Policy

Perverted Justice

Sex offender laws represent the triumph of outrage over reason.

|

"If we had been aware of his record," says Maureen Kanka, "my daughter would be alive today." She is referring, in a statement on the website of an anti-crime group she founded, to Jesse Timmendequas, a neighbor in Hamilton Township, New Jersey, who raped and murdered her 7-year-old daughter, Megan, in 1994. Three months later, the state legislature enacted Megan's Law, which created a publicly accessible registry of sex offenders.

"Without the registry," says Shirley Turner, "he would still be alive today." She is referring, in a 2006 interview with Human Rights Watch, to her 24-year-old son, William Elliot. He was murdered that year by a pedophile-hunting Canadian gunman who found his name and address in Maine's online database of sex offenders. Elliot's crime: When he was 19, he had sex with his girlfriend, who was three weeks shy of 16, the age of consent in Maine.

The panic that followed Megan Kanka's murder produced an alarm system that often fails to distinguish between dangerous predators like Timmendequas, who had a record of assaulting little girls, and nonviolent lawbreakers like Elliot, who posed no discernible threat to the general public. They are all mixed together in the online registries of sex offenders that every state is required to maintain as a condition of receiving federal law enforcement funding—a mandate imposed by another Megan's Law, enacted by Congress in 1996.

Registration only rarely leads to murder, but it routinely ruins relationships, triggers ostracism and harassment, and impedes education and employment. These burdens are compounded by state and local laws that ban sex offenders from living near schools, parks, day care centers, and other locations where children congregate. Such restrictions, which often apply even if an offender's crime had nothing to do with children, can be so extensive that entire cities are effectively off limits. In Miami local residence restrictions have given rise to a colony of more than 70 sex offenders who live under the Julia Tuttle Causeway, a bridge that crosses Biscayne Bay.

Some sex offenders, including nonviolent ones, will not live to see the underside of a bridge because they receive sentences that keep them behind bars until they die. Two decades of ever-more-punitive legislation have produced sentencing rules so bizarre and byzantine that the punishment for possessing images of sexually abused children can be more severe than the punishment for sexually abusing them. And even prisoners who complete their sentences may not go free, since the federal government and about half of the states have laws authorizing the indefinite civil commitment of sex offenders who would otherwise be released.

American policies regarding sex offenders mark them as a special category of criminals for whom no stigma is too crippling, no regulations are too restrictive, and no penalty is too severe. This attitude, driven by fear and outrage, is fundamentally irrational, and so are its results, which make little sense in terms of justice or public safety. Like the lustful predators of their nightmares, Americans pondering the right way to deal with sex offenders seem captive to their passions.

'I Am on the Registry for Having Premarital Sex'

The public branding of sex offenders through online registries is a reaction to horrible, highly publicized crimes, such as Megan Kanka's murder, in which strangers abduct, rape, and kill children. But this sort of crime is exceedingly rare. Data from the Justice Department's National Crime Victimization Survey indicate that more than 90 percent of sexually abused minors are assaulted by relatives or acquaintances—people they trust. (According to the same survey, strangers commit just one in four sexual assaults on adults. They commit only 14 percent of sexual assaults reported to police.) Furthermore, according to a 1997 Justice Department study, nearly nine out of 10 people arrested for sex offenses have no prior convictions for this category of crime, so they would not show up in sex offender registries.

Meanwhile, the people on sex offender lists may pose little or no threat. A 2007 report by Human Rights Watch found that "at least 28 states require registration as a sex offender for someone convicted of having consensual sex with another teenager, if the offender was either age 17 or two years older than the other party." Eleven states set no minimum age difference. "It's one thing if you are a 40-year-old having sex with a 13-year-old," says the report's co-author and editor, Jamie Fellner, senior adviser to the U.S. Program of Human Rights Watch. "It's another thing if you're a 17-year-old boy having sex with your 16-year-old or 15-year-old girlfriend. Registration as a sex offender is just completely inappropriate there, does nothing to promote public safety, but ruins lives."

