Two ways to classify sex offenders and the battle over which is better

A couple of weeks ago — the City Beat forgets when because time flies when you’re writing lots of stories — I wrote about North Dakota’s failure to comply with the Adam Walsh act, better known around here for its inclusion of the Dru Sjodin national sex offender registry.

I have a bunch of links for those interested in the issue, which I have to get off my computer. They’re driving me insane.

Basically, the only jurisdictions that have complied are: Delaware, Florida, Ohio, the Confederated Tribes of the Umatilla Indian Reservation and the Confederated Tribes (in Oregon) and Bands of the Yakama Nation (in Washington state). Everybody else is holding out, possibly until the drop-dead deadline next July or, in some cases, forever.

The Adam Walsh act is called that by most people, but in Department of Justice jargon it’s also known as the Sex Offender Registration and Notification Act, or SORNA. I say this because I’ve gotten some confused replies when I asked about SORNA. They are one and the same and the people charged with ensuring implementation are the fine people at the SMART office, which is short for Office of Sex Offender Sentencing, Monitoring, Apprehending, Registering and Tracking. It’s an awesome acronym, like the spy agencies from those old TV shows and films: UNCLE or CONTROL or SHADO or SPECTRE.

But I digress. SORNA is meant to create a unified system for tracking and ranking sex offenders at the national level. Unified is key because, currently, it’s a patchwork. (For a briefer overview of SORNA, see this link.)

My story was about one reason why that patchwork endured because some states, North Dakota included, think a few aspects of SORNA go against their interests: They think their system for assessing the risk that individual sex offenders will re-offend is better and offers more protection.

In other words, this is partly a conflict between two systems of risk assessment.

The system used by the protesting states, generally called the “actuarial approach” to risk assessment or “risk-based” classification, involves multiple factors that includes everything from the age of the victim to the offender’s current living conditions. I’m not sure if it’s based on any kind of science, but I’m certain it’s based on what law enforcement people would consider years of practical experience.

The system that’s supposed to be used under SORNA, generally called the “offense-based approach,” bases risk on one factor: The crime the sex offender was convicted of.

While actuarial systems may vary depending on the state, sex offender statutes do not vary as much. The feds say that making exceptions for the protesting states would undermine the unity of a national system. The point is for a sex offender to have exactly the same risk ranking from one state to another so everyone is on the same page about just how much of a threat he is.

Why N.D. thinks our system is the best

The argument for North Dakota’s system, as articulated by Attorney General Wayne Stenehjem is this: Our system is better at identifying risk because we look at many, many factors instead of just the conviction, which can be misleading.

As we all know, the justice system is imperfect and, at times, it can be overburdened. For this reason, plea bargaining is extremely common — some sources say 95 percent of felony convictions occur for this reason — which means that, many times, the conviction doesn’t fit the crime. This could be because the prosecution isn’t sure it can get the conviction it wants or because some poor schmuck got intimidated into accepting a conviction. I won’t discuss the pros and cons here. Suffice to say that the conviction may not always fit the crime 100 percent.

The actuarial system used by North Dakota generally sidesteps the convictions and bases assessment on what occurred in the incident, among other things. The list of mitigating and aggravating factors is long and you can find it here starting on page 6. The state’s Sex Offender Risk Assessment Committee, or SORAC, uses this as a guideline.

A consequence of using this system is that sometimes a sex offender will get a higher risk rating than the conviction warrants because, let’s say, he refuses to go into sex offender treatment and has indicated he’d re-offend. It also means that sometimes a sex offender will get a lower risk rating because, let’s say, he has a stable living situation and is in such poor health that he isn’t much of a threat.

Potential problems with the system

There are two potential objections to this approach that occurred to me as I wrote the story. Here I’ll digress a little because I think these are issues worth thinking about.

The first objection is that some committee could lower a sex offender’s risk in a way that might contradict the seriousness of the crime he’s convicted of committing. A reasonable person might ask if it isn’t better to be safe than sorry in that sort of circumstance.

From what I’ve gathered, that’s actually completely unreasonable from a law enforcement perspective.

Inflating the number of higher risk sex offenders would effectively dilute the registry with so many of them that “high risk” would cease to mean high risk to the public, leading to complacency.

On the other hand, it could also increase public persecution of sex offenders that weren’t very high risk to begin with and turn him into the very thing we feared him to be. Experience informs law enforcement that risk levels can drop if a sex offender has a stable life, with a job, some kind of social network to support him and isn’t on drugs.

