Gun Violence, Profiling and Faulty Social Science

Photograph Source: Bob Dass – CC BY 2.0

Another day, another mass killing at a school.  This time Louisville.  But unfortunately and predictably there will be another mass shooting. Predictably there will be the same polarized political debate that results in a policy stalemate. And predictably the debate around guns will center on the usual stereotypes and myths that continue to confront mass killings in America.

The NRA will trot out that the problem is not too many guns or the type of guns but that there are not enough in the right hands.  It will join a US Supreme Court in contending that the Constitution protects our right to possess guns (especially handguns) in our homes and perhaps on the streets for self-defense. Progressives will demand new gun control legislation, calling for bans on assault weapons, purchase delays, and background checks.  Across these debates, facts take a backseat to myths.

Social science research can both inform and muddle the debate on guns in America and reinforce and dispel the myths or pop culture views informing what our policy responses and options should be.  Let’s use good research here.  What do we know?

The Problem of Mass Killings

Mass millings such as what happened in Louisville are tragic.  But mass shootings pale in comparison to other forms of gun violence.

Mother Jones maintains a database on mass killings, defined as four or more individuals killed in a public place. They began assembling the database in 1982 and updated it to Louisville. During that 41-year period, there have been 1101 killed, and another 1885 wounded.  Of the 141 incidents, 22 are school-related.  Sixty-six of the assailants had some history of mental illness.  Finally 99 of the incidents involved weapons labeled as semiautomatic and seven as assault.

Gun Deaths in America

Let’s do some comparisons.

According to the Center for Disease Control, in 2020, there were 24,292 suicides that involved guns.  In one year there were 22 times more suicides with guns compared to mass killings in America in 41 years.

Alternatively, the  Gun Violence Archive focus on four or more shootings, even if no one dies.  Since 2013 they calculate 144 mass shootings with the total dead and wounded at 7,740. One year of gun-related suicides is still more than 3.1 times the number of mass shooting injuries in a ten-year period.

Handguns are the gun of choice for 69% of males and 88% of the time for females when committing suicide.   In crimes, among the approximately 19,400 homicides in America in 2020, about 57% were probably with handguns.  Simply put, of the 45,222 gun deaths in America in 2020, less than one percent were mass shootings that took place with assault-type or semi-automatic weapons.

Finally, there is little evidence that guns are used for self-defense and instead are more likely to be used against another member of a household or for suicide than to thwart an intruder.

All gun violence is bad, but mass shootings with assault guns pale in comparison to handgun suicides and other forms of homicide. If we really wanted to affect gun violence in America we would be better served targeting resources on handguns and addressing the issues of suicide and the reasons why people kill in general.

The Myth of the Mentally Ill

The Mother Jones database alone should dispel the mentally-ill bias.  Sixty-six of the approximately 141 shooters had a history of mental illness, 17 did not, and the remainder is indeterminate.  Given these statistics, there is nonetheless a belief that mental illness is connected to mass shootings.  Yes, that is partially true, but this does not provide a complete picture of the connection between mental illness and gun violence.

Based on mass shooter research, there is a profile that such individuals are mentally ill and therefore we need to screen for this when it comes to access to guns.  Additionally,  in seeking to craft a profile of a mass shooter, studies seek to look at other variables, such as perhaps childhood trauma, unstable family structure, stress, or neighborhood characteristics.  There are problems in making these claims.

The first is a problem of reverse engineering.  By that, we do case studies of mass shooters and create a profile of who they are.  The goal here is to develop a screen and identify possible future shooters.

Reverse engineering here is dangerous.  What counts as trauma or unstable family structure, for example, is not well defined.  With that, moving from saying all mass shooters had X traits runs the same risks of stereotyping as do other forms of profiling. Profiling originated in efforts to predict who would hijack airplanes in the 1970s to eventually racial profiling at auto stops in the 1990s and then Muslim profiling after  9/11.

Not everyone who has a mental illness is violent or dangerous.  The National Institute of Health estimates that in 2021 57.8 million adults in the US have some form of a mental illness.  At most, only a small fraction are violent.   The American Psychological Association notes that if a person has a mental illness there may be other risk factors that are associated with violence, but mental illness alone is a poor predictor of violence.  Of those with a severe mental illness, 2.9% had committed a previous act of violence within the past four years compared to 0.8% for the general population.  Note the emphasis on severe mental illness, not all forms of mental illness.  For all individuals with all forms of mental health issues, barely one-percent are violent—about the same rate as the general population.

Moreover, a better risk factor tied to violence, according to the APA, is substance abuse.  The Center for Disease Control lists many factors that lead to violent behavior, including substance abuse, emotional problems, family, and neighborhood factors. But here is the problem, not everyone who had childhood trauma, grew up in bad neighborhoods, or faced stress is violent and becomes a mass shooter.  In fact, the vast majority of these individuals do not exhibit this behavior.

Using case studies with small samples to determine the profile of mass shooters and then making that a predictive tool to inform public policy is highly problematic.  It is not the problem of the ecological fallacy—extrapolating general population characteristics to make claims about individuals—but instead the atomistic fallacy of making incorrect assumptions about a population based on the traits of some individuals.  This is an inductive problem of rendering false generalizations.  It is also the classic mistake of confusing correlation with causation.  Mental illness may appear to correlate with mass shooters, but bad sampling and faulty assumptions render bad social science and predictions. It is using anecdotal and low-number case studies to render broader conclusions.


There are several conclusions.  Mass shootings are a problem but relatively insignificant in terms of gun violence.  Handguns are a far bigger problem than assault weapons.  Using mental illness as a predictor of gun violence or mass shooter profiles is under and over-inclusive of who commits such violence, and the same may be said of other forms of profiling.  At best we may have some factors that are associated with gun violence, but using them as a predictive screen is at best porous.

Contrary to Second Amendment defenders, we probably need to address the issue of handgun availability because simply relying on profiles will not address overall gun violence in America.  And contrary to those who focus on assault weapons in mass killings, this fails to capture the broader and more significant problem of gun violence in America and who commits it.

David Schultz is a professor of political science at Hamline University. He is the author of Presidential Swing States:  Why Only Ten Matter.