TruthIsAll
02-05-2011, 08:13 PM
http://richardcharnin.com/TrueVoteModeling.htm
True Vote Modeling: Essentials for the Layman
In the eleven elections since 1968, there have been approximately 80 million net uncounted votes (an average of 7 million per election). Net uncounted votes declined from 10.6 million (10%) in 1988 to 5.4 million (5%) in 2000 to 3.4 million in 2004 (3%). The vast majority of uncounted votes are Democratic. Therefore, the Democratic recorded vote must always be lower than the True Vote. That is one reason why Democratic presidential candidates always do better in the unadjusted and preliminary exit polls than in the recorded vote.
But since HAVA (2002), there are more compelling reasons than uncounted votes for the discrepancies between the True Vote and the recorded vote. The percentage of net uncounted votes has declined steadily since 1988; they are no longer the dominant factor in the True Vote discrepancy.
Electronic voting machines have become institutionalized. Touch screen computers (DREs) produce unverifiable results. Optical scanned paper ballots are rarely hand-counted. In addition, the invisible central computers that tabulate total votes for each district/county are vulnerable to malicious programming; calculations based on DREs are lost in cyberspace and cannot be verified. Since 1998 Oregon is the only paper ballot state which mandates hand-counts of randomly-selected counties. Its vote-by-mail system has resulted in much higher voter turnout and nearly fool-proof elections.
In the1968-2008 elections, the average presidential recorded vote share was 49-45% in favor of the Republican. We will show that the average presidential True Vote share was 49-45% in favor of the Democrats.
We distinguish between the True Vote (how people actually tried to vote) and the official, recorded vote provided by the media. It is an undeniable fact that in every election, True Vote shares do not equal the recorded vote shares. This is self-evident since the number of votes cast is never equal to the number recorded and therefore the corresponding shares of total votes cast cannot be equal to the recorded vote shares.
It follows that the first step in calculating the True Vote is to estimate the number of uncounted votes. The Census Bureau surveys total votes cast in every election (the margin of error is less than 0.5%). We have the simple formula:
Net Uncounted Vote = Census Total Votes Cast - Official Recorded Vote
The Net Uncounted Vote is greater than zero when the number of uncounted votes exceeds the number of stuffed ballots.
Net Uncounted Vote = Uncounted Votes – Stuffed ballots
Key parameters in calculating the True Vote are a) the number of returning voters from the prior election, b) new voters and c) corresponding exit poll vote shares. In order to calculate a robust estimate of returning voters, we must consider the mathematical constraints. It is obvious that the number of returning voters must be less than the number who actually voted (the Final National Exit Poll which is forced to match the recorded vote would have us believe otherwise).
We need to estimate voter mortality and turnout of prior election voters. An estimated 5% of voters pass on in the four years from the previous election (based on mortality tables). Vote shares are hardly affected by changes in the rate. The turnout of previous election voters can be estimated from registered voter turnout and can vary from 90-98%, depending on voter interest. It is estimated that in 1992 and 2004, 98% of previous election voters turned out to vote.
Given voter mortality and turnout of living voters from the previous election, we can now calculate an estimate for the number of returning voters.
Method 1: Returning voters based on previous Recorded vote
Returning voters = (Previous election Votes Recorded – voter mortality) * Turnout rate
Given returning voters, we easily calculate the number of new voters in the current election:
New voters = Total Votes Recorded in the current election – Returning Voters from the previous election (recorded)
Note that by calculating returning voters based on total recorded votes, we understate the Democratic vote share, since the calculation does not include heavily Democratic uncounted votes. This method is analogous to the exit pollsters designing a sample based on the previous election voting demographis,
Method 2: Returning voters based on previous Total Votes Cast:
Returning voters = (Previous election Votes Cast – voter mortality) * Turnout rate
Given the number of returning voters, we easily calculate the number of new voters in the current election:
New voters = Total Votes Cast in the current election – Returning Voters from the previous election (cast)
Next we need to calculate the number of returning voters for each candidate
Democratic total votes cast in prior election = Democratic recorded vote + 75% of uncounted votes in previous election
Republican total votes cast in prior election = Republican recorded vote + 25% of uncounted votes in previous election
Method 3: Returning voters based on previous Exit poll
In Method 2, we assume that total votes cast for each candidate is exactly the sum of the recorded and uncounted votes. We can also calculate the votes cast for each candidate by applying the previous unadjusted exit poll to total votes cast. This approximates the shares calculated using Method 2.
