Sunday, April 7, 2013

Risk screening worthless with juvenile sex offenders, study finds

Boys labeled as 'sexually violent predators' not more dangerous

Juveniles tagged for preventive detention due to their supposedly higher level of sexual violence risk are no more likely to sexually reoffend than adolescents who are not so branded, a new study has found.

Only about 12 percent of youths who were targeted for civil commitment as sexually violent predators (SVP's) but then freed went on to commit a new sex offense. That compares with about 17 percent of youths screened out as lower risk and tracked over the same five-year follow-up period.

Although the two groups had essentially similar rates of sexual and violent reoffending, overall criminal reoffending was almost twice as high among the youths who were NOT petitioned for civil commitment (66 percent versus 35 percent), further calling into question the judgment of the forensic evaluators.

Because of the youths' overall low rates of sexual recidivism, civil detention has no measurable impact on rates of sexual violence by youthful offenders, asserted study author Michael Caldwell, a psychology professor at the University of Wisconsin and an expert on juvenile sex offending.

The study, just published in the journal Sexual Abuse, is one in a growing corpus pointing to flaws in clinical prediction of risk.

It tracked about 200 juvenile delinquents eligible for civil commitment as Sexually Violent Persons (SVP's). The state where the study was conducted was not specified; at least eight of the 20 U.S. states with SVP laws permit civil detention of juveniles, and all allow commitment of adults based on offenses committed as a juvenile.

As they approached the end of their confinement period, the incarcerated juveniles underwent a two-stage screening process. In the first phase, one of a pool of psychologists at the institution evaluated them to determine whether they had a mental disorder that made them "likely" to commit a future act of sexual violence. Just over one in every four boys was found to meet this criterion, thereby triggering a prosecutorial petition for civil commitment.

After the initial probable cause hearing but before the final civil commitment hearing, an evaluator from a different pool of psychologists conducted a second risk assessment. These  psychologists were also employed by the institution but were independent of the treatment team. Astonishingly, the second set of psychologists disagreed with the first in more than nine out of ten cases, screening out 50 of the remaining 54 youths. (Only four youths were civilly committed, and a judge overturned one of these commitments, so ultimately all but three boys from the initial group of 198 could be tracked in the community to see whether or not they actually reoffended.)

Evaluators typically did not rely on actuarial risk scales to reach their opinions, Caldwell noted, and their methods remained something of a mystery. Youths were more likely to be tagged for civil detention at the first stage if they were white, had multiple male victims, and had engaged in multiple instances of sexual misconduct in custody, Caldwell found.

However, no matter what method they used or which factors they considered, the psychologists likely would have had little success in predicting which youths would reoffend. Even "the most carefully developed and thoroughly studied" methods for predicting juvenile recidivism have shown very limited accuracy, Caldwell pointed out. This is mainly due to a combination of youths' rapid social maturation and their very low base rates of recidivism; it is quite hard to successfully predict a rare event.

Indeed, a recent meta-analysis revealed that none of the six most well-known and best-researched instruments for appraising risk among juvenile sex offenders showed consistently accurate results. Studies that did find significant predictive validity for an instrument were typically conducted by that instrument's authors rather than independent researchers, raising questions about their objectivity.

"Juveniles are still developing their personality, cognitions, and moral judgment, processes that reflect considerable plasticity," noted lead author Inge Hempel, a psychology graduate student in the Netherlands, and her colleagues. "There are still many possible developmental pathways, and no one knows what causes persistent sexual offending."

Caldwell agrees with Hempel and her colleagues that experts' inability to accurately predict which juveniles will commit future sex crimes calls into question the ethics of civil commitment.

"From the perspective of public policy, these results raise questions about whether SVP commitment laws, as written, should apply to juveniles adjudicated for sexual offenses," he wrote. "If SVP laws could be reliably applied to high risk juvenile offenders, the benefit of preventing a lifetime of potential victims makes for a compelling case. However, the task of identifying the small subgroup of juveniles adjudicated for sexual offenses who are likely to persist in sexual violence into adulthood is at least extremely difficult, and may be technically infeasible."

* * * * *

The articles are:

Michael Caldwell: Accuracy of Sexually Violent Person Assessments of Juveniles Adjudicated for Sexual Offenses, Sexual Abuse: A Journal of Research and Treatment. Request it from the author HERE.

Inge Hempel, Nicole Buck, Maaike Cima and Hjalmar van Marle: Review of Risk Assessment Instruments for Juvenile Sex Offenders: What is Next? International Journal of Offender Therapy and Comparative Criminology. Request it from the first author HERE.

1 comment:

  1. This is not surprising in the least. The US Sentencing Commission found that current guidelines for risk assessment were based on very little in the way of science and were about as useful as a coin flip.

    This "industry" has no interests in assessing actual risk, only putting people in jail and then civil commitment. It's time to let mental health professionals way in on how to treat mental health problems!

    ReplyDelete

 
Real Time Web Analytics