US Count Votes Response to Dan Tokaji, Assistant Professor of Law, Ohio State University
http://moritzlaw.osu.edu/blogs/tokaji/2005/01/mea-culpa-rejected.htmlBlumenthal and Tokaji state in their critique of USCountVotes study of the Edison/Mitofsky report
http://uscountvotes.org/ucvAnalysis/US/USCountVotes_Re_Mitofsky-Edison.pdf that “the argument that Kerry voters participated in exit polls at higher rates than Bush voters” is not refuted by US Count Votes. This misstates our point; there was nothing to actually refute. The burden of proof is on Edison and Mitofsky to provide statistical evidence supporting their assertion of exit poll inaccuracy, and neither they (nor Tokaji and Blumenthal) have done so. The data we cited on page 36 simply cast additional doubt on their unproven assertion.
Tokaji and Blumenthal note that there were also widespread discrepancies between exit polls and the official results of the 1992 Presidential election. Their implicit argument appears to be that two large unexplained exit poll discrepancies in a period of twelve years is simply additional evidence that exit polling is unreliable. The alternative interpretation that two episodes of widespread vote counting errors may have occurred within a period of twelve years is equally coherent. There is no reason to presume that widespread vote counting errors are a brand new phenomenon. Rather than simply selecting one of the three alternative explanations for exit poll discrepancies on the basis of personal whim or bias, US Count Votes prefers to subject all suspicious exit poll and voting discrepancies to objective, precinct-level statistical and demographic analysis. It should also be emphasized that an unquestioned predisposition to automatically discard US exit polls whenever they show significant variance from official vote tallies guarantees that widespread fraud, if it does occur, will be certain to go unchallenged.
Tokaji said” …the fact that the exit polls were close to accurate among the few voters who still use hand-counted paper ballots proves, well, pretty close to nothing.”
To the contrary: we believe an indication that all methods of recording votes, except for paper, are subject to large and unexplained exit poll variances that consistently favor the same candidate and are greater than the margin of error clearly deserves further investigation. Had Edison/Mitofsky released their raw precinct-level data, their assertion that there were no significant correlations between voting machine type and Within Precinct Error (WPE) could be independently tested.
In our response to Edison/Mitofsky our primary concern was to note there are unexplained anomalies that may indicate pervasive nationwide problems with official vote counts, and that these are certainly worthy of further statistical investigation. However, statistics can only identify the presence of vote counting anomalies. It would take a much different investigation to determine the root cause, and that is not our mission.
However, Tokaji raises the specter of a "grand conspiracy" to impeach our credibility, and we must object. His assertion that bribery or large-scale conspiracy is required to cause widespread corruption of the official vote count is simply mistaken. US Count Votes believes that it is not difficult to identify credible risks of systematic vote manipulation that clearly justify our proposal to analyze the accuracy of the national vote. There are at least two coherent scenarios where pervasive corruption of vote counting could occur without a massive, coordinated conspiracy.
The first coherent scenario of pervasive corruption of vote counting involves manipulation electronic voting systems. Modern voting systems are geographically-dispersed distributed computing systems, and computer security professionals regard all such systems as inherently vulnerable. Regardless of the nature and type of pre-release testing and certification, the process of replicating software to the hundreds or thousands of devices to be deployed in the field is inherently vulnerable to systematic manipulation. It would only take a few malicious insiders to corrupt the master copies of voting software; thereby undetectably altering the behavior of thousands of voting machines. Worse yet, rogue insiders have first-hand, detailed knowledge of any internal security mechanisms and are thus ideally equipped to subvert them. Since just three companies tally upwards of 80% of the US vote, the potential risk of systematic nationwide manipulation of vote counting equipment by a small group of rogue insiders cannot be dismissed out-of-hand. Banks know that they face a far more serious threat from embezzlers than from gunmen in the lobby; that’s why banks have extraordinarily comprehensive and elaborate audit mechanisms that are verified on a regular basis by insiders and independent outsiders.
