DID ANYONE EVER QUESTION EXIT POLL ACCURACY PRIOR TO SELECTION 2000?

[TruthIsAll]

Is there any documented evidence of exit poll error prior to 2000?

Why are exit pollsters even used at all if they are so inept?

Can anyone explain why it is that only in the U.S. does the media make the claim that the exit polls were wrong?

Can anyone explain why other nations rely on exit polls to protect against fraud and we seem to rely on the vote counters to prove that exit polls are wrong? Doesn't that seem odd?

Can anyone explain why, even after the selection 2000 fiasco in Florida, the media would never even consider the possibility of programmed election fraud, but instead would fault the exit polls?

[jwmealy]

1. Trenchant questions, which urgently need answers.

I have one for you. Can you either put me in touch with Jonathan Simon, or help me get my hands on the actual screenshots he used to collect his exit poll data (published here on DU)?

I am preparing to go public with my analysis of the national pattern in redshift (www.selftest.net/redshift.htm), and the weak link is that I am totally reliant, for the exit poll information, upon althecat's published tables of Jonathan Simon's information here on DU. My analyses are being gone over by a professional statistician (PhD level) today and tomorrow, and if my consultant gives me the green light I plan to publish.

Is there any exit poll information that is now being made publicly available by the pollsters themselves? Last I heard, they did not plan to release their raw data until the Spring.

[TruthIsAll]

5. Sorry, can't help you on that.

I work independently. If I see data posted, I use it.

Creating a workbook for publishing is a good idea.
I might create one, even if it's just to organize my data and charts. Everything analytical that I have posted is Excel-based.

[davidwparker]

2. It's only questionable when the Bush Crime Family is involved.

If the fascists can get people to doubt exit polls, then they will go away. Then, all the fraud going on with black box voting can't be detected.

http://blackboxvoting.org

Dean!!!!!!!! 2008.

[RaulVB]

3. I only know of one case before the Ukraine fraud and the American one

Chile in 1989, right after Pinochet was outed.

The government's TV network, by then controlled for Pinochet's opposition, asked for exit polls (first time for Chile). They "showed" that Pinochet's followers did NOT WIN A SINGLE SEAT IN THE NEWLY ELECTED CONGRESS.

Turned out that the TV station lied about the numbers and made them public before the election was over that day so the candidates of the Chilean right wing party were to be damaged by the "news", while the people was still voting!

It was their "payback" after 17 years of suffering under Pinochet's rule.

[bigmustelid2]

4. Yes, Exit polls have often been wrong

http://www.washingtonmonthly.com/archives/individual/2004_11/005178.php

1988: Dukakis vs. Bush - exit poll discrepancy = 8.3% in favor of Democrats

1992: Clinton vs. Bush - exit poll discrepancy = 7.2% in favor of Democrats

1996: Clinton vs. Dole - exit poll discrepancy = 6.2% in favor of Democrats

2000: Gore vs. Bush - exit poll discrepancy = 1.8% in favor of Democrats

[TruthIsAll]

6. Just proves the Repukes have been rigging the votes for years.

Are they early or final exit polls? I cannot believe they were that far off, unless the votes were rigged.

Having said that, these numbers are not state exit polls. Isn't that what we really want to see?

Until I see the states, I will not trust these numbers.

[Yupster]

55. In 1988

the networks called Pennsylvania for Dukakis based on the exit polls. Bush actually won by 100,000 + votes.

The last time Senator Bob Smith of New Hampshire won reelection (1992? 1996?)all three networks called him a loser within a minute of the polls closing based on the exit polls. He won a close race.

These are just from my memory.

There wasn't any internet back then and the networks were not real keen on publicizing their mistakes.

[RaulVB]

7. The article also says something that you don't mention

The cases of 2000 and 2004 are troubling. Well, WE DON'T KNOW THE RESULTS OF FLORIDA in 2000. You may remember that.

Then, the NATIONWIDE exit polls and the STATEWIDE exit polls are two separate issues.

The author of the article refuses to consider the possibility of fraud or manipulation in vote counting.

The element of RAW AND UNWEIGHTED DATA plus methodology used must be taken into consideration.

The people doing the polling refuses to say how they worked it out on November 2nd.

