Defensive Gun Use and the Difficult Statistics of Rare Events

In the mid-1990s, Kleck and Gertz (1995) estimated that in a typical year about 1.3% of US adults used a gun for self-defense against another person. Kleck and Gertz’s estimate, which came from a survey of nearly 5000 people, implied that there were millions of defensive gun uses every year.

Following Kleck and Gertz’s 1995 paper, the CDC added a question about defensive gun use to their Behavioral Risk Factor Surveillance System (BRFSS). In 1996, 1997, and 1998 the CDC asked:

“During the last 12 months, have you confronted another person with a fire arm, even if you did not fire it, to protect yourself, your property, or someone else?”

But here is the surprise. The CDC buried the question and the results. Only recently was the data discovered and made public by Kleck in a new paper.*(see addendum) So what were the results? You will perhaps now not be too surprised that the CDC’s survey supports Kleck and Gertz’s original finding, about 1% of survey respondents reported a defensive gun use, implying millions of such uses over a year.

The case isn’t closed on defensive gun use, however, because of a statistical conundrum.

The CDC asked 12,870 individuals about defensive gun use over the three samples.That’s a relatively large sample but note that this means that just 117 people reported a defensive gun use, i.e. ~1%. In comparison, 12,656 people (98.33%) reported no use, 11 people (0.09%) said they didn’t know and 86 people (0.67%) refused to answer. People answering surveys can be mistaken and some lie and the reasons go both ways. Some people might be unwilling to answer because a defensive gun use might have been illegal (Would these people refuse to answer?). On the other hand, mischievous responders might report a defensive gun use just because that makes them sound cool.

The deep problem, however, is not miscodings per se but that miscodings of rare events are likely to be asymmetric. Since defensive gun use is relatively uncommon under any reasonable scenario there are many more opportunities to miscode in a way that inflates defensive gun use than there are ways to miscode in a way that deflates defensive gun use.

Imagine, for example, that the true rate of defensive gun use is not 1% but .1%. At the same time, imagine that 1% of all people are liars. Thus, in a survey of 10,000 people, there will be 100 liars. On average, 99.9 (~100) of the liars will say that they used a gun defensively when they did not and .1 of the liars will say that they did not use a gun defensively when they did. Of the 9900 people who report truthfully, approximately 10 will report a defensive gun use and 9890 will report no defensive gun use. Adding it up, the survey will find a defensive gun use rate of approximately (100+10)/10000=1.1%, i.e. more than ten times higher than the actual rate of .1%! Those numbers are, of course, approximately what the CDC survey found which doesn’t prove that Kleck’s interpretation is wrong only that very different interpretations are also plausible.

The bottom line is that it’s good to know that the original Kleck and Gertz survey replicated–approximately 1% of adult Americans did report a defensive gun use in the 1990s–but the real issue is the interpretation of the survey and for that a replication doesn’t help.

Addendum: The paper has since been taken down perhaps because in addition to the issue of interpretation that I raised the survey may not have been national. Robert VerBruggen has further details.

Comments

The survey should require a description of the incident (time, place, person(s), whether reported to the police, witnesses (is the incident alone or with others, and who else witnessed) etc.

Correct!!! Unless the term is adequately defined and some manner of confirmation present, the survey is not particularly useful. One would expect that in the cases of personal defense there would be a police report in the majority of cases.

I would expect police reports of defensive use to be atypical to rare. No upside, lots of potential downside.

For similar reasons, I would expect government survey results to materially undercount actual events.

'No upside, lots of potential downside.'

Not in the following list of states - 'The states that have legislatively adopted stand-your-ground laws are Alabama, Alaska, Arizona, Florida, Georgia, Idaho, Indiana, Iowa, Kansas, Kentucky, Louisiana, Michigan, Mississippi, Missouri, Montana, Nevada, New Hampshire, North Carolina, Oklahoma, Pennsylvania, South Carolina, South Dakota, Tennessee, Texas, Utah, West Virginia, and Wyoming.' https://en.wikipedia.org/wiki/Stand-your-ground_law

Of course, such laws tended to have much less relevance two decades ago.

Stand your ground laws? So what. There is still no upside for reporting.

No upside for being honest on a survey where there is no legal liability?

In that case, make up whatever numbers suit you best by using this German concept - Dunkelziffer.

Clockwork is triggered because he doesn't like triggers.

The simple fact is that guns in honest people's hands save lives. Perhaps a couple million times a year in the U.S. the 2nd amendment prevents crimes and saves lives.

5479 times a day? Oooookay.

I understand, big numbers are tricky. 330,000,000 people. Perhaps billions maybe even trillions of human interactions between two or more people everyday. BIG numbers! Scary! But indeed not all interactions between people are civil. Many of them are hostile and threatening and in that context 5479 hostile interactions a day is not surprising or even out of line.

Even in states with stand your ground laws, I would expect people to be reluctant to voluntarily involve the police after defensive gun use where there were no actual shots fired, except where the assailant can be identified.

And of course in other states, the risks are even higher.

England has reached the sad state where self defense inside against home invaders inside your own home can end up putting you in jail. While the US is not generally this crazy, prudent people may well avoid inviting the police to make after the fact judgements.

'I would expect people to be reluctant to voluntarily involve the police'

I was responding to this, however, though not obviously enough - 'For similar reasons, I would expect government survey results to materially undercount actual events.' The wisdom of keeping American police as far from your affairs as possible is undisputed, however.

You mean like after firing at a child who knocked on your door to ask directions, claiming it was in self defense?

Or shooting through a door and killing a motorist in need of assistance, claiming it was in self defense?

Wrong, you were responding to this: "I would expect police reports of defensive use to be atypical to rare."

No, I was responding to this - 'For similar reasons, I would expect government survey results to materially undercount actual events.'

As noted by the response to Dave Smith, which said 'No upside for being honest on a survey where there is no legal liability?'

But sure, as also noted in that comment, I did not make it fully clear what I was responding to in Engineer's comment.

Now do rapes, and all other crime stats. Most of these incidents do not get reported to the police because there is no information on the perp and in many cases, the incident is in a high crime area which tend to be gun controlled (think about that a for a second).

Of course Alex's comments apply to rapes, and indeed to all crimes. I don't know why he didn't mention that. (He usually has more intellectual integrity than most academic bloggers.)

That was a response to Bill who wanted witnesses and a whole description of the event on the survey.

DGU has potential criminal and civil liability. The same isn't true for reporting rape or other crimes.

Exactly, now do "black lives matter" or campus "sexual assault" claims.

The standards of proof demanded at this blog vary widely depending on the author's affiliations. Tyler in particular will grab the most ridiculous quasi-statistic to support whatever social analysis happens fashionable at GMU at the moment. The most shameless social justice warrior PR disguised as economics.

Sexual crimes usually involve people who know each other.