Standard security best practices have evolved over the last fifty years to mitigate ever-present risks of insider manipulation. To support the highest degree of integrity, many of the most important financial transactions are still officially tracked on, or reconciled to, paper records. The key point is to always require vulnerable systems to have the capability to be independently audited, and to actually audit them on a regular basis. Today's standards for security and audit control of voting equipment that are considered acceptable to many county, state and federal election officials would be deemed unacceptably risky in other domains – especially in the banking and financial services industries. It is a fact that if a publicly-held corporation deployed a financial system featuring the same level of security controls as that provided by many popular electronic voting systems, it would expose that company's Board of Directors to prosecution under the provisions of the Sarbanes-Oxley financial accountability law.
These considerations are so obvious to computer professionals that an overwhelming majority of 95% of the computer scientists and software engineers polled by the ACM computer society in 2004 opposed deployment of unauditable electronic voting systems. The vendors of electronic voting systems and their advocates in academia and industry are out-of-step with the mainstream. Computer security professionals know that human ingenuity knows no bounds when the prize is sufficiently large; in fact, anecdotal reports of problems with computer systems that tally votes date back as far as the late 1960s (see
http://www.ecotalk.org/BevHarrisBook2.pdf for original primary source citations).
There is a second coherent scenario where pervasive vote counting problems could occur without a massive, coordinated conspiracy: given that there are smart, ruthless people in any sizeable organization that are always on the lookout for signals that it's permissible to do anything in pursuit of wealth and power, a sense of impunity may be all that is needed to establish a climate that tolerates unethical behaviors. Enron is a classic instance where lawlessness flourished without the need for explicit oversight and coordination; indeed, oversight may even be discouraged to provide plausible deniability. No political party is immune to the temptations of power; US Count Votes intends to statistically identify and report vote counting anomalies favoring any political party.
Tokaji: “Blumenthal notes here that errors favoring Kerry tended to occur in places where the interviewer was under 35, inexperienced, and had a graduate degree -- and that 'it is not hard to see the underlying attitudes and behaviors at work might create and exacerbate the within-precinct bias.'' According to Blumenthal, Bush voters approached by young interviewers could mistake them for Kerry supporters, especially if they were approached by poll takers stationed near groups of electioneering Kerry volunteers.
It is hard to take this scenario seriously. The within precinct errors (WPEs) varied less between age groups than the overall WPEs. No one party has a monopoly on electioneering. We’re sure that there are many conventionally-dressed, friendly and unthreatening poll takers who are also under 35, inexperienced and with a graduate degree. Assuming it’s not a joke, this hypothesis can’t be independently tested without access to raw precinct-level exit poll data correlated with poll taker demographics.
Simply because one can make a reasonable hypothesis that appears consistent with the data does not mean that it is also the correct hypothesis. Consider that in the recent Ukraine election, the poll results and the election results differed just like ours, and there also were numerous anecdotal reports of election irregularities in both countries. Reasonable hypotheses to explain why the poll results were incorrect in Ukraine included (a) opposition voters were “more likely to respond” - the same behavior ascribed to Kerry voters here; or (b) the poll results were weighted toward the more urban and western districts, where opposition support was stronger or (c) the poll results understated higher turnout in eastern and rural areas. Each of these is a simple hypothesis that "explains" the exit poll discrepancy, but the hypothesis actually adopted by the U.S. government was fraud. This choice was made apparently without first specifying a coherent theory of how such a fraud could have been committed.
A consistent theory is just that: a consistent theory. Whether it is valid as well depends entirely on whether it holds up to empirical analysis. Rather than grasping at straws, given a host of well-documented risks and vulnerabilities and persistent and troubling anecdotal reports of vote counting irregularities, the possibility that systematic corruption of official vote counts may actually have happened simply cannot be dismissed out of hand.
Tokaji's unquestioning belief in the integrity of the vote count should be systematically tested as proposed by US Count Votes.
Bruce O'Dell, US Count Votes, Vice President
Printer-friendly format
Email this thread to a friend
Bookmark this thread
(1000+ posts)
Send PM |
Profile |
Ignore
