New Applications and Controversies in Survival and Trajectory Models

Session Submission Type: Complete Thematic Panel

Abstract

This session is about models for the analysis of longitudinal data in criminology, including survival models and latent trajectory models. The first paper reassesses recent uses of simulation methods to evaluate group-based trajectory, generalized mixture, latent class, and mixed membership models. The second paper re-examines data from the National Supported Work Demonstration, which conducted an $82 million experiment that randomly assigned disadvantaged individuals to jobs versus no jobs. The paper explores using instrumental variables to model departures from randomization within a survival model framework. The third paper estimates an event history model of desistance from gangs. Using data from the Denver Youth Survey, it examines the role of gang organization, low self-control, and gang identity on desistance. The fourth paper estimates a competing risks event history model of “crimmigration”—the process by which immigrants leave detention.

Sub Unit

• Area XII: Methodology / 71: Advances in Quantitative Methodology

Individual Presentations

• Work and Crime: A Causal Treatment Effects Survival Model - Jun Hwang, University of Washington; Ross L. Matsueda, University of Washington; Elena A. Erosheva, University of Washington; Christopher Uggen, University of Minnesota

• Trajectory Models Revisited: Assessing Recent Simulation Studies - Ross L. Matsueda, University of Washington; Bryan D. Martin, University of Washington; Elena A. Erosheva, University of Washington

• A Competing Risk Analysis of Immigrant Detention and Case Outcomes - Ryan Larson, University of Minnesota; Christopher Uggen, University of Minnesota; Linus Chan, University of Minnesota

• Tell Me Who You Hang with and I’ll Tell You What You Are: Gang Identity, Organization, and Desistance - John Leverso, University of Washington

Chair

• Daniel Nagin, Carnegie Mellon University