[TruthIsAll]

8. Agreed, Raul. The fact that fraud is not even mentioned is odd.

In any case, we need the detailed state exit polls.

There is no way in hell that legitimate exit polls would be off by more than 2% nationwide.

Look at exit polls overseas. In the last three German elections, they were correct to within 1% of the vote tallies.

Are they better at it than we are? I doubt it.

[RaulVB]

9. The Germans are honest

That's the difference.

For what I know, exit polls may be off 1 or 2 points, as you said, but they will NEVER state the opposite of the actual vote tally, that is giving the largest percentages to the LOSER of the election.

That is just nonsense.

[bigmustelid2]

11. Ah but they did

"For what I know, exit polls may be off 1 or 2 points, as you said, but they will NEVER state the opposite of the actual vote tally, that is giving the largest percentages to the LOSER of the election."

In 1988 according to this avowedly Democratic source:

http://www.washingtonmonthly.com/archives/individual/2004_11/005178.php

the exit polls predicted Dukakis win by 50.3% over Bush's 49.7%. The exit polls were wrong - Bush won by 7.7%.

Never say "NEVER".

[RaulVB]

12. Did I write Germans?

Did you read that I was talking about the Germans, mainly?

Dukakis was the big media "darling" back then. I hear some people saying that they wrote what they wanted to be true and then Dukakis got killed by the Republican's "dirty" job.

"I am going to make Willie Horton the running mate of that SOB..."

* Statement by Lee Atwater, "operative" working for Bush senior (Atwater is dead now, I thank God)

[TruthIsAll]

18. That is pure FICTION .Bush was never in doubt of losing.

He won with 54-46%.

Dukakis led in the early July polls right after the convention for about two weeks - until the GOP convention. Bush was ahead in the polls all the way to the election.

I never saw that exit poll mentioned. I followed the election very closely. Why did we not hear of it at the time?

If the exit poll had Dukakis winning, then for a Red Shift 8.3% move to Bush to happen, he had to steal it.

That's not such a surprise, come to think of it.

[RaulVB]

19. Thanks, TIA

I knew that the polls were made earlier that year and, yes, Bush's victory was not in doubt.

[Yupster]

57. The Saturday night before the election

Al Franken was on Saturday Night Live and drew a map of the USA on an easel and drew in each state's electoral votes to show that Dukakis could win. I think he was hoping more than explaining.

Anyway, I thought it was a weird skit for a comedy show.

It was before I knew how partisan he was.

[bigmustelid2]

21. Please address your accusations of lying

to the source - it is an avowedly Democratic source, by the way:

http://www.washingtonmonthly.com/archives/individual/2004_11/005178.php

[RaulVB]

22. And then YOU address your missing response

You have a good handling of numbers. Explain your thesis.

[Name removed]

25. Deleted message

Message removed by moderator. Click here to review the message board rules.

[tritsofme]

31. Doesn't this just prove the flaws of unweighted raw exit poll data?

There were no public opinion polls in 1988 that didn't show Bush with a comfortable lead over Dukakis, and yet you would seem to argue that the raw exit poll data poll would be more reliable in this case.

[Name removed]

33. Deleted message

Message removed by moderator. Click here to review the message board rules.

[tritsofme]

34. Believe me I know...

some people on this forum seem to swear by the unweighted exit poll

[RaulVB]

36. Well you two guys...

Happy you found each other!

[intelle]

48. TIA....agreed.

As I mentioned in a previous post, the news did not call the election for Dukakis at any time based on exit poll data.

[aquart]

30. You mean the exit polls failed in ANOTHER Bush election?

Are there any known instances of exit poll failure in an election that does NOT feature a Bush?

[bigmustelid2]

32. Yes - Clinton vs. Dole

1996: Clinton vs. Dole - exit poll discrepancy = 6.2% in favor of Democrats

[RaulVB]

39. Just for the sake of intellectual honesty

In 1996, Clinton was given a margin of 12-15 percentage points of advantage over Dole BY THE MEDIA during the entire campaign.

John Zogby entered the polling business on that year. He did a great job by predicting that the final margin would be 8% difference Clinton winning over Dole.

The final result was Clinton's victory by 8.1%

[Name removed]

40. Deleted message

Message removed by moderator. Click here to review the message board rules.

