TruthIsAll
09-09-2009, 11:40 PM
BIG 43/37 CONTRADICTION HERE
FEBBLE
a) A REQUIRED ADJUSTMENT....
"The fact that Bush's retrospective margin was not much inflated in the unadjusted poll, is if anything, evidence that ADJUSTMENT WAS REQUIRED - that Bush 2004 voters had been UNDERSAMPLED.
or
OTOH
b) AN OVERSTATEMENT....
"there is no clear limit to what percentage of 2004 voters could have claimed to vote for Bush in 2000. The apparent OVERSTATEMENT is in line with overstatements of previous winners' vote shares in other exit polls"
FEBBLE
I share with TIA his indignation at this systematic disenfranchisement of largely Democratic voters. As for his second point - TIA read it in a book somewhere. It is not something that can, or should, be generalised to exit polls. There are many sources of evidence, including direct experimental evidence that in exit polls, Democratic voters tend to be over-sampled.
3) Exit poll response was high in Bush states and low in Kerry states. Here's GRAPHIC MATHEMATICAL PROOF USING LINEAR REGRESSION ANALYSIS:Overall response rates are not a proxy for response bias. Response bias occurs when the response rates for one set of voters differ from the response rate for the other set of voters, whether the two rates are 15% and 20% or 60& and 80%. Moreover, selection bias will not show up in response rates - and may even be associated with higher response rates. If more willing voters are being selected, completion rate will go up.
*****************************************************************************
FEBBLE: That people recall their previous votes correctly (there is excellent evidence that they do not, and that they tend to misreport having voted for the previous winner)
******************************************************************************
TIA: If you consider that the FINAL National Exit poll is ALWAYS MATCHED TO THE RECORDED VOTE, WHICH MEANS THAT THE WEIGHTS ARE FUDGED TO MATCH THE VOTE. I CALL IT FUDGING; YOU CALL IT VOTER MISREPRESENTATION. Witness the 2004 Final NEP, which Bush won by 51-48%: 43% of 2004 voters were Bush voters and 37% were Gore voters? This was a FUDGE necessary to match the recorded vote. In the 12:22am poll, which Kerry won by 51-48%, the mix was 41 Bush/39 Gore.
Mathematically, the MAXIMUM Bush 2000 representation weighting was 39.8%, which is the ratio of Bush 2000 voters still alive n 2004 by the 122.3mm who voted.Do you want MORE of this evidence? Look at the 2006 NEP. The 7pm poll had the 2004 weighting as 45 Bush/46 Gore. The FINAL had it 49 Bush/43 Gore/8 Other. Where did the OTHER 8% come from?
Third-party 2004 voters comprised 1% of the vote. Where did the excess 7% come from? Kerry. Here’s why:Are we to believe Bush voters outnumbered Kerry voters by 6%? Even if you believe the 2004 FINAL, which we have proven bogus, the Bush margin was 3%. In reality it should have been 50 Kerry-47 Bush, after deducting 1% for voter mortality overthe 2-year period. Now 50% = 43% + 7%. There is your Kerry vote.
I really can't be bothered to explain this to TIA again. He needs to read Mark Lindeman's paper:
http://inside.bard.edu/~lindeman/too-many.pdf
Febble (1000+ posts) Sat Nov-18-06 05:59 PMResponse to Reply #3940. Response to TIA
1. I do not have a fixed agenda. This false.
2. I am quite aware that uncounted Democratic votes may contribute to a discrepancy in the exit polls. Indeed, in my work for Mitofsky I specifically looked for evidence of this, and found some that, while not conclusive, is suggestive that it may have played a role. The measure of precinct level discrepancy I used, developed together with Mark Lindeman, meant that it was more sensitive than the traditional measures to bias in extreme precincts, and this may have been why my analysis was able to detect something.
The finding was that in urban precincts, particularly largely black urban precincts, those precincts in which older technology was used (levers and punch cards) the discrepancy was greater then where digital technology was used. It was a small effect, and may have been confounded by other factors, but it was of interest. However, because uncounted Democratic votes tend to be concentrated in extreme Democratic precincts, it is unlikely that even large numbers of residual votes would have a large impact on the exit poll, simply because extreme Democratic precincts are not heavily represented in the precinct sample.
