Mental Illness and Violence: About that MacArthur Study…

Ruth Johnston
10 min readApr 26, 2019


Image by Clker-Free-Vector-Images from Pixabay

I’m talking with Lynn Nanos, a social worker specialized in evaluating serious mental illness. Lynn has worked in emergency care, including in-home evaluations, for over ten years, and recently she wrote Breakdown: A Clinician’s Experience in a Broken System of Emergency Psychiatry. Not many professionals are willing to speak out about the need for reform, so she is such a treasure and a resource for the rest of us. In our previous conversation, we talked about the reasons why it’s so hard to get help for someone with severe mental illness. In this chat, I want to focus on a chapter in Breakdown that looks into a famous study that’s often cited, the MacArthur Violence Risk Assessment Study conducted in the 1990s. [See also our third chat, “Psychiatric Treatment Culture: The Effects of Losing Hope”]

Ruth: One of the first things the internet will tell you about mental illness and violence is that there’s no significant link. In fact, site after site will tell you that “studies show” there’s no real link. And yet my personal experience, at least, says there is a very strong connection. My son was not violent or angry, which is why we were so shocked when he actually killed my mother. The sole reason was that his delusional beliefs had constructed a world in which his action made sense. His neuropsychiatric illness, and nothing else, caused his violence. Lynn, in your work you have met some of the sickest of the sick [click here to read about our thoughts on psychiatric treatment culture]. How have you observed the link between delusions and violence?

Lynn: I like how you described Levi’s illness as “neuropsychiatric.” Though there’s still a split between neurology and psychiatry, we know of structural differences between the brains of healthy people and those with schizophrenia. Yes, delusions can result in violence. I recall a case in which a psychotic patient violently held staff members hostage at his state-funded residential program because he believed doing this would prevent government officials from harming them. When interviewing psychotic patients to determine if they qualify for inpatient, I look closely at the content of the delusion. Some delusions will not result in violence while other delusions will. Persecutory delusions can be dangerous if the patient feels the urge to defend himself or herself against attack, especially when the patient begins to believe people are not really who they are. For instance, a patient believed that I was a childhood friend of his, despite that I introduced myself.

Ruth: One part of Levi’s delusion was that he didn’t think his grandmother was really herself anymore. He thought some hostile evil spirit was living in her body. It seemed obvious to me that the belief itself might lead to harm, whether he had made past threats or not.

Lynn: Indeed, William Bruce killed his mother, Amy, because of his belief that she was part of al-Qaeda. Breakdown details their tragic story.

Ruth: In Breakdown, you told us some surprising facts about the major study that usually gets cited to say that there’s no link. Let’s start by saying that the MacArthur Foundation did a large, expensive, carefully planned study in the early 1990s. Among the things they got right, they checked people’s self-reports against other people who knew them, and against public records like arrests. That’s great. But they concluded that mentally ill people are not more likely to commit violence than most people, and for a lot of folks, that’s the last word.

Lynn: When I testified in favor of Assisted Outpatient Treatment (mandated outpatient treatment usually ordered by courts) at the Massachusetts State House, a powerful government official used the MacArthur study to oppose it.

Ruth: What happened? How could this Cadillac of studies conclude something that seems so wrong compared to our personal experience?

Lynn: If you review the study closely, it’s obvious that the participants most prone to violence were not chosen.

Ruth: You said in Breakdown that they started by recruiting people who were being released from inpatient stays at three major psychiatric hospitals. Of course, some people would just say no thanks to a year-long intrusive study. What do we know about who said “yes”?

Lynn: Well, the majority of the participants were not psychotic. And common-sense tells us that those not willing to participate more likely don’t recognize that they’re ill, and thus are more sick than those who cooperate.

Ruth: So they had major depression, for example?

Lynn: In fact, depression was the most common disorder among the study participants. There is no link between depression and violence toward others! And among the psychotic patients, the study excluded the schizophrenic patients who could not be found during the follow-up. These patients were more likely to be violent than those who cooperated with the researchers.

