We should all rejoice in the promotion of the rBr Liepothesis.

[TruthIsAll]

It just shows their desperation.

It just makes our case stronger.

Keep your eye on the ball.

Focus on the Voting identity:

Total 2004 Vote = 122.26 = Living(Gore + Bush + Nader) voters + New voters.

We know the total number of 2004 voters.

We know the maximum number of Gore 2000 voters who could have voted in 2004.
We know the maximum number of Bush 2000 voters.
We know the maximum number of Nader 2000 voters.

Therefore we know the MINIMUM number of NEW voters who voted in 2004.

These NEW voters consisted of the following:
1) Brand new voters who NEVER voted before.
2) Voters who voted before 2000 but sat out that election.

That's it.
That's all you need to know.

Oh, and don't forget.
Kerry won a clear majority of BOTH groups of voters (57%).
Kerry won a clear majority of Nader voters (70%).

The margin of error for the 3168 respondents to the question "How Did You Vote in 2000" is 1.77%

That means that the probability is 97.5% that Kerry's actual percentages were greater than 55.2% for New voters and 68.2% for Nader voters.

Focus on the FACTS.
FACTS have a way of winning over MYTHS and LIES.


Its basic arithmetic.

Its the Voting Identity (VI).
Its NOT a hypothesis.
It's a fact.



Where did Bush get 13 million new votes?
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x367638

Even if 100% of Bush voters turned out and Gore voters stayed home,
Kerry still won. Check the scenarios.

The Graph


The Numbers
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x367600

[whistle]

1. Could you list those facts (numbers) here using the outline ....

...you provided so that the rest of us can see what you know.

[Peace Patriot]

2. You're right. "rBr" is stupid and silly, and oblivious of facts...

...and that's all they've got--that, and Wally O'Dell & Co. counting all the votes in secret.

30 people in Waynesville, NC, last week were thrown out of their Baptist church for refusing to pledge support for Bush. Who would be the "reluctant responders": the majority of people in that church who voted for Bush and who don't think others are entitled to an opinion, or the 30 people-minority in that church who may well have voted for Kerry despite the bullying of the majority--bullying so severe that the dissenters have now been thrown out of the church?

Follow one of those 30 dissenters to the polling place a few blocks away from the church. She's been thinking about things a lot, and praying a lot. She thinks Jesus is for peace, not war. She thinks poor people should get help from the government, not the rich. She believes in "live and let live" on social issues, and it'll all be sorted out in Heaven by a Higher Justice than we can ever know. She therefore votes for Kerry, and against all this Bush stuff--invading other countries and benefiting the rich and hating homosexuals. She knows what her pastor thinks. He's a Bushite--and not shy about it. She knows what the majority in the church think--Bush is Lord, and Kerry voters are witches and devils. She votes for Kerry because that's what her conscience tells her to do, and she knows that her choice is protected by the secret ballot. A stranger approaches her as she leaves and asks how she voted. She doesn't know who it is, or who they might tell. She walks away--she doesn't respond--or, she tells a little white lie that she voted for Bush, because she doesn't know who it might get back to, and she doesn't want repercussions in a church that is about to boil over with controversy, six months later.

Reluctant Responder, circa 2004.

Draw me a picture of a Bushite voting for Bush in BushWorld 2004, who would be reluctant to admit a vote for Bush. Then I might think twice about "rBr." Because I can't picture it. Numbers aside, it makes no sense in our present political climate. What would such a voter have to fear in a country dominated by Bush, and in a subculture of intimidation and coercion in which support for Bush is expected, and rewarded? (Numbers included, there is no data to support it--except PAST models when Republicans were suspicious of "leftist media"--something they have no reason to be suspicious of today-- and who, in any case, have already been factored into the models (possibly even creating an erroneous pro-Bush skew in the models in '04.))

I can't picture ANY voter more likely to be afraid and reluctant to state their vote publicly than a Republican Kerry voter in a Bush stronghold. And I can't picture any OTHER voter who would be afraid and who would walk away and not respond, or lie about their vote, except maybe DEMOCRATIC voters for Kerry ALSO living in Bush strongholds, and not wanting their boss to find out, or their minister, or not wanting their tires to get slashed. And then there are the black Kerry voters in Democratic strongholds who got anonymous phone calls from threatening southern, white male voices who promised a return to jail if they showed their faces at the polling place, and those who faced such types of people challenging their right to vote within the polling place itself. They come outside, after that harrowing experience, and some youngster they don't know comes up to them and asks them how they voted. They might just lie, or walk away.

All of these are picturable and seem quite possible to me as "reluctant responders"--with Republican Kerry voters in Bush strongholds being the sharpest picture of all: dissenters from the Bush paradigm within a social and political context that would punish them for such a vote.

