Naive, biased, yet Bayesian: Can juries interpret selectively produced evidence?

In an idealized model of civil litigation, interested parties incur costs to produce statistical evidence. A subset of this evidence is then presented to a naive decisionmaker (e.g., a jury). The jury is naive in that it views evidence as a random sample when in fact the evidence is selectively produced. In addition to being naive, the jury is also biased by prior beliefs that it carries into the courtroom. In spite of the jury's naivete and biasedness, a full-information decision is reached as long as both litigants choose to produce evidence. Our results suggest that criticisms of the jury process based on jury bias or the jury's use of simple or heuristic rules may be overstated, and underscore the potential importance of competitively produced evidence in legal decision-making.