Inferring the winning party in the supreme court from the pattern of questioning at oral argument

It is no longer a secret that a lawyer arguing a case before the Supreme Court is more likely to lose if he is asked more questions than his opponent during oral arguments. This paper rigorously tests that hypothesis and the related hypothesis that a lawyer is more likely to lose if he is asked longer questions (measured by words per question) than his opponent. Using regression analysis, we find strong evidence for both hypotheses: the number of questions asked and the number of words per question asked are both negatively correlated with a party's likelihood of winning. Although the paper is primarily empirical, we also explore the theoretical basis for these results. We analyze the role of deliberation in appellate courts and explain that because formal deliberation is often quite limited, judges use oral argument as an alternative way to express their opinions and attempt to influence other judges.