DNA Databases and the Arrest of Repeat Offenders: A Quasi-Experimental Test of DNA Database Expansion

Abstract Forensic DNA databases are designed to identify repeat offenders by matching crime-scene evidence to profiles of individuals who have previously entered the criminal justice system. If they operate as intended, expanding the enrolled population should increase the share of arrested defendants with prior felony records. We test this proposition using Arizona’s January 2004 law requiring DNA collection from all convicted felons. The analysis includes 8,296 felony defendants prosecuted in Maricopa and Pima Counties across eight waves of the State Court Processing Statistics program from 1994 to 2009, along with 107,639 defendants from 65 comparison counties. While a conventional pre-post comparison suggests a 71% increase in the odds that defendants had prior felony arrests after the expansion, this association disappears after accounting for preexisting trends. Likewise, a difference-in-differences estimate is not robust to violations of the assumption of parallel trends, alternative trend specifications, or placebo tests. The DNA database effectiveness perspective also fails to predict offense-specific and criminal history patterns. These findings support the limited-effect perspective and suggest that expanding DNA database eligibility alone is unlikely to substantially increase repeat offender identification.