Gun Control: Does Biased Research Foster Workplace Danger?

[Drizzt]

Gun Control: Does Biased Research Foster Workplace Danger?
Posted by Howard Nemerov on March 27, 2007 - 12:31.

To bolster their hypothetical link between concealed carry and workplace violence, Brady Campaign references a paper published by researchers from the University of North Carolina:

As a result of the NRA’s shall-issue laws, companies that have not taken affirmative steps to keep guns out of the workplace and off company property have faced an increased risk of workplace violence. Indeed, a study published in May 2005 in the American Journal of Public Health concluded:

“[W]orkplaces where guns were specifically permitted were 5 to 7 times more likely to be the site of a worker homicide relative to those where all weapons were prohibited.”1

Science or Opinion?

Brady cited a paper entitled Employer Policies Toward Guns and the Risk of Homicide in the Workplace. The authors, Loomis and Marshall, surveyed companies in North Carolina about policies regarding firearms in the workplace and concluded:

Workplaces where guns were permitted were about 5 times as likely to experience a homicide as those where all weapons were prohibited…The findings suggest that policies allowing guns in the workplace might increase workers’ risk of homicide.2

The authors introduce their topic by stating their real agenda as an assumption: “we hypothesized that policies allowing guns in the workplace may increase the risk of homicide for workers.”3 A more appropriate hypothesis would have been: “Do policies allowing guns in the workplace have an impact on the risk of homicide?”


The authors admit:

This study was limited by the nature of the data available on worker’s exposure to guns. We generally did not know how often employees had guns at work, whether worker’s guns were used during the fatal events, and whether perpetrators came armed or used the victims’ own weapons. The inability to examine worker’s or perpetrators’ actions limited the ability of the current study to look beyond employers’ policies.4

Relevant data points are missing. The authors do not know if concealed carry licensees were involved in any of the shootings, or if these licensees had any moderating effect on the attacks. The authors do not know how often or how many employees were bringing firearms to work, so they do not know if greater presence of firearms at the time of incident had a positive or negative effect on the attack. All the authors can address is the relative restrictiveness of an employer’s policy regarding firearms possession on the premises.

Even though they conclude there is a correlation between an employer’s liberal firearms policy and shootings, they ignore the causal relationship: did the employer allow trusted employees to carry on the job because they recognized their business was more at risk for an attack? The authors note that workplaces with security “control measures” such as locked entrances, bright lighting and alarms had the greatest probability of homicide, and also note that high-risk workplaces have the second-highest probability.5 Curiously, the authors derogate their own findings:

Although we collected data on workplaces’ experience with robbery and violent crime, we did not control for it in the models presented here because adjustment for a determinant of exposure generally is not appropriate.6

Nor do they answer another relevant question: If firearms in the workplace lead to more shootings, why do we not read reports of incidents occurring in police stations, where virtually everybody is armed?

Poison Fruit
Anti-rights organizations often cite such research papers to promote their agenda. In 1986, Kellermann and Reay studied gunshot deaths in King County, Washington. They concluded: “We noted 43 suicides, criminal homicides, or accidental gunshot deaths involving a gun kept in the home for every case of homicide for self-protection.”7

This conclusion is often cited when gun control organizations promote their policies. For example, two representatives of Ban Handgun Violence, in their opinion piece supporting San Francisco’s proposed firearms ban, stated:

The New England Journal of Medicine found that a handgun in the home makes it 43 times more likely that a friend, family member, or acquaintance will be killed than an intruder.8

However, the National Research Council of the National Academy of Science (NAS) had this to say about this conclusion:

Kellermann and Reay find that there were nearly 5 times as many homicides and 37 times as many suicides as perpetrators killed in self-defense. They go on to conclude, “The advisability of keeping a firearm in the home for protection must be questioned.”

