A study performed by the RAND corporation has been picked up by leftists on social media. The claim is for the number of households who have a firearm in the house. Read:
I replied that I think these numbers are wildly inaccurate. For example, Florida has a claimed household firearms ownership rate of 35.3%. Florida has approximately 8,752,810 households, so the claim is just over 3 million households have a firearm in it. Here is why that number is demonstrably inaccurate:
- Florida does not require concealed weapons permits, yet 2.5 million people in the state have a CCW.
- Since the Brady Act begin requiring NICS checks in 1998, Florida has seen more than 30 million of them performed.
Those two data points alone point to more than 3 million households with a firearm. Still, I pointed out there is no way to be sure, because no one keeps a registry of firearms and their owners. The leftists made the the claim that the numbers were close to accurate because:
- Itβs likely from firearms purchased through an FFL where data is collected through the 4473.
- The FBI does the background check. A transfer to you is in their database.
- It could be a household of five all with five firearms but it is still just one household with a firearm.
So to boil it down: The ATF has your data because they somehow have a record of your gun purchases because 4473. Even though that is illegal, and once a firearm is sold by a dealer, no one knows for sure who it was sold to, unless it somehow winds up being resold to a dealer. That’s why the left pushes so hard for universal background checks.
I pointed out the law in the second- the FBI doesn’t do NICS checks, the state police do. He replied how the FBI is better and more trustworthy with your data than the FDLE. Whatever, that still doesn’t validate the above state level data, which actually came from a RAND report created from summing up telephone surveys.
Then there is the antigun trope of claiming just a small minority of people owning a large number of firearms creating an illusion of widespread firearm ownership. That’s simply not true. I have already blogged about that ridiculous claim:
the top twenty percent of all gun owners actually owned 55% of the guns. In terms of absolute figures, ten million people owned 105 million guns β for an average of ten guns per person, and the remaining 87 million guns were owned by 34 million people β for an average of 2.6 guns per person. The population of the U.S. was 263 million in 1994 β indicating that only 16.7% of Americans had actually owned a weapon.
IΒ wrote about this in June of 2016, and I guessed at the time that there were somewhere between 400 and 500 million firearms in the US. Since that time, there have been another 310 million NICS checks.
Again, let’s look at Florida’s demographics- there are just over 8.7 million households in Florida, and those households contain 19 million adults. With over 2.5 million CCW permits, 200,000 known NFA items, and approximately 100,000 active law enforcement officers, there are just about 3 million people right there. That doesn’t account for people with granddad’s hunting rifle, criminals who have a firearm but aren’t going to tell some random survey caller, and others. On the other hand, there are households like my sister and her crotch fruit who don’t own any guns at all. Of course, none of them have a CCW, either.
Looking at the numbers, they seem to support a household ownership rate closer to 60% than to 35%. However, I also admit that my estimate is just as much guesswork as the RAND corporation’s report, because no one can know for sure.
9 Comments
Cederq · September 4, 2026 at 7:47 am
They also don’t take into account those that refuse to participate in “surveys” or acquire fire arms outside of FFL dealers, but still legal sales.
J J · September 4, 2026 at 7:57 am
I, like a lot of other people I know, would never tell some stranger on a phone call about my firearms ownership. If I actually owned any. π
Tsgt Joe · September 4, 2026 at 8:11 am
Dont know if snowbirds would affect your numbers much but a fair percentage of the midwestern snowbirds tend to have at least an old revolver in their rvβs.
bob · September 4, 2026 at 8:23 am
Reality-Firearms-min 1 billion,Ammo-min 10billion rds floating around,does’nt include reloaders
Joe Blow · September 4, 2026 at 8:43 am
I guarantee you that’s not accurate. TN allows firearm transactions between private parties without registration. They have NO IDEA how many firearms are out there. Then we can have the conversation about home-made AR’s that came from kits and never once saw a registration form…
Bob · September 4, 2026 at 8:54 am
Alabama has no idea that I’m even here much less if I have firearms.
landroll · September 4, 2026 at 12:15 pm
I’m like Bob. Any statement of the number of firearms in the wild is just a WAG.
hh475 · September 4, 2026 at 3:34 pm
The thing about these is that almost none are direct measures. Here’s how RAND says they did it:
“First, [the researchers] used a small-area estimation technique to create state-level ownership estimates for each of 51 nationally representative surveys assessing household firearm ownership rates. They then used structural equation modeling to combine these survey-based estimates with administrative data on firearm suicides, hunting licenses, subscriptions to Guns & Ammo magazine, and background checks into the final measure of household firearm ownership. ”
They are all inferential statistical models that are exquisitely tied to basic assumptions about how one statistic implies something about another statistic. If you look at the RAND study, the basic dataset is a “small area estimation technique” set. This means that rather than looking at gun ownership at site X, they took national data where they are more confident of the results, They then *infer* the household rate based on demographics, and then further *infer* the data about the states on the basis of various proxy measures, e.g. NICS data, Guns and Ammo subscriptions, etc. This is interesting because some of these proxies are pretty bad. In one study, the rate of Guns and Ammo subscriptions was *negatively* correlated to gun ownership.
Not only do you have to worry about whether or not the proxy data implies what you think it does, but you *also* have to worry about the quality of the proxy data itself.
Worse, this gets very circular when models like these are used to make correlations between things like crime and gun ownership. For instance, the RAND study uses gun suicide, homicide, and assault rates as proxies for gun ownership. Imagine a study that naively uses these estimates to correlate suicide and gun ownership!
Treefarmer · September 5, 2026 at 6:47 pm
60.1% for Idaho? We don’t know of a single household in ID without a firearm(s). Even the token democrats up here are armed. Hell, even the mennonites are armed.
Comments are closed.