Wednesday, August 16, 2006

More from the NASC conference - Sex Offender Research

This has been a busy year for sex offender legislation (several laws went into effect July 1) – much of it concerned with protection of our most vulnerable population, children. There are many issues raised by our drive to protect children and others from victimization.

Regarding state activity in this area, many states have passed a Megan's Law, broad-based community notification of the presence of dangerous sex offenders (information/links to sex offender registration sites are available here and here.) As John Tierney of the NY Times noted on June 10, 2006, there's also Ohio’s Nicole's Law, New Mexico Katie's Law, Minnesota Katie's Law….

A Georgia 2006 law set the bar high, perhaps impossibly so, attempting to keep convicted sex offenders from getting out of prison and re-settling near children. In addition to banning resettlement near schools, etc. the law also extends the no sex-offender zone to within 1,000 feet of one of the state’s 150,000 school bus stops. That would make most residential areas, and even entire counties, unavailable to sex offenders, as noted here.

According to, other states are cracking down on sex offenders by:
- Implementing Global Positioning System devices to monitor sex offenders (eight states with Georgia, Kansas, Virginia and Indiana laws taking effect July 1.)
- Nevada will publish the names, pictures and addresses of almost 2,000 higher-risk sex offenders online.
- Alabama will begin testing those charged with a sex crime for sexually transmitted diseases.
- Repeat violent sex offenders in Idaho can now be sentenced to life imprisonment.

What do we know about sex offenders, and do these laws represent a step forward in safety and justice? The topic of sex offender research was a focus of the 2006 NASC conference, and brought together some notable researchers in the field (R. Karl Hanson, Kim English, and Karol Lucken, moderator: Meredith Farrar-Owens).

First, some of what we know about sex offenders. Over five percent of sex offenders (5.3 %) released from prison are re-arrested for a sex offense within three years, but are more likely to commit a new sex offense than other types of offenders released from prison (1.3 %) as reported in a recent BJS study. However, such studies are plagued by an under-reported crime (see here for a recent review of reporting). Further, many registration laws require lifetime reporting, so a longer follow-up period would be of interest.

R. Karl Hanson (Corrections Research, Public Safety Canada) addressed the NASC conference and discussed his longer term studies. He estimates that for all sexual offenders (specifically following Canadian sex offenders) the re-conviction rate rises from single digits to 24 % when tracked over 15 years. More importantly, not all sex crimes have the same rate of re-offending. Using the same measure of recidivism (reconviction within a 15 year period), child molesters of boys are far more likely to re-offend (35 %), followed by rapists (24 %), child molesters of girls (16 %) and incest offenders (13 %). In short, while much recent legislation paints with a broad brush, a more targeted approach to certain kinds of sex offenders is possible and perhaps advisable.

Speaking of taking a more targeted approach, Dr. Hanson has developed and studied risk scales (The Rapid Risk Assessment for Sex Offense Recidivism or RRASOR, and STATIC-99/STATIC-2002) to estimate the likelihood of recidivism for sex offenders fitting a particular profile. This approach is known as actuarial risk assessment because it predicts an individual’s membership in a subgroup that is correlated with future offending, and cannot predict a particular person’s likelihood of re-offending (just as insurance companies classify automobile drivers for purposes of setting auto insurance rates, not purporting to predict a particular driver’s chances of an accident). The RRASOR scale includes only four factors (male victims, unrelated victims, prior sex offenses, and under 25 years of age). The STATIC-99 scale has 10 factors, and can successfully distinguish low risk offenders, whose average observed recidivism rate is 10 % when tracked over 15 years, from medium high (37 %) and high (52 %) risk offenders over the same period. His and other actuarial instruments, while not perfect predictors, have proven far more accurate than clinical judgment.

Bottom line: More targeted approaches to combating repeat sex offending are possible, and these approaches are being used selectively in sentencing (for example), parole and corrections settings, but not necessarily considered in broad-brush legislation.

Further thoughts: There are many other issues involved in sweeping legislation imposing sex offender registration and other restrictions. Among the most important topics is the problem of low reporting of sex crimes (all the figures given above are an undercount of actual crimes), the problem of non-stranger sex crimes (the majority of all sex crimes but presumably little affected by registration laws since the victims already know their offender), the problem of driving sex offenders underground through visible campaigns that target them, etc. Many of these problems are not resolved or even addressed by sex offender registration, but there are other approaches.

No comments: