Less Guns, Less Crime- Debunking the Self-Defense Myth

Posted On August 21, 2013
539264_10151555914761833_416157604_n[1]
August 21, 2013

Myth:

Opponents of concealed-carry laws, such as Mayor Michael Bloomberg of New York, insist that the presence of armed citizens at mass shootings would result in even more deaths because of crossfire. But armed civilians prevented mass shootings on many occasions – in schools, a mall, and other public places – and there is no instance on record of a permit holder’s accidentally shooting a bystander – John Lott

None of the criminologists or economists who have studied concealed-carry handgun laws has found an increase in murder, suicide, or accidental death. Indeed, the vast majority of studies have found the opposite. – John Lott

An armed society is a polite society. Manners are good when one may have to back up his acts with his life. – Robert Heinlein

  • We see it on the news every day, yet another instance of some heroic armed citizen forcing criminals to flee from a robbery, a woman in a dark alleyway able to protect her virtue against a man twice her size, a husband protecting his wife and children from armed intruders.
  • Over the past several decades, the number of guns Americans own has skyrocketed. Simultaneously, gun violence has receded to levels not seen since the 1960s.
  • This contention rests on the idea that guns are the best means of defense and are a deterrent against criminals.
  • This idea is based heavily on research done by John Lott, whose research indicated that concealed carry laws reduced violence and overall crime, and that states with higher levels of gun ownership are less violent.
  • Linked to the idea that an armed society is a polite society.

Reality:

One of the most common assertions in defense of firearm use is the argument that guns are used millions of times a year in self-defense and are therefore effective crime deterrents.  There are two theoretical mechanisms that could explain this effect: 1) an attempted crime may be stopped due to the presence of a firearm; 2) a potential crime may be deterred due to a criminal’s caution in pursuing areas with high firearm ownership rates. As it turns out, neither of these explanations bear out in the data.

Why Guns Per Person is the Wrong Way to Think About Gun Control:

Before debunking both of the mechanisms outlined above, first let’s be clear about the proper metric in discussing the relationship between guns and crime.

As the graph above shows, the number of guns in the US has increased dramatically over the past few decades. This increase in the number of guns has corresponded to a decrease in violent crime. Therefore, gun advocates conclude that more guns = less crime.

However, this thought process is merely an extension of deterrence theory. We illustrated that the MAD Theory of Guns is incoherent when applied to mass shootings, and it is just as illogical when applied to violent crime.

First, let’s examine the apparent contradiction between two ostensibly incongruent facts: the number of guns in the United States has been increasing, while the gun ownership rate is decreasing. This is easily explained by the fact that many if not most gun owners buy more than one gun. Not only do hunters need several types of hunting rifles to match their prey, but also a large segment of the gun-rights advocate population has been stockpiling arsenals for a while now in preparation for social unrest or government despotism (I find neither of these arguments particularly convincing myself, though I might be persuaded about the necessity of preparing for the Zombie Apocalypse).

A person owning one gun is just as dangerous and likely to commit a homicide as a person with a dozen guns, ignoring that owning multiple guns might be a proxy for other sorts of behavioral characteristics that might be related to criminality.  What we see in the data is that the United States has reached a saturation point with guns. Most people who are going to buy a gun for whatever reason have already bought one, and buying a second does not increase the gun ownership rate.

Think of it this way, 100 people with one gun each is much scarier statistically speaking than 1 person with 100 guns.This is why the gun ownership rate is the statistic to worry about much more than the overall number of guns. And to further demonstrate that it is highly unlikely that the increase in the number of guns caused the decrease in the homicide rate, consider this: If the deterrence theory is correct (which I will demonstrate is not the case), a person with one gun is just as likely to stop a crime as a person with 100 guns. The extra 99 guns don’t help.

It is merely the fact that a person owns a gun, not how many, that matters with regard to the crime debate. As gun ownership has not increased in tandem with the number of guns, it is not possible for the increase in guns to have contributed to the decrease in violent crime. The only effects that can stem from this surge in guns are deleterious. With hundreds of thousands of guns stolen every year, the stockpiling of weapons only increases the likelihood that they end up in the wrong hands.

To be clear, then: though the “number of guns” has been increasing, this is not a useful metric in empirical analysis when talking about crime. It is the per capita ownership rate (or household ownership rate), not the total number of guns that matters most.

The NRA hasn’t updated its sources since 1995:

It’s important to highlight that the only academic source which substantiates the claim that guns are used millions of times a year in self-defense is a 1995 publication by Gary Kleck and Marc Gertz from Northwestern. This paper forms the entire basis behind the National Rifle Association’s talking point that “Americans use guns for self-protection as often as 2.1 to 2.5 million times a year.” Even if this paper had statistical merit, it’s rather curious that the NRA, Congress, and the media are still using a statistic computed from data that starts in 1981, the peak of the post-Vietnam crime wave, and is clearly inconsistent with a modern reading of literature on the subject .

However, even if you felt that a paper from 1995 still has social relevance, you should know that the entire paper has since been eviscerated by scholars who have pointed out that Kleck and Gertz’ paper suffers from errors so severe that their entire estimate is useless.

For example, according to Kleck and Gertz’s paper, guns were used in self-defense in 845,000 burglaries. However, from reliable victimization surveys, we know that there were fewer than 1.3 million burglaries where someone was in the home at the time of the crime. From surveys on firearm ownership, we know that 42% of US households own firearms (at the time of the survey), 33% of which contained occupants who weren’t sleeping at the time of the burglary. In order for the burglary statistic cited in Kleck and Gertz’s paper to be true, burglary victims would have to use their firearm in self-defense more than 100% of the time. Or, burglars could only rob houses of gun owners, and those gun owners would have to use their weapons in self-defense every single time they are awake.

Another absurd conclusion made in order to get to the ‘2.5 million’ claim, is that victims of rape and robbery are more likely to use a gun in self-defense than the offender is to use a gun against the victim. This clearly is impossible as most citizens don’t carry guns with them, and criminals choose the time and place of an attack.

David Hemenway, an economist and director of the Harvard Injury Control Research Center, concludes that “one small reason to expect even a tiny percentage of responders to over report (the number of times they used a firearm in self-defense) may be enough to lead to a substantial overestimate.”  Indeed, not one scholar since has been able to externally validate any of the claims made by Kleck and Gertz.

Guns are rarely used in self defense:

Now that I’ve given ample reason to be skeptical of the basis upon which the self-defense position is founded, let’s look at actual data on firearms used in self-defense.

To challenge the first way a firearm may potentially decrease crime, that is, attempted crimes may fail to materialize due to the presence of a firearm, we will use the most recent evidence from four national data sets, as provided by the Violence Policy Center.

In 2010, according to the most recent data on justifiable homicides from the FBI’s Uniform Crime Reporting (UCR) program, there were 230 justifiable homicides involving a private citizen using a firearm in self-defense during either an attempted or a completed crime. In the same year, there were 8,275 firearm homicides. This means that, for every one justifiable firearm homicide, there were 36 criminal homicides. Contrary to the gun lobby’s claim that, between 2007 and 2011, guns were used 12.5 million times in self-defense, the most reliable data on this question clearly show that firearms were used only 338,700 times in self-defense, and this includes off-duty police. Clearly, then, despite living in a country with 300 million guns, the use of firearms in self-defense appears to be an exceedingly rare phenomenon.

Two other areas of crime—violent crime and property crime—are analyzed by the Violence Policy Center, and cast serious doubt on the argument that guns are used regularly in self-defense. Between 2007 and 2011, only 0.8 percent of violent crimes involved the intended victim using a firearm in self-defense. During the same five year period, only 0.1 percent of attempted or completed property crimes involved the intended victim using a firearm in self-defense. Given that between 40-45% of American households own a gun, and less than 0.1 percent of victims of property crime end up using a gun to stop a crime, it’s impossible to suggest that guns are being effectively used in self-defense. Rather than guns serving as a useful deterrent, they instead helped to directly facilitate crime: 232,400 guns were stolen each year from U.S. households between 2005 and 2010.

You might be thinking, as many of our readers did, that 338,700 defense gun uses in a five year period (DGU) is not ‘rare’, and that, according to our own data, guns are used in self-defense 7 times as much as they are used in criminal homicides.  As Devin and I have made clear in the comments section, however, this comparison is disingenuous:

The 338,700 DGUs is primarily composed of self-defense uses in which the victim would not have died in the absence of a gun.  We know this to be true because the majority of firearm violence in which the victim is unarmed does not result in the death of the victim.   It should be clear, then, that comparing defensive gun uses to offensive gun deaths, commits a category mistake because you are comparing variables of different magnitude.  Either uses should be compared to uses, or deaths should be compared to deaths.

Thus, two potential comparisons are more appropriate:

  1. Justifiable homicides vs criminal homicides – (deaths vs. deaths) the reason this is more accurate is that both result in a guaranteed death, and justifiable homicides function as a serviceable substitute for situations in which the death of the victim is more likely in the absence of a gun. In this case we have 230 justifiable homicides in 2010 vs 8,275 criminal homicides in 2010, which clearly weighs in favor of gun control advocates.
  2. Defensive gun use vs crimes committed with a firearm -(uses vs uses).  Again, guns were used in self-defense 338,700 times between 2007-2011.  In that same five year period, there were 2,277,000 crimes committed with a firearm.  Let’s be clear about this: every time a gun was used in self-defense, a criminal committed fatal or non-fatal firearm violence ~7 times.Either way you look at it, offensive gun use far outweighs defensive gun uses, and there is very little evidence validate the claim that guns reliably de-escalate a criminal encounter once it happens.

