Friday, April 26, 2013

Lies, Damned Lies and (Gun Control) Statistics


Anyone who has been a sports fan for any length of time knows how statistics can be used to make a false argument about a player or to find “hidden value” in a player that by appearances might seem less attractive to a team. A great example of the former would be a Pitchers Win/Loss totals or in football, the Quarterback Rating. Neither of which by itself is a reliable metric of a player’s value. There are some examples of the latter found in modern “Advanced” statistics such as WAR, BABip etc that can show a player to be of greater or lesser value than he seems to the naked eye.

In sports as in many other areas of “debate” statistics can be very useful. Provided the data set is of sufficient size to provide a representative sample, provided the statistic actually measures what is being evaluated and provided the statistic is being accurately represented. Equally, statistics can be used to distort the truth - particularly if the above provisions are ignored or not met and that is the topic of this post.

The post Sandy Hook Gun Control debate rages on in many states but has thankfully due to cooler heads prevailing, lost steam at the National level. There were two widely repeated statistics used by the Anti-Second Amendment crowd that I’d like to discuss.

40% of all gun sales are completed without a background check.

90% of Americans (including gun owners) support background checks.


It’s important I think to first point out the origin of these statistics before discussing how they’ve been used to distort the truth. The 40% stat comes from a 1997 report by the Instituteof Justice written by Phillip Cook (Duke University) and Jens Ludwig (University of Chicago). The report is based on a 1994 survey that contained among others a question regarding firearm transfers (NOT sales – that’s important) in 1993 and 1994.

The question asked in the survey was if the firearm was obtained from a licensed dealer. The affirmative responses totaled ~64% meaning that ~35% did not obtain their firearms from a licensed dealer. They then rounded the figure up and made the assertion that “40% of firearms are not purchased from licensed dealers.”

The rounding is not the problem here. The question (validity of the metric) the data set (adequate sample size) and representation, however are.

The question is a problem in that it asks if the firearm was obtained from a dealer which sounds benign enough until you delve deeper into the study. In looking at the data of the survey it shows that a significant number of the acquisitions were from gifts, inheritances and prizes and that removing those from the totals brings the number from “40%” to 26.4% that were actually obtained by purchase from a non licensed dealer.

The data set too causes problems with the validity of the statistic. This survey was a random-digit-dial survey of 251 people with a response rate of ~50%. The sample size is 251. No, seriously. 251

The last problem with this is in how it’s presented and what its supposed to be representative of. The salient point in that is the timing of the survey and subsequent report. Follow along with the timeline here……Again, the report was published in 1997. The study took place in 1994. Why are those dates relevant? Quite simply, because the “Brady Bill” was enacted in 1994. One of its components, the creation of the National Instant Criminal Background Check System (NICS) wasn’t launched until November of 1998.

The survey covers a statistically irrelevant sample size of gun transfers (not sales) that occurred BEFORE there was a legal mandate for background checks on firearm sales. The survey asks too few people the wrong question, at the wrong time.

The 40% number is often used to support the assertion of the “Gun Show Loophole”. To assume that it’s true (which it clearly isn’t) requires a pretty massive suspension of disbelief. In 2012 alone there were 19.6 Million firearms purchased via licensed dealers using the NICS system. So what the anti-gunners are trying to tell you is that an additional 12 Million firearms were sold by private face to face sales. Why on earth would vendors pay for a booth at a gun show, surrounded by competitive dealers if 40% of the people in attendance were there to buy guns privately? Because it isn’t true. Which begs the question of why would anti-Second Amendment advocates misrepresent irrelevant and outdated data to support their position? Because it isn’t a position, it’s an agenda.

The second statistic bears equal scrutiny – “90% of Americans (including gun owners) support background checks.”

The source of this statistic is a CBS/New York Times poll of 1110 people conducted in early January of 2013. Though an awfully small sample size it does appear that the poll itself was conducted properly. However there remains a problem. The most obvious problem here is with the question and how the results are misrepresented. In this case it is a question that assumes the answer. The question asked in the poll was:
“Do you think anyone buying a firearm at a gun show should undergo a background check?”
Setting aside the fact that the overwhelming majority of firearm purchases made at gun shows are transacted through FFL (Federal Firearms License) Dealers and thus ARE accompanied by NICS background checks, the question omits the actual topic under discussion – Universal Background Checks. The question is intentionally worded in such a way as to provide a desired answer. The poll didn’t ask;
 “Do you support Universal Background Checks on ALL firearms transfers AND the full National firearms registration that the DOJ states would be required in order to enforce the program?” 
This is a question that undoubtedly would provide a much different answer. Thus, the poll provides an intentionally misleading result and should not be considered relevant.

With regard to how the result of the poll, however attained, is represented we’ve only to look at some of the sound bytes that have come out of Washington over the last few months.

“90% of NRA members support our proposal” ~Joe Biden
    • Actually, the poll says 85% of NRA members support background checks. The proposal Biden refers to is for Universal Background checks which the poll did not specifically identify.
“Right now, 90 percent of Americans — 90 percent — support background checks that will keep criminals and people who have been found to be a danger to themselves or others from buying a gun. Think about that. How often do 90 percent of Americans agree on anything?” ~President Obama
    • Again, the politician refers to a poll result that asks a question that is not specific to the assertion being made. Disingenuous and highly misleading.
“90% of Americans support this bill” ~ Sen Joe Manchin
    • The Senator refers to the “Toomey Manchin” background check expansion bill which would broaden NICS to all firearm transfers. The Bill was not mentioned in the CBS/NYT poll as it had not yet been written.

As alluded to above, the greater concern here is how this statistic is used. When a politician stands before us and makes a claim such as this (regardless of the validity of the statistic) they are using argumentum ad populum. This is a logical fallacy that concludes that the proposition is true or just because many or most people agree with it. It is used in much the same way as geocentrics used consensus to suppress heliocentrics such as Galileo, the way the early church used consensus to brand Da Vinci a heretic and the way those opposed to emancipation used consensus to argue against the 13th Amendment to the Constitution.

Using consensus as a means of forming public policy or codification of law is an example of “pure democracy” a model that is not allowed for by virtue of how the Constitution is constructed. While there are examples of it’s use at the State and Local levels, (California’s public initiative process as an example) it is not and should not be a matter of course at the National level. To do so sets us on a very slippery slope that I rather think we shouldn’t tread for were we to depend entirely upon consensus, or “mob rule” the 13th Amendment would not have passed, nor would have suffrage for blacks and women.

If you tell me you are opposed to the Second Amendment because you just don’t like guns or you are afraid of them or you just don’t think people have a need for them that’s fine. I can respect your statement of opinion and values even if I wholeheartedly disagree with them. But when you, or a politician begins using logical fallacy, intentionally misleading statistics or outright falsehoods to support the anti-Second Amendment agenda the respect ends.

3 comments:

  1. Some actual research and fact-checking during these times of mass hysteria?

    As a fellow Calgunner, I must commend you. :-)

    ReplyDelete
  2. Love your blog Damn True. Keep it up!

    ReplyDelete
  3. Here's a different source for the "90%" number, FYI: http://www.quinnipiac.edu/institutes-and-centers/polling-institute/national/release-detail?ReleaseID=1877

    They show their data and polling methods, and if further food for thought.

    ReplyDelete