Application II: Crime Recidivism

Abstract

In the second application, a data set on crime recidivism is used. The data set has been previously analyzed and obtained by Schmidt andWitte (1988). The data was obtained from files on inmates of a North Carolina prison who were released between July 1979 until June 1980. It can be freely downloaded from the webpage https://www.icpsr.umich.edu/icpsrweb/ICPSR/studies/8987?q=ICPSR as of 06/27/2014. The data set was assembled in 1984, the observation windows equals 81 month. The original sample is of size 9327, after deleting missing values 4628 cases remain. Failure time is defined as time until return to prison. Due to rounding, the failure times are interval-censored. Failure times are rounded to the nearest month, so that return to prison after 15 days would be coded 1, return after 16 days until the 15th of the following month would be coded 2. The month of release in addition to the month of return to prison are available, so the exact interval in which return to prison occured can be reconstructed.