E.g., on first date and didn't consent. Relatives outside of immediate family. A teacher or boss.

Are you serious Bill? Nothing happens unless there are witnesses and a report?

I know two people who anecdotally have similar stories, and what is even more interesting they didn't have a weapon. Simply giving the impression that they were armed changed a situation.

There are YouTube channels that seem to come up with a security camera video of defensive gun use every day; it can't be that rare.

Here's an example:

https://www.youtube.com/channel/UCsE_m2z1NrvF2ImeNWh84mw

It's a bit cheesy, and the videos are not exclusively American. Some observations from a total amateur: get away if you can (a shocking number of people just mill about in the presence of violent crime, perhaps thinking the perpetrator is not focused on them); fighting back with bare hands usually goes badly; keep your gun on your person, not in a drawer or a purse; keep your gun ready to shoot - safeties or other actions required before shooting are hard to execute under pressure; ambush, don't shoot it out; perpetrators often bolt at the first sign of armed resistance; do not follow a perpetrator who bolts; do not count on one shot to stop a threat; practice with your firearm; tasers are useless, except maybe for restraining someone who's already on the ground.

Keeping a gun ready to shoot and too accessible can result in the gun being discharged accidentally, wounding or even killing people. No, guns should be stored safely and in ways where accidental use (including by young children) is unlikely.

I believe I said to keep it on your person and not leave it laying around.

'keep your gun ready to shoot - safeties or other actions required before shooting are hard to execute under pressure'

You do realize that an unsafed loaded gun is about the worst way to carry around a weapon, right? That includes combat zones (outside of being engaged in combat) by the way.

For the love of god prior... most police-issue sidearms don't even have the old-fashioned style safety anymore, because of the problems I mention above.

If you are talking about a Glock, it is not true that they are not safed when being carried - https://us.glock.com/technology/safe-action

Nobody with any familiarity with guns ever walks around with an unsafed loaded gun unless they are expecting to discharge the weapon in the immediate future. And anybody with any familiarity with guns never wants to be around someone with a loaded unsafed weapon except in certain specific situations - such as engaging in combat.

This is really, really basic gun safety, by the way. Of the sort drilled into people who actually use weapons. Know any Marines?

Because this is the text of the first page from the DIV 24 Pistol Weapons Handling_Media.pdf

SAFETY RULES

1. TREAT EVERY WEAPON AS IF IT WERE
LOADED

2. NEVER POINT A WEAPON AT ANYTHING
YOU DO NOT INTEND TO SHOOT

3. KEEP FINGER STRAIGHT AND OFF THE
TRIGGER UNTIL YOU ARE READY TO FIRE

4. KEEP WEAPON ON SAFE UNTIL YOU
INTEND TO FIRE

Though I will be honest, the Marines have a reputation for not being defensive gun users - they are proud to be thoroughly offensive oriented gun users, and leave the defensive gun use to the people they are attacking.

On the other hand, I will be absolutely shocked if you find a single U.S. military manual or regulation that says walking around with a loaded unsafed weapon is ever allowed outside of active combat situations.

The people who use weapons professionally are fully aware of how dangerous they are, after all.

Firmly pull the trigger on the Glock and it will fire. Drop it and it won't. It's a modern handgun. It doesn't require flipping a switch to enable firing.

So the first part of your comment is just wrong, or at least misleading.

The second part is non-sequitur. I have no idea what it's supposed to add.

'your comment is just wrong, or at least misleading'

You did see that the link was to Glock explaining its safety system, right? No one walks around with a loaded unsafed Glock.

'I have no idea what it's supposed to add.'

That no one who uses weapons professionally is especially interested in weapons that cannot be reliably safed (a trigger safety is a partial safety - it does not prevent a weapon from being discharged accidentally, for example by a drawstring from a coat getting caught at the trigger). Which might just explain why Glock sells a number of models with this feature - 'The GLOCK Safety lock* is a development by GLOCK and is the first safety cylinder lock for pistols worldwide located conveniently in the back of the pistol grip. GLOCK’s modern key technology permits the best possible adaptation to your requirements.' https://eu.glock.com/en/technology/glock-safety-lock<https://eu.glock.com/en/technology/glock-safety-lock

People who routinely use weapons to kill people - you know, like Marines - tend to be real sticklers for weapons safety. Extending to the sorts of requirements they have in the weapons they purchase.

LOL. Do you think anybody carries a weapon in a state requiring a *key* to prepare it to fire?

Almost every modern striker-fired handgun, like a glock, do not have an external safety. This includes many modern hammer-fired gun like the Sig Sauer P226/229 series. In the case of handguns that do not have external safties the holster will act as the safety. Every cop, special agent, professional LEO carries their firearm with a loaded chamber in a holster regardless of type of handgun. Holsters are the safties for handguns.

Both wrong. Yay pluralism.

Lord A,

No one walks around with a semi automatic weapon (select fire) on "fire" unless he is a complete moron. A dangerous moron. Any trained adult can select fire while drawing a weapon. And no sane adult walks around in the firing position (high ready) all day. Not in combat zones nor anywhere else. The one exception to this is in clearing houses in urban areas. The weapon is drawn (at the stack) and depending on the unit the clearing team will select burst on the weapon system (upon stack) to avoid the need to select twice (safe to semi to burst/auto).

Clockwork,

Technically wrong. The M249 Saw (and certain open bolt weapon systems in general) are famous for not being reliably "safed." Weapon is charged or not. The backup to not having a reliable safety on an open bolt weapon system is to not charge the weapon in the first place (pulling the charging handle back to load the linked round into the open bolt). Riflemen and grenadiers can manage the first 1 second of contact and charge their weapon system without the SAW gunner crawling and jumping with a loaded and charged LMG with a notoriously unreliable safety mechanism.

Yes, people walk around with semi autos on "fire" all the time, in the sense that a huge percentage of modern weapons do not even have such a selector and the mere pulling of the trigger will result in a shot being fired.

Please list these modern weapons that do not have a safety mechanism.

Your implication is that a "huge percentage" of firearm holders are walking around with a round loaded in the chamber "all the time" with nothing but the trigger pull to prevent accidental discharge.

If so I may need to change my priors from "strong belief in the 2nd amendment and its necessity to a free society, to gun owners are irresponsible idiots and it's worth the risk to limit guns"

There is not a Glock in existence with a safety. It is the #1 supplier of LEO firearms.

Holy smokes, on behalf of all Glock owners and CCW holders, would clockwork_prior and Hmm kindly stick a sock in it.

Notwithstanding your other delusions of gun familiarity, "select fire" is not what you think it is. Hmm is so uniquely pompous and wrong at the same time that I wonder if TrollMe has moved on to another screen name.

I know! It's actually amazing how he writes with such authority while betraying his total and complete ignorance of the topic. He's probably a shocking good con man.

