Developing Non-Arbitrary Metrics for Risk Communication: Percentile Ranks for the Static-99/R and Static-2002/R Sexual Offender Risk Tools

Abstract

The aim of this article was to advance risk communication by examining percentile ranks as a non-arbitrary metric for quantifying risk. Although percentile ranks have a simple meaning, their calculation is complicated by ties (i.e., more than one offender having the same score). The strengths and weaknesses of percentile ranks are discussed, as are the options for calculating and presenting them in applied risk communication. As a demonstration, percentile ranks for Canadian sexual offenders were computed for the most popular sexual offender risk assessment tools (Static-99, Static-99R, Static-2002 and Static-2002R). The distribution of Static-99 scores was highly stable in international comparisons of sexual offenders from Canada (1990 to 2005; n = 2,011), Sweden (1993 to 1997; n = 1,278) and California (2008 to 2010; n = 37,600). The major limitation of percentile ranks is that they measure the “unusualness” of scores in a particular reference group, and may not correspond to other indicators of relative or absolute risk. Consequently, we recommend that evaluators presenting percentile ranks should consistently provide recidivism base rate information so that decision makers do not confuse the rarity of a score with estimates of absolute recidivism risk.

Keywords:

• risk assessment
• percentiles
• relative risk
• sexual offenders
• Static-99

Acknowledgments

The views expressed are those of the authors and not necessarily those of Public Safety Canada or the Wisconsin Department of Health Services.

We would like to thank the following researchers for sharing their data with us: Jacques Bigras, Sasha Boer, Andy Haag, Niklas Långström, Janet Neeley, and Jean Proulx.

Notes

Static-2002 originally had 13 items. To increase clarity in the coding manual (Phenix et al., 2009), one item was divided into two separate items. However, this does not change the total score (the summed score of the two separate items in the new coding rules is identical to the score for the old item).

We would like to thank Kelly Morton-Bourgon for locating these data.