A man who was convicted of statutory rape when he was 16 for having consensual sex with his 14-year-old girlfriend told Human Rights Watch: "We were in love. And now we are married. So it's like I am on the registry for having premarital sex. Does having premarital sex make me a danger to society? My wife doesn't think so."

The Human Rights Watch report also found that at least five states required registration for offenses related to adult prostitution, at least 13 required registration for public urination, and at least 32 required registration for exposing one's genitals in public. And from the information given in a registry, which typically is limited to a vague legal description of the offense, it is often hard to tell what someone did to end up there. "Without any further information, it is difficult to provide reasonable steps that people can take to help keep themselves safe," says Maia Christopher, executive director of the Association for Treatment of Sexual Abusers. "Just knowing that there is someone living next door to you who's committed a sexual offense doesn't necessarily give you enough information to know what you're supposed to do about it." In fact, when the U.S. Supreme Court upheld Connecticut's sex offender registry in 2003, it did so partly because the state expressly disavowed any claims about the "current dangerousness" of the people in its electronic pillory, which meant they did not have a due process right to a hearing on that question.

Consider the case of Tony Washington, a promising college football player whose professional career was derailed by a conversation-stopping offense he committed almost a decade ago: At the age of 16, he had consensual sex with his 15-year-old sister. A 2010 profile of Washington in ESPN magazine explained the context of this forbidden liaison: a troubled, dispiriting childhood in the rougher sections of New Orleans, where Washington was constantly threatened by violence and had few sources of emotional support. Although he overcame a deprived background to become a star player at Abilene Christian University in Texas, his taboo-breaking transgression has deterred professional teams from drafting him and will mark him until the day he dies.

If you search for Washington's name in the Texas sex offender registry or the U.S. Justice Department's nationwide database, you will see photographs, a physical description, his date of birth, and his home address. His offense is listed as "prohibited sexual conduct," which most people, given the context, will assume refers to some sort of predatory crime. Few people will bother to look up the Texas statute explaining that consensual sex with several different kinds of relatives, including adopted siblings and first cousins, qualifies for this label, triggering the same lifelong registration requirement that applies to rapists and child molesters. (On its face, the law even covers sex between first cousins who were legally married in one of the 25 states that allow such unions.) Whatever you may think of Washington's crime, it hardly marks him as a public menace whom women and children should fear, let alone as someone who will be a danger to others even when he is old and infirm.

Washington's case illustrates another way in which the legal treatment of sex offenders is unusual. Although the records of juvenile offenses typically are sealed, sex registration is public, and it applies even to people who, like Washington, committed their offenses as teenagers or children. According to The Dallas Morning News, the sex offender registry in Texas, where Washington lives, includes about 4,000 people who were minors when they committed their crimes, a quarter of whom were under 14. Human Rights Watch interviewed the father of a 10-year-old boy accused of touching his 5-year-old cousin's genitals. "My son doesn't really understand what sex is," he told the group,"so it's hard to help him understand why he has to register as a sex offender." This policy of tarring minors as sex offenders undermines a central aim of the juvenile justice system by burdening people with the mistakes of their youth for the rest of their lives.

'No Area in Town Is Available to Live In'

Like registration, residence restrictions are ostensibly aimed at protecting potential victims from known sex offenders, in this case by creating a geographical buffer. But the logic of these rules, which have been adopted by more than 20 states and hundreds of municipalities, is hard to understand. "I don't know of any research that suggests the residency restrictions are effective," says Christopher. "People don't necessarily offend where they live."

In 2005 Iowa banned people convicted of sex offenses involving minors from living within 2,000 feet of a school or day care center. Almost immediately, the ban prompted complaints from police and prosecutors, who worried that the residence restrictions were so burdensome that they discouraged sex offenders from pleading guilty and from registering after conviction, making them impossible to track. In a 2006 statement, the Iowa County Attorneys Association said the law "does not provide the protection that was originally intended" and called for its repeal, citing "the cost of enforcing the requirement and the unintended effects on families of offenders." After the law took effect, the number of sex offenders whose whereabouts were unknown more than doubled. The prosecutors reported that "the residency restriction is causing offenders to become homeless, to change residences without notifying authorities of their new locations, to register false addresses or to simply disappear."

