How are the risk scores displayed and what do they mean?
This screen shot shows the comparative risk score for reconviction of a felony. This estimates the percent of offenders at a lower risk for reconviction of a felony within 3 years of release from prison or imposition of probation based upon the risk scores of all felony offenders that have been released from prison or sentenced to probation in the last 5 years. The comparative risk shown in the screen shot means that the offender's actual risk score for reconviction of a felony is higher than 85% of felony offenders that have been released from prison or sentenced to probation in the last 5 years.
This screen shot shows the actual risk score for reconviction of a felony and the community corrections classification based on the actual risk score. This is viewed by clicking on "View community corrections classification for this person". The Public Safety Checklist also displays risk scores for a new person arrest and a new property arrest. The actual risk for reconviction of a felony estimates the likelihood the offender will be convicted of a new felony within 3 years of release from prison or imposition of probation. This estimate is based upon the recidivism patterns of 345,000 offenders that have been released from prison or sentenced to probation since 1980. The actual risk shown in the screen shot means that the offenders' likelihood of new felony conviction within 3 years of release from prison or imposition of probation is 39%. The community corrections classification is shown as low, medium or high based on the actual risk score and cut-off levels determined by OACCD (Oregon Association of Community Corrections Directors).
What is the Public Safety Checklist?
The Public Safety Checklist is an actuarial risk assessment tool that uses offender characteristics to predict recidivism.
What does the Public Safety Checklist do?
The Public Safety Checklist provides a quick, objective, validated assessment of the probability an offender will be re-convicted of a felony or re-arrested for a person or property offense based on the offender's following characteristics:
- Current age
- Gender
- Age at first arrest
- Severity of current crime
- Multiple custody cycles
- Total number of statutory arrests
- Total number of property arrests
- Total number of person arrests
- Total number of statutory arrests in last 5 years
- Total number of property arrests in last 5 years
- Total number of person arrests in last 5 years
- Prior incarceration
- Prior theft conviction
- Prior revocation
- Previous sentence type (probation or incarceration)
How was the Public Safety Checklist developed?
The Oregon Department of Corrections and Oregon Criminal Justice Commission analyzed actual re-arrest and re-conviction data from 55,000 offenders in Oregon from 2000 to 2005, and used logistic regression analysis to determine which offender demographic and criminal history factors are predictive of a felon being reconvicted or rearrested. This model of risk assessment was then applied to all 350,000 offenders sentenced to probation or released from prison from 1980 through present to make sure the risk tool accurately and validly categorized these offenders' risk to be re-arrested or re-convicted.
What measures of recidivism are used in the Public Safety Checklist?
Reconviction for any felony within 3 years of release from prison or being sentenced to probation, re-arrest for any person felony or class A person misdemeanor (32% are misdemeanors, not including DUII and misdemeanor Assault 4), and re-arrest for any property crime listed in ORS 137.717(2) (39% are misdemeanors including Theft 2).
Is the ethnicity of the offender included as one of the factors that predict the offender's risk score?
No. An offender's ethnicity is not included in the measurement of an offender's risk to be re-arrested or re-convicted.
How is a risk score produced?
An algorithm is used to weigh the offender's predictive factors and give a "composite score" for an offender with these characteristics. This is the same procedure that is used in life insurance actuarial tools or heart attack risk factors (although the factors are, of course, different).
What are the limitations of the Public Safety Checklist?
The Public Safety Checklist uses Oregon data and does not include out-of-state, federal or juvenile arrests and convictions. The Public Safety Checklist does not account for dynamic factors (factors that can change) that could possibly predict recidivism. These include factors such as the offender's willingness to change or current participation in a drug treatment program. The Public Safety Checklist database includes offenders that have been convicted of a felony or serious misdemeanor. It also includes offenders that have a SID number and arrest in LEDS.
Is this type of tool currently used elsewhere?
Yes. The Virginia and Missouri sentencing guidelines currently incorporate the risk assessment into their recommended sentence. The same actuarial assessment is commonly used in the medical world when identifying "risk factors" for a heart attack or in the insurance world in setting premium rates for drivers based upon demographic information and past driving record.
How accurate is the Public Safety Checklist?
The area under the curve is a statistical measure of a model's fit or predictability. An area under the curve of 0 indicates no predictability at all and 1 is perfect predictability. The area under the curve for the public safety checklist is over 0.70. This is comparable accuracy to risk assessment tools used in Washington, Virginia, and Missouri.
Why use a Public Safety Checklist?
Many studies have shown that a professional using an actuarial tool does a better job than clinical judgment alone in assessing the likelihood of recidivism. The Public Safety Checklist provides increased objectivity, is quick and easy to use, and no training is required. It provides better information which in turn leads to better decisions.
Does the Public Safety Checklist tell the judge what the sentence should be?
No. The Public Safety Checklist gives scientific information about the likelihood of recidivism for an offender based upon analysis of thousands of previous Oregon offenders. That information is not dispositive of what the sentence should be in an individual sentencing proceeding. There are four principles that guide the punishment of crime articulated in Oregon's Constitution: protection of society, personal responsibility, accountability for one's actions and reformation. The Public Safety Checklist informs a judge about the two principles that relate to increasing public safety after the sentence is imposed: protection of society and reformation. The judge will always need to look at all aspects of the criminal conduct, its impact on the victim, and any steps the offender has taken at taking responsibility before imposing a sentence that serves the principles of accountability for one's actions and personal responsibility. Assessing the likelihood of improving public safety through treatment in or out of custody depends in part upon the offender's needs, which are addressed by other instruments but not by the risk assessment instrument.
Does the Public Safety Checklist eliminate the need for professional judgment?
No. It informs but does not replace professional judgment. However, many studies have shown that a professional using an actuarial tool does a better job than clinical judgment alone in assessing the likelihood of recidivism. So, it should not be ignored, but should be overridden in situations where other factors than those captured make this offender different or unique from most offenders.
How well does it differentiate between high, medium, and low risk felons?
The Public Safety Checklist defines low risk felons as those in the lowest 20% for risk to recidivate, medium as the next highest 60%, and high as the highest 20%. For offenders released from prison or sentenced to probation in the last 5 years, the low risk offenders had a recidivism rate of 10%. Recidivism is a new felony conviction within three years. The medium risk offenders had a recidivism rate of 28% and high risk offenders had a rate of 55%.
