Methodological Approaches to Modeling Criminal Trajectories

Criminal offenders compose a heterogeneous population. Criminal trajectory research aims to capture this heterogeneity in terms of the frequency or severity of offending. This chapter describes the concept a criminal trajectory and the statistical technique used to derive trajectories from longitudinal data. Both the semiparametric group-based trajectory modeling (SGBTM) and latent growth mixture modeling (GMM) approaches are described in nontechnical terms, and the differences between them are noted. Despite some similarities, these approaches are also distinguished from conventional growth curve modeling. Guidelines and factors to consider in building and testing trajectory models are discussed. Last, extensions of SGBTM and GMM are presented.