Re: There are YouTube channels that seem to come up with a security camera video of defensive gun use every day; it can't be that rare.

Er, one incident a day would translate into 365 incidents a year. OK, of course there are plenty of incidents that don't end up on YouTube, but still your statement does not prove that there are immense numbers of such incidents.

Sure, and I'm confident the vast majority of them don't end up with good video - we're only seeing the ones that occur in stores, in front of police cars, or in countries where video surveillance of the street is common.

Widespread video means we know there's no such thing as Sasquatch, curve balls really curve, and defensive gun use is commonplace.

Also, the idea that the bad guy will grab your gun appears to be a myth. Guns, it turns out, are designed to be held on to by the grip. Trying to grab a gun from an armed person is suicide.

It's suicide. I knew about that incident when I wrote what I wrote. It does not generally work. It is an extremely bad idea.

Which makes one wonder why so many criminals seem to be able to grab guns from police officers when it is so suicidal. Of course, some do die in the attempt, and maybe such people are desperate enough not to care about their chances of dying. Nonetheless, it does not seem excessively suicidal to grab a police officer's gun, if news reporting over decades is to be trusted.

They often grab officer's holstered weapon in a scuffle, not guns aimed at them.

As "8" said.

In a scuffle over a contested gun, the grip-holder moves back, this tends to draw the gun into line with his contestant's body. This is why grabbing a gun is usually a bad idea from a simple kinaesthetic/MMA POV. There are better disarming techniques.

"Tell that to this guy "

The gun was empty in that case. The shooter was trying to reload.

I think Bill's point is that a lot of things might fit this description which don't really qualify as "self defense". You naturally think about self defense with a firearm as something like confronting home invader or mugger. But it's possible that this kind of incident is in the minority.

For instance:
-Some neighborhood kids are playing on your property and you saunter onto your porch with a shot gun.
-Someone makes a pass at your wife at a bar so you reveal your firearm.
-You get into a fender bender and you pull out your gun because you feel like the aggrieved party is being "hot headed".

Without hearing the specific stories, is hard to tell what is legitimate self-defense and what's glorified bullying.

"In the mid-1990s, Kleck and Gertz (1995) estimated that in a typical year about 1.3% of US adults used a gun for self-defense against another person. Kleck and Gertz’s estimate, which came from a survey of nearly 5000 people, implied that there were millions of defensive gun uses every year."
I can not imagine what life must be under such conditions. Life as a constant, nasty, brutish, short deadly struggle against one's fellow citizens.

Me neither. But then, I'm American.

Yet, in America, cvilians must fend off deadly agression on a daily basis.

Apparently, in Brazil civilians don't manage to fend off deadly aggression.

"In 2017, Brazil had a murder rate of 39.2 per 100,000 population. There were a total of 56,101 murders in Brazil in 2017. Another study has the 2017 murder rate at 32.4 per 100,000, with 64,357 homicides."

https://en.wikipedia.org/wiki/Crime_in_Brazil#By_Brazilian_states

Many of those are not "civilians" by no stretch of imagination. Cops (0.4% of the murdered were cops from Rio de Janeiro City alone), drug dealers, robbers, suspects, etc. The civilian murder rate must be a half of that.
My home state actually experienced a 60% decline in murders from 1998 to 2012. Brazil has achieve unprecedent good outcomes in crime fighting. https://en.m.wikipedia.org/wiki/List_of_Brazilian_states_by_murder_rate
Some states of the Northeast are suffering the effects of late urbanization and the boom-bust cycle. The general situation is basically under control. Measures are being taken to deal with special situations. Rio de Janeiro City are under military intervention. The tanks are rolling as we speak. As Churchiill said in 1942, "Now this is not the end. It is not even the beginning of the end. but it is, perhaps, the end of the beginning."

Representative Bolsonaro has promissed to make crime fighting the cornerstone of his
Administration and hit hard the criminals.

Murder is out of control in Brazil. You would have to pull the trigger.

The crime plug is being plugged. More and better border controls, more Armed Forces participarion in crime fighting, more intelligence coordinations. Thougher punishments are under studt. All in all, the situations is under control.

Obligatory question: why are we discussing 20-year old statistics? Where's the more recent research?

In the case of the CDC, they have been prohibited from collecting data by Congress until earlier this year. One might be able to collect such information from police departments but there is no centralized way of doing this.

No, they can collect data. They have not been allowed to fund pro-gun control efforts. Why did they stop? Well that and that they buried the fact that they confirmed Kleck and Gertz have the same answer.

And that data has it's own problems as well as the problem of integrating and introducing error via the integration processing.

True. The CDC has the answer - confiscate guns. It simply needs millions to fund statisticians to gin up a bunch of standard deviations, etc. and kill a bunch of trees with a 600 page report replete with fake data and conclusions to trash the typical American as a closet homicidal maniac and unconstitutionally repeal of the Second Amendment.

Plus, how many instances are not (would not be) be officially reported?

Owning relatively standard guns for the purpose of volunteering in a state militia for pro-freedom purposes is different from bringing handguns to the mall.

"relatively standard guns" are on the chopping block under the most recent gun control proposals by Democrats. Why pretend that the choice is between legal "standard guns" and criminalizing banning CC at shopping centers? The choice is between criminalizing everything except licensed flintlocks for everyone less influential than Chelsea Handler and Nancy Pelosi, or legal standard guns.

I didn't say that the positions are representative of what choices are available.

For example, some people advocate for some background checks as opposed to no background checks. Others advocate for a mandatory cooling off period, where you have to wait at least a few days between processing a background check and purchasing a firearm.

Yeah I would say it's significantly worse (also irrelevant)

The mom defending herself from her two year old child was reported.

I guess that argues for more guns in the hands of mothers for self defense.

I know the idea of balancing costs and benefits is incomprehensible to you, but it's not to anyone else so please stop.

The real story is the cdc burying inconvenient results. It calls into question all of their surveys.

Yeah. I really don't doubt that the surveys, both by Kleck and the CDC, over-reported at some level.

But the data hiding? What we've got here is the smoking gun to show that the CDC was doing policy advocacy rather than science when Congress cut off the funding. Does the rot go any deeper than firearms research?

Did they suppress all results that had this kind of statistical problem (basically such a low rate of reporting that it's hard to know whether it's meaningful)? If the CDC did this consistently, then there's no partisan data hiding going on and no scandal and they were just doing their job as well as they knew how. If they only suppressed results they didn't like for some policy reason, then it's a huge problem.

“We need to revolutionize the way we look at guns, like what we did with cigarettes ... It used to be that smoking was a glamour symbol—cool, sexy, macho. Now it is dirty, deadly—and banned.”

