SORA: The human cost of junk science

 

Of the nearly 40,000 persons on New York’s sex offender registry, 9,679 are displayed on its public website as Level 3, a warning that he or she presents the maximum risk of committing a sex crime of maximum seriousness. 14,087 persons are displayed as Level 2, meaning they’re moderately likely to commit a moderately serious sex crime.

With so many Frankensteins at large, it’s a wonder anyone dares leave the house.

How does the State know these people are so dangerous? Because they’ve each had a SORA hearing where a judge added up the numbers on a “Risk Assessment Instrument” form (RAI) and classified them accordingly.

Given the severe consequences attached to being publicly identified as someone who’s likely to commit a sex crime any minute, you’d expect courts to use a reliable, scientifically validated method of risk prediction. But the RAI’s only validation is the say-so of its authors: five employees of the Departments of Probation and Parole picked by the Governor.

The RAI is a mishmash of pseudo-science and law enforcement folklore. The scientific articles it cites, apart from being long outdated, don’t remotely support its sweeping generalizations about what factors predict recidivism. Although the RAI claims to be an actuarial instrument, its scoring system has no discernable methodology except for being heavily weighted towards a finding of maximum risk. Despite its scientific pretensions, the RAI is a caboodle of moral judgments based on the wholly unscientific assumption that “if he did it before, he’ll do it again.”

SORA was a reaction to the rape and murder of little Megan, an unattended 7-year old invited into the home of a man who turned out to be a sex offender. The statute was enacted on the assumption that sex offenders are a species of murderous pedophiles who, unlike other convicted offenders, remain incorrigible even after they’ve served their sentences. They were thought to be a kind of Pod People, described by one assemblyman urging passage of the bill, as “the most dangerous kind of criminals. . . . who look just like you and I, who live in homes just like we do, and from their outward activities, seem just like we are. But in fact, Madam Speaker, they are not.”

Another lawmaker argued, “We have passed bills all the time requiring labels on chemicals, products and toys to inform consumers of potential dangers. On a pack of cigarettes and bottles of alcohol there is a warning label.” And since sex offenders are “the human equivalent of toxic waste,” he reasoned, they should similarly be labeled as a warning to the public.

The notion of “sex offender” as an existential category is a myth, as is the belief that anyone convicted of sexual misconduct has a higher likelihood of reoffending than other offenders. Researchers such as R. Karl Hanson, who developed a widely used, independently validated risk assessment tool called the Static-99, have shown that recidivism prediction can’t be simplistically correlated to the moral seriousness of the offense. Nor is there any empirical basis for assuming that a reoffense would be the same as the past offense.

Defense lawyers have repeatedly presented uncontroverted expert evidence at SORA hearings showing that the RAI doesn’t use a valid methodology and isn’t based on the scientific consensus about what factors are predictive of recidivism. The courts have responded by keeping their heads firmly in the sand.  New York’s highest court has even declared that the RAI gives “presumptively correct” results.  How does the court know? Because the RAI says so.

The only judge who has publicly criticized the RAI is Justice Daniel Conviser in Manhattan. In a scathing 100-page opinion called People v McFarland, he found that the RAI, by relying on factors irrelevant to recidivism, is so arbitrary as to violate due process. Although he felt constrained by legal precedent to use the instrument, his analysis was the basis of a bill in the Legislature requiring the use of a scientifically validated instrument for SORA risk level classifications. It died in the Senate.

The Legislature and courts protest – perhaps too much – that SORA isn’t additional punishment for the past offense. That’s hard to explain to the thousands of persons who can’t “live in homes just like we do,” because of being branded as sex offenders.  But SORA is merely regulatory, we are told. No different in principle from labeling a pack of cigarettes.

In reality, there’s no instrument capable of predicting an individual’s future risk accurately enough to justify the lifelong humiliations and deprivations inflicted on SORA registrants. Even the Static-99 claims to be only moderately predictive. But if courts want to maintain the fiction that SORA is purely a public safety measure and not a vindictive targeting of an unpopular class, there’s no justification for using junk science.

Update: SORA courts reject validated instruments even as grounds for downward departure. See The Crime Report 1/25/2018

And for a detailed critique of the RAI, see Dr. Laurie Guidry’s article in “The Alliance.”

Guidry.Alliance