Well, ok, since I'm doing pictures, I CAN picture a Republican voter for Bush in a Democratic stronghold feeling a bit outnumbered--especially in THIS election, with Democrats voting in droves and lining up at the polls for hours to get Bush out. But it's hard to picture such a person being shy of pollsters. I picture such a person as more pugnacious in BushWorld 2004, and not in the least shy about his or her views. They hold to their views despite what the majority might think, and would feel quite righteous about it (is my picture--I'm thinking of a member of my family in just such a situation within the family--not shy!). In any case, what does such a voter have to fear from disclosing their vote to a pollster? Economic repercussions? Democrats don't tend to inflict them--they don't decide on a grocer or a doctor by their political affiliation--whereas I've heard that a number of rightwing groups do just that. Social repercussions? Democrats tend to be more open-minded than Republicans, at least Bush Republicans--more tolerant of dissent. Being bullied, rejected, thrown out of church? When did Democrats ever exhibit such behavior?

Maybe a few rBr's in Democratic strongholds. There are always a few in any category you could name. But a significant phenomenon, in THIS election? I don't think so. And compared to the other possibilities--especially Republican defectors to Kerry in Bush strongholds, who have real and palpable things to fear (fears that in Waynesville came true), the odds of rBr's in Democratic strongholds turning a 4 million vote Kerry victory (the exit polls) into a 2.5 million Bush victory (official tally) are just not there.

And who else is there--except voters in circumstances of intimidation or strong social pressure who might have caused the exit polls to err in Bush's favor (understating Kerry's victory)? That is, a) Republican defectors to Kerry in Bush strongholds with good reason not to talk to pollsters; b) Democrats voting for Kerry in Bush strongholds with similar reasons; c) Minority voters who were bullied, or heard of others who were threatened, but who managed to vote anyway; d) Women voters for Kerry in disagreement with their Bushite husbands--and other such picturable voters who would lie about their vote, or walk away from a pollster, for reasons of fear.

The conditions of fear-mongering and fear, and intimidation and social pressure in Bush strongholds that would come to bear on defectors voting for Kerry are far more picturable than their opposite (fearful Bush voters). And those conditions are still in effect (as Waynesville shows), and would be doubly reinforced by Bush still being in power, thus suppressing any effort at crying foul on the part of Kerry voters in Republican strongholds who suspect that their votes were stolen.

2004 is not 2000, nor 1996, nor even 1980. 2004 is a "brave new world," with "rBr" vanishing in the haze, and fear and loyalty tests back in style in the subculture of Republicanism (much like the McCarthy era of the 1950s). Some basic assumptions about our democracy are no longer true--including giving the benefit of the doubt to the official vote tally.

[autorank]

3. One of your best yet...

I just exhausted my "febble" mind (there I said it again, may the Lard strike me down) comment on the atavistic Liepothesis from a common sense point of view (Bush voters are not "reluctant"...duH)
http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x367600#368256
so I'm in need of the vapors to recuperate.

This is just perfect.

Rumors of my departure are greatly exaggerated...by me." autorank

[bleever]

4. Agreed. The best explanation they have

is really quite weak, pathetically so.

But: their power resides in making sure that nobody's even looking for or caring about an explanation. Then it doesn't matter if we're right.

So we'll just keep talking, and talking, and talking about it.

This crime isn't going away.

Neither are we.

[snot]

5. TIA, I want to say thanks, & I hope history remembers you.

[Febble]

6. TIA one thing you really need to know

The MoE tells you the confidence interval for a result, given sampling error.

Sampling bias is something completely different.

No amount of statistical power will make sampling bias go away.

Even if your samples were so huge that the margin of error was infinitesimal, you could still have sampling bias.

Sampling bias arises when you violate the assumption that the people who refuse to take part in a survey are drawn from the same population as the people who agreed to take part.

The lower the response rate, the greater the likelihood that this assumption will be violated. Response rates in the exit poll were low.

There is therefore absolutely no a priori reason to assume that the assumption was not violated.

This has absolutely nothing to do with sampling error, or with the MoE, which is a way of expressing sampling error.

Proving that a result was outside the MoE does absolutely nothing to refute the possibility of sampling bias.

You may be right - that the Great Exit Poll discrepancy was due to fraud, not sampling bias. But no amount of math (not that we can do anyway, regression analyses of the actual data might do it) will tell us which it was.

Which leaves us with actual evidence of irregularities which is fine. They are excellently documented in Conyers' report.

But the math that you keep posting is only relevant inasmuch as it tells you the size of the discrepancy. Not what caused it.