Mark Rosenberg, then the director of the CDC’s National Center for Injury Prevention and Control

"The actual amendment sponsored by Jay Dickey, a congressman from Arkansas, did not explicitly forbid research into gun-related deaths, just advocacy. But the Congress also lowered the CDC’s budget by the exact amount it spent on such research. Message received. It’s had a chilling effect on the entire field for decades."
https://www.theatlantic.com/health/archive/2018/02/gun-violence-public-health/553430/

Hmm. Weird.

The federal government falls prey to politics?!? Say it ain’t so!

Does either data, or the underlying analysis using them, control of people in either law enforcement or the military?

Also, would be better to also include a scenario where the lying is asymmetric -- more will lie about not using defensive force than will falsely claim they did.

@Thiago -- Really, 1% (or less if Alex's imagination was correct) gets us to a Hobbesian state of nature? Thought you were in one of the big Brazilian cities? Those are not known as wonderlands of personal safety and universal good will towards one another. You could also try visiting the Philippines, particularly Manila but suspect it's maybe even truer Mindanao. Armed guards everywhere at the Mall. (They are actually very nice and polite -- the police or military personal are a bit more reserved and don't want their time wasted by tourists.)

Adding about the "statistical conundrum" (https://amstat.tandfonline.com/doi/abs/10.1080/09332480.1997.10542033?needAccess=true&journalCode=ucha20)
Perhaps just not in the visible portion but they seem to say the sample show guns being used 1.5% of the time in a defensive way (and that twice as many times the gun was taken from the victim by the robber suggesting 4.5 % that it was attempted to be done).

But that is only going by reported to police cases. If a gun were successfully used as a deterrence and no really property damage (under the deductible) how many people would simply not report the event?

At the end of the day we only know the numbers in any given study and they are what they are -- nothing more or less. In most cases whatever the aggregate statistic come out to be the are of limited value for informing on policy decision (doubly so when we don't really know what they actually mean as both (all three?) sets seem to suffer from).

With all the recent finds about how statistics really have a fundamental design flaw that is seldom or never accounted for (or even understood) by the scientist, including the social scientist, you would think we'd start moving away from inquiries like these and on to more useful inquiry.

"In the mid-1990s, Kleck and Gertz (1995) estimated that IN A TYPICAL YEAR about 1.3% of US adults used a gun for self-defense against another person."
I can not imagine any situation aside full-scale military mobilization unheard in Brazil since the War Against the Paraguayan Aggressor in the 1860s that would make more than 1% of Brazil's population see action avery year. We could reocupy the Cisplatine with much less effort. Also, my city, particularly my neighborhhod, is reasonably safe. Lots of university kids, professors, managers, retired people, small business-owners, etc.

Also, my city, particularly my neighborhhod, is reasonably safe. Lots of university kids, professors, managers, retired people, small business-owners, etc.

IOW, the key to being safe is being affluent enough to afford affluent neighbors.

Brazil's murder rate is four to five times higher than that of the US. You have more and shittier neighborhoods.

"IOW, the key to being safe is being affluent enough to afford affluent neighbors."

It is a shame criminals (usually) have legs and have worked out the benefits of transportation. But the posh students, retired people, small business owners and I thank you for your concern anyway. The entire city's criminal activity is at a record low. Not thanks to you.

"Brazil's murder rate is four to five times higher than that of the US."
America's suicide rate is more than twice Brazil's. You have more people, but they lead sh***ier lives.

Suicide rate is not a good measure of how shitty a population's lives are. Aren't Scandinavian suicide rates high?

Actually, if Wikipedia is ro be believed, Sweden is tied with the USA, and they have socialized healthcare and better furniture.
https://en.m.wikipedia.org/wiki/List_of_countries_by_suicide_rate
They just look like theynare about to kill themselves. Brazilian wruter Otto visited Sweden and said he was strongly tempted to ask people there, "When will be the suicide?". They really looked depressed. It is just how they are.

Thiago has reported previously that he lives in one of the smaller cities on the northeast Atlantic coast. (Vitoria, maybe? I'm not going to go to the trouble of sifting through old comment sections...) My impression is he lives in the equivalent of a nice college town.

I was born in Vitória, a few blacks from the sea, and go back often to visit family, but spend most of the year in Campinas nowadays. Espírito Santo (of which Vitória is the capital) actually experienced a 60% decline in murders from 1998 to 2012. A steady hand is it what takes to deal successfully with crime.
Campinas is (on average) reasonable safe. And my neighborhood is almost a college town of its own with low populational density, at least three medium to big universities around, lots of students without much money to spend and people who settled there before the universities came. It is not affluent, it is simply quiet and lacking many social ills.

"Does either data, or the underlying analysis using them, control of people in either law enforcement or the military?"

Both Kleck and Gertz and the CDC specified outside of one's work duties.

After the Iraq War, the Lancet estimated about 1 million excess deaths, extrapolated from about 30 survey responses. The suggestion that there might be any imprecision in those results, let alone any dishonesty, was denounced by opponents of the war.

Kudos to Alex for taking a different approach and being willing to ask difficult questions of the data in this case, even when the results might be considered politically desirable.

Yeah. But in addition a bit of shoddy stats on confidence intervals, the real story there was that the actual survey data was clearly faked by the Iraqi researchers in the field; (some) of them were clearly making it up.

The Lancet leadership didn't look too closely at their own data and rushed to publish because it told a story they liked. Then they tried to withhold the data after the facts. Really quite a disgraceful partisan episode that wrecked that publications credibility for me.

The most entertaining thing about the Lancet study was all the innumerate talking heads/columnists who suddenly acquired a strongly-held opinion on the validity of cluster-sampling, which just happened to align with their prior views on the wisdom of the Iraq war.

800,000 and 1.2 million are both large numbers too. The context is one where for all of history, it has been known that civilian often pay the largest price in the form of suffering due to deprivation during conflict. If you are aware of credible debate about the numbers, please share a link.

1% is a small percentage of respondents to potentially be lying or differently understanding the question.

A significant proportion of the survey responses were faked.

For example, 80% (iirc) of the survey deaths supposedly came with a death certificate from Iraqi government (to show it was "real" death...). BUT the Iraqi Ministry of Health had only issued ~300k certificates against the 800k+ supposed deaths uncovered by the survey. What good luck that 80% of deaths in this "representative" survey came from the documented 40% of the general population, right?

Of course not. We can all do binomial stats. The survey was faked by (some) of the field staff (there's tons of other give-aways from copied entries to weird demographics and distribution of deaths). It's GIGO.

Not sure. But if a sample is not representative, it is within good statistical practice to adjust numbers to make the sample representative.

The population of Iraq being 35 million or so, only 300k death certificates within a single year would be equivalent a life expectancy of 100 years. I.e., the 300k figure you mention is extremely obviously not a correct number, regardless of how many death certificates were issued by medical professionals during the warring period of interest.

Well, the government writ didn't include Kurdistan and some of the Sunni triangle, and I'm not sure in the census period was 1 year or a shorter duration.

