METHODS

Six handgun dealers from each of the 20 largest cities in the US that had 10 or more dealers participated in the research. (A power analysis was conducted before data collection to determine the number needed to detect meaningful differences.) A list of US cities, ranked by size, was used to identify potential cities.¹⁷ We identified dealers listed in an internet business directory under the keywords of “guns”, “gun dealers”, and “firearms”.

Cities with fewer than 10 dealers listed were eliminated and replaced with the next largest city. Six firearm dealers were randomly selected from the listings for each city. Additional dealers were randomly selected when necessary (for example, when a dealer did not sell handguns). Dealers were grouped by region of the country into Northeast (Baltimore, New York City, Philadelphia), South (Memphis, Nashville, Jacksonville, Oklahoma City, Houston, Dallas, San Antonio, El Paso, Austin, Fort Worth), Midwest (Cleveland, Indianapolis), and West (Denver, Seattle, Phoenix, Los Angeles, San Diego). Population sizes ranged from over eight million (New York City) to about one half million (Oklahoma City).

Dealers within each city were randomly assigned to one of two interviewers, one male and one female. Interviewers called the sampled firearm dealers posing as potential handgun purchasers. Interviewers followed a predetermined script that began with “Hi, do you sell handguns?” If not, the dealer was thanked and the call was terminated. If the dealer said yes, the interview continued with one of three a priori randomly assigned conditions, specifically: (1) “I’m looking to buy a handgun for myself,” (2) “I’m looking to buy a handgun for my girl/boyfriend for her/his birthday,” or (3) the intentionally ambiguous situation of “My girl/boyfriend needs me to buy her/him a handgun”. The interviewer continued with “I’ve never done this before. What do I need to know?” The script and answers to anticipated questions (for example, “How much do you want to spend?”) were pilot tested with 10 gun dealers across the US who were not from the sampled cities. Each pilot call was observed by the other interviewer and an attempt was made to standardize tone, style, and other speech patterns.

The callers took notes during all interviews and completed a brief questionnaire immediately after each call. Participation rate was 100%. Clerks who answered the phone were not made aware that they were participating in a study.

Attempts to persuade can reasonably be expected to be used by potential purchasers, therefore, when a clerk was unwilling to sell a handgun to the caller, the interviewers were instructed to gently attempt to persuade him or her (for example, “Is there any way I can do this—is there any way we can work it out?”). Five of the 28 conversion attempts were successful; we classified the converted cases (n = 5) as willing to sell. Ten clerks gave a response that could not be classified as a clear yes or no which was recorded as a “might” or “maybe”. Given that an equivocal response is not a likely outcome in an in-person transaction, we adopted a conservative approach and required an affirmative response for inclusion in the “willing to sell” category; when a dealer indicated that s/he “might” sell, it was coded as a “no”.

Frequencies and Fisher’s exact tests were calculated to examine differences related to the manipulated variables, that is, interviewer gender and sales condition. Multivariate logistic regressions that took into consideration study variables (that is, interviewer gender and sales condition) and geographic region were used to predict dealer willingness to sell. Likelihood ratio tests were used to determine whether adding geographic region to the regression significantly improved the fit of the model. We also conducted descriptive and multivariate logistic regression analyses examining the role of the legality of the sale.

Under federal law, licensed firearm dealers may legally sell a firearm to any person who is not a prohibited purchaser (for example, a convicted felon) including guns that will be given as a gift.^7,18 In most cases, however, a licensed dealer may not knowingly sell a firearm to someone who the dealer knows is not the intended possessor, independent of the intended possessor’s eligibility.¹⁹ To determine whether it is legal under state laws for a dealer to sell a handgun when purchased as a gift, we contacted nine law enforcement agencies in seven different states. Responses ranged from the openly unsure (for example, “That’s a tricky question right there”) to the tentative (for example, “I’m pretty sure it’s legal”) to the accusatory (“I’ve been on the job 12 years and never had a person ask the law if they weren’t going to do something illegal”), to the incorrect (for example, “Federal law says no for handguns”). Given this inconsistent information, we sought the assistance of two attorneys, a former Assistant United States Attorney and an attorney specializing in firearm policy. They reviewed relevant federal and state statutes^18,20 and judged whether it was legal for the dealer to sell a handgun under each condition in the specific state. The independent evaluators differed only on whether the gift condition was illegal in Indiana. Data were analyzed with Indiana classified each way; substantive findings were the same. To facilitate presentation of the findings, results are presented with Indiana classified as allowing gift sales.

The research was reviewed and approved by UCLA’s Internal Review Board.

