Professor develops method to predict repeat sex offenders

Megan Mcgurk

Identifying sex offenders who may re-offend could get easier, thanks to the work of one ISU professor.

Doug Epperson, associate professor of psychology, is using a statistical measuring tool with the ability to help predict which sex offenders are likely to commit the same sort of crime once they get out of prison.

“The research has been so productive that it’s being used in many states — so many I’ve lost count,” he said.

Epperson said he has been working on this research for nine years and is getting good results.

“It’s nice when you can do something that proves to be socially useful,” he said.

Epperson’s research looks at many variables about the offender while making the prediction, including personal stability, demographic data, criminal history and the offense itself.

Craig Anderson, professor and chairman of psychology, said Epperson’s measuring tool is proving to be effective.

“It’s a simple means of predicting with pretty good accuracy,” he said.

Anderson also said this instrument is more effective than clinical predictions. “Clinical judgment is not always good,” he said.

As explanation, Epperson said clinical predictions don’t often weigh important factors, including personality stability and criminal history.

“Clinical predictions can’t articulate how they came to this conclusion,” he said.

Epperson said his predictive model has originated from public uproar about the number of sex offenders who re-offend.

“Legislatures wrestled with this in response to the outcry,” he said.

Because of this, legislators across the country saw a need for a predictive model such as the one Epperson developed, he said.

“Legislature saw the need for objective ways to identify who was at high risk of re-offending,” he said.

The research has been cross-validated with a second sample to prove validity, and Epperson said he’s been getting better outcomes than he hoped for.

“There’s still a very good correlation,” he said.

Epperson said he will be wrapping up this project and moving onto similar research, in which he will use predictive models dealing with juvenile sex offenders.