This is from a list for historians that may be of interest.

------------------------
From: Beth Salerno <BSalerno@Anselm.Edu>

From: James Lindgren [mailto:jlindgren@worldnet.att.net]
Sent: Sunday, October 07, 2001 1:07 PM

This post by Robert Churchill (Princeton) to H-OIEAHC on September 19 contains much new substantive information about his research into problems with ARMING AMERICA that has never been discussed on any other list. For context, it was prompted in part by Michael Bellesiles' claim that the false reports that had been on his website since February were actually written by a hacker. --Jim Lindgren, Northwestern U., jlindgren@att.net

[Editor's Note (Salerno): I have edited the following post by removing three paragraphs which related to materials discussed on H-OIEAHC but not H-SHEAR.

I have put ellipses to indicate the missing material, which did not change the argument as laid out below.]

-------------------------------------
[9/19/2001]
Dear Colleagues:

I have followed the controversy over Michael Bellesiles' work on probate records in some detail over the past few months. My review of the ARMING AMERICA will be appearing in REVIEWS IN AMERICAN HISTORY sometime this month [for those with MUSE access, the review can be found at http://muse.jhu.edu/journals/reviews_in_american_history/toc/rah29.3.html]

Given Michael's latest claim that his report of his research on Vermont probate inventories is in fact the product of some unidentified third party, I thought it was time to review the controversy.
-
ARMING AMERICA'S claims regarding very low rates of gun ownership rely on three types of data: probate inventories, militia returns, and anecdotal evidence. Of these, the anecdotal evidence, often in the form of requests that colonial and state governments provide guns to beleaguered populations, is the least reliable, statistically speaking, because such evidence is generated only by a state of scarcity. This leaves the militia returns and the probate analysis as the empirical foundation on which Bellesiles' chronological argument rests.
-
Jim Lindgren's analysis of the book last January made its most important contribution in providing a detailed statistical analysis of probate inventories in Providence, Rhode Island, published in the Providence Town Records, and of Alice Hanson Jones' published compilation of inventories. At the time, all evidence indicated that Bellesiles had relied upon both of these samples. Bellesiles has since claimed that he never used Jones' sample, despite stating in his 1996 JAH article that he had. Six years after Bellesiles published his findings, those of us engaged in the professional responsibility of evaluating his work are still guessing at the composition of his sample. I will return to this point shortly.
-
Bellesiles could not claim that he had never used the Providence inventories, as he discussed them in detail in ARMING AMERICA. What Lindgren's analysis demonstrated was that Bellesiles had misrepresented the original records, mischaracterizing the sample and altering the language of the original documents. Specifically, Bellesiles had claimed that half of the guns were listed as old, and "a great many" were listed as "Queen's arms," or state property. Lindgren found slightly more guns, fewer inventories, that only 9% of the guns were listed as old, and only one was a Queen's arm. He also found that Bellesiles had in the book attributed this Queen's arm to the wrong inventory. In May Bellesiles posted a revised analysis of the Providence records on his website that conceded on all of these points.
-
Prior to this, in February, Bellesiles had posted an analysis of guns in Vermont probate inventories from 1770-1790. This analysis had a specific purpose. When Lindgren circulated his essay at the turn of the year that in part analyzed Alice Hanson Jones' probate sample, he found guns in over 40% of northern inventories and over 60% of southern inventories. Before I was aware of Lindgren's work, I made a similar count that confirmed his findings. As Bellesiles had reported finding guns in 15% of northern inventories and 18% in the South, his work was called into question. Bellesiles' purpose in posting the essay on Vermont was to shore up his claim of lower rates of gun ownership than Jones had found. He accordingly found guns in 44 of 342 "probate records." Once a large number of state owned guns were removed from the sample, guns were present in only 11.4% of the sample.
-