Although the facts are in no doubt, the conclusions do not seem to follow. Certainly, effective defensive gun use need not ever lead the perpetrator to be wounded or killed. Rather, to assess the benefits of self-defense, one needs to measure crime and injury averted. The particular outcome of an offender is of little relevance.9

The NAS panel actually challenged Kellermann’s and Reay’s conclusions, saying they were of “little relevance” because there is no established relationship between homicide/suicide and defense against criminal perpetrators.

This also highlights the problem of sampling error, where the small group studied does not reflect the actual experience of the entire population. Right-to-carry (RTC) states–where law-abiding gun owners may publicly carry concealed handguns–are considered by Brady to be more dangerous. Yet in 2005, they had a 43.8% lower murder rate and a 27.8% lower violent crime rate than non-RTC states, showing Kellermann’s hypothesis to be inaccurate when applied to the national population.10


Loomis and Marshall based their conclusions on 296 businesses,11 but the Small Business Administration notes that in 2003–the latest data available–there were 166,070 employer businesses in North Carolina alone.12 Loomis and Marshall based their conclusions on less than 0.2% of all North Carolina businesses, and highlights how fragile such conclusions are when compared to national data sets.

Loomis and Marshall cite other Kellermann research as the basis for their assumption that guns in the workplace increase homicide risk.13 The NAS panel found Kellermann’s conclusions in the Loomis and Marshall reference “are not tenable.”14 Furthermore, they quote Kellermann’s own disclaimer: “it is possible that reverse causation accounted for some of the association we observed between gun ownership and homicide.”15 He acknowledges uncertainty over his own conclusion that gun ownership in the home causes more homicide.

Conclusion

Loomis and Marshall rely upon discredited research to justify making a biased conclusion at the outset of their study. They then ignore their own findings as inappropriate when they conflict with their predetermined outcome. Basing policy upon such a questionable publication invites its own risk.




About the Author

Howard Nemerov is an investigative analyst for NRA News. He can be reached at HNemerov [at sign] Netvista.net.

Endnotes

[1] Legal Action Project, Forced Entry: The National Rifle Association’s Campaign To Force Business To Accept Guns At Work, Brady Campaign to Prevent Gun Violence, November 2005, page 1. http://www.bradycampaign.com/xshare/pdf/forced-entry-report.pdf

2 Dana Loomis, PhD, Stephen W. Marshall, PhD, and Myduc L. Ta, MPH, Employer Policies Toward Guns and the Risk of Homicide in the Workplace, American Journal of Public Health, Vol. 95, No. 5, May 2005, abstract, page 830.


3 Ibid.

4 Ibid, page 831.

5 Ibid.

6 Ibid.

7 Arthur L. Kellermann and Donald T. Reay, Protection or Peril?: An Analysis of Firearm-Related Deaths in the Home, 314 New Eng. J. Med. 1557-60 1986, page 1560.


8 Bill Barnes and Burke Strunsky, NRA out of S.F., San Francisco Bay Guardian.http://www.sfbg.com/39/15/x_oped.html


9 Charles F. Wellford, John V. Pepper, Carol V. Petrie, et al, Firearms and Violence: A Critical Review, National Academy of Science, copyright 2004, page 118.


10 Compiled from Table 4 – Crime in the United States by Region, Geographic Division, and State, 2004-2005. http://www.fbi.gov/ucr/05cius/data/documents/05tbl04.xls and Table 5 – Crime in the United States by State, 2005. http://www.fbi.gov/ucr/05cius/data/documents/05tbl05.xls Email request for spreadsheet.


11 Dana Loomis, PhD, Stephen W. Marshall, PhD, and Myduc L. Ta, MPH, Employer Policies Toward Guns and the Risk of Homicide in the Workplace, American Journal of Public Health, Vol. 95, No. 5, May 2005, page 830.

