A growing number of researchers have found that approximately 10 percent of police officers cause 90 percent of the problems faced by law enforcement agencies. Being able to identify these problem officers early in their careers can help prevent serious incidents that can harm an agency's image and mission in the community it serves. EWSs have three basic phases: selection, intervention, and postintervention monitoring. This article focuses on the selection phase, as it identifies variables linked to the emergence of an officer's problem behaviors. In identifying variables linked to officer misconduct, researchers have placed them in two broad categories: officer characteristics and job characteristics. Regarding officer characteristics, "antisocial-egocentricity" as measured by the Minnesota Multiphasic Personality Inventory has been found to be related to citizen complaints against an officer and to insubordination. Other officer characteristics related to problem behaviors include a preference for using force in the performance of duties, lack of self-discipline, and poor performance in training exercises and testing. Regarding job characteristics as factors in officer problem behaviors, officers who have not had the benefit of close association with exemplary officer role models are at risk for developing problem behaviors. This suggests that agencies must be particular in the selection of field training officers, whose behaviors and actions can shape new recruits' future conduct. EWSs are likely to be effective in preventing serious officer misconduct when they are based in an information system that captures relevant information on an officer's performance. Relevant information pertains to factors that research has shown to be related to a pattern of problem behavior. 13 notes