The 2004-2008 County Presidential True Vote Database Model
March 23, 2012
The 2004-2008 County True Vote Database Model has been restructured. Just enter the state code in cell A2 of the new “Input” sheet.
The objective of the model is to determine the most fraudulent counties in 2008. No model is perfect, but the TVM provides a good estimate of election fraud as measured by vote share and vote count discrepancies in margin from the recorded vote.
The following states are currently in the database. More states will be added:
AZ CA CO FL GA IA IL IN MA MI MO NC NM NV NY OH OR PA TN TX WI
In 2004, Kerry lost the national recorded vote by 50.7-48.3%. He had 51.1% in the unadjusted state exit polls (76,000 respondents) and 51.7% in the unadjusted National Exit Poll (a subset with 13,660 respondents). The True Vote Model sensitivity analysis provides convincing evidence that Kerry won the election easily.
In 2008, Obama won the national recorded vote by a 52.9-45.6% margin. He had a 58% share in the unadjusted state exit polls (83,000 respondents) and a whopping 61% in the unadjusted National Exit Poll (a subset with 17,836 respondents).
The key stats are shown in the “Input” sheet. These include the state True Vote table and discrepancies between Obama’s True county vote margin and the recorded margin. Calculation details are displayed in the “Model” sheet.
The 2008 Election Model projection exactly matched Obama’s 365 electoral votes and was within 0.2% of his 52.9% share (a 9.5 million margin).
But the model understated his True Vote. The forecast was based on final likely voter (LV) polls that had Obama leading by 7%. The registered voter (RV) polls had him up by 13% – before undecided voter allocation. The landslide was denied.
The Final 2008 National Exit Poll was forced to match the recorded vote by indicating an impossible 103% turnout of living Bush 2004 voters and 12 million more returning Bush than Kerry voters. Given Kerry’s 5% unadjusted 2004 exit poll and 8% True Vote margin, one would expect 7 million more returning Kerry than Bush voters – a 19 million discrepancy from the Final 2008 NEP. Another anomaly: The Final 2008 NEP indicated there were 5 million returning third party voters – but only 1.2 million were recorded in 2004. Either the 2008 NEP or the 2004 recorded third-party vote share (or both) was wrong.
The True Vote Model determined that Obama won by over 22 million votes with 420 EV. His 58% share was within 0.1% of the unadjusted state exit poll aggregate (83,000 respondents).
True Vote margins are calculated based on the returning voter method: The default method is that 2004 voters return to vote in 2008 in proportion to the 2004 state exit poll shares. Optionally, set code 1 in cell A5 to calculate returning voters in proportion to the 2004 recorded vote shares.
If the recorded vote option is used, county vote discrepancies will be lower than they would be if the default method was used. But since the 2004 recorded votes were fraudulent, unadjuted 2004 exit polls (the default) should be used to calculate returning voters to provide a better estimate of the true discrepancy.
The model automatically adjusts state and county vote shares based on the differential between the unadjusted state and national exit polls.
The user has the option of overriding the returning voter mix as well as the vote shares. Enter incremental percentage changes to a) Kerry’s returning vote counts (an automatic offsetting change is made to Bush’s vote count) and b) to Obama’s shares of new and returning voters (an automatic offsetting change is made to McCain’s vote shares.
Default 2004 living voter turnout in 2008 is set to 97%.
The model indicates that the following counties were the most fraudulent:
Obama recorded share: 56.2%, Exit Poll: 63.3%, True Vote Model: 58.8%
Obama recorded share: 51.4%, Exit Poll: 56.3%, True Vote Model: 56.2%
Obama recorded share: 50.9%, Exit Poll: 52.1%, True Vote Model: 54.7%
Palm Beach, Miami-Dade, Broward, Brevard, Hillsborough
Obama recorded share: 62.8%, Exit Poll: 71.5%, True Vote Model: 68.3%
Nassau, Suffolk, Erie, Queens, Westchester
Obama recorded share: 54.5%, Exit Poll: 63.8%, True Vote Model: 62.0%
Philadelphia, Allegheny, Bucks, Westmoreland, Montgomery, Delaware
The correlation ratio is a statistical measure of the relationship between Obama’s recorded vote share and the True Vote discrepancy. In general, there is a strong negative correlation between the two variables. This indicates that as Obama’s recorded county vote share increases (decreases) the discrepancy decreases (increases). This is an indication that GOP counties are the most fraudulent (measured by vote share margin discrepancy).
The correlation ratio is in the range from -1 to +1 (-1 is a perfectly negative correlation and +1 is perfectly positive). A near zero correlation indicates little or no relationship. A positive value indicates that the variables generally move in the same direction: as one variable increases (decreases), the other also increases (decreases). A negative value indicates the opposite: as one variable increases (decreases) the other decreases (increases). The correlation is a strong one if it is higher than 0.50 (positive) or lower than -0.50 (negative).
For example, in Ohio 2004 the -0.82 correlation indicated that Bush counties were more fraudulent than Kerry counties (based on vote share margin discrepancies). In 2008, the -0.50 correlation was not as strong but still significant.
North Carolina was an unusual exception. In 2004 there was a near-zero (-.01) correlation, indicating no relationship between county partisanship and election fraud. But in 2008, the strong negative (-0.72) correlation indicates that election fraud was more prevalent in GOP than Democratic counties.
County Correlation Ratios between the Democratic Recorded Vote and
the True Vote Share Margin Discrepancy
State 2004 2008
NC -0.01 -0.72
WI -0.70 -0.50
OH -0.82 -0.50
NY -0.62 -0.45
FL -0.43 -0.79