8/28/2014 feature story
Using data on 130,000 felons, Jeffrey Morenoff and Jeffrey Smith examine the effects of incarceration on mortality risk by type of sentence and across demographic subgroups and causes of death.
Much research on the social determinants of health investigates how aspects of the social environment associated with race/ethnicity and SES are linked to biological processes related to health effects – and how these effects may cumulate over the life course. Over the last four decades in the U.S., a social change that has had significantly disproportionate impacts on minorities and the poor is the 705% increase in the size of the incarcerated population. Yet relatively little research has investigated the effects of incarceration on health and health disparities, in part because the incarcerated population is underrepresented in most data used to study health and mortality. Moreover, prior studies in this area have been limited by the difficulty in finding appropriate comparison groups and their vulnerability to unobserved confounders that threaten causal inferences. This project examines the effects of incarceration on the risk of mortality using data on the population of all individuals sentenced to felonies in Michigan from 2003 to 2006 (n=130,000) and matching these individuals to death records from the National Death Index. Causal effects rest on comparisons between convicted felons who received different types of sentences (e.g., prison time, probation) and are estimated using (a) an instrumental variables design that leverages the random assignment of felony cases to judges, and (b) a design that capitalizes on "regression discontinuities" in the probability of being sentenced to prison based on the operation of the Michigan Sentencing Guidelines. Analyses will also explore effect heterogeneity in the impact of incarceration on mortality among felony offenders across demographic subgroups and causes of death. This project is significant in its focus on the health effects of incarceration, its use of a quasi-experimental design that allows for stronger causal inferences, its compilation of longitudinal data on a large statewide cohort of felony offenders, and its comparisons across offenders receiving different types of sanctions.
Jeffrey Morenoff, Jeffrey A. Smith
- News & Events
- Info Services