[TruthIsAll]

7. Febble, a few things you really need to know..

I Focus on the FACTS.
FACTS have a way of winning over MYTHS and LIES.
Its basic arithmetic.

Its the Voting Identity (VI).

Its NOT hypothesis.
It's a fact.
It's simple arithmetic.

And I challenge you to prove otherwise.

EVERYONE CAN UNDERSTAND VI.
ESPECIALLY THOSE WHO NEED TO KEEP WITHIN A BUDGET.

HERE THE BUDGET IS 122.26 MM VOTES.
YOU ARE TRYING TO RE-ALLOCATE THE BUDGET WITH A BOGUS HYPOTHESIS.

IF ANYTHING, IT WOULD BE KERRY VOTERS IN BUSHVILLE WHO WOULD BE RELUCTANT, NOT THE OTHER WAY AROUND. BUSH PEOPLE CAN BE QUITE PUGNACIOUS.

DID YOU SEE THEM AT THE DADE COUNTY RECOUNT RIOT IN 2000?

BYE, BYE rBr..

[Febble]

8. So if what you are quoting are facts

what does this mean?

"The margin of error for the 3168 respondents to the question "How Did You Vote in 2000" is 1.77%"

[torque]

9. Game over feeble. You can't refute the laws of mathematics

[Febble]

10. Oh yes I can

It's called garbage in, garbage out.

Nothing wrong with TIAs math, it's his decisions as to what are facts that I question.

Math can't tell you whether the data inputs are correct.

[TruthIsAll]

12. Febble, be specific. Which data inputs do you object to?

The total number of 2004 votes?

Gore 2000 total?
Bush 2000 total?
Nader 2000 total?

The 13047 Sample MoE?
The 3168 Sample MoE?

At least you agree the math is correct.

[Febble]

13. Well, what do you mean by

"The margin of error for the 3168 respondents to the question "How Did You Vote in 2000" is 1.77%"?

Are you assuming that because the MoE is small, the data is correct?

If so, that is an unwarranted assumption.

As I have explained on another thread, the MoE tells you nothing about sampling bias, only sampling error.

If there was sampling bias, then the facts you quote are not facts.

If there wasn't, then the facts you quote are facts, and there was fraud.

But you cannot argue that because these are "facts" therefore there was fraud.

It is a circular argument

[autorank]

14. I'm glad you're not in law enforcement. No one would be prosecuted.

The constant questioning of assumptions and methods truly begs the question. There is more than enough evidence to satisfy a politically free prosecutor to pursue a case of election fraud. The evidence ranges from reports of fraud, circumstantial evidence of fraud, and top-down pointers to fraud through the polling analysis. Quibbling on the margins of polling analysis does nothing to further the general cause. You said you were outraged about actual fraud in Ohio, why not get some publicity on that, why not push those facts in forums that are open to change (we're convinced here)?

Given the behavior of the Bush administration, the callous disregard for domestic laws and foreign relations, there is no fraud to great that they would use to perpetuate their power. This is the key point. Do you know that Bush promulgated regulations to allow Mercury into the domestic water supply? This is just one fact that is part of a mosaic of facts that show things are just awful here (making the bLiar government look semi-civil, which takes great effort).

There was fraud, it was multi level, and there is enough evidence for any federal prosecutor to pursue a general conspiracy case. That's the point of all this.

[Febble]

16. Fine

If you've got enough evidence, prosecute.

Evidence is what we need.

Although I might say there has been plenty of questioning of my assumptions and methods around here, so give me leave to do a bit of questioning myself when the assumptions and methods look iffy.

I completely agree with you about the polls being irrelevant. I think they are useless as evidence as they are totally inconclusive. In fact I think they are worse than useless, as they discredit the very real evidence you guys actually have.

And I have done my level best to get publicity for the outrages in Ohio, inasmuch as it is possible for a British data analyst with a love of democracy to do. My computer is full of analytical evidence which I have emailed to lawyers, Representatives (Conyers) and media outlets. I have diaried them on DKos, and cited them elsewhere.

And the entire reason I got involved in the first place is that I think it is a disaster for the world that Bush became president (I won't say elected, because he was certainly not elected in 2000).

So yes, forget the exit polls. They are a turkey.

[autorank]

19. Corrections -- My corrections printed in bold.

If you've got enough evidence, prosecute. I'm not a prosecutor. I'm sure you know that. My point in the post is that the mosaic of evidence without full investigative prosecutors is stunning. I also pointed out that the people who would prosecute are all Bush appointees. The only way a Bush appointee is going to presecute is with a united front and a clear understanding that we have enough analysis and evidence to push for a quick, intense investigation which would, in all probablity, lead to action.