The Right to Carry Battle:

The clearest test of the MAD Theory of Guns is whether Right to Carry Laws (RTC) deter crime. RTC laws are the ultimate expression of MAD Theory. Those who do choose to carry are the quintessential “good guys with guns.” To get a permit, one has to go through a background check and take several hours of coursework. Even better, concealed carry means the bad guys have no idea who is armed and who isn’t. There could be a “good guy with a gun” around any corner, ready at a moment’s notice to bring a criminal to justice. Criminals realize that any illegal act could easily be their last, and so the presence of concealed carrying citizens should imbue criminals with caution, leading to a decrease in the crime rate.  If we cannot find supporting evidence of MAD Theory here, then it is safe to assert that the theory is intellectually bankrupt.

At first glance, the research appears promising for gun advocates. In his widely cited work “More Guns, Less Crime,” John Lott lays out the case for RTC laws using a comprehensive data set and the veneer of sophisticated statistical analysis. His analysis shows that RTC laws are very effective at lowering violent crime rates. At the surface level, MAD Theory seems to work.

However, even a brief glance at Lott’s findings reveals a multitude of peculiar findings that bear no resemblance to reality. As Albert Alschuler explains in “Two Guns, Four Guns, Six Guns, More Guns: Does Arming the Public Reduce Crime,” Lott’s work is filled with bizarre results.

According to Lott’s data, rural areas are more dangerous than cities. FBI data clearly shows this is not the case. In his data, the number of African-American females over the age of 65 is more highly correlated with the murder rate than the number of African-American teenagers. The number of killings committed by a stranger increases while the number of intra-family killings decreases after the passage of RTC laws. Lott also finds that there is only a weak deterrent effect on robberies, the most common street crime.

One critique of Lott’s work found that, according to Lott’s own data, RTC laws result in a positive statistically significant effect on crime in some states, and a negative statistically significant effect in others. The jumbled, contradictory, and quirky nature of these findings renders Lott’s analysis incoherent at best. Something is very wrong with his methods. To accept the statistical conclusion blindly, even in the face of obvious empirical realities, is not a good practice if we wish to create thoughtful policies on gun reform.

The suspicious nature of Lott’s findings are confirmed by other academic research. In response to John Lott’s book, “More Guns, Less Crime” a sixteen-member panel of the United States National Research Council convened to analyze the findings and statistical methods of Lott’s book.  They found that the existing evidence did not support the more guns, less crime hypothesis. In 2010, a reexamination of the panel’s work concluded the following: “the most consistent, albeit not uniform, finding to emerge from the array of models is that aggravated assault rises when RTC (Right to Carry) laws are adopted. For every other crime category, there is little or no indication of any consistent RTC impact on crime.”

The most recent available meta-analysis on firearm availability and homicide, surveying over fifty international and domestic studies, finds the consistency of evidence in favor of gun regulation overwhelming:

The available evidence is quite consistent. The few case control studies suggest that households with firearms are at higher risk for homicide, particularly firearm homicide. International cross-sectional studies of high-income countries find that in countries with more firearms, both men and women are at higher risk for homicide, particularly firearm homicide. This result is primarily due to the United States, which has the highest levels of household ownership of private firearms, the weakest gun control laws, and the highest homicide rates. Time series studies of particular cities and states, and for the United States as a whole, suggest a positive gun prevalence-homicide association. Finally, perhaps the strongest evidence comes from cross-sectional analyses of U.S. regions and states. Again, places with higher levels of gun ownership are places with higher homicide rates.

In a gigantic, 120-page paper in Stanford Law Review entitled, “Shooting Down the ‘More Guns, Less Crime Hypothesis”, the authors find that the model used in Lott’s paper was riddled with inaccurate assumptions. Most problematically, Lott’s model estimates an aggregated effect for all states adopting shall-issue laws. This is faulty for at least two reasons: 1) it presupposes that shall-issue laws have a uniform effect across all states, when the effect is certainly variable; 2) the aggregated effect gives unfair weighting to states which are early adopters of shall-issue laws, and almost no weight to late adopters. We know from reliable victimization surveys that late adopters tended to experience crime increases, and so any aggregated effect will be biased to show that crime decreased after shall-issue laws were implemented.

The paper finds that, after controlling for the selection effects of very late and very early adopters, and introducing state level trends into a disaggregated model, the conclusions “largely eviscerate the more guns, less crime hypothesis.”

In short, the entire intellectual framework supporting RTC laws, and more generally The MAD Theory of Guns, has been utterly discredited  That being said, although it is empirically irrefutable that more guns does not equal less crime, does the opposite hold? Do less guns mean less crime, or is there simply no significant relationship between the two?

More Guns, More Crime:

The effect of gun ownership on crime is theoretically ambiguous—it could be the case that gun ownership increases the cost of crime and thus deters potential criminals, or it could also be the case that high gun ownership rates makes it easier for criminals to obtain guns, and thus causes more lethal crimes to occur. Extricating the exact relationship between firearms and crime has proven difficult, because of the need for a reliable proxy for gun ownership that is not correlated with the crime rate.

One of the most frequently cited papers challenging the “more guns, less crime” hypothesis is a seminal paper by University of Chicago Professor Mark Duggan in which he shows that gun ownership increases the crime rate.

In this paper, Duggan found that, between 1993 and 1998 gun homicides dropped 36 percent, while non-gun homicides dropped only 18 percent.  He found that one-third of the decline in gun-related homicides since 1993 can be explained by a decline in the gun ownership rate.   Duggan used a clever proxy in order to estimate the rate of gun ownership, the state and country-level sales rate for the Guns & Ammo magazine.

He found that gun magazine subscription data was strongly related to the demographic characteristics that were associated with gun possession, but not correlated to the crime rate.  Using this instrument, Duggan was able to show that guns cause an increase in crime, refuting the ‘reverse causality’ hypothesis in which individuals in high crime areas may hypothetically respond to the perceived threat of criminal activity through buying more firearms. Duggan therefore concludes that “increases in gun ownership lead to substantial increases in the overall homicide rate. This is driven entirely by the relationship between firearms and homicide in which a gun is used, implying that the results are not driven by reverse causation nor by omitted variables.”

To conclude:

  • Statistically speaking, guns are rarely used in self-defense, and thus cannot be defended on the grounds that they can reliably defuse crimes while they are happening.
  • The NRA bases its claim that guns are used millions of times a year in self-defense on a discredited study from 1995 that has not been validated in a single academic paper.
  • Concealed Carry Laws are not associated with decreases in crime, and sophisticated analyses show that, in some cases, there is an increase in aggravated assaults associated with concealed carry laws.
  • The best studies to date, using proxies to estimate gun availability, show that more guns lead to more crime.

 


*correction: Previously the article mentioned that the NRC panel met twice. They only met in 2004, and it was a separate study reexamining the NRC’s findings that was issued in 2010.

About Evan DeFilippis

Evan DeFilippis graduated from the University of Oklahoma with a triple degree in Economics, Political Science, and Psychology. He wrote for the university's opinion page, and received an outstanding contributor award twice for his publications on student apathy. He was the University of Oklahoma's valedictorian in 2012, he is one of the nation's few Harry S. Truman Scholars based on his commitment to public service, and is a David L. Boren Critical Languages scholar, fluent in Swahili, and dedicated to a career in African development. He will be starting work as a Project Associate for Innovations for Poverty Action working on a project called the Kenya Life Survey Panel in September.

60 Comments

  1. Devin Hughes   August 21, 2013 12:53 am / Reply

    We welcome all coherent, respectful comments. Since it is impossible to provide thoughtful commentary on something you haven’t bothered reading in full, comments where this is readily apparent will not be posted. We review all comments before they appear, so there will be a delay even if it meets the preceding criteria. Although we do not typically respond to comments and have no time to engage in prolonged debates, we will attempt to answer questions about our data or research.

  2. Surplus_Knowledge   August 22, 2013 3:50 am / Reply

    Most of the basic facts you use in this article are sourced, and sound, but I take issue with a number of your conclusions:

    Statistically speaking, guns are almost never used in self-defense, and thus cannot be defended on the grounds that they can reliably defuse crimes while they are happening.

    “Almost never?” That’s ridiculous. The VPC source you cited listed 338,700 self-defense uses of guns from 2007-2011. 338,700 is a pretty significant number compared to “never,” which would be 0. You’d consider a 338,700 digit gap “almost?”

    I’m assuming you’re claiming they’re “almost never” used for self defense compared to other statistics about gun use, so let’s look at those. In that 4 year period that number encompasses, there were 46,313 gun murders. The statistics that you claim show that guns are “almost never” used for self-defense actually show that a defensive gun use is 7 times more likely than a gun murder.

    When you look at the amount of non-fatal firearm victimizations/incidents, the defensive gun uses clearly pale in comparison. If your objective is to minimize gun self defense usage compared to non-fatal firearm victimizations/incidents, then you’d be correct. Your conclusion falls apart when comparing the likelihood of gun murders vs gun self defense, though.

    Concealed Carry Laws are not associated with decreases in crime, and the most sophisticated analysis actually shows an increase in aggravated assaults.

    To quote your own passage on this:

    In 2010, [The NEC panel] concluded the following, “the most consistent, albeit not uniform, finding to emerge from the array of models is that aggravated assault rises when RTC (Right to Carry) laws are adopted. For every other crime category, there is little or no indication of any consistent RTC impact on crime.”

    You’re using a catch-all statement of essentially “CCW is associated with an increase in aggravated assaults,” when your own source admits that this is a consistent, but not uniform finding.