But the point is the incidence of claimed certification was TOO HIGH; the survey takers were inflating the death numbers without realising that they should have used lower certification rates to be credible.

If death rates are lower during a war than before a war, when calculated on the basis of death certificates officially issued by officially recognized physicians, then almost certainly many deaths are not being captured through the process of official death certificates issued by officially recognized physicians.

How does it correlate with the display of ANY weapon in self defence?

Of course, the real issue is not how many aggressors were deterred by display of a weapon present, but by how many aggressors were deterred because they feared a weapon might be present....

Right. I guess this number (how many armed interactions) is certainly a lower bound on the more interesting one (how many deterred). And if 1% is true, then this bound is already higher than most anti-gun types would admit. 1% of adults every year is a lot.

Yes.

I don't carry any particular love for the 2nd, and think the US may be in a sub-optimal equilibrium here in several ways. BUT defensive guns must still have some deterrence effect, right? If we assume that criminals have at least bounded rationality? After all, we readily accept criminals are deterred by other factors before the fact, so why not guns? Let's try and honestly measure that effect!

Instead we have anti-gun types who are reduced to the anti-scientific position of proclaiming that there can be no deterrent effect at all, and hence that criminals take no account of their target before engaging in aggression. It's intellectually dishonest.

If I shoot some young trespassers with rock salt and they don’t come back does that count as deterrence?

Yes? I assume "weapon use" is a subset of "weapon show" and so also counts as deterrence.

'How does it correlate with the display of ANY weapon in self defence?'

Precisely - I would guess that the defensive use of one's hands (particularly with the survey definition of 'use') is considerably higher than that of a gun.

'how many aggressors were deterred because they feared a weapon might be present'

Apparently, American criminals are different than German ones, as in Germany, the lower crime rates are clearly not due to a large number of armed citizens deterring crimes. http://www.nationmaster.com/country-info/compare/Germany/United-States/Crime

Number of police, on the other hand, might be relevant in that comparison.

Germany has lower crime rates because it has a high percentage of pacifist Germans. As the non-German, non-pacifist population has increased, so have your crime rates.

In sum, people make countries, not the other way around.

'As the non-German, non-pacifist population has increased, so have your crime rates.'

Well, one hopes you can read German, because you are completely wrong - 'Deutschlandweit wurden den Angaben zufolge fast zehn Prozent weniger Verbrechen registriert als noch 2016. Der "Welt am Sonntag" zufolge ist das der stärkste Rückgang seit 1993 und insgesamt der tiefste Stand seit der Wiedervereinigung. Demnach gab es im vergangenen Jahr rund 5,76 Millionen Straftaten - mehr als 610.000 weniger als 2016.' https://www.mdr.de/nachrichten/politik/inland/kriminalitaet-osten-statistik-gefaehrliche-staedte-100.html#sprung5

German crime rates are currently at the lowest level since re-unification, basically.

The first google entry: http://www.dw.com/en/german-study-links-increased-crime-rate-to-migrant-arrivals/a-42006484

According to the study, which was conducted by the Zurich University of Applied Sciences and paid for by Germany's Ministry of Family Affairs, police witnessed an increase of 10.4 percent in reported violent crimes in 2015 and 2016. More than 90 percent of the increase (not of total violent crimes) was said to be attributable to migrants.

A lower link says 2017 decreases. 1 million immigrants causing that kind of increase in crime in a country of 82 million people is a huge change.

Shame you do not read German, as the figures quoted above are from 2017, which had a total decline of reported crimes of 10% compared to 2016.

Further, the current crime rate in Germany is now the lowest since German reunification - in other words, Germany crime rate is roughly the level it was in 1990, even after the huge change of accepting a large number of refugees.

However, just as like before reunification, Germany is now host to a large number of single men. But considering that the Americans and Russians stationed here were also tolerated, along with the inevitable sorts of crime that go along with single men of that age group, it does not seem as if Germany will find itself facing serious problems. 2018's crime rate will certainly provide more information, even it may not come up in the first position of an English language google search.

Absolute levels of crime rates aren't a valid comparison, better to see what types of crimes are being committed and the trends in those rates. Gun ownership shifts criminal behavior towards property crime. American burglars (the subset that is not very stupid) do not want to enter an occupied home. Hearing a homeowner say, "I have a gun" is enough to send most of them fleeing.

Obviously your hands are not a weapon, no one would interpret the question this way

Yet how often are hands used to deter crime? Undoubtedly much more often than guns, even if a fist punching you in the face does not fit your definition of weapon.

"Of course, the real issue is not how many aggressors were deterred by display of a weapon present, but by how many aggressors were deterred because they feared a weapon might be present...."

Ok, if knowledge someone has a gun derpters threats, then I should consider all reports of police offices killed by guns to be fake news.

Guns deter violence, cops always carry guns, so no one is ever killed when cops are around, especially cops are never killed.

Fake News from the failing NBC:

"The official story about how two Florida sheriff's deputies were gunned down in a Chinese restaurant shifted slightly Friday when a deputy standing in for Gilchrist County Sheriff Bobby Schultz announced that they were shot inside the eatery — not through the window as officials had first reported.

"We know now that the assailant went inside the building and assaulted and killed them," said Lt. Scott Tummond, who works for the Levy County Sheriff's Office."

FAKE!

Fake News from failing CNN:

"Two Ohio police officers were fatally shot Saturday while responding to a 911 hangup call involving potential domestic abuse, authorities said.

Westerville Police Officers Eric Joering, 39, and Anthony Morelli, 54, were shot as they entered an apartment in Westerville, Division of Police Chief Joe Morbitzer said. Westerville is a northern suburb of Columbus.
"Both officers gave their lives in the protection of others," Morbitzer said through tears. "Those are true American heroes."

FAKE NEWS!

"Ok, if knowledge someone has a gun derpters threats, then I should consider all reports of police offices killed by guns to be fake news."

To clarify, you're arguing that the existence of any officer with a gun being shot demonstrates that there is no deterrence effect. And you are furthermore doing this on a economics blog called *Marginal* Revolution,

Will you next be claiming that price increases do not deter purchases because demand does not drop to zero?

Mulp,

I've never seen anyone give 10 paragraphs entirely over to a single logical fallacy of composition before. Fascinating.

Are you kidding? He usually does 15-20 paragraphs of that. I wonder if he's feeling under the weather.

>the real issue is the interpretation of the survey

I'd say the real issue is that you don't like the facts that the survey supports.... so much so, in fact, that you had to write this post and tell us what the "real issue" is.

As Alex says, statistics is useless in these circumstances. Even if the result is true, you can't trust the numbers on a statistical basis.

Anecdotal information is all there is.

