Document Type


Publication Date



Removal jurisdiction, Forum selection, Win rates


Civil Procedure | Jurisdiction | Litigation


General Observations on Interpreting Win-Rate Data Properly. Many empirical legal studies use data on plaintiffs' rate of success, because of those data's ready availability and apparent import. Yet these "win rates" are probably the slipperiest of all judicial data. Win rates are inherently ambiguous because of the case-selection effect. The litigants' selection of the cases brought produces a biased sample from the mass of underlying disputes. The settlement process, usually conducted by rational and knowledgeable persons who take into account and thereby neutralize the very factor that one would like to study, produces a residue of litigated cases for which the win rate might indicate nothing more than the percentage of successful plaintiffs in this peculiar and nonrandom sample of cases.

Nevertheless, careful research and theorizing can sometimes tease out an explanation of win-rate data by isolating case-selection effects to reveal meaningful non-case-selection effects, such as the effect of the forum chosen. This artful process of win-rate explanation involves controlling for and otherwise investigating multiple variables to see which of the possible explanations conform to the additional evidence, and then applying a plausibility screen to the surviving explanations.

Specific Study of Removal Jurisdiction. Plaintiffs' win rates in removed cases are very low, compared to cases brought originally in federal court and to state cases. For example, our data reveal that the win rate in original diversity cases is 71%, but in removed diversity cases it is only 34%. In a regression controlling for many case variables, this "removal effect" remains sizable and significant. The explanation for this phenomenon could be the ready one based on the purpose of removal: by defeating the plaintiffs' forum advantage, defendants thereby shift the biases, inconveniences, court quality, and procedural law in their own favor. Alternatively, the explanation might lie not in forum impact, but instead in case selection: removed cases may simply be a set of weak cases involving (i) out-of-state defendants who have satisfied or settled all but plaintiffs' weakest cases or (ii) plaintiffs' attorneys who have demonstrated their incompetence by already exposing their clients to removal.

Our analysis indicates that both case selection and forum impact are at work. Thus, forum really does affect outcome, with the removal process taking the defendant to a much more favorable forum.

Publication Citation

Published in: Cornell Law Review, vol. 83, no. 3 (March 1998).