Don't Read Too Much into Exit Polls
November 04, 2008Election '08
In a matter of hours, you are going to start hearing talk of exit polls.
But, before you get ahead of yourself, there's a question that needs to be asked: Should you trust the exit polling data? The short answer is: No. The longer answer is: Noooooooooooo. Right now, if there's one memory that remains -- stinging -- to a nation of Democratic voters it's the memory of a slate of crazy Kerry-leaning exit polls that made it look like Bush was going down to defeat at about 4:30pm on Election Day. It didn't turn out that way.
Marc Ambinder sums up the exit poll phenomena, thusly: First - "anyone who claims to have exit poll data before then is either lying or has really, really good sources." Second - "The problem with the exit polls has never really been a problem with the exit polls. They've been a problem with people incorrectly interpreting the exit polls; people who don't know what the exit polls actually are." I think that this is more or less true. Exit polls could be astoundingly accurate or terrifyingly wrong, but either way, there's one thing that everyone who pimps them has in common: none of them really have any idea what they are talking about.
Nate Silver, the proprietor of FiveThirtyEight.com, has a very good primer on exit polling up on his site that's worth a long look, but to summarize: exit polls have a much larger margin of error than regular polls, they tend to skew to the Democratic candidate, they proved to be really bad predictors this year, they actually miss a ton of voters, and, ultimately, they can never be reliably sourced. (Believe it or not, I ran across a website yesterday that purported to have exit poll data!)
So, if you must indulge in exit polling data, feel free. Accept it as another part of the election scenery, and do not put a lot of stock -- or pin too many of your hopes -- to these results. I cannot caveat this emptor enough, people.
In a matter of hours, you are going to start hearing talk of exit polls.
But, before you get ahead of yourself, there's a question that needs to be asked: Should you trust the exit polling data? The short answer is: No. The longer answer is: Noooooooooooo. Right now, if there's one memory that remains -- stinging -- to a nation of Democratic voters it's the memory of a slate of crazy Kerry-leaning exit polls that made it look like Bush was going down to defeat at about 4:30pm on Election Day. It didn't turn out that way.
Marc Ambinder sums up the exit poll phenomena, thusly: First - "anyone who claims to have exit poll data before then is either lying or has really, really good sources." Second - "The problem with the exit polls has never really been a problem with the exit polls. They've been a problem with people incorrectly interpreting the exit polls; people who don't know what the exit polls actually are." I think that this is more or less true. Exit polls could be astoundingly accurate or terrifyingly wrong, but either way, there's one thing that everyone who pimps them has in common: none of them really have any idea what they are talking about.
Nate Silver, the proprietor of FiveThirtyEight.com, has a very good primer on exit polling up on his site that's worth a long look, but to summarize: exit polls have a much larger margin of error than regular polls, they tend to skew to the Democratic candidate, they proved to be really bad predictors this year, they actually miss a ton of voters, and, ultimately, they can never be reliably sourced. (Believe it or not, I ran across a website yesterday that purported to have exit poll data!)
So, if you must indulge in exit polling data, feel free. Accept it as another part of the election scenery, and do not put a lot of stock -- or pin too many of your hopes -- to these results. I cannot caveat this emptor enough, people.