I agree with TIA, and with Greg Palast (though I would still like details of where he gets his numbers) that uncounted Democratic votes, together with voter suppression, cost the Democrats many votes in each election, and, indeed, cost Gore the presidency. But the exit polls are unlikely to reflect this loss except at precinct level, and discrepancies between overall exit poll estimates and counted results are unlikely to index, or even reflect, the magnitude of this problem.
3. What I said is that TIA does not factor in the probability that his own assumptions are in error.
4. As many others have argued, translating a generic ballot into seats is difficult, given the nature of congressional district boundaries, and, indeed, the number of seats up for election. Nonetheless several pundits attempted it. None that I was following came up with TIA's optimistic projections. Larry Sabato pretty well nailed it.
**********************************************************************************
5. As TIA knows, ****retrospective inflation**** of the winners margin is a routine phenomenon in exit polls, regardless of the popularity of the incumbent. It even happened with Nixon, who was not even president by then. These facts are not on TIA's side.mom cat (1000+ posts) Sun Nov-19-06 03:33 AMResponse to Reply #4043.
************************************************************
TIA response to FEBBLE and other interested DUers.
DUers may not be familiar with these facts. Given: Bush 2000 vote: B= 50.45mm 2004 total vote: N= 122.3mmAnnual U.S. mortality rate: R= 0.87%Final NEP: W= Bush 2000 voters/N = 43%Calculate: 1 The number of Bush 2000 voters (D) who died prior to 2004: D = 4*R*B2 The maximum number (X) of Bush 2000 voters who could vote in 2004: X = B–D3 The maximum weighting (W) of Bush 2000 voters who voted in 2004: W = X/N
If the calculated W does not equal the Final NEP W of 43%, which one is correct? What does this tell us about the "false recall" theory, that a significant percentage of exit poll respondents falsely report who they voted for in the prior election? What does this tell us about the Reluctant Bush Responder (rBr) theory, that for various reasons Democrats are more likely to be exit-polled than Republicans?
Febble (1000+ posts) Sun Nov-19-06 08:10 AMResponse to Reply #4344. Well, TIA knows the answer to this one or if he doesn't, it's not for want of being told.If the calculated W doesn't equal the Final NEP W, then the probability is that, as usual, people ****misreported**** their previous vote in favor of the winner.
*****************************************************************************
This "retrospective inflation" of the winner's margin is evident, as Mark Lindeman showed here:
http://inside.bard.edu/~lindeman/too-many.pdf
******************************************************
In every single presidential exit poll since 1976, and applies whether the previous winner is Republican, Democrat, running, losing, or even still in office (Nixon). The fact that Bush's retrospective margin was not much inflated in the unadjusted poll, is if anything, evidence that *****adjustment was required - that Bush 2004 voters had been undersampled*****.
The adjusted margin is more inline with the kind of retrospective margin inflation expected. TIA makes exactly the same error that O'Dell and Simon just made in their recent "Landslide Denied" paper.So that is the answer to TIA's question.
*******************************************************
It tells us that ************Kerry voters were probably sampled at a higher rate than Republicans (as is evident from other analyses)**** and ****that people misreported their previous vote in favor of the winner**** in the same kinds of proportions as would be expected, given the pattern we observe in the historical data.
*******************************************************************
Febble, once again you stubbornly avoid the FACTS and the IMPLICATIONS of the FACTS.First of all, you didn’t do the math. Because if you did, you would know that at MAXIMUM, only 48.7mm Bush 2000 voters could have voted in 2004. And therefore the MAXIMUM Bush weighting was 39.8% = 48.7/122.3%. The 43% weighting was IMPOSSIBLE. This is a MATHEMATICAL FACT.
Therefore, if you were willing to accept this MATHEMATICAL FACT, you would have to agree that the Final NEP 43% Bush weighting was a FICTIONAL ARTIFICE AND NOT A SAMPLED RESULT. IT WAS AN ARBITRARY FUDGE WHICH WAS REQUIRED TO FORCE THE FINAL NEP TO MATCH THE BUSH RECORDED VOTE. Even you have agreed that the Final NEP is ALWAYS matched to the recorded vote.Why, for heaven's sake, is it not yet clear to YOU that matching to the recorded vote count ONLY makes sense IF the vote count is ACCURATE and there is ZERO FRAUD? This should be clear to everyone, even a third grader who cheats on his arithmetic test.