    It would be accurate instead to say that “in some cases, CCW is shown to associate with increase in aggravated assault.” Your use of absolutes in this statement is similar to the use of the popular anti-gun phrase “Guns are made to kill,” when there clearly are some models of firearms not made for killing in any form, thus invalidating the statement.

    The best studies to date, using proxies to estimate gun availability, show that more guns lead to more crime.

    Except that your entire first section was about how this concluding statement is essentially not the right statement to make or explore, as it’s more about per capita ownership rate vs the actual number of guns compared to crime. Seeing as national gun crime has hit record lows over the past decade as the number of owned guns has increased, at its face, more guns clearly do not equal more crime.

    • Evan DeFilippis   August 22, 2013 4:16 am / Reply

      Well-worded reply. I’ve modified the verbiage in the post to be less aggressive, as my aim is to be as accurate and articulate as possible with respect to the data.

      One point to mention is that comparing the 338,700 self-defense uses of guns from 2007-2011 to the 46,313 gun homicides statistic is slightly disingenuous. In only a small fraction of 338,700 self-defense uses would the victim have been killed in the absence of a gun, and we can estimate this probability by looking at the percentage of property theft and violent crime victims who are murdered when they don’t have a weapon– and this fraction is very small. By comparison 100% of homicide victims die, obviously.

      A fairer comparison might be to look at justifiable homicides, rather than self-defense uses. The reason being that justifiable homicides might be a better proxy for controlling for situations in which the death of the victim was likely in the absence of a gun. Given that there were only 230 justifiable homicides in 2010 and some 11,000 criminal homicides, the difference becomes much more salient.

      • Surplus_Knowledge   August 22, 2013 5:07 am / Reply

        I’m still wondering the exact reason why you believe 338,700 of something qualifies as that “almost never” happening. Why did you choose these words? I see you’ve edited this to “rarely,” but rarely compared to what? I don’t see any other edits addressing my other points, either.

        My comparison to gun murders was based upon this blog (and many in the gun control movement) bringing up the amount of gun murders as being an important reason for more gun control and gun control laws, despite the fact that said gun murders happened 7 times less than gun self defense usage in the period under scrutiny. The point isn’t homicide itself, but instead what is more likely to happen with a gun: murder, or self defense.

        • Evan DeFilippis   August 22, 2013 6:19 pm / Reply

          My apologies, many of the edits did not go through.

          Self-defense use by a firearm happens “almost never” relative to the claim cited by the NRA that guns are used in self-defense 12.5 million times a year (Kleck, 1995).

          Also, I tried to make it clear in the above post that it is academically disingenuous to compare all firearm self-defense uses between 2007-2011 (338,700) to the 46,313 homicides during the same period. This is because the 338,700 is primarily composed of self-defense uses in which the victim WOULD NOT have died in the absence of a gun.

          Two potential comparisons are more appropriate:

          1) justifiable homicides vs criminal homicides (the reason being that both result in a guaranteed death, and justifiable homicides are situations in which the death of the crime victim was likely). In this case it’s 230 justifiable homicides in 2010 vs 11,000 criminal homicides in 2010…

          OR self-defense uses per year (55,000-80,000) versus the number of crimes committed with a gun each year (450,000-600,000) according to the National Crime Victimization Survey. Each way, offensive gun use far outweighs defensive gun uses.

          • Scott in Phx AZ   December 11, 2013 11:39 pm

            Wasn’t Klecks claim 2.5 million times per year?

  3. Surplus_Knowledge   August 23, 2013 2:53 pm / Reply

    Also, I tried to make it clear in the above post that it is academically disingenuous to compare all firearm self-defense uses between 2007-2011 (338,700) to the 46,313 homicides during the same period. This is because the 338,700 is primarily composed of self-defense uses in which the victim WOULD NOT have died in the absence of a gun.

    And I tried to make it clear in my above post that I’m not comparing the likelihood of someone dying during a defensive situation, but instead comparing the parent theory of “what happens” when a gun comes into play; murder or self defense. The anti-gun lobby continuously vilifies guns as weapons of murder, but my point is that this vilification is unfair because they’re 7 times more likely to be used in a self defense situation (which may or may not result in a death, that’s beside the point), than used in a criminal murder.

    • Paul Morphy   August 23, 2013 5:23 pm / Reply

      Again, the comparison you make is unfair. Guns are 7 times more likely to be used in a self defense situation relative to CRIMINAL MURDER. That’s an insane comparison. A more appropriate comparison would be to compare self-defense uses to crimes aided by a weapon, for which it’s already been pointed out weighs heavily in favor of gun opponents.

      • Surplus Knowledge   August 23, 2013 7:44 pm / Reply

        Guns are 7 times more likely to be used in a self defense situation relative to CRIMINAL MURDER. That’s an insane comparison.

        I fail to see why you and the other poster find this such an incredulous statement. Mass shootings and “gun deaths” are the rallying call behind gun control laws these days. No one talks about how we need to regulate magazine sizes or “assault weapons” because of how criminals possess firearms during the commission of a non-fatal crime. It’s all about the deaths. Bloomberg’s summer gun control tour was themed “No More Names,” featuring Slate’s list of individuals who have died since the Newtown massacre, not a list of armed robbery victims.

        Yet despite all these firearm deaths, according to statistics, self defense is still more prevalent a use of a gun than causing a death. How can gun control activists attack guns as “instruments of death” when they’re used way more often for non-fatal self defense usage?

        A more appropriate comparison would be to compare self-defense uses to crimes aided by a weapon, for which it’s already been pointed out weighs heavily in favor of gun opponents.

        I fully admitted that guns are used more often for what were termed “non-fatal firearm victimizations/incidents” than self defense. The statistics support this. Self defense is still a more common use of a gun than murder, though, which puts a whole new perspective on what is more likely to happen with a gun: murder, self-defense, or “non-fatal firearm victimizations/incidents.”

        • Devin Hughes   August 24, 2013 2:31 pm / Reply

          The reason why Evan and the other dude find your comparison so incredulous is that you are comparing two things of different magnitudes. You can either compare self-defense gun deaths to homicides, or compare self-defense uses to uses in crimes. The first compares deaths to deaths, the second uses to uses. Comparing deaths to uses makes no sense. To provide an admittedly on the fly analogy:

          Imagine that there has just been a battle, and the general asks his aide how his troops fared. The aide replies, “Well, we suffered 11,000 thousand casualties.” The general is disheartened by the news, but asks “How many casualties did we inflict on the opposing army?” The aide proudly replies “We pointed our guns at the enemy nearly 100,000 times, which is more than 7 times the number of casualties they inflicted on us.” The general facepalms in disgust.

          The moral: comparing uses to deaths provides no valuable information. Either compare uses to uses, or deaths to deaths.

          • Jeremy Smith   October 24, 2013 2:19 pm

            Saying the only accurate way to measure self defense uses it by comparing justified homicides to non-justified, would require that only instances of where the victim and offender were armed at the same time. To say guns are never used in self defense, and then use instances where no gun was available for the victim, of course you come out with the conclusion you have.

          • Devin Hughes   October 24, 2013 3:01 pm

            Jeremy: You would be absolutely right if I actually said what you claimed I said. I didn’t. To quote myself: “Either compare uses to uses, or deaths to deaths.” Either is perfectly valid. We did both in the above post, and both comparisons supported our conclusions.

          • Jeremy Smith   October 26, 2013 6:03 am

            Well unfortunately the conclusion reached in this blog is loaded. They specifically chose a counterpoint that was weak and used data provided, by a biased sourced, that supported what they wanted it to. In fact the more of this blog i read the more disenfranchised I am from accepting any of their sources. They even contradict themselves from this post and their recent one regarding gun myths. In this post they claim you can’t look at guns per capita in compare it to a drop in homicides, but in the recent post they use a study that claims guns per capita have a direct impact on homicide rates? Then they ignore solid studies even done by the CDC that show Defensives uses of firearms are more common the violent crimes committed with a firearm. http://www.slate.com/articles/health_and_science/human_nature/2013/06/handguns_suicides_mass_shootings_deaths_and_self_defense_findings_from_a.html . You missed my point in order to have a better more accurate analysis of wether guns are used more defensively, would require also to take in account how often was a gun in the hands of the victims of the crime. (example) you do a study of ten violent crime committed with a firearm, and only in 3 of those crimes were the victims armed. Then out of those three, only two were successful in prevent the crime with their firearm. and then your conclusion is firearms are only used 4 timed to commit crimes then for defense. That is not a very accurate study. Which is one of the many reason we can not have an accurate calculation of how they are bing used especially if your gonna argue that only 35% of homes have guns in them.

          • Devin Hughes   October 26, 2013 3:29 pm

            Jeremy: The reason we ignore the CDC “study” (it wasn’t actually a study, but was rather a review of the literature on the subject) is that the portion you reference was written by Gary Kleck, who was citing himself when he arrived at those numbers. The underlying studies, as we point out above, are absurdly flawed. Garbage in (Kleck’s studies), garbage out (the portion of the CDC study you reference).

            This is why we focused on property crimes. It would be near impossible to formulate a study that accurately showed how many people carry guns around with them outside their house (at least I have yet to see one, as people with a gun may or may not choose to carry it with them on that particular day), but we know with a very high degree of confidence how many households have guns in them, and we know with a relatively high degree of confidence how often those guns are used in self-defense during a break-in. As we mention above, with these two statistics, it is impossible to say that guns are an effective deterrent.

          • freeballer   November 6, 2013 2:05 am

            Its amazing how many gun advocates cite the new cdc study as some sort of nail in gun control’s coffin.. Its clear if you read it, its a stepping stone to further research, funding and education. It says several times the data is conflicting, and more study is required… But i imagine when more studies come out, the nra and gun advocates will simply shrug it off as more “liberal” propoganda.

