Do Guns Reduce Crime Through Posing a Threat to Criminals?

Last Updated on November 14, 2021 by Shaun Snapp

Executive Summary

  • Gun advocates often propose that guns reduce crime by serving as a deterrent.
  • How much evidence is there for this claim?


The claim is frequently made that guns reduce crime. What follows is an analysis of a claim to this effect, which relies upon an often quoted study to make this case.

Our References

See our references for this article and related articles at this link.

A Good Example of This Claim

An excellent example of this claim is found in the following Quora comment by James Lott.

Assertion #1: Armed Gun Owners Save Others and Themselves

How often does an armed gun owner stop an assault or save an innocent person?
Daily. The problem is you generally don’t hear about it.

It certainly does. But with a population of 330 million people, many things happen daily that are very low in incidence. And it is not mainly a problem that one does not hear about, although I have heard about people fighting off attackers with guns — particularly those that work in convenience stores. However, the number can still be estimated.

Validity of This Assertion: True, but Not Sufficiently Descriptive of a Claim


Assertion #2: The CDC Performed Research into Gun Violence?

Back when Obama was still president, if you’ll recall, a big “issue” was made of the fact that Congress had passed measures that stopped the CDC from doing research concerning firearms – something that was blamed on the NRA. Obama made a case to Congress to repeal that. When it was done, he then ordered the CDC to do so. He wanted more gun control and this was his method of going about it.

There is a lot about this that is odd. First, the CDC is responsible for diseases and mortality statistics, but they don’t have any expertise in gun violence. I use CDC murder and injury statistics later in this article, but that does not mean the CDC has expertise on what happens before the person is hit by a bullet. Remember, the CDC provides guidelines for how doctors should fill out death certificates. The CDC then tabulates the deaths and reports on them by different causes.

So getting back to the storyline of how this report came about.

Obama ordered the CDC to perform this analysis, but as the CDC has no expertise, (which it seems that Obama should have known, but apparently he did not know what the CDC’s preview was) so the CDC outsourced the job.

This is explained in the following quotation from the report.

The CDC and the CDC Foundation2 requested that the Institute of Medicine (IOM), in collaboration with the National Research Council (NRC), convene a committee of experts to develop a potential research agenda focusing on the public health aspects of firearm-related violence— its causes, approaches to interventions that could prevent it, and strategies to minimize its health burden. – Priorities for Research to Reduce the Threat of Firearm-Related Violence (2013)

One of the early problems with these commenters is that Lott continually refers to the final product being from the CDC, which, aside from being a pass-through, the CDC had nothing to do with. That is a clue that Lott does not have very high attention to detail when it comes to what he writes about what he reads. It states in the report that it was not produced by the CDC.

Validity of Assertion: Half True


What Is the Report and Who Was In Charge of It?

I have read a lot of research in my life. This report is 129 pages, and it is essentially a review of other research. This report is just a committee report that the IOM and NRC (National Research Council) convened some people in the field to perform. The National Research Council comprises the National Academy of Sciences, the National Academy of Engineering, and the National Academy of Medicine. The IOM is the Institute of Medicine.

The most important feature to understand about these parts of the government is that none of the entities have any background or interest in this topic.

This entire chain of involved parties does not seem to know what to do with this research request. This is not a comment on the Lott but a comment on the discombobulated nature of our government.


  1. LESHNER (Chair), American Association for the Advancement of Science, Washington
  2. DC LOUIS ARCANGELI, Georgia State University, Atlanta
  3. ALFRED BLUMSTEIN, Carnegie Mellon University, Pittsburgh, PA
  4. C. HENDRICKS BROWN, University of Miami Miller School of Medicine Miami, FL
  5. DONALD CARLUCCI Picatinny Arsenal, Rockaway Township, NJ BG (Ret.)
  6. RHONDA CORNUM, TechWerks, North Middletown
  7. KY PAUL K. HALVERSON, Indiana University Richard M. Fairbanks School of Public Health Indianapolis
  8. STEPHEN W. HARGARTEN, Medical College of Wisconsin, Milwaukee
  9. RONALD C. KESSLER, Harvard Medical School, Boston
  10. MA GARY KLECK, Florida State University, Tallahassee
  11. JOHN A. RICH, Drexel University School of Public Health Philadelphia, PA
  12. JEFFREY W. RUNGE, Biologue, Inc., Chapel Hill
  13. NC SUSAN B. SORENSON, University of Pennsylvania, Philadelphia
  14. DAVID VLAHOV, University of California, San Francisco

There are several problems with this list.