THEREFORE, THE FINAL NEP IS A FRAUD. When will you accept the FACT that the RECORDED 2004 VOTE was BOGUS and that BushCo used massive FRAUD to STEAL the election? In fact, BushCo has successfully stolen EVERY election since 2000 -except for 2006. They were stopped in 2006 only because of the Democratic TSUNAMI.The reality-based community must therefore conclude that your “retrospective” exit poll argument HOLDS NO WATER AND IS JUST ANOTHER RUSE TO DEFLECT FROM THE FINAL NEP'S EGREGIOUS MATCHING TO FRAUDULENT, MISCOUNTED VOTES.
If you were a true analytical investigator, you would not employ TWISTED LOGIC TO DEFEND THE FINAL NEP AND CONSTANTLY CRITICIZE THE ACCURACY OF PRE-ELECTION AND PRELIMINARY EXIT POLLS WHICH CLEARLY POINT TO FRAUD. If you were a true analyst, you would stick to the mathematical facts.
OnTheOtherHand (1000+ posts) Sun Nov-19-06 09:28 PMResponse to Reply #4748. why this is wrong
*************************************************************************
Short version: as has been explained many times, there is no reason to expect exit poll respondents to report their past votes accurately. Therefore, while it may be true that no more than 39.8% of 2004 voters could have voted for Bush in 2000, there is no clear limit to what percentage of 2004 voters could have claimed to vote for Bush in 2000. The apparent overstatement is in line with overstatements of previous winners' vote shares in other exit polls.It must be frustrating to believe that the entire reality-based community should have looked at the exit poll tabulations in November 2004 and realized that Kerry won the election. All those people blithely writing as if Bush won -- are they innumerate? careless? craven?Maybe they just don't take polls as literally as, apparently, you do. Skepticism about polls could be reality-based, don't you suppose?
Graphical Election Myths
http://www.RichardCharnin.com/DebunkingElectionMyths.pdf
***********************************************
FEBBLE
a) A REQUIRED ADJUSTMENT....
"The fact that Bush's retrospective margin was not much inflated in the unadjusted poll, is if anything, evidence that ADJUSTMENT WAS REQUIRED - that Bush 2004 voters had been UNDERSAMPLED.
or
OTOH
b) AN OVERSTATEMENT....
"there is no clear limit to what percentage of 2004 voters could have claimed to vote for Bush in 2000. The apparent OVERSTATEMENT is in line with overstatements of previous winners' vote shares in other exit polls"
FEBBLE
I share with TIA his indignation at this systematic disenfranchisement of largely Democratic voters. As for his second point - TIA read it in a book somewhere. It is not something that can, or should, be generalised to exit polls. There are many sources of evidence, including direct experimental evidence that in exit polls, Democratic voters tend to be over-sampled.
3) Exit poll response was high in Bush states and low in Kerry states. Here's GRAPHIC MATHEMATICAL PROOF USING LINEAR REGRESSION ANALYSIS:Overall response rates are not a proxy for response bias. Response bias occurs when the response rates for one set of voters differ from the response rate for the other set of voters, whether the two rates are 15% and 20% or 60& and 80%. Moreover, selection bias will not show up in response rates - and may even be associated with higher response rates. If more willing voters are being selected, completion rate will go up.
*****************************************************************************
FEBBLE: That people recall their previous votes correctly (there is excellent evidence that they do not, and that they tend to misreport having voted for the previous winner)
******************************************************************************
TIA: If you consider that the FINAL National Exit poll is ALWAYS MATCHED TO THE RECORDED VOTE, WHICH MEANS THAT THE WEIGHTS ARE FUDGED TO MATCH THE VOTE. I CALL IT FUDGING; YOU CALL IT VOTER MISREPRESENTATION. Witness the 2004 Final NEP, which Bush won by 51-48%: 43% of 2004 voters were Bush voters and 37% were Gore voters? This was a FUDGE necessary to match the recorded vote. In the 12:22am poll, which Kerry won by 51-48%, the mix was 41 Bush/39 Gore.
