The continuous debate over the relationship between unemployment and crime has shown the theoretical importance of economic conditions in criminology. David Greenberg has argued that prior research on this relationship has been flawed with the incorrect statistical model developed by Cantor and Land and the lack of appropriate operationalized hypotheses in testing his strain theory. Greenberg claimed that the use of cointegration analysis provided the better framework in the analysis of time series data. This article addressed Greenberg’s arguments and claims through a discussion on the utility and the substantive meaning of the cointegration approach for analyzing time series data, the theoretical and empirical limits of cointegration analysis, and the possible testing of Greenberg’s strain theory. It was determined that the best time series model was dependant on the theoretical issues being tested. The cointegration approach might provide a more informative set of results when the focus of the analysis was long-term equilibrium in the levels of unemployment and crime. However, the first-differenced regression equations represented an appropriate alternative when the focus was on explaining changes in crime rates. In regards to Greenberg’s strain theory, it was suggested that in order to examine the utility of a theory the testing of its propositions should be encouraged. References