RRASOR

The Rapid Risk Assessment for Sexual Offense Recidivism, abbreviated as the RRASOR (pronounced like the cutting tool), is an actuarial scale designed to assess different levels of sexual recidivism risk for convicted sexual offenders. This scale was the first empirically validated actuarial instrument specifically designed for the assessment of sexual offense recidivism. As such, it served both to change the manner by which people conducted risk assessments and as the stepping stone for the development of second-generation actuarial instrumentation, though it is still being used because of its unique contribution to risk assessment technology. This overview describes the instrument’s development, summaries of tests of reliability and validity, how it has been used in furthering risk assessment knowledge, and the reason why it is still used despite the availability of newer instruments.

Development of the RRASOR

1. Karl Hanson was the sole developer of the RRASOR. He started (in 1997) with a set of seven empirically derived correlates to sexual recidivism selected from the findings of a then recently completed meta-analytic study. These seven risk factors were selected based both on their correlation with sexual recidivism and the presumed ease in finding the relevant information in prison records. A regression analysis using measures of these seven variables across six aggregated samples found that a subset of four of the seven variables accounted for nearly all the relevant variance of sexual recidivism. These four risk factors (prior sex offenses, offender age being younger than 25 years, ever having a male victim, and ever having an extrafamilial victim) were then combined to form an instrument called the RRASOR. A cross-validation test of the new scale with an independent seventh sample documented supportive evidence of the scale’s utility.

Scoring for 3 of the items is dichotomous, while scoring for the “prior sex offense” item involves four levels, with higher item scores always representing high recidivism risk. Each additional sign of risk is scored 1 point. The total scale score for the instrument is computed through simple addition across the 4 item scores. The final form of the instrument offers recidivism rates for each of six rank-ordered score levels (i.e., scale total scores of 0 through 5) for 5- and 10-year follow-up periods.

Although scores of 6 were possible, no one with that score was found among the nearly 2,600 subjects used to develop the instrument. Offenders with RRASOR scores of 6 have been found since by other researchers and risk evaluators, though that score is so rare that its associated degree of recidivism risk is still unknown.

The developmental study found the RRASOR to have an ROC (Receiver Operating Characteristic) ranging from .62 to .77 (relative to sexual recidivism) across the seven samples used to develop it. This range of ROCs is very much similar to the ROCs found from cross-validation tests of the instrument.

Reliability and Validity Findings

There have been between two and three dozen empirical tests involving the RRASOR to date. These have occurred using samples across at least eight countries (Belgium, Canada, England, the Netherlands, New Zealand, Sweden, the United States, and Wales). Interrater reliability tests have typically found the instrument’s reliability to fall in the .90 to .94 range— very high, though not surprising given the small number of items on the scale.

The RRASOR’s predictive accuracy (as measured by the ROC) is regularly found to be equal to that of other actuarial instruments when tested with individual samples. When viewed across a host of samples in a meta-analytic test by R. Karl Hanson and Kelly Morton-Bourgon, the RRASOR showed a lesser degree of relationship with sexual recidivism than did a second-generation actuarial scale called the STA-TIC-99 (described briefly below). The RRASOR’s predictive validity relative to sexual offenders’ non-sexual violent recidivism, however, has regularly been found to be poor, indicating the instrument’s effectiveness is specific to sexual recidivism.

Use of RRASOR in Furthering Risk Assessment Technology

As a first-generation instrument, the RRASOR served to demonstrate that actuarial procedures could be effective in assessing sexual re-offending risk. The development of a second-generation risk assessment instrument, the STATIC-99, was completely based on combining the 4 items from the RRASOR with the items from one other scale, for a total 10-item scale. Given that the STA-TIC-99 has become the most widely used actuarial instrument for assessing sexual recidivism risk, the RRASOR served well toward its own improvement.

The RRASOR has also served as a covariate in analyses to scale out the possible effect of different a priori recidivism risk levels (such as relative to the effect on sexual recidivism rates of aging, of the inability to suppress deviant responding on the penile plethysmograph, and of sex offender treatment). These occasions represented the first time researchers could statistically control for different risk levels in an empirically validated way.

Why the RRASOR Is Still Used

Research involving dimensions of risk (i.e., risk driven independently by offenders’ sexual deviance, their general antisociality, and other dimensions) has documented the utility of the RRASOR and fueled its continued use. The 3 nonage items from the RRASOR appear regularly as correlates to measures of sexual deviance and as not correlating strongly with measures of general criminality. Advocates of multidimensional risk assessment procedures often still use the RRASOR when assessing sexual recidivism risk to address risk driven by subjects’ sexual deviance, while assessing general antisociality-driven risk with other instruments.

References:

1. Doren, D. M. (2002). Evaluating sex offenders: A manual for civil commitments and beyond. Thousand Oaks, CA: Sage.
2. Doren, D. M. (2004). Stability of the interpretative risk percentages for the RRASOR and STATIC-99. Sexual Abuse: A Journal of Research and Treatment, 16(1), 25-36.
3. Doren, D. M. (2004). Toward a multidimensional model for sexual recidivism risk. Journal of Interpersonal Violence, 19(8), 835-856.
4. Hanson, R. K. (1997). The development of a brief actuarial risk scale for sexual offense recidivism. Retrieved June 26, 2015, fromhttps://www.publicsafety.gc.ca/cnt/rsrcs/pblctns/dvlpmnt-brf-ctrl/index-eng.aspx
5. Hanson, R. K., & Thornton, D. (2000). Improving risk assessments for sex offenders: A comparison of three actuarial scales. Law and Human Behavior, 24(1), 119-136.

Return to the overview of Violence Risk Assessment in Forensic Psychology.