Jerry Behn, the Republican state senator who introduced the law, conceded that he might have gotten carried away. "If you draw a map, pretty soon you can make it so no area in town is available to live in," Behn told The Atlanta Journal-Constitution in 2006. "It would have been better if we had put it at 1,000 feet." In 2009 the legislature replaced the 2,000-foot rule with a law prohibiting anyone convicted of "a sex offense against a minor" from working at a school or day care center, visiting an elementary school or day care center without permission, or loitering within 300 feet of a facility intended primarily for use by children.

Iowa's unhappy experience with residence restrictions did not deter Georgia legislators from enacting an even more onerous set of rules in 2006. Their law, which extended existing residence restrictions, banned all registered sex offenders (not just those who had committed crimes against children) from living, working, or loitering within 1,000 feet of schools, churches, child care facilities, or other places where minors congregate, including parks, playgrounds, swimming pools, skating rinks, and school bus stops. Even a sex offender who did not currently live within one of these exclusion zones could be forced to move in the future, depending on how his neighbors decided to use their property. For the law's sponsors, its indiscriminate breadth was a feature, not a bug. Georgia House Majority Leader Jerry Keen (R-St. Simons) said he hoped sex offenders "will want to move to another state."

Under the 2006 law, all 490 registered sex offenders in DeKalb County, most of them men who as teenagers had consensual sex with younger girls, were required to move because their residences were within 1,000 feet of a covered location. The law applied even to sex offenders dying in nursing homes. One Georgia woman, labeled a sex offender because she performed fellatio on a 15-year-old boy when she was 17, had to move in 2005 because she was too close to a day care center. When the legislature added school bus stops to the list of prohibited locations in 2006, her new home became illegal as well.

In 2007 the Georgia Supreme Court struck down the residence restrictions, citing the perpetual insecurity they created. The court was responding to a lawsuit by Anthony Mann, who in 2002 was sentenced to four months in jail and five years of probation for flashing two minors during a business trip to North Carolina. In 2003 Mann married and bought a house in Hampton, Georgia. At the time, it was a legal location. But then a day care center opened nearby, rendering his residence illegal. "Under the terms of that statute," the state Supreme Court noted, "there is no place in Georgia where a registered sex offender can live without being continually at risk of being ejected." Concluding that the law "precludes appellant from having any reasonable investment-backed expectation in any property purchased as his private residence," the court unanimously ruled that it violated the Fifth Amendment's ban on uncompensated takings of private property.

The Georgia law also prevented Mann from working at a barbecue restaurant in Lovejoy that he co-owned, since it was within 1,000 feet of a day care center that opened after the restaurant was established. But because Mann did not present enough evidence of economic harm, the Georgia Supreme Court did not overturn the law's work restrictions.

'A 90 Percent Likelihood of Recidivism'

Constitutional issues aside, closing off employment opportunities for sex offenders, who are already handicapped by criminal records and public registration, does not seem like an effective way of encouraging them to put their lives in order. Neither does forcing ex-convicts to cluster in the boondocks, far from jobs, family, churches, and treatment programs.

"Once someone is zoned out that far, a lot of the resources that help reduce the risk to reoffend are taken away," says Maia Christopher. "People who are coming out of prison, who have whatever deficits and risk factors they have, are put into situations that really high-functioning people would find very challenging. I don't think anybody finds it particularly easy to be living under a bridge. When you have someone who's already got a lot of issues, part of what we're doing is increasing that challenge. Once you make the restrictions so broad, you tend to take away a lot of the resources that we know help keep people safe, such as community stability, employment, family services, spiritual support—"the relationships that help people stay crime-free."

Human Rights Watch's Jamie Fellner likewise argues that the rules aimed at discouraging sex offenders from committing new crimes can produce the opposite effect. "One of the most powerful things that keep people from breaking the law is a sense of what they will lose if they do," she says. "If you're treating someone like a cur, a dog that is being kicked out of the village, how are you strengthening that person's desire or ability to follow the law?"

