Executive summary of
the report:
The authors list several reported problems with the election. They claim the exit polls are the only reliable way to evaluate the integrity of the election due to these discrepancies. Exit polls suggest that Kerry should have received 5.5% more of the popular vote.
The organizations that actually performed the exit polls claim that the error was with their own sampling procedures, and that more Kerry voters volunteered to be . The authors summarily dismiss this due to lack of evidence, then submit that Bush voters were actually overrepresented such that the margin for Kerry should have actually been greater than the 2.5% suggested by the polls. They offer no evidence of this.
The authors then consider three possible causes for the discrepancy between exit poll results and actual election results:
1. Statistical sampling error ? or chance
2. Inaccurate exit polls ? Kerry supporters responded in greater numbers than Bush supporters.
3. Inaccurate election ? the voters? intent was not accurately recorded or counted.
They, and the polling firms, ruled out #1, since the polls were conducted in accord with standard procedures. They do some handwaving to throw out three results that are completely and grossly out of line with the actual vote count (>4 standard deviations off), then claim that the rest of the results are pretty close to a normal distribution. To me, the fact that three states were this far off says that something is wrong with the methodology. The authors then throw out some extraneously large numbers to overstate how small a chance there is of some of the disparities, again assuming that the proper methodologies were used. So, now we're left with a normally distributed result that is shifted one standard deviation from the expected mean (bias towards Kerry in the exit polls). Fair enough.
#2 - The polling companies rule out possible sources of bias due to exit polling error except for the random voter selection process. ?While we cannot measure the
completion rate by Democratic and Republican voters, hypothetical completion rates of 56% among Kerry voters and 50% among Bush voters overall would account for the entire Within Precinct Error that we observed in 2004.? The authors then present a graph that they claim counters the report by showing that more exit poll participation was received in precincts that went for Bush. The graph is a shoddy piece of work, besides being clearly fashioned to support the agenda of the authors. If I put a figure like that in a lab report, I would likely get kicked out of school - it's that bad.
That aside, the conclusions that they draw from this single graph are questionable at best, bordering on complete conjecture. But, to this point, the authors state that it is unlikely that these two factors had anything to do with the skew of exit poll results versus real election results. They pin the blame solely on #3.
#3 - The authors state "The many anecdotal reports of voting irregularities create a context in which the possibility that the overall vote count was substantially corrupted must be taken seriously." Appealing to anecdotal evidence to declare vote fraud in a presidential election is a bit of a stretch. They complain that the polling companies did not publish a number of important statistical testing results that could give a better picture, but it does not appear that any attempt was made to contact the polling companies to get the data to test themselves. Sounds to me like they would prefer to sweep it under the rug of 'conspiracy' rather than get to the bottom of it, as these tests are really trivial to complete (I could do it by myself on my computer in a matter of minutes).
And, the closing line: "We invite all those who care about democratic processes in this country to join us in fully investigating and explaining what really happened in the 2004 Presidential election."
It looks like a couple PhD's from around the country wanted to raise a ruckus so they spent an afternoon throwing together a 5 page summary of someone else's report. They didn't offer any statistical evidence of their own, instead just conjecturing on the soure of disparity. It should be pretty clear to anyone who reads this paper that the authors knew their conclusions before they started their work. Sorry, but I'm definitely not impressed.
Facebook
Twitter