[bigmustelid2]

10. If anything, the national exit polls should

be much more precise than the state exit polls - because of much higher sample size.

You are the math wiz. Please calculate what probability there is that in 1988 the national exit polls would be wrong by whopping 8.3% - in favor of Dukakis. And if it comes out to be 1 in XX billion chance - do you think Bush win over Dukakis was fraudulent as well?

[RaulVB]

14. Different methodology back then

And I am going to suspend my conversation with you until you break down the Republican vote for me in 2004 and you show me how BUSH gained 9 millions of new voters respect to 2000 (don't include the Florida 2000 uncertified results in the package), after LOSING SUPPORT IN EVERY SINGLE DEMOGRAPHIC GROUP except for "rural christians."

I would really appreciate your help with this.

I'll see you then.

[bigmustelid2]

15. Um - Raul

you are aware, of course, that US has secret ballot voting. This means that no one can tell you how anyone has voted. The only thing that can be done is to poll people and *ask* them how they voted. Just like exit polls (as I showed in that article I cited) this methodology is only as accurate as people are honest.

I will give you one example: let's say, for argument's sake, there is a Democrat who, for whatever reason (let's say momentary weakness or fear), voted for Bush in 2004. He gets called or visited by a pollster and asked who he voted for. He tells them Kerry. Why? Who knows. Maybe his wife is nearby and he doesn't want her to know. Maybe he is afraid his response will get recorded somewhere under his name and will hurt him somehow.

Enough such people (and how would you know that there weren't?) and you get a poll that shows landslide for Kerry while the "real" results were for Bush.

The only legitimate "poll" is the ballot. Anything else and you run into weighting problems, sampling problems, people lying, question skewing, etc. etc. etc.

[bobbieinok]

16. exit polls used to question Ukraine's election...correct in 2000 except

for FL

[RaulVB]

17. Not responding the question

Is a simple, direct question.

You seem to have a pretty clear picture on your mind. Please, communicate that to us on this forum.

Is JUST ABOUT NUMBERS THAT DON'T MATCH ACTUAL VOTING PATTERNS in the last 24 years of American electoral history.

Your "bending over" to portray millions of people "afraid" of responding to how they voted doesn't hold any water here in the US.

I know something about it, try again.

[Name removed]

20. Deleted message

Message removed by moderator. Click here to review the message board rules.

[RaulVB]

24. So your "thesis" is:

EXIT POLLS ARE NOT A VALID TOOL.

POLITICAL SCIENCE IS NOT A DISCIPLINE THAT SHOULD BE PART OF THE ACADEMIC WORLD.

NOBODY CAN SAY HOW PEOPLE VOTES AROUND THE WORLD.

WE JUST CAN GUESS, HOPE AND PRAY.

OK, now I seem to "get it"

Nice talking to you, good night!

[Name removed]

26. Deleted message

Message removed by moderator. Click here to review the message board rules.

[TruthIsAll]

29. Still spinning those worn out excuses? There's no basis for them.

I give up.

[reality_bites]

61. Sorry, but you're wrong about this

It is extremely easy to find statistical studies on the web supporting the phenomenon whereby, one side or the other, depending on the election, will make false claims about who they voted for, or not respond at all. I'll even give you a link to one of the studies right here:

http://www.cpsa-acsp.ca/papers-2004/Docherty%20and%20others.pdf

[aquart]

37. How are these exit polls conducted?

The one I worked on for the Asian American Legal Defense and Eduction Fund gave voters a form to fill out. Their names were not on the form, and it went into a box after we scanned it for one question (Did you have a problem voting?)

If the polls are anonymous, and I have never seen a name connected with one, your exit poll theory falls apart as it deserves to.

[bigmustelid2]

38. I was talking about telephone after-election polls -

but even with an election exit poll, if asked, I'd tell the pollster that it was none of his business. I bet they get a lot of that.

[RaulVB]

41. And they will put you down as part of that group

"None of your business" responses are an element of their work.

You are too funny!

I really took you seriously earlier.

[bigmustelid2]

42. I don't think so -

If you look at the exit poll data, they do not tell you how many people told them to take a hike and refused to answer.

[RaulVB]

43. So you're saying...

Exit polls are good but...

