Is there a smoking gun in the smoking gun? TIA and Others?

[davidgmills]

The CNN poll shows that only 38% of the respondents voted for Gore in 2000, yet 41% of the Respondents voted for Bush in 2000.

Questions:

1. Shouldn't these numbers have been about equal?

2. Since they weren"t, weren't Republicans overpolled by about 1.25% and Democrats underpolled by about 1.25%?

3. When Respondents were ID'ed for party affiliation, 38% claimed to be Democrats and 36% Republicans. Zogby maintained all along that the ratio was 39% Democrat and 35% Republican and that is what he used in his polls. Add one percent for the Democrats and take away one for the Republicans and you get Zogby's ratio of 39% to 35%. Was Zogby right after all?

4. Now for the Blockbuster question. If the Republicans were over polled, and the Democrats underpolled, was this intentional? Did Mitofsky cherry pick Republican precincts from 2000 to give Bush an edge in the exit polls? Does Mitofsy want to take the Fifth Amendment?

5. What would have been the percentages for Kerry and Bush had the Republicans not been overpolled and the Democrats underpolled and what does this translate to in actual numbers of votes?

[Fiona]

1. One thing to remember

is that people often are mistaken or lying when asked who they voted for the last time. Some percentage of people will always claim to have voted for the winner.

[davidgmills]

2. The Winner last time was

Gore

[MelissaB]

3. Do you really think they would

lie right after voting? They didn't know who the winner was then.

[Fiona]

6. the question was

who did they vote for in 2000.

[lonestarnot]

4. This is just ridiculous theory!

[Fiona]

5. No it's not

but your factual refutation and well-stated argument are impressive.

[lonestarnot]

7. Well I don't have time to refute of every ignorant post I see. There is

much concrete evidence here. Seek and you shall find truth.

[Name removed]

8. Deleted message

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

[lonestarnot]

11. Contrair bitchy? WTF All I said was lack time, and seek truth...bitchy?

Not...

[Name removed]

16. Deleted message

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

[lonestarnot]

19. Didn't call you ignorant, just a post that doesn't appear to have been

well thought out. I am not a polster and would venture a guess that you are also not a polster. They are very mathmatically complicated. I do not believe majority of Americans would lie to polster re vote. In fact most REPULSIVES I have spoken with since the election, would prefer to have Bushitler removed, when it it demonstrated, as it will be, that people did NOT LIE when exit polled.

[Fiona]

21. You are simply wrong in your belief, then

A post below describes how many more people claim to have voted AT ALL than actually voted. And yes, it's known that more people claim to have voted for the winners than actually voted for him or her.

Seek the truth, blah blah blah.

[lonestarnot]

22. Ah blah blah blah; You chose to believe Americans are liars; I choose to

believe they are not. Shall we agree to disagree?

[Fiona]

23. If it were simply a matter of opinion

we could do that. But it's not. It's established fact, well-known by pollsters.

[lonestarnot]

24. Not a fact!

[Fiona]

26. there are posts below

proving that people lie to pollsters. Perhaps "lie" is too strong a word. Perhaps they "misremember".

If you polled Americans in 1975 and asked how many of them voted for Nixon in '72, you would've gotten a lot fewer Nixon voters than actually voted for him.

People lie. But it's nice to know there are people in the world who are unaware of that fact.

[lonestarnot]

28. Unaware of the fact that people lie? Now your getting a little more off

base. Case closed!

[Fiona]

31. Yes

your factual argument has convinced me. :eyes:

[lonestarnot]

32. No facts or citations in your argument, just arrogance and hostility.

[Fiona]

43. I referred you twice to posts below

with some citations.

"arrogance and hostility".... you crack me up.

[davidgmills]

14. I have heard this claim as well

But I seriously question it. I'm a lawyer. We have to judge credibility all the time. To me, there really is no motive to lie about who one voted for in a Presidential election. Yes, there always will be people who feel like they needed to be on the winning side, but I just don't see the evidence for it in Presidential polling. There is just no motive to lie about it and nothing to be gained by lying.

But in the 2000 election, the question of who won is really a cloudy issue. Gore won the popular vote and Bush (thanks to the SC) won the electoral vote.