Ruth: So they had the initial data, but then these people just dropped out. We have no idea if they did anything violent later. They must have had some level of homelessness to just vanish.

Lynn: Then the researchers did not include prisoners or recent prisoners, anyone jailed, anyone involuntarily engaged in treatment, anyone in forensic hospitals, or anyone committed to extended-care inpatient units. These populations tend to be more violent than other populations. Subjects were excluded if they had been hospitalized for at least twenty-one days. Those who are most ill and violent often need more inpatient time.

Ruth: So basically anyone who had been arrested recently was excluded? Nor did they include anyone whose family had to press to get them admitted to the hospital against their will?

Lynn: That’s right.

Ruth: It sounds like the study recruited people who had a pretty good understanding of their problem and were probably on medication if they needed it. How did this change the study’s outcomes?

Lynn: They selected patients who were least likely to become violent.

Ruth: Additionally, you wrote in Breakdown that the MacArthur study excluded incidents of violence that did not result in bodily harm. In other words, if a victim successfully ran away from an attempted assault, this incident was not counted. The researchers also didn’t consider incidents of property damage. Punching through a wall didn’t count, nor did bashing someone’s windshield or setting a fire. We don’t know if these things happened, but if so they weren’t counted — and yet these are things most of us would consider pretty scary violence.

Lynn: For example, Eric Bertrand, homeless and showing signs of paranoid schizophrenia, set a fire in the Forbes Library in Northampton, Massachusetts, and the fire resulted in an estimated $100,000 worth of damage. Eric would not be considered violent in this study.

Ruth: Unbelievable. Okay, so over the course of a year, they checked in with the people in the study, asking if they had committed any violent acts. They were diligent to check on the truth of these statements, which is praiseworthy. Then they compared their study subjects with a group of people who had not been hospitalized or diagnosed with a neuropsychiatric illness. How did they find this group of people?

Lynn: They chose people who had not been hospitalized but who were from the same neighborhoods where some of the study subjects lived.

Ruth: That sounds like a good idea, since it meant they were comparing people who were similar except for having a mental illness.

Lynn: But these neighborhoods were economically impoverished, and many seriously mentally ill people lived there. Additionally, drug abuse was prevalent in these neighborhoods in the 1990s, and countless studies have shown that substance-misuse increases violence.

Ruth: Why were so many of the study subjects living in impoverished, dangerous neighborhoods? We know that brain illness doesn’t only strike the poor, it can show up anywhere.

Lynn: It’s true that mental illness affects all socioeconomic classes, but there are reasons why mental illness leads to poverty. The onset of serious mental illness is usually in either late adolescence or early adulthood. Delays in such illness getting sufficiently treated contribute to lack of normal functioning. Most people don’t have the luxury of using family funds to purchase homes, cars, clothing, healthcare, food, and everything else that are so easily taken for granted. Untreated or insufficiently treated serious mental illness interferes with the development of major milestones. Breakdown describes Lily, who dis-enrolled from Dartmouth College because of her psychosis. Concentration in college classes becomes impossible. Erratic sleeping patterns that are so common in mental illness lead to too many missed classes. When grades plummet, either the student or administration decides it’s not working out. Lack of higher education usually correlates with low wages. Regardless of educational background, mental illness can also interfere with the ability to appropriately perform in employment. Disability payments through Social Security are minimal, and Section 8 housing is often in poor neighborhoods. It just adds up.

Ruth: Also, I think work experience determines the amount of payment from the government. So doesn’t that mean that someone who is struck with schizophrenia hard and early in life will receive lower payment than someone who develops an injury, leaving them unable to work, later in life?

Lynn: Exactly. Lack of lots of employment experience means less Social Security income, which is typically just enough to cover basic necessities, such as medication. If people receiving this improve their ability to function to a point where they want to gain employment, the Social Security Administration limits the amount of money and number of work hours. Exceeding these limitations risks loss of benefits, including government-funded health insurance.

