A re-evaluation of New York City criminal justice projects of Fishman study

Since forming homogeneous strata depends upon the data at hand, it is possible that the Cum f rule may occasionally provide the most homogeneous strata. Nonetheless, the re-evaluation indicates that it is prudent to use the present method whenever a small number of projects are grouped into strata as was done by [Fisman, R. (1977a). Criminal Recidivism in New York City: An Evaluation of the Impact of Rehabilitation and Diversion Services. New York: Praeger Publishers.].