So I find this argument even more frought with danger in this instance.

[Darkhawk32]

9. All avenues must be explored...

Truth is stranger than fiction...

Raising the question shouldn't be grounds for rudeness.

[SammyWinstonJack]

10. How could anyone not remember who they voted for ?

And why would anyone lie about who they voted for, winner or not? Makes no sense. OTOH, I might lie about voting for the Chimp, it would be embarrassing to admit a vote based on greed, stupidity or self righteousness.

[SicTransit]

13. Do you think people would remember whether they voted or not?

From Kevin Drum:

http://calpundit.blogspot.com/2002_10_06_calpundit_archive.html

Using the resonable assumption for 2002 that once again about 62% of registered voters will actually show up at the polls, the conclusion is that people lie when pollsters ask them questions. Lots of them. And if you think "lie" is a little strong, consider the results of a Princeton Survey Research Associates poll taken after the 2000 election. 72% of respondents said they voted, while actual voter turnout was 51%.

[davidgmills]

15. That's far worse than who is going to be a "likely voter"

Pollsters spend lots of time on who is going to be a likely voter and they do it based upon reported past voting records of their respondents. They are very accurate with their predictions of who will be a likely voter so this one analysis does not square with the record that pollsters have for coming up with likely voters.

[SicTransit]

17. What the quote I posted means is

people massively lie to pollsters about who they voted for or whether they voted at all.

[davidgmills]

18. And I am arguing with that conclusion.

[SicTransit]

20. You're arguing with well established research

"Response error tied to over-estimation of voting is one of the oldest and most persistent types of response error to be documented...

reports that such response errors tend to range between 12 and 16 per cent, with the error tending to be larger the closer a survey was done to the election".

Robert H Prisuta, A post-election Bandwagon Effect, 1992

Stanley Presser: Can Context Changes Reduce Vote Over-reporting; Public Opinion Quarterly, Wier 1990

[davidgmills]

25. That's ONE analysis ONLY

If it is so established, how do pollsters get so accurate with their predictions of "likely voters?"

If this were universally true, pollsters would never be able to predict with any accuracy who "likely voters" are.

That is why I question the study's conclusions.

[SicTransit]

27. Pollsters cannot predict with any

accuracy who "likely voters" are as it is. They guess.

[davidgmills]

30. So let's do away with polling altogether because people can't be trusted

That's the conclusion I draw from your arguments.

Where's your article that says they guess?

[davidgmills]

33. Back to My original Questions Which no One Has Answered

Assuming that people don't massively lie to their pollsters about who they voted for last time, where do these numbers take us?

I'm tired of the discussion of whether people lie to pollsters. If you want to discuss this issue further please make your own post and have your own discussions about it.

[SicTransit]

35. There is a quote from Abraham Lincoln

"If you call a tail a leg...how many legs does a dog have?

Four.

Calling a tail a leg does not make it a leg."

-Abraham Lincoln


The lesson: do not fool yourself into making bad conclusions if you have bad assumptions.

[davidgmills]

36. No shit

Why don't we argue about whether there is a God?

Adds about the same to the argument.

[anaxarchos]

44. Smug and Scientific both start with 'S' but they are not the same word...

And don't fool yourself into thinking that anything you've said has anything whatever to do with science...

1) Your citation concerns opinion polls, not exit polls - "Calling a tail a leg does not make it a leg".

2) Your citation actually supports the case for "fraud" as no "Bandwagon Effect" whatever is discernible in opinion polls before and after this election, thus your source actually reinforces the case for the accuracy of exit polls in 2004. Try Ken Maxwell on precisely the same citation (hmmm... EXACTLY the same citation at that):

http://www.jamaicaobserver.com/columns/html/20041106T200000-0500_68984_OBS_A_LOBOTOMY_FOR_DEMOCRACY.asp

"Response error tied to over-estimation of voting is one of the oldest and most persistent types of response error to be documented. . reports that such response errors tend to range between 12 and 16 per cent. with the error tending to be larger the closer a survey was done to the election". ( Robert H Prisuta, A post-election Bandwagon Effect 1992 and Stanley Presser: Can Context Changes Reduce Vote Over-reporting?; Public Opinion Quarterly, Wier 1990)

"In this case, and as far as I can discover, only in this case does the percentage claiming to have voted for the winner fall below the percentage actually voting for him.