Politicians who push ever-harsher laws routinely argue that sex offenders are almost certain to commit new offenses anyway. "The rate of recidivism for these crimes is astronomical because these people are compulsive," said Rep. Jennifer Dunn (R-Wash.) in 1994, making the case for the federal Megan's Law. In 2005 Rep. Mark Foley (R-Fla.), later notorious for sending sexually suggestive email messages to teenaged pages, claimed: "There is a 90 percent likelihood of recidivism for sexual crimes against children. Ninety percent. That is the standard. That is their record. That is the likelihood. Ninety percent." His source may have been California Assemblyman Bill Hoge (R-Pasadena), who in 1996 told The New York Times that child molesters, upon being released from prison, "will immediately commit this crime again at least 90 percent of the time."

There is no basis for these numbers. "Though often thought of as the most persistent and dangerous criminals, sex offenders are among the least likely criminals to recidivate," write Florida Institute of Technology psychologist Timothy Fortney and three co-authors in a 2007 article published by the journal Sexual Offender Treatment. A 2003 Justice Department study of 9,700 sex offenders found that 5 percent were arrested for new sex crimes within three years of being released from prison. (By comparison, 23 percent of burglars were arrested for new burglaries, and 22 percent of people who had served time for nonsexual assault were arrested for new assaults.) Studies that cover longer periods find higher recidivism rates for sex offenders, but still nothing like those claimed by panic-promoting politicians. Two meta-analyses of studies involving a total of 29,000 sex offenders, published by the Journal of Consulting and Clinical Psychology in 1998 and 2005, found a recidivism rate of 14 percent after four to six years. A study of 4,700 sex offenders, published by Public Safety Canada in 2004, found that 24 percent were charged with a new sex crime over a period of 15 years.

Since the National Crime Victimization Survey indicates that most sex crimes go unreported, these numbers do not tell the whole story. But Karl Hanson, a senior research officer at Public Safety Canada who co-authored many of the recidivism studies, believes the unreported sex crimes largely fall into two broad categories: those deemed too minor to bother calling the police (such as a drunken groping at a party) and those involving relatives or trusted members of the victim's social network—neither of which fit the pattern that legislators have in mind when they argue that registration, public notification, and residence restrictions can help protect children and women from attacks by strangers.

The research on recidivism helped change the mind of at least one prominent advocate of sex offender registries. In 1989 Patty Wetterling's 11-year-old son, Jacob, was kidnapped by a masked gunman while riding his bike home from a convenience store in St. Joseph, Minnesota. He has not been seen since. The crime inspired the 1994 Jacob Wetterling Crimes Against Children and Sexually Violent Offender Registration Act, the federal law that created the state registries that Megan's Law made publicly accessible two years later. "The high recidivism rates I assumed to be true do not exist," Patty Wetterling told Human Rights Watch in 2007. "It has made me rethink the value of broad-based community notification laws, which operate on the assumption that most sex offenders are high-risk dangers to the community they are released into."

Looking vs. Touching

People convicted of molesting children have demonstrated a dangerous proclivity, even if they are less likely to repeat the crime than is popularly believed. But what about people who are convicted of possessing child pornography? Although conventional wisdom assumes child-porn consumers are undiscovered or future molesters, that assumption is also wrong.

This year Michael Seto, a psychologist who advises the Integrated Forensic Program of the Royal Ottawa Health Care Group, published a study of this question, co-authored by Karl Hanson and Kelly M. Babchishin of Public Safety Canada, in the journal Sexual Abuse. Seto, Hanson, and Babchishin performed meta-analyses of 24 studies that looked at the criminal histories of "online offenders" (mainly consumers of child pornography) and eight studies that calculated their recidivism rates. They found that one in eight had an official record of committing a contact offense. In the six studies that included self-report data (drawn from treatment sessions and polygraph examinations), one in two child pornography offenders admitted to having sexual contact with children.