There were very good reasons the Polish freedom movement waqs called SOLIDARITY. That's required when your advocating a difficult cause. SOLIDARITY!

I completely agree with you about the polls being irrelevant. I NEVER SAID THE POLLS ARE IRRELEVANT. I said that as part of the direct, circumstantial, and top-down polling evidence, they are highly relevant.

And the entire reason I got involved in the first place is that I think it is a disaster for the world that Bush became president (I won't say elected, because he was certainly not elected in 2000).Nor 2004. If you are so concerned about Ohio, which you've said, but you don't think Kerry was actually elected, what's the point of all this. The "quibbling around the margins" is used and distorted by "election fraud debunkers" to discredit the entire movement.

So yes, forget the exit polls. I DIDN'T SAY FORGET THE POLLS. I said with only public media access and some digging, there is an abundance of evidence that the exit polls point to wrong doing. Proof, that's for a full investigation when everyone opens up in the interests of our most sacred function here, electing public officials, everything starts there.

[Febble]

22. I take your point about Solidarity

I don't agree with it, but I take your point.

I don't agree with it because I think a weak argument can undermine a strong argument, and I think that is what the exit poll argument does. In fact, I think the exit poll argument discredits the entire movement.

I am concerned about fraud in Ohio, but I think the evidence is better for massive voter suppression than of massive fraud. And people who couldn't vote, even if it was because of illegal activity, can't have those votes restored.

I am not a lawyer, but it strikes me that Ken Blackwell, and probably Damschroeder have committed prosecutable offences. I've done my best to provide legal teams in Ohio with my little bit of the puzzle, but I'm hoping that those on the ground are still working on it.

[Sawyer]

24. Hmm

I also pointed out that the people who would prosecute are all Bush appointees.

Every Attorney General in every state is a Bush appointee? I was not aware of that.

How about NY AG Eliot Spitzer? That guy does not seem to balk from going after sacred cows. There are people that claim that New York was one of the fraud centers to pad Bush's popular vote - why isn't Spitzer going after it if the evidence is so rock solid? Do you think he is "in" on it?

[autorank]

30. Federal Prosecutors.

You know, The Voting Rights Act.

I think Spitzer is an amazing public official. I'm not too worried about the results in NYS in 2004. Are you? Since he has no standing in Ohio, New Mexico, or Florida, to name the most prominent fraud areas, his interest is irrelevant.

There are people on this forum who may have experience with their state's attorney general. I'd be interested to hear from them. Whether they're Republicans or Democrats, they are political animals. Why would any state attorney general go forward when faced with a movement that had quibbling and sniping from the rear by "debunkers," naysayers, and debate team wannabes? That's why it's necessary to put the quibbling to rest and move forward in SOLIDARITY!!! Just a thought.

[Sawyer]

35. The voting in the US is really not a federal matter, afaiu,

the US Constitution does not even require that people actually vote for President. It is a state matter, and state AGs are the ones who would prosecute the fraud in the states, I would think.

[autorank]

39. US DOJ: "Voting Rights/Fraud:..prosecute violators vogorously..whereever..

From:

Fact Sheet

PROTECTING VOTING RIGHTS AND

PREVENTING ELECTION FRAUD

Background Information on the Justice Department’s Mission

to Protect Voting Rights and Prevent Election Fraud

§ The Voting Access and Integrity Initiative. On October 1st, the Attorney General ordered the creation of a Department-wide Voting Access and Integrity Initiative. The goals of this initiative are to:

ü enhance our ability to deter discrimination and election fraud; and,

ü prosecute violators vigorously whenever and wherever these offenses occur.

§ Civil Rights and Criminal Division Leaders Heading Initiative. To protect the vote of Americans, resources from across the Department of Justice have been committed to this effort. Assistant Attorney General of the Civil Rights Division Ralph Boyd, Jr. and Assistant Attorney General of the Criminal Division Michael Chertoff are leading this initiative. Ralph Boyd, Jr. has met with dozens of individuals and civil rights organizations to discuss our commitment to enforcing voting rights and ensuring voter participation.

§ Designating Federal Prosecutors to Address Voting Matters. The Attorney General directed all 94 U.S. Attorneys to designate a District Election Officer responsible for our efforts on Election Day. These officers are experienced Assistant U.S. Attorneys with knowledge of the laws that:

ü prohibit racial discrimination and voter intimidation;

ü govern absentee voting for uniformed and overseas citizens, the elderly, and citizens with disabilities; and,

ü protect the voting rights of minority language Americans.