I’ve read stories of paranoid laws, demanded by a paranoid public, that makes it illegal for sex offenders to be anywhere close to places where children might be, effectively driving sex offenders into the streets. The laws have, as a result, created a group of people who have nothing to lose and who are difficult for law enforcement to track, exactly the opposite of what the public would actually want.

Higher risks would then require more frequent checks by law enforcement. Given that there are only so many police officers and sheriff’s deputies that we as taxpayers can afford to hire, asking them to spread their resources thin means that they aren’t able to focus on the real high-risk sex offenders.

The second objection is opposite of the first, and it has to do with a committee elevating a sex offender’s risk above the seriousness of the crime he’s convicted of committing. Sex offender status is the equivalent of a scarlet letter and, even though the main purpose is simply to alert the public, the public scorn that it earns is essentially an additional punishment on top of the sentence that a sex offender would already have served.

I don’t necessarily have a problem with that — too bad for them, but they shouldn’t have done it — but I did wonder if the ability to elevate a sex offender’s risk isn’t essentially a way for the state to get after the sex offender not for the crime he was convicted of but for the crime it wished he were convicted of.

That’s probably overstating the case a little. From what I understand, SORAC starts with various reports — drug and alcohol records, pre-sentencing investigation, sentencing reports, criminal records, police reports, psychological evaluations, prison disciplinary records — to establish a baseline risk level. It then raises or lowers that risk level based on other factors, some of which I’ve already mentioned, such as willingness to go through sex offender treatment or having a steady job.

Out of all that, the one thing that stands out to me is the police report because that’s just what police officers say they think happened. And, in plea bargained cases, the veracity of those reports haven’t been challenged in a court of law. Perhaps a committee could balance all this out, but, you know, it still raises a flag for me.

Here ends the digression.

Who else likes the actuarial approach

Now, North Dakota isn’t the only state that uses the actuarial approach. One source, the testimony of a Seattle expert on this approach, pegs the number at about 20. A slightly older source, a legal journal article, pegs it at 32, including the District of Columbia. (That article, by the way, has some extensive discussion about the actuarial approach. I didn’t see it until now and don’t have time to read in-depth. But if you’re very curious, it looks like a good read, though it sounds rather critical of the approach.)

To get the big picture with SORNA I did some Googling to see what kinds of dicussion are out there. A key source for me was a House Judiciary Committee hearing from March 2009. It brought together both supporters of SORNA and supporters of the actuarial approach.

Here are some choice quotes. Because they are so long, I’ve underlined stuff that seems relevant to me to save you time.

Madeline M. Carter, principal, Center for Sex Offender Management at the Center for Effective Public Policy in Silver Spring, Md. CSOM is sponsored by the Department of Justice.

Perhaps one of the most illuminating research findings relates to the label “sex offender.” One of the fundamental problems in our field is that we tend to paint all sex offenders with the same brush when professionals in the field have long recognized key differences among these offenders. These differences relate to the types of crimes they commit and the victims they target, the pathways that lead to their abusive behavior, the degree to which they are motivated to change, their risk for recidivism, and the types of interventions that will most likely reduce their risk for reoffense.

These key differences have important implications. For example, among adult sex offenders, research tells us that some are at higher risk to reoffend than others. While some are extremely dangerous others can be safely managed in the community. Research further distinguishes adult sex offenders from their juvenile counterparts: Juveniles are developmentally different, have lower recidivism rates, and seem to respond well to treatment.

These research findings suggest that a “one size fits all” approach to sex offender policy is inappropriate. Instead, a more tailored and strategic approach is called for. I respectfully recommend that this Committee support further examination of the differences between juvenile and adult sex offenders, and the treatment, supervision, and other supports needed to prevent specific sub-populations of offenders from committing new crimes….

Risk assessment is an important tool in our management arsenal. If a one size fits all approach is not appropriate, we need a way to distinguish among sex offenders. Until recently, we had no choice but to categorize offenders primarily on the basis of the specific offense they had committed. Risk assessment instruments offer a scientifically-based method to distinguish important differences among individuals. While these tools are not perfect, they have been consistently demonstrated to be more reliable than professional judgment.

Given the significant advances in research-both in terms of our understanding that sex offenders are not all alike, and in terms of our ability to distinguish sex offenders from one another through the use of risk assessment tools—a tailored approach to sex offender management, based upon risk to reoffend, should be employed to all of our sex offender management strategies.

