Abstract
The authors examine the potential effect of accounting for exposure time by examining the arrests of 272 serious offenders who were paroled at age 18 and followed through age 33. The authors describe the overall change in the arrest rate over the 16-year period, with and without adjustments for exposure time. The authors also estimate latent class models that decompose the heterogeneity of arrest rate trends, with and without variation in exposure time. Two results are noteworthy: (a) conclusions about the level of arrest activity did depend on adjustments for exposure time, but the overall trend in arrest activity did not depend on these adjustments; and (b) latent class analysis without exposure time adjustments suggested that more than 92% of the sample exhibited their highest level of arrest activity in late teens and early 20s; then offending declined during the late 20s and early 30s. When adjusted for exposure time, the analysis revealed that about 72% of the sample exhibited this decline; the remainder remained quite active in offending.
