Settling once and for all the validity of the EXIT POLL SAMPLE

[Firespirit]

Sorry to toot my own horn here, but with althecat's exit poll data, we can determine whether the national poll sample was good or not. Here's the idea that came to me on this thread (http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=203&topic_id=214701):

The poll is for the House and the Presidential races.

The House race doesn't break it down into Congressional districts, so it can't be used to validate individual House races. BUT, if we get the total votes cast for all House candidates, and break it down by region in the same way the poll did, we can prove whether it was a good sample. "The sample was bad" seems to be the main defense against the exit polls, and up to now we haven't been able to disprove that.

Conyers' people have the data, as do the CASE Ohio folks.

Any insiders from either group here? If so, and if you can spill this, are they already doing such an analysis?

If not...

It should be done. And I volunteer to help do it.

[CalebHayes]

1. Great idea

kick :kick:

[ISUGRADIA]

2. If it was a biased sample wouldn't this also show up in the

senate and governor exit polls (I know this does not cover all states but many of the key ones)

[Firespirit]

3. I am thinking it would be everywhere.

If the same people were polled. Which, by all accounts, was true.

This is really a gift; it has such a huge sample size (>13,000) that standard sampling error is pretty much insignificant. It covers the entire U.S. And it includes a second race -- the overall Congressional vote -- that can validate the Presidential race. What's more, everyone had a House contest on the ballot, which was not the situation with Senate and Governor races. I mean, this is a friggin' gold mine.

If the same people were polled, then what's good for one is good for the other.

To be honest, I'm hoping somebody with Conyers or CASEOhio will say that they're already doing this. I don't want to waste my time duplicating their work.

[TruthIsAll]

4. Breakthrough.You have the Firespirit. I'll leave that one to you.

Go get that mutha.

Put those naysays out of their misery.

[Firespirit]

5. OK, consider me on it

It's a lot of grunt work, gathering the results, but nothing too error-prone or tricky.

Can't promise when this will be done, but I'm shooting for early Monday.

[missouri dem 2]

6. Great concept and method to prove that the exit polls were right.

[izzie]

7. Exit polls I think were done by few people.

I wondered what the women was doing and finally asked what her job was and she told me. So I took it.No one else did while I was in the room to vote and I stopped at the tables they had with information and food. I do not know any one else that voted there that did. They told me they did not know they took an exit poll at that place. Also my answer was values as my values said I could not vote for Bush. To me it put all the things like jobs and wars in line.I just do not think it means what the news keeps saying.

[DoYouEverWonder]

8. How would you classify the precinct where you live?

Is it mostly REP or DEM? Is the population in your precinct representative of other precincts in your county? Or does it lean further right or left of the other precincts?

I'm trying to collect a list of precincts that the polling was done at to see if there was a bias in how they picked precincts. Thanks.