The road to moving the criminal justice system from an offense-based to a risk based system, not only for sex offenders but also with other offender types, has been a long one. Today, many states use actuarial risk assessment to differentiate between offenders; resource allocation and management strategies are deployed accordingly. I encourage this Committee to consider establishing a commission to examine the use of actuarial risk assessment tools to guide the tiering of sex offenders for registration and notification purposes.

Robert Shilling, Seattle Police Department, Sex and Kidnapping Offender Detail, Sexual Assault and Child Abuse Unit:

Washington State has been in the national forefront of sex offender management and in ensuring public safety from sex crimes. We have an End of Sentence Review Committee that looks at the risk each sex offender poses to the community prior to their release from prison. We have a highly regarded sex offender treatment program within the prison system, and statewide certification of sex offender treatment providers in private practice. We do actuarial risk assessments on each of our sex offenders in an effort to identify those who are most likely to re-offend. This helps put precious public safety resources where they are needed the most; monitoring the highest risk offenders. We proactively educate our community about sex offenders. We want the public to be able to protect themselves from known sex offenders, as well as those who haven’t been caught yet. We also educate the community that it’s in the best interest of public safety to be invested in the offender’s success when they are released….

The SORNA mandates offense based tiering, which is a faulty alternative to actuarial risk based tiering used in over 20 states. Citizens have grown used to level one sex offenders being low risk, level 2 moderate risk, and level three high risk. Under SORNA, most sex offenders will be tier 3. That will cause great confusion and anxiety for the citizens, as they believe each of these offenders is a high risk to re-offend. That just is not true. Sex offenders differ greatly in their level of impulsiveness, persistence, risk to the community, and their desire to change their deviant behavior. Assigning sex offender tiers based on crime of conviction tells us very little about who this sex offender is and what his or her risk for re-offense may be. In Washington State, I have the ability to aggravate someone’s risk level if dynamic risk factors indicate an escalation in risky behavior. I won’t have that ability under SORNA. Their tier is their tier. It is not an effective way of doing business with the public.

Amy Borror, public information officer, Office of Ohio Public Defender:

The transition from a risk-based classification system to an offense-based system has turned Ohio’s sex offender registry upside down.

Prior to adopting SB 10, Ohio had a risk-based sex offender classification system. After a conviction of or plea to a sexually oriented offense, a hearing was held to determine whether the offender was likely to commit another sex offense in the future. While these proceedings were deemed to be civil in nature, the Ohio legislature recognized that the offenders needed procedural protections. At the hearing, the offender and the prosecutor could present evidence, call and examine witnesses and expert witnesses, and cross-examine witnesses and expert witnesses. The offender had the right to be represented by counsel and, if indigent, to be provided counsel at state expense. The state had the burden to prove, by clear and convincing evidence, that the offender was likely to reoffend. And, the offender had the right to appeal an adverse ruling.

Simplifying Ohio’s risk-based classification system a bit, offenders could be classified into one of three categories. An offender who had been convicted of or pled to a sexually oriented offense, but who had not been found likely to re-offend, was labeled a sexually oriented offender. An offender who had a prior conviction for a sexually oriented offense, but had not been found likely to re-offend, was labeled a habitual sexual offender. And an offender who had been convicted of or pled to a sexually oriented offense, and had been found likely to commit another sex offense in the future, was labeled a sexual predator. These three categories roughly translate, in duration and requirements of registration, to the Adam Walsh Act’s Tier I, Tier II, and Tier III, respectively.

The state’s risk-based classification system had resulted in a registry that looked much like what scientific research tells us about the likelihood of sex offender recidivism: 77% of Ohio sex offenders were classified as sexually oriented offenders, 4% were labeled habitual sexual offenders, and 18% were labeled sexual predators. After implementing SB 10, Ohio’s registry became top-heavy: only 13% of offenders are classified in Tier I, 33% are in Tier II, and 54% are in Tier III.

The number of people in the highest tier of Ohio’s registry has tripled. Nearly 8,000 of Ohio’s sex offender registrants were moved from one of the two lower classification levels into Tier III — not because they had committed a new crime or because of new evidence of their future dangerousness, but only because of the crime of which they had been previously convicted.

Ohio’s old registry was, potentially, a useful public safety tool. It included more than 22,000 offenders; however, only 4,000 of those offenders were labeled sexual predators. Those 4,000 offenders, found by a judge to be likely to reoffend, would rightly garner the most attention from the public and require the closest supervision by law enforcement. Now, however, Ohio’s registry includes more than 12,000 people labeled as Tier III offenders. Their propensity to reoffend is not known, but the public will certainly perceive them as dangerous, and law enforcement must expend tremendous resources to supervise them.