Exit polls are not good but...

They are good and bad but...

They are bad and good but...

N/A

None of your business...

[bigmustelid2]

44. I showed you that exit polls

are historically not very accurate. Thus, basing your "fraud" argument on exit polls will not carry much weight with anyone that bothers to check past exit polls' data.

[PennyMan]

52. Interesting Site On Exit Polls

Here Is A Very Interesting Site About Exit Polls I Found To Be Helpful. http://ustogether.org/election04/supporting_doc/Freeman__2_Expldiscrpv00oPt1.pdf

[Yupster]

59. Another example

An exit pollster goes up to an angry white old guy who's a Rush Limbaugh listener and a Bush supporter. The exit pollster asks him who he voted for and he curses him out.

The exit pollster knows he needs a white old guy so the next white old guy is a Kerry supporter who answers honestly and courteously that he voted for Kerry.

The pollster has his correct demographics, but the poll isn't accurate any more.

[TruthIsAll]

27. Depending on the Exit Poll MOE....

It's a 4.15% deviation from the exit poll.

If we assume a 2% MOE, the probability P is:
P= NORMDIST(0.4985,0.54,0.01,TRUE)
P= 0.0000166334
or 1 out of 60,000

If we assume a 1% MOE, the probability becomes ZERO.
National exit polls in Germany are accurate to within 1%.
Take your pick.

[bigmustelid2]

28. you skipped my question

do you think Bush win over Dukakis was fraudulent as well - based on those exit poll numbers and on your probability calculations?

(If you're wondering about MOE - I would presume, because of the huge sample size, that national exit polls' MOE has to be around 1%)

[TruthIsAll]

35. What do you care what I think? Draw you own conclusions.

What I believe should not be the issue here.

Now YOU just focus on the mathematics, use some common sense, and combine that with the accuracy of exit polls - and the Bush family history.

And draw your own conclusions.

.

[Yupster]

56. How do exit polls account for early voting?

Does anyone know?

In Texas 1/3 of all the votes were cast before election day. Same in Florida I think.

If that's not taken into account, the polls would of course be crap. It must be taken into account somewhere I assume.

[reality_bites]

62. Absentee voting is factored into...

.. ACTUAL the exit poll result, but is not a factored in the RAW exit poll numbers. This is one of the reasons that RAW exit poll numbers can't be used to make any intelligent statments about anything, let alone election fraud.

[jamboi]

13. I've heard that they weren't ??ed before 2000.I think I heard Zogby

say that. I'd want to take a look at the numbers used to derive the above stat before going with it. I'd believe Zogby before this other fellow.

[thanatonautos]

45. Hmm. Those numbers look impressive enough but ...

one should look at the actual source. See page 5.

The author is clearly skeptical of the idea that discrepancies
with raw exit poll numbers, if those are what were
published, prove fraud. This is fair enough, as far as it goes.

But he leaves more or less open the question:

Exactly how raw were the numbers that were published?

On this question he has no more of a clue than any of us,
including you.

In otherwords, your own premise is not satisfied by the
numbers you post. Those numbers do not prove
that exit polls have often been wrong in the past.

Further, what is not in dispute, so far as I know, is that
Florida in 2000 was originally called for Gore, by
professionals at the various networks, in part on
the basis of exit polls, some taken specifically
in Palm Beach County, which showed sufficient numbers
of voters left the polls believing they had voted
for Gore.

We know what happened in Palm Beach County, and we
know that the Florida exit polls in 2000 were, in fact,
basically correct.

More Palm Beach voters, and more Florida voters did cast
their votes for Gore ... it was the butterfly ballot in
Palm Beach and ballot spoilage elsewhere in the state that
produced a false Bush victory.

So the exit poll controversy ... while it certainly does
not prove fraud, certainly is absolutely worthy of full
investigation.

[sharman]

49. FL 2000 exit poll

I'm pretty sure I read that the networks called FL for Gore so early because there was a palpable--but obvious and quickly corrected--error in data input, that had Gore winning in a blowout.

Therefore, correcting for that error (which was in fact quickly discovered and corrected), the 2000 FL exit poll predictions were accurate: the state was very close, nor a clear Gore victory that could be called as soon as the polls closed.