The US press in its cocoon of fantasy, pretends to believe that this result is possible and accurate. No one can - without his consent - be deprived of his rights. It says so even in Third World constitutions."

3) There is not a large body of work on respondents lying to exit pollsters. There is precisely 1 case in the UK where Conservative Party voters are purported to have been too "ashamed" of their vote to admit it. The evidence for even this case is largely anecdotal. There is not a single piece of empirical evidence supporting this thesis in the 2004 election. Instead there is a whole lot of TALK by mystery pollsters that this is what "must of happened" because the actual vote count "can not possibly be wrong". That would not be "plausible".

4) The bias in #3 above is what makes exit "pollsters" unreliable in this case. The craft (no, it is not a "science") takes the voting results as the empirical data and therefore judges exit polls based on conformance. This is exit pollsters assumption, their business, and their art. That is why they "adjust" exit polls at all - so that polling can more accurately represent "reality". If there is a conflict here, they immediately start explaining why the poll "must have been wrong". Blumenthal, in particular, reeks of this. But that doesn't mean it is not circular reasoning. As far as election fraud is concerned, any lawyer or anyone else is as much an "expert" as anyone else, resume mongering notwithstanding.

5) Now, on to the original poster's question. There is evidence that suggests Democrats repeat vote less often than Republicans. This is directly caused by the poverty of some Democratic voters (they move a lot more) and the way that voter rolls are "purged" in many states based on residency. I don't remember if it accounts for the kind of numbers you point to. Would you like me to dig a little?

[davidgmills]

50. Please Do

I have been arguuing with Blumenthal on Mystery Pollster for some time and he wilts under cross examination.

You obviously have done your homework.

If you look at my posts on this board you will see some of my criicisms of Blumenthal. Many in the last 48 hours or so. On his site as well.

I understand the concept of clustering and the design effect phenomenon. I have said that I can see how the concept applies to clusters of marbles in a jar. But it is a stretch from there to precinct exit polls. I understand the general rule, but I also understand that general rules have exceptions when the facts warrant.

This post is a good example. If somone wants to make the point that the general rule is that people lie about the loser they voted for, I can except that as a general rule. What I can't accept is its application to every situation, especially the 2000 election. Who won? Depends on who you ask.

Of course you have even given me pause to accept the "general" rule, which my nature and training makes me suspicious of anyhow.

But to the point at hand. Since MOE's are so crucial in this debate, it would be nice to know if we have some wiggle room based on underpolling of Democrats and overpolling of Republicans. I can see your argument that democrats move because of poverty and are dropped from voter rolls. Makes sense on a local level. However, I would expect those voters to vote somewhere else. Also, this still would not explain overpolling of Republicans. Chances are the democrats would not be moving out of Republican precincts because they could not have afforded to be there in the first place. They are not upwardly mobil. I could see the opposite happening. Upwardly mobil Republicans moving out of lower class neighborhoods.

[anaxarchos]

53. The issue is more basic ...

In poor neighborhoods there is less voter education, less government infrastructure, even fewer phone numbers which are more likely to change. People are registered by their church, by a community organization, or in a voter registration drive. Then, for whatever reason, they are dropped off the rolls. They don't know how to get back on, or they lose touch with voter transportation, or the mechansism that got them to the polls goes away. The net result is that the re-vote is less in poor urban precincts and, of course, this is precisely where the core of the Democratic vote comes from. Take a look at this citation from the Kennedy School at Harvard (a very good site for a lot of background info):

http://www.vanishingvoter.org/Releases/release111104.shtml

The number one reason (one third of the sample) for not voting is not "voter apathy" as the press would have it, but "Because I moved recently and hadn’t registered at my new location" (32%). Also look at "Because I didn’t have any way to get to the polls on election day" (20%), "Because I didn’t quite know how to go about registering to vote" (13%), "Because I thought I was registered, but when I went to vote I learned I wasn’t and could not vote" (6%), "Because I was worried that my right to vote would be challenged and I would not be allowed to vote (3%), etc. Overall, the difficulty of registration and voting logistics issues trump political issues 120% to 83% (responents were allowed to cite multiple reasons).