A similar proportion of survey respondents admitted to homosexual contacts in the era of those surveys. Given Tabarrok's analysis, shouldn't we believe that very few people are actually queer, and most of the supposed homosexuals in America are just straight survey-mocking pranksters?

In fact, Gary Kleck addressed all of Alex Tabarrok's objections long ago.

This +100. It would be really helpful if AT took the time to familiarize himself with the discussion about this very topic that took place at the time. This subject was discussed, Kleck has responded at great length about which way the errors go and at what magnitude. Maybe Kleck is wrong, but it does no one any good to simply pretend the discussion never took place.

It sounds high. And though unscientific my reasoning is as follows: I know many people and many gun owners. Outside professional life (law enforcement, military etc. ) I know zero people who have a story where they've pulled a gun on someone to stop something.

Adding the quesiton "Do you know anyone who, during the last 12 months, confronted another person with a fire arm, even if he/she did not fire it, to protect his/her self, his/her property, or someone else? If so, how many?” Given the rarity of the event and the low probability of survey respondents knowing the same people, this would expand the pool of information quite a bit. You know people who use their weapon in self-defense aren't going to be quiet about.

"You know people who use their weapon in self-defense aren't going to be quiet about." [CITATION NEEDED]

They would be if they feared the defensive gun use was illegal, such as owning the weapon illegally. There are lots of reasons people might want to be quiet about a DGU.

"I know many people and many gun owners." How many of them are poor or live in bad neighborhoods? I can tell you from experience, the likelyhood of using a firearm in self defense goes way way up if you live in a high crime neighborhood.

Ripping off your drug dealer.

Robbing the 7-11.

Defending your gang territory.

Chasing off residents protesting your corner boy drug dealing, or your whores soliciting customers.

None of those things are a defensive gun use.

Anecdotes aren't a great source for information here, but I know two people who have pulled weapons (once a gun, once a knife) and successfully chased away would-be attackers.

Adding to this: I have been told by zero people that they've used a gun defensively. And that's not something many would keep secret.

I know a half dozen people who have used guns defensively. I knew some of them for many years before I learned of it.

YMMV.

Either you have jumpy friends or you live in a fairly violent area and I would suggest moving.

I know three people. One, a friend of mine, who didn't pull his gun until the 3 guys going to rob him knocked him to the ground (not jumpy). Second is a guy who lives in a mediocre neighborhood. 3x stopped his roofing business shop from being robbed, once got robbed on the street, but didn't have his gun. And number three was an older police detective that shot an armed guy trying to rob him getting off the bus while he was coming home from work.

All of that was over 30 years though. I live in a safe neighborhood, but not everyone does, and a lot of people need to travel into less safe neighborhoods occasionally.

Closest thing to self defense with a gun I know of is an incident that happened to my step mother and sister back in the 90s. A burglar climbed through their kitchen window one night, perhaps thinking the house was unoccupied. Ma was still awake reading in bed and hearing a noise she got up to investigate thinking maybe the cat was up to to mischief. Instead she found the burglar standing in the hall at her door. She tried slammed her door shut, but the guy grabbed her arm, so she had his arm through the door holding on to hers. She screamed for my step sister to wake up, and specifically "get the gun!" When the burglar realized there might be be another person home, and they might be armed, he fled. That did not involve actual weapons use or even display: the gun could have been nonexistent (it wasn't, but I'm not sure it was available in my step-sister's bedroom either).

If you live in a major urban area such as NYC or Chicago, telling someone about your illegally owned gun could get you sent to prison.

"I have been told by zero people that they've used a gun defensively. And that's not something many would keep secret". How many people have you asked whether they've used a gun defensively? Do you expect most people you know would volunteer that information outside of a conversation on the topic? I know people who have used a gun for defensive purposes, and I knew a couple of them for years before discovering that information when it came up in a relevant conversation. Most people that know me don't know that I've used a gun for defensive purposes because (1) it wasn't that big of a deal (nothing worth bragging about) and (2) it would just be weird of me to bring it up outside of proper context. That said, I thought Alex' analysis was excellent, and cannot begin to imagine the ~1% figure as close to representative of the general population (excluding police, military, violent criminals).

Actually considering the accusations of sloppy research or outright fraud that are regularly directed at researchers who come up with findings supporting positive effects of gun ownership, the fact that the CDC backed up their findings (and then hid it) is huge.

Tabarrok has criticized police union privileges

https://marginalrevolution.com/marginalrevolution/2018/01/police-union-privileges.html

but he has also called for doubling the number of police in the United States based on his own work showing the effectiveness of police

https://marginalrevolution.com/marginalrevolution/2012/12/police-crime-and-the-usefulness-of-economics.html

Art Deco is off base.

So by the same reasoning, I should revise downward my estimate of the number of trans individuals in the US? After all, we are talking about a smaller rate than Kleck reports and plenty of people who might lie for social justice reasons or just gits and shiggles.

Or perhaps I should increase my estimate of falsely reported rapes? After all the highest estimates of rape rates still allow for more women lying that they have been raped than those lying that they have not been raped?

Exactly when should I apply this heuristic? Hate crimes rates? Voter suppression?

It seems awfully selective to use this argument only for gun statistics and not for all the other statistics floating around that are of similar magnitude and power.

Yes, as Scott Alexander puts it:

"When we’re talking about very unpopular beliefs, polls can only give a weak signal. Any possible source of noise – jokesters, cognitive biases, or deliberate misbehavior – can easily overwhelm the signal. Therefore, polls that rely on detecting very weak signals should be taken with a grain of salt."

http://slatestarcodex.com/2013/04/12/noisy-poll-results-and-reptilian-muslim-climatologists-from-mars/

Which is why Kleck took such great pains to weed out false positives.

How many of those incidents involved someone using a gun defensively when there was no actual threat? E.g., telling a salesperson to get off your property? drunkenly confronting someone over a minor dispute? Chasing off someone who was never really a threat to being with? When we hear about some unfortunate person shot because they rang the wrong doorbell while looking for help, does that not also count as defensive use? That is what the person doing the shooting thinks.

Also, if you simply extrapolate from that statistic, you would assume that half the population has used a gun defensively over a 40-50 year period, and that over a lifetime, the majority of people would have used a gun defensively, but obviously that cannot be correct. The number/percentage of gun owners does not seem to corroborate the straight extrapolation of defensive gun uses. I.e., there are not enough gun owners to assume that half the population has used a gun defensively. The number of people who used a gun defensively is limited by the number of gun owners.

Anecdotally, I'm in my 40's and I can only think of one person who *may* have used a gun defensively. But maybe my context is safer, or more gun-free, or both, or something else altogether. I would guess there is a lot of regional variation in defensive gun use.

