The California Department of Corrections and Rehabilitation (CDCR) is using one of two studies on crime, incarceration, and recidivism as management tools to assess the risk level of inmates in California state prisons.
The studies developed by the University of California, Irvine (UCI), and the PEW Center on the States used an evidence-based approach to assist the CDCR to identify levels of risk that inmates pose to public safety.
“After decades of experience managing offenders and analyzing data, practitioners and researchers have identified key factors that can help predict the likelihood of an individual returning to crime, violence or drug use,” the PEW study stated.
PEW described a risk/needs assessment tool as “essentially a uniform report card that measures offenders’ criminal risk factors and specific needs.” And further said, “When developed and used correctly, these risk/needs assessment tools can help criminal justice officials appropriately classify offenders and target interventions to reduce recidivism, improve public safety, and cut costs.”
Unlike the work of UCI, the PEW study recognizes “changeable (dynamic) and unchangeable (static) risk factors related to criminal behavior.” The seven changeable risk factors are Antisocial Personality Patterns, Pro-criminal Attitudes, Social Supports for Crime, Substance Abuse, Poor Family/Marital Relationships, School/Work Failure, and Lack of Prosocial Recreational Activities.
Unchangeable or static risks, linked to recidivism are factors such as the age of an offender at his or her first arrest, the number of prior convictions, and the current commitment offense.
This distinction between “dynamic” and “static” risk assessment is important to note because:
In 2009, the CDCR adopted the California Static Risk Assessment (CSRA) as both an instrument to determine the rehabilitation needs of inmates, and as an indicator of those inmates who are at risk of returning to custody within three years of being released.
The CSRA’s assessment method consists of “four major steps. They are: 1) prior felony and misdemeanors; 2) the counts of prior convictions and age at release, and gender; 3) calculations from the second step used to create subscale scores; and 4) measures of predicting the accuracy of a subsequent conviction.”
Researchers at UCI developed the CSRA tool for CDCR based on a model created by the Washington State Institute for Public Policy (WSIPP). “The Washington tool was chosen for several reasons…most important was the tool that used static items only.”
CDCR’s data did not contain “dynamic” factors for every offender. As a result, factors like education and drug use could not be developed for risk assessment in a timely manner.
According to UCI, “CDCR felt that Washington state offenders would be similar enough to California offenders that replication would result in a valid tool…The project began in October, 2007 and produced the CSRA tool by the end of January 2008.”
Development of the CSRA began with a data sample of 103,603 California inmates released from the CDCR during the 2002-2003 fiscal year.
The California Code of Regulations, Title 15, Section 3768.1 reflects the CSRA. It was filed as an emergency on January 7, 2010 and says, in part: “The tool produces a risk number value that will predict the likelihood that an offender will incur a felony arrest within a three-year period after release to parole.”
“Risk groups were developed based on cut points for each of the scales, resulting in five different groups,” UCI said. The risk groups numbered from low to high are: 1, Low Risk; 2, Moderate Risk; 3, High Drug Risk; 4, High Property Risk; and 5, High Violent Risk.
“The three high risk groups have the highest overall recidivism rates,” UCI noted. “Whites and ‘others’ are more likely than Hispanics and blacks to be included in the Low Risk category.” However, in the High Risk category for drug or violent crimes, whites and ‘others’ were just as likely to re-offend as Hispanics and blacks.
The CSRA uses information from automated California Department of Justice criminal records (“rap sheets”) to calculate risk group assignment. The use of an automated tool has the advantage of being faster and more consistent than manually scoring risk assessment.
However, according to UCI, metric tests used to assess how well the CSRA tool predicts recidivism show mixed results. Predictions for a felony arrest within three years of release were better (or more accurate) than predictions for actual convictions for felonies. This makes sense because parolees are frequently arrested for parole violations that do not lead to criminal convictions, according to prison officials.
The PEW study has concluded that, “There is no one-size-fits-all risk assessment tool. Risk/needs assessments cannot predict an individual’s behavior with absolute precision. Inevitably there will be lower-risk offenders who reoffend and higher-risk offenders who do not reoffend.”
The UCI study stated, “The Center for Evidence-Based Corrections is collaborating with CDCR on a number of enhancements of the CSRA,” and promised, “resulting refinements of the CSRA will be documented in additional reports” in the future.