Predictors of Recidivism Among Incarcerated Elders: Do Sex, Drugs, or Violent Offenses Matter?

Background and purpose: The crisis of aging people in prison is rapidly becoming a national economic, moral, and public safety concern, especially for the release of elders with histories of serious sex or violent offenses. Evidence suggests that older people are less likely to recidivate than their younger counterparts but older people with violent offense histories are often denied parole. There is a gap in the literature that examines what factors predict recidivism, especially sex and violent offense histories.  The current study fills a gap in the literature by exploring the role of criminal and substance use histories and other potential explanatory factors, such as life course trauma and mental health in predicting risk of recidivism among a sample of older adults.

Methods:  This study used a cross-sectional correlational design with 677 adults aged 50 and older (650 men; 24 women) in a Northeastern prison system. It explored group differences on older adults with histories of sex, violent, or drug offenses as well as other potential factors, such as cumulative trauma (as measured by the Life Stressors Checklist-Revised), PTSD (as measured by the Post Traumatic Stress-Civilian Scale) and multiple indicators of substance abuse including diagnosis, under the influence at the time of the offense, and drug offense history. The Dillman et al. (2009) method for mailed surveys was used to maximize response rates. Chi square and binary logistic regression analysis were used to assess which variables significantly increased the likelihood of recidivism.  

Findings: Both groups (those with sex offenses and nonsexual offenses) who recidivated had higher mean LSC-R scale scores (p<.001).  However, no significant differences were found for the PTSD scales for recidivists and non-recidivists. Results of chi-squares tests for violence, drugs, alcohol use and mental issues showed that recidivists scored significantly higher on all subscales except for mental health. The sex offense group who recidivated were more likely to have committed a violent offense (x²₍₁₎ = 35.0, p<0.001) compared with nonsexual offenders, and were more likely to have a drug offense (x²₍₁₎ = 13.7, p<0.001) and be under the influence drugs at the time of the offense (x²₍₁₎= 12.6, p<0.001) compared with non-recidivists. Finally, using binary logistic predicting the probability of being a recidivist among those with sex offenses, having a violent offense (p=.007) and the LSC-R total score (p=<.001) were significant.

Conclusions: Using a life course perspective, of relevance to risk assessment and policy for all ages, for the sex offense group use of drugs and having a violent offense should be considered, and appropriate community based prevention and interventions efforts should be used with lower risk justice involved individuals in accordance with risk, need, responsivity approaches and restorative justice strategies.  Since substance use and drug offense histories were a significant predictor for all offense group on recidivism, substance use prevention and treatment is warranted.