Session Submission Type: Policy Panel
This thematic panel reviews recent advances in the harmonization and integration of criminal justice administrative records to support evidence-based policy making. The Criminal Justice Administrative Justice System (CJARS) is a collaboration between the University of Michigan and the U.S. Census Bureau with the aim of building a data infrastructure with nationwide coverage of the major events that occur in the justice system (i.e., arrests, criminal court case filings, and terms of probation, incarceration, and parole). The unique advantage of CJARS is that criminal justice records can be linked at the individual-level with extensive survey and administrative data held by the U.S. Census Bureau. In this thematic panel, we will discuss recent advances made by CJARS to: 1) develop a benchmarking validation system that quantifies the quality of data harmonization algorithms against published aggregate statistics from the Bureau of Justice Statistics, 2) demonstrate a novel hierarchical machine-learning tool for text-based offense classification, 3) highlight key research projects being conducted using CJARS data (inequalities in the justice system, impacts of fines and fees, and labor market outcomes for justice-involved individuals since the Great Recession), and (4) details on how to gain access to CJARS data through the Census Bureau’s FSRDC network.
Benchmarking the Criminal Justice Administrative Records System’s Data Infrastructure - Jordan Papp, University of Michigan; Michael Mueller-Smith, CJars University of Michigan
Developing Modern Criminal Offense Classification Techniques - Jay Choi, University of Michigan; Michael Mueller-Smith, CJars University of Michigan; David Kilmer, Measures for Justice; Sema Taheri, Measures for Justice
Advances in Research Using the Criminal Justice Administrative Records System - Michael Mueller-Smith, CJars University of Michigan; Keith Finlay, U.S. Census Bureau; Elizabeth Luh, University of Michigan
How to Access Criminal Justice Administrative Records System Data - Shawn Ratcliff, U.S. Census Bureau