To quote the study, "In many democracies, registration is renewed automatically when a registered voter moves to a new residence. America’s registration system, which places the burden of registration on the individual rather than, as in Europe, on officials, depresses turnout in other ways, too. One in eight of non-voters with an interest in voting said they simply do not know how to go about registering."

The effect is "Although it is difficult to estimate precisely the combined effect of these various factors, there is little question that several million more Americans would have cast a ballot on Election Day if the country’s registration and voting system were more welcoming."

And of course, it all goes against the Democratic Party vote disproportionately.

The use of voter registration and voter qualification issues to not only implicitly "depress" but to explicitly suppress the poor vote in America is as old as the republic itself and has always been very well understood. In the last 75 years, it was one of the most important pillars of southern segregation.

More recently, however, it has become one of the staples of local Republican election strategies. Because of the non-uniformity of registration and election laws and the fact that registration rolls can often be purged from the state OR county levels, many different mechanisms are used. The effect however is the same.

Almost invariably, the purge targets poor, urban African-American precincts. The reason is not simply racial but tactical as well. The black vote is the single most important constituency in the United States and votes 9 to 1 Democratic. It is also a geographically and demographically concentrated population which has the advantage of taking out disproportionately Democratic urban Hispanic and poor white voters as well. You come up with a reason to purge the rolls and count on simple economics and logistics to keep those wrongly purged from re-registering. The reasons given are sometimes related to "re-districting", sometimes a change of voting system, sometimes economic, and almost always to prevent "vote fraud" but always at the expense of disenfranchising dozens of legitimate voters for every "phantom" vote that is prevented. The statewide purge of (wrongly identified) "ex-felons" in Florida in 2000 and the attempt to do the same in 2004 are examples of this.

All in all, this is one of the "dirty little secrets" of American politics and whatever it is, it ain't Democracy.
The irony is that the gap among re-voters is not much larger than that which you report but it is also true that poor voters are more likely to want to vote and to try to vote than the general population. I can probably provide citations on this as well (my memory is failing me at the moment).

Please forgive my rudeness for not replying sooner but I was away for the holiday.

[davidgmills]

55. Much of this discussed in Freeman's new report

A real indicment of the voting system.

[Doctor O]

54. Prior to the election, I remember that:

The pollsters were cautioning the networks AGAINST using the exit polls to predict the winner. and:

There was much speculation with the pre election polls that one large problem all the pollsters were having was the predicted voter turnout.

It seems that they were speculating on shifts in the % of Reps / dems /. Independents and they were concerned about not being able to hit the right mix. in order to predict the models.

In 2000 the networks in claimed the exit polls were flawed, and hired mitofsky to build a new exit polling process and scrapped the old one. 1992 and 1996 exit polls while right in selecting the winner but had poll results favoring democrats by up to 8 percentage points deviance from the final results.

So if Exit Polls are so lose why did the netwroks fire their pollsters in 2000, and rebuild the system. Only to claim to have problems this time under the new improved system.

I also remembers hearing speculation that the pollsters were unsure of the models when the turnout hit 120 million. I beklieve the comment of one pollster was "Once the turnouts hits 120 million, all bets are off."

If these were in fact comments by respected pollsters, would this further damage reliance on exit polls for evidence.

[shraby]

47. I think the bigger question is

who did they say they voted for THIS time..and the majority said they voted for Kerry at least until they jiggered the polls to match what the machines were saying.

[SicTransit]

34. There were tons of articles before the elections

that discussed the inability of the pollsters to predict accurately who the "likely voters" are.

And what's the point of me bringing you an article? You will just dismiss it because you think it's wrong. That seems to be your style of argument.

[Freddie Stubbs]

29. Some will even claim to have voted when they didn't

People often don't want to admit to being bad citizens.