IMO, intended votes gave Gore about a 55,000 victory, but those votes were lost to an unprecedented amount of ballot spoilage. But 55,000 is still a very close win (1%). The early call for Gore was based on a prediction that Gore would win by about 5%.

Again, this validates the exit polling science. It was spot on in 2000 as well--except for one obvious bit of human error that was quickly found and corrected.

Therefore: not until 2002 and 2004 have we had reason to doubt predictions based on the exit polls

[thanatonautos]

64. Thanks for the explanation ...

I think you have it right about the data entry error,
I must have misremembered that part of the story.

This article agrees with what you have to say about it.

For a certainty, the margin for Gore statewide was
many thousands of votes when spoilage was corrected
for. The little booklet Jews for Buchanan has a really
excellent analysis of the 2000 results in Florida.

[intelle]

47. But, it is not certain if the percentages are unweighted

and, the explanation provided by Mitofsky is bogus.

So, this article tells us nothing really substantial. I know I do not ever remember the news channels being as wrong as they were in 2004.

[proudbluestater]

54. I read the article, but where did he get this info?

He does not give a source for this exit polling data from past elections. It's been common knowledge that exit polling data never was off until 2000, and ever since then. Until I see some source to back up his claim, I'm not sure I'm ready to buy it.

[ClintCooper2003]

23. I'm going to start an exit polling firm so we can

get to the bottom of this. Care to join me?

[myschkin]

46. there's something wrong with this data

Just my opinion on this: I think the Freeman paper handled with WEIGHED exit poll data. I think, the difference in 96/92/88 is impossible - the whole exit poll effort would be ABSURD - nobody would believe in them. But the media and also the stock market believed in them this year... I think they must have some accuracy. Also were the Zogby Polls far off this year... and this TOGETHER with the Exit Poll discrepancy is very strange...

Even Tony Blair believed in the Exit Polls! ;-)

[adolfo]

50. Polls were correct

Exit polls are used partly in detecting fraud. The polls were accurate until they "adjusted" the results with the final tally. Obviously, it seems like they are covering their tracks.

[texpatriot2004]

51. Not that I know of, there was never a problem before 2000 n/t

[PennyMan]

53. Very Interesting Site About Exit Polls

Here Is A Very Interesting Site About Exit Polls I Found To Be Helpful. http://ustogether.org/election04/supporting_doc/Freeman...

[dargondogon]

58. Complain to Mitofsky;'s peers in opinion research

Republican pollster Frank Luntz got smacked down by the American Association for Public Opinion Research for failing to disclose his methodology.

Perhaps we should complain to AAPOR about Warren Mitofsky.

Time and again I've read about researchers complaining they can't get the raw data or a clear explanation of methodology from Edison/Mitofsky's National Election Pool exit polling. If true, that may conflict with AAPOR's Code of Professional Ethics and Practices.

From AAPOR's news release:
"In particular, the AAPOR inquiry focused on Luntz's reporting, prior to the November elections in 1994, that his research showed at least 60 percent of the public favored each of the elements in the GOP "Contract with America."
When later asked to provide some basic facts about this research, Luntz refused.

"AAPOR holds that researchers must disclose, or make available for public disclosure, the wording of questions and other basic methodological details when poll findings are made public. This disclosure is important so that claims made on the basis of opinion research findings can be independently evaluated. Section III of the AAPOR Code states: "Good professional practice imposes the obligation upon all public opinion researchers to include, in any report of research results, or to make available when that report is released, certain essential information about how the research was conducted."

http://www.aapor.org

Incidentally, the AAPOR web site makes passing reference to other associations (CASRO and NCPP) that apparently have similar guidelines for disclosure of polling data and methodology. Here's a paste from the site:

Disclose all methods of the survey to permit evaluation and replication.

Excellence in survey practice requires that survey methods be fully ddisclosed and reportedin sufficient detail to permit replication by another rresearcher andthat all data (subject to appropriate safeguards to maintain privacy and confidentiality) be fully documented and made available for independent examination. Good professional practice imposes an obligation upon all survey and public opinion researchers to include, in any report of research results, or to make available when that report is released, certain minimal essential information about how the research was conducted to ensure that consumers of survey results have an adequate basis for judging the reliability and validity of the results reported. Exemplary practice in survey research goes beyond such standards for "mminimaldisclosure," promulgated by AAPOR and several other pprofessionalassociations (e.g., CASRO and NCPP) by (a) describing how the research was done in sufficient detail that a skilled researcher could repeat the study, and (b) making data available for independent examination and analysis by other responsible parties (with appropriate safeguards for privacy concerns).

