Exit Polls and Voter Fraud: A User-Friendly Explanation
by SEAN SABATINI
Electronic voting machines were supposed to save us from the nightmare of hanging chads. The day after the election, a lot of Americans learned for the first time that most of these machines are owned by private companies who refuse to divulge exactly how they work; that computer security experts have been highly critical of them; that they 've already experienced serious failures, and that many of them leave no paper trail for backup.(1) Since then, online forums have been jammed with claims that the vote was electronically hijacked. Right now, the mainstream media is looking the other way, but members of Congress have already launched an investigation. And it all started with exit polls, those pesky little interviews done with voters right after they cast their ballots.
Exit polls actually consist of a long list of questions about a voter 's choices and the various factors that influenced those choices. They provide a wealth of information for political strategists, and have long been considered an accurate gauge of the American electorate, as they poll actual voters, and are not influenced by assumptions or guesswork. And they have a good record of picking winners. In some countries, election monitors have even used them to help measure the validity of official election results.
In the U.S., exit polls are now done by an outfit called the National Election Pool, which is financed by a consortium of big media outlets. The media receive the numbers throughout the day, analyze them, and use the results to start calling election winners almost as soon as the polls close. Later, the data is made public.
But this year, a couple of remarkable things happened. One, the results were leaked early, and spread throughout the internet. Two, the exit polls said John Kerry won. In the wee hours of the morning, the poll numbers were revised to better match the official election results. (How and why is another story.) But by then the damage was done. Accusations flew that the polls were right, and the electronic voting systems had been compromised.
Were the exit polls really that far off? Here 's what happened:
Over the course of election day, several successive sets of exit polls were released by various websites, covering the swing and semi-swing states. The information was somewhat spotty, as not all polls were available for all states. The numbers also got a little mixed up, and they differed from one site to the next, but not by much. As one editor at Slate.com put it: "In some states the sources disagree about the specific margin by which a candidate leads, but never about which candidate is out front. " After the polls closed, CNN posted exit poll data on its website.
A look at how the polls differed from the official results in each swing state should make things pretty clear. Note that we 're looking at original numbers here, not the revised data now posted on major news sites.
First off, we have to throw out New Hampshire, which Kerry won by one point. Early polls showed him with an unrealistic 17 point lead , and while it later dropped to 10, that 's still an improbable lead for a state that Bush won in 2004. Sadly, the numbers for New Hampshire are too unreliable to analyze properly. But here are the other states, with percentage points rounded off:
In Arkansas, the polls had Bush up by 7-9 points, depending on which numbers you use. He won by 9 points.
In Missouri, Bush closed up 5-8 points. The official result: Bush by 8 points.
On the Kerry side, we have Maine, a state where they still use paper ballots, and even count 35% of them by hand. It 's hard right now to get the exit poll numbers for this state, but they seem to have matched Kerry 's 8 point victory pretty well. So far, so good.
Moving over to Iowa, things get a little more interesting. Here we see Kerry either tied, or up by 1-2 points, all day. But in the official tally, Bush gains 2 points. Not a big shift, but enough to win a state that he lost in 2000.
In Colorado, the exit poll numbers fluctuate a little more widely throughout the day. But Bush is always leading. The final polls show him up 1-3 points. He wins by 5, another gain of at least two points.
Wisconsin is just the opposite: Kerry leads all day and ends with a 3-5 point advantage. But he wins by just one point, another shift of at least 2 points--again in Bush 's direction.
In Louisiana, exit polls give the President a whopping 11-13 point lead. But he does a little better still, winning by 15.
In hotly-contested New Mexico, Kerry is slightly ahead all day, ending with a slim 1-2 point lead. When the votes are counted, though, Bush comes from behind and wins by one point. The 2-3 point shift gives Bush another state that he lost in 2000.
Up in Michigan, Kerry maintains a solid lead all day, ending 4-6 points ahead in the polls. But he wins by a single point, a 3-5 point gain for Bush.
Minnesota produces some unrealistically high pro-Kerry numbers early on, but by the evening they have settled down, possibly because of corrective steps taken by the polling company. Kerry ends up with a 6-10 point lead--again, depending on which numbers you use. He wins by 3; another gain of at least 3 points for Bush.
What is going on here? Republicans have said that these numbers fall within a certain margin of error, that the sample size may be too small, and so on. This is nonsense. Random fluctuations would not all benefit the same candidate. Executives of the company that ran the exit polls have offered up their own excuses: the pollsters couldn 't get too close to the voting machines (so what?), and Bush supporters just didn 't want to speak up (since when?). If there was some inherent flaw in this year 's polls that produced a 2 to 3 point skew for Kerry, it has yet to be explained.
And it gets worse.
In West Virginia, the polls show Bush trouncing Kerry by 9 points. When the votes are counted, Bush is up by 13, having picked up another 4 points.
A far closer race occurs in Nevada, where the candidates run neck and neck all day, with the lead shifting between them. It 's a good test for Nevada 's new all-electronic voting system, which produces a voter-verifiable paper record in case of recounts. Kerry ends the day just barely up by one point in the polls, but Bush wins by 3. Another impressive gain of 4 points for Bush, and, incidentally, just enough of a margin of victory to keep Nevadans from asking for a recount.
Pennsylvania, like Minnesota, starts off with some out-of-whack numbers in the morning, but then settles down. By the end of the day, Kerry has a solid 7 point lead. He wins, but only by 2 points--a disturbing 5 point jump for Bush.
In the crucial state of Ohio, we see Kerry with a steady lead all day. When the polls close, all eyes are on Ohio, and the unauthorized numbers on the internet have Kerry ahead by 2 points. Then CNN posts exit poll data showing Kerry up by a good 4 points. Somehow, Bush pulls off a mysterious 6 point gain and wins by 2 points.
North Carolina is a real beauty. It was a pro-Bush state, but Democrats had hoped to swing it their way by having favorite son John Edwards on the ticket. It didn 't work. Bush was ahead in the polls all day, and ended with a strong 4 point lead. Amazingly, though, he ended up winning the state by 12 points, an inexplicable gain of 8 points.
Meanwhile, down in Florida, 1.6 million more people were voting for President than had done so in 2000, the year that Bush and Gore essentially tied. Exit polls showed more new voters were going for Kerry, as predicted, and far fewer voting for Nader this time. Not surprisingly, Kerry was slightly ahead all day and ended with a 2 point lead. But of course, Bush won. And not by one or two points, as he did in Ohio and New Mexico. In Florida, where vote-tallying machines have been seen counting backwards, Bush gained a full 7 points on his exit polls and beat Kerry by 5 points.
Detailed mathematical analyses of these and other data are now available on the web, but you get the idea.
What does it all mean? It means that Kerry conceded too soon. These figures by themselves may prove nothing, but they raise a tremendous red flag. And now analysts are identifying voting patterns in Florida that border on the impossible.(2) Of course, computerized theft is certainly not the only means of manipulating the vote. Muckraking journalist Greg Palast has written an excellent piece on how plenty of Kerry 's supporters in Ohio and New Mexico had their votes trashed the old-fashioned way.(3) It 's also now apparent that the initial, frenzied attempts to link exit poll discrepancies to states without paper trails was misguided. Most states use a combination of voting methods, both electronic and manual. But even paper ballots are usually counted electronically, and the data is transmitted to central computers, often over phone lines. It 's a system that computer experts from Stanford, MIT and Johns Hopkins have criticized as weak and easily corrupted.(4)
What needs to be done? First, we need to shake off the notion that widespread voter fraud could never happen in America. Computer hacking may be a relatively new phenomenon, but crooked elections are not. And corruption is as old as politics. Then, we have to discover the truth. As traumatic as it may seem, we need manual recounts, wherever possible, in Ohio, New Mexico, Iowa, Nevada, and Florida. We need to have computer security experts examine every aspect of the electronic voting system for signs of tampering. With the evidence that 's coming in now, we may well need a full-scale criminal investigation in Florida.
What are the chances of any of this happening before the electoral college meets to crown George Bush? Not so good. The Democratic National Committee has stuck its head in the sand, for reasons that have a lot to do with politics and little to do with the good of the country. John Kerry, who promised to make every vote count, amassed an army of lawyers and 50 million dollars to pay for recounts and legal expenses. Then he packed up his tent and went home. He told the people who worked long and hard for him that he wanted to put his arms around them all, but many are now asking why he doesn 't fight for them instead. After all, Nader and Cobb can ask for recounts, but somebody has to pay for them.
Still, whatever happens in the next few weeks, this issue is not going to die. MoveOn is starting to step up, and actions are being taken on the local level in many states. A group called Black Box Voting (5) is filing Freedom of Information acts across the country to get computer logs, and is seeking more volunteers and more money.
The fight goes on, because this is not just about just about who wins today. It 's about ensuring the integrity of the American electoral system. Republicans will blame it all on conspiracy nuts and sore losers, but it doesn 't matter. In the digital age, we have to make e-voting secure and honest, or we can kiss democracy goodbye.
Sean Sabatini is an antiwar activist and writer who lives in Nevada. He can be reached at firstname.lastname@example.org
Add your comments below