12 Office of Advocacy, Small Business Profile: North Carolina, Small Business Administration, 2006, page 2. http://www.sba.gov/advo/research/profiles/06nc.pdf


13 Dana Loomis, PhD, Stephen W. Marshall, PhD, and Myduc L. Ta, MPH, Employer Policies Toward Guns and the Risk of Homicide in the Workplace, American Journal of Public Health, Vol. 95, No. 5, May 2005, page 832. See their notes 3 and 4.

14 Charles F. Wellford, John V. Pepper, Carol V. Petrie, et al, Firearms and Violence: A Critical Review, National Academy of Science, copyright 2004, page 118.

15 Charles F. Wellford, John V. Pepper, Carol V. Petrie, et al, Firearms and Violence: A Critical Review, National Academy of Science, copyright 2004, page 119.

http://newsbusters.org/node/11663

 

[glummer]

“[W]orkplaces where guns were specifically permitted were 5 to 7 times more likely to be the site of a worker homicide relative to those where all weapons were prohibited.”

What sort of workplace has a policy of “specifically permitting” guns?? Every ordinary workplace I know of either prohibits weapons, or has no particular policy at all. Some places, like police stations, or security companies, might require weapons carry, but who would have a policy of “specific permission?”

We noted 43 suicides, criminal homicides, or accidental gunshot deaths involving a gun kept in the home for every case of homicide for self-protection.

The intellectual-dishonesty/stupidity level in that statement literally makes me want to vomit.
That is not a statement about the risks of firearms; it is a statement about the relative exposure of outsiders & residents, to things in the home.
Similar proportions would apply to the knives in your kitchen.
Or the bathtub in your bathroom, for that matter – when was the last time someone used a bathtub in lethal self-protection? But plenty of people die in them otherwise.

 

[PedalBiker]

Good people = safe work place.
Good people + a bad apple = one off crime - rare event
Bad people = unsafe workplace

Guns are irrelevant, as is most gun control. Well executed gun control (like Japan) will have no effect on crime. Poorly executed gun control (like Chicago) will increase crime. Just having guns in the general population doesn't help if they're not principaled and decent people with enough common culture to prevent civil splits.

It's a mistake to try to mix gun laws and safety. Social norms and morals are key to crime reduction. Gun laws aren't. It's dangerous to make promises we can't keep. In most cases liberal CCW rules and widespread gun ownership is a good thing, but it's due to the quality of the citizens of our country, not the guns. If we continue to allow personal responsibilty to fade and morals and mores to degenerate that may not always be the case.

That's why I teach my kids it's more important to be good than smart, and it's more important to be honest than rich. It's better to die for good than to live for evil. If these principals are forgotten guns won't matter.

That's why studies like the guns vs workforce violence are such a sham. The guns are largely irrelevant. If the wacko knows their fellow workers are armed he can easily change tactics to accomodate the fact.

Stealth and 20lbs of propane can make your CCW worthless.

 

[migoi]

The very first hour of ...

all the statistics classes I ever took (and pretty much daily thereafter) it was drilled into our heads... correlation does not equal causation. It seems our esteemed "researchers" forgot this axiom.

This does seem to be the least of the problems associated with this "study" though.

migoi

 

[Standing Wolf]

What sort of workplace has a policy of “specifically permitting” guns??

The gun shop where I work expects all employees to carry.

 

[obxned]

The only places where mass shootings occur are places where handguns are not allowed except for your garden variety homicidal maniacs and religious nut-cases.

 

[LiquidTension]

Workplaces where guns were permitted were about 5 times as likely to experience a homicide as those where all weapons were prohibited…The findings suggest that policies allowing guns in the workplace might increase workers’ risk of homicide.

Could it be that guns are allowed at these places because there is a higher rate of crime there, not the other way around? What? The article doesn't address a possibility that doesn't mesh with their predetermined outcome? You don't say

Funny, my office allows over 3/4 of the workers to carry guns openly and we don't have any problems with workplace violence. The author of this article clearly doesn't understand what a causal relationship is.

