cpRossi {penPHcure}

Criminal Recidivism Data

Description

A sample of 432 inmates released from Maryland state prisons followed for one year after release (Rossi et al. 1980). The aim of this study was to investigate the relationship between the time to first arrest after release and some covariates observed during the follow-up period. Most of the variables are constant over time, except one binary variable denoting whether the individual was working full time during the follow-up period.

Usage

cpRossi

data(cpRossi,package="penPHcure")

Format

A data.frame in counting process format with 1405 observations for 432 individuals on the following 13 variables.

id

integer. Identification code for each individual.

(tstart, tstop]

integers. Time interval of the observation (in weeks). Observation for each individual start after the first release.

arrest

factor with 2 levels ("no", "yes"). Denote whether the individual has been arrested during the 1 year follow-up period or not.

fin

factor with 2 levels ("no", "yes"). Denote whether the inmate received financial aid after release.

age

integer. Age in years at the time of release.

race

factor with 2 levels ("black", "other"). Denote whether the race of the individual is black or not.

wexp

factor with 2 levels ("no", "yes"). Denote whether the individual had full-time work experience before incarceration or not.

mar

factor with 2 levels ("yes", "no"). Denote whether the inmate was married at the time of release or not.

paro

factor with 2 levels ("no", "yes"). Denote whether the inmate was released on parole or not.

prio

integer. The number of convictions an inmate had prior to incarceration.

educ

factor with 3 levels ("3", "4", "5"). Level of education:

• "3": <=9th degree;

• "4": 10th or 11th degree; and

• "5": >=12 degree.

emp

factor with 2 levels ("no", "yes"). Denote whether the individual was working full time during the observed time interval.

Source

The Rossi dataset in the RcmdrPlugin.survival package (Fox and Carvalho 2012) is the source of these data, which have been converted into counting process format.

References

Fox J, Carvalho MS (2012). “The RcmdrPlugin.survival Package: Extending the R Commander Interface to Survival Analysis.” Journal of Statistical Software, 49(7), 1–32. http://www.jstatsoft.org/v49/i07/.

Rossi PH, Berk RA, Lenihan KJ (1980). Money, Work, and Crime: Experimental Evidence. New York: Academic Press. doi: 10.1016/C2013-0-11412-2.