An examination of the problems of drawing causal inferences about deterrence from models of crime rates and criminal justice activity accompanies an estimation of four different specifications of a model for police spending and violent crime rates. An examination of the relationship between self-reports of past behavior and predictions of behavior focuses on its relevance to deterrence theory and to estimations of delinquency correlates. A study of causal order in the perceptual deterrence literature uses a two-wave panel design, based on interviews with college students regarding past involvement in specific illegal acts and perceptions of their risk of arrest. A literature review describes evaluations of three types of legal interventions related to drinking and driving: the adoption of strict laws similar to those in Scandinavia, special enforcement campaigns based on these laws, and increases in the threatened severity of sanctions. Most of the studies used interrupted time-series analyses. An examination of criminal justice agencies' effects on juvenile clients emphasizes methodological issues, with illustrations from the Law Enforcement Assistance Administration's (LEAA) deinstitutionalization of status offenders program. A discussion of the causal modeling of time-series data describes the Pierce-Haugh method and illustrates its use in deterrence research. Figures, tables, notes, and chapter reference lists are provided. For individual articles, see NCJ 88196-88201.