            … But why take my word for it? read it, in its entirety.. Don’t rely on “conservative” / libertarian media…

            “CONCLUSION

            The research agenda proposed in this report is intended as an initial— not a conclusive or all-encompassing—set of questions critical to devel- oping the most effective policies to reduce the occurrence and impact of firearm-related violence in the United States. No single agency or re- search strategy can provide all the answers. This report focuses on the public health aspects of firearm violence; the committee expects that this research agenda will be integrated with research conducted from criminal justice and other perspectives to provide a much fuller knowledge base to underpin our nation’s approach to dealing with this very important set of societal issues.”

  4. Douglas   October 9, 2013 3:22 pm / Reply

    Duggan’s study is inherently flawed.

    “Duggan (2000, p. 1110): Duggan makes the adjustment for the standard errors in column 2 of table 12. Murder and violent crime show statistically significant drops after the adjustment, but Duggan knows that there are also typos for his rape and assault results. Simply divide the coefficients for rape (=-.052/.0232) and assault (=-.0699/.0277) and you will see that they have t’s greater than 2. Thus for all the violent crime categories but robbery the adjustment does not change the conclusion. In addition, there is the issue of looking at before and after averages versus before and after trends, with the symmetry in the changes in trends before and after the before and after averages do not show a big change even though the change in trends is very big, especially for robbery.

    About half of his violent crime rate estimates show statistically significant drops in violent crime from right-to-carry laws and none of his results show a statistically significant increase.”

    • Evan DeFilippis   May 2, 2014 5:07 am / Reply

      Good find with that paper. I hadn’t seen it before. There are some other papers out there validating the use of Guns and Ammo magazines, but you are essentially right– there are better ones out there.

      For now, the best proxy is a combination of the fraction of firearm-related suicides (FS/S) and hunting licenses per capita. The combination of these two proxies has an R-squared value of close to .95. Studies using that proxy came to the same conclusion as the Duggan study.

      • Davon   May 2, 2014 5:14 am / Reply

        but I quote: “Since Duggan regresses
        GA on G, he is estimating the inverse of β , which implies that β = 2.82 . He finds that
        the long run elasticity of homicide to GA (bβ) is .255, while the long run elasticity of GA
        with respect to homicide (d/β) is .052 (p. 1097). However, the implied true elasticity of
        homicide with respect to guns is
        dH dG b / ( )(1/ ) (.255)(.354) .090 == = β β
        Similarly, the true elasticity of guns with respect to homicide is
        dG dH d / ( / ) (2.82)(.052) .147 == = β β .
        Duggan, therefore, overestimates the effect of guns on homicide, and
        underestimates the effect on homicide on guns, by a factor of 3. The estimated elasticity
        of guns with respect to homicide is actually greater than the elasticity of homicide with 5
        respect to guns. However, given the standard errors involved, the implied elasticities are
        not significantly different from each other.”

        what is your answer to this?

        • Evan DeFilippis   May 2, 2014 7:40 am / Reply

          What? I’m not arguing with you. Like I said, I think there are better proxies out there. At the time of the Duggan paper, FS/S was not in use. Now we have FS/S and hunting licenses, and papers which use that proxy also show that more guns –> more crime.

          As an aside, Kleck doesn’t have the best track record with good papers on the gun debate. I thoroughly debunk his self-defense paper in this article.

          • Davon   May 2, 2014 12:02 pm

            I don’t think you understand, you claimed that this refutes the reverse causation claim but it turned out Duggan significantly underestimated it, and his paper turned out to be extremely flawed, it’s widely expected that when violent crime increases people acquire moee guns for self defrnse. Thus if you want to claim that more crime doesn’t lead to more guns, you cannot use it to ‘debunk’ something that is yet not “debunked”

            If Kleck is refuted then does that apply to Kellermann when it was found out that a majority of the homicides he looked at did not involve household guns? Kellermann himself has admitted to this. What about the other Kellermann study which compared Vancouver to Seattle and came to the conclusion that Seattle’s loose handgun laws were the cause of its higher overall homicide rate? The only flaw that while both cities hsd similar percentage of non-white residents, the demographics were different? Seattle had a much higher black and hispanic population while Vancouver had a higher Asian population, so what happened when the non-hispanic white homicide rate of both cities were compared to instead? Seattle non-hispanic whites actually had a lower homiciee rate (not significantly though) than whites in Vancouver despite owning guns at a much higher rate?

            How about Branas’ et al study which was proven to be methodologically prpblematic by Wintemute?

            http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2866589/

            http://hsx.sagepub.com/content/5/1/64

            Why are these studies still in your ‘database’?

          • Evan DeFilippis   May 2, 2014 12:23 pm

            Why are you so angry? I’m not disagreeing with you. When I wrote this column nearly a year ago (when the blog first started), I wasn’t aware of a lot of the best papers on the subject. I have since updated this argument multiple times in a number of different columns with better evidence.

            Again, studies which use FS/S demonstrate the strong, positive relationship between guns and violence. FS/S disconfirms reverse causation because people don’t commit suicide in response to heightened levels of crime in their neighborhoods.

            The most recent and sophisticated study on this subject came out a couple weeks ago. I would use this as a starting point for your inquiries: http://www.ncbi.nlm.nih.gov/pubmed/24740937

            When I write columns, I survey a lot of literature on the subject. Not all of the papers cited are equally good or sophisticated. I’d encourage you to challenge the best arguments. You can start with the above link.

  5. Pingback: The Inevitability Argument: Why Conservatives Love It, And Why That’s Just the Way Things Are | Armed With Reason

  6. Pingback: How Gun Advocates’ Blind Focus on “Freedom To” is Destroying Lives by Ignoring Our Right to “Freedom From” | Armed With Reason

  7. Pingback: Guns Used For Self Defense - Page 42

  8. Davon   May 2, 2014 1:53 pm / Reply

    Here’s a study which says the opposite: http://www.jstor.org/discover/10.2307/20062076?uid=3739448&uid=2&uid=3737720&uid=4&sid=21103955516417

    Your study doesn’t address reverse causation, though. Nice try.

    • Davon   May 2, 2014 2:54 pm / Reply

      Furthermore in my initial post the first study I linked which examined all proxies in comparison indicates that hunting licenses is a poor proxy for household gun ownership by the GSS. Is your study comparing this new proxy that includes hunting licenses to the GSS results, or to the FS/S?

      For example in the 2013 study by Siegel, he reports that Mississippi has the highest household gun ownership. In contrast a survey by the CDC found that Wyoming had the highest rate:http://www.washingtonpost.com/wp-srv/health/interactives/guns/ownership.html

      • Evan DeFilippis   May 2, 2014 4:06 pm / Reply

        Are you deliberately being obtuse?

        Of course FS/S studies address reverse causation. The entire point of an instrument IS TO TEST REVERSE CAUSATION (otherwise studies would just use GSS measures). How could you not know that?

        Perhaps you need a refresher: FS/S correlates highly with the independent variable (Gun ownership) but is not correlated with the dependent variable (gun crime and violence). This is the minimum requirement for an econometric instrument. The reason it allows you to evaluate reverse causation is because if, in the regression model, higher rates of FS/S produce higher rates of gun crime, then what we are really saying is that higher rates of gun ownership produce higher rates of gun crime (because, again, FS/S is a substitute for gun ownership).

        However, as I stated in my previous post, the REVERSE explanation cannot hold true. It is definitively NOT THE CASE that higher rates of crime cause people to commit more suicides with a firearm , and this was evaluated by the Siegel 2014 paper. Therefore, reverse causation does not describe the reality of the situation.

        The most recent paper on proxies for gun ownership (http://www.ncbi.nlm.nih.gov/pubmed/23956369) finds that a combination of the FS/S and hunting licenses yields a proxy that follows 95-97% of the variation in gun ownership rates. I agree with you, using hunting licenses by themselves is rather useless. Thankfully, no paper published after 2005 does that. Even the Kleck paper you cited shows that FS/S is the best measure. You should ONLY be citing papers that use that measure to prove your point. As it stands, every single measure using FS/S finds that more guns –> more crime.

    • Devin Hughes   May 2, 2014 4:05 pm / Reply

      Couple points. First, check the date of your most recent study by Moody and Marvell. It was written in 2005, and used data from 1977-1998. The study Evan provided is from 2013, with data from 1981-2010. Moody and Marvell’s study may have been correct back in 2005 (though unlikely given that multiple studies have disagreed with them) with the available data, but it isn’t today with the current data. Science over time improves, and the Siegel proxy is one of those improvements.

      Further, the Moody and Marvell study you provided is deliberately misleading even on page 1. They state that there is a rough equilibrium between DGUs and criminal uses. This is blatantly false. There hasn’t been a single survey that looked at both and didn’t come to the conclusion that criminal uses outweighed DGUs by at least a factor of 4. Even further, Moody and Marvell don’t consider that a large portion of DGUs are actually illegal and not beneficial for society. They cite a Cook and Ludwig study that they purport supports their position, but it does no such thing. While it does arrive at the “more than 500,000″ DGUs number, the study goes into great detail why that number should be treated as bogus. They also cite Kleck, but his numbers have been thoroughly debunked as well. So just on page 1, the study you reference already has a couple crippling errors. Which shouldn’t be surprising, as Moody and Marvell have a history of providing studies of dubious quality, as seen in the back and forth Lott debate.

