As the early election returns start coming in this evening, all of the news media will begin discussing exit polls.  While there is validity to exit polls, in a close election, like this Presidential election, we need to be very careful in how we interpret the information.  First of all, exit polls really aren’t meant to pick the winner of an election.   They are designed to determine voting tendencies; how women appear to be voting, how men are voting, how people with various incomes and education are voting.  They provide information about voters, not who they voted for.  Of course the responses can and do give an indication but we need to look at those results with a critical eye.  The problem with exit polls is, they are often inaccurate at predicting the outcomes of elections. For example in 2004, exit polls projected John Kerry the winner, but in the end, there was almost a 3.0% swing and Bush won by greater than 2.0%.  Then there was the debacle, recall election in Wisconsin this past June.  Exit polls projected Tom Barrett tied or slightly in the lead in the election against Scott Walker.  Not only wasn’t he victorious, but Walker beat him convincingly, at 53% to 46%.  Even in the Republican primaries we saw flaws in exit polling.  Rick Santorum was predicted to finish third in Mississippi and Iowa; he won both states.    So as the networks are discussing the exit polls tonight, be a little skeptical.  It would be pretty unfortunate for a candidate to be presumed a winner, only to have the actual votes contradict that.