This article reports the results of the application of interactive analyses to three related hypotheses of disparity. The typicality hypothesis proposes that women are treated with chivalry in criminal processing, but only when their charges are consistent with stereotypes of female offenders. Selective chivalry predicts that decision makers extend chivalry disproportionately to white females. Differential discretion suggests that disparity is most likely in informal decisions such as charge reduction rather than in formal decisions at final sentencing. Data for the analysis derived from 9,966 felony theft cases and 18,176 felony assault cases disposed in California in 1988. Gender disparity was evident in findings that females with no prior record were more likely than similar males to receive charge reductions, and this enhanced females' chances for probation. The only indication of selective chivalry was a greater tendency to change charges of assault to nonassault among white female defendants than among minority females. Pivotal decisions concerning charge reduction provided partial support for the notion of differential discretion. The findings provided no clear support for the typicality thesis. Results of related studies are discussed briefly. Suggestions for future research are presented. Tables, notes, references