These instruments are the probability of being arrested, given that the arrestee is not incarcerated; the probability of being incarcerated; the length of incarceration; and the length of postprison probation. Empirical analysis provides measures of the crime control impact of these sanctions on individual UCR offense rates. It uses aggregate, cross sectional data, as well as data relating to individual UCR offenders newly admitted to prison, from Georgia (1978) and North Carolina (1979). The strongest support for the existence of a deterrent effect related to the risk of incarceration. Negative coefficients were obtained, often with magnitudes exceeding twice their standard errors. The variation in the incarceration rate coefficient across models and estimating procedures was very large, thereby considerably limiting the value and practical significance of the inferences to be drawn from the data. The existence of an inverse relation between offense rates and arrests whose outcome was not incarceration was also established, but with somewhat less certainty than for the incarceration rate relation. The evidence concerning the deterrent effect of postprison probation, limited to the Georgia sample, suggested that the probation component of split-sentencing acted as a deterrent to UCR offenders. The evidence concerning the sentence length variable was not consistent with the deterrence (or incapacitation) hypothesis. Tables and about 75 references are included. Appendixes present data sources, definitions, estimation procedures, and miscellaneous data tables.