While early arrest studies tended to be more qualitative in nature and focus on the nonarrest process, more recent quantitative research has centered on the factors most likely to predict arrest. Using observational data, the current study draws on both lines of inquiry by examining incidents of nonarrest where there was a threshold of evidence that would support an arrest. A multivariate model of nonarrest decision making is presented, and qualitative narratives are analyzed to disentangle actions that police took in lieu of arrest, as well as reasons for alternate actions. The findings indicate that nonarrest behavior is much more prevalent than arrest, irrespective of evidence strength, and that several situational factors are statistically related to nonarrest decisions. The findings also show that alternative arrest actions are not demonstrably different than those noted by descriptive studies some 40 years ago, although the reasons for nonarrest behavior are substantially more varied.