Ruth: I see. So, I’m thinking about the study’s design and how control groups work. You want them to be the same as the study subjects except for one factor, which is the thing you’re studying. The MacArthur study could have set it up differently, like choosing one random person who was matched to every study subject in age, sex, race, and educational level, but that random person might live anywhere —right? And that would have been another way to do it.

Lynn: Yes. But instead, they chose people only from the same neighborhoods.

Ruth: All right, now for the big reveal: do we know what the violence rates for the study group actually were?

Lynn: During this year, 27.5 percent of the participants engaged in physical violence toward others.

Ruth: That’s one in four! Most of us would call that a fairly poor record of peacekeeping. And the community control group?

Lynn: There was not much difference in rates of violence between the control and experimental groups. And worse, the MacArthur team did a follow-up one year later: over 50 percent of the study subjects had engaged in “menacing behavior” by then. There was a strong association between violence and the continuation of delusional thinking.

Ruth: When people quote “no more likely to be violent than anyone else,” I doubt they realize that the real numbers were so discouraging. There are reasons most of us try not to live in neighborhoods where 25 percent of the residents engage in violence! So here’s a question: has anyone done a similar study in which they only include people who are from safe neighborhoods?

Lynn: I don’t think so, though there are a number of other studies and meta-studies (studies of studies). One survey of over 200 studies suggested that psychosis, particularly psychosis expressed as delusions, may cause a 49 to 68 percent increase in likelihood of violence.

Ruth: The effect of ignoring the link between some brain illness and some violence is that public policy will then ignore it and, in the end, allow it to happen more by not preventing it. I think they’re afraid that if we say openly that untreated psychosis leads to violence, maybe the public will be mean to those with mental illness. Or, perhaps worse, that the public would demand that state hospitals should get larger budgets and add more beds. It has to be a lot cheaper to just insist that there’s no real link. Nothing to see here, folks, move along. Then the cost is shifted to homeless shelters, jails, and scared families.

Lynn: I think that those who ignore the link between brain illness and violence are afraid that stigma will be confounded. They seem to embrace the perspective that any connection between mental illness and violence should be downplayed or denied. Perhaps they believe that stigma is the biggest barrier to treatment, having been influenced by the massive amount of anti-stigma campaigns.

Ruth: By “stigma,” they mean “a sense of shame and unacceptability of admitting you have this illness”? Something like that?

Lynn: Yes, stigma suggests that the mentally ill person caused the illness.

Ruth: Ugh. That’s definitely not true.

Lynn: But I think stigma has been overemphasized as a problem, creating the illusion that it is the most significant barrier to accessing treatment. But we know that anosognosia, lack of awareness of being ill, is the most common reason for treatment nonadherence. Inaccurately identifying the problem stigmatizes the mentally ill population. Rather than promoting the mantra that the mentally ill are no more likely than the rest of the population to become violent, I propose we emphasize that there exists some link between violence and mental illness, especially among those with untreated psychosis and anosognosia.

Ruth: That’s perfect. We want to shout to the world, “Hey, we’re talking about untreated illness!” I don’t want to increase stigma — I want to increase treatment, giving families more power to insist that people who aren’t rational enough to make choices shouldn’t be given the choice to stay sick. Lynn, thank you so much for your time and attention. Readers, there’s so much more where this came from in Breakdown. It’s a great sourcebook for advocates like me, and of course I find the stories fascinating. Lynn and I also have another conversation about the way our current laws have shaped the culture of psychiatric treatment.

Link: “Five Reasons Why We Can’t Get Help for Mental Illness: Interview with Lynn Nanos

Link: “Psychiatric Treatment Culture: The Effects of Losing Hope



Ruth Johnston

I'm the author of Re-Modeling the Mind: Personality in Balance; and sometimes I write from family experience about better ways to treat schizophrenia.