RESULTS

Willingness to sell

Most (70.8%) dealers indicated that they were willing to sell a handgun to the caller regardless of the stated end user of the gun. Willingness to sell differed by sales condition (p = 0.003). As shown in table 1, dealers were willing to sell a handgun 87.5% of the time when the caller stated that the handgun was for him/herself. (Dealer refusal to sell under this condition occurred when the interviewers, when asked, revealed that they did not meet residency or license requirements for the purchase.) When the caller stated that the handgun would be a gift, 72.5% of the dealers indicated they would sell the gun. One half (52.5%) indicated they would sell a handgun to the caller when told it was for a girl/boyfriend “who needs it”.

Dealers were equally willing to sell a handgun to male and female callers (70.0% v 71.7%). (When “maybe” was included as a category, male callers were more likely to be given an outright “no” (28.3% v 13.3%), whereas females were more likely to be told that she “might” be able to get the gun (15.0% v 1.7%).) Although the proportion of dealers who were willing to sell a handgun when it was for the girl/boyfriend was identical (62.5%) for the interviewers, the proportion differed according to the rationale provided: a higher proportion were willing to sell to the male than the female interviewer when the gun would be a gift (80.0% v 65.0%) but a lower proportion were willing to sell to him than her when it was for a girl/boyfriend who needs it (45.0% v 60.0%).

A finding that emerged from the data, not postulated a priori, is that of geographic region. Dealers in the Northeast were substantially less likely to agree to sell than dealers in other regions of the country (27.8% v 78.4%; p = 0.001), and those in the South were more likely than those elsewhere to agree to sell a handgun (81.7% v 60.0%; p = 0.015).

Handgun dealers and the law

Selling a handgun would be illegal under the “need” condition (n = 40), yet one half (52.5%) of the dealers were willing to do so. If the sale was legal, 80.0% were willing to do so. The legality of the sale was related to dealer willingness to sell (p = 0.003). Some dealers appeared to be aware of, but willing to work around, limitations placed by the law, as evidenced by the following exchanges: Interviewer: “Is there a problem with me buying it for my girlfriend?” Dealer: “As long as we don’t know about it. It’s personal business”. Or, “First thing, don’t ever tell anyone you’re buying a gun for someone else because it’s against the law”. In several cases, dealers suggested that the caller contact another specific dealer, indicating that they may be aware of someone who might be willing to sell under these circumstances.

Multivariate analyses

Multivariate logistic regressions were used to assess the independent effect of the type of sale, caller gender, and geographic region. The first regression tests the role of interviewer gender and type of sale (see table 2, model I). Dealers were less likely (adjusted odds ratio (AOR) = 0.38) to say that they would sell a handgun intended to be a gift than for personal use. Although this finding is substantively important, it is not of statistical significance (p = 0.11). Dealers were less willing (AOR = 0.16; p = 0.001) to sell a handgun if the caller indicated it was a purchase intended for a girl/boyfriend because s/he “needs it”.

Adding geographic region to the regression (see model II) improved the fit of the model substantially (p = 0.001) and the added variables were statistically significant. Dealers in the Midwest, South, and West were more willing than those in the Northeast to sell regardless of the recipient of the handgun or the gender of the caller (AOR = 21.30, p = 0.003; AOR = 18.74, p = 0.000; AOR = 8.93, p = 0.004, respectively). “Need” remained significant (AOR = 0.09, p = 0.001).

All “need” sales were illegal; the others were legal. Replacing sale type with sale legality in models I and II resulted in an AOR of 0.28 (95% confidence interval (CI) 0.12 to 0.63) for illegal sales; the AOR for interviewer gender remained the same. When region was added to the regression, the AOR for interviewer gender and each region did not change substantially. Illegal sales remained statistically significant (AOR = 0.20, 95% CI 0.08 to 0.53).

DISCUSSION

Firearm dealers who sell handguns are generally willing to sell to a potential customer regardless of whether the gun is for his or her own use or for use by another person. Gender of the purchaser was generally not relevant. Dealers in the Midwest, South, and West were more likely than those in the Northeast to be willing to sell a handgun regardless of interviewer gender and sales condition. Although dealers were less willing to make a sale when it would be illegal, more than half were willing to sell a handgun even when it would be illegal to do so.

