Accusing someone of using pseudoscience will seem to some a harsh accusation, putting the accusee in the same camp as Biblical creationists, astrologers, and followers of the late L. Ron Hubbard. So I'll start with a presumption of innocence, that the paper recently published by the Brady Center is what it appears to be at first glance: A scientifically sound analysis of real data regarding use of "assault weapons" in crime.
The new article,
On Target: The Impact of the 1994 Federal Assault Weapons Act - gets the following introduction on
http://www.bradycampaign.org"As the Senate prepares to vote on an amendment that would renew the Federal Assault Weapons Act for another ten years, a new study shows a substantial drop in the use of high-firepower assault weapons like UZIs and AK-47s in crime, despite industry efforts to evade the statute."That sets the stage. Please take a look:
http://www.bradycampaign.org/xshare/200403/on_target.pdfForst off, who is Gun Crime Solutions LLC? There is no Web site, and the work is credited to one person, Mr. Gerald A. Nunziato, who in the past served as "Special Agent in Charge of the Bureau of Alcohol, Tobacco, Firearm’s National Tracing Center". OK, we can assume Mr. Nunziato is a top expert in tracing guns but does he have the academic qualifications to do a statistical analysis on use of weapons in crime?
Mr. Nunziato is a retired ATF agent who now makes a living doing paid appearances on TV and radio shows about gun control, and also gets paid as an expert witness. He works for a company called the Law Bulletin Publishing Company. See
http://www.juryverdicts.com/experts/nu1.html , look for his name near the top, and follow the links.
Mr. Nunziato is a criminologist, not a statistician. There is no evidence presented that he is qualified to do the analysis in the Brady Campaign's report.
To be taken seriously by scientists, a scientific paper has to be published in a recognized journal. Before publication any paper is subjected to deep scrutiny by the journal's editors and to a rigorous process of
peer review by a committee of qualified individuals, people who are knowledgable in the disciplines from which the article derives.
The Brady Center's article is published by the Brady Center. That would be seen as a conflict of interest in the academic world. No peer review was done. That doesn't mean it isn't science, but it does mean it cannot be assumed to be good science.
Moving on to the article itself,
To evaluate the questions below, the Brady Center to Prevent Gun Violence asked
Crime Gun Solutions LLC to review and analyze national crime gun trace data
maintained by the Bureau of Alcohol, Tobacco, Firearms and Explosives (ATF). This
data represents guns nationwide that have been illegally possessed, used in a crime, or
suspected of being used in a crime, thereafter recovered by law enforcement, and then
traced to learn about the sales history of the gun.The analysis was done not on actual crime data but on gun trace data, which includes guns known to have been used in crimes but also guns only suspected of being used in crimes. We aren't given any information about the breakdown of real crime data vs. assumed crimes. In science writing this is called a "proxy measure". In honest work, proxy measures are used only when direct measures are unavailable. In less than honest work a proxy measure may be substituted for real data that does not support some conclusion (foregone) that the author wishes to present.
Whenever a proxy measure is used in science writing, the authors are expected to disclose that the measure is a proxy, explain why the proxy was used, and provide some evidence that the proxy has some connection to the missing direct measures. Mr. Nunziato didn't do any of those, in fact the word "proxy" doesn't appear anywhere in the article.
If I was a member of a peer-review committee for a respectable journal I'd stop right here and not waste any more time on this article. It does not meet academic standards for publication.At this juncture I can say the article no longer qualifies as good science. It may be bad science, or it may be something less than science, but if you care about real knowledge it's probably not worth reading. Why did Mr. Nunziato not use actual crime data? Why did he choose this particular proxy? What real data was available? We don't know, and we as readers may form different and contraductory assumptions about the use and validity of this proxy measure. If I saw this in a scientific journal I'd fire off a scathing letter to the editor.
For brevity I'll restrict the rest of this post to pointing out a couple of glaring logical fallacies that will provide a chuckle but no enlightenment. We already know that this piece was written by a paid hack and has not been subjected to any kind of academic scrutiny. It may be a "report" but it is not a "study" in the generally accepted meaning of that word.
Here is the first of Mr. Nunziato's two core findings:
In the five year period (1990-1994) before enactment of the Federal Assault
Weapons Act, assault weapons named in the Act constituted 4.82% of the crime gun
traces ATF conducted nationwide. Since the law’s enactment, however, these assault
weapons have made up only 1.61% of the guns ATF has traced to crime – a drop of
66% from the pre-ban rate. Moreover, ATF trace data shows a steady year-by-year
decline in the percentage of assault weapons traced, suggesting that the longer the
statute has been in effect, the less available these guns have become for criminal
misuse. Indeed, the absolute number of assault weapons traced has also declined.Indeed, indeed! Here he's talking about "assault weapons" in an absolute sense here - Weapons defined under the 1994 federal AWB as such and grandfathered. In other words, what the gun community refers to as "pre-ban" weapons like the original Colt AR-15 and not post-bans like the Colt Sporter, Olympic Arms PCR, Bushmaster XM-15, etc.
What has happened to those genuine, pre-ban "assault weapons" in the last 10 years? Not a whole lot, except:
- Their value on the secondary market has double or tripled. Compare the price of a pre-ban Colt AR-15 in as-new condition to a brand-new functionally identical just-as-good Bushmaster. Real "assault weapons" are collector's items now because their number cannot be increased under the present ban. Real "assault weapons" are either squirrelled away in peoples' closets or have been sold for profit to affluent individuals. Either way they're unlikely to be used in crimes BECAUSE OF THEIR MONETARY VALUE.
- People who owned them in 1994 are now 10 years older and therefore statistically less likely to be involved in crime.
- A few have been lost by attrition - Burned up in fires, run over, lost in the woods, fallen out of canoes, seized by law enforcement, etc. But most of them are probably still out there.
These three factors would seem to all suggest less likelihood of a pre-ban "AW" being used in crime. But remember, the Brady Center report concerns
gun trace requests, not crimes. Mr. Nunziato makes no effort to control for these factors. It would appear that crime may be more a function of people than of weapons.
How many Duesenberg automobiles have been involved in car crashes recently? I haven't heard of a single incident in my lifetime. None have been manufactured since the 1930s, and the ones that exist today are in the hands of wealthy collectors. The situation is quite analogous to that for pre-ban "assault weapons". But there are plenty of post-Duesenberg automobiles available today. They along with Cords and Auburns were the Infinitis and Lexuses of their day, the "luxury/performance" cars. Today's fancy cars do get involved in crashes, but like the post-ban "assault weapons" they tend to be owned by people who don't get into much trouble.
And here is Mr. Nunziato's second major howler, er, finding:
After the Assault Weapons Act was passed, gun manufacturers sought to evade
the ban by producing weapons with minor changes or new model names. The Act was
designed to prevent this occurrence by defining assault weapons to include “copies or
duplicates” of the firearms listed in the ban in any caliber, though this provision has
never been enforced. Yet, even if copycats of the federally banned guns are
considered, there has still been a 45% decline between the pre-ban period (1990-1994)
and the post-ban period (1995 and after) in the percentage of ATF crime gun traces
involving assault weapons and copycat models.What's this?
Mr. Nunziato is no longer talking about "assault weapons" in the legally proper sense. He's talking about a mix of pre-ban and post-ban weapons
not covered by the assault weapons ban and claiming that even if you decide to consider them to be "AWs" the ban reduced THEIR use in crime as well.
Now we have entered the realm of pseudoscience, mixing apples and oranges and claiming to have created something new. Any resulting numbers are too fuzzy to have any discernable, verifiable meaning. There's no attempt to explain why the existence of functionally identical post-bans in numbers far higher than pre-bans has not completely negated the effects of the AWB.
Mr. Nunziato has not considered the obvious question of HOW MANY post-bans have been manufactured since 1994. Some data about the numbers of firearms made by manufacturers exist, but you have to look around for it. I find it interesting that our friends at the Violence Policy Center stopped collecting statistics after 1995. See
http://www.vpc.org/resource/index.htm - It seems that if they really wanted to collect numbers to support gun control they would have kept up this effort. I wonder why they dropped it.
Mr. Nunziato has also not considered the effects of changes in overall crime rates and shifts in crime patterns in his (let's remember this one more time) gun TRACE REQUEST figures.
In short, this piece is pseudoscientific propaganda written by a hired gun who lacks academic qualifications to write a real research paper. If I had turned this in as a term paper for an experimental design and analysis course as an undergraduate it would have gotten a D at best.