 

[PedalBiker]

Here's another clue. They said the places were at risk for the occurance of a homicide. That just means a dead body. They forgot to include if the homicide was justified or not. I'd much rather work at a place with a higher risk for homicide if that risk applies to criminals rather than the workers.

 

[MaverickNH]

Sorry for the long post, but I thought this might be of interest, as I have been addressing the same issue:

1. NYTimes Article

Workers’ Safety and the Gun Lobby

http://www.nytimes.com/2007/03/30/op...=1&oref=slogin

"The aura of invincibility that has legislatures bowing before the gun lobby is running into a commendable challenge from corporate America. Two conservative powerhouses — gun fanciers and business leaders — are facing off in statehouses over the gun lobby’s attempt to stop employers from exercising their property rights and barring workers from carrying firearms to work.

Bills to deny this common-sense right to workplace safety were initially approved in three states. But they failed last year in such gun-friendly states as Florida, Georgia, Indiana and Virginia after business interests rose up in active opposition. The National Rifle Association is back at work harder than ever in a dozen states. But so are Chambers of Commerce and corporate executives, warning of the danger — and business liability — of forcing companies to allow workers to carry guns.

There is no debate that doing so endangers workers. Workplaces that tolerate guns are five to seven times more likely to suffer homicides than job sites that ban firearms, according to a 2005 study in The American Journal of Public Health. The notion that self-defense mandates keeping guns in office drawers or out in parking-lot glove compartments is a dangerous fantasy.

The employers’ challenge was first thrown down in 2002 by an Oklahoma company that fired workers who refused to leave their weapons at home. That sparked the gun lobby’s mischievous campaign. With court fights under way, the employers’ cause was bolstered by the American Bar Association’s approval last month of a resolution defending the primacy of “traditional property rights” and federal laws mandating safe workplaces.

The private sector is showing good traction. The feisty Florida Chamber of Commerce doesn’t shy from mockery in warning lawmakers against a “Take Your Gun to Work Day” mentality. In Georgia, a conservative legislator with an A-plus N.R.A. rating, buttressed now by actively concerned corporations, dared to denounce the gun lobby’s “bullying” threats (“We will spare no effort to work against you,” vows the N.R.A).

The nation’s welfare needs more of this man-bites-John Wayne news."

2. My Letter to the Editor to NY Times

The editorial “Workplace Safety and the Gun Lobby” stated, “Workplaces that tolerate guns are five to seven times more likely to suffer homicides than job sites that ban firearms, according to a 2005 study in The American Journal of Public Health.” This editorial infers that studies prove more employees will be shot dead at work if they can carry guns during their workday. In reality, these researchers showed the risk of being killed at work is higher for workers of non-European ethnicity, employed in jobs like taxi and pizza delivery driver and bartender, who work alone at night, on weekends. This is not new news in the big city. This editorial calls to withhold from honest citizens, lawful means of self-defense during their workday, by negligently twisting the facts. The honest facts remain, as published by John Lott in “More Guns, Less Crime”, that legal carry of guns reduces crime.



3. My email to Loomis (the author) on 2005 Study

“The public is being presented with an interpretation of your work as proving that, in companies not prohibiting guns on the premises, gunfire breaks out amongst employees and employees are shot and killed 5-7 times more frequently than in companies banning guns on the premises.

I don’t believe this is what you had concluded. The publication referenced did indeed disclaim knowledge of whether 1) the victim was killed by another employee, 2) the victim was shot to death. Am I reading your published work correctly?

In your prior publications, it appears that your work indicates occupations such as taxi driver, gas station attendant and pizza delivery drivers were victimized at a rate making them significant contributors to study outcomes. As well, it seemed that you had indicated companies with employees of non-European descent composed a significant part of the at-risk population. One of your recommendations was that solo work be considered an at-risk practice as well. These conclusions drive in directions inconsistent with the motives of the popular press in attempts to influence public/private policy debates.

