625
Analysis of Arrest Rates
Subir Sinha
Abstract:
We investigate the offenses committed by felons in the
United States (U.S.). The two counties that we consider are
Los Angeles (California) and Maricopa (Arizona). The data
were collected from convicted felons in the above mentioned
two counties during 1986 and 1990. The primary aim of the
proposed analysis is to compare arrest rates of different
kinds of offenders to assess the crime control potential
associated with recent increases in drug offenders within
the criminal justice system. The statistical analysis for
addressing this issue involves a Bayesian analysis in which
the posterior distribution cannot be obtained in closed form
and we use instead Markov chain simulation, which produces a
sequence of random variables distributed approximately from
the posterior distribution. These random variables can be
used to estimate the posterior or features of it like the
posterior expectation and variance. In addition to the arrest
rates of offenders, there are other covariates which we
incorporate through model selection and use them to fit a
general regression model. Again the posterior cannot be
obtained in closed form and so we use Metropolis within
Gibbs to get the posterior of the regression coefficients
corresponding to the previously selected covariates which
answers some more questions of interest.
Here is the full postscript text for this
technical report. It is 1829259 bytes long.