Under Ohio’s old law, a person convicted of rape for consensual sex with a person four years and one day his junior might have been classified a sexually oriented offender, if that person had not been found likely to commit another sex crime. Also under Ohio’s old law, a person convicted of sexual imposition, a misdemeanor, might have been classified a sexual predator, if a judge found him likely to reoffend. Now, however, Ohio courts are mandated to classify the person convicted of rape as a Tier III offender and the person convicted of sexual imposition as a Tier I offender.

The person convicted of rape could lead a law-abiding life and could even, as happened in at least one Ohio case, marry the “victim” of his offense and have a family, but he would forever be labeled a Tier III offender, the supposed worst of the worst. Even though the person convicted of sexual imposition is likely to commit future sex offenses, a judge would not be able to classify that person into a higher tier until that person committed and was convicted of a subsequent sex offense. Instead of being able to properly label a high-risk offender, the court must instead wait until another offense is committed and another victim is created.

Sex offender registration and notification laws are supposed to be forward-looking, aimed at protecting the public from future crimes. Risk-based systems, like Ohio’s prior scheme, do a much better job of addressing the stated aim of sex offender registries: protecting the public from future criminal acts.

In its position paper on the Adam Walsh Act, the National Alliance to End Sexual Violence (NAESV), a victim advocacy organization that conducts the public policy work of state sexual assault coalitions and rape crisis centers, states that, “over-inclusive public notification can actually be harmful to public safety by diluting the ability to identify the most dangerous offenders and by disrupting the stability of low-risk offenders in ways that may increase their risk of re-offense. Therefore, NAESV believes that internet disclosure and community notification should be limited to those offenders who pose the highest risk of re-offense.”

The Adam Walsh Act, however, is not concerned with the likelihood of future crimes. It looks only at past offenses and labels offenders based on those past offenses, without considering what those offenders might do in the future.

Why the feds don’t want the actuarial approach

The bottomline for the feds is… Well, let me point out that there are two feds we’re talking about here. The first is Congress, which wrote SORNA and was very specific about certain things, among them the offense-based approach to sex offender risk assessment. The second is the Department of Justice, by way of the agents of SMART, which is tasked with enforcing the law the way it was written with limited flexibility.

It appears that the two are on the same page. I distinguish between them because it’s worth noting that states primarily deal with SMART when it comes to implementing SORNA, but to get an exception made for the actuarial approach requires lobbying Congress. So far, I’ve not heard of any major lobbying effort and I’m guessing that’s probably because there’s still a year to go before states are penalized for not implementing SORNA. (The penalty is a 10 percent reduction in Byrne grants, but that penalty is so small for North Dakota it means pretty much zip. Wayne Stenehjem said in fiscal year 2008, we would’ve risked $31,000, in FY 2009 $82,000 and in FY 2010 $78,000.>)

Anyway, the feds don’t really have specific objection to the actuarial approach. They seem to only want something that can be implemented nationwide and, it seems to me, the offense-based approach is the easiest because it’s the simplest and cheapest. Some also question the efficacy of the actuarial approach, but I think the main objection is that such an approach would create a patchwork of risk assessment processes that undermine the goal of having a unified national policy.

I called Sen. Byron Dorgan, D-N.D., to talk about this. He authored the section on the Dru Sjodin Web site. He didn’t go into a lot of details, but insisted that there has to be some national system and what we have now doesn’t work at that level. If states have a problem, he said, they should implement SORNA first and then Congress can work with them to make necessary modifications.

I can see how the states would look at this. If you believe the offense-based approach will be less safe for your residents, why should you compromise and go for the lowest common denominator? And second, how long would it take before Congress will revisit the issue and make the changes you desire, if ever?

Anyway, here are more voices in support of SORNA and the offense-based approach.

Laura Rogers, former director of SMART:

There has been much progress throughout this nation in the implementation of the Sex Offender Registration and Notification Act (SORNA). However, the momentum with which this progress is being made stands to be undermined if special-interest groups’ and individual jurisdiction’s myopic criticisms of the law is allowed to change the statutory language of SORNA. Individuals who do not have a national perspective do not understand the significance of the jurisdiction-specific modifications they seek.

Congress intended to give this country and its citizens a comprehensive system for sex offender registration and notification under SORNA. SORNA recognized that every jurisdiction is unique, with distinct systems and issues, and SORNA provides significant flexibility that will allow for the comprehensive nature of the Act to be achieved, while still requiring jurisdictions to meet or exceed equivalent minimum standards.

