We normally don’t post a blog entry for a single academic paper, but this one is important because it is the largest well-conducted survey concerning gun ownership and self-defense with guns (as well as other topics) to date. ¹ The survey size makes most of the data unassailable, and thus “gold standard.”

Take-aways

• 54,000 people contacted, 16,708 gun owners within that group.
• Low-side concurrence with other surveys concerning defensive gun use (DGU).
• New insight into public carry and self-defense, both public and private.
• Scale data for assault weapons (AR-15s specifically) and magazine capacities more than 10 rounds.

Why this paper matters

Much of what we know about guns has some limitations:

ESTIMATES FROM MIXED SOURCES: The best estimates we have had traditionally about the number of guns in private hands came from (a) estimating the “original” stockpile from dusty manufacturing records and (b) adding import/export/manufacturing records the federal government started keeping in the late twentieth century. In short, this was a very educated guess, but a guess nonetheless.

SMALL SURVEYS: It costs money to survey people. Underfunded criminologists traditionally have conducted surveys with the smallest viable sample size to gain insight. A typical survey might include 2,500 people, whereas this survey covers nearly seven times that amount, which adds precision.

INDIRECT MEASURMENTS: Suicidal Americans principally use guns (though the USA suicide rate is only slightly above the international mean). As such, for state-by-state gun ownership estimates, the suicide rate was used as a proxy for gun ownership rates. Now we have much more solid state-level estimates about all varieties of guns.

The point here is that this work will likely be a cornerstone of research going forward. Anyone not using it as reference data might be trying to conceal something.

Gun ownership

There is a difference between individual and household gun ownership rates.

Criminologists traditionally have measured household gun ownership rates (“does one of more people in your household own a gun”). The reason was that for home self-defense, the primary motivation for gun buyers, one gun could be used by anyone in the household.

This study measures individual gun ownership rates, ignoring households. Currently in the United States, there are about two adults age 21 or over in each household, so we could do some guesswork about household rates, but that is for another day.

What we see is that nearly 32% of all adults in the USA own one or more guns. Naturally, some own more and some own many.

The question is, does this agree with measures of household gun ownership rates? One example, provided by the paper’s author, was: “If you have two households, each with a husband and wife, and only one of those four people owns a gun, then if you polled them all you would find 25% individual ownership and 50% household ownership.” He also opined that his data would likely lead to a roughly 45% household gun ownership rate, which is aligned with those recorded by Pew, Gallup, ABC, and other sources.

The author polled for how many of each type of firearm people owned, and we’ll just say the range goes from one gun to over 100. What’s important is that the tally of the public stockpile of guns can be well estimated. At least for handguns, this closely aligns with the estimates the Gun Facts project has been tracking for a number of years (our latest estimate was 173 million).

The last ownership item to mention, because it feeds into the next section, is magazine capacities. Forty-eight percent of gun owners have, at one time or another, owned magazines that hold more than 10 rounds; that’s about 39 million people. Additionally, for handguns at least, about two thirds of the magazines owned by any given individual held above 10 rounds.

This is a little slippery because some states have banned magazines with larger than 10 rounds capacity. Some people may have surrendered or destroyed them, while others may have stored them out of state. The author suggests, and we agree, the 48% figure for 10+ round magazine capacities might be an upper bound.

So why do so many people own magazines with higher capacities? It is because of…

Self-defense

They want larger capacity magazines for self-defense purposes, thanks to the uncertainty of what they might face, and the extremes that some do face.

About half (51.2%) of the surveyed gun owners who have performed a DGU faced more than one perpetrator/assailant. This is quite profound in terms of preparedness for self-defense. Americans appear to sense that in the face of both uncertainty and a demonstrated recurrence of multiple attackers, it is better to have and not need extra rounds of ammunition, than to need and not have the same.

As seen in other polls on the subject, the number of DGUs and the lack of bloodshed in them is why there may be a public misperception of the net-positive effect of guns vis-à-vis crime. In this survey, 31.1% of gun owners (25 million people) have performed a DGU, but in 81.9% of the cases they never pulled the trigger.