      And you are right, no one study here determines causation. Nor would even 5 or 6. But once you get into the range of dozens or even hundreds, all of which use slightly different techniques and focus on different areas yet mostly show the same outcome, we can confidently assert some degree of causality. Especially when the the proxy used (FS/S) shows the reverse causality story CANNOT explain the effect. The type of proxy used eliminates the potential for capturing an effect from reverse causality. Is a proxy 1-on-1 substitute for the actual ownership rate? No, but it doesn’t have to be to have great statistical value. Further, gun ownership rates change over time. The CDC data you provide is from 2001, and those surveys are few and far between. Using the 2001 gun ownership data and then extrapolating that over 30 years would be foolish.

      • Davon   May 2, 2014 5:08 pm / Reply

        Actually Cook and Ludwig give an estimate higher than 108,000, a number of 256,500-373,000 to be precise because the NCVS doesn’t cover all crimes where a DGU is applicable.

        http://scholarlycommons.law.northwestern.edu/jclc/vol87/iss4/7/

        “Moody and Marvell don’t consider that a large portion of DGUs are actually illegal and not beneficial for society.”

        This comes from the Hemenway study that had a very small sample size to work with, and the five judges that looked at the incidents hailed from California, Massachusetts, and Pennsylvania., two states which are heavily anti-gun. Hemenway mentions that it’s a convenience rather than a random sample.

        The potential for bias here is huge, what is a “legal” DGU in Vermont would probably not fly in California. Consider this, the majority of defensive gun uses actually involve the killing of the criminal or injury to the criminal (which pretty much always get reported in the media)a re found to be legal and justified by the courts. The victims are almost always cleared of any wrong doing. Thus it wouldn’t be unreasonable to conclude that the majority of the 108,000 DGUs are legal. The NCVS controls for false positives because it asks the respondents if they’ve been a victim of a crime first before inquiring about their response to the crime.

        “Moody and Marvell have a history of providing studies of dubious quality”

        Citation? Does this apply to 1993 Kellermann when it was found out that the majority of homicides did not involve guns belonging to the household? So much for family fights turning deadly.

        The BFRSS data has been used in recent studies, it is still considered to be valid. I never said that the data should be extrapolated to the period covered by Siegel, I’m just pointing out the fact that no survey has ever pegged Mississippi has having the highest gun ownership.

        Here’s data from 2002: http://m.pediatrics.aappublications.org/content/116/3/e370/T1.expansion.html

        Siegel says Hawaii has a gun ownership percentge of 25%. Are we really to believe that is true when every survey puts it at 6-10 percent?

      • Devin Hughes   May 3, 2014 3:46 am / Reply

        Several points:

        First, where did I say that Cook and Ludwig gave an estimate of 108,000? Nowhere. I agree wholeheartedly with Cook and Ludwig that NCVS underestimates the number of DGUs. However, their estimate, which you correctly cited, is much lower than what Moody and Marvell provided, indicated they were either being intentionally deceptive or absurdly careless, neither of which bode well for their research. This is not the only time they have been either lying or sloppy: http://econjwatch.org/articles/yet-another-refutation-of-the-more-guns-less-crime-hypothesis-with-some-help-from-moody-and-marvell

        “This comes from the Hemenway study that had a very small sample size to work with, and the five judges that looked at the incidents hailed from California, Massachusetts, and Pennsylvania., two states which are heavily anti-gun.”

        First, Hemenway’s sample is larger than ALL of the original Kleck studies showing the astronomical rates of DGU’s (and is a quite typical size for social science experiments). Further, if you actually read the study, each judge read each case, and on a majority of the cases came to a unanimous decision. Even further, the study directly states that the study sample is random, very explicitly. ALL of your concerns seem to be deeply unfounded here.

        Also, the VAST majority of DGUs involve no injury or death. The ones that do escalate to the point of injury are the ones that are the most likely to be justified by a long shot. It would be a logical travesty to then assert that because the DGUs that are most likely to be valid are declared legal, therefore most of the others are as well. If anything, Hemenway’s results are most likely an underestimate, as it relied on self-reporting, and people are going to report their actions in the best light possible.

        As for the critique of Kellerman, Kleck’s analysis was extensively lacking in understanding. Kleck completely ignores the fact that the gun in the home doesn’t have to be the murder weapon to make the family less safe. People were more likely to die trying to use their gun to ward off the attacker than successfully use the gun in self defense. Had they not tried to “heroically” confront the intruder, it is very likely they would have lived. Further, people were more likely to lose their gun to the intruder than use it successfully as well. Also, similar studies (by different researchers) have been conducted and show the same result. The attack on Kellerman was fueled by a bungled critique by Kleck and exacerbated by politics. If there was a substantive critique of the study, I would reconsider that particular study’s merits. Even if that one proves to be suspect though, there are dozens more that have reached similar conclusions and whose methodologies have not been called into question.

        And for the last time, FS/S does not have to have a 1-to1 correlation with gun ownership to be an excellent proxy. As Evan mentions, the newest proxy follows 95-97% of the variation in gun ownership rates. In the social sciences that is as close to 100% as you are going to get, and far more than sufficient for academic use. This particular proxy is especially useful in that it directly counteracts the reverse causality hypothesis you keep positing. The proxy is more useful than the actual gun ownership data. Statisticians drool over proxies like this. They are awesome.

        • Davon   May 3, 2014 7:42 am / Reply

          “First, where did I say that Cook and Ludwig gave an estimate of 108,000? ”

          I never said you did. That was the initial estimate by the NCVS, you should have read my paper fully.

          “Next, research by Cook and Ludwig
          suggests that perhaps 16-42% of DGUs involve crimes not covered by
          the NCVS. 41 Adding in these would raise DGUs to 256,500-373,000.”

          Which is significantly higher than the 108,000 estimate.

          “Even further, the study directly states that the study sample is random, very explicitly. ALL of your concerns seem to be deeply unfounded here.”

          Nope, I was referring to the Judges, two of which hail from states that are very anti-gun. Again the potential for bias here is massive. What constitutes a ‘legal DGU’ in one state may not be considered one in another. They explicitly mention that it was not a random sample.

          “Only five judges, from three states, assessed the self defense gun incidents from the surveys; they were a convenience rather than a random sample, and the sample is too small to be confident of the stability of the aggregate ratings we report here.”

          That’s straight from the study, are we reading the same thing?

          “As for the critique of Kellerman, Kleck’s analysis was extensively lacking in understanding. Kleck completely ignores the fact that the gun in the home doesn’t have to be the murder weapon to make the family less safe. People were more likely to die trying to use their gun to ward off the attacker than successfully use the gun in self defense. Had they not tried to “heroically” confront the intruder, it is very likely they would have lived.”

          This is the same EXACT copy and paste reply Kellermann gave, here’s what’s wrong with it:

          First off the underlying assumption of the study was that the majority of the homicides involved household guns, that is domestic incidents turn deadly and the victims gun is used against them. Kellermann even goes on to state that people should strongly be discouraged from keeping guns in their homes, Not people should be discouraged for keeping a gun for the purpose of self-defense in their households.” This study is extensively cited by gun control advocates pertaining mainly to this context. Kellermann was being incredibly disingenuous with his work when multiple researchers asked him how many of the homicides involved household guns, I quote:

          “Soon after the article’s publication, the New England Journal of Medicine published a series of letters commenting on the articled, including one from a group of students in a college statistics class (The Students of Dr. Mark Ferris’s mathematical statistics 460 1994) The students pointed out that although Kellermann et al. (1993) were arguing that guns in the household raised one’s risk of being murdered, the authors had not stated how many homicide victims in their sample had been killed with a victim gun rather than a gun that was brought to the scene by the perpetrator.”

          Kellermann had clearly gathered this data because in his subsequent study published in 1998, he reports:”Of fatal and non-fatal gunshot woundings, showed that only 14.2% of the shootings involving a gun whose origins were known, involved a gun kept in the home where the shooting occurred. (Kellermann, et. al. 1998. “Injuries and deaths due to firearms in the home.” Journal of Trauma 45:263-267) (“The authors reported that among those 438 assaultive gunshot woundings, 49 involved a gun ‘kept in the home where the shooting occurred,’ 295 involved a gun brought to the scene from elsewhere, and another 94 involved a gun whose origins were not noted by the police”

          Kellermann did release his data set regarding his 1993 study in 1997, a few years later. It still didn’t include the origins of the gun used to murder the decedents in his study.

          The partial dataset is as follows:

          Behavior Odds Ratio Confidence Interval

          1) Victim Drank Alcohol 2.6 1.9-3.5

          2) Drinking Problems (house) 7.0 4.2-11.8

          3) Drinking Problems (work) 10.7 4.1-27.5

          4) Victim Drink Prob (work) 20.0 4.9-82.4

          5) Housemember drugs 9.0 5.4-15.0

          6) Victim uses drugs 6.8 3.8-12.0

          7) Physical fights (drinking) 8.9 5.2-15.3

          8) Medical attention (fight) 10.2 5.2-20.0

          9) Any household arrested 4.2 3.0-6.0

          10 Victim arrested 3.5 2.4-5.2

          ed 3.5 2.4-5.2
          Environmental Factors

          1) Home Rented 5.9 3.8-9.2

          2) Victim lived along 3.4 2.2-5.1

          3) Security access 2.3 1.2-4.4

          4) Gun(s) in home 1.6 1.2-2.2

          Handgun 1.9 1.4-2.7

          Shotgun 0.7 0.5-1.1

          Rifle 0.8 0.5-1.3

          Any gun unlocked 2.1 1.4-3.0

          Any gun loaded 2.7 1.8-4.0 (This is where the the quoted figure comes from)

          Guns kept mainly for self-defense 1.7 1.2-2.4

          Here we can see that the risk factor Kellermann identified was solely attributable to handguns, rifles and shotguns were not correlated. Are we to believe that shotguns and rifles are not owned at home for the purpose of self-defense.