Mathematically, the MAXIMUM Bush 2000 representation weighting was 39.8%, which is the ratio of Bush 2000 voters still alive n 2004 by the 122.3mm who voted.Do you want MORE of this evidence? Look at the 2006 NEP. The 7pm poll had the 2004 weighting as 45 Bush/46 Gore. The FINAL had it 49 Bush/43 Gore/8 Other. Where did the OTHER 8% come from?
Third-party 2004 voters comprised 1% of the vote. Where did the excess 7% come from? Kerry. Here’s why:Are we to believe Bush voters outnumbered Kerry voters by 6%? Even if you believe the 2004 FINAL, which we have proven bogus, the Bush margin was 3%. In reality it should have been 50 Kerry-47 Bush, after deducting 1% for voter mortality overthe 2-year period. Now 50% = 43% + 7%. There is your Kerry vote.
I really can't be bothered to explain this to TIA again. He needs to read Mark Lindeman's paper:
http://inside.bard.edu/~lindeman/too-many.pdf
Febble (1000+ posts) Sat Nov-18-06 05:59 PMResponse to Reply #3940. Response to TIA
1. I do not have a fixed agenda. This false.
2. I am quite aware that uncounted Democratic votes may contribute to a discrepancy in the exit polls. Indeed, in my work for Mitofsky I specifically looked for evidence of this, and found some that, while not conclusive, is suggestive that it may have played a role. The measure of precinct level discrepancy I used, developed together with Mark Lindeman, meant that it was more sensitive than the traditional measures to bias in extreme precincts, and this may have been why my analysis was able to detect something.
The finding was that in urban precincts, particularly largely black urban precincts, those precincts in which older technology was used (levers and punch cards) the discrepancy was greater then where digital technology was used. It was a small effect, and may have been confounded by other factors, but it was of interest. However, because uncounted Democratic votes tend to be concentrated in extreme Democratic precincts, it is unlikely that even large numbers of residual votes would have a large impact on the exit poll, simply because extreme Democratic precincts are not heavily represented in the precinct sample.
I agree with TIA, and with Greg Palast (though I would still like details of where he gets his numbers) that uncounted Democratic votes, together with voter suppression, cost the Democrats many votes in each election, and, indeed, cost Gore the presidency. But the exit polls are unlikely to reflect this loss except at precinct level, and discrepancies between overall exit poll estimates and counted results are unlikely to index, or even reflect, the magnitude of this problem.
3. What I said is that TIA does not factor in the probability that his own assumptions are in error.
4. As many others have argued, translating a generic ballot into seats is difficult, given the nature of congressional district boundaries, and, indeed, the number of seats up for election. Nonetheless several pundits attempted it. None that I was following came up with TIA's optimistic projections. Larry Sabato pretty well nailed it.
**********************************************************************************
5. As TIA knows, ****retrospective inflation**** of the winners margin is a routine phenomenon in exit polls, regardless of the popularity of the incumbent. It even happened with Nixon, who was not even president by then. These facts are not on TIA's side.mom cat (1000+ posts) Sun Nov-19-06 03:33 AMResponse to Reply #4043.
************************************************************
TIA response to FEBBLE and other interested DUers.
DUers may not be familiar with these facts. Given: Bush 2000 vote: B= 50.45mm 2004 total vote: N= 122.3mmAnnual U.S. mortality rate: R= 0.87%Final NEP: W= Bush 2000 voters/N = 43%Calculate: 1 The number of Bush 2000 voters (D) who died prior to 2004: D = 4*R*B2 The maximum number (X) of Bush 2000 voters who could vote in 2004: X = B–D3 The maximum weighting (W) of Bush 2000 voters who voted in 2004: W = X/N
If the calculated W does not equal the Final NEP W of 43%, which one is correct? What does this tell us about the "false recall" theory, that a significant percentage of exit poll respondents falsely report who they voted for in the prior election? What does this tell us about the Reluctant Bush Responder (rBr) theory, that for various reasons Democrats are more likely to be exit-polled than Republicans?
