Overused Statistic #738: Mental Illness and Violent Victimization

So there’s a particular bit of overused truth whose use I want to challenge. Again, it’s not that it’s not true. AND it’s not that we shouldn’t be telling people that it’s true, BUT it seems to only ever be used in contexts where it doesn’t mean what people think it means.

Persons with mental illness are more likely to be victims of violent crime than to be perpetrators of violent crime. This is true.

Persons without mental illness are also more likely to be victims of violent crime than to be perpetrators of violent crime. This is also true.

But wait! How is that possible?

It is possible – and, my two blessed readers, it is actually true – because people who commit violent crimes often commit more than one and against more than one victim over the course of a lifetime. If the average person willing to commit violent crime does so against not only one person but against two, then on average and without victim overlap there are twice as many victims in the population as a whole than there are perpetrators. Given what we know about how domestic violence works, with many persons surviving violent relationships and escaping to new (hopefully healthier) relationships, we can confidently say that abusers get into new violent relationships after old violent relationships end. This doesn’t even consider children who might also be victimized.

Even if the only violent criminals who ever victimized two or more different people were child abusers who had multiple children and domestic abusers who start new relationships after the old ones end, we would have more victims in society than violent criminals. But of course muggers and pimps and bar-fighters and mass shooters and serial killers all exist as well and frequently (or even by definition) victimize more than one other person. Of course, we could also consider victim overlap, but as long as victimization rates are below 100% then at least some victims after a perp’s first are going to be persons never previously victimized. This can bring the pool of victims down closer to 1:1, but in the case of crimes of the nature of child abuse and domestic violence, it will never reach 1:1, much less a state where there are more perps than victims. This is important because while certain crimes (e.g. a gang attack on an isolated individual) can have a nature that creates more perps than victims, the most common violent crimes in our society are adult assaults against children or adult assaults against a domestic partner (including, but not limited to, a spouse). This disparity in types of violent crimes guarantees that we will always have more victims than perps in society generally and in every large demographic group.

So when someone says, “People with mental illness are more likely to be victims than perpetrators,” they are saying something so duh, obvious! as to be completely meaningless. It’s not that this statement is wrong, but if you’re relying on this statement to, for instance, tell people that they shouldn’t be blaming violence on the mentally ill (which is very often when this bit of truth is deployed) then any reasonable person will immediately discount your argument.

After all, this is what people are saying:

  1. There are more victims among the mentally ill than perpetrators among the mentally ill.
  2. One should not blame the victim (this is, often, silently implied rather than stated but is necessary to the argument)
  3. Therefore one should not blame the mentally ill.

However, what if statistics showed that every single person who commits violent crime is mentally ill? What if, simultaneously, those mentally ill perps each have an average of 4 victims, 2 mentally ill and 2 not? Since the population of non-mentally ill persons is larger than the population of mentally ill persons, we can then conclude that:

  1. The mentally ill are far more likely to be victims than perps
  2. The mentally ill are far more likely to be victimized than the non-mentally ill are to be victimized. In other words, the mentally ill suffer overwhelmingly disproportionate effects of societal violence.
  3. There is no public policy at reducing violence among the non-mentally ill that could ever have any positive effect, since all perps are mentally ill.
  4. All prevention measures should assume that any perp is mentally ill (since they all are).
  5. Reasonable prevention measures could be aimed at reducing mentally ill persons’ access to weapons.
  6. It is neither irrational nor stereotyping to portray perps as mentally ill without investigating the specific background of any specific perp.

It thus drives me crazy to see writers deploy, “the mentally ill are far more likely to be victims than perps” and “the mentally ill are far more likely to be victimized than the non-mentally ill” (numbers 1 & 2 above) to counter messages like 3-6. Without a showing that non-mentally ill people commit violence statements 1 & 2 are entirely consistent with the unjust bullshit of statements 3-6. We need to counter 3-6 because they are wrong. They are unjust bullshit. But neither statement 1 nor statement 2 does anything at all to prove that 3-6 are unjust bullshit.