First, most of these people work in medical schools or schools of public health.

David Vlahov works out of the school of nursing presently at Yale. Nursing helps treat bullet wounds. They don’t specialize in understanding what drives bullet wounds.

The most qualified person I could find was Alfred Blumstein, who is a criminologist. I find it very likely that the vast majority of the report was written by just a few people on this committee.

As I was reading the profiles I found myself repeatedly asking  “what does this person’s background have to do with analyzing gun violence?”

Secondly, there are 14 people on this committee. Were there 14 people required to write and direct this report? And that was not the extent of the bloat. Look at the report managers and overseers, which are the IOM and NRC Staff.

  1. BRUCE M. ALTEVOGT, Senior Program Officer, Board on Health Sciences Policy
  2. MARGARET A. MCCOY, Program Officer, Board on Health Sciences Policy
  3. JULIA K. HOGLUND, Research Associate, Food and Nutrition Board
  4. KATHERINE M. BLAKESLEE, Global Program Advisor, Board on Global Health
  5. MORGAN E. HELLER, Assistant to the IOM President for Special Projects
  6. BRADLEY A. ECKERT, Research Assistant IOM Executive Office
  7. PATRICK W. KELLEY, Senior Board Director, Boards on Global Health and African Science Academy Development
  8. ARLENE F. LEE, Board Director, Committee on Law and Justice

That is another eight people. 8 + 14 = 22 people.

Looking at the report, I could have written this report by myself. The people on this committee are very senior and have full-time jobs. It is doubtful most of them spent much time on the report.

Getting Analysis from the Wrong Group of People?

Something else that struck me as odd is that the titles of some of these resources are not what would suit such a report. Several of these people, like Bruce Altevogt and Margaret McCory, work in Health Sciences Policy. Julia Hoglung works in food and nutrition. Health care and medicine is the only area that makes up the IOM and NRC.

The Punting of the Report Responsibility

So if we look at this report, it began with the CDC, then was punted to the IOM and NRC, which had no idea what to do with it, so they assembled a committee of people who are primarily from health care and also did not know what to do with it.

The report itself is just an overview of other research with requests for funding to perform some actual research.

The story of how this report circuitously came into existence is a more interesting story than anything in the report. A better approach would have been to find some prominent criminologists working in academics and restrict the report to just a small group of these.

Assertion #3: Did the Study Backfire?

First, I’m afraid I have to disagree that this was “a study.”

A study is when some new analysis is done or a survey is sent out. What the IOM and NRC committee of health care specialists produced was a report. And the report authors appear to agree with me, as they call the work a report in their discussion of it in the document, as the following quotation explains.

The project that is the subject of this report was approved by the Governing Board of the National Research Council, whose members are drawn from the councils of the National Academy of Sciences, the National Academy of Engineering, and the Institute of Medicine. Additional copies of this report are available for sale from the National Academies Press…

So the first problem is Lott’s continual assertion of framing that the report contains new insights. However, let us continue.

Lott states that the study backfired, as he states.

Interestingly enough, it backfired beautifully.

I do not understand why Lott says the study/report backfired. I don’t know, and I don’t see the evidence presented that Lott knows that the report was designed to be used to suppress gun rights. It does not appear designed to promote gun control. It certainly does not read that way.

Validity of Assertion: False or Exaggerated


Assertion #4: Did the “Study” Find that Between 500K and 3M CItizens Uses Firearms to Defend Themselves and Those Around Them?

Lott states the following:

The CDC found that 500k to 3 million US citizens used personally owned firearms to defend themselves and those around them. The problem is that the vast majority, well over 90%, never have to fire a shot to do so. Typically, criminals back off and retreat when faced with an armed citizen. If fact, criminals disengaging from the conflict when faced with an armed citizens happens far more often than even criminals who are facing the police. That this happens isn’t surprising.

“Why?”, you ask? Because even criminals are aware of the fact that armed citizens kill far more criminals than the police.

This section is a gross misrepresentation of what is in the report. The first thing to make clear is that this is not a report that has any original research. Therefore, this report could not and did not…

“find that 500k to 3M citizens USD personally owned firearms to defend themselves and those around them.”

It did not find this because it was not tasked with and did not perform primary research. This is a misrepresentation by Lott. This estimate referenced in the report is taken from a very old report which I covered in the article How the Number of Blacks Predicts Any State’s Murder Rate.

Let us review exactly what the report said to see how misleading Lott’s description is.