          Now here’s the problem, why does the risk factor decrease of the gun was unloaded? This would only make sense if the majority of decedents were murdered with loaded and unlocked firearms kept in their own homes. The issue is that they weren’t.

          Why would only handguns and not shotguns or rifles be a significant risk factor? Are we seriously to believe that people do not use shotguns and rifles for defense? I quote: “Virtually all of this risk involved homicide by a family member or intimate acquaintance.” These were not random intruders.

          “People were more likely to die trying to use their gun to ward off the attacker than successfully use the gun in self defense. Had they not tried to “heroically” confront the intruder,”

          Except this still wouldn’t explain why few of the homicides involved guns belonging to victims of the household. It’s true that few cases such as above were mentioned by Kellermann, but they were far from the majority, they only made up 5% of the deaths he looked at. Again, quote: “In our original report, we identified at least 21 incidents (5% of the study’s total) in which the victim died while unsuccessfully attempting to use a gun in self-defense (Kellermann et al., 1993)

          “In the first place, of the original 420 homicides committed in the homes of victims, only 209 (49.8%) of them were by firearm. 26.4% were by cutting instrument, 11.7% by blunt instrument, 6.4% by strangulation or suffocation, and 5.7% by other means. Of those 209 cases, Kellermann does not report how many were shot with their own firearm as opposed to a firearm carried into the home. Sociologist Gary Kleck, however, has used Kellermann’s data and some additional assumptions to try to determine what percentage of homicide victims were killed in their own home using a gun “kept in the home where the shooting occurred.” He concludes that as few as 9.7% and as many as 14.2% of gun homicides were committed in the victims’ home with a gun kept there (Kleck, “Can Owning a Gun Really Triple the Owner’s Chances of Being Murdered? The Anatomy of an Implausible Causal Mechanism,” Homicide Studies 5 [2001], pp. 69-70). So, 209 gun homicides x 0.142 (proportion own gun, own home) = 30 cases. This leads to two conclusions:

          “Of the total number of homicides committed in the homes of victims, only 7.1 percent (30 of 420) were committed using a gun kept in that home. 92.9 percent were committed using a gun brought into the home or another mechanism of death.
          Of the total number of homicides committed in the study area, only 1.6 percent (30 of 1,860) were gun homicides committed in the victim’s home using a gun kept there. 98.4 percent we either outside the home, were not gun homicides, or did not use the victim’s gun. People in the case sample are 62 times more likely to be killed under these other circumstance than to be killed in their own home with a gun kept there.”

          There’s absolutely no way you can claim this is a casual factor in the slightest, it simply does not make sense to claim if if you own a gun, you are 2.7 times more likely to be murdered by gun that isn’t yours. You’re much better off using the infrequency of justifiable homicides that the study finds instead.

          ” Also, similar studies (by different researchers) have been conducted and show the same result.”

          The Cummings. P et al. study that looked at handgun ownership and homicides, correct? It also did not examine the origin of the gun that was used to murder the decedents. Neither did the Wiebe or Dahlberg study as well. Replication here doesn’t mean much because the end result is still the same. Gun homicides in the victims household are rarely committed with guns belonging to a member of the household, family fights turning deadly because of a gun in the house were not the majority, furthermore because most of the homicide offenders were “acquaintances” to the victim (very broad term which can include everything from a friend to a loanshark) it strongly implies that these victims were threatened or at an increased risk of homicide beforehand, and acquired a handgun for the purpose of self-defense, this is the easiest and most plausible explanation. Thus it would be an absolute mistake to claim that the household guns was the casual factor here.

        • Davon   May 3, 2014 7:46 am / Reply

          “And for the last time, FS/S does not have to have a 1-to1 correlation with gun ownership to be an excellent proxy. As Evan mentions, the newest proxy follows 95-97% of the variation in gun ownership rates. In the social sciences that is as close to 100% as you are going to get, and far more than sufficient for academic use. ”

          Which is why the Harvard researchers mentioned the survey as the ‘gold standard’ for measuring gun ownership, correct?

        • Devin Hughes   May 4, 2014 4:15 am / Reply

          “I never said you did. That was the initial estimate by the NCVS, you should have read my paper fully.”

          I have read both Cook and Ludwig as well as the NCVS data. I am quite familiar with both and the bias inherent with the NCVS data. The only reason I again mentioned the Cook and Ludwig study was to highlight the fact that it directly contradicted what Moody and Marvell said. As it seems you are no longer interested in defending Moody and Marvell, we can leave this portion of the discussion aside.

          “Nope, I was referring to the Judges, two of which hail from states that are very anti-gun. Again the potential for bias here is massive. What constitutes a ‘legal DGU’ in one state may not be considered one in another. They explicitly mention that it was not a random sample.”

          My bad. I thought your comment about the random sample referred to the study sample, not the judges. However, you missed my points about the judges as well. In a majority of the cases the judges were unanimous. In more than 90% of the cases, the ruling was either 4-1 or unanimous. The potential for bias here is limited at most. But in fact there is room for massive bias in the study, but it is in the exact opposite direction that you posit. 58% of the DGUs were reported by a grand total of 3 people (50, 25, and 15 uses in the previous 5 years). It is extremely likely that many of these uses were aggressive rather than defensive (and thereby not legal). Further, as I mentioned in the previous comment, people tend to portray themselves in the best light (whether consciously or not). The judges were told to take the self-reported instances as fact. The actual circumstance would almost certainly be much less favorable to each use. So it is EXTREMELY likely that Hemenway’s numbers of illegal DGUs are an UNDERESTIMATE. Your concern of a bias the other way is very clearly unfounded.

          Now for Kellerman.

          Yes, my response came from Kellerman’s response, just as your argument is basically copy and pasted from Kleck. As I mentioned, Kleck’s argument is far more flawed than even the alleged flaws of Kellerman’s study. Let’s take a closer look.

          “Here we can see that the risk factor Kellermann identified was solely attributable to handguns, rifles and shotguns were not correlated. Are we to believe that shotguns and rifles are not owned at home for the purpose of self-defense.”

          In the original study, 42.9% of the homicides were committed with a handgun, 2.4% with a rifle, and 3.6% with a shotgun. There weren’t enough rifle or shotgun homicides in the sample to even determine statistical significance. Further, the study was conducted in the most populous counties of Tennessee and Ohio (translation: urban). Handguns are a far more popular means of self-defense highly populated areas than rifles or shotguns. In rural areas, this may be different given the necessity of owning a rifle or shotgun for hunting or protecting livestock from animals. But for exclusively self-defense, the handgun is the weapon of choice. This is unsurprising.

          “Now here’s the problem, why does the risk factor decrease of the gun was unloaded? This would only make sense if the majority of decedents were murdered with loaded and unlocked firearms kept in their own homes. The issue is that they weren’t.”

          Incorrect, they only have to be a significant enough minority to sway the statistics, which they are.

          I will stop quoting from what you said from now on as it would take up a large amount of room and merely present my rebuttal.

          First, Kleck arrives at the 14.2% by taking data from Kellerman’s 1998 study and then extrapolates the 9.7% number from a study published in 1958. The problem with using the 1958 study are twofold. First, its ancient. Second, Kleck uses it to imply that cases where the owner in the house is not the victim are irrelevant. This is absurd. If there is an argument, and the owner of the house pulls out the gun and shoots the person he invited in, that certainly counts and does nothing to undermine Kellerman’s study (keep in mind that only 14% of the cases in the original study had signs of forced entry, meaning the other person in the house was overwhelmingly allowed in the house.

          Also, using the 14.2% number is deeply suspect. The 1998 study looked at both assaults and homicides, whereas the original study looked at just homicides. Guns are more deadly than other methods of assault, that is simple fact. Therefore, we would expect that this percentage should be higher when only homicides are included. But, for the sake of argument, let’s just assume that the 14.2% stat is accurate.

          Here is where Kleck makes another major logical error. He assumes that this is the only way a gun in the house could have a causal role in the homicide. This is just silly. There are 30 cases where the victim is shot with a gun kept in that home. There are also 21 cases where the victim unsuccessfully tried to use the gun and was killed. So there are at a minimum 51 cases where the household gun directly contributed to the death. At a bare minimum. This is ignoring the other ways the gun could have contributed (e.g. the victim went for the gun but never got to it).

          Now Kleck makes a massive category error. He compares these 30 cases to ALL of the homicides in the study area, and uses this as his damning evidence against Kellerman. He completely overlooks the fact that the study only focused on homicides in or around the home, and makes no claim about overall homicides. This is a stunning oversight. Even had he just limited his comparison to all of the homicides, it would have still been flawed. The correct way to look at the data would have been to focus on how these 51 cases (the bare minimum proper number) relate to the rest of the homicides in homes with guns. So, there are 191 cases in the study of homicides in homes with a gun. of these, at least 51 are a result of the gun in the house. The presence of the gun in the home then at the very least increase the odds of homicide by 1.36 (or 36%). Astoundingly, these rough calculations provide the exact same number Kleck arrives at in his own study, though he desperately tries to downplay his findings as insignificant. Keep in mind that this is the absolute bare minimum (and also best case scenario for Kleck) effect a gun in the house would have.

          Before continuing, it is worth noting that Kellerman’s warning against owning a gun is justified in every way. First, this analysis excludes suicides, and there is overwhelming evidence of a causal link between gun ownership and the rate of suicide. Second, look at the likelihood of effectively using the gun versus being killed by it. In 3.6% of the cases the homicide was justified. That represents about 16 cases (leniently assuming that all of these justified homicides were with guns). The odds in this case are 2 to 1 that your own gun is turned against you. Add in the other 21 cases (died trying) and this becomes even worse. Another Kellerman study found twice as many people lost their gun to the intruder than used it successfully. These odds are really bad for gun owners.