Looking forward, Seto says, "C.P. offenders are relatively unlikely to commit contact offenses in the studies that have followed them." Over all, the recidivism studies indicate that only 2 percent of child pornography offenders committed a sexual offense involving physical contact during the follow-up period, which ranged from 18 months to six years. In short, says Hanson, "there does exist a distinct group of offenders who are Internet-only and do not present a significant risk for hands-on sex offending."

Why would anyone look at this horrible stuff if he was not inclined to imitate it? Troy Stabenow, an assistant federal public defender in Missouri who is a prominent critic of child pornography sentencing policies, put it this way in a 2009 interview with ABA Journal: "People who watch movies like Saw and Friday the 13th are being titillated by the act of torture and murder. That doesn't mean that they're going to go out and commit torture and murder."

Dean Boland, an Ohio defense attorney specializing in child pornography cases, says a substantial share of defendants were themselves victims of sexual abuse as children and look at these images as a way of working through the trauma. He recalls one client, a 65-year-old former pastor who received a sentence of more than 17 years, saying, "When I'm looking at these images, I'm not envisioning myself as the adult. I'm envisioning myself as the kid."

Yet the legal treatment of people caught with child pornography is so harsh that they can end up serving longer sentences than people who actually abuse children. Under federal law, receiving child pornography, which could mean downloading a single image, triggers a mandatory minimum sentence of five years—"the same as the penalty for distributing it. Merely looking at a picture can qualify someone for the same charge, assuming he does so deliberately and is aware that Web browsers automatically make copies of visited sites. In practice, since the Internet is almost always the source of child pornography, this means that viewing and possession can be treated the same as trafficking. The maximum penalty for receiving or distributing child porn is 20 years, and federal sentencing guidelines recommend stiff enhancements based on factors that are extremely common in these cases, such as using a computer, possessing more than 600 images (with each video clip counted as 75 images), and exchanging photos for something of value, including other photos.

In a devastating 2008 critique of these sentencing policies, available on his office's website, Stabenow shows that Congress ratcheted the penalties for looking at child pornography upward through a series of ill-considered, undebated dictates driven by little more than public outrage and disgust. The upshot: Between 1997 and 2007, the number of people sent to federal prison for possessing, receiving, or distributing (but not producing) child pornography quintupled, from 238 to 1,170, while the average sentence more than quadrupled, from 21 to 91 months. Among the baffling results of these policies: A defendant with no prior criminal record and no history of abusing children would qualify for a sentence of 15 to 20 years based on a small collection of child pornography and one photo swap, while a 50-year-old man who encountered a 13-year-old girl online and lured her into a sexual relationship would get no more than four years. The comparison, Stabenow writes, "demonstrates the absurdity of the system."

The absurdity has not gone unnoticed by the judiciary. In a 2010 survey by the U.S. Sentencing Commission, 70 percent of federal judges said the recommended penalties for possessing or receiving child pornography are unreasonable. Although the Supreme Court has ruled that the guidelines are only advisory and not mandatory, judges still must justify deviations in written explanations that are subject to review by appeals courts. Many have not been shy in expressing their opinions about the fairness and wisdom of the penalties they are asked to impose.

In a 2008 child pornography case, Robert Pratt, a U.S. district judge in Des Moines, gave the defendant a sentence of seven years instead of the recommended 18. The guidelines "do not appear to be based on any sort of empirical data," Pratt said, "and the Court has been unable to locate any particular rationale for them beyond the general revulsion that is associated with child exploitation-related offenses."

The following year, Lynn Adelman, a U.S. district judge in Milwaukee, cited Stabenow's critique of the sentencing guidelines when he gave a middle-aged funeral director who was caught swapping child pornography a six-year sentence instead of the 18 years sought by prosecutors. "The flaws identified by Stabenow were certainly evident in this case," Adelman wrote in his sentencing memorandum. "I could not conclude that under the circumstances of this case, given all of the flaws in the guideline discussed above, that the range deserved deference."

