I examine the variables that predict arrest in rape cases based on hypotheses derived from the feminist-conflict theory, the consensus perspective, and the liberation hypothesis. Feminist-conflict theorists argue that extralegal variables influence the decisions of the police, irregardless of legal variables. Conversely, the consensus perspective argues that legally relevant variables will have the greatest impact on police decisions. The liberation hypothesis suggests that the influence of extralegal variables on police decisions depend on the strength of evidence and crime seriousness. The results from a logit regression analysis on arrest using police archival data do not support the liberation hypothesis. The feminist-conflict theory correctly predicts a decrease in the likelihood of an arrest as the intimacy between the suspect and victim increases. However, there is more support for the consensus perspective for predicting arrest as evidentiary strength is the strongest predictor of arrest.