The data used in this analysis were collected as a part of a larger evaluation of police services conducted in 1977. The data are comparable to previous observational data on police behavior in that 5,688 police-citizen contacts were observed and recorded by trained civilians riding on 900 patrol shifts. The organizational properties examined in relation to the arrest decision were (1) the level of bureaucratization, (2) the average educational level of police officers in a given department, (3) the proportion of black officers on the force, (4) the average level of officer job satisfaction, (5) the mean number of years of officer experience, and (6) the proportion of encounters involving a supervisor. The analysis indicates that arrest decisions are influenced by the demands of the immediate situation, certain organizational properties of police departments, and the organizational context in which police-citizen encounters occur. By far the most important factors influencing the arrest decisions were certain aspects of the police-citizen encounter itself, such as the dispositional preferences of complainants, the seriousness of the offense, the relational distance between suspect and victim, and the involvement of a supervisor in the encounter, Three properties of police agencies -- mean officer education, mean officer experience, and the degree of bureaucratization -- directly influence the probability of arrest, but the strength of these effects is uniformly small. The influence of several situational variables on the arrest decision is conditional on the organizational context in which the police-citizen encounter occurs. Implications of these findings for the measure of police performance are discussed. Eight notes and 44 references are provided, along with tabular data.