§ District Election Officers. These officers are election experts in their Districts, and implement the Department of Justice's response to election fraud and campaign financing crimes under the overall supervision of the Public Integrity Section pursuant to 9 U.S.A.M. 85.210.

ü Terms of a District Election Officer last through the end of July in the next even-numbered year.

ü District Election Officers met with senior election administrators within their Districts to offer federal assistance in the investigation and prosecution of electoral crimes that occur during the forthcoming federal election.

###
http://www.usdoj.gov/opa/pr/2002/November/02_at_641.htm

Time to pay the rent!

:hi:

[eomer]

15. A discrepancy that is outside the MOE

is proof, statistically speaking, that sampling bias, official count bias or both are present.

Once that is proven, as is the case presently, then the next step is to investigate to determine whether the discrepancy stems from the sampling process or from the official count being wrong.

That's where we are at now and the next step is to do the investigation.

[Febble]

17. Exactly

I couldn't agree more.

[eomer]

21. I'm sure you're aware that's the main premise of the USCountVotes paper

The end of their Introduction:

We consider here the three possible explanations for a discrepancy between the official vote count and exit
polls:
1. Statistical sampling error – or chance
2. Inaccurate exit polls – Kerry supporters responded in greater numbers than Bush supporters.
3. Inaccurate election – the voters’ intent was not accurately recorded or counted.

We agree with Edison/Mitofsky that the first possible cause, random statistical sampling error, can be ruled
out. The second possible cause, that inaccurate exit polls were biased towards Kerry, is a hypothesis that is
compelling only if one dismisses the third, that official election results may have been distorted³.

And from their Summary:

The Edison/Mitofsky report states (p.12), “We need to do more investigation into the causes of the statistical
skew in the exit poll data for the general election.” USCountVotes agrees, and we suggest that that
investigation extend to the official vote count tallies. In this context, USCountVotes affirms our mission to
create and analyze a database containing precinct-level election results for the entire United States in order
to do a thorough mathematical analysis of the 2004 election results.

http://uscountvotes.org/ucvAnalysis/US/USCountVotes_Re_Mitofsky-Edison.pdf

Just wanted to point out that this is something that Febble, USCountVotes and maybe even Edison/Mitofsky agree on.

[Febble]

25. We do indeed.

And it was why, until recently, I was a member of the USCV team.

[TruthIsAll]

18. Febble, I could show you a thousand instances where the numbers don't add

In fact, I already have.

Now you resort to that cop-out: "But it's not proof".

Taken in context with all the other confirming evidence, it is just a small part of the total fraud mosaic.

Of course, we already have the thousands of documented cases of voter intimidation and touchscreen vote switching: 99%+ from Kerry to Bush. Will you once again resort to "But it's not proof"?

Where is a shred of rBr proof?
We anxiously await the UCSV response to the Febble function.

[Febble]

20. I didn't, in that post, say "it's not proof"

I asked you a question. What do you mean by the MoE?

Confirming evidence is fine. It's what we need. I just don't think we need circular arguments along the lines of:

If this is fact, therefore fraud.

It is a fact, therefore fraud.

Because you could just as easily argue:

This is not a fact: it results from sampling bias

It results from sampling bias, therefore it is not a fact.

If you think there was fraud because you have supporting evidence, fine. But your claim in your post that these are all the "facts" you need.

Which completely begs the question as to whether or not the facts are facts.

You might also like to address my question regarding the MoE.

[TruthIsAll]

23. Febble, the MoE stands by itself.

MoE = 1/sqrt(n), where n= sample-size.

MoE =1/sqrt(3168) = 1.77%

Assuming the poll was a legitimate random sample, as E-M stipulates, that is the MoE.

Since according to the 13047 exit poll respondents, Kerry won 57% of the new voters. That will be correct to within +/1.77% 95% of the time.

And the probability is that Kerry had greater than 55.2%.
Of course, you would rather believe it was due to rBr and faulty polling.

What about rKr?
Have you considered that as a hypothesis?
Of course not.

Oh, by the way, you never responded to a recent post of mine.
Here is what MP wrote before the election concerning reluctant responders:

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x365678

[Febble]

26. Oh, I know how to calculate an MoE

I am asking you what you think it means.

[TruthIsAll]

28. You are starting to sound a lot like Saywer. Oh, here he comes...

I gave you your answer.

[Febble]

31. No you didn't.

You gave me a formula.

A formula is meaningless unless you know what it means.

I want to know what you think it means. Unless you can tell me, I will assume you don't know.

And if you don't know, you had better look it up, because most of your arguments depend on an understanding of what the MoE actually is.

How it is calculated is not at issue.

[Sawyer]

27. "Assuming the poll was a legitimate random sample, as E-M stipulates,"

That is precisely what E-M does not stipulate.

