Effect of judges’ gender on rape sentencing: A data mining approach to analyze judgment documents

This study took advantage of newly released court sentencing documents to evaluate the effect of judges’ gender on juridical decision making in China. More specifically, it focused on whether the gender of judges and gender composition of a collegiate bench exert effects on sentencing outcomes after controlling for legal circumstances. All rape sentences issued from 2012 to 2015 and uploaded to China Judgements Online (CJO) were retrieved. Each judge’s gender was estimated using the naive Bayesian classifier. To compensate for possible misclassifications of gender and the gender composition of a collegiate bench, a measurement error model (MC-SIMEX) was implemented for both count (length of sentence in months) and binary (whether the defendant received probation) outcomes. Although our results reveal no consistent differences in sentencing outcomes between male and female judges, we found femaledominated collegiate benches to be associated with shorter sentences. This article not only furthers scholarly knowledge of the role a judge’s gender plays in the juridical process in China, but provides analysts with an effective way of applying text-mining techniques to social science research.