Modeling uncertainty in latent class membership: A case study in criminology

Social scientists are commonly interested in relating a latent trait (e.g., criminal tendency) to measurable individual covariates (e.g., poor parenti ng) to understand what defines or perhaps causes the latent trait. In this article we develop an efficient and convenient method for answering such qu estions. The basic model presumes that two types of variables have been mea sured: response variables (possibly longitudinal) that partially determine the latent class membership, and covariates or risk factors that we wish to relate to these latent class variables. The model assumes that these obser vable variables are conditionally independent, given the latent class varia ble. We use a mixture model for the joint distribution of the observables. We apply this model to a longitudinal dataset assembled as part of the Camb ridge Study of Delinquent Development to test a fundamental theory of crimi nal development. This theory holds that crime is committed by two distinct groups within the population: adolescent-limited offenders and life-course- persistent offenders. As these labels suggest, the two groups are distingui shed by the longevity of their offending careers. The theory also predicts that life-course-persistent offenders are disproportionately comprised of i ndividuals born with neurological deficits and reared by caregivers without the skills and resources to effectively socialize a difficult child.