Polling 101, Vote Maps, and Punditology

Kari Chisholm FacebookTwitterWebsite

These days, it seems everyone's obsessed with polling. Maybe it's part of the post-2000 hangover, but it seems that everyone want to know which polls are swinging which way, what the state-by-states are doing, blah blah blah.

I thought I'd share a few resources. First, dig into William Saletan's Consumer Guide to Polls. Good stuff. Then, go back to the BlueOregon archives and read Jeff Alworth's excellent How to Read a Poll

Once you're up on polling, dig into the state-by-state maps. I've collected a bunch of them at my personal blog Politics & Technology in a post helpfully entitled Looking for Electoral Vote Maps? I'm also keeping a running average of all the maps I can find.

Another approach is to boldly declare "Screw the Polls!" and instead ask pundits and other observers. Sometimes, the collective feeling in the gut is closer to right than wildly swinging polls. To that end, check out the Punditology Challenge, wherein you can make your own best calls on who will win and who will lose. Once you make your picks, you'll see the collective wisdom. (Note: The group is almost always right, except when we're wild-ass wrong.)

Finally, there's the Iowa Electronic Market where real people bet real dollars on the outcome of the election. Right now, it's Bush 53, Kerry 47. That means it's VERY close. That's a proxy for probability not the actual vote margin. (Clinton led Dole by around 90/10.)

  • Betsy (unverified)
    (Show?)

    True polling wonks should also check out a relatively new blog, Mystery Pollster, where pollster Mark Blumenthal gets into the nitty-gritty hardcore details, including a six part series about the whys and hows of polling 'likely voters.'

  • the prof (unverified)
    (Show?)

    Add this one:

    www.uselectionatlas.org

    It is better than the electoral link provide by Kari because it averages across the most recent polls rather than taking just the most recent.

    Both provide nice electoral college predictions.

    Here's another list of sites: http://academic.reed.edu/poli_sci/electionsites.pdf

  • (Show?)
    <h2>Prof... I'm not sure which one you're talking about. My site doesn't average any polls at all. It's averaging lots of different electoral maps - each using different methodologies. Admittedly, that can introduce a new kind of bias. But, it's probably better than looking at one single electoral map.</h2>

connect with blueoregon