It is indeed important to gain recognition that immigrants, who often take solo jobs in the services industries, such as taxi, gas station and fast-food/delivery, are at greater risk of fatal victimization. But is appears to be a misinterpretation of your work by the popular press to suggest that in companies that do not prohibit guns in a standard office environment, gunfire will break out more frequently among employees. This emphasis leaves the true victims unrecognized, furthering the harms of socioeconomic inequity, while distracting focus and action on other workplace environments not indicated as significant by the outcomes of your data gathering and analysis. In effect, your work is being used by others to further an agenda not supported by your conclusions.

How can this be corrected?"


4. Loomis replies

"To my reading, the editorial in the Times does not misrepresent the findings of the study we published in American Journal of Public Health in 2005. It says simply that workplaces that permitted guns were 5-7 times more likely to have killings, which is a concise, accurate account of what the study found. The editorial writer’s assertion that there is “no debate” as a result of our study is arguable, but it’s clearly his/her opinion. As a scientist, I tend to think that one study is rarely definitive, and that the issue should be studied in other places using other methods.

Future studies could certainly improve on ours. You are correct that we were not able to conduct analyses to identify possible differences in the perpetrators or motives of homicides in workplaces that did and did not allow guns. We do have data on the means of death and we know that over 80% of all of the deaths were caused by firearms. However, we did not look for differences in the frequency of firearm deaths between workplaces that did and did not allow guns, and when guns were used, we do not know who brought them to the workplace.

Unfortunately, authors have no control over the interpretation of their work once it is in the public domain. This particular study has been misrepresented by the NRA and perhaps by anti-gun interest groups, too, although I have not seen it myself. Such misrepresentation is regrettable, but it usually gets corrected as part of the public decision-making process.

Dana Loomis, PhD
Professor and Chair
Environmental and Occupational Health
School of Public Health / 274
University of Nevada, Reno
Reno NV 89557-0036, USA
Phone: +775-682-7103
Fax: +775-784-1340"

5. My 2nd email to Loomis

“Thank you for taking the time to reply to my message. A Google of the internet does indeed show that your 2005 publication has been taken up by parties on both sides of the gun debate, with varying degrees of misrepresentation. On the anti-gun side, the paper is most often cited accurately, but mingled in amongst other text that leads one to picture employees gunning down the boss and other employees a the brick-and-mortar workplace (office, workshop, warehouse, etc.) rather than the all-to-common news report of a cabbie, pizza driver or “Stop & Rob” convenience store robbery turned deadly. This is the misrepresentation that I find most disquieting.

I am not surprised that > 80% of the deaths reported in your study were perpetrated using guns – a much lesser figure would have cast doubt on the validity of the data, as guns are the primary mode of homicide in the US. You seem to have taken pains to gather the most complete, accurate and balanced data set possible - much better than most, by a long shot. In the end, the data are what they are, as they say.

I think you make a useful point in the concluding lines of your 2005 paper: “These findings bear directly on policy for workplace safety. In light of the evidence, it is reasonable to question the costs and benefits of policies permitting firearms in the workplace.”

Your present work focuses on the potential costs of permitting firearms in the workplace – do any of your data reflect on the potential benefits? In assessing a rcost/benefit ratio, both sides of the equation are necessary.

While it appears the initial motive for gathering the dataset was driven by assessing homicide determinants and the efficacy of safety measures, perhaps there are some questions asked that would reveal benefits. Were any questions of the controls directed toward employee use of a weapon in an act of self-defense, either in transit to, from, or at the workplace? I ask about transit, as a prohibition of guns on workplace premises would entail a prohibition of carry to and from work, including in the vehicle, which is the workplace for cabbies and food delivery drivers. With many of the test cases detailed by proxies, this question may be of lesser utility (especially where the sole employee is dead).

