Criminal Risk Assessments in Colorado
Judicial use of algorithms in sentencing and bail-setting has some downsidesBy Judy Malmon, J.D. | Reviewed by Canaan Suitt, J.D. | Last updated on May 1, 2023 Featuring practical insights from contributing attorney Thomas R. Ward
Use these links to jump to different sections:
- How is the Report Used in Sentencing Decisions?
- Attorney Concerns Over Transparency
- Biases in the Assessments
- Sentencing Should Be Individualized
More and more courts are utilizing risk assessment tools to inform decision-making about bail and sentencing. These risk assessment instruments incorporate a range of data based on questions answered regarding each defendant.
In Colorado, when someone has been charged with a crime, the court will generate a Pre-Sentence Investigation (PSI) report, listing risk factor information such as:
- Current charges;
- Prior convictions or criminal history;
- Work history;
- Marital status;
- Financial status;
- Educational background;
- Substance abuse history; and
- Sentencing recommendations.
It will also include an LSI (Level of Service Inventory) Report—a document that ranks the defendant’s risk of recidivism within one year as low, medium or high.
How is the Report Used in Sentencing Decisions?
“It’s difficult to tell how much this risk score is used in sentencing decisions,” says Tom Ward, a Denver criminal defense attorney. “I don’t hear judges talking about it at hearings.”
Not talking about it doesn’t mean the algorithmic assessment isn’t a factor in a judge’s decision—merely that it adds another layer to the already unknowable determination process.
Attorney Concerns Over Transparency
“Transparency is a huge concern for lawyers who are representing defendants. To the extent that judges are taking these risk assessment scores into consideration in determining sentences, you want to be able to explain why a score was high,” Ward adds.
“If I don’t have any access to the raw data and I don’t know how these instruments work, I can’t go to a sentencing hearing on behalf of my client and challenge the weight put on one factor, like maybe lack of education. There can be a bigger story surrounding that particular defendant’s reason for why he didn’t go to school—maybe he had to drop out to work to support his family. But these software instruments are just taking the data and crunching the numbers and spitting out a result. It’s very difficult to identify what we should be clarifying.”
Biases in the Assessments
In addition to this murkiness, the assessment instruments appear to be incorporating the kinds of information that judicial rulings have explicitly moved away from considering—factors like race, gender and socioeconomic status.
“If a judge were to say in a sentencing hearing, ‘This defendant comes from a poor socioeconomic background and is not highly educated, and therefore a harsher sentence is warranted,’ you would think that judge was off their rocker,” Ward says. “With these instruments, poor socioeconomic background and lack of education can go into determining that this person has a higher risk to reoffend.”
Moreover, the temptation to regard the seemingly objective assessment as scientifically accurate may be difficult to resist. “It’s a little bit like the danger of expert witness testimony to a jury,” says Ward. “Expert testimony can be very helpful, but where juries start to focus on it as a sort of talismanic pronouncement that’s absolutely infallible, it can be dangerous. The danger here is if judges assume the assessment score must be right.”
Sentencing Should Be Individualized
The ultimate effect can be to limit the range of information considered in each individual’s case, particularly in light of heavy caseloads that are the norm in most courtrooms today, as well as pressure on judges to get it right. A high risk score in the record can look bad for a judge who gave someone a shot who then reoffended.
“Sentencing is supposed to be this very individualized exercise,” says Ward. “We look at all of the things that make this person unique, all of the circumstances that led to this crime being committed. If we’re dropping people into one of these baskets—high, medium or low risk—determined from these instruments, you’re not going to be doing a good job recognizing other important factors in sentencing.”
It’s worth noting that the risk assessment tools currently in use were not originally intended to be so heavily leveraged in bail and sentencing decisions, but were in fact designed to help identify needed supports for those released to the community. For this purpose, Ward acknowledges that the tools can be useful: “If you use the algorithm to look at this person’s needs as far as interventions, that makes some sense.”
If you are facing prosecution for any crime in the state of Colorado or any Colorado counties, don’t try to overcome a risk score on your own. Be sure to work with a reputable and experienced criminal defense attorney. For additional information on this area of law, see our overview of criminal defense.
Additional Criminal Defense articles
Find top lawyers with confidence
The Super Lawyers patented selection process is peer influenced and research driven, selecting the top 5% of attorneys to the Super Lawyers lists each year. We know lawyers and make it easy to connect with them.Find a lawyer near you