Log in / Register
Home arrow Philosophy arrow Sex Crimes and Sex Offenders: Research and Realities

Actuarial Assessments

Actuarial models are relied upon more often than clinical judgments alone because actuarial models are based on empirically derived scales that have been validated (Monahan, 1996). They also include a cut-off score for various risk levels. Actuarial models have outperformed clinical judgments by accurately predicting outcomes (i.e., sexual recidivism) in more cases than clinical judgment (Goggin, 1994).

Structured Anchored Clinical Judgment (SACJ-Min)

The Structured Anchored Clinical Judgment (SACJ-Min) model was relied upon for the development of several other risk-assessment scales (Craig et al., 2004). The first two stages make up the SACJ-Min score, and are based on a person's conviction history and aggravating factors (stranger victim, male victim, never married, convictions for non-contact sex offenses, substance abuse, placement in residential care as a child, deviant sexual arousal and psychopathy). There can be a third stage, which includes an assessment of current behavior, including response to a treatment program (Grubin, 1998).

Rapid Risk Assessment of Sex Offender Recidivism (RRASORj

Those who have received training and have experience in actuarial risk scales can administer the Rapid Risk Assessment of Sex Offender Recidivism (RRASOR). Hanson found that general-offender recidivism scales do not work well in predicting sexual recidivism among sex offenders. Based on a large sample of approximately 2,500 sex offenders four items were highly correlated with sex-offender recidivism and became the RRASOR. The scores range from zero to six, with a higher score indicating high risk. This is one of the most widely used actuarial tools and is effective at accurately predicting sexual recidivism. Given that the items are static and there are only four of them, it is possible to score an offender with only the person's administrative records. The items that make up the RRASOR are also used in the Static-99R (Harris, Phenix, Hanson, & Thornton, 2003).

Static-99R and Static-2002

Those who have completed a Static-99R training session can use the Static-99R. It is the most widely used sex-offender assessment scale in the world (National Institute of Corrections, n.d.). The questions assess the following: age at release, history of living with a romantic partner for at least two years, non-sexual violence committed during the index offense, convictions for non-sexual violence prior to index offense, number of prior sentencing dates, convictions for non-contact sex offenses, unrelated victims, stranger victims, and male victims.

The questions, although quite simple on their face, can be complicated to accurately score. The authors provide an 80-page manual explaining how to score each item. For example, the second item may be difficult to score, if the offender has lived with someone for two years, but it was a same-sex relationship. The authors clarify situations such as this, by noting that same-sex relationships should be counted in this category. Other situations are also clarified, such as the offender being incarcerated for much of his adult life. This is scored as not having lived with a lover for at least two years. Also, if the offender lived with a lover for two years, but with different persons, this item is scored as not living with a lover for two consecutive years. Clarifications are made for each item (Harris et al., 2003).

Each of the items is based on empirical findings showing that it is associated with risk of recidivism. For example, the first item is based on a correlation between offender age and the likelihood of recidivism. The younger the offender is at the time of the risk evaluation, the more likely he is to recidivate. It should be noted that the Static-99R is intended for male adults at least 18 years old (Beck & Harrison, 2008).

Each answer is given a score. The scores are then added for a possible range from 0 to 12. Based on their score, they are considered either a low, low-moderate, moderate-high, or high risk for sexual recidivism (Harris et al., 2003).

The Static-2002 is based on similar items to the Static-99R and was developed to increase conceptual cohesion and clarification (Hanson & Thornton, 2003). A weighted score is calculated for each of the five sets of questions. Those scores are then totaled to yield one of five possible risk levels: low, low-moderate, moderate, moderate-high and high. This assessment predicts sexual recidivism slightly better than the Static-99R (Phenix, Doren, Helmus, Hanson, & Thornton, 2008), yet both are still widely used today.

Sex Offender Risk Appraisal Guide (SORAG) and Violence Risk Appraisal Guide (VRAG)

The Sex Offender Risk Appraisal Guide (SORAG) and Violence Risk Appraisal Guide (VRAG) were developed by Quinsey, Harris, Rice, and Cormier (2006) and based on empirical research. When an offender commits a sexual offense, both should be used. For those without a sex crime, only the VRAG should be used. The purpose of the SORAG is to predict violent recidivism among sex offenders. Violent recidivism includes sexual recidivism. The VRAG includes 12 items and includes many of the items from the SORAG, but also asks about female victims and harm to victims. It also includes a childhood and adolescent component, which assesses childhood problems, including alcohol problems and conduct-disorder symptoms. The test items are scored, totaled, and given a risk level of either low, medium, or high.

Risk Matrix 2000/S (RM2000/S)

The Risk Matrix 2000/S (RM2000/S) was developed by Thornton and is a sexual offense prediction scale. It was developed in conjunction with two other scales: RM2000/C, a prediction scale for non-sex offenses, and RM2000/V, which is a combination of the other two scales, predicting sexual and non-sexual violence. The RM2000/S was developed from an existing scale, the Structured Anchored Clinical Judgment (SACJ). Additional factors from the literature were added to the scale and tested for predictive accuracy on a sample of offenders, which led to the development of the SACJ-Min in 2000. Based on substantive findings in the literature and for the purpose of seeking a parsimonious model, the SACJ was modified to include three variables that were combined to identify four levels of risk. Aggravating factors in the SACJ were included in the model, which includes male victim, stranger victim, non-contact sex crime, and never married. All of these factors have consistently been found to correlate with subsequent sex crimes.

The RM2000/S scale assesses age, number of appearances for (any) crimes, number of appearances for sex crimes, and aggravating factors (male victim, stranger victim, non-contact sex crime, and never married). The scores are totaled according to the guidelines established by the author of the scale and yield low, medium, high, and very high risk levels (Thornton, 2007). This scale was first validated in 2003 (Thornton et al., 2003).

Minnesota Sex Offender Screening Tool-3.1 (MnSOST-3.1)

The work for the Minnesota Sex Offender Screening Tool (MnSOST) began over two decades ago, and the tool has undergone several revisions since then, resulting in the latest version, Minnesota Sex Offender Screening Tool-3.1 (MnSOST-3.1) (Duwe & Freske, 2012). It is also one of the most widely used assessments. It includes nine items: (1) sentences for predatory offenses, (2) number of felony sentences, (3) number of harassment sentences (harassment, stalking, violate protective order, violate no-contact order, violate restraining order), (4) number of disorderly conduct sentences in previous three years, (5) age at release, (6) released without correctional supervision, (7) completion of both sex-offender and substance-abuse treatment, (8) sentences involving male victims, and (9) committed a sex/sex-related charge in public (Minnesota Department of Corrections, 2012). The totaled score results in one of three risk levels (Level 1, 2, and 3).

Found a mistake? Please highlight the word and press Shift + Enter  
< Prev   CONTENTS   Next >
Business & Finance
Computer Science
Language & Literature
Political science