A policy capturing analysis of federal district and appellate court sexual harassment cases

A policy capturing analysis of 206 federal district and appellate sexual harassment cases showed that federal judges were moderately consistent in their judgments, with 29% of the variance in case outcomes accounted for by the regression model. Judges used Victim's Reaction, Form of Harassment, and Coercion, to make their ruling, and these three cues accounted for 20% of the variance in judgments. Appellate judges were slightly less consistent than district judges, and there were differences in judgment policy between appellate and district level cases. These findings are consistent with key aspects of the EEOC's Guidelines on Sexual Harassment and suggest that federal judges use the Guidelines to make their judgments. The model could be used by potential plaintiffs or defendants in sexual harassment cases to generate a prediction of the likely outcome of their case, to decide whether to pursue the case in court or seek a settlement.