This is, to our knowledge, the first study in the peer reviewed literature on factors that might influence firearm dealers’ willingness to sell a handgun. The strengths of the study include its design (for example, see Campbell and Stanley²¹). The population included all dealers listed in the yellow pages of an internet directory, an easily accessed source of information for many persons wishing to purchase a firearm. Participating dealers were selected from large US cities in states that contain over half (53%) of the US population²² and 35.5% of the FFLs (federal firearm licensees) in the US.¹

Several matters warrant comment when interpreting the findings. First, dealers’ stated intent may not correspond to their actual behavior. Although not necessarily a usual business practice, some dealers may say that they would sell a handgun while on the telephone but not do so if the potential customer was on-site. Alternatively, some dealers might resist persuasion attempts on the telephone but might yield to in-person pressure. Second, employees likely vary in their knowledge of and compliance with firearms laws such that if an interviewer had spoken with a different clerk at the same retailer, s/he might have received a different answer. Although personnel training may vary across and within stores, one “bad” clerk can implicate the entire dealership. Third, the effect of price on willingness to sell is not clear. If asked how much they were willing to spend, interviewers stated “about $300”. Handguns are available for less and for substantially more. Perhaps dealer behavior would differ depending upon the amount of money a customer was willing to spend. Fourth, the effect of dealer location (that is, rural or urban) is unknown because the sample included only urban dealers. Fifth, the effect of state level regulation of firearm dealers is not examined in this study. Sixth, we did not attempt to ascertain whether the participating dealers held FFLs. One could assume that, if listed in a telephone directory, the dealer is “regularly involved in the business of selling firearms at wholesale or retail” and should have an FFL.² It was not possible, however, to ascertain this information without arousing undue suspicion. And, finally, although only two or three dealers seemed to be suspicious of the interviewers, skeptics may wonder whether the dealers were simply “playing along” with the caller.

To address the latter concern, we made 20 additional calls after the study was complete. A dealer was randomly chosen from each city and randomly assigned to each interviewer. The interviewer opened with “My girl/boyfriend needs me to buy her/him a handgun because s/he isn’t allowed to”. That is, the caller was explicit about wanting to buy a gun for a prohibited purchaser. In 16 of the 20 calls, the dealer responded with an unequivocal “no” and commented about such a purchase being clearly illegal, a straw purchase, etc. Each of the four who agreed to sell a handgun appeared to recognize that the sale would be illegal. They said: (1) “As long as you have no record, you can come down here and pick one up and put it in your name”; (2) “You can do whatever you want after you walk out the door”; (3) “What you do with it is your business. Legally you’d be responsible for it, you’re more than welcome to buy one. You can’t transfer it to him—I assume he’s been turned down”; and (4) “She can’t come in, pick one out and you buy it. That’s against the law”. Interviewer: “I’d come, just me”. Clerk: “I’d have no problem with that”. These comments suggest that, even when expressly prohibited by law, dealer judgment enters into their sales.

Implications for prevention

Buying a handgun for another person may be a generous act intended to improve the recipient’s ability to protect him/herself. It also can provide legal cover for the recipient of the firearm because s/he is not subject to a background check or other regulatory mechanisms. Background checks are far from perfect,²³ but they are the primary means for screening those seeking to buy a firearm. It might be worth considering whether to prohibit purchasing or selling guns as gifts or, more generally, to prohibit purchasing or selling a firearm for someone else. Individuals are already prohibited from purchasing a firearm for a felon or other unauthorized purchaser; not allowing any firearm purchase on behalf of another would be useful when the purchaser or dealer is unaware of the recipient’s relevant background. Alternatively, it might be worth considering changing the law to require a background check of the recipient of the firearm as well as the purchaser.

Regulating gun transfers (for example, requiring transfers between private parties to go through an FFL), as some states already do, is another potentially promising method to decrease gun injuries and fatalities.²⁴ Such approaches necessitate consideration given that current law enables people to break, with relative ease, the link between a firearm and the individual who uses it. If firearms were registered, transfer of ownership could be tracked more easily, and options and considerations such as those listed in the previous paragraph would be less relevant.

Educating retailers and enforcing existing laws are two other ways to reduce illegal sales. Decoy operations in which law enforcement agents impersonate a customer could help identify errant dealers. Such approaches have been used in cigarette and alcohol sales to minors.

Recent data indicate that 57.4% of crime guns can be traced to 1.2% of FFLs,⁷ and sales volume, although an obvious covariate, appears not to be a determining factor.²⁵ Likewise, in the aforementioned General Accounting Office investigation, almost all FFLs adhered to federal and state firearm purchase laws.²³ Our findings differ. About half of the dealers were willing to sell a handgun when such a sale would be illegal. In addition, a few dealers seemed to be aware that such action would be illegal in that they indicated that they would be willing to skirt, if not openly violate, the law or they referred a potential customer to another retailer who may be so inclined.