This agrees with an older and respected survey of DGUs ²that estimated 92% of DGUs involved merely brandishing or firing a warning shot. Likewise, another older study ³ showed that 68.7% of the time brandishing scared off a perpetrator, and if that didn’t work, 32.4% of the time pointing the gun at the perp, but not pulling the trigger, did the job (most people who did the latter, first did the former).

click to enlarge

The author’s estimate for annual DGUs is 1.67 million, which is a little lower than the previous aggregate two million DGUs in previous studies. However, the author notes that this is a conservative estimate due to the way the question was asked (“Have you ever defended yourself or your property with a firearm, even if it was not fired or displayed? Please do not include military service, police work, or work as a security guard.”). They then estimate the annual rate from the lifetime number of DGUs.

This is our only nit to pick with this study. A detour question after this, asking about DGUs in the past 12 months, would have avoided the calculations required to derive their 1.67 million estimate. In their calculation they had to adjust for age issues, but most importantly, the responses only came from current gun owners. This makes their estimate low by some undeterminable degree. But that this  estimate falls in the center of all other studies helps confirm the general rate of annual DGUs in America.

Interestingly, only 66% of the respondents who had performed a DGU did it more than once in their lives. Nearly 33% had done so twice, 12.6% thrice, and 7.8% five or more times. That said, while most people (68.9%) have never performed a DGU, most of those who have performed a DGU (44.1%) have done so only once.

The big surprise was women. They are not far off from men in DGUs, with 27.3% of them having performed a DGU compared to 33.8% of men – meaning men perform DGUs a mere 24% more often than women. And while we are on demographics, somewhat unsurprising given the conditions within American inner cities, blacks performed DGUs at a rate of 44.3% compared to whites at 29.7%.

Where DGUs occurred is also informative. A full 83.9% occurred in the home or on personal property, or the the homes and property of someone else. But 53.9% occurred outside of the home but on one’s own property. This is a difference with no meaning because a homeowner, in most states, can step out of their house to initiate a self-defense intervention. Likewise, someone who gets jumped after pulling into their driveway fits this category of DGU location.

One the flip side, 9.1% of DGUs occurred in public. If we take the older aggregate of studies that showed two million DGUs per year, this equates to 182,000 public DGUs annually. That alone is an astounding statistic inasmuch as this rarely enters the public perception. Likewise, the public is largely unaware of…

Public carry

… how many people legally carry guns in public. This study concludes that 20.7 million people carry handguns in public. Of gun owners, 56.2% indicate they carry/carried at some point. An interesting addendum is that of those who carry, and ignoring incidents where a gun was used, 31.8% felt that the mere presence of the gun deterred a crime. Ponder that – no brandishing, no pointing the gun at a perpetrator, no firing the gun. Maybe the gun “printing” under a shirt or just showing that a gun was in a waistband was enough.

The whole DGU issue gets to the last interesting datum. Of all the motivations recorded for owning larger capacity magazines, and aside from a barely larger vote for recreational shooting, defense in and out of the home topped the list – 41.7% for outside the home and 62.4% for at home. This result aligns with other studies about why people buy guns at all.

A note about methodology

The author used an outside polling company to do the work, which is good. They constructed the survey instrument well and tested it on one state before going nationwide.

This company has an online “panel” of respondents who can ignore requests to take a survey in the same way someone can ignore a “random digit dialing” call from a survey taker. However, people who join panels for survey companies are likely not in low-income brackets or disadvantaged neighborhoods. This might seem like a flaw, but those same people don’t often take telephone surveys, so this might be a wash.

It is the panel approach and conducting the survey online, instead of on the phone, that makes this a more robust approach… though some will disagree. The panel has the time to take long surveys at their leisure, which produces a very low drop-out rate (less than 1%).

Factors combined, this allowed for a comparatively huge sample size, which makes the outcome both more robust and the numbers for ownership or DGUs to be slightly conservative.

A final note

click to enlarge

We pulled the state-by-state gun ownership data from the paper and fetched gun homicide rates from the CDC. The covariance calculation of gun ownership rates and gun homicide rates shows there is none (R² of 0.01, where 1.0 is complete covariance and 0.0 is no covariance).