An interesting pattern was observed in a small sample of 50 risk assessments done by our state’s correctional system and I would like to hear your views on it.

Risks for incarcerated sex offenders are determined within the Utah prison system by trained, licensed sex offense therapists by combining Static-99R and Stable-2007 scores. We studied raw scores and resulting predicted risks for a small random sample of inmates. (For privacy reasons, we had only the raw scores [numbers] and the reported risks.) The scores are for people who have had at least some therapy.

It has been stated that these two risk assessment instruments are designed to give about the same average risks, so that combining the Stable score with the Static should improve accuracy, but on average it will neither raise nor lower risk values as compared to the Static risks alone. The data sample did not satisfy this condition. For 82% of the people on the lower end of the risk range, Stable scores appear to increase risk substantially. For most of the highest-risk people (five of the six), the Stable appears to reduce risk. And for four people (8%), risk stayed about the same. I made a graph, attached, of risk average vs static score. I am also sending the raw data in an Excel file.