[Febble]

32. Exactly

Thanks Sawyer!

The debate is essentially about whether it was a random sample.

The MoE is entirely irrelevant.

[TruthIsAll]

41. Febble, I'm reaaly NOT surprised you would say that..

I though we were way passed this.

Now you question whether it was indeed a random sample?

Well, to be precise, it was a randomly-selected sample.
Of 13047 voters.

Do you really want to dispute Mitofsky on this?

I direct you to the bottom of the screen.

[autorank]

42. The replies stop. TIA in a TECHNICAL KNOCK-OUT.

[Sawyer]

43. autorank, it would help if the cheerleaders

actually understood the rules of the game.

E-M specifically, unequivocally, stipulated that his sampling was not "legitimally random" when he stated his RBR hypothesis. TIA can claim otherwise until cows come home, it won't change that fact.

[autorank]

44. Too late Sawyer. Go tell people DOJ doesn't care about election fraud.

The results are official! Take your objections to the WBC (World Bullshit Council).

[TruthIsAll]

46. Mitofsky said it at 12:22am - before he massaged the poll to match Bush.

Its right there at the bottom of the screen in black and white.

12:22am
1.0% MoE
Randomly-selected sample of 13047 voters.

Who should we believe?
You or our lying eyes?

Sawyer, have we met before?

Your voice sounds very familiar.

[FULL_METAL_HAT]

29. A Few Neutral Questions for Clarity: MOE & SB simulations

Despite the political capitol of both Larry and Curly, it seems the point of argument keeps coming back to Moe... ;)

I had a thought experiment to ask Febble and TIA to consider, and apologies if this has been covered elsewhere:

Sampling bias
When data is influence in one way or another so that the data no longer represents the entire population.

Given there is "Sampling bias" which Febble notes is distinctly different from "Margin of Error", maybe we can simulate the effects of various levels of sampling bias? Especially in relation to the "Voted for in 2000". Can we get a little reply on "confidence interval/level" as well?

#1) What could be the MAXIMUM "sampling bias" that typical, professional statisticians could allow? This answer would allow for simulation of various levels of biased response to compare to the "official WEIGHTED (aka "professionally vetted) results"
a) up to 1%
b) up to 5%
c) up to 10%
d) up to 25%
e) more?

#2) Does the "Margin of Error" calculations INCLUDE perceived threats of "Sampling Bias", or does it assume a totally bias-free sampling? If it does, is there a standard value?

#3) There are other factors involved in the production of these statistics like the "confidence level". Have these factors been brought into the argument to help clarify the factors that are static and those that are variable?

#4) Simply, what actual level of "Sambling Bias" would be needed to achieve the "official results"?

I've always found that quantifying points is always helpful to firm up understandings, even if the quantifications are typically "understood" and seeming needless for an argument at hand. Removing "a lot" from conversation and replacing it with concrete numbers seems to help.

If anyone could chime in on this, please do!!

BTW, here's a good primer on the concepts that I think sometimes are getting crossed. I'd love to hear any commentary on the inter-relational importance, relevance or irrelevance of any of these other "concepts", most especially "Sampling error", "Confidence level (or level of confidence)", and "Confidence interval".

B. Sampling Concepts
Population/Target population: This is any complete, or the theoretically specified aggregation of study elements. It is usually the ideal population or universe to which research results are to be generalized. For example, all adult population of the U.S.

Survey population: This is an operational definition of the target population; that is target population with explicit exclusions-for example the population accessible, excluding those outside the country.

Element (similar to unit of analysis): This is that unit about which information is collected and that provides the basis of analysis. In survey research, elements are people or certain types of people.

Sampling unit: This is that element or set of elements considered for selection in some stage of sampling (same as the elements, in a simple single-stage sample). In a multi-stage sample, the sampling unit could be blocks, households, and individuals within the households.

Sampling frame: This is the actual list of sampling units from which the sample, or some stage of the sample, is selected. It is simply a list of the study population.

Sample design: This refers to a set of rules or procedures that specify how a sample is to be selected. This can either be probability or non-probability.

Sample size: The number of elements in the obtained sample.

Sampling error: This is the degree of error to be expected for a given sample design or the difference between the sample mean and the population mean.

Sampling bias: This refers to the notion that those selected are not "typical" or "representative" of the larger populations that have been chosen from.

Margin of error refers to the precision needed by the researcher. A margin of error of 5 percent means that the actual findings could vary by as much as 5 points either positively or negatively.

Confidence level (or level of confidence) is a statement of how often you could expect to find similar results if the survey were to be repeated, or the degree of certainty of obtaining the same results. It often informs about how often the findings will fall outside the margin of error.

