By Mike Donatello
Testimony from Baltimore Baltimore County, Md., Police Chief James Johnson:
“From November 2011 to November 2012, an estimated 6.6 million firearm transactions occurred
without a background check. Up to 40 percent of firearm transactions occur through private
individuals rather than licensed gun dealers. Allowing 40 percent of those acquiring guns to
bypass background checks is like allowing 40 percent of airline passengers to board a plane
without going through airport security.”
The gun-control bunch likes to argue that four of every ten firearm purchases in the U.S. are made through channels in which the buyer is not subject to a background check through the federal NICS system. That’s a lot guns being purchased in absence of a background check, they argue, with many coming through the infamous “gun show loophole.” As with many “facts” offered by gun grabbers, however, this one’s not only misleading, it’s based on a suspect source.
Advocates for gun restrictions toss around statistics like they’re magic talismans which confer instant credibility in any debate. The favored source for factoids on gun purchases – including the proportion of transactions subject to background checks – is the National Institute of Justice report, “Guns in America: National Survey on Private Ownership and Use of Firearms.” Published in 1997, the report is based on an even-older piece of research, the “National Study of Private Ownership of Firearms in the United States,” a survey sponsored by the Police Foundation in 1994.
You read that correctly: this is research that is two decades old.
Given the changes in Americans’ shopping habits, economic conditions and other potential influences since 1994, the mere age of the study would render it suspect, assuming the data did, in fact, support the anti-gunners’ fabrications. In truth, the data support no such assertions.
As social-science research goes, the 1994 work was conducted in keeping with generally accepted guidelines for survey-based studies. Interviewing was conducted by telephone among U.S. adults, with data weighted to reflect some basic demographics and produce estimates projectable to the country’s adult population. The study boasted a response rate of “approximately 50 percent,” which was reasonable in 1994 and would be looked upon quite favorably if obtained in a study conducted today.
Those basics aside, however, there are some concerns with the mechanics of the study that should preclude its use as a Delphic source by the Obama-Biden-Feinstein crew.
First, there is no indication that researchers took into account geographic or social variation in how guns are obtained. Are gun shows more popular in the South or West than in the Northeast? Are firearms more likely to be obtained from family or friends in rural areas than in urban centers? My personal experience would say that is the case for both examples, but I don’t have any hard numbers to back that up and, apparently, neither do politicians relying on this study. So, instead of useful information to help us understand the nuances of our diverse nation, we instead have a few weighted averages.
Second, the actual number of interviews of gun owners, from which pols make pronouncements about the volume of gun purchases gone unchecked, is relatively small: just 251 interviews among the total of more than 2,500 completed interviews in the study. Although, in general, one can draw conclusions from a subsample of that size, the resulting estimates are subject to markedly greater error than would be the case with more interviews available for analysis. That’s why, when survey researchers want to delve deeply into results from specific subgroups of respondents (e.g., looking at where gun purchasers are obtaining their firearms), they will typically sample substantially greater numbers of respondents from those subgroups, and use statistical methods to weight the results back in line with total U.S. population patterns.
The result of that exercise, called “oversampling,” is more robust – that is, more reliable – estimates of details like purchase source, without accompanying distortion of figures caused from interviewing too many members of one particular subgroup. Again, however, the 1994 study does not appear to have oversampled among gun purchasers in a way that would have bolstered reliability of the data, opting instead for a simple minimum quota of approximately 250 interviews. So, to continue our previous example, if breakdowns by state were of interest, we’d be stuck with a potential average of five interviews per state, because no one worked to ensure that the study was conducted in a way that would allow meaningful drill-down on specific topics of interest. Not good. And, although the gun-control crowd is well aware of the limitations of the research – remember, it’s been around for almost 20 years – they still persist in drawing conclusions that the study was never designed to facilitate.
Perhaps most disturbing isn’t the technical details of the sampling scheme, as those can always be couched as a trade-off between lesser-quality data and cost savings, or the vintage of the study. The worst part is that the gun-ban crowd flagrantly misrepresents the proportion of transactions in which a background check would have occurred, creating a “40% of gun purchases are unchecked” myth by simply lying about the data.
The table below displays the source-of-acquisition detail from the 1994 report (i.e., “Exhibit 5. Methods and Sources for Gun Acquisition in 1993 and 1994”). From the upper section, we know that six of 10 guns purchased by respondents to the study came from a traditional retail setting of some sort. No problem there.
| Source | Proportion of Total Guns Obtained |
| Gun store | 43 |
| Pawnshop | 6 |
| Other store | 11 |
| Subtotal from retail | 60% |
| Gun show or flea market | 4 |
| Through the mail | 3 |
| Member of the family | 17 |
| Friend or acquaintance | 12 |
| Other | 4 |
| Total | 100% |
Note, however, that only four percent of owned firearms were purchased at a gun show or flea market – hardly a huge source of unchecked purchases. So what’s left? The ambiguous “other” four percent of purchases (about which no detail is offered in the report) and the 29% of gun acquisitions obtained through family, friends or acquaintances.
The respondents of the survey were anonymous and there was no way to verify if the data they provided are true.
It’s clear that the universal-check lobby is claiming that a gun purchase not made on a retail premises is not subject to a background check – an assertion that is false on its face because, of course, any gun purchased from an FLL-licensed dealer at a show or flea market still would be subject to NICS checks. Likewise, any “through the mail” purchase would have been processed by an FFL as well, if it was a dealer-associated sale. What’s also clear is that the true motive here appears to be regulation of private-party transactions, as those are the only major source of gun acquisitions that are not covered by the NICS system.
So, despite potential issues with the research and the likelihood that the original study has passed its useful life expectancy, the real offense – committed with intent to mislead – lies in quoting numbers out of context, claiming empirical justification for expansion of regulation where no such justification exists.
Mike Donatello’s 27-year career in opinion and marketing research includes senior roles at USA Today, the Washington Post Company and other organizations. He is currently completing a doctorate in media research at the University of North Carolina at Chapel Hill.