College of Liberal Arts and Sciences at Iowa State University

Keeping tabs

Douglas Epperson has developed an actuarial tool to help predict recidivism of sexual offenders.
The headlines screamed the news. A recently released sexual offender had sexually assaulted another victim. Many times, the neighborhood was not aware that a sexual offender was even in their midst.

Reacting to public pressure, every state passed sexual offender registration and community notification statues. In addition 13 states have also passed and implemented civil commitment or sexual predator statues that permit the state to commit the most dangerous sex offenders to forensic hospitals instead of releasing them upon completion of their sentence. However the constitutionally of these statues largely rests on the states' ability to accurately identify those individuals at the greatest risk of reoffending upon release.

But there was no good way to accurately predict that. Douglas Epperson, associate professor of psychology, and colleagues with the Minnesota Department of Corrections have changed that.

"Although many released sex offenders are never arrested for another sex offense, there was a critical need for an empirically-based prediction tool, one that would identify who is less likely or more likely to reoffend," he said.

Epperson's research has led to such a tool --one that predicts recidivism of sexual offenders. That tool, Minnesota Sex Offender Screening Tool Revised (MnSOST-R), has led to important improvements in the implementation of the civil commitment/sexual predator and community notification statues since it was presented at a national conference just over a year ago.

"The field was so hungry and desperate for actuarial tools that states started to use it almost immediately. Before we knew it, it was being used all over the country," Epperson said.

All 13 states with civil commitment statues are utilizing the MnSOST-R, with nine using it extensively. It is also used in countless other states in conjuntion with community registration/notification statues. In addition, some psychologists in England, Ireland and Canada are using the tool, and European psychologists have asked to translate the MnSOST-R into their native language and validate it in their country.

The MnSOST-R is very similar to tools that insurance companies use when determining the level of fees an applicant will pay for life insurance. Instead of lifestyle questions, these focus on areas relevant to sexual recidivism.

"It matches individuals with groups of people for whom there are known outcomes," Epperson said.

Epperson's research concluded that 12 historic/static variables were predictive of sexual recidivism. By the same token, previous offenders are credited for solid behavior.

There are currently three other actuarial tools available to states in predicting sexual recidivism. But there is a major difference between those and the MnSOST-R.

"All of the tools are somewhat similar," Epperson said. "We each key in on the same type of static variables. What sets the MnSOST-R apart is the dynamic variables that we have included."

The dynamic variables focus on the offenders' history while they were incarcerated, including completion of chemical dependency and sexual offender treatments and their disciplinary behavior in prison.

MnSOST-R scores give states risk levels for sexual offenders, ranging from low to high probabilities, enabling states to better implement commitment and notification statues.

"More importantly," Epperson said, "we view the MnSOST-R as a prevention and resource management tool. States simply don't have the resources to commit all dangerous sex offenders or even to release them all under intensive community supervision. It becomes critical that the available resources are focused on the most dangerous sex offenders.

"Our research indicates that the MnSOST-R is very useful in indentifying those people."