|By Ann Coppola, News Reporter|
As gas prices rise to seemingly unlimited new heights and stocks plunge to cringe-worthy lows, the one word that seems to be on the lips of anyone watching the United States economy is “risk.” According to recent comments from the Federal Reserve, the U.S. “can reduce risk in the financial system” by following a four-point plan that predictably requires an economics degree to understand. One thing, though, is clear: risk assessment is a vital tool during volatile times. The calculation of risk affects almost every aspect of our lives, whether it’s a financial or personal matter, or even one of public safety.
“We are in the risk management business,” says Calvin Johnson, director of research and evaluation for the District of Columbia’s Court Services and Offenders Supervision Agency. CSOSA is using statistical data analysis to take its supervision of more than 15,000 parolees, supervised releases and probationers to the next level.
“We’re able to employ advanced analytic models like those used in the insurance business or finance industry,” Johnson explains. “The difference is we’re looking at things like the probability of an offender being rearrested for a violent crime, the probability they will test positive for drugs three or more times over the next month, or the probability their parole is going to be revoked.”
CSOSA’s current system, SMART-STAT, gathers data from local, state and federal data sources. Analysts study offenders’ backgrounds and performance under supervision to predict what intervention strategies would be most helpful for certain offenders. SMART-STAT utilizes data from court hearings, dispositions and sentences, nightly police booking reports, prison release schedules, disciplinary actions, drug testing, treatment participation, and even DNA.
“We supervise 15,000 offenders daily that present all types of social characteristics, offenders from the mental health community, offenders with little to no family support,” Johnson adds. “These kinds of factors in many cases place them in a certain risk category. Just like an insurance agent, we manage the risk that these offenders will re-offend.”
SMART-STAT is modeled after COMP-STAT, the New York City Police Department’s organizational management tool that utilizes various software and geographic information systems to map crime and create crime intervention strategies. COMP-STAT is widely credited for reducing New York City’s crime rate in the 1990s. Now, CSOSA is using it as a guide to reduce the District’s recidivism rate.
“We’re looking for an 80 percent success rate, which means a 20 percent recidivism rate for our offenders,” says CSOSA IT Associate Director Bill Kirkendale. “Right now, our success rate is around 70 percent. Employment, housing, and staying off drugs are the key things for an offender’s success.”
Obtaining that success rate basically translates to knowing more about offenders’ lives than anybody else.
“What we do is very similar to the consumer marketplace,” Kirkendale explains. “When you check out with your grocery store card, they know what to put on sale, what to send you in the mail, what to put on the shelf, what to advertise on TV. They know everything about you. We know everything about these offenders, but we’re looking at their propensity to succeed or fail.”
That knowledge puts CSOSA in the position to match certain categories of offenders with the treatment regimen or supervision strategy that will most improve their chances for success.
“We have really rich data,” says Kirkendale. “Offenders have particular dynamics, such as their age, criminal history, history of drug use, and so forth. We look for trends and patterns to better refine supervision tactics, strategies, and programs. Is the offender employed? If they go into a 28-day facility, do they relapse? How long until folks with this history versus that history relapse?”
SMART-STAT does more than help CSOSA put offenders on the right path in the community. The data keeps the agency itself on track by creating a “structure of accountability.”
“We can see if we’re not moving at a rate of pace we think we should be,” says Johnson. “We can drill down through the agency to the branch, to the team, to the officer. We can go even further to specific offenders who may be creating the sluggish movement.”
“It could be because one offender is not responding to the intervention services we’re offering them,” he adds. “And so because we can drill all the way down, we can start holding the entire system accountable, from the offender to the officer to supervisors to the branch chief to the deputy.”
The self-accountability has paid off. CSOSA recently was honored at the SAS Institute, Inc. Executive Conference for Government for its innovative use of data. SAS software is utilized by the banking, retail, aerospace, and insurance industries to name a few. One of the breakthrough data sources for CSOSA has been geographic information on its offenders who are tracked using global positioning systems (GPS).
“We can do analysis that allows us to see how the GPS monitoring is affecting the offender’s performance to conditions of their parole or probation,” Kirkendale says. “We can find out if it is helping them or hurting them. For some it might prove to be a hampering effect. For example, it’s very tough to get a job with a bracelet around your leg.”
For supervising offenders in the community, Kirkendale says the three biggest predictors of success or failure are employment, housing, and drug test performance.
“No one has better data than we do,” he says, “and we’re monitoring all of that business intelligence to reduce violence and improve public safety.”
Like any industry in the intelligence or future forecasting business, CSOSA needs to calculate and understand its risks in order to meet its goals. CSOSA will continue honing its data warehouse, and maybe one day the agency’s entire performance pie chart will have only one category, labeled “success.”
Watch video on CSOSA’s data management
Forbes looks at SAS
IN CASE YOU MISSED IT