Past research on the self-esteem of bullies has produced equivocal results. Recent studies have suggested that the inconsistent findings may be due, in part, to the failure to account for bully/victims: those children who both bully and are victims of bullying. In this longitudinal study, we examined the distinctions among pure bullies, pure victims, bully/victims, and noninvolved children in a sample of 307 middle school students. Analyses of cross-sectional and longitudinal results supported the importance of distinguishing between pure bullies and bully/victims. In addition, results revealed some interesting sex differences: girls in the pure bully and bully/victim groups reported significant increases in self-esteem over time, with girls in the pure bully group reporting the greatest increase, whereas boys in these groups reported no significant changes in self-esteem over time.