Official Arrest Trajectories Across A 27-Year Span in At-Risk Men

• In Event: Life-Course Studies of Trajectories in and Age-Period-Cohort Effects on Crime

Abstract

Purpose and design: Eggleston et al. (2004) demonstrated that length of follow-up and inclusion of exposure time and mortality information can impact findings of semi-parametric group-based modeling analyses. Few other studies have investigated these issues. The aim of this longitudinal study was to model arrest trajectories across a 27-year span for 206 at-risk young men from the Oregon Youth Study. Prior research with this sample (Wiesner et al., 2007) only covered a 17-year span and did not control for mortality of the men. Annual counts of official arrests were derived from juvenile and adult court records. Findings and conclusions: Semi-parametric group-based modeling using Stata (Nagin, 2005) identified three arrest trajectories from ages 10-11 to 37-38 years (controlling for mortality and exposure time): rare offenders (62.8%), low-level chronic offenders (21.6%), and high-level chronic offenders (15.7%). Classification quality was high (e.g., entropy: 0.88, all OCC > 20). Findings showed that length of follow-up, exposure time, and mortality information altered some trajectory attributes when compared to the prior study, especially for high-rate offenders. The arrest trajectory groups also differed in terms of types of crimes committed (e.g., violent offenses). Theoretical and policy implications of these findings will be discussed.

Authors

• Margit Wiesner, University of Houston

• Deborah M. Capaldi, Oregon Social Learning Center

• Joung Yon Choi, University of Houston

• David M. Day, Toronto Metropolitan University

• Weiwei Wu, University of Houston