Modification to SORNA will not resolve all hurdles to substantial implementation. Modifications to SORNA will create new and different issues. As the SMART Office currently does, each jurisdiction must be worked with individually to achieve success in a unique way….

There is no workable alternative to a system like SORNA. SORNA is an evidence-based system that requires registration based on the fact that the sex offender has ALREADY been convicted of assaulting a real person. There is a movement afoot however, to remove the evidence based component of SORNA and replace it with a soft (and unproven) artifice called “risk assessments.” Congress wisely recognized that risk assessment tools should not used to determine if a convicted sex offender should register — by guessing whether they will re-offend. [Remember that supporters of the actuarial approach think that it has, in fact, been proven. Somebody must be blowing smoke up somebody's something.] Rightly so, Congress recognized that risk assessments are not foolproof and are not useful for juveniles. However, “risk assessment” tools remain available for treatment purposes. Currently, only a minority of jurisdictions use them for registration purposes, and it should remain that way for good reason. For one reason, besides the obvious (they are not reliable) there are an insufficient amount of trained professionals available to appropriately administer risk assessment tools to all the sex offenders in the United States.

Evelyn Fortier; vice president for policy; Rape, Abuse and Incest National Network:

SORNA’s Emphasis on Uniformity is Positive. Title I of the Adam Walsh Act creates a comprehensive national system for the registration of sex offenders, defining three tiers of sex offenders (depending on the severity of their crimes). A uniform sex offender registration system, if implemented successfully by all jurisdictions, promises to eliminate inconsistencies in the various states’ laws. Before the passage of the Adam Walsh Act, Congress heard about legal loopholes that enabled sex offenders to “forum shop,” i.e., find jurisdictions with less stringent laws to evade sex offender registration and notification requirements. Prior to the enactment of the Adam Walsh Act, Congress also heard that as many as one in five sexual offenders who were required to register would eventually go “missing” from the system. By encouraging uniformity across jurisdictions, the Adam Walsh Act should help prevent sex offenders from evading detection. The Adam Walsh Act also provides an avenue for states to share data about sex offenders, which is a positive feature of the Act.

An Objective Offender Classification System Promotes Fairness. SORNA classifies offenders into three categories (tier I, tier II, or tier III), depending upon the severity of their crimes. For example, the tier III offender, considered the most serious of the three categories, will have committed an offense that is punishable by more than one year in jail and is at least as severe as certain listed offenses, and involves kidnapping a minor or occurs after the offender becomes a tier II offender. The tier III offender is subject to lifetime registration under SORNA.

Some have argued that classifying offenders using an actuarial risk assessment system would be preferable to SORNA’s offense-based classification system. We note, however, that an offense-based classification system is far more objective than a risk-based assessment scheme. An objective system may be the best way to achieve fairness for all. Also, an objective system may also be far less costly to administer than a subjective, risk-based assessment system because the objective system does not obligate police departments across the nation to hire psychologists or other professionals to individually assess every offender’s risk after they have already been tried and convicted.

Rep. Louie Gohmert, R-Texas; ranking member; Subcommittee on Crime, Terrorism and Homeland Security; Judiciary Committee:

Some States take issue with SORNA’s offense-based approach of categorizing sex offenders by their crimes and requiring individuals who committed similar crimes to have similar registration obligations. These States advocate a risk-assessment approach to registration that utilizes actuarial tools to predict recidivism by taking an individual’s criminal history, victim profile, and age into account.

However, there is little consistency to these various programs. They are not uniform in the criteria they apply or in who performs the assessments. This creates discrepancies over which sex offenders should be tracked nationwide.

Despite these discrepancies, risk-assessment States allege their approach is better than SORNA’s offense-based approach. Washington State uses the risk-assessment approach, but it cannot properly track Darrin Sanford, a convicted sex offender with a history of failure in registering as a sex offender. Sanford had been identified as a person with a high likelihood to re-offend, so much so that he was forced to wear a GPS tracking device. Although Mr. Sanford was under Washington’s highest level of supervision, this did not stop him from assaulting and killing a 13-year-old girl in Walla Walla last month, a crime that he confessed to committing. Until there is some uniformity to these risk-assessment programs and they demonstrate a better track record, the most reliable approach is to track offenders by offense and to lock those up who fail to register.

[This seems to be a very bad example. The system knew he was a high risk and it was proven correct!]

As you can see, this is a pretty complex issue. I found it interesting and enjoyable to research because, the more you dig into it, the farther away you get from fear mongering and closer to serious policy discussions.