The POPN observational data were collected by trained observers who accompanied patrol officers on their shifts. Survey data on officer characteristics were collected from structured interviews with just over 95 percent of the patrol officers. The characteristics of the neighborhoods served by the officers were determined from the 1990 census. The study compared the actions of Black and White officers in physical conflicts and verbal disputes with citizens. The dependent variables addressed the level of force used to resolve the conflict as well as the number of supportive actions taken by officers. Independent variables pertained to officer characteristics, situational characteristics, neighborhood characteristics, and three interaction terms (officer race and citizen race, officer race and predominantly White neighborhoods, and officer race and diverse neighborhoods). The study found that Black officers were generally more likely than White officers to use coercion in responding to the various types of conflicts; however, Black officers were also more likely to use supportive actions in predominantly Black neighborhoods. Black and White officers behaved similarly toward Black and White citizens, although situational variables were generally better predictors than officer characteristics in determining officer responses, such as citizen's role in the conflict, demeanor, emotional state, and evidence strength. Social disadvantage in a neighborhood apparently influenced police supportive actions, but did not affect coercive actions. This suggests that intensive training in the fair application of justice since the 1980's has possibly achieved results in officer actions toward citizens. The implications of these findings for policy and future research are discussed. 3 tables, 7 notes, and 76 references