This essay was crucial because it offered a second specific set of records with which to evaluate Bellesiles' work. And it is here that Bellesiles' failure to reveal the precise details of his sample is critical. Lindgren and Justin Heather went to Vermont, and found that once again Bellesiles had misrepresented the original records. Furthermore, he had done it in precisely the same manner that he has now been proven to have utilized with the Providence records. Without revealing the details of Lindgren's newest work, which he has shared with me but has not yet published, the pattern is clear: more guns, fewer inventories, and language altered to make guns appear old or state owned, and thus advance the overall thesis.
-
In my own research into late eighteenth and early nineteenth-century militia returns, I have also found that Bellesiles has mischaracterized, misinterpreted, and sometimes grossly misrepresented the original records on which he relied. I will discuss three examples here, not because they are particularly egregious, but because they fit the pattern that Lindgren has discovered. Bellesiles cites a 1744 militia return from Worcester County, Massachusetts. He claims that 8 of 21 companies that "filed a report on their firearms" reported that they were "entirely deficient." (ARMING AMERICA, 150) In the original document the colonel of the regiment reported the state of the arms and ammunition of each company. He noted that four of the companies were "entirely deficient as to arms." He reported the other four as "entirely deficient as to ammunition." Bellesiles has thus altered the language in the original to advance his thesis of gun scarcity. A photocopy of this document is in my possession, and I will be happy to provide it to anyone interested.
-
Another example of this tendency is Bellesiles' use of an anecdote from Harold Selesky's WAR AND SOCIETY IN COLONIAL CONNECTICUT, pp. 228-229. In Bellesiles telling, when Benedict Arnold heard the news of Lexington and Concord, he "inspected his troops and found them largely unarmed. He threatened to break into the town arsenal in order to arm his men, but the town's selectmen relented and opened the doors to his militia, with Arnold supervising the distribution of Brown Besses." (ARMING AMERICA, 181) But in Selesky's account, and in the source upon which Selesky relies, Arnold demanded the keys to the powderhouse, so as to secure ammunition for his men. The compelling details of gunlessness and the distribution of public arms appear to be an invention designed to advance the thesis.
-
Finally, in the Table 3 of the appendix of ARMING AMERICA, Bellesiles purports to show the levels of "Private gun ownership in Massachusetts." For a series of sample years from 1789 to 1839, he reports the "number of privately owned muskets or rifles," the state's population, and the "percent of population with guns" for both total population and white males 16 and older. The problem with this analysis is that Bellesiles has taken the number of guns from George D. Moller's approximations of the number of guns returned in the Annual Return of the Massachusetts Militia for the given year, published in Moller's MASSACHUSETTS MILITARY SHOULDER ARMS. These are thus the approximate number of guns brought by militiamen to the annual inspection musters held around the state. They are not the total number of privately owned guns in Massachusetts, a number that no state government sought to keep track of in this period. In comparing this number to total population and to all white men over 16, and not to the number of militiamen appearing at muster, Bellesiles artificially inflates his sample in order to report low rates of gun ownership.
-
Which brings us to Bellesiles' claim that the inaccuracies in his essay were inserted by a hacker.[...] If Bellesiles' website has been hacked, it has been hacked twice. The hacker would have to have changed the language Bellesiles reported for individual gun estates, while leaving intact Bellesiles' claims that guns were present in only 11.4 % of estates. That overall finding was on the website when I first viewed it a week after it was created, on February 16. The hacker would have to have inserted alterations entirely consistent with the alterations Bellesiles made in the Providence records. These alterations must have been made within weeks of the original posting. The hacker must then have struck again in the last week or two to add the references to pornography that Bellesiles mentions but that did not exist on the site on September 3.
-
[...]

Sincerely,
Robert Churchill
Rutgers University, Ph.D., 2001
Lecturer, Princeton University
rchurchi@princeton.edu

---------------End of Original Message-----------------

===================================================================
Constitution Society, 7301 RR 620 N #155,276, Austin, TX 78726
512/531-0767 Date: 10/09/01 Time: 10:46:37
http://www.constitution.org/ mailto:jon.roland@constitution.org
===================================================================