Priest and Klein’s 1984 article, “The Selection of Disputes for Litigation,” famously hypothesized a “tendency toward 50 percent plaintiff victories” among litigated cases. Despite the article’s enduring influence, its results have never been formally proved, and doubts remain about their meaning, validity, and generality. This article makes two main contributions. First, it distinguishes six distinct hypotheses plausibly attributable to Priest and Klein. Second, it mathematically proves or disproves the hypotheses under a formalized and generalized version of Priest and Klein’s model. The Fifty-Percent Limit Hypothesis and three other hypotheses attributable to Priest and Klein (1984) are mathematically well-founded and true under the assumptions made by Priest and Klein. In fact, they are true under a wider array of assumptions. More specifically, the Trial Selection Hypothesis, Fifty-Percent Limit Hypothesis, Asymmetric Stakes Hypothesis, and Irrelevance of Dispute Distribution Hypothesis are true for any distribution of disputes that is bounded and both positive and continuous near the decision standard, even if the parties’ prediction errors are not independent. The Fifty-Percent Bias Hypothesis is true when the parties are very accurate in estimating case outcomes, but only sometimes true when parties are less accurate. As shown in Klerman and Lee (2014), the No Inferences Hypothesis is false.


Civil Procedure | Courts | Law | Law and Economics | Litigation

Date of this Version