The Impact of the Structured Assessment of Violence Risk in Youth (SAVRY) in Juvenile Corrections

Purpose: Risk assessment has become an important tool of forensic social work. As incarceration may increase a juvenile’s exposure to criminogenic influences, and lead to a greater disconnect from pro-social supports, the implementation of validated risk assessment tools may increase the Parole Board’s confidence to release more juveniles early on parole, rather than relying on habits, traditions, and guesswork.  One such risk assessment instrument, The Structured Assessment of Violence Risk in Youth(SAVRY), was adopted by New Jersey’s State Parole Board in 2009. The SAVRY instrument guides assessors through a list of 24 risk factors grouped under the domains of historical risk, social/ contextual factors, individual/ clinical factors, and six protective factors. Completion of the instrument leads to the generation of an overall “risk summary,” where the assessor concludes that the juvenile poses a low, moderate or high risk of future serious violent behavior.

Despite the potential utility of risk assessment tools, empirical research has not yet examined the actual impact of risk assessment instruments upon practice.  Given the limitation in literature, this quasi-experimental study evaluates the outcome of this statewide risk assessment implementation, and hypothesizes that juveniles who were assessed by SAVRY would be more likely to receive parole than those who were not.

Methods

Study data were drawn from 445 parole case files, of which 236 juveniles had received a SAVRY risk assessment, and 209 had received no assessment.  The two groups were drawn from all juveniles who were evaluated for parole between 2009 and 2011.  The groups were matched on age, gender, ethnicity, sentence length, seriousness of offenses, and risk of reconviction. 

A survival analysis using cox proportional hazard modeling was employed to assess whether SAVRY-assessed juveniles were likely to be released earlier than non-assessed juveniles, controlling for juveniles’ offense seriousness, risk of reconviction, age at release, and race/ethnicity being covariates.   

Results

The majority (93.7%) of juveniles were male. On average, they were released at age 18.2 years. African Americans comprised 67.6% of the sample, followed by Hispanics (20.2%), and Caucasians (11.7%). The average sentence received was 22.9 months, with 42% of juveniles committing a violent offense, 19.8% a drug-related offense, and 15.5% an offense of robbery without physical contact with the victim.

Survival analysis results show that juveniles assessed by SAVRY were 1.71 times (95% CI, 1.300, 2.258; p< .0001) more likely to receive parole than those who were not assessed.  Age at release was a significant predictor, as was the risk of reconviction, and offense seriousness score.  The ethnicity of the juvenile was not a significant predictor of parole release.

Implications: These findings support the use of risk assessment instruments in juvenile justice decision-making. Juveniles who received a risk assessment are more likely to be released early on parole than those who did not. The SAVRY instrument appears to guide Parole Board decision-makers along a carefully constructed path of risk factors, encouraging probabilistic rather than possibilistic thinking.