How public registries obscure sexual recidivism risk: Evidence from state administrative data.

Abstract

Public sex offender registries were created on the assumption that people with a history of sexual offending present a high and enduring risk for sexual recidivism. Analyzing data from everyone on Michigan’s sex offender registry (n = 45,145), we found that recidivism risk was very low for many registrants—and predictably declined in systematic ways that are relevant to registry design. The sexual recidivism rates were 3%–5% after 5 years and 10% after 20 years. Risk of future offending predictably declined the longer people remained sex-offense-free in the community. Assuming a typical distribution of initial risk levels and the observed time-free in the community, the likelihood of sexual recidivism was very low for approximately half of the registrants: out of 35,000 registrants in the community in Michigan, the risk level for 17,000–19,000 individuals was no different from the likelihood of first-time sexual offending in the general male population (<2% lifetime). Michigan’s registry provides no means for the public to differentiate between the noise presented by the large number of very low-risk individuals and the signal representing the minority of registrants who may be genuinely high risk. Major reforms are needed if public sex offender registries have any potential for advancing community safety. However, instead of placing our hopes in large, passive, and stigmatizing registries, we are more likely to reduce sexual recidivism by investing in individualized supervision, treatment, and other support for people at high risk for sexual recidivism. (PsycInfo Database Record (c) 2025 APA, all rights reserved)