ones, again other things equal. But this literature has done little to bring these two conclusions together. This paper employs simulations to compare the respective contributions of size and difference to optimal committee performance. It demonstrates that the contributions depend dramatically upon bias. In the presence of low bias, committee composition matters little. In the presence of high bias, it can matter a great deal; optimal committee performance, however, does not vary dramatically between low- and high-bias committees.