Exit polls are almost never wrong. They eliminate the two major potential fallacies in survey research by correctly separating actual voters from those who pretend they will cast ballots but never do and by substituting actual observation for guesswork in judging the relative turnout of different parts of the state.
snip
So reliable are the surveys that actually tap voters as they leave the polling places that they are used as guides to the relative honesty of elections in Third World countries. When I worked on Vicente Fox's campaign in Mexico, for example, I was so fearful that the governing PRI would steal the election that I had the campaign commission two U.S. firms to conduct exit polls to be released immediately after the polls closed to foreclose the possibility of finagling with the returns. When the polls announced a seven-point Fox victory, mobs thronged the streets in a joyous celebration within minutes that made fraud in the actual counting impossible.
But this Tuesday, the networks did get the exit polls wrong. Not just some of them. They got all of the Bush states wrong. So, according to ABC-TV's exit polls, for example, Kerry was slated to carry Florida, Ohio, New Mexico, Colorado, Nevada and Iowa, all of which Bush carried. The only swing state the network had going to Bush was West Virginia, which the president won by 10 points.
http://www.thehill.com/morris/110404.aspx