Confidence interval is a range in which we are fairly certain that the population value lies.

Parameter-the summary description of a given variable in a population.

Statistic-the summary description of a given variable in a sample.

Statistical inference:The process of reasoning by which information about a population is extracted from sample data.

from http://www.uic.edu/classes/socw/socw560/Sampling1.htm

[TruthIsAll]

33. Febble and Sawyer might be interested in this.

http://www.democraticunderground.com/discuss/duboard.php?az=view_all&address=203x366149

[TruthIsAll]

34. Want to see how it works? You want an MoE? Design your own Sample -size.

http://www.democraticunderground.com/discuss/duboard.php?az=show_topic&forum=203&topic_id=356832

[Febble]

38. I'm increasingly intrigued

by your reluctance to explain what you mean by a margin of error.

I have defined it myself more than once in these threads, but you have yet to do so in response to any of my questions.

If all you can do is provide me with links and formulae, then I begin to suspect that you do know know what the margin of error means.

If you do, all you have to do is tell me. Not how to calculate it, but what it means.

And while you are at it, perhaps you would like to tell me the difference between sampling error and sampling bias.

If you will do that, I will get off your case.

[LightningFlash]

51. Actually, I can do that easily. But it refutes your argument...

Are you certain then you want to see what MOE represents?

This data set is properly sampled and weighted to reflect the views of all voters nationwide in 2002, but it should not be used to analyze voting behavior in individual states. As in previous years, VNS in 2002 developed a national sample and, in a separate process, created statewide samples with different questionnaires and weighting procedures in each state. Nearly all of those state-level interviews are not included in this data set and no valid analysis of individual states can be made without them. Do not select respondents in a specific state and analyze their responses separately; any attempt to do so could easily result in invalid and misleading results.

http://www.ropercenter.uconn.edu/usvns2002_2.html#states

ftp://ropercenter.uconn.edu/United_States%5CVNS/USVNS2000-NATELEC/version2/usvns2000-natelec.pdf

Full ranged tables and accumulated data, from Mitofsky's own VNS webservers.

"And no valid analysis of individual states can be made without them."

Apparently the motto is, if it doesn't allign with reluctant responder theory it isn't needed period. Response bias for a fact happened. Uniform computer level accurate response bias? Not a chance. And not in all of these elections.

[Febble]

37. I love Moe

He gets me drunk....

Yes, we can simulate sampling bias, and it's what I did in my model. We can't measure it though, and we can't predict it. We can minimise it with good survey design, and with high response rates. But the lower the response rates, the less justified we are in assuming our sample is random. Yes we can check if we think the refusers were female, ethnic minority, of certain age, and do a little weighting to compensate using the voting patterns of those demographics in our sample, but what we don't know is why they refused. And the fact that they refused makes them a different group from the group who agreed to take part in the study.

The margin of error does not include sampling bias. It includes a factor for clustered sampling, which means that instead of taking a random sample of voters from each state, the polls take a random sample of precincts, and then a random sample of voters from those precincts. This increases the MoE, but the sample sizes are so large, the MoE remains small. There is reason to believe, from the E-M report, that sampling error was even larger than allowed for, possibly due to inadvertent clustered sampling at the precincts (if a group of Kerry voters comes along during a thunderstorm for example, you might get low coverage of Kerry voters at that precinct) but this does not explain bias.

Bias is entirely to do with the non-random nature of the sampling, and thunderstorms are random. They'd just increase the sampling error.

Not much sampling bias is required to explain the results. If the ratio of Kerry voters to Bush voters sampled was 1.12 to 1, that would account for the discrepancy. It's only slight, and both groups of voters were "reluctant" - response rates were only 53%. The hypothesis is simply that Bush voters may have been slightly more reluctant. 1.12 to 1 is not a large ratio for such a low mean response rate. It was not enough for any state to be "called" erroneously. It was simply apparent in the difference between the early predictions and the late predictions, and in the within-precinct error rate (WPE), which would reflect voter sampling (but would also reflect precinct level fraud, or differential vote spoilage, which also is well-documented in past elections.)

I probably haven't addressed all your questions, but the important take-home-message is that sampling BIAS has nothing to do with sampling ERROR. Sampling error can be minimised by large samples. Sampling bias can only be minimised by very high response rates and absolutely random protocol. Neither were true of the E-M exit poll, and of interest is their finding that bias was greatest where good protocol was most difficult to adhere to. This is the support for their conclusion that a differential response rate of 1.12:1 amongst Kerry versus Bush voters was the most likely explanation for the discrepancy.