A comprehensive list of the elements proposed for disclosure by one or more ssources which in combination, exceed the "standards for mminimumdisclosure" proposed by any one of the professional organizations includes:

* who sponsored the survey, and who conducted it;
* the purpose of the study, including specific objectives;
* the questionnaire and/or the exact, full wording of all questions asked, including any visual exhibits and the text of any preceding instruction or explanation to the interviewer or respondents that might reasonably be expected to affect the response;
* a definition of the universe the population under study which the survey is intended to represent, and a description of the sampling frame used to identify this population (including its source and likely bias);
* a description of the sample design, including cluster size, number of callbacks, information on eligibility criteria and screening procedures, method of selecting sample elements, mode of data collection, and other pertinent information;
* a description of the sample selection procedure, giving a clear indication of the methods by which respondents were selected by the researcher, or whether the respondents were entirely self-selected, and other details of how the sample was drawn in sufficient detail to permit fairly exact replication;
* size of samples and sample ddisposition theresults of sample implementation, including a full accounting of the final outcome of all sample cases: e.g., total number of sample elements contacted, those not assigned or reached, refusals, terminations, non-eligibles, and completed interviews or questionnaires;
* documentation and a full description, if applicable, of any response or completion rates cited (for quota designs, the number of refusals), and (whenever available) information on how non respondents differ from respondents;
* a description of any special scoring, editing, data adjustment or indexing procedures used;
* a discussion of the precision of findings, including, if appropriate, estimates of sampling eerror withreferences to other possible sources of error so that a misleading impression of accuracy or precision is not cconveyed and a description of any weighting or estimating procedures used;
* a description of all percentages on which conclusions are based;
* a clear delineation of which results are based on parts of the sample, rather than on the total sample;
* method(s), location(s), and dates of interviews, fieldwork or data collection;
* interviewer characteristics;
* copies of interviewer instructions or manuals, validation results, codebooks, and other important working papers; and
* any other information that a layperson would need to make a reasonable assessment of the reported findings.

[reality_bites]

60. Questions to and answers for the master

"Is there any documented evidence of exit poll error prior to 2000?"
I assume this worded incorrectly.

"Why are exit pollsters even used at all if they are so inept?"

Even innaccurate exit polls contain valuable (even accurate) subgroup demographic data. It is this demographic data that those why pay for exit polls are interested in. Google using the following criteria, if you wish to understand how demographic data in a exit poll can be useful even if the entire exit poll is off: Google this:

"exit poll" statistics random -kerry -2004

"Can anyone explain why it is that only in the U.S. does the media make the claim that the exit polls were wrong?"

Maybe foriegn media outlets have lower standards. Also, Mitafsky has noted a dozen times in the last 3 weeks that RAW exit poll numbers are useless to show election fraud. He is a expert in this field so why would the US media discount him and favor the work of ameteurs ?

"Can anyone explain why other nations rely on exit polls to protect against fraud and we seem to rely on the vote counters to prove that exit polls are wrong? Doesn't that seem odd?"

Please provide some examples and/or links.

"Can anyone explain why, even after the selection 2000 fiasco in Florida, the media would never even consider the possibility of programmed election fraud, but instead would fault the exit polls?"

I think the media would be more than willing to conclude there is a problem if there was some evidence of fraud. Please don't respond with your tired old claims that your work means something.

[liberal43110]

63. Just saw a plausible reason in another thread.....

One reason that exit polls in the U.S. frequently skew Democrat may be related to ballot spoilage and the higher than normal spoilage rate for Democrats, especially the poor, those with little education, and those with poor/no English skills. The polls show the electorate's true intentions, but ballot spoilage (and I mean this in a "generous" sense, because I do not think we do enough to try to read many of these ballots, but the rationale is always that the vote margin is larger than, blah, blah) reduces the actual vote tallies for Democrats.

Just a thought....