In June of 1994, a convicted child molester named Charlie Taylor moved into a small apartment in downtown Hamilton, Ontario, Canada, across the street from a community center. He had no family. He had no parole officer. At the time, sex offenders deemed too dangerous to be let out of prison early were, paradoxically, released at the end of their sentences with no ongoing oversight or treatment from the Correctional Services of Canada.
What he did have was a friend. Harry Nigh, a Mennonite pastor at The Welcome Inn church had invited Charlie Taylor to town. Without knowing Taylor personally, without any real training in dealing with sex offenders, Nigh welcomed Taylor and brought him home for supper. Nigh's son, then a third grader, would later recognise Taylor's face on warning fliers handed out in the local school. Meanwhile, Nigh lived in fear — not of Taylor, but of his own furious neighbours. "I thought I was going to get run out of town," he says.
Most of us would probably sympathize with the people carrying torches and pitchforks. In fact, laws in the U.S. and Canada are increasingly designed to exclude sex offenders from society, with residency restrictions, monitoring programs, community notification programs, and civil commitments that effectively keep people locked up forever, even after they have served their term in prison.
The plan Nigh came up with for Charlie Taylor was the opposite of all that. Since then, it's come to be known as Circles of Support and Accountability, or CoSA. The model is used internationally to reduce the chances that freed sex offenders will fall back into their old crimes. Two decades of research suggests it can be effective — more so, at any rate, than knee-jerk safety measures like notifying the public about sex offenders in their midst, which can actually increase the risk of reoffense.
But back when Harry Nigh decided to bring Charlie Taylor to Hamilton, he didn't have any proof that his idea would work. What he had was a belief that criminal justice should be restorative justice, helped along by a community. And he believed that every crime, even a heinous one, grew out of a hunger to be seen and replace something lost.
Today, a CoSA is very similar to the impromptu group that formed around Charlie Taylor in 1994. Four to six community volunteers dedicate themselves to befriending a recently-released sex offender who is deemed high risk. They help that person become a part of society by meeting regularly, helping the offender find work or a place to live, and inviting them to Christmas. These programs exist around the world, but they aren't wildly popular. In the U.S., for instance, only seven states have functioning CoSAs. In Canada, CoSAs recently lost federal funding.
CoSA is a program backed by evidence, yet governments are ignoring it in favour of laws that, by and large, aren't supported by evidence. It would be easy to frame CoSA's cause as a fight against the forces of anti-science. But it's also a great example of how implementing evidence-based policy can be messy in the real world.
A Circle of Friends
One of the first nights after Charlie Taylor arrived in Hamilton he went to a barbecue. It was a welcoming party in Harry Nigh's backyard, and Charlie invited the cops. Earlier that day, police officers had accompanied him to court, and he'd gotten to like them. The police did wind upcoming to the house, but not to celebrate. They wouldn't even come inside. Instead, they sat in their cars and watched. Charlie Taylor was under 24-hour surveillance.
Bob Maxwell was one of the officers assigned to Charlie. At the time, he was a staff sergeant in homicide, and he went to Charlie's house every morning to check in on him. Watching Charlie meant that Maxwell was also watching Nigh and the other parishioners who made up his circle. Suspicious at first, he wondered about their ulterior motives. Maxwell worried that they were naive, and would add stress to Charlie's life. They'd never spoken to the police about their ideas, and he had no idea who they were.
But he did know who Charlie was. "His family was dysfunctional," Maxwell says. "He was put into an industrial home when he was 5. Basically, jail for kids. But he'd done nothing wrong." Charlie Taylor was sexually assaulted there, numerous times, by several different people. When he became a teenager, he started doing the same things to other children. Most sex offenders are not the bogeymen we see on Law & Order. It's actually rare to find a sex offender who snatches and assaults children he doesn't know, who does this repeatedly throughout his life, and who can't function in normal society.
Charlie was the exception that proved the rule. According to Robin Wilson, a psychologist who has studied CoSAs since he worked with the Correctional Service of Canada on Charlie's case, Charlie Taylor was first sent to prison at age 18 and had been in and out after that — always for sexual crimes against children. He was 41 when Harry Nigh took him in. He likely had dozens of victims, boys and girls.
Everybody thought Charlie would reoffend. Maxwell thought so, as he followed Charlie to the barber and sent plainclothes officers to take down vigilante signs posted in Charlie's neighbourhood. Nigh thought so, as he listened to the harrowing stories of Charlie's childhood, watched Charlie try to warn the Road Runner about Wile E. Coyote, and held his hand late one night when his cat was hurt and all the vets were closed.
But Charlie never reoffended again.