So it is not true to say there hypothesis is unsupported. It is not proven, but it is certainly not unsupported.

[mopinko]

11. i nominate Liepothesis

as the greatest new word of the new century. my only regret is the overwhelming need for it.

[Name removed]

36. Deleted message

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

[FULL_METAL_HAT]

40. To be polite, if the news came out with evidence would you still say

"the election is over"?

It's very true that November 2, 2004 is 192 days ago but if there were some kind of "problem" found, would it matter if it was 1, 19, or 1,920 days ago?

I know you're just trying to help, but at what point is the "statute of limitations" tip over for you? Of course we're not talking about a typical legal definition of such a limitation, since regular elections are constrained by them, but given whats being alleged in this thread, don't you think that if there were results from this discussion that the time passed would be less important than the new evidence?

[bleever]

45. As a corollary to the original post, I'd like to thank all the defenders

of E-M for their heavy lifting, and the lengths they will go to in order to find a leg for this fiasco to stand on.

While acknowledging the good faith (in my opinion) of febble and anyone who seeks to keep a focus on objective facts and their ensuing implications, I see more than a hint of knee-jerk naysaying raising its head here, which usually indicates that someone is nervous that the zombie of the stolen election just won't stay dead.

[autorank]

48. This is significant coming from you since you were quite gracious

to "the pollster(s) who came to dinner."

I like that..."someone is nervous that the zombie of the stolen election just won't stay dead.

What are the royalties if I want it for my sign line?

[kster]

47. kick.nt

[LightningFlash]

49. Seems to me this ends the discussion.

If you read and check everything below, it comes out pretty clear I'd say.

http://uscountvotes.org/ucvAnalysis/US/exit-polls/USCV_exit_poll_simulations.pdf


One could probably consider this argument moot. There is no integrity to the discussion that a bush response bias can outpace its own data, to the point that the original data Edison-Mitofsky holds no longer is valid in itself. It's what I call double reasoning. Plus on top of that, there's no way that such a sampling on purpose or mistake could be orchestrated at this level. At least not in multiple elections and all using the same affect.

http://www.votefraud.org/News/2000/7/071800.html

I think the theory is effectively dead. Was there response bias, most absolutely. Just none so phenomenonal it somehow lines up exactly with Mitofsky's exit poll statistics. But my real question is, why hasn't Mitofsky given all of the precinct data over to the PHDs and statisticians? Why has he not released it to USCountVotes and the entire team? What is his motive or operandi, that while anyone who comes out of nowhere like ESI and even Liddle can view some of the real data....yet fully hard working and credentialed master PHD scientists like Kathy Dopp, Stephen Freeman, and Ron Baiman can not?

What exactly does he hide? It's needed to be proven Bush won, isn't it? How will anyone know if all the data isn't checked and released? How come Bluementhal and others got to see the data in 2002, 2000, and other elections before when the "responder bias" was documented, yet nobody within the scientific community even got a glimpse?

Wonders will never cease in this day and age.

[Joe for Clark]

50. Good for you Truth.

Keep posting your message.

Bush did lose this one, the more people voicing it, the sooner it ends. It will, sooner or later his popularity ratings will catch up with the shenanigans. Americans can be fical that way, I guess.

Anyway, I think your numbers are pretty good. Don't know who you are, but clearly you do understand the probabilities here. Karma is a bitch sometimes. And I sure hope Mr Bush finds out this "reality" pretty soon.

Keep posting. I think you have the right tact on this.

Joe

[LightningFlash]

52. It should be brought to full attention...

That there certainly for a fact was response bias. But in any of the calculations was there uniform response bias of every degree?

The answer is no, especially on the computer-accurate level that is being shown here. To be quite exact, it's happened in 4 seperate elections or more.

For it to be that accurate and pristine, a computer must be responsible.

ftp://ropercenter.uconn.edu/United_States%5CVNS/USVNS2000-NATELEC/version2/usvns2000-natelec.pdf

This data set is properly sampled and weighted to reflect the views of all voters nationwide in 2002, but it should not be used to analyze voting behavior in individual states. As in previous years, VNS in 2002 developed a national sample and, in a separate process, created statewide samples with different questionnaires and weighting procedures in each state. Nearly all of those state-level interviews are not included in this data set and no valid analysis of individual states can be made without them. Do not select respondents in a specific state and analyze their responses separately; any attempt to do so could easily result in invalid and misleading results.

http://www.ropercenter.uconn.edu/usvns2002_2.html#states

Of course, Mitofsky repeatedly stated VNS never announced to the news that their database crashed then too. It was conveniently omitted, but the evidence is now in full display during the same exit-poll.

[kster]

53. Kick to the evening crowd.nt