A Final Word on Exit Polls

Well, it looks like Cautious Networks 1, Eager Bloggers 0 on the use of exit polling data to predict the result. But I was a little incredulous when Bill Schneider, CNN’s “polling expert” said with a hint of schadenfreude that exit polls were never meant to predict, just to show demographics. Bill, if it’s a skewed sample, using a subset of it is even sillier than merely presenting the headline result. About the only exception to this that I can see would be if the networks presented not the vote for men and women, but the gender voting gap. But they don’t. Once we knew that the exit polls were badly off on the headline result, it’s hard to see why CNN kept running them. Schneider’s no fool, so he must’ve known he was blathering.

This entry was posted in Uncategorized. Bookmark the permalink.

2 Responses to A Final Word on Exit Polls

  1. Neil Hickey says:

    Hi Andrew,
    Good comment about the exit polls – what are they paying these guys for? Maybe we should give them credit for not calling it for Nader … or Dukakis!
    My experience conducting exit polls have been a lot more accurate though. The only exit polls we do are at the prepolling – and there’s only venue in each electorate to do that. So people of all (legal) ages, gender, occupations, political affiliations are all walking out of the one venue. It’s normally a very representative spread. When we did exit polling for the Queensland election this year we correctly tipped four out of four seats and got three out of four margins correct.
    It’s the only polling I trust. To me, trying to establish a trend by doing exit polling on the morning of election day is pretty flukey.
    Cheers,
    Neil Hickey

  2. Geoff Robinson says:

    Very dimly I seem to recall that in the 1984 Oz election exit polls were tried. They predicted a strong swing to Labor!

Comments are closed.