Substantial evidence exists that the overall justice system deters crime. This essay examines the existence of an association between sentencing severity and levels of crime. A useful way of testing a particular supposition is by distinguishing between the research hypothesis (that a relationship between two variables exists) and the null hypothesis (that there is no association between these same factors). The problem with this approach is that one cannot logically prove the absence of a phenomenon. Social scientists have been cautious in accepting the null hypothesis that sentence severity does not reduce levels of crime. A large number of studies are examined to determine the deterrent effects of variations in sentence severity, and discern trends over time, providing insights into the stability of results. The majority of these reviews do not support the claim that harsher sanctions deter. The studies that have found support for the notion that harsher sentences deter are relatively few in number. They also suffer from one or more serious methodological, statistical, or conceptual problems that render their findings problematic. The null hypothesis that variation in sentence severity does not cause variation in crime rates should be conditionally accepted. Deterrence-based sentencing makes false promises to the community. As long as the public believes that crime can be deterred through harsh sentences, there is no need to consider other approaches to crime reduction. Sentencing systems that do not subscribe to general deterrence already exist in several nations such as Finland and Canada. 5 tables, 70 references