Defensive use of guns by crime victims is a common occurrence, although the exact number remains disputed (Cook and Ludwig, 1996; Kleck, 2001a). Almost all national survey estimates indicate that defensive gun uses by victims are at least as common as offensive uses by criminals, with estimates of annual uses ranging from about 500,000 to more than 3 million (Kleck, 2001a), in the context of about 300,000 violent crimes involving firearms in 2008 (BJS, 2010). On the other hand, some scholars point to a radically lower estimate of only 108,000 annual defensive uses based on the National Crime Victimization Survey (Cook et al., 1997). The variation in these numbers remains a controversy in the field. The estimate of 3 million defensive uses per year is based on an extrapolation from a small number of responses taken from more than 19 national surveys. The former estimate of 108,000 is difficult to interpret because respondents were not asked specifically about defensive gun use. – Priorities for Research to Reduce the Threat of Firearm-Related Violence (2013)

Did you notice anything about this quote versus Lott’s representation of the report’s “findings”?

First, there is a reference for the 500k to 3M estimate. That means this report did not “find” this. The report states that this is previous research. Second, you notice how there is a second estimate, which is 108,000? Might I ask why the commenter only listed the 500k to 3M number and not the 108k number? Could it be that Lott is cherry-picking items from the report to support a pre-existing bias?

Secondly, did Lott not find it odd that the range between the estimates is (500k + 3M)/2 = 1.75M versus roughly 100K, or a range of 17.5x? Didn’t that range strike the Lott as a bit large?

But let us now review the 500k to 3M number.

This study is by Gary Kleck of Florida State University and his colleague Marc Gertz. randomly asked 5,000 Americans if they or another member of their household had used a gun for self-protection in the past year. Just over 1 percent said yes, and the researchers extrapolated this percentage for the entire US population, giving them up to 2.5 million annual instances of defensive gun use.”

There are only 16,000 murders in the US every year (2020 was a very unusual year with 19,380 in part driven by the Geoge Floyd protests and the drive to defund the police). Secondly, the US population was a much smaller 266M in 1995 when the study was performed. Scaling up for the population in 2021 would be 330/266 = 1.24. 1.24 * 500K is 620k and 1.24 * 3M = 3.72M. The quote above shows the upper bound at 2.5M. Kleck performed a study in 1992 and then again in 1995. I think 3M is the upper estimate from 1992, and the 3M number is from 1995.

Let us take a mid-value of the numbers as scaled up for the larger population since 1995. That would be 620K + 3.72 / 2 = 2.17M.

Before I analyze this math, let us compare this to the other studies that have been one in this area.

Another study from the same period, the National Crime Victimization Survey (NCVS), estimated 65,000 DGUs annually. The NCVS survey differed from Kleck’s study in that it only interviewed those who reported a threatened, attempted, or completed victimization for one of six crimes: rape, robbery, assault, burglary, non-business larceny, and motor vehicle theft. That accounts for the discrepancy in the two results. A National Research Council report said that Kleck’s estimates appeared to be exaggerated and that it was almost certain that “some of what respondents designate[d] as their own self-defense would be construed as aggression by others” (Understanding and Preventing Violence, 266, Albert J. Reiss, Jr. & Jeffrey A. Roth, eds., 1992). DOJ study reported 83,000 annual defensive gun uses from 1987-1992. During same period, there were more than 135,000 total gun deaths and injuries in the U.S. annually. What is also interesting is that the study notes that “In most cases victims who used firearms to defend themselves or their property were confronted by offenders who were either unarmed or armed with weapons other than firearms.” Specifically, only 35% of those who used a firearm in self-defense actually faced an offender who had a gun. DOJ makes no judgments in this study on whether the level of force employed by these individuals was appropriate or consonant with the threat they faced. – The Virgina Center for Public Safety

So let us review the different estimates together.

  1. National Crime Victimization Survey (in 1995): 65,000
  2. DOJ: 83,000 annual defensive gun uses from 1987-1992 – This averages to 83,000 / 5 = 16,600 per year.
  3. Kleck/University of Florida: 1.75M per year (as of 1995)

Yet, the report presented this massive range without much discussion except to say that there is controversy. This makes me question the integrity of this report. Why did the report even take the Kleck study seriously? It is very lazy work to allow such a low integrity study into the mix, without commenting on its specifics and then allows such a massive range in estimates to go unevaluated. Furthermore, the report made it sound like there were only two studies, Klerk’s and the National Crime Victimization Survey, but there are more than this. And the bulk of the studies are grouped around each other with Kleck’s study being an extreme outlier. Extreme outlines need to be explained. I do not understand how a criminologist could have been on this committee and allowed that discrepancy to not be addressed.