          And I would not assert causality based on the Kellerman study alone. However, not admitting that there is a very high likelihood of causality when there are 5 other studies that have examined this study and have come to a similar conclusion would be silly. Here is a very recent meta analysis that looked at each of these studies and arrived at the same conclusion: http://annals.org/article.aspx?articleid=1814426

          Oh, and the “bought the gun in response to a threat” logic would only make sense if the presence of a gun in the house also significantly increased the risk of non-firearm homicide. It doesn’t. Only one study (Cummings if you are curious) came close to a statistically significant positive effect, but not even it reached significance. Rather unsurprisingly, Kleck attempts to hype this effect as much as he attempts to downplay his own findings (and not very persuasively given his other massive errors in the response to Kellerman). Further, only in 14% of the cases was there forced entry, meaning in the vast majority of cases the acquaintance was certainly and acquaintance (or at least somebody you knew/trusted enough to let into your house, a label that would not apply to loan sharks wanting their money back).

          Given the strong evidence supporting Kellerman and the rather lackluster case brought by his critics, it would be foolish not to think that guns in the house have at least some significant harmful causal role on homicides in the home.

          • geb   May 4, 2014 9:36 pm

            what if the owner doesn’t own a handgun, or a gun for self defense and keeps it unloaded and locked in a safe at all times? it would seem a lot amount of risk is mitigated as a result. it’s about your attitude. I agree that the best self defence method is avoidance

            the problem seems to be handguns, im from canada and I support our licensing and training, I keep my rifle locked at all times, it isn’t for self defence from intruders or intimate acquaintances. according to the study illicit drug usage and alcohol was far bigger risk factors than the handgun as well

            it just seems that the 2.7 figure only applies to people who keep loaded handguns for self defense in the home. no?

          • Devin Hughes   May 4, 2014 10:06 pm

            geb: I would agree with that assessment for the most part. If you have a gun securely locked away it is very unlikely to make an encounter lethal that otherwise would not have been. And alcohol or illicit drug usage is definitely a larger risk factor for homicide. Guns and alcohol are a very lethal combo. Further, people who own a handgun for self-defense are probably far more likely to rely on armed confrontation rather than avoidance (that’s why they bought the gun after all), with overwhelmingly tragic results.

            I would qualify your last statement and say if applies to anybody with an easily accessible loaded gun (not enough cases to determine significance for easily accessible rifles and shotguns). That stat would also probably be higher when only looking at the easy handgun population, as the 2.7 figure is applied to all gunowners, not just the ones with the easily accessible loaded handguns.

  9. geb   May 4, 2014 10:36 pm / Reply

    i would assume that the risk factor for long guns is substantially less than handguns, if you looked at the suicide study by kellermann (risk factor for any gun was 4.8, long guns only was around 3.0)

    even in the duggan paper you linked a 10% increase in handgun ownership was tied to a 2% increase in homicides. handguns are the choice of guns for homicides, generally not rifles or shotguns. for example FBI statistics indicate that knives are used more in homicides than long guns

    the problem seems to be mainly handguns, i just think the 2.7 figure isn’t really applicable in cases such as mine. again no handgun and I keep my rifle locked up constantly when not in use and I even store the ammo separately. its a recreational hobby for me, nothing more and I would never dream of using my gun on another being

    by the way what do you think of smart guns? do you think such technology will be able to mitigate a lot of unintentional shootings?

    • Devin Hughes   May 4, 2014 11:19 pm / Reply

      I agree the risk factor for long guns would be substantially less, if only for the reason most people don’t keep them for purpose of self-defense and often store them. If somebody actively used a rifle for self-defense, my guess is their risk factor would be quite similar to that of handgun owners. On this though there simply isn’t enough data to determine.

      And I completely agree that for a licensed, trained, responsible gun owner such as yourself the risks are minimal.

      I am thrilled by the idea of smart guns. I feel that if we were to implement licensing for ownership (which would include a full background check (including mental health) and mandatory training) along with mandating some sort of smart gun technology (once the technology is sufficiently developed) many of our gun problems would be greatly diminished. Gun theft would no longer be that big a problem (a gun is practically worthless if you can’t fire it), accidental shootings and suicides would markedly decrease, and homicides would almost certainly decrease as well. Of course we would still have 300 million or so non-smart guns in circulation, but it would at least be a much needed leap in the right direction. Unfortunately, here in the US we have states like Georgia trying to make it as easy as possible for criminals to obtain firearms and shoot people, so such regulation is a pipe-dream. We are stuck trying to enact universal background checks, the most basic form of regulation. Tragic.

      • geb   May 4, 2014 11:47 pm / Reply

        excellent points, I’m glad we can agree. I do hope the technology continues to advance, anybody opposed to such a thing is illogical, the most extreme anti-gun or pro-gun person alike

        training, licensing with background and mental checks, as well as secure storage is what I support. no more and no less.

        perhaps you can write a piece on smart guns in the future, I would greatly look forward to reading it

  10. Tofiy   May 6, 2014 3:17 pm / Reply

    I agree with geb here, for example in your annal paper link, even Hemenway acknowledges the fact that most homicides do not involve the victim’s weapon: http://annals.org/article.aspx?articleid=1814430&resultClick=3

    So Kleck was certainly correct that the vast majority of decedents in Kellermann’s study were murdered with guns brought outside the home. Ultimately we will never know what percentage that is with certainty because Kellermann has yet to tell us so regarding his 1993 study.

    The ‘casual’ relationship is shaky at best because it hinges on others having access to a gun, not the gun the victim owns. I come from a country where a license is required to own a firearms and handguns are strictly regulated and a safe is mandated. The majority of the risk was attributed to loaded handguns and being killed by someone else’s gun, also it is certainly possible that some of the decedents acquired a handgun for the purpose of self defense because an acquaintance threatened them with a gun.

    I have a long gun that I keep in a safe stored securely and ammo kept elsewhere, it is not a weapon for self-defense. Also no criminal record, violent history or drug or alcohol usage.

    I should add that in the 1992 paper by Kellermann as geb mentioned the risk was 4.8 for suicides with any loaded gun, this dropped to 3.0 when only long guns were considered, and then to 2.0 for an unloaded and locked long gun.

    Whatever the risk factor is, it certainly is much less than 2.7 and less than 2.0 as estimated by anglemyer (wiebe estimated 1.4 for males, 2.5 for females, dahlberg 1.9, cummings 2.2) for an unloaded and locked long gun not owned for self-defense. It absolutely pales in comparison to other risk factors such as consuming alcohol, illegal drugs, history of an arrest, physical fights in the home, etc, etc.

    As long as you are a responsible owner, keep your gun locked up at all times when not in use, store the ammo separately and not own a handgun or any gun for self defense the risk for homicide is quite minimal, it is quite safe.

    Suicide however is an entirely different question though because most of it is committed with the gun the person owns.

    • Tofiy   May 6, 2014 4:17 pm / Reply

      I just wanted to say though I fully support licensing and safe storage, I don’t use my gun for the purpose of self defense

    • douglas   May 10, 2014 10:18 am / Reply

      ditto, we can only speculate on how many homicides involved household guns as the data has never been released. if the relationship is hingent on external guns and confronting the threat instead of avoidance it is quite a poor one. it could also be that some of the decedents were threatened specifically with a gun and thus acquired a handgun for the purpose of self defense, or it could be multiple factors

      I too do not own a gun for self defence and I rely on avoidance and deescalation.

      • Weer'd Beard   May 11, 2014 12:44 am / Reply

        One big factor missing from this argument is how many LEGALLY held guns never harm a single human being either physically or by presentation (legal or not). Factoring that into the equation further makes the case that banning and restricting lawful gun ownership is a fool’s errand that only harms good people.

        I have a few guns that were military issue rifles that are over 100 years old, chances are they were never on the front lines of the many military battles their issuing country (or countries in some cases) fought, and were likely only fired by soldiers training, or civilians who bought the guns as surplus after the military upgraded their arms.

        I have one gun that lived it’s long life as a civilian arm of personal protection. It’s a Colt 1908 .25 pocket pistol that came off the assembly line in Hartford Connecticut in 1917. The chances this gun was even PRESENTED in a defensive gun use is pretty small, let alone fired at another human being, let alone fired and HIT another human being.

        Most guns aren’t as well preserved as this one after such a long service life, many fall into neglect and are scrapped or kept as trinkets that will be never fired, let alone kept for defense.

        The idea that even most guns during their service life will harm ANYBODY for good or ill is rather foolish, and most of these studies make that implication.

        Also I strongly rely on avoidance and deescalation as well, and odds are that like many others it will be more than enough. Still given the likelihood that any of my guns will harm anybody for ill is so small, it seems smart to keep one close at hand whenever possible in the unlikely event that my life, or the lives of any of my family or friends is threatened.

        It’s foolish to also claim that EVERYBODY who is gravely harmed or killed by a violent attacker somehow could have avoided the event if they had acted smarter.

        The same idea is why I have fire insurance, smoke detectors (several which are wired to my home security system), and fire extinguishers.

        The chances of my home burning down is extremely small these days, but the unlikely event that one of these rare fires happens in my home, without any of these items could mean financial destitution, death, or serious injury.

        With the small inherent risk and cost of these contingency plans, it seems foolish to act any other way.