This statistic, even if "acccurate" doesn't say very much. I would also want to know

-What is the regional variation in reported incidents?
-How does that correlate with gun ownership? (I would expect areas of higher gun ownership to have higher defensive use)
-How does defensive gun use correlate to overall crime patterns? (Can anything useful be learned from that correlation? Perhaps defensive gun use (and the prospect thereof) naturally limits/inhibits crime. How do try to isolate for effect?)
-What was the perceived threat in the use cited?
-Did the use of the gun actually prevent a crime from happening?
-was there only one time a gun was used defensively? Or has the responsdent used a gun defensively multiple times?
-what is the breakdown between peoople who have a used a gun once versus more than once?
-what is the age of the person at the time of use?
-did defensive use use include use against wildlife?
-Also, for comparison, how many times has a person used some other weapon? How about a car? Defensive use of a car? It sounds ridiculous, but I'm willing to bet the answer is not zero.

"Also, if you simply extrapolate from that statistic, you would assume that half the population has used a gun defensively over a 40-50 year period, and that over a lifetime, the majority of people would have used a gun defensively, but obviously that cannot be correct."

Perhaps DGU's are not distributed evenly amongst the population. Perhaps the people who do have a DGU all have a similar set of characteristics, like living in a bad neighborhood or working a job like convenience store clerk where there is a higher probability of crime.

Your right that what you said obviously cannot be correct, but it is incorrect because you assume that DGU's fall evenly across the populace.

In other words, most gun use in parts of Chicago and Philly, etc, and justified defensive use of guns, and if anything, these areas need lots more guns to stop gun violence by increasing the fear more people have guns so guns are used even more in self-defense??

Yes any gun use, even one that is clearly self defense, would have been a crime in Chicago or New York or Philly at the time. It is therefore reasonable to conclude that legit self defense usages in those places would be under reported.

So when you tell me that 1% of people are LGBT, it might actually be .1%? Or if you tell me that .1% of people are trans, it's actually .01%?

Kleck's study took great pains to control for false positives, by asking many follow up questions, carefully noting details and having other researchers call back later and confirm that the details still match.

Again, it would be helpful if AT at least took the time to understand what Kleck's reasoning was rather than just assume that he did not even contemplate the problem of studying rare events.

Here is the origional Kleck study:

https://pdfs.semanticscholar.org/91da/afbf92d021f06426764e800a4e639a1c1116.pdf

Here is Kleck's response to David Hemmingway, who made the exact same claims about 20 years ago:

https://web.archive.org/web/20130903165308/http://www.saf.org/LawReviews/KleckAndGertz2.htm

"It is worth emphasizing how difficult it was for our RS to falsely report a completely nonexistent event as a DGU. Unlike the UFO example that H insists is somehow parallel to reports of DGUs,[16] a respondent who wanted to falsely report a nonexistent DGU could not qualify as having had such an experience merely by saying "Yes." Rather, respondents had to provide as many as nineteen internally consistent responses covering the details of the alleged incident. In [Page 1450] short, to sustain a false DGU claim, RS had to do a good deal of agile mental work, and stay on the phone even longer. On the other hand, all it took to yield a false negative was for a DGU-involved R to speak a single inaccurate syllable: "No." The point is not that false positives were impossible, but rather that it was far harder to provide a false positive than a false negative."

We have been through all of this.

"The deep problem, however, is not miscodings per se but that miscodings of rare events are likely to be asymmetric. Since defensive gun use is relatively uncommon under any reasonable scenario there are many more opportunities to miscode in a way that inflates defensive gun use than there are ways to miscode in a way that deflates defensive gun use."

Also covered in the original study. You cant simply assume that false positives are common and false negatives are rare, especially when dealing wtih something in which the respondent might feel they broke the law.

Again, from the original Kleck study:

"Even under the best of circumstances, reporting the use of a gun
for self-protection would be an extremely sensitive and legally controversial
matter for either of two reasons. As with other forms of forceful
resistance, the defensive act itself, regardless of the characteristics
of any weapon used, might constitute an unlawful assault or at least
the R might believe that others, including either legal authorities or
the researchers, could regard it that way. Resistance with a gun also
involves additional elements of sensitivity. Because guns are legally
regulated, a victim's possession of the weapon, either in general or at
the time of the DGU, might itself be unlawful, either in fact or in the
mind of a crime victim who used one. More likely, lay persons with a
limited knowledge of the extremely complicated law of either self-defense
or firearms regulation are unlikely to know for sure whether
their defensive actions or their gun possession was lawful."

"an R who reports a DGU may believe that he is placing
himself in serious legal jeopardy. For example, consider the issue
of the location of crimes. For all but a handful of gun owners with a
permit to carry a weapon in public places (under 4% of the adult
population even in states like Florida, where carry permits are relatively
easy to get)28 , the mere possession of a gun in a place other than
their home, place of business, or in some states, their vehicle, is a
crime, often a felony. In at least ten states, it is punishable by a punitively
mandatory minimum prison sentence. 29 "

Great post. Important and often overlooked statistical consideration.

In other news, approximately 1.4% of Americans surveyed reported having beaten up a whole room full of bad-ass gang dudes with guns using their kung fu skills at least once in the past year. And 2.3% of American men surveyed report having junk that is at least 14 inches long if not way longer.

And 25% of American female college students report having been raped.

... erm, I think maybe you failed to read the main post, which is about small-likelihood (1-2%) events and how self-rep survey data can’t capture anything meaningful about them. 25% is ... well, it’s not 1-2%. Not sure what else to say.

I don't know a single person who has used a gun defensively. I would think the majority of people who use guns defensively live in "bad" neighborhoods with high crime rates, so that rather than extrapolating that over a lifetime, 40-50% of people would defensively use a gun, we would see a much smaller % of people who use guns defensively several times. Also, I would suppose that the majority of people who have defensively used guns are also guilty of offensive gun use, but those people are less likely to answer surveys.

Some people lie about it to survey takers, when it DID NOT happen because a) it supports gun rights to lie about it and b) maybe they think it sounds tough or cool or something.

And other people lie about it to survey takers when it DID happen because it really did happen and they don't tell the survey taker.

Which effect is larger?

