UCR Versus Locally Acquired Arrest Data as Weighting Targets

Michael Yang, National Opinion Research Center
Dean R. Gerstein, NORC at the University of Chicago
James C. Hendrickson, National Opinion Research Center
Bruce Taylor, National Institute of Justice

Due to the challenges of survey implementation, most studies of arrestees are based on convenience samples from which statistical generalization is problematic. A primary goal of the Arrestee Drug Abuse Monitoring (ADAM) program is to estimate various characteristics of the arrestee population with known precision based on probability samples of booked arrestees in 35 metropolitan areas. Post-sampling stratification adjustments are performed to take into account facility types within sites, variations in arrestee transport patterns, release procedures for different offense types, barriers to continuous 24-hour sampling in most locations, and other similar considerations. ADAM sites currently collect the needed arrest data on all bookings from sampled facilities during the survey field period. This paper explores the possibility of supplementing or substituting these locally acquired booking data with Uniform Crime Report (UCR) data for some or all of the respective ADAM sites, which might permit reliable derivation of weights through statistical means at substantial savings in survey cost. We construct statistical models to examine the relationship between the UCR and locally acquired arrest data and draw conclusions pertinent to ADAM, other arrestee surveys, and similar survey sampling problems.

(Return to Program Resources)

Updated 05/20/2006