TruthIsAll
01-06-2009, 10:49 AM
Jan. 6, 2009
This is a belated response since I just read the comments 30 minutes ago.
Thanks, auto, for your "defense" of my work.
http://www.smirkingchimp.com/thread/18513#comment-144297
_______________________________________________________________________
The Naysayer:
What is the math based on, though?
What I see in your "Urban Legend" piece is you working backwards from the notion that the actual vote tallies were incorrect, that the national and state exit polling raw numbers are more reliable than either the weighted polling numbers or the actual vote counts, and then you work the poll data in such a way that it proves the vote tallies must be wrong. Yeah, it's math, and I can see your work - but that doesn't mean I think it's good math.
Similarly, I look at TruthIsAll's work, and find that he must have done something similar with the 2004 results to come up with his "TrueVote" model. He then takes the estimates of what the turnout was expected to be this year, assumes that the high-end poll estimates of Obama's support in the final weekend were gospel, applies the "TrueVote" and comes up with a number that he claims were the actual votes cast, a large percentage of which has been "disappeared" because it doesn't match the figures on CNN. That's one hell of a lot of assuming, estimating and extrapolating, don't you think? The only actual number in any of that is the reported returns - everything else is smoke and mirrors.
There's a reason why "lies, damned lies, and statistics" is such a cliche.
_______
It could be that the purpose of your life is only to serve as a warning to others.
Submitted by Michael Sheridan on November 7, 2008 - 5:25pm.
________________________________________________________________________
My response:
Jan. 6, 2009
This is a belated response since I just read the comments 30 minutes ago. Your interpretation of the analysis is wrong.
There are LIES, DAMNED LIES and a solid Statistical ANALYSIS to expose the LIES.
The official vote count is 131.37 million, up 10.16m since election Day. Obama has won 59.2% of the late votes. As a result,his official vote share has climbed from 52.3to 52.9%. But the recorded vote is NEVER equal to the True Vote. Read on and you will see how some very basic statistical analysis indicates that his True Vote share is 57%.
Note: The Election Model (EM) projected that Obama's expected EV was 365.3 (based on the latest state polls adjusted for undecided voters). Obama had 365 EV. He won 5000 of 5000 Monte Carlo Electoral vote simulation trials - a 100% win probability. The model also projected he would win by 53.1% of the vote; he is at 52.9%. But the model was WRONG; it underestimated Obama's True Vote share by 4% (see the Election Calculator model).
Why did the EM underestimate Obama's True share?
1. The final state pre-election polls were LV (likely voter) polls which are subsets of RV polls. The LVs exclude many newly registered voters; RV polls include new registerd voters (Obama won 71% of new voters according to the National Exit Poll).
2. I projected 60% of undecided voters for Obama; typically, the challenger (Obama) gets 75-90%. In fact, that's the Election Calculator result based on National Exit Poll vote shares (probably conservative)and a FEASIBLE returning voter mix (the NEP returning Bush/Kerry 46/37% voter mix is impossible).
3. The Oct. 30 PEW RV poll had Obama leading by 52-36%. Allocate 50% of the 10% undecided and it's 57-41%. With a 60% UVA, it's 58-40%. It's 59.5-38.5% with a 75% UVA! So to say he had 56-57% is by no means a stretch.
4. Final pre-election state polls are conducted several days before the election; late undecideds and independents broke sharply on Election Day for Obama.
To summarize:
The Election Model used the latest state polling results, adjusted for allocation of undecided voters.
http://www.geocities.com/electionmodel/2008ElectionModel.htm
The Election Calculator used returning prior election voters and the National Exit Poll vote shares to a) forecast the election in May 2008 and b) determine the True Vote (post-election) based on the 2008 National Exit Poll vote shares.
http://www.geocities.com/electionmodel/2008ElectionAnalysisLinks.htm
They are two distinct models; you are confusing the methodologies.
First, consider these election FACTS:
http://www.progressiveindependent.com/dc/dcboard.php?az=show_topic&forum=120&topic_id=4055
To supplement Collin's Urban Legend analysis, see chapter 13:
http://www.geocities.com/electionmodel/FurtherConfirmationOfaKerryLandslide.htm
The True Vote Model is actually the Election Calculator (EC) model. I used it to analyze the 2004 election based on the Preliminary National Exit Poll vote shares. The Final NEP 43/37% Bush/Gore weights were mathematically impossible, so I revised the 2000 returning voter mix based on 2000 returning voters adjusted for voter mortality, uncounted votes and turnout in 2004. Kerry won the True Vote with 53+%.
In May 2008, the EC projected 2008 assuming 2004 voter turnout based on mortality, uncounted votes, etc. Obama was a 54% winner.
