Abstract
I use three separate data sets and designs to determine whether sex offender registries are effective. First, I use state-level panel data to determine whether sex offender registries and public access to them decrease the rate of rape and other sexual abuse. Second, I use a data set that contains information on the subsequent arrests of sex offenders released from prison in 1994 in 15 states to determine whether registries reduce the recidivism rate of offenders required to register compared with the recidivism of those who are not. Finally, I combine data on locations of crimes in Washington, D.C., with data on locations of registered sex offenders to determine whether knowing the locations of sex offenders in a region helps predict the locations of sexual abuse. The results from all three data sets do not support the hypothesis that sex offender registries are effective tools for increasing public safety.