Failed exit polls suggest fraud occured in NH, if you ask me

[antiimperialist]

One thing is pre-vote polls getting it totally wrong, but another thing is exit polls, where respondents have no reason to lie about who they voted, getting it this wrong.
There are no undecideds in an exit poll.
There are no last-minute events that could influence people's votes, because they already did so.
There is no "Bradley effect" in an exit poll.
If one fellow DU'er is not lying, CNN and Fox both predicted a 4-5 victory by Obama, yet Clinton won by 3%.
In areas where Diebold was used, Clinton dominated. In hand counted areas, Obama dominated.
Exit poll responses are facts. You voted for who you voted.
Entrance polls and pre-primary polls only tell you about who the voter THOUGHT he/she would be voting for.

To those who are not afraid to be called conspiracy theories: Do not stop asking questions. Do not stop doubting.

[ORDagnabbit]

1. yep keep asking the question

why they seem to only poll correctly in hand counted areas.

polling to hand counts = correct
exit polling to hand counts = correct


polling to machine counts = error
exit polling to machine counts = error

[LSK]

2. where do the exit polls have it wrong?

[antiimperialist]

3. Fox had Obama up by 5%

and a fellow DU'er posted a graph saying that CNN predicted Obama +4. I haven't confirmed the CNN one.

[creeksneakers2]

13. How do you know Fox had Obama up by 5%?

There was only one exit polling outlet and the original results matched the final outcome. I'd love to see a transcript or tape of Fox claiming that but if they did they were lying.

[OnTheOtherHand]

18. whoa there, whoa

Do you have inside access to the exit poll data?

Because you're stating with great confidence what could easily be untrue. And I say this as someone who thinks that the NH count was probably accurate. The NH exit poll, there I wouldna care to give an opinion.

[creeksneakers2]

23. I have three good reasons to know the original data was close

to the actual result.

The first is this link to preliminary results. They had it Obama 39% and Clinton 38%.
http://blogs.usatoday.com/onpolitics/2008/01/exit-poll-obama.html?csp=34

The second is that NBC declared the race too close to call at 8PM, so the results then must have been close.

The third is the male/female breakdown. The vote percentages within each gender were assessed, then the proportion of each gender that voted was tested. The resulting prediction was very close to the final result.

People here keep saying that exit polls change as the night goes on to reflect the final outcome. That's true, but they can only change in regard to total number of votes and where votes come from. No other new information is given from official sources. The official sources don't provide updates on male/female breakdowns.

One cannot update without new information. If the male/female breakdown couldn't have been updated, the proportions given in the exit poll must be the original ones. They matched the results.

I found sources of information given to discredit the exit poll results. One, the 35% number, comes from a single line in a British newspaper that had no direct access to the data. Another false reading came from Fox News, who claimed to have exclusive data. If they did their data wasn't very good. Another claim by the conspiracy theorists is that Chris Matthews said the original results had Obama either way ahead or ahead by five points. Here's BradBlog's quote of what Matthews actually said:

" So what accounts for Hillary Clinton's victory in New Hampshire? What we don't know is why the victory is so much different in fact, then the polling ahead of time, including what we call the Exit Polls were telling us. Obama was ahead in those polls by an average of 8 points, and even our own Exit Polls, taken as people came out of voting, showed him ahead . So what's going on here?


So what Matthews said doesn't contradict an initial result of Obama 39% /Clinton 38% that the USA Today reported.

[OnTheOtherHand]

24. I like your first one best, although

New Hampshire being New Hampshire, I have no way to be sure that that projection wasn't already updated based on vote totals that came in after 7. Wait, let me restate that: I would be startled if it wasn't. The projections should be updating in close to real time; I can't think why a projection at 8 wouldn't incorporate any and all data then available. The tabulations posted at that time might or might not have, but in fact they seem to match, for what that's worth.

There is no way they would call a race based on exit poll data at 8:00 if the exit poll showed, say, a 6-point lead. If they had a bunch of vote counts to support it, maybe.

I'm not sure what you're saying about the male/female breakdown. The percentage of female voters shouldn't change, but the estimates of Clinton and Obama support among men and among women can be weighted any which way.