        • douglas   May 12, 2014 4:09 am / Reply

          “(Il)legal Guns and Homicide: A Case Study of New Orleans, by Jessica Doucet (Francis Marion University), Julia D’Antonio-Del Rio (Louisiana State University), and Chantel D. Chauvin (Lousiana State University)

          Ingeniously, these authors set out to move the guns and crime debate forward by distinguishing between the effect of legal and illegal guns on homicide. They hypothesize that presence of legal and illegal guns affect homicide rates, but in different ways. Legal guns will reduce gun homicide rates (supporting Lott’s more guns, less crime argument), while illegal guns will increase gun homicide rates (supporting Cook’s more guns, more crime argument).

          Looking at the 177 Orleans Parish, Louisiana Census Tracts, the authors measure legal gun access using the number of concealed carry permits issued from 2007-2009 according to the Concealed Handgun Permit Unit. They measure illegal gun access using the number of reported gun violations from 2007-2010 according to the New Orleans Police Department. The dependent variable is the number of gun homicides from 2007-2010 according to the NOPD.

          Controlling for “resource disadvantage” (including poverty, race, female-headed households, education, unemployment), residential instability, religious organizations, and population size, the authors find the following:

          More guns that are legally on the streets (represented by number of concealed carry permits) are associated with lower rates of gun homicide.

          More guns that are illegally on the streets (represented by the number of gun violations) are associated with higher rates of gun homicide.

          Illegal guns have a bigger effect on increasing gun homicide rates than legal guns have on decreasing them.”

          I believe it, why? Because in Canada gun homicides (but not overall homicides!) and gun licenses per 100,000 per province and territory are negatively correlated…

          http://www.rcmp-grc.gc.ca/cfp-pcaf/facts-faits/index-eng.htm Licenses per 100,000

          http://www.statcan.gc.ca/pub/84f0209x/84f0209x2009000-eng.pdf Assault (homicide) by discharge of firearms (X93-X95) 2009

          http://www.statcan.gc.ca/pub/84f0209x/84f0209x2008000-eng.pdf Assault (homicide) by discharge of firearms (X93-X95) 2008

          http://www.statcan.gc.ca/pub/84f0209x/84f0209x2007000-eng.pdf Assault (homicide) by discharge of firearms (X93-X95) 2007

          see for yourself…

  11. Tony   May 19, 2014 6:21 pm / Reply

    1. There are risks with keeping a gun in your home like many other things This is undeniable..

    2. Because so little homicides involved guns belonging to the household, and because I do not own a gun for self defense (hunter), and keep mine locked up, Im going to reasonably conclude that any homicide risk is minimal until proven otherwse. In such situations I too will retreat and not confront the threat.

  12. John   May 21, 2014 5:07 pm / Reply

    “Given the strong evidence supporting Kellerman and the rather lackluster case brought by his critics, it would be foolish not to think that guns in the house have at least some significant harmful causal role on homicides in the home.”

    Not really. If you were to read Hemenway’s response he admits to the fact that so few homicides involve the household weapons. If the risk comes from a gun that is brought into the household and doesn’t belong to the house. Kellermann tries to offer an explanation along the lines of that the household guns do not need to be the murder weapons, and that a false sense of security can also be a factor. Failed self-defense is not sufficient enough to explain the homicides, if it did Hemenway would have noted it in his response. Rather the relationship is “uncertain” which implies another factor.

    But let’s assume that all of the cases save for a few in Kellermann’s study were failed attempts at self-defense. This is the worst casual effect I’ve ever heard. It’s hardly Doll-esque linking smoking to cancer, and because Kellermann refuses to tell us the exact percentage, the 14% is the best figure we have. I would even suggest that the true percentage is even lower just from Kellermann’s unwillingness.

    In terms of direct casual effects on homicide, household guns pose a negligible risk.

  13. John   May 22, 2014 1:30 pm / Reply

    Check this out: http://www.sciencedirect.com/science/article/pii/S0047235211000377

    “The estimated effect of perceived risk and prior victimization changed from being nonsignificant when household gun ownership was the dependent variable (as in most prior research) to being increasingly strong, and statistically significant, when gun ownership of the individual respondent for defensive reasons was measured. Further, once the causal order issue was side-stepped, risk and victimization showed even stronger, significant positive effects on planning to get a gun.”

    Looks like crime DOES effect guns, thus reverse-causation is absolutely true and still real. Duggan is wrong.

    • Evan DeFilippis   May 22, 2014 4:06 pm / Reply

      Hey John– I think you’re misreading that study. Kleck’s study did not include crime rates in his analysis. Instead, he used survey measured of PERCEIVED crime rates that may or may not be related to the actual crime rates. The paper demonstrated the rather obvious fact that paranoid people own more guns. Studies that use FS/S as a proxy for gun ownership have unequivocally disproven the reverse causality hypothesis.

    • Evan DeFilippis   May 22, 2014 4:06 pm / Reply

      Hey John– I think you’re misreading that study. Kleck’s study did not include crime rates in his analysis. Instead, he used survey measures of PERCEIVED crime rates that may or may not be related to the actual crime rates. The paper demonstrated the rather obvious fact that paranoid people own more guns. Studies that use FS/S as a proxy for gun ownership have unequivocally disproven the reverse causality hypothesis.

  14. John   May 22, 2014 4:25 pm / Reply

    And here’s another: http://link.springer.com/article/10.1007/s10940-012-9185-7

    That one uses FS/S as well.

    • Evan DeFilippis   May 22, 2014 4:36 pm / Reply

      Hmm, I’m not familiar with either of the models used in the cited studies. I’ll have to read up and get back to you. Thanks for the information!

      -Cheers,
      Evan

  15. Jeff from FL   May 25, 2014 12:46 pm / Reply

    First, thank you for a thought-provoking article and kudos to you for one of the best discussions on this topic I’ve read!

    With that said, in addition to all of the other counterpoints, I see this as a fundamental flaw of reasoning:

    “4. Perhaps it’s dropping more than any of it’s counterparts because it’s so much higher than all of it’s counterparts? I’m interested in maximizing the rate of downward change… simply pointing to an exogenous effect says nothing about the meaningfulness of gun control.”

    To discount EMPIRICAL evidence as “exogenous” seems a bit disingenuous.

    Jeff

  16. Ricardo Sharpe   June 3, 2014 4:41 am / Reply

    There are a lot of long answers here but let me make it simple. Guns require a great deal of care, experience and proficiency to be useful in self defense. I like your post. Thanks for sharing post on that to Know about using weapons for self defense purposes.

  17. Kyle   June 8, 2014 9:43 am / Reply

    This article makes a few mistakes from what I can see thus far:

    1) Assuming that gun ownership rates are declining. There is no real way to know this as there is no national registry. They can conduct surveys, but a lot of people may lie on those surveys, saying that they do not own a gun when they really do.

    2) People owning multiple guns is not a proxy for behaviors that relate to criminality. People own multiple guns because such people like guns.

    3) Assuming that defensive gun uses involve firing the gun. There are plenty of defensive gun uses in which the gun is not fired at all.

    4) Using Violence Policy Center data. The VPC is a very, very pro-gun-control organization. This doesn’t mean that their data is wrong, but take it with a grain of salt. That would be like a pro-gun rights person using data from the NRA to argue a point.

  18. Pingback: Yet another shooting - Page 8

  19. Joseph   July 18, 2014 2:50 am / Reply

    Approximately 260 million unsuspecting individuals so far, who were disarmed by their government under the pretense that it would make them safer, longed for their firearms at time of death. According to many self-proclaimed ‘experts'; gun owners should just go ahead and “turn ‘em all in”. “When the people fear the government there is tyranny, when the government fears the people there is liberty” ~T. Jefferson. Tell me Evan; have you read, I mean really read the Constitution and could you even begin to understand why it was written to begin with? Someday soon I think you’ll come to understand…

  20. Pingback: The NRA = Not Really American : Quiet Mike

  21. Jeff C   September 22, 2014 3:47 pm / Reply

    I am dismayed to read some of your findings. First, in regards to right to carry and aggravated assualts, I believe this was a single study correct? Where there any further studies into this or was this study just and outlier?

    Furthermore, are you willing to make the statement that:

    “Persons with concealed carry licenses commiting a statistically significant number of aggravated assualts”

    Otherwise, I find it difficult draw anything but a weak correlational relationship between these two, akin to more ice cream sold = more crime.

    Furthermore, I am dubious with your comparason with justifiable homicide vs murder. The motivations for each are distinct. In justifiable homicide the amount of force used can ONLY be enough to stop the threat while in murder the objective is to kill. Surely, you recognize the two differences can have a difference on the justifiable homicide numbers.

  22. Doug R   November 19, 2014 5:15 am / Reply

    You haven’t debunked anything. You’re using the (declining) household ownership percentage rate and treating it as if it somehow corresponded with the number of guns being carried.

    Most gun-owning households of yesterday had nobody licensed to carry. The percentage of gun-owning households who now have one or more residents licensed to carry has risen dramatically. So your choice of data was a poor one, and your analysis was therefore very misleading.

    The relevant questions regarding deterrence are how many guns are being carried at a time when they could be used to deter crime, what does that number look like over time, and how does number that relate to various crime rates?

    If you want to compare a valid and meaningful statistic that relates to the number of guns actually being carried, you should at least begin with the number of concealed carry licenses held over time.

    Certainly not everyone with a license actually carries on a random day, that number is likely a low percentage of licensees. However it makes intuitive sense that those licensees who live in or travel through known dangerous areas may be aware of their circumstances and be more likely to actually carrying during those times. Thus the percentage being carried when it really counts might be higher than the percentage being carried overall. For meaninfgul analysis it may be sufficient to start and end with simply the number of concealed carry licenses held.

Leave a Reply

%d bloggers like this: