Blackwell Tried to Invalidate OHIO 2004 Exit Polls with >100 Feet Rule

[McCamy Taylor]

Remember on election day 2004 how we woke up to find that Ohio SOS Blackwell had tried to keep exit pollsters far, far away from polling places, but that the news networks had gone to court? At that time I knew he was up to something. Well, the gun is smoking even more. The story about his unsucessful attempt to interfere with the exit polls that would turn out to the smoking gun of that election has all but been scrubbed from the internet. (Google the key words and then check the links if you dont belive me). Yahoo and all the regular news sites have taken the story down, but some "little" news sites still have it up, so here is one:

http://www.unknownnews.org/041109comvot.html

"Five television networks -- ABC, CBS, CNN, Fox News and NBC -- and the Associated Press filed the lawsuit in U.S. District Court in Cincinnati, seeking additional access at the polls Tuesday. The organizations have formed a consortium to collect exit-polling data in Ohio and other states.

"The networks and the AP said exit-poll reporters were allowed to conduct interviews within 100 feet of polls as recently as the March primary.

"A Blackwell spokesman said Monday that in February, media interest in the Ohio election was much less. Anticipating a crush of voters and media on Tuesday, Blackwell told boards to enforce a state law that bars anyone besides voters, election officials, challengers and witnesses from inside a 100-foot limit."

Why on earth would Blackwell get it into his busy, scheming head to keep exit pollsters more than 100 feet away from polling places? Could it be that he was afraid that exit polls would not match the vote tallies? Could it be that someone familiar with exit polls told him that the more distance there is between the exit pollsters and the polling place the more unreliable the result of the poll is. And the 100 feet is a critical number?

Check out this document which was put together to refute Mitoksy's Reluctant Responder Theory. I was looking up the part about the theory being inoperative in precincts with hand counted paper ballots (which to me means it is inoperative where GOPers cant cheat) when I noticed that if the pollsters are greater than 100 feet from the polls, the exit poll results are thrown into chaos(!!!!). (go to pp 17-18)

http://electionarchive.org/ucvAnalysis/US/Exit_Polls_2004_Edison-Mitofsky.pdf

This is the kind of thing that someone who has made of study of exit polls would probably know all about. And it raised a possibilty. One of the difficult things to explain about the Reluctant Responder Theory is why Mitofsky would put his name on such a piece of career ruining crap if he really didnt believe it, since his rep. is now shot to hell for all time. However, what if he first got involved because someone asked him "Is there a way to invalidate exit polls?" and he said "Well, yes there is. If the pollsters are kept more than 100 feet away from the polling place, then the results become less reliable. And discrepancy could be attributed to that." "And would you be willing to testify to that?" "Certainly."

Only, the courts would not let Blackwell get away with it, the exit polls were done by the book, and suddenly Mitofsky was called upon to do something he had not bargained for, and he was not given a chance to decline.

Does anyone know enough about exit polls to know if the effect of distance had been studied before this election and if so, by whom? Could Blackwell have asked for the 100feet deliberately in order to invalidate the results?

[bleever]

1. Very interesting. K&R, and thanks.

[Febble]

18. Self-delete

sorry replied to wrong post

[robinlynne]

2. Well, there is the simple and obvious. In order to do an exit poll,

you have to ask every person who comes out of the voting booths to participate. that way only people who don't want to participate won't. In my experience a huge majority will answer. If you are further away, you will get a smaller sample, which will be less accurate.

[OnTheOtherHand]

3. actually not every person (usually)

I'm referring to the national exit polls; I do realize that many exit polls operate exactly as you describe. But in the national exit polls, there is an interviewing rate for each polling place: the interviewers are directed to approach every Nth voter, where N could be anything from 1 to 10. That's more complicated for the interviewer. And the further away the interviewer is from the entrance, the more leeway voters have to steer toward or away from the interviewer, which tends to mess up the sampling. Not only do you end up with fewer completed interviews, but there is more possibility of bias. There's nothing magical about 100 feet. A 1995 study of the 1992 exit polls showed larger errors for interviewers 10-50 feet from the polls than interviewers inside or right at the door (over 50 feet was even worse).

Contrary to the OP, this isn't some secret that only Mitofsky would know -- and if it were, how on earth does Kathy Dopp know it? And if the exit pollsters were actually scheming to be over 100 feet away in Ohio, why didn't they manage not to win their lawsuit? (And the premise that Mitofsky's reputation is ruined because he believes in survey bias doesn't hold up, sorry. That might work if DU financed the exit polls, but last I checked, not.) But as for Blackwell's motives, I don't know. I never even looked to see whether he offered a rationale. It's certainly not the first effort to interfere with an exit poll.

[Tigress DEM]

4. At polls there is a general rule of 100 feet to keep out those who....

would try to influence voters going in to vote. If pollsters were voluntarily remaining at 100 feet in order to not appear to be in conflict with the general rule, but were being pushed out even farther...

But the rule shouldn't apply to pollsters since they are asking for results only, and probably have people answer a short questionaire. Without a verbal announcement it would not be invasive so there is no real reason to keep these people 100 feet away from the voters in line to vote...

However, for those news channels that are talking about poll results, they might discourage people from voting if they think there is too big of a gap for their vote to matter, so those types of media persons might have had problems.

[Febble]

14. Greater redshift

kicks in at 25 feet. Most interviewers in most states were within 100 feet.

[OnTheOtherHand]

16. some more background

I don't have the whole story, but Ohio had previously allowed interviewers within 100 feet; just a few days before the election, Blackwell instructed county BoEs not to do that any longer. It doesn't even seem that the networks ever got their hands on an explicit statement from Blackwell, which might or might not offer a rationale.

Here's something Blackwell wrote in April 2005, apparently striving for maximum feasible confusion:
http://www.sos.state.oh.us/sos/electionsvoter/directives/2005/mainDocs/Dir2005-09.pdf

Ohio law stipulates that people must not "loiter or congregate" within 100 feet.

Folks can find a bunch of the filings at http://moritzlaw.osu.edu/electionlaw/litigation/st-ohio.php#abc

[McCamy Taylor]

8. To clarify: Blackwell wanted invalid polls. The networks didnt want

to waste their money and get crap results. After all, they need to be able to study their methods so they can improve them for later elections. So they were gonna compromise. Do valid polls but promise to sit on the results. Remember how none of the news networks would release their poll numbers? However, a disgruntled employee at one of the news networks saw what was happening and released the poll results to the world over the internet (which everyone should have seen coming). This meant that there were accurate, valid exit poll numbers out there, something that Blackwell and Rove had not planned for when they did their math. It was this discrepancy that would later lead Powell to his passive aggressive act of declaring the election in the Ukraine suspect based on exit polls--a big slap in the face of Bush, and potentially politically deadly had not the entire MSM of America been on Bush's side at that time. Because of this exit poll discrepancy, by December 2004, 25% of Americans believed that election 2004 had been stolen. We even got a (one!) senator to contest the election.

This was why they needed Mitofsky's ridiculous Reluctant Responders Theory.

If Blackwell had been able to get those exit pollsters to stand behind the line at 100 feet, when their results did not match the vote tally, a whole bunch of experts would have said "The exit polls are wrong. They stood too far away. End of story."

And we would not be talking about Election 2004. This Topic Forum would not exist.

[OnTheOtherHand]

10. but that makes no sense

The reality is, a whole bunch of experts do say that the exit polls were wrong. And yet, this topic forum exists.

There's another level of weirdness here: if you accept the premise that interviewer distance actually can interfere with the accuracy of the exit poll, then why are you so sure that the exit poll was accurate?

Do you have documentary evidence of Colin Powell "declaring the election in the Ukraine suspect based on exit polls"?

[OnTheOtherHand]

21. whoops, just spotted another myth

Remember how none of the news networks would release their poll numbers? However, a disgruntled employee at one of the news networks saw what was happening and released the poll results to the world over the internet (which everyone should have seen coming).

What on earth?

CNN.com announced in advance that they would post the preliminary exit poll tabulations once the polls closed in each state. And they did. It had nothing to do with "a disgruntled employee at one of the news networks."

I guess the kernel of truth here is that the tabulations don't incorporate actual projections, which is why there was a small cottage industry of estimating vote shares based on % male / % female and so forth. But the premise that Blackwell and Rove were somehow shocked that the tabulations were posted doesn't make sense.

Go back and look at the DU threads. People expressed shock when the tabulations were updated -- but I haven't noticed any who expressed shock when the tabulations were initially posted. 'Oh my God, how did the networks let this get out?' Didn't happen, did it? (I was reading the exit poll screens that night, not DU.)

[Febble]

22. This is almost totally

fiction (make a nice movie though).

There are certainly accurate, valid exit poll numbers out there, I saw them on CNN myself (so they did release the numbers), and E-M published the details last January (so they did too). So Bush and Rove forgot to tell CNN, it seems (I don't remember any other networks - I can't get Fox, even if I wanted to).

And I think the Ukraine elections were considered suspicious because people were seen pouring acid into the ballot boxes, and some other stuff - oh yes, one of the candidate was poisoned, and nearly died, and when he recovered, his face looked horrible on TV. So I guess that would have raised suspicions (I think there were two exit polls, and they didn't even agree with each other, never mind with the result, so I expect the poison and the acid might have been the clincher).

And then of course, Mitofsky's Reluctant Responder wasn't actually ridiculous, but a well-known, experimentally confirmed phenomenon, and supported by the actual data.

But sure, if you were going to steal an election, messing with the exit polls would be quite a smart move. That can be the McGuffin, maybe.

[Awsi Dooger]

23. Why didn't Blackwell fix the exit polls to favor Bush?

That's another thing I've never understood since late '04. Rove and the Republicans are given full credit for planting the national guard documents to get rid of Dan Rather and nullify that story. They rig machines nationwide, seemingly knowing ahead of time exactly where Kerry will finish in the vote, and therefore how much they need to turn the dial.

But they have no clue how to get their own guys hired by E-M, so the exit polls are neutral at worst, or even favoring Bush. Remarkable.

They didn't even need the damn jobs. Just make sure your voters, even planted ones, swarm the exit pollsters en masse and tell them they never considered a Republican before but voting for Bush was the highlight of their life. 100 feet or 5 miles, who cares? Tilt 'em, just like "DU this Poll."

This thread is one of the first to declare exit polls were considered beforehand. Freeman and others seem to believe the GOP ignored the exit polls while pulling off the heist. Exit polls are the gotcha.

But wouldn't Bush be better served if the exit polls asserted he won by greater margin than the actual vote indicated? Talk about nullifying the issue. That would have reversed it. There would be election forums on the conservative sites, not here. Plus, it would go along with 2000, where the early exit polls in Iowa and Wisconsin indicated Bush led Gore. Pennsylvania was a tie. Yet Gore won all three states. Conservatives could claim the exit polls in 2000 and 2004 showed votes were being thieved from Bush, and no doubt downballot Republicans as well.

Maybe I'm not thinking clearly. The deviation from actual numbers to guesswork earns a daily headache, and not a tiny one . Maybe Republicans are too proud to fix exit polls in their favor. Everyone has a border they won't cross.

[bleever]

9. Will Mitofsky continue in exit polling?

Not that I expect you to necessarily have personal knowledge about his plans, but my guess is that the exit poll paradigm (as we knew it) has been crashed, and that Mr. Mitofsky could not continue in this line of work.

[OnTheOtherHand]

11. Mitofsky has continued in exit polling

Edison/Mitofsky polled in Azerbaijan in November 2005; there's a pretty interesting article about it at http://www.publicopinionpros.com/features/2005/nov/mitofsky.asp . The National Election Poll has stated on the record that they are conducting exit polls this year; I assume that E/M is conducting them. (I have to admit that I haven't paid much attention.)

The exit poll paradigm as I know it actually hasn't been crashed. Mitofsky warned the NEP sponsors on Election Day afternoon that the election was too close to call, and E/M didn't make any call contrary to the official results, nor did any network. When Mitofsky started exit polling for CBS, it was CBS policy not to use the exit poll interview results in the projections at all. At some point they decided that was too pure, but the pollsters and networks are still very cautious about relying upon the interview results in close state races. (Both Florida calls in 2000 incorporated vote counts.)

[Tigress DEM]

5. Like he scrubbed his pic at the Council for National Policy?

I have it on my desk top and they say that is a group that is online with that whole 30 year agenda put out by the reich wing to dump America in the toilet.

It's supposed to be some sort of really secret group that doesn't want the public to know who it's members are.

[Awsi Dooger]

6. I've read that increased distance led to higher WPE

Mystery Pollster had a list, from the Edison/Mitofsky report, indicating WPE toward Kerry was greatest in precincts where these factors were present:

*An interviewer age 35 or lower
*An interviewer with a graduate degree
*A larger number of voters, where a smaller proportion were selected
*An interviewer with less experience
*An interviewer who had been hired a week or less prior to the election
*An interviewer who said they had been trained "somewhat or not very well."
*In cities and suburbs
*In swing states
*Where Bush ran stronger
*Interviewers had to stand far from the exits
*Interviewers could not approach every voter
*Polling place officials were not cooperative
*Voters were not cooperative
*Poll-watchers or lawyers interfered with interviewing
*Weather affected interviewing

[Febble]

13. Yes

Beyond 25 feet the redshift was greater. It just didn't happen to be a particular problem in Ohio, though other things were.

[Tigress DEM]

7. And the way the signing statements have been scrubbed from public view?

Does anyone have all 720 of those somewhere?

[Febble]

12. Well, whatever he tried to do

he didn't have much luck.

Most of the interviewers in Ohio were either inside the polling place or right outside.

[OnTheOtherHand]

17. but on the other hand (TM)

Our suit was filed the day before the election. We were successful in overturning the Secretary of State’s ruling, but the court ruling did not occur until 10:30 PM on the night before the election. Although we were able to contact all of our interviewers before the polls opened, many Ohio election officials at our polling places did not know of this ruling when the polls opened and many of our interviewers in Ohio were delayed in starting their interviews until the local election official was informed of this ruling. The last local election official did not permit our interviewer to begin work until close to 5 PM.

- E/M evaluation report at p. 70

So I would say that he had some luck.

[Febble]

19. Good catch.

I hate discrete variables.

Your point remains good though - if Blackwell buggered up the exit polls in order to hide fraud, they remained buggered up, so we can't tell whether he did it to hide fraud or not.

[Febble]

15. Oh, and regarding that paper

you linked to:

The argument it makes doesn't work for a fairly large number of reasons. If you want to know more I can take you through them, but it was hashed out fairly thoroughly a year or so ago, and at least some of the authors revised their views.

The "reluctant responder" theory is not actually "crap", it is well-attested, there is actual experimental evidence for it, there are vast books written on the subject, and there is excellent evidence that was a major contributory factor to the exit poll discrepancy in 2004. I recently read an analysis of the exit poll for 1996, where it was also a major contributory factor.

There are two myths that need to be debunked here:

1. US exit polls are not particularly accurate - in 1992 the discrepancy was almost as great as in 2004.
   
2. Non-response bias and selection bias in surveys are not inventions - they are extremely well researched phenomena, and, specifically, Democratic bias in exit polls has been demonstrated experimentally.

[Febble]

20. Taking this in bits

as I wait for a program to debug....

OK, paper ballots. We have an apples and oranges problem here.

The reason the argument doesn't work is that paper ballots were only used in rural and suburban areas and small towns. In these areas, the overall discrepancy was a lot less anyway, and in fact, when you analyse these precincts together, there is no difference between the average discrepancy in paper ballot precincts and the average discrepancy in precincts using any other voting technology (also, remember an average of zero doesn't mean zero - it just means the discrepancies cancelled out). When you look at precincts serving urban areas (pop>50,000) the discrepancy is greater overall, and in fact, it turns out that the precincts with the biggest discrepancies used older technology (levers, punchcards), whereas precincts using digital technology (DREs, optical scanners) had smaller discrepancies.

So if you compare apples with apples. i.e. precincts with paper with similar precincts using other methods, paper is no better or worse than the other methods.