Returning voters = (Previous election Votes Cast * Exit Poll – voter mortality) * Turnout rate
Given returning voters, we easily calculate the number of new voters in the current election:
New voters = Total Votes Cast in the current election – Returning Voters
Democratic total votes cast in prior election = Democratic unadjusted exit poll * total votes cast in previous election
Republican total votes cast in prior election = Republican unadjusted exit poll * total votes cast in previous election
Method 4: Returning voters based on previous True Vote
The most logical method is to use the previous election True Vote shares as the basis for calculating returning voters. This is a recursive process, whereby the current True vote is a function of the prior True vote. This method is the most accurate.
Returning voters = (Previous election Votes Cast * True Vote – voter mortality) * Turnout rate
Given returning voters, we easily calculate the number of new voters in the current election:
New voters = Total Votes Cast in the current election – Returning Voters from the previous election (cast)
A demonstration version of the Recursive True Vote Excel Model is available as a Google Docs spreadsheet.
https://spreadsheets1.google.com/ccc?hl=en&key=tZC-vfYAE-API5cYs6dklWA&hl=en#gid=0
Vote shares can be calculated for all elections since 1972 using four methods:
1. Final National Exit Poll (forced to match)
2. Returning voters based on previous election recorded vote
3. Returning voters based on previous election votes cast
4. Returning voters based on previous election True Vote
Row 15 displays the True Vote for all elections 1968-2008.
Note that the Excel command used to create "what-if" sensitivity analysis tables is not available in the Google Spreadsheet Doc.
These default input assumptions may be changed:
1) Prior election 5% voter mortality (cell A3).
2) Turnout of prior election living voters in the current election (cells B4,B5,B6).
To run the model:
1. Choose the election by entering the election code (1-10) in cell A1:
Note the codes: 10=2008, 9=2004,...1=1972
2. Enter the calculation method (1-4) in cell A2:
DO NOT SAVE THE SPREADSHEET OR EDIT CHANGES WILL REMAIN IN EFFECT.
CLICK X IN THE TAB TO CLOSE THE SCREEN AND EXIT.
True Vote Modeling: Essentials for the Layman
In the eleven elections since 1968, there have been approximately 80 million net uncounted votes (an average of 7 million per election). Net uncounted votes declined from 10.6 million (10%) in 1988 to 5.4 million (5%) in 2000 to 3.4 million in 2004 (3%). The vast majority of uncounted votes are Democratic. Therefore, the Democratic recorded vote must always be lower than the True Vote. That is one reason why Democratic presidential candidates always do better in the unadjusted and preliminary exit polls than in the recorded vote.
But since HAVA (2002), there are more compelling reasons than uncounted votes for the discrepancies between the True Vote and the recorded vote. The percentage of net uncounted votes has declined steadily since 1988; they are no longer the dominant factor in the True Vote discrepancy.
Electronic voting machines have become institutionalized. Touch screen computers (DREs) produce unverifiable results. Optical scanned paper ballots are rarely hand-counted. In addition, the invisible central computers that tabulate total votes for each district/county are vulnerable to malicious programming; calculations based on DREs are lost in cyberspace and cannot be verified. Since 1998 Oregon is the only paper ballot state which mandates hand-counts of randomly-selected counties. Its vote-by-mail system has resulted in much higher voter turnout and nearly fool-proof elections.
In the1968-2008 elections, the average presidential recorded vote share was 49-45% in favor of the Republican. We will show that the average presidential True Vote share was 49-45% in favor of the Democrats.
We distinguish between the True Vote (how people actually tried to vote) and the official, recorded vote provided by the media. It is an undeniable fact that in every election, True Vote shares do not equal the recorded vote shares. This is self-evident since the number of votes cast is never equal to the number recorded and therefore the corresponding shares of total votes cast cannot be equal to the recorded vote shares.
It follows that the first step in calculating the True Vote is to estimate the number of uncounted votes. The Census Bureau surveys total votes cast in every election (the margin of error is less than 0.5%). We have the simple formula:
Net Uncounted Vote = Census Total Votes Cast - Official Recorded Vote
The Net Uncounted Vote is greater than zero when the number of uncounted votes exceeds the number of stuffed ballots.
