How do psychologists assess the risk that a sex offender will strike again?
Through a combination of clinical judgment and statistics. A forensic psychologist interviews the convict, speaks with his family, and reviews police reports and prison records. The psychologist combines this research with actuarial assessments designed to predict whether an offender will commit another crime.
These actuarial tools have become common only in the past 10 years; today, most forensic psychologists recognize their value in predicting recidivism. Researchers design the assessments by examining sex offender databases. They look for the variables that best correlate with repeated criminal activity, then whittle these down to the most important factors.
The simplest tool for evaluating sex offenders is the Rapid Risk Assessment for Sex Offense Recidivism. The RRASOR (pronounced "razor") includes just four items. A sex offender gets one point for being under the age of 25 at the time of his release from prison, another if any of his victims were male, and a third if he wasn't related to his victims. He gets up to three more points depending on how many sex crimes he's been charged with.
The most dangerous offenders, then, are young adults who have committed multiple sex crimes against boys they've never met before. According to the RRASOR's table of probabilities, these six-point cases have more than a 73 percent chance of committing another crime within 10 years. . . .
Let's say that the RRASOR test seems like an accurate predictor, and in particular sex criminals (overwhelmingly male ones) who attacked a boy are found to be substantially likelier to reoffend than those who attacked a girl. May parole boards and sentencing judges take the test's result into account? (1) Would this be sex discrimination, on the theory that it discriminates based on the sex of the victim, though not of the offender? (2) Would it therefore violate the Equal Protection Clause? (Note that the Equal Protection Clause has been interpreted as presumptively barring discrimination based on sex, though generally not discrimination based on sexual orientation.)
Keith: What about Craig v. Boren's condemnation of discrimination based on statistics?
But I think for a parole board to adopt a simple policy of "high RRASOR score = parole denied" would be an Equal Protection violation. (I do, by the way, think that Equal Protection should apply to sexual orientation; I agreed with O'Connor in Lawrence.)
If that's not a sufficiently bright-line standard, than I'd lean toward saying that parole boards can't take a test like RRASOR into account at all, even indirectly, until the "Nothing in this Constitution shall be construed as" anti-pedophile amendment is passed. Supermajority support for that one should be a political no-brainer.
I'm tired of having judges and parole boards release these guys and then pretend that they're surprised when they strike again.
The parole process is a way to get out of a full sentence, so it's not a right or an entitlement. The standard, then, is lower than if the felon were entitled to leave but RRASOR were keeping him/her inside. The other factors like multiple offenses, age or whatever else are fair game, since parole is a subjective matter. In general, I'd say statistical discrimination is a useful tool for bigots and morons to use against unpopular groups (in addition to its legitimate uses), and should be avoided in situations like these, but parole is (again) very subjective. It's troubling to decide someone's fate because other people like him have done good or bad things.
It is sex discrimination, since it discriminates based on sex, but the 14th just says "equal protection of the laws" and doesn't explicitly say all sex discrimination is out (the draft is still men-only, as are most combat roles in the military). Still, it seems to me that the 'gay' pedophiles have a least a shot at making a case for discrimination relative to 'straight' pedophiles (since they're getting edged out of parole for abusing boys and not girls). The sexual orientation discrimination is a lot closer, assuming one accepts that the 14th forbids it.
The real problem is parole itself. It's going to get messy introducing a subjective system wherein a few people can attempt to project their own values (however popular or unpopular ) and judgment (however sound or unsound) onto possibly reducing a prisoner's sentence. I think we need to scale back or eliminate the parole system itself, rather than rearranging the manner in which it is provided. That goes extra for sexual offenders, many of whom deserve capital punishment.
One issue is specific versus general deterrance. People who kill their spouses are unlikely to go around killing other people, but the system should treat them severely to discourage other people from killing their spouses.
Meanwhile, if the RRASOR researchers find that sex offenders who had Swedish surnames at the time they committed the offense were differently likely to re-offend upon release than the general population, can that highly-suspect-but-not-always-absolutely-forbidden aspect be taken into account?
I think the criticism of statistical use in Craig vs. Boren included the issue that 2% of males (the unlucky folk who couldn't drink beer between 18 and 21) were DUI; the court didn't like a 100% rule applied to 2% of the people. They were also concerned that the statistics were flawed, in that Oklahoma troopers in the 1970's might not be eager to arrest drunk teenage girls.
In some of the sex cases, the person is extraordinarily likely to reoffend; the "She was hot, she was 8, and she wanted me" folk are problematic and recidivist. This isn't a 2% situation, and the harm of a single sex crime vastly exceeds the harm of a non-injury DUI.
Certainly, basing ongoing incarceration based on psychiatric testimony and statistical data is an interesting and difficult question. Each state has its own parole rules (or, in California and some other states, civil commitments for Sexually Violent Predators) and its own evidentiary structure; I don't think the use of tests generally violates Equal Protection. I do, however, think if your primary source is the RRASOR, you're being lazy - there are a fair number of far more involved and more accurate and informative tests.