Polls Are Like Horoscopes

In response to Ezra’s great compilation of polls from the 2000 election, I was going to track down some horoscopes from mid-October 2000 that indicated that Gore would sail to victory or whatever. The moral of the story would be that astrology is more reliable than polling and we all would have had a good hearty laugh at Gallup’s expense. But I encountered two problems :

  • It’s a lot harder to find archives of horoscopes from four years ago than you’d think. If I didn’t know any better, I’d almost think that they didn’t want you to fact-check their asses.
  • After reading two or three horoscopes, I started feeling sick. Not in a “Dammit, I wish I had gotten a flu shot” way, but more of a “If I have to read any more of this new age bullshit I’m going to vomit all over my keyboard” way.

    Sorry guys, no astrology post for you. It probably would have been funny, but it just wasn’t worth reading through all the crap about star signs and moons over my hammy. If I wanted to subject myself to vague metaphysical nonsense based on the pseudo-spiritual writings by people who don’t understand science and are just looking for an easy way to explain the world, I’d go to church.

    Just to push home the point about how unreliable polling is, however, here’s what Ezra wrote about the polls from Oct. 18, 2000 :

    No matter who you think they’re favoring, the polls are proving wildly unreliable, and it occurred to me that, no matter who wins in November, the polling companies are going to have a lot to answer for if they want to be taken seriously in the future. But I spoke too soon. In checking my thesis, I went back to 2000 and looked at the polling released during the exact same period.
    . . .
    Note that not a single one of these polls gave any advantage to the actual winner of the popular vote. So while the media crows about this or that metric, remember that they were all wrong at this point in 2000. While the polls did tighten in the ending days of the race, only 3 of the 20 polls taken in the final two days gave Gore any sort of lead. And we’re much closer than he was, with much more money and a ground organization he could only dream of. We’re in good shape, folks.

    While we’re on the subject of polling, are there any statisticians out there with the 411 on the Census Bureau’s American Community Survey and whether or not it’s prone to the same flaws as we’re seeing here? I know the sample size is supposed to be much bigger, but if they do the same questionable “normalizing” stuff, then we could be in trouble.


    • posted by greg on October 19, 2004 @ 10:31 am

    2 comments

    1. If I wanted to subject myself to vague metaphysical nonsense based on the pseudo-spiritual writings by people who don’t understand science and are just looking for an easy way to explain the world, ….

      I thought you were going to say you’d read some BushCo policy initiatives.

      Comment by eRobin — October 19, 2004 @ 11:28 am

    2. polls are part of the reality-based community. Bushco creates their own reality, so even if he loses, he could claim he won. would he stoop to a ‘bloodless coup?’ refuse to move out of the white house? change the locks? my favorite bumper sticker this year is “Bush – let’s not elect him in 2004 either.”

      Comment by freedombychoice — October 19, 2004 @ 7:41 pm

    Copy link for RSS feed for comments on this post

    Sorry, the comment form is closed at this time.