Lankford's paper is found to be so flawed in such peculiar ways that the only reasonable conclusion is that Lankford either is incompetent or purposely intended to mislead.
Lankford's paper is a good example of the kind of flawed study used by anti-gunners to further their agenda without regard to the truth.
Here is the introduction to Lott's and Moody's paper.
--------------------------------------------
Adam Lankford (2016) asserted that the United States accounted for 31 percent of the world’s public mass shooters over the 47 years from 1966 to 2012. The news media around the globe widely publicized Lankford’s claim as soon as he started circulating his unpublished paper in 2015. Yet, despite numerous requests from researchers and the news media over four years, Lankford refused to provide a list of his cases or explain how he compiled them (see Lott 2018b). In responding to our research (Lott and Moody 2019), Lankford (2019) finally provided an appendix listing the 292 cases upon which he says he based his 2016 article.3 The extreme difference between his findings and ours, we now know, is driven by Lankford not following the definitions that he says that he was using. While we are still missing the data for the regressions that he ran for his 2016 paper, we at least now know what cases his sample included and excluded.
Lankford (2016, 190–191) claimed that he followed the FBI, Department of Homeland Security, and NYPD traditional definition of public mass shootings, but we discover otherwise. He included cases for the United States that do not fit those definitions, and he excluded hundreds of cases from around the world that do. Both errors greatly exaggerate the United States’ share of these attackers.
We also discovered that Lankford only included cases with just one shooter, except when he includes cases with two shooters. The only case involving two shooters in the United States that he counted was the 1999 Columbine attack, and the only such case outside the United States was from Russia.
Unlike Lankford, we immediately provided as part of Lott and Moody (2019) our entire list of public mass shooters as well as the news stories and sources that we relied on to put the list together. Even if Lankford thought there was a justification for studying only attacks with one or two shooters, it would have been easy to go through our list and explain why his list of such cases differed from ours. For example, Lankford excludes, without explanation, 37 foreign public mass shootings involving just one shooter and another 40 foreign cases involving two shooters. Furthermore, he does not justify the additional cases for the United States that he included that do not fit the FBI, Department of Homeland Security, and NYPD definition of public mass shootings. Both errors greatly exaggerate the United States’ share of these attackers.
We took care to exhibit, at length, official definitions (Lott and Moody 2019, 41–42). Nowhere do any of those sources confine the definition of public mass shootings under examination to cases with just one shooter. Indeed, the NYPD included cases involving up to ten shooters. Lankford’s response to our extensive demonstration of official definitions is to ignore that demonstration.
In his original paper, Lankford states, “For this study, attackers who struck outdoors were included; attackers who committed sponsored acts of genocide or terrorism were not. This is consistent with the criteria by the Federal Bureau of Investigation (FBI) in its 2014 active shooter report” (2016, 190–191). Nowhere in Lankford’s original paper does Lankford mention that he limited the attacks to one or sometimes two shooters. We invite readers to obtain a PDF of Lankford (2016) and search on ‘one,’ ‘alone,’ ‘lone,’ ‘solo,’ and ‘single’ to confirm that he nowhere reveals that he has confined his definition to cases with one shooter (except when he includes two). The first example that he provides of a public mass shooting on the first page of his paper is an exception—the Columbine attack, which had multiple shooters—so that example especially obscures that, aside from Columbine and a Russian case, his list is confined to cases of lone shooters. A more complete discussion of Lankford’s decision to include attacks with one or two shooters is provided below.
Finally, we discuss whether Lankford excluded terrorism cases as Glenn Kessler (2018) guessed he did, and we point out that even if all terrorism cases from outside the United States are excluded while those in the United States are included, the United States would account for less than 6 percent of the world’s public mass shooters—less than one-fifth of the rate that Lankford claims.
----------------------------------
Here is the conclusion.
--------------------------------------
Lankford’s study makes it extremely clear how important it is for researchers to provide their data to others or at least tell people their data sources and how their data were collected. For four years, Lankford refused to do either, and his misleading and error-filled research received much attention worldwide. It shows that the press ought to be very skeptical of studies from scholars who refuse to provide others with their data. Even a quick look at Lankford’s list of cases would have made it very clear that there were significant errors in both the list of United States and foreign cases.
As it is, despite repeated requests, we are still missing the rest of the data that Lankford used to run his regressions, and we have been unable to replicate anything close to his estimates even when using his flawed list of public mass shooters. Lankford has also declined to even answer any questions about how that other data set was put together.
Lankford uses neither the FBI nor NYPD definitions that he continually said that he used. We wonder whether the New York Times, Washington Post, and USA Today would have been somewhat reticent to use Lankford’s results if they had known that he had not in fact conformed to those definitions. Now we know that his definition excluded all but a few of the non-U.S. public mass shootings. Further, his dataset contains many errors and he doesn’t use any definition consistently.
We would have no quarrel with Lankford studying lone-wolf shooters, though he didn’t do that, but if he makes an international comparison concerning the number of ‘shooters’ without the qualification ‘who acted alone,’ then he must take them as they come, often in groups. The United States has less than three percent of the world’s public mass shootings or people killed in those incidents (Lott and Moody 2019, 53, Table 1). Using the NYPD definition, the U.S. has much less than its share of public mass shooters. Allowing other researchers to examine his data would help provide answers to the remaining questions concerning the mysterious Lankford datasets.
No comments:
Post a Comment