Kleck responded to this a bit: (H is David Hemmingway)

" He insists that the predominant bias surrounding DGU reports is a "social desirability response bias"[44] with RS making false reports of DGUs to present themselves as "heroic."[45] H ignores the information we provided in our article on the distinctly unheroic character of the DGU accounts provided.[46] What was most striking about the reported events was their banality. If H's speculations had merit, false portrayals of heroism should have involved frequent claims of facing down gun-wielding bad guys and exciting shootouts. In fact, RS reporting DGUs claimed to have faced adversaries with guns in only one-in-six cases,[47] claimed involvement in a shootout (both parties shooting) in just 3% of the cases,[48] and usually reported opponents with no weapons at all.[49] Likewise, they rarely boasted about their deadly shooting, with only 8% even claiming to have wounded an adversary.[50]

In any case, H is again focussing on a red herring. The issue is not whether some RS might think DGUs are heroic (this is undoubtedly true for at least a few people), but rather whether this sentiment is so strong and pervasive that it would, on net, outweigh the seemingly more common and natural tendency to conceal one's illegal behaviors from strangers who call on the phone. "

When I was in highschool taking the first class in statistics, my teacher told us that for survey questions that are controverisial like this, or if the responders worry that answering the question truthfully may put them in danger (like, are you having an affair?), the smart people of America have an amazing approach: they first produce a biased coin that turns head 55% of the time, ask the responder to toss the coin, if it's head, they have to lie, if tail, tell the truth.
He went on to explain that in this manner, and if we explain clearly to them, the responders understand that on individual level, his answer is just random noise, so there is no need to lie and it's not possible to influence the result, so they are more likely to do it honestly. Collectively, the correct rate can be inferred by simple math.

Then I got here and nobody in the US seems to be doing things that way, so I figure that the teacher might lie to us. But it seems to be legit, why it is not used when there are so many surveys out there tainted by lying?

2.5 million seems correct. 800k Cops, 1.2 million military. 500k Zelous Patriots. THat stat sounds correct and worth further study.

Both studies excluded GDU's from cops, military, security, etc.

Related:

"The Border Patrol counted an incident in which people allegedly threw objects at seven agents as "126 assaults." The math: 7 agents, multiplied by 6 perpetrators, multiplied by "3 different types of projectiles (rocks, bottles, and tree branches)."

https://t.co/g6d2QKRl3x

One way to get SOME kind of handle on bogus reporting is to ask essentially the same question, but with different time periods. i.e. Ask whether folks have used a gun for defensive purposes within the last 12 months. Ask also (either on the same or a different survey) about usage within the last 5 years. If most of the "yes" responders are miscoded, then the rate should be fairly similar to both questions (i.e. ~1% or whatever).

If the yes respondents are mostly genuine, then extending the time horizon to 5 years should substantially increase the yes responses. (Probably not by 5x for various reasons, but by some quite significant amount...)

This would help a lot.

It might not tell you exactly what's actually going on, but it could certainly flag inaccuracy in the responses.

Kleck did that in his original study. I dont believe the CDC did, but i dont know for sure.

I think they should also ask
In the questionnaire
Whether the respondent
Has been
Visited by Martians.

And did they chase the Martians off with their gun.

They were undoubtably protected from this by the “Martian Free Zone” signs that are so common among safe spaces theses days.

It's a safe zone,
No doubt,
Because they defend themselves
With
Spaceguns.

Problem: One man's "self-defense" is another man's provocation!

Looks like some of it may be in part due to a Lizardman's Constant.

- “Four percent of Americans believe lizardmen are running the Earth”
http://slatestarcodex.com/2013/04/12/noisy-poll-results-and-reptilian-muslim-climatologists-from-mars/

Again, if they did that, it would at least require a fair amount of effort on the liar's part to keep the details consistent across multiple followup questions and follow up calls.

That last part is important, in order to believe that Kleck's results are driven by malicious respondents, you have to believe that on a unsolicited random call, someone would invent a story about a defensive gun use (mostly banal story, by the way, see above), keep the details of that lie consistent over several follow up questions, then remember all those details for a follow up call at a later date, all for the purpose of messing with some research that they would likely never see.

Its not impossible, but it seems pretty unlikely that you would get a large number of these type of respondents.

This +1. It seems like, at least, a first-order attempt to answer the method problem and deserves to be taken seriously.

In the mid-1990s, Kleck and Gertz (1995) estimated that in a typical year about 1.3% of US adults used a gun for self-defense against another person.

On the face of it, this statistic would mean that most Americans use a gun for self-defense against another person in their lifetime. This does not seem credible.

So either most of those 1.3% were macho bragging idiots, or some small fraction of Americans are far more likely to need a gun to defend themselves than others. The people most likely to fall into that catagory are criminals, defending themselves against other violent criminals.

Here's some 2016 data, from Justice Department's Bureau of Justice Statistics, on the incidence (about 1.3% of the population per year) of violent crimes, and how often they are reported by the victims of the crimes (42%). I suppose that this would mean, by the reasoning above, that most Americans are the victim of a violent crime in their lifetime.

"In 2016, U.S. residents age 12 or older experienced 5.7 million violent victimizations, including rape or sexual assault, robbery, and aggravated and simple assault. This was a rate of 21.1 violent victimizations per 1,000 persons. An estimated 1.3 percent of U.S. residents experienced one or more violent victimizations in 2016.

The NCVS collects data from residents on crimes both reported and not reported to the police. Fewer than half (42 percent) of the violent victimizations committed in 2016 were reported to police. Aggravated assault (58 percent) and robbery (54 percent) were more likely to be reported to police than simple assault (38 percent) and rape or sexual assault (23 percent). Sixty percent of the 480,940 nonfatal firearm victimizations were reported to police in 2016."

https://www.bjs.gov/content/pub/press/cv16pr.cfm

“. . . miscodings of rare events are likely to be asymmetric. Since defensive gun use is relatively uncommon under any reasonable scenario there are many more opportunities to miscode in a way that inflates defensive gun use than there are ways to miscode in a way that deflates defensive gun use.”
No, opportunities to miscode are independent events and so equally likely to inflate or deflate defensive gun use. What Alex is actually judging is people’s motivation in making their reports – he suggests that inflating a report of defensive gun use is somehow more satisfying to the reporter than deflating such report. What evidence?

If you accept that "defensive gun use is relatively uncommon under any reasonable scenario" then there are many more people who have the opportunity to falsely claim that they used a gun defensively than there are people who have the opportunity to falsely claim they did NOT use a gun defensively. On the assumption that people are equally likely to lie in either direction, it's overwhelmingly likely that a sample has more people lying about using a gun than lying about not using a gun. To come to a different conclusion you need to make judgements about people's motivations and rates of lying in one direction or the other. What evidence?

Americans are not a violent culture. Indeed, with all those guns out there, I'm amazed that there are any surviving Americans. If I owned a gun, I would have taken out more than my share of the assholes. Americans have to be the most disciplined and peace-loving people on the planet.

Nope, even you would have acted like an adult if you actually had a gun in your hand. This is more evidence that you live in a fantasy land.

The definition of "self-defense" may be an issue. Some people may think it's self-defense and they need an AR15 to properly threaten the neighbors' kids to blow their brain out when they come to their property and pick up that ball. I certainly think it's not.

I see nothing in that twitter thread that speaks to what AT is talking about.

My biggest problem with using Kleck's research to estimate DGUs is that survey is too old to be useful, as it was administered at a time when violent crime rates were much higher than they are today. It seems implausible given our historically low crime rates today that there would be anywhere near 1,000,000 DGUs a year.

Comments for this post are closed