It was Wonkette writer Stephen Robinson (whom I generally respect) that set off this particular incarnation of this rant, when in concluding an article about Dana Loesch’s most recent craptastic media appearances, he wrote this:

Mentally ill people are more likely to be victims of violence than perpetrators. Loesch and the gun-obsessed are looking for scapegoats. They won’t accept the simple solution, which is restricting access to guns. Literally locking up the “weird kid” in school isn’t going to prevent gun violence. It’s also a far greater infringement on civil liberties than flat out getting rid of guns. Besides which — and you never ever ever saw this coming — the NRA has been fighting “red flag laws,” which temporarily remove guns from people who may be a threat to themselves or others, for as long as it has been arguing the real problem is crazy people.

But how do we know that locking up the weird kid isn’t going to prevent gun violence?

We know that not because mentally ill people – like every large demographic group – contains more victims than perpetrators. We know it because extensive research shows that persons with nearly any diagnosed mental illness are about as likely to commit violence as anyone from the general public when not under treatment and less likely to commit violence than anyone from the general public while under treatment.

The only exceptions are one large category, persons with an addiction to an intoxicant, and a couple categories that are much smaller than most expect. Even narrow categories such as “schizophrenic persons” are not, on average, more violent than the general public. Rather, there is a small subset of delusional persons where two different types of delusions co-occur that are found to have increased rates of violence. The types of delusions are persecutory delusions (someone is out to get me/ people I care about/ the whole world) and grandiose delusions (I can accomplish things no other person can). Typically these do not co-occur. “I’m the greatest painter ever!” doesn’t usually co-occur with, “and the Nixon family is secretly paying art collectors not to buy my work.” There are many people who are delusionally full of themselves, some of those even to the point of mental illness. This does not in any way indicate that it’s likely that they also feel persecuted, especially when they are not persecuted and are deluded about the existence of that persecution. There are a few different disorders characterized by delusions, and I’m not expert in psychology generally much less any of those particular disorders, but as far as I know any of them has the possibility of manifesting both grandiose and persecutory delusions. Until someone does manifest both types, however, research has shown that they are no more likely than the general public to commit violence – and less likely while under treatment.

This “less likely while under treatment” bit is really quite important. Why? Because we often look for treatment as a red flag, but it is apparently just the opposite, a green one if you will.

In the meantime, someone will inevitably bring up personality disorders. There is some disagreement about whether or not personality disorders are mental illnesses, but historically there was a distinction made and for a very important reason. Personality in psychology is simply the sum of your observable tendencies in choice making. Personality disorders are diagnosed from your tendencies in choice making. Personality disorders that show increased rates of violence all have acts of violence as an example of a particular type of flawed choice making used to diagnose the personality disorder.

In other words, personality disorders don’t result in violence. We simply call choosing to commit violence an example of bad choice-making and use it as the basis for a concept we call a personality disorder. The violence is thus the cause, not the effect.

It is only this set of facts, not the rather banal and universal truth that the mentally ill, like all other large demographics, are more victims than perpetrators, that allows us to rationally say that focussing on the violence committed by the mentally ill wastes effort and produces injustice.

So the next time people are talking about who commits violence and does so by blaming the mentally ill, don’t change the subject to who is victimized unless and until you’ve corrected the record about who actually commits the violence. The disproportionate victimization of the mentally ill is both true and important for public policy to address, but if we only discuss that, then we effectively leave unaddressed a lie.

Mentally healthy people commit the vast majority of violence, and mentally ill people who access treatment are actually less likely to commit violence than any random neighbor on any random street. Before we start mentioning who is victimized, let’s mention that: because if we ever managed to shift violence so that it no longer disproportionately targets the mentally ill, every single act of violence would remain unacceptable.

The focus should always first fall on who is committing the violence and why, or people are going to continue to believe that folks with mental illness are dangerous, however much we are also victims.



  1. Pierce R. Butler says

    “Mentally ill” – a category including the dangerous, the socially neutral, and the vulnerable.

    Rather like “Floridian”, except we probably constitute more of a menace, taken in toto.

    We have an over-/false- generalization problem: a mental error with unhealthy consequences, though not a “mental health” issue as such, somehow.

    Mentally healthy people commit the vast majority of violence…

    Lots of redefinition seems called for here.

  2. ridana says

    This disparity in types of violent crimes guarantees that we will always have more perps than victims…

    I can be pretty dense, but isn’t this backwards? It seems to run counter to everything you said preceding it in the paragraph.

  3. EnlightenmentLiberal says

    To Crip in OP.
    Thanks. Well said. Remember kids – statistics is hard.

Leave a Reply

Your email address will not be published. Required fields are marked *