A note on exit polls before the deluge
posted at 1:41 pm on November 6, 2012 by Ed Morrissey
Election Day has arrived, and all of the seriousness and silliness that goes with it. We’ll bring you updates during the day on the front page and in our Green Room, and later this evening we’ll have open threads for each time zone and poll closings. Starting at 6 pm ET/3 pm PT, Guy Benson and I will appear on the Hugh Hewitt Election Night Marathon, which will be streamed live at Hot Air.
One particular piece of silliness will get a lot of attention: exit polling. If you watch television, listen to radio, or read news sites, you’ll start getting data from exit polls by about mid-afternoon. When you do, be sure to keep one fact about them in mind in order to put them into their proper perspective:
Oh, I don’t mean completed exit polling. We use that all of the time, but we use it for its intended purpose of demographic analysis. Pollsters ask people leaving voting stations to complete surveys, and take the data from those willing to share it. Later, those reports are matched to the actual vote in order to weight the raw data properly, since the sample is so uncontrolled, especially since different precincts will have different response rates, and not every precinct is covered by pollsters. Only later, when all of the votes have been counted and the raw data adjusted, can even the demographics be considered reliable.
Exit polls aren’t really intended to predict outcomes ahead of the vote being counted. Media organizations will use the raw data in specific ways — generally to gauge outcomes in bellwether precincts — and match them to vote totals, earlier polling, and historical election results to make calls on which candidate will win the state. Even then, though, they can’t do that without (a) the actual vote counts being reported after the polls have closed, and (b) complete sets of raw exit-polling data.
The “early exit polls” released in the afternoon are nothing more significant than a tease for data-starved media outlets, and data-starved media consumers. Taking them seriously leads to embarrassing outcomes, as the adviser to John Kerry who started calling him “Mr. President” discovered in 2004 when the early exit polls supposedly showed a five-point victory in the offing for Democrats.
So, enjoy the game of Wild Speculation that follows the release of early exit poll data, but don’t take it seriously. Remember that we’ll know within a few hours of poll closings who won this election, and the game of Wild Speculation won’t change that a bit by this point in the election cycle.