As the 51mo study period (1/1/94 – 3/31/98) included 23mo before and 28mo after NC enacted Right to Carry legislation 12/1/95, you might find a near 50/50 mix of before/after data (based on 28 cases in each of 1994 and 1995, as reported in your 2001 AJE paper). NC state data http://www.ncgccd.org/pubs/systats/spring97.pdf
indicates that over 20,000 CCW permits were granted in less than a year following enactment, so the ramp-up seems to have been relatively sharp. More recent data shows the trend has continued, with further increases in CCW permits issued by NC http://sbi2.jus.state.nc.us/crp/public/other/conceal/Sept302004stats.pdf. Analysis using such figures might be another way to understand how guns in the workplace modulate worker safety. As the number of Shall-Issue/Unrestricted CCW states has increased from 30 in 1997 to 38 in 2006, this would be a very relevant comparison.

Further, as NC law allows for signage prohibiting guns on the premises “…where notice that carrying a concealed handgun is prohibited by the posting of a conspicuous notice or statement by the person in legal possession or control of the premises.”, should your dataset include data on such signage before and after enactment of NC’s CCW laws, that would an interesting analysis as well. Elsewhere, I have seen taxis with the “No Guns” stickers, so presumably even some cases with no brick-and-mortar worksite scene of shooting could be analyzed. While I did not find an assessment of such signage in your 2002 JAMA paper, do you have such data available? To that question, is your dataset publicly available or available on request?

Curiously, the number of fatal work injuries has drop 12.3% in the most recent 4 years of BLS data in comparison to the 1994-1997 BLS data – your study period. The rate per 100,000 has dropped from ~5 to ~4. Homicides have dropped from 2nd place to 4th place. BLS data http://www.bls.gov/iif/oshwc/cfoi/cfch0004.pdf
show the decline in homicides accounts for this drop in fatality rate, even offsetting rises in other categories. In NC, 2005 BLS data indicate worker shooting homicides rank 5th for cause of death.

One just has to wonder what significant changes have lead to a reduction in the ranking of firearms-related worker homicide in the last ten years, especially given the rise in CCW rates over that period. Indeed, the murder rate in NC has dropped from a range of 8-11 in the 1994-1997 study years to 6-7 in the most recent years reported (2002-2005).

Certainly the hypothesis proposed by John Lott, that of “More Guns, Less Crime”, is a subject of debate almost 10 years since his original proposals and 1998 book, even if the editor of the NY Times article considers guns-in-the-workplace a settled matter. If and how such an effect, as Lott proposes, might help explain the apparent data shift in NC would be a telling analysis.

Based on a review of your broad publication record, it’s clear that you make efforts to let data tell the story, even if you might have personal opinions (as we all do) on the wide range of matters your work addresses. As a scientist myself, I try to let data guide my decisions, even if I am troubled by data that challenges my expectations and current views of the world. I’ll look forward to following your work on this topic.

I note that Nevada flipped to Shall-Issue at about the same time as NC, although the reported CCW rate is almost half of NC (which may not account for the more frequent open-carry I have observed in the outback). I hope you are enjoying your new surroundings.”

 

[MaverickNH]

... and look at how the Brits wrote up the article in the Britsih Medical Journal http://www.bmj.com/cgi/reprint/330/7499/1043

"Workers in petrol stations, grocery stores, and other locations who take weapons with them to work are three to seven times as likely to be murdered at work than workers in equivalent jobs in which weapons are prohibited, a study has found (American Journal of Public Health 2005;95:6-8)"

Nowhere did the AJPH paper say the victims actually had guns when the were killed - in fact, most were solitary workers who were robbed at gunpoint and shot dead in companies that did not ban CCW of firearms. Are we to suppose they all were robbed by unarmed thugs who took their guns and killed them with their own guns?

The medical establishment is very willing to negligently fabricate misinterpretations of very limited data to "prove" guns in the workplace result in employees shot dead. To see this in the most respected British medical journal is disgusting.

The author of the BMJ article, Janice Hopkins Tanne, is supposedly an award-winning science writer. I can't find her contact information to send a scathing letter.