Exit poll at my NC precinct today

[mnhtnbb]

This was in Chapel Hill, NC--and I've never seen an exit poll rep there at any other election.
Chapel Hill is a bastion of liberalism in the state.


After I voted, I went outside to find a media rep (consortium of media outlets) asking voters
to fill out exit poll surveys. I gladly participated. BUT--you put a folded up completed survey
in a box that could easily be rifled for responses and have surveys removed before being tabulated.
In no way was it secure.

Dem or Repub--handed different survey

Asked on the Dem survey I completed:

voted Obama or Hillary?

Was vote influenced by race or gender?

Demographic data of age, sex, race, family income, education level and religious affiliation/frequency of church attendance of voter requested

Would you be happy if your candidate did not win the primary?

Would you vote for McCain or Hillary in general?

Would you vote Obama or McCain in general?

Did you trust: Obama, Hillary, both, neither?

Who was best qualified for job of president?

Who was best qualified to handle the economy?

What issue is of most concern to you: Iraq War, economy, health care issues

Which candidate (Hillary or Obama) was most likely to beat McCain in general?

When did you make up your mind about who you were voting for: <3 days ago; 1 week ago; and some other
longer time periods

Did the Rev Wright controversy affect your vote?

There may have been a couple of other questions--but I think I've summarized most of them.

Anybody else fill out an exit poll today?


On edit: I'm paraphrasing questions. They didn't call Hillary by just her first name or Obama by just his last name. I don't remember whether they were each referred to as Senator, or just by given first and last names.

[lisa58]

1. That's really interesting - thanks for posting that...

did a lot of voters want to participate?

[mnhtnbb]

6. Two people were filling it out when I came out; one older gentleman also

agreed to fill out while I was completing mine. He didn't seem to want to admit whether he'd voted Dem or Repub, but accepted the Dem survey. I thought that was a little odd.

[AllentownJake]

2. I got exit polled in PA

Similar questions but substitute Rev Wright with Bitter.

According to the Exit Polling Obama won the ward I was in by 2%. He ended up losing by 15% in that ward.

I'll never trust Exit Polling again after seeing that.

[OnTheOtherHand]

3. FWIW, CBS News will probably post the questionnaires

For instance, here is the Ohio Republican form (the link to the Dem form seems to be broken).

At some point after the polls close, the Indiana exit poll link should show up here, and that page should provide a link to the questionnaire (possibly later in the evening). Same idea for NC.

Several exits have asked whether one or both candidates attacked the other unfairly -- did you see that question?

[mnhtnbb]

5. Yes, that question was on there. Thanks for the reminder.

[flpoljunkie]

11. Do you have the link to the Dem exit poll?

I tried substituting Dem or Rep in the address you gave and it did not work.

[OnTheOtherHand]

15. sorry I was offline

As I said, the Dem link from Ohio seems to be broken, although you can pretty well use the tabulation to reconstruct what questions were on it. There's no obvious rhyme or reason to which ones they have properly posted; here is PA.

[flpoljunkie]

18. Thanks for the link!

[mnhtnbb]

4. I found it most disturbing that the surveys weren't secured. Really makes you consider

exit polling to be unreliable. I wonder if that's part of their intention to give people the sense
that exit polls should be taken with a grain of salt.

Anybody know how exit polls have been done in previous elections?

[OnTheOtherHand]

7. I think it's about cost and convenience

As I understand it, the interviewers call in the results three times during the day. Providing them with lockboxes might not be very cost-effective.

I would venture that out of all the things that survey researchers worry might go wrong with exit polls, actual tampering with the forms is not in the top five. It could happen, of course; it just isn't a priority concern. The decision desks are probably more worried that some of the interviewers will make up data on the spot and the desks will get sucked into a bad call -- although I don't know that that is a top-five concern either.

[anigbrowl]

10. It's just a poll, has no legal meaning so no especial need for security

Don't be paranoid - nobody has anything to gain from looking at individual poll results. The situation with exit polls is that they're basically very reliable, but since our election cycle has become so ridiculously poll driven, people fuck with the polls just for the 'amusement' value in distorting the result. It makes them feel powerful to be able to put one over on the media, then they can snicker to themselves about how much smarter they are than those guys on the teevee.

[mnhtnbb]

12. Up until 2004, the exit polls were actually quite reliable in predicting outcomes.

The disparity in the exit polls to the outcome of 2004 is one of the factors that caused people to start
looking seriously at why, all of a sudden, the exit polls were so far off in predicting results and whether that meant, as you say, people were distorting their answers for amusement, or whether there was tampering with the voting machines. Makes a big difference to democracy depending upon which answer you get.

[OnTheOtherHand]

16. the exits were almost as inaccurate in 1992

But no one noticed because it didn't affect the calls, only the margins. The notion that the exits are super-accurate may be based on the fact that networks rarely call races wrong -- but networks usually avoid calling close states until lots of votes are in. (Florida in 2000, the first time, was a spectacular counterexample, but even then the networks were using vote counts, not just exit poll data.)

Of course it's conceivable that there was massive miscount in 1992, too. The problem is -- well, there are lots of problems, but one of them is that the exit poll discrepancies don't correlate with differences from pre-election polls.

[flpoljunkie]

8. Very interesting. Thanks for taking the time to post a report on your exit polling.

[mnhtnbb]

9. Just remembered another question regarding how liberal you consider yourself

with categories ranging from very liberal to not very liberal.

[tledford]

13. No exit polling at my Greensboro precinct

In Northwest Greensboro at my precinct, there was no exit polling in evidence. Greensboro, although hardly as "liberal" as Chapel Hill, is definitely blue -- went for Gore in 2000 by ten or twelve points and Kerry in 2004 by seven or eight.

I wonder what methodology they're using to choose the precincts where they do exit polling? God only knows, but if it is run by one of the TV networks, you can rest assured that it will be bullshit.

[flpoljunkie]

14. Apparently, NEP is doing the exit polling for ABC, AP. CBS, CNN, Fox and NBC.

This is according to the pdf document in an earlier post here.

Wish there was a working link to the NEP Dem Primary exit poll!

[OnTheOtherHand]

17. they might generally interview in 30 to 50 precincts

In NC it was probably on the high side. (I think the full range in the 2004 general was 15 to 55.)

There's nothing wrong (that I can see) with their precinct sampling. They make sure to spread the precincts around the state and get a mix of sizes while still having a true random sample. But since there aren't a lot of precincts, they're somewhat vulnerable to bad luck or bad interviewers.

[TragedyandHope]

19. Thanks, it's interesting to hear about the other side of the polls

after seeing them on the news all night.