Febble (1000+ posts) Sun Nov-19-06 08:10 AMResponse to Reply #4344. Well, TIA knows the answer to this one or if he doesn't, it's not for want of being told.If the calculated W doesn't equal the Final NEP W, then the probability is that, as usual, people ****misreported**** their previous vote in favor of the winner.
*****************************************************************************
This "retrospective inflation" of the winner's margin is evident, as Mark Lindeman showed here:
http://inside.bard.edu/~lindeman/too-many.pdf
******************************************************
In every single presidential exit poll since 1976, and applies whether the previous winner is Republican, Democrat, running, losing, or even still in office (Nixon). The fact that Bush's retrospective margin was not much inflated in the unadjusted poll, is if anything, evidence that *****adjustment was required - that Bush 2004 voters had been undersampled*****.
The adjusted margin is more inline with the kind of retrospective margin inflation expected. TIA makes exactly the same error that O'Dell and Simon just made in their recent "Landslide Denied" paper.So that is the answer to TIA's question.
*******************************************************
It tells us that ************Kerry voters were probably sampled at a higher rate than Republicans (as is evident from other analyses)**** and ****that people misreported their previous vote in favor of the winner**** in the same kinds of proportions as would be expected, given the pattern we observe in the historical data.
*******************************************************************
Febble, once again you stubbornly avoid the FACTS and the IMPLICATIONS of the FACTS.First of all, you didn’t do the math. Because if you did, you would know that at MAXIMUM, only 48.7mm Bush 2000 voters could have voted in 2004. And therefore the MAXIMUM Bush weighting was 39.8% = 48.7/122.3%. The 43% weighting was IMPOSSIBLE. This is a MATHEMATICAL FACT.
Therefore, if you were willing to accept this MATHEMATICAL FACT, you would have to agree that the Final NEP 43% Bush weighting was a FICTIONAL ARTIFICE AND NOT A SAMPLED RESULT. IT WAS AN ARBITRARY FUDGE WHICH WAS REQUIRED TO FORCE THE FINAL NEP TO MATCH THE BUSH RECORDED VOTE. Even you have agreed that the Final NEP is ALWAYS matched to the recorded vote.Why, for heaven's sake, is it not yet clear to YOU that matching to the recorded vote count ONLY makes sense IF the vote count is ACCURATE and there is ZERO FRAUD? This should be clear to everyone, even a third grader who cheats on his arithmetic test.
THEREFORE, THE FINAL NEP IS A FRAUD. When will you accept the FACT that the RECORDED 2004 VOTE was BOGUS and that BushCo used massive FRAUD to STEAL the election? In fact, BushCo has successfully stolen EVERY election since 2000 -except for 2006. They were stopped in 2006 only because of the Democratic TSUNAMI.The reality-based community must therefore conclude that your “retrospective” exit poll argument HOLDS NO WATER AND IS JUST ANOTHER RUSE TO DEFLECT FROM THE FINAL NEP'S EGREGIOUS MATCHING TO FRAUDULENT, MISCOUNTED VOTES.
If you were a true analytical investigator, you would not employ TWISTED LOGIC TO DEFEND THE FINAL NEP AND CONSTANTLY CRITICIZE THE ACCURACY OF PRE-ELECTION AND PRELIMINARY EXIT POLLS WHICH CLEARLY POINT TO FRAUD. If you were a true analyst, you would stick to the mathematical facts.
OnTheOtherHand (1000+ posts) Sun Nov-19-06 09:28 PMResponse to Reply #4748. why this is wrong
*************************************************************************
Short version: as has been explained many times, there is no reason to expect exit poll respondents to report their past votes accurately. Therefore, while it may be true that no more than 39.8% of 2004 voters could have voted for Bush in 2000, there is no clear limit to what percentage of 2004 voters could have claimed to vote for Bush in 2000. The apparent overstatement is in line with overstatements of previous winners' vote shares in other exit polls.It must be frustrating to believe that the entire reality-based community should have looked at the exit poll tabulations in November 2004 and realized that Kerry won the election. All those people blithely writing as if Bush won -- are they innumerate? careless? craven?Maybe they just don't take polls as literally as, apparently, you do. Skepticism about polls could be reality-based, don't you suppose?
Graphical Election Myths
http://www.RichardCharnin.com/DebunkingElectionMyths.pdf
***********************************************