The Illogic of the Math of the Kleck Study

The Kleck study does not seem realistic when one triangulates other statistics on gun violence. The idea is that 2.17M altercations are stopped by a person showing or using a weapon.

This is 5,945 altercations involving guns every day!

Let us review these statistics from the CDC’s National Center for Injury Prevention and Control’s Web-based Injury Statistics Query and Reporting System. There is variance year per year, but on average, these are the statistics they publish.

  • 14,062 are murdered
  • 76,725 people survive gunshot injuries
  • 34,566 are intentionally shot by someone else and survive

These are the stats that interest us. As we have to combine Item #1 and #2. So 14,062 / 365 = 38 per day. 34,566 / 365 = 95.

This means there are 38 people murdered with guns per day and 95 people shot, but that survive the event.

Let us combined these two numbers. 38 + 95 = 133 / day.

Now let us return to the estimate from the Kleck study, which is 5,945 per day.

This means that (if we accept Kleck’s study as accurate), there are 5,945 / 133 = 44x the number of altercations stopped by a gun owner than actual people shot and killed and shot and injured. Doesn’t that strike anyone as a bit disproportionate?

The Kleck study is one of the worst I have ever read. It places the survey participants in the position of guessing and as a proxy for first-hand experiences of others, with no way of verifying this experience. If Kleck were to work in sex research, he would determine sexual intercourse by having the study participants project how many times per week their neighbors are having sex.

And yet, the study is routinely quoted, not just by the NRA, but by normally conservative websites to support the idea that guns are commonly effectively used for defensive purposes.

Here is one of the titles of an article on the study.

The Criminologist Whose Self-Defense Research Destroyed Gun Control Arguments

In the process, he managed to shatter previous survey claims. He found that guns are used far more often for self-defense than they are used to commit crimes.

For every use of a gun to commit a crime, there are three to four cases of guns used in self-defense. –

Yes, he may have proved that — if Kleck’s study had any validity. However, it didn’t, so Kleck proved nothing. The primary insight from Kleck’s study is how people will repeat something, no matter how illogical, or how out of line with other estimates as long as it conforms with their biases.

Validity of Assertion: False and False on Multiple Counts

Assertion #5: Obama Found the Report Contradicted His Assumptions?

Now, why didn’t you hear about it? Because when the study was released, Obama read it. He read it and found out how wrong he was, and his answer to that was intellectually dishonest. Rather than addressing it and gracefully admitting he was wrong, his answer was to just ignore it and pretend he never said abd did those things. And it wasn’t just Obama that did so. So many far left leaning media outlets reported on his CDC address to Congress and blatantly supported him. When they study came out? All those media outlets? They didn’t address it either. They didn’t say anything at all – just ignored it – pretending it never happened.

I don’t know how Toll knows that Obama found out how wrong he was. What did he need to address? Furthermore, the report was just an overview and did not contain anything noteworthy. When I read it, I recognized the same things I had seen many times before. The report is not notable, so the way Toll presents the idea that a left-wing conspiracy somehow suppressed a very nondescript report because its contents were so illuminating is bizarre. I can’t tell if this is because Toll is himself new to reading research on gun violence — but the references within the text should have been a clue that the report was referring to previous work.

The report is not particularly interesting and as I have said already is not original. Anything the media covered in the report would just be repeating coverage on research that had been done in most cases many years ago. Furthermore, there are a lot of reports like this that receive little attention in the media. I have a number of much more interesting reports than this report that are little covered in the media. The media is much more likely to cover a study on how consuming three glasses of wine per week improves sexual endurance or reduces constipation than any government report.

Toll continues.

The study the CDC released? Its findings entirely contradicted every anti-gun sentiment. Their findings, that study – it ended up supporting the 2nd amendment.

I did not get that from reading the study.

Toll continues.

Now, there were a few that did report it – but even they ended up retracting their articles. Those focused on the singular finding that “The U.S. rate of firearm-related homicide is higher than that of any other industrialized country: 19.5 times higher than the rates in other high-income countries.” However, as noted by the CDC and even every single other unbiased properly performed study on the matter, if figures are excluded from such anti-gun bastions as Illinois, California, New Jersey and Washington, D.C., “The homicide rate in the United States would be in line with any other country.”