You're right that the Matthews quote proves nothing. I'm just trying to keep an open mind about what I don't know.

[creeksneakers2]

26. There would be no way to update the estimates of Clinton

and Obama voters for each gender. That information is never officially released. Therefore, the original estimate can't be updated. They have to go with what the poll originally predicted.

You could argue that say, final result predictions showed 40% Obama and the poll had 36% men for Obama and 38% female for Obama that both could be raised. But how much for each? You wouldn't know if female was actually 38% and men 42% or the other way around. Anyway, you wouldn't know from any data that became available after the polls closed.

[OnTheOtherHand]

31. the original estimate can be updated, and is

The infamous "screen shots" from 2004 demonstrate this -- and New Hampshire is a good place to look, since there was a double-digit exit poll discrepancy and a large partial recount that supported the original results.

http://www.exitpollz.org/exitpolls/NHam1224CNNScreen0035.pdf
shows men 52/47 Kerry, women 58/41 Kerry

http://www.cnn.com/ELECTION/2004/pages/results/states/NH/P/00/index.html
shows men 47/52, women 54/45

(the discrepancy was larger than the first tab indicates, because the initial projections incorporate pre-election expectations)

You're of course absolutely right that they have no way of knowing whether the original estimate was "off" for men, women, or both, in what proportions. They don't adjust the tabulations individually; they adjust the weights for each respondent. In this case, that should mean that the estimates for men and for women change about equally (which is true, although given rounding error, we can't tell exactly how true it is).

Insisting that we know the original exit poll was right is the wrong way to go, and will blow up in your face -- if not this time, then next time the exit poll is wrong.

[truedelphi]

25. The procedure with the exit poll change is that after

The votes are tallied up and the final tally is arrived at, then the exit poll data is "re-aligned" with the outcome.

This happened in the wee hours of Nov 3rd 2004 - the exit polls clearly showed that Kerry had won, and yes, as you say, this data SHOULD only be changed in regard to the addditional votes as the votes come in. But then when it was all over, the networks (who are the ones paying for the polls) went in and changed the exit poll data. Even though they should not have.

Many DU'ers kept running screen shots of all this data at the time. There was much speculation about it.

And you are not wrong - these polls and their data SHOULD NOT be changed. But apparently in this corporate controlled state, life in Amerika demands that we have our media manufacture consent and the "appearance" of democracy.

[creeksneakers2]

28. I'm not saying that the exit polls weren't changed

I'm saying I have a link to exit poll information BEFORE it was changed, and other evidence points to that original result. That result was not updated and it came very close to matching the actual results.

I'm also saying that only two kinds of information are released after the polls close. That is vote totals and geographic information about where the votes came from. So the final result prediction could change from either matching trends or reweighting estimates based on percent of the outcome expected from each geographical location. Only the final projected total vote changes.

There is no additional data given for subcategories so those subcategories could not be changed. If they report that 51% of voters who make over $25,000 a year voted Clinton then they have to keep that estimation. There is nothing to refute it later. If the total vote counts are higher they won't know if they were wrong on voters over $25,000 or under $25,000. So they have to leave those numbers as they were.

The subcategories must be the original reports and they predict an outcome very like the actual outcome from counting votes.

[truedelphi]

30. That's good information to know. Thanks for posting it. n/t

[terisan]

4. Are you going to act on your belief?

If you start by contacting the person in charge of the election you can follow their instruction on how to contest. Also the other candidates may want to contest.

[MethuenProgressive]

5. The pre-vote polls got BO's numbers right. They said mid-30s, he got 36%.

All the rest is whine.

[hugo_from_TN]

6. People can lie to exit pollsters

I've done it myself after a general election.

[tridim]

15. Why did you lie?

And please be specific. Did someone ask you to lie?

[hugo_from_TN]

16. Because it wasn't their business who I voted for....

An overly eager (and young) pollster wouldn't accept when I declined to be polled because he really wanted to get a 'good sample of voters'. Being generally against polling on principle, I decided to go ahead and fill out the form and just make up all the responses. I guess I'm just subversive at times.

I think misleading of pollsters is going to be high this election cycle due to the emotions involved and the racial and gender issues that are involved in the candidates.

[Joe Chi Minh]

17. Most people would want to tout their candidate, all the more so on

such an occasion, so your thesis doesn't stand up in my opinion.at.all.

[OnTheOtherHand]

20. or avoidance

The circumstantial evidence is pretty good that some Bush voters in 2004 just kinda steered around the earnest young interviewers.

[jpljr77]

7. Or, the numbers suggest failed methodology on the part of the pollsters.

I have no doubt that the spectre of true vote fraud has reared its head in the U.S. election process with the introduction of the Diebold voting machine. And I think that there may have been shenanigans in each of the past 3 general elections.

But we also have to look at why polling methodology has failed so miserably starting with the 2004 election. I think the formulas the pollsters use reflect an older reality. Now that populations have become more nimble and transient, it's harder to pin down the calculus required to allow for variables like registered party and demographic voting preference.

In short, I think exit polling, and pre-polling, is at its worst stage ever. Until they figure out how to effectively crunch the numbers, the results will be off.

[Bonn1997]

19. Please explain

"Or, the numbers suggest failed methodology on the part of the pollsters."

It would have to be failed methodology on the part of pollsters *only in areas with diebold voting machines*. That sounds far-fetched!

[OnTheOtherHand]

21. no, not at all

We don't have any information on how accurate or inaccurate the exit polls were in various towns. There's no way to say that the exit polls were inaccurate "only in areas with diebold voting machines" (i.e., scanners).

I do know that if one controls for Kerry's performance in the 2004 primary, the 2008 NH op-scan results don't look systematically different from the hand-count results.

[Bonn1997]

22. You're right (oops)

You're right. I made a dumb inference from other statistics I saw.

[foo_bar]

8. "Exit poll responses are facts. You voted for who you voted."

see: http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=389x2648205

There is no "Bradley effect" in an exit poll.

Source?

This reluctance to give accurate polling answers has sometimes extended to post-election exit polls as well.
<...>
The original term Bradley effect derives its name from a 1982 campaign involving Tom Bradley, the long-time mayor of Los Angeles, California. Bradley, who was black, ran as the Democratic party's candidate for Governor of California against Republican candidate George Deukmejian, who was white. The polls leading into the day of the election consistently showed Bradley with a lead.<4> In fact, based on exit polls on election day, a number of media outlets projected a Bradley win that night; early editions of the next day's San Francisco Chronicle even had a bold "BRADLEY WIN PROJECTED" headline. However, Bradley narrowly lost the race. Post-election research indicated that a smaller percentage of white voters actually voted for Bradley than that which had said they planned to vote for him, and that voters who had been classified as "undecided" had gone to Deukmejian in statistically anomalous numbers.<2>

http://en.wikipedia.org/wiki/Bradley_effect

[truedelphi]

27. Thanks Foo bar.

And I'll never forget a leson learned: if there is a discrepancy and it occurs in the Ukaraine, our government will go to tremendous lengths to ensure that the truly winning candidate takes office.

But it has to happen in the friggin' Ukraine.

[Muddy Waters Guitar]

29. Exactly

Exit polls aren't predictive, they're post-hoc and measure actions, not predictions of actions. They're essentially raw data.

[Jane Austin]

9. Didn't MSNBC's exit polling have him up 8%? n/t

[creeksneakers2]

14. No

The original results matched the final count.

[laconicsax]

10. I seem to remember

that the media outlets didn't want to project a winner on the Dem side based on the exit-polls because the numbers were too close. I've been looking for the post that claimed this, but I haven't found it yet.

[cloudythescribbler]

11. GREAT POST -- how long till it gets banished to "election reform"

Obviously, you have drunk the koolaid of the 'tin foil' hats folk.

No room for healthy skepticism (as one DUer posting had it) on this question.

VERBOTEN!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

I want to use this post as an example of why banishing this theme of discussion should be opposed by DUers

[EFerrari]

12. The ER is a great little forum. I was hoping this would be done

so all those screaming people could keep talking about the next thing and we could keep trying to figure this out in peace.