I recently received a response from Professor Walter Mebane to my questions about his group’s portions of the DNC 2004 Ohio Election Report. Professor Mebane’s group is responsible for Sections VI and VII of the DNC report. I consider Section VI to be the most critical section of the report, because that is the section that concludes that the analysis strongly suggests that there was no widespread fraud in the Ohio 2004 Presidential election.
My intention is to go over the response to my questions with some statisticians prior to posting a summary of the response on the DU. That is because I am unable to fully understand some key parts of the response. However, in the meantime, I think that it is worth while to post some key points of the response because I know that a lot of you are writing letters to our representatives or to the DNC, and I think that some parts of Professor Mebane’s response may be helpful to that effort.
Before listing those points I just want to say a general word about Professor Mebane’s response. He responded to my e-mail to him within four hours, and the response is extremely detailed and thorough – though quite difficult to read for non-statisticians. In that regard, and because some of his points I believe could be useful our effort, I think that many of you would feel some gratitude towards him for his prompt and thorough response. On the other hand, he still sticks by his conclusion that “The precinct data
provide strong evidence against the claim that there was widespread misallocation of votes from Kerry to Bush.” In that respect I’m sure that many of you would be profoundly disappointed in him, to say the least.
I myself am still unable to understand the rationale for coming to that conclusion, and that is one reason why I want to go over the response with some statisticians. Anyhow, here are the points that you might find useful:
1. Relationship of Issue 1 (ban on gay marriage) to turnoutOne very perplexing (to me) part of his report was that there was a positive correlation between turnout and voting “yes” on Issue 1, which is responsible for an increase in turnout of one half percent to 2%. That was perplexing to me because we have always been told that a large turnout favors Democrats, but if voting “yes” on Issue 1 led to increased turnout, then that would appear to mean that increased turnout favored Bush.
So I asked if the reason for the apparent correlation between turnout and voting “yes” on Issue1 could be due to voter suppression or electronic manipulation to reduce the vote in Democratic precincts, since this would make it
appear as if there was low turnout in Democratic precincts.
The response I received to this question was basically that it was impossible to tell because there was no data available on party affiliation, and with regard to the possibility that electronic fraud played a role in this, there was no evidence found for this “but the kind of data we have are not really suitable to digging into that. A forensic examination of administrative records would be needed to make the case for or against.”
2. Voter suppression due to insufficient voting machine allocation in Franklin CountyItem # 3 in my thread where I post my letter to Howard Dean discusses in depth the problem of voter suppression due to insufficient voting machine allocation in Franklin County:
http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=203&topic_id=380878&mesg_id=380878. I asked Professor Mebane about that, and whether that could be part of the reason for the apparent “low voter turnout” in Democratic precincts. He did confirm the problem in Franklin County by saying that: "There were also proportionally fewer voting machines in Franklin County's minority neighborhoods than in predominantly white neighborhoods." I don’t believe he specifically addressed the issue of whether that could have contributed to the apparent low voter turnout in Democratic precincts. But I don’t see any way that it could NOT have contributed to that. I mean, it seems to me that that’s just common sense.
3. Strange findings in Cuyahoga County I asked about some of the strange findings in Cuyahoga County, which I describe in detail in item # 2 of my letter to Howard Dean (see link above). Briefly, this includes implausibly low voter turnout in numerous precincts, coupled with the strange finding noted in the DNC report that in Cuyahoga County the normal expected positive correlation between voter turnout and machines per voter was ABSENT.
Here’s what Professor Mebane had to say about that: “I don't know what went on in Cuyahoga County. As I wrote in several places in the DNC report, there were many anomalies in the data from Cuyahoga County that warrant further investigation.”
4. Late vote surge in Miami CountyI also detailed in my Dean letter thread, item # 5, the late vote surge of 19,000 votes in Miami County,
after 100 % of precincts had reported, giving Bush an additional cushion of about 6,000 votes.
In response to my pointing that out, Professor Mebane noted that there were four precincts in Ohio that were outliers with respect to turnout, as predicted by support for Issue 1. Three of those precincts were in Miami County, two high outliers and one low outlier.
5. Other anomaliesI also pointed out other anomalies and suspicious findings, including electronic vote switching from Kerry to Bush in Mahoning County, and several anomalies in southwestern Ohio, including a swing of 37,000 votes to Bush (compared to his 2000 vote) from the three large counties in southwestern Ohio (Warren, Butler, and Clermont), unexpectedly poor performance of Kerry compared to the relatively unknown liberal Democratic candidate for Chief Justice of the Ohio Supreme Court, the infamous “lockdown” in Warren County to prevent anyone other than Republicans from observing the vote count, and the tremendously high increase in voter registration in these counties, despite the fact that the DNC report finds that this was supposed to be associated with Democratic gains. These are all detailed in item #s 4 and 6 of my Dean letter thread.
Professor Mebane’s response to all this was: “The problems you review for Mahoning and the other counties you mentioned seem to me to call for investigation, regardless of what is in the precinct data we analyzed.
6. Possible problems with central tabulators I asked Professor Mebane if it is possible or likely that if one or more of these episodes (i.e., the several findings that I describe above) do in fact represent fraud, that the fraud was perpetrated through manipulation of central tabulators, and therefore was not necessarily manifested by wide variance in precinct data?
His answer to that was: “The mysteries of central tabulation are important to understand and eliminate. Section VII and especially Section VIII of the DNC report address this. The DNC team did not have access either to the tabulation hardware or software or to the original ballots (where paper ballots were used).”
Well, that’s all I have for now. Not as much as we would like by any means, but I hope that there are at least
some useful points here. Some of Professor Mebane's responses that I've summarized here were already included in the DNC report, but they weren't included in the Executive Summary, which is what most people read.
Printer-friendly format
Email this thread to a friend
Bookmark this thread
(1000+ posts)
Send PM |
Profile |
Ignore