To be more specific, the homicide rate would be in line with Europe without blacks and Hispanics. These cities are not just “anti-gun bastions” for no reason because they are filled with liberals. Wherever there are large numbers of blacks or Hispanics, but in particular blacks, they restrict gun rights.

This is a dishonest attempt by a conservative to lay gun deaths on anti-gun laws. He has the causality reversed. This is demonstrable because blacks, with only 13% of the population, commit roughly 1/2 of all violent and gun-related crimes.

However, pro-gun advocates don’t like to bring this up, so instead, they prefer to call black cities “liberal,” which is what causes their gun violence. Well, Vermont is quite liberal, with a large percentage of the dogs in the state wearing tie-dye bandanas, the home of Ben and Jerry’s and Bernie Sanders, but gun violence is not an issue in the state as the state is 94% white and 1.4% black. So it seems states can be both liberal and very low in violent crime — but only if they are predominantly white. There have been calls to improve Vermont’s condition of low violent crime, by making it more “diverse.”

Toll repeatedly asserts that the IOM report is some type of groundbreaking item that has been suppressed. As evidence against this, I submit some of the most interesting quotes I could find from the report.

In 2010, incidents in the United States involving firearms injured or killed more than 105,000 individuals; there were twice as many nonfatal firearm-related injuries (73,505) than deaths. Nonfatal violence often has significant physical and psychological impacts, including psychological outcomes for those in proximity to individuals who are injured or die from gun violence.

Gun type and intended use vary; so do the manifestations of firearm violence. Some firearm violence results in death, but most does not. There are important disparities across socioeconomic and ethnic groups in overall mortality rates from firearm violence.

Further, there is substantial variation within each type of violence: suicide, homicide, unintentional injuries, and fatalities.

Inclusive of homicide, suicide, and unintentional death, African American males have the highest overall rate of firearm-related mortality: 32 per 100,000,8 twice that of white, non-Hispanic males (at 16.6 per 100,000),9 and three times that of Hispanic and American Indian males (at 10.410 and 11.811 per 100,000, respectively). The rate of mortality by firearm for Asian/Pacific Islander males is 4.2 per 100,000.12 The rates of mortality for females are much lower, ranging from a low of 0.6 per 100,000 for Asian/Pacific Islander females13 to 3.3 per 100,000 for African American and 3.0 for white, non-Hispanic females.14

Even when defensive use of guns is effective in averting death or injury for the gun user in cases of crime, it is still possible that keeping a gun in the home or carrying a gun in public—concealed or open carry— may have a different net effect on the rate of injury. For example, if gun ownership raises the risk of suicide, homicide, or the use of weapons by those who invade the homes of gun owners, this could cancel or outweigh the beneficial effects of defensive gun use (Kellermann et al., 1992, 1993, 1995).

Most firearm-related deaths are suicides. Fifty percent of suicides are by firearm and 60 percent of firearm deaths are suicides (Law Center to Prevent Gun Violence, 2013). Research demonstrates that the proportion of suicide by firearm is greater in areas with higher household gun ownership (NRC, 2005). Further, two studies found “a small but significant fraction of gun suicides are committed within days to weeks after the purchase of a handgun, and both [studies] also indicate that gun purchasers have an elevated risk of suicide.

I will allow the reader to decide how different any of this information is from any other report on the topic.

Validity of Assertion: Insufficient Evidence Presented to Support the Claim


The Kleck study was very poorly constructed and should be dismissed as providing any insight into the use of guns for defense.

  • First, the study does not hold up to comparative analysis versus other statistics on gun violence in the US.
  • Secondly, its estimate is extremely large versus other estimates — yet the other estimates are rarely mentioned.

The Kleck study is repeated as solid evidence for gun ownership and the 2nd Amendment. Gun advocates do not need to rely on personal defense arguments for the right to keep and bear arms. In the article The Meaning of the Militia: Analysis of Hamilton, Jay, and Madison’s Federalist Papers, I cover that the 2nd Amendment is not based upon the argument of personal defense. This is a major blind spot for gun advocates. The 2nd Amendment is a special declaration of the right to keep and bear arms which is were put into place for the common defense and putting down insurrections, which were to both be performed by state militias. It is separate from the general right to keep and bear arms that existed for over a hundred years prior to the Bill of Rights.

This personal defense argument has been inaccurately proposed as historical support for the 2nd Amendment, but also for general gun rights by the NRA, who have also been promoting the Kleck study. The Kleck study, nor any similar study is required to support either the special or general right to keep and bear arms.