Societies and organizations spend considerable resources in search of the truth or to identify the true state of the world. For example, it must be decided whether a defendant is guilty, an economy is in recession or a given parcel of land bears oil. Usually, the decision is made in a state of relative ignorance because no one knows the truth. The question then arises: By what rule should the society decide? In particular, is the society more likely to succeed by delegating the task of discovery to one of its members or to a jury (committee) which follows majority-rule voting? If the latter, who should be on the jury? When the number of alternatives is two, Condorcet's jury theorem (CJT) states that under certain conditions a majority of jurors would do better than any one individual. Unfortunately, these conditions are difficult to satisfy: CJT assumes that jurors (i) share a common goal, and (ii) vote independently. In reality, assumption (i) may not be met, and assumption (ii) is seldom met. This research investigation seeks to advance Condorcet's jury theorem in various directions. It synthesizes the preceding work on CJT by permitting simultaneously both conflict of interest and correlated votes. It studies choice situations where the alternatives are not fixed but arise from interactions among the voters. In particular, the research seeks to specify conditions under which a relatively uninformed electorate attains the outcome it would choose with full information.
