After an intensive period of conflict resolution training, a selected group of 18 officers operated as generalist-specialists in one precinct of New York City with a population of about 90,000. In 21 months, the officers intervened in 1,388 disturbances in 962 families. Results showed there were no assaults upon any of the 18 officers despite a statistical probability that there would be; no homicides occurred in any family known to the officers; family assaults diminished; the officers made referrals to a range of social and mental health agencies, while the few referrals made by the comparison precinct were mostly to family court; and a favorable community response was perceived. Further, a quasi-experimental study was conducted with the police of the New York City Housing Authority. Introducing elements of control that were not possible in the previous feasibility study, the second research was conducted in three public housing projects over a 1-year period. In addition to confirming the findings of the original demonstration, officers trained in conflict management were determined to be measurably superior in all police performance to officers not so trained. A third study was conducted in Norwalk, Connecticut, to determine how police officers not specially trained in conflict management typically intervene in interpersonal conflicts (not only domestic conflicts). The technique of police intervention was found to depend on the circumstances of the dispute, the motivation, the background and training of the officer, police department policies, and community laws and resources. Surprisingly, all three studies showed a low percentage relationship between alcohol use and the interpersonal conflicts investigated. Further, assaults were less common when alcohol had been used. The most effective approach to police intervention in interpersonal conflicts is to give discretion to an officer trained in conflict management, while having available a broad range of options for each situation.