Net Uncounted Vote = Uncounted Votes – Stuffed ballots
Key parameters in calculating the True Vote are a) the number of returning voters from the prior election, b) new voters and c) corresponding exit poll vote shares. In order to calculate a robust estimate of returning voters, we must consider the mathematical constraints. It is obvious that the number of returning voters must be less than the number who actually voted (the Final National Exit Poll which is forced to match the recorded vote would have us believe otherwise).
We need to estimate voter mortality and turnout of prior election voters. An estimated 5% of voters pass on in the four years from the previous election (based on mortality tables). Vote shares are hardly affected by changes in the rate. The turnout of previous election voters can be estimated from registered voter turnout and can vary from 90-98%, depending on voter interest. It is estimated that in 1992 and 2004, 98% of previous election voters turned out to vote.
Given voter mortality and turnout of living voters from the previous election, we can now calculate an estimate for the number of returning voters.
Method 1: Returning voters based on previous Recorded vote
Returning voters = (Previous election Votes Recorded – voter mortality) * Turnout rate
Given returning voters, we easily calculate the number of new voters in the current election:
New voters = Total Votes Recorded in the current election – Returning Voters from the previous election (recorded)
Note that by calculating returning voters based on total recorded votes, we understate the Democratic vote share, since the calculation does not include heavily Democratic uncounted votes. This method is analogous to the exit pollsters designing a sample based on the previous election voting demographis,
Method 2: Returning voters based on previous Total Votes Cast:
Returning voters = (Previous election Votes Cast – voter mortality) * Turnout rate
Given the number of returning voters, we easily calculate the number of new voters in the current election:
New voters = Total Votes Cast in the current election – Returning Voters from the previous election (cast)
Next we need to calculate the number of returning voters for each candidate
Democratic total votes cast in prior election = Democratic recorded vote + 75% of uncounted votes in previous election
Republican total votes cast in prior election = Republican recorded vote + 25% of uncounted votes in previous election
Method 3: Returning voters based on previous Exit poll
In Method 2, we assume that total votes cast for each candidate is exactly the sum of the recorded and uncounted votes. We can also calculate the votes cast for each candidate by applying the previous unadjusted exit poll to total votes cast. This approximates the shares calculated using Method 2.
Returning voters = (Previous election Votes Cast * Exit Poll – voter mortality) * Turnout rate
Given returning voters, we easily calculate the number of new voters in the current election:
New voters = Total Votes Cast in the current election – Returning Voters
Democratic total votes cast in prior election = Democratic unadjusted exit poll * total votes cast in previous election
Republican total votes cast in prior election = Republican unadjusted exit poll * total votes cast in previous election
Method 4: Returning voters based on previous True Vote
The most logical method is to use the previous election True Vote shares as the basis for calculating returning voters. This is a recursive process, whereby the current True vote is a function of the prior True vote. This method is the most accurate.
Returning voters = (Previous election Votes Cast * True Vote – voter mortality) * Turnout rate
Given returning voters, we easily calculate the number of new voters in the current election:
New voters = Total Votes Cast in the current election – Returning Voters from the previous election (cast)
A demonstration version of the Recursive True Vote Excel Model is available as a Google Docs spreadsheet.
https://spreadsheets1.google.com/ccc?hl=en&key=tZC-vfYAE-API5cYs6dklWA&hl=en#gid=0
Vote shares can be calculated for all elections since 1972 using four methods:
1. Final National Exit Poll (forced to match)
2. Returning voters based on previous election recorded vote
3. Returning voters based on previous election votes cast
4. Returning voters based on previous election True Vote
Row 15 displays the True Vote for all elections 1968-2008.
Note that the Excel command used to create "what-if" sensitivity analysis tables is not available in the Google Spreadsheet Doc.
These default input assumptions may be changed:
1) Prior election 5% voter mortality (cell A3).
2) Turnout of prior election living voters in the current election (cells B4,B5,B6).
To run the model:
1. Choose the election by entering the election code (1-10) in cell A1:
Note the codes: 10=2008, 9=2004,...1=1972
2. Enter the calculation method (1-4) in cell A2:
DO NOT SAVE THE SPREADSHEET OR EDIT CHANGES WILL REMAIN IN EFFECT.
CLICK X IN THE TAB TO CLOSE THE SCREEN AND EXIT.