Jack Weinstein, a U.S. district judge in Brooklyn, has been fighting for years, sometimes through rulings of questionable legality, to spare a married father of five not only the 11-to-14-year sentence recommended by the guidelines but the five-year statutory minimum for receiving child pornography. "Imprisonment of at least five years for this defendant is cruel," Weinstein wrote in a 2008 opinion.

State penalties for possessing child pornography can be even harsher. In Arizona, one count of possessing child pornography carries a 10-year mandatory minimum sentence, each image qualifies as a separate count, and the sentences must be served consecutively. That's how Morton Berger, a former high school teacher with no criminal record, ended up with a 200-year sentence in 2003.

In 2006 the Arizona Supreme Court upheld Berger's sentence, rejecting his argument that it violated the Eighth Amendment's prohibition of cruel and unusual punishment. Writing in dissent, Vice Chief Justice Rebecca Berch noted that "Arizona's sentence for this crime is by far the longest in the nation and is more severe than sentences imposed in Arizona for arguably more serious and violent crimes." For example, "the minimum sentence for possession of an image of child pornography is longer than the presumptive sentence for rape or aggravated assault. A presumptive sentence for possession of two images of child pornography…is harsher than the sentences for second degree murder or sexual assault of a child under twelve….For molesting a child, one might receive the same sentence that Berger has received for possessing one picture."

It is hard to make any sense out of such a sentencing scheme. Ostensibly, the law punishes people for possessing child pornography because their demand for this material encourages its production, which necessarily involves the abuse of children. That is the main rationale cited in Osborne v. Ohio, the 1990 Supreme Court decision that said mere possession of child pornography can be banned without violating the First Amendment. But the minimal, indirect role that any one consumer of child pornography plays in creating a market for the stuff can hardly justify sending him to prison for years or decades, and this argument has little relevance now that people who look at child pornography typically get it online for free. "They are not protecting a single child," says Dean Boland, the defense attorney. "They are throwing people in prison for having dirty thoughts and looking at dirty pictures." The harsh treatment of people who like to look at child pornography seems to rest on the mistaken premise that they are equivalent to child molesters. In essence, they are being punished for the crimes they are expected to commit.

Preventive Detention

The same could be said of sex offenders who are confined indefinitely to mental institutions after they have completed their prison sentences. In the 1997 case Kansas v. Hendricks, the Supreme Court upheld this policy of post-sentence detention, ruling that a state law authorizing civil commitment of "sexually violent predators" was not punitive and therefore did not violate the Constitution's Double Jeopardy Clause or its ban on ex post facto laws. The Court also concluded that the criteria for commitment satisfied the requirements of substantive due process.

Under the Kansas law, an offender can be committed if a jury decides he is likely to engage in "predatory acts of sexual violence" due to a "mental abnormality" or "personality disorder." The law defines "mental abnormality" as "a congenital or acquired condition affecting the emotional or volitional capacity which predisposes the person to commit sexually violent offenses in a degree constituting such person a menace to the health and safety of others." Writing for the majority, Justice Clarence Thomas said these criteria "serve to limit involuntary civil confinement to those who suffer from a volitional impairment rendering them dangerous beyond their control." Thomas added that the "lack of volitional control, coupled with a prediction of future dangerousness, adequately distinguishes [sexually violent predators] from other dangerous persons who are perhaps more properly dealt with exclusively through criminal proceedings."

The thing is, the child molester who brought this case, Leroy Hendricks, did go through criminal proceedings, a fundamental premise of which was that he was able to control his behavior and could have chosen to act differently. Based on that premise, he was convicted of "taking indecent liberties" with two 13-year-old boys and served a 10-year sentence, whereupon the state declared that Hendricks could not go free because he was unable to control his behavior.