[lonestarnot]

12. Is Ms. Jones a relative?

[dlaliberte]

37. New voters didnt vote for anyone in 2000

Note that it says of those who voted in 2004, 38% had voted for Gore and 41% for Bush. That sounded fishy to me also, since Gore won the popular vote (by a narrow margin). But there were also 17% new voters in 2004 (6 out of 10 voted for Kerry), and those who did not vote for either Bush or Gore also voted 3 out of 4 for Kerry. So my intuitive math says this works about right.

[davidgmills]

39. The Numbers Add Up to 100%

But the issue is did they overpoll Republican precincts and underpoll Democratic precincts.

[L. Coyote]

38. Make enough SMOKE and someone might think there's a fire

SMOKE SMOOCK

A bit more level-headedness and careful thought about these thing would be useful. Thanks for raising some "real" questions.

Before getting too excited about playing these numbers to meet expectations, it is important to remember the assumptions being made.

Assumption 1. The people who voted in 2000 are precisely represented by the 2004 exit poll.

People who voted in 2000 and did not vote in 2004 are not being exit polled in 2004!!

Until this is taken into account, extrapolations from the data are meaningless. In other words, METHOD FLAWED!! Forget it.

There are other assumptions.

SCREAMING SMOKE IN ALL CAPS DOES NOT MAKE A FIRE. (Try lower case, please.)

Next, people like to vote for a winner. Basic psychology tell us that after the votes are counted, some people lie and say they voted for the winner, moreso than the opposite. What they say about their 2000 votes will be skewed. How much??? Good question. Probably less that the 3% here.

[davidgmills]

40. All lower case I think in my original post

Also you say:

"People who voted in 2000 and did not vote in 2004 are not being exit polled in 2004!! "

Agreed. But wouldn't one expect that the voters in 2004 who voted in 2000 be more equally split? If not? Why not? The only reason offered so far is that voters lie about who they last voted for.

Are we to assume that 3% of Gore voters chose not to vote at all this time? Otherwise, we must assume that Gore voters and Bush 2000 voters would be equally split.

[L. Coyote]

48. Re: Assumptions and UPPER CASE

The upper case comment is directed at those using it, not you. Upper case in E-media is bad etiquette, like screaming at a party!

Re the assumptions question. My point is that we are not to make conclusions without considering our assumptions. People take a set of numbers and start screaming FIRE. They see smoke everywhere. But their methods are flawed. They don't even reflect on what their assumptions are.

I'm not weighing in on the question. I'm questioning the methods of those seeing smoke everywhere. If they refined their methods, they might clear the theater after all. Why listen to them meanwhile?

[Occulus]

41. Was the info here debunked yet?

Or are these files legit?

http://www.scoop.co.nz/stories/pdfs/Mitofsky4zonedata/

[Truman01]

42. See how accurate polls are?????? nt

[Time for change]

45. I think it's a moot point because

as you can see from these two links, the exit polls show Kerry winning the popular vote from 2.6 - 3%.

http://www.scoop.co.nz/stories/pdfs/PopularVotePaper181_1.pdf
http://www.scoop.co.nz/stories/pdfs/Mitofsky4zonedata/US2004G_3798_PRES04_NONE_H_Data.pdf

[ClintCooper2003]

46. He didn't overpoll Republicans, but I think he may have

overpolled Independents who voted for Bush last time. I think that 38% of the people polled were Democrats and 36% were Republicans.

[davidgmills]

51. That's a plausible explanation.

But I was curious because Zogby has always maintained that Democrats have been casting 39% of the vote and Republicans 35%. It just works out that if you add 1 to the Dems and take one away from the Repubs, then you get Zogby's 39/35 ratio.

[Land Shark]

49. Political Scientists say maybe people lied about 2000, then...

Two political scientists presented this data at a university near me and said it didn't add up at all, suggesting that people may have forgotten who they voted for in 200 (unlikely) or lied about it (also unlikely). Then I brought up the possibility of the election data being wrong and they had to admit it was indeed also a possibility. It's just too bad they didn't mention all the logical possibilities up front.
Land Shark

[davidgmills]

52. I would tend to suspect

the data being wrong before I would believe lying and memory lapse. Especially given what we are beginning to suspect about systemic voter fraud being on the increase since computers and partisan corporations usurped the voting process.