The Myth of the Reoffender
Here's the thing, though. It turns out this is actually pretty normal. Though Charlie Taylor had a history of reoffense, sex offenders, as a whole, are less likely to reoffend than other types of criminals — even without any kind of help or treatment at all. It's hard to pin down exact numbers; rates differ from place to place and change over time, and you have to also remember that sex crimes are underreported. But a 2012 study that analysed 23 previously published research papers, incorporating the data of more than 8000 sex offenders, found that, after five years, you could expect between 4-12 per cent of sex offenders to reoffend. After 10 years, between 6 per cent and 22 per cent.
That complicates efforts to understand treatments that work and treatments that don't. Did Harry Nigh's Circle of Support and Accountability prevent Charlie Taylor from reoffending? Nigh believes it did. So does Maxwell, who eventually came to see the circle as a powerful force for good.
In 2005, Robin Wilson published a report for the Correctional Service of Canada that found CoSAs reduced the rate of sexual recidivism by 70 per cent. To Wilson, an atheist, this is a place where belief and logic intertwine. "Maybe there is something to having the faith to spend time with someone who won't get anything else," he says. "But there's also a certain evidence base to that, that we see all over the healthcare and impulse-control literature, that says people with problems always do better when they have strong support."
In the case of CoSAs, though, it's been difficult to prove that definitively. That's because they're hard to study. Because most sex offenders don't reoffend, you need to look at lots and lots and lots of cases if you want to see whether a given intervention worked — more cases than is usually feasible. Also, CoSAs are volunteer groups, not research programs run by a lab. So studying them usually means looking at past results, rather than experimentally following circles in real time. Both of those things result in a quality of evidence that scientists consider less reliable.
So far, there has only been one randomised controlled trial of CoSA — meaning that sex offenders who wanted CoSA help were randomly assigned to get it or not.
The study, conducted by the Minnesota Department of Corrections, took 62 sex offenders and split them into two groups, one that got CoSA and one that didn't. After three years, none of the people in the CoSA group had reoffended sexually. And only one person in the control group had. That difference was so small that the study had to look at other signs of success, like whether people were arrested or convicted of any new crimes, not just sexual ones, and at the cost-benefit analysis.
There, CoSA got results. There were 23 rearrests, for any offence, in the CoSA group and 51 in the control group. By reducing the societal and administrative costs of rearrest and conviction, the study estimated that the CoSA saved the state more than $15,000 per participant over three years.
Evidence and Belief
But there is evidence to support other kinds of interventions, as well. Some of these stand in direct opposition, ideologically, to CoSA. For instance, the same team that conducted the randomised controlled trial also did another study that found evidence supporting the effectiveness of civil commitment — a program of continuing to hold sex offenders after their prison sentences have technically ended, sometimes for the rest of their lives. Over four years, this program was found to reduce the sexual recidivism rate by 12 per cent. Civil commitment is expensive, and the programs in Minnesota and Missouri were recently found to be unconstitutional. But, especially for the highest-risk offenders, it could also work.
And that brings us back to the murky issue of personal belief. Evidence suggests CoSA works. Evidence suggests civil commitment works. All the evidence is flawed in various ways. Which you choose to pay more attention to probably has a lot to do with what you already believe to be true. Even when the cost-benefit analysis suggests that CoSA has big practical monetary benefits, belief still plays a role in deciding how much we value that money and what it's being used to do. You can come to an evidence-based conclusion that CoSAs are the best way to deal with somebody like Charlie Taylor. You can come to an evidence-based conclusion that civil commitment is preferable. And the fundamental disagreement between those two points of view is not an argument about science vs. anti-science.
The trouble with our conception of evidence-based policy is the idea that simply having evidence will automatically show us what policies we ought to choose — that all rational people will easily see a right side and a wrong side in that evidence. CoSA shows us that that isn't true. Evidence can tell us something valuable about the world, but it doesn't always make our decisions for us.
When Harry Nigh decided to start the first Circle of Support and Accountability, he made that decision based on personal values and belief, not evidence — and a lot of people in his community couldn't understand his choice. Now, we have evidence, but there are still a lot of people who wouldn't make the choice Nigh made. In fact, the researchers at the Minnesota Department of Corrections say that one of the biggest limitations to CoSAs is finding enough volunteers.
Charlie Taylor died in 2005. Harry Nigh found him on Christmas Day. He'd had a heart attack in his living room. The TV was still playing Cartoon Network. He had gone 11-and-a-half years without harming a child. But he never did have a big, cinematic insight about his own actions, Nigh says. It was all too knotted together with the abuse he'd suffered as a child, himself.
From Nigh's perspective, Charlie went to his grave feeling more wronged by what had been done to him than wrong in what he had done to others. And depending on what you feel in your gut, you will see that as proof for whatever you want to believe.
Illustration by Jim Cooke