That was not the only apparent contradiction in the case. The Court ruled that indefinite commitment was not tantamount to a life sentence because the prisoner-cum-patient is "permitted immediate release upon a showing that the individual is no longer dangerous or mentally impaired." At the same time, the Court said the state is not constitutionally required to provide treatment that might help eliminate the danger. Indeed, the state legislature suggested that sex offenders covered by the law cannot be successfully treated. The preamble to the statute called them "a small but extremely dangerous group of sexually violent predators…who do not have a mental disease or defect that renders them appropriate for involuntary treatment" under the state's general civil commitment statute. These offenders, the law said, "generally have anti social personality features which are unamenable to existing mental illness treatment modalities."

Predictably, offenders confined under sexually violent predator laws after completing their prison sentences are almost never declared well enough to release. In 2007 the Washington State Institute for Public Policy counted discharges by the 20 states that had implemented such laws at the time. Of the 4,534 offenders who had been committed since the first such law was passed in 1990, 188 (about 4 percent) had been released based on a "program staff recommendation." According to 2011 legislative testimony by Rob Siedlecki, acting secretary of the Kansas Department of Social and Rehabilitation Services, only two out of 247 prisoner/patients (less than 1 percent) have successfully completed that state's Sexual Offender Treatment Program since it was created in 1994. Taxpayers are paying a pretty penny for these dismal results. The Washington state study found that committing a sex offender cost an average of $97,000 annually, compared to $26,000 for a year of prison, a gap that a 2007 investigation by The New York Times attributed to "higher costs for programs, treatment and supervised freedoms."

Contrary to the beliefs of legislators, there is not much evidence that psychiatrists or psychologists can predict which individuals will commit sex crimes. While sex offenders can be sorted into high-risk and low-risk groups based on their criminal histories, says Karl Hanson, the Public Safety Canada researcher, "the best that we can do for identifying high-risk offenders is identifying groups that have approximately 50 percent observed recidivism rates."

What's theoretically possible, of course, is not necessarily what government agencies actually achieve. A 2006 report from the Washington State Institute for Public Policy concluded that the methods used by the state to classify registered sex offenders by risk had "little or no accuracy in predicting sex offender recidivism." By contrast, a 2004 study by Lynn University psychologist Jill Levenson, reported in the International Journal of Offender Therapy and Comparative Criminology, found that sex offenders recommended for commitment in Florida "scored significantly higher on actuarial risk assessment instruments" and were more likely to have "other risk factors that have been empirically correlated with sexual recidivism" than sex offenders who were recommended for release.

Even when evaluators do the best they can, per Hanson's estimate, the most they can say is that half the offenders who meet certain criteria will commit new offenses if they are released, which means half will not. Such probabilistic assessments, even for crimes that have already occurred, are not usually considered sufficient grounds for depriving people of their liberty. "This is prison," says Jamie Fellner. "We shouldn't pretend otherwise. Basically, this is a form of preventive detention."

When you strip away the quasi-medical language, what states are really saying when they indefinitely commit odious individuals like Leroy Hendricks to mental hospitals is this: "Whoops. We should have given this guy a longer sentence." But it is no mere formalistic quibble to point out that a defendant's sentence should be imposed at the time of his conviction as determined by a judge within the parameters set by statute. These are basic requirements of due process and the rule of law, and we make exceptions to them at our peril. Clarence Thomas may be confident that preventive detention won't be extended to "other dangerous persons," but I am not. It seems to me that all it would take is a new law attached to a new scientific-sounding label invented by legislators or grabbed from the Diagnostic and Statistical Manual of Mental Disorders. How many convicted criminals could qualify for a diagnosis of, say, anti-social personality disorder?

In a 2004 Criminal Law Bulletin article, William Mitchell College of Law professor Eric Janus argued that "sexual predator laws provide a model for undercutting…constitutional protections." The process, Janus said, starts with a universally despised group of people who, like suspected terrorists, attract no public sympathy. He warned that "we are at risk of becoming a 'preventive state,' in which the paradigm of governmental social control has shifted from solving and punishing crimes that have been committed to identifying 'dangerous' people and depriving them of their liberty before they can do harm." To most Americans, I fear, this prospect is not nearly as scary as the possibility that a sex offender lives down the street.

Senior Editor Jacob Sullum (jsullum@reason.com) is a nationally syndicated columnist.