Assessing risk of recidivism among juvenile offenders: the development and validation of the recidivism risk instrument

Lela Williams, Craig Lecroy, John P. Vivian

Research output: Contribution to journalArticle

3 Scopus citations


A recidivism risk instrument was developed and validated on a sample of juvenile offenders (N = 1,987) based on the need to classify juveniles by their likelihood of re-offense. Female recidivism (R(2) = 27%) was predicted by younger age at first expulsion from school, history of parent incarceration, gang involvement, felony class offense, and firearm use. Male recidivism (R(2) = 12%) was predicted by younger age at first adjudication, referrals, school suspensions, history of maternal incarceration, firearm use, running away, gang involvement, and destroying property/stealing. Cross-validation analyses indicated that high-risk offenders recidivated at more than five times the rate of low-risk offenders.

Original languageEnglish (US)
Pages (from-to)318-327
Number of pages10
JournalJournal of Evidence-Based Social Work
Issue number4
StatePublished - 2014



  • Adolescent
  • gender
  • predictive validity
  • stepwise regression

ASJC Scopus subject areas

  • Human Factors and Ergonomics
  • Medicine(all)
  • Sociology and Political Science

Cite this