After the election, the EC applied 2008 NEP vote shares to a FEASIBLE (and PLAUSIBLE) returning 2004 voter mix. Obama was a 57% winner.
Both models used the best available data: pre-election state and national polls, national exit polls, voter mortality rates, U.S. 2000 and 2004 Census votes cast (includes uncounted votes), 2008 recorded votes and assumed 95% turnout of 2004 election voters.
_______
truthisall.net
Submitted by TruthIsAll on January 6, 2009 - 12:42pm.
Just in case it's still not clear why the Final NEP is bogus.
You seem to have a naive belief that the recorded vote count is sacrosanct. These facts should cause you question that notion.
1- The Final National Exit Poll is always forced to match the official, recorded vote count.
2-The Final 2004 NEP closely matched the 50.7-48.3% Bush recorded margin.
3-The Final 2008 NEP closely matched the 52.9-45.6% Obama recorded margin.
4-The number of returning voters from the prior election MUST BE LESS than the number who actually voted due to mortality and turnout.
5-To match the recorded vote, the 2004 and 2008 Final NEP both indicated that there were MORE returning Bush voters than were still alive.
6-In 2004, the Final NEP indicated 52.6 million returning Bush voters from 2000.
7-But Bush only had 50.5 million recorded votes and approximately 2.5m died.
8-In 2004, Kerry won the unadjusted state exit polls by 52-47% and the preliminary National Exit Poll (12:22am, 13047 respondents) by 51-48%.
9- In 2008, the Final NEP indicated 60.3 million returning Bush voters from 2004.
10- But Bush ONLY had 62.0 million recorded votes and approximately 3.0m died.
11- In 2008, the Final NEP indicated 11 MILLION more returning Bush voters than Kerry voters. BUT Bush won by 3.0 MILLION RECORDED (not TRUE) votes.
12- The returning voter anomalies had the effect of DECREASING the Kerry and Obama vote shares by 3-4%.
13- If the Final NEP indicates an IMPOSSIBLE number of returning voters, then simple LOGIC dictates that the Final is impossible.
14- If the Final NEP is impossible, since it is MATCHED to the official vote count, then the official vote count must ALSO be impossible.
15-If the Final NEP is impossible, then ALL of the demographic crosstabs are IMPOSSIBLE.
16- Since the 2004 and 2008 Final NEP were both impossible and matched to the recorded vote, the RECORDED vote counts were IMPOSSIBLE.
If you disagree with any of these FACTS, tell us.
_______
This is a belated response since I just read the comments 30 minutes ago.
Thanks, auto, for your "defense" of my work.
http://www.smirkingchimp.com/thread/18513#comment-144297
_______________________________________________________________________
The Naysayer:
What is the math based on, though?
What I see in your "Urban Legend" piece is you working backwards from the notion that the actual vote tallies were incorrect, that the national and state exit polling raw numbers are more reliable than either the weighted polling numbers or the actual vote counts, and then you work the poll data in such a way that it proves the vote tallies must be wrong. Yeah, it's math, and I can see your work - but that doesn't mean I think it's good math.
Similarly, I look at TruthIsAll's work, and find that he must have done something similar with the 2004 results to come up with his "TrueVote" model. He then takes the estimates of what the turnout was expected to be this year, assumes that the high-end poll estimates of Obama's support in the final weekend were gospel, applies the "TrueVote" and comes up with a number that he claims were the actual votes cast, a large percentage of which has been "disappeared" because it doesn't match the figures on CNN. That's one hell of a lot of assuming, estimating and extrapolating, don't you think? The only actual number in any of that is the reported returns - everything else is smoke and mirrors.
There's a reason why "lies, damned lies, and statistics" is such a cliche.
_______
It could be that the purpose of your life is only to serve as a warning to others.
Submitted by Michael Sheridan on November 7, 2008 - 5:25pm.
________________________________________________________________________
My response:
Jan. 6, 2009
This is a belated response since I just read the comments 30 minutes ago. Your interpretation of the analysis is wrong.
There are LIES, DAMNED LIES and a solid Statistical ANALYSIS to expose the LIES.
The official vote count is 131.37 million, up 10.16m since election Day. Obama has won 59.2% of the late votes. As a result,his official vote share has climbed from 52.3to 52.9%. But the recorded vote is NEVER equal to the True Vote. Read on and you will see how some very basic statistical analysis indicates that his True Vote share is 57%.
Note: The Election Model (EM) projected that Obama's expected EV was 365.3 (based on the latest state polls adjusted for undecided voters). Obama had 365 EV. He won 5000 of 5000 Monte Carlo Electoral vote simulation trials - a 100% win probability. The model also projected he would win by 53.1% of the vote; he is at 52.9%. But the model was WRONG; it underestimated Obama's True Vote share by 4% (see the Election Calculator model).
