The 2004 presidential exit polls were wildly off the mark in swing states; the difference between the expected and actual results was not randomly distributed, it was all in Bush’s favor.
Because of these discrepancies, I studied the election results in Colorado, Florida, Iowa, Michigan, Minnesota, Nevada, New Hampshire, New Mexico, Ohio, and Pennsylvania. What I found were not answers, but more questions.
Because of the amount of data, I made two assumptions to simplify my inquiry. The first was to examine the states in the order of the reported magnitude of the difference between the expected and the actual results; I started with New Hampshire where the difference was 9.5 percent. And, I did not consider anecdotal evidence; I disregarded reports that a voting machine had malfunctioned in a particular precinct, for example, and, instead, looked for systemic failures.
New Hampshire: Kerry was projected to win by 10.8 percent and actually won by 1.3 percent, 9,274 votes. Bush narrowly won New Hampshire in 2000, but this time Democrats expected to win as they had conducted an aggressive registration drive. However, New Hampshire permitted same day voter registration and there were 96,000 registrations on Nov. 2, about 15 percent of the turnout; Democrats accused Republicans of taking advantage of the rules and bringing in questionable new voters. It’s not clear how these voters were profiled by party registration but exit polls indicated that 44 percent of voters said there were Independent, 32 percent Republican, and 25 percent Democrat—fewer Democrats than were expected.
An independent assessment of voting results in areas serviced by optical scanning equipment indicated irregularities. For example, in 2000 Al Gore carried Newton Township by 126 votes, whereas in 2004 Kerry lost by 57 votes. There will be a recount in 11 precincts that produced anomalous results.
Ohio: Kerry was projected to win by 4.2 percent and lost by 2.5 percent—using unofficial numbers. There are have been many reports of voting irregularities; Greg Palast noted some at www.gregpalast.com/detail.cfm?artid=395&row=0. Most counties in Ohio continued to use antiquated punch-card equipment; the problems experienced were similar to those that plagued Florida in 2000.
Forty percent of Ohioans registered as Republicans, 35 percent as Democrats, and the balance Independents. Kerry got more votes and a higher percentage of the total vote than Gore did in 2000, but still lost. Former Ohio Senator John Glenn said that Republicans won because they did a better job getting out the vote.
Pennsylvania: Kerry was projected to win by 8.7 percent and actually won by 2.2 percent. There were remarkably few voting problems reported in Pennsylvania, which has a mixture of old and new equipment.
While Pennsylvania voter registration roles show that 48 percent are registered Democrats—versus 41 percent Republican and the balance varieties of Independent—exit polls showed that only 41 percent of those who voted said they were Democrats.
This suggests that either fewer Democrats showed up to vote than were expected, or those who did voted as if they were Independents, i.e. crossed over and voted for Bush. Pennsylvania exit polls indicated that voters who made their Presidential decision at the last moment favored Bush, a characteristic that was not true in the other swing states, where last-minute voters favored Kerry.
Minnesota: Kerry was projected to win by 9 percent and won by 3.5 percent. There was an unusually large turnout, 77 percent, and Kerry’s plurality was almost 100,000.
Minnesota is one of the states that permit same day voter registration and this may have been a factor in the final results. Relative to the 2000 election, Bush picked up support among both Republicans and Democrats.
There were no major problems reported with voting equipment—most counties used optical-scan equipment but rural areas still have paper ballots.
Florida: Bush was projected to win by 0.1 percent and actually won by 5 percent, a margin of approximately 380,000 votes.
There are many articles on the Internet concerning Florida vote irregularities. The most recent, and damming, was conducted by UC researchers and released on Nov. 17, http://ucdata.berkeley.edu (under “voting”). The carefully conducted study concludes that electronic voting machines—touch-screen devices—appear to have systematically over counted in Bush’s favor. There is another study found at http://blog.democrats.com/florida that shows problems in counties with optical-scan equipment. There are several ongoing investigations in Florida; including one led by Bev Harris of blackboxvoting.org.
I don’t have space to chronicle all that I found in the other five swingstates, but in each case there were anomalies; the most worrisome occurred in New Mexico, which has yet to declare its official results.
The bottom-line is that there are troubling problems in at least three swing states and issues in others. The anomalous exit-poll results may have resulted from a very effective GOP get-out-the-vote effort that brought more Republicans than expected to the polls. Unfortunately, the exit-poll discrepancies may also indicate political chicanery in some states. It is in the best interests of the American people for all these matters to be carefully investigated.
Bob Burnett is a Berkeley writer and computer scientist best known as one of the executive founders of Cisco Systems.