That’s just about all of us if you believe the press, except that according to Politicus USA, that’s only if you have a lot of guns.

This week an article was circulating around social media with a headline that would draw cheers from many, though by no means all, progressives: “Bombshell: Gun Owners Tend to Be Angry, Unstable, Impulsive.” If one clicks on the article, there is indeed some bombshell research contained within it. However, the headline does not accurately reflect the research study’s findings and misses the chance to inform, as well as advance, a policy argument. It is unfortunate because it doesn’t do progressives any favors to circulate headlines that are erroneous. It ends up putting us on the defensive as we have to defend why statements are being made that don’t reflect reality.

Researchers at Duke and Harvard conducted research with results that can help form social policy, indicate to family and friends when a loved one may be dangerous, and destigmatize a group that has long been maligned as unsafe. The social scientists who completed this research used the National Comorbidity Study to get their findings. The National Comorbidity Study has been conducted since the 1990s as a study of mental health and substance abuse in the community, most recently using a nationally representative sample of 5,563 people with the purpose of determining the percentages of mental illness and substance abuse in the population. The calculation of whether randomly selected research participants have either problem was done through an extensive, standardized, face-to-face interview. There are also questions about the use of professional services, etc. The most recent iterations have included questions related to gun ownership. This is where the “bombshell” research results come from.

The researchers found that 8.9% of those who reported being impulsively angry also reported gun ownership. This is not a small percentage if we, as Americans, are trying to keep guns out of the hands of people who could easily be at sudden risk of shooting any number of people. On the other hand, it also means that over 90% of the people who get impulsively anger did not have a gun. Because of the way the study is structured, the converse—over 90% of gun owners were not impulsively angry—is also true. Oops, there goes the bombshell idea that “gun owners tend to be angry, unstable, impulsive.”

What made this study so useful is that they were able to pinpoint who is at risk for violent behavior. It comes as absolutely no surprise that people who owned six or more guns or who had a history of violent behavior were those who reported the riskiest behaviors of “outbursts of anger,” “getting into fights,” or “smashing or breaking things.”  These are the people who are at risk for killing someone. The vast majority of these people don’t have a diagnosable mental illness, and reported never having been committed, so the study’s authors point out that the focus on keeping guns away from people with serious mental illness who have been committed is misplaced. It ends up stigmatizing a group further that already struggles heavily with stigma.

So if you own six or more guns, you are at risk for killing someone.  Six is the magic cutoff – not five, and not seven.  It’s six.  So the “researchers” and “social scientists” say.

Actually, her data proves nothing of the sort.  There isn’t any cutoff, there isn’t any proof.  This is nothing.  In order to have a meaningful analysis, an individual would have to have (a) a number of very large sample populations, (b) assurance that the sample population wasn’t biased, (c) a reliable means of extracting true information (torture is illegal and unreliable and gun owners are known to give false or incomplete information concerning their ownership of guns due to [justifiable] mistrust of the government), (d) repeatability and verifiability (in other words, is there fully independent means of verifying that when they said they got violent they actually did so, were charged with it, and found guilty of a crime), and finally (and this may be the most important), (e) enough studies over a large enough set of sample populations that they could meet the central limit theorem and create a standard distribution in results, with a fractional standard deviation of, say, around 0.05 or less.

None of these things obtain, and they will never do so.  And if you believe that there is such a thing as social “science,” then you may as well believe in witch doctors and voodoo.  If you want to do science, then study the Navier-Stokes equations or the Boltzmann transport equation.  Then maybe I’ll have some respect for you.  Until then, you’re all just witch doctors to me.