The cases were selected from police records and consisted of those incidents where either the client was accused of assaulting a police officer or the citizen had filed charges of excessive use of force against one or several officers. Firearms were not involved in any of the incidents. The aim of the study was to identify factors that might help in preventing such incidents. The study also focused on the value of using interview statements from parties involved in an incident compared to the use of documentary sources. The violence was found to occur at five stages of case processing: apprehension, transport, arrival at the police station, immediately after arrival, and during custody. Information on the incidents was solicited from all parties involved through nonstandardized, free-flowing, topically controlled interviews. The interviewers found most of the police officers to be tense and controlled in their responses to questions; whereas, the citizen interviewees were eager to volunteer information that bolstered their accusations against the police. This dynamic of the interviews was probably due to the fact that the police may have felt they had much to lose from any possibly incriminating statements in the interviews, and the citizens had little to lose by their statements. This reveals a dilemma that police officers face; officers must protect themselves when attacked, but when responding to violence directed against them they risk damaging their careers should they use excessive force. This dilemma creates stress for police officers and must be addressed in training and ongoing stress management guidance. Also, officers must be so supervised and trained in how to manage suspects in all phases of case processing so as to minimize stimuli that tend to trigger violent behavior by the parties. 15 references