Why did the EM underestimate Obama's True share?
1. The final state pre-election polls were LV (likely voter) polls which are subsets of RV polls. The LVs exclude many newly registered voters; RV polls include new registerd voters (Obama won 71% of new voters according to the National Exit Poll).
2. I projected 60% of undecided voters for Obama; typically, the challenger (Obama) gets 75-90%. In fact, that's the Election Calculator result based on National Exit Poll vote shares (probably conservative)and a FEASIBLE returning voter mix (the NEP returning Bush/Kerry 46/37% voter mix is impossible).
3. The Oct. 30 PEW RV poll had Obama leading by 52-36%. Allocate 50% of the 10% undecided and it's 57-41%. With a 60% UVA, it's 58-40%. It's 59.5-38.5% with a 75% UVA! So to say he had 56-57% is by no means a stretch.
4. Final pre-election state polls are conducted several days before the election; late undecideds and independents broke sharply on Election Day for Obama.
To summarize:
The Election Model used the latest state polling results, adjusted for allocation of undecided voters.
http://www.geocities.com/electionmodel/2008ElectionModel.htm
The Election Calculator used returning prior election voters and the National Exit Poll vote shares to a) forecast the election in May 2008 and b) determine the True Vote (post-election) based on the 2008 National Exit Poll vote shares.
http://www.geocities.com/electionmodel/2008ElectionAnalysisLinks.htm
They are two distinct models; you are confusing the methodologies.
First, consider these election FACTS:
http://www.progressiveindependent.com/dc/dcboard.php?az=show_topic&forum=120&topic_id=4055
To supplement Collin's Urban Legend analysis, see chapter 13:
http://www.geocities.com/electionmodel/FurtherConfirmationOfaKerryLandslide.htm
The True Vote Model is actually the Election Calculator (EC) model. I used it to analyze the 2004 election based on the Preliminary National Exit Poll vote shares. The Final NEP 43/37% Bush/Gore weights were mathematically impossible, so I revised the 2000 returning voter mix based on 2000 returning voters adjusted for voter mortality, uncounted votes and turnout in 2004. Kerry won the True Vote with 53+%.
In May 2008, the EC projected 2008 assuming 2004 voter turnout based on mortality, uncounted votes, etc. Obama was a 54% winner.
After the election, the EC applied 2008 NEP vote shares to a FEASIBLE (and PLAUSIBLE) returning 2004 voter mix. Obama was a 57% winner.
Both models used the best available data: pre-election state and national polls, national exit polls, voter mortality rates, U.S. 2000 and 2004 Census votes cast (includes uncounted votes), 2008 recorded votes and assumed 95% turnout of 2004 election voters.
_______
truthisall.net
Submitted by TruthIsAll on January 6, 2009 - 12:42pm.
Just in case it's still not clear why the Final NEP is bogus.
You seem to have a naive belief that the recorded vote count is sacrosanct. These facts should cause you question that notion.
1- The Final National Exit Poll is always forced to match the official, recorded vote count.
2-The Final 2004 NEP closely matched the 50.7-48.3% Bush recorded margin.
3-The Final 2008 NEP closely matched the 52.9-45.6% Obama recorded margin.
4-The number of returning voters from the prior election MUST BE LESS than the number who actually voted due to mortality and turnout.
5-To match the recorded vote, the 2004 and 2008 Final NEP both indicated that there were MORE returning Bush voters than were still alive.
6-In 2004, the Final NEP indicated 52.6 million returning Bush voters from 2000.
7-But Bush only had 50.5 million recorded votes and approximately 2.5m died.
8-In 2004, Kerry won the unadjusted state exit polls by 52-47% and the preliminary National Exit Poll (12:22am, 13047 respondents) by 51-48%.
9- In 2008, the Final NEP indicated 60.3 million returning Bush voters from 2004.
10- But Bush ONLY had 62.0 million recorded votes and approximately 3.0m died.
11- In 2008, the Final NEP indicated 11 MILLION more returning Bush voters than Kerry voters. BUT Bush won by 3.0 MILLION RECORDED (not TRUE) votes.
12- The returning voter anomalies had the effect of DECREASING the Kerry and Obama vote shares by 3-4%.
13- If the Final NEP indicates an IMPOSSIBLE number of returning voters, then simple LOGIC dictates that the Final is impossible.
14- If the Final NEP is impossible, since it is MATCHED to the official vote count, then the official vote count must ALSO be impossible.
15-If the Final NEP is impossible, then ALL of the demographic crosstabs are IMPOSSIBLE.
16- Since the 2004 and 2008 Final NEP were both impossible and matched to the recorded vote, the RECORDED vote counts were IMPOSSIBLE.
If you disagree with any of these FACTS, tell us.
_______