US 2008 Presidential Election: Detecting Fraud

There is no such thing as a "fair" election, there are only elections that are more fair than others.

Most Americans would call an election "fair" if the two major parties spend $1 billion dollars on campaigning, third parties are excluded from debates, third parties had a difficult time getting on the ballot, 10 million undocumented residents were not permitted to vote (or the 7 million of them who are of voting age), people convicted of a felon were not permitted to vote (depending on the state - 4 million are affected), 16 and 17 year olds cannot vote (9 million people), economic inequality ensures that the rich can vote with their money - thus carrying many times more influence in the outcome than the poor or middle class, that elections are on a weekday (instead of a weekend or a national holiday), that US voter registration rates are low and turnout is even lower, and that third party votes almost never elect anyone (eg. all my votes in US elections have gone to losers).

If undocumented residents could vote, they'd probably vote for Obama by a 70/30 margin. Young people would vote for Obama (65/35). Convicted felons would vote for Obama (70/30). The net result of all of these policies is that working class and people of color are less likely to vote. So to narrowly win a "fair" election, Obama needs the support of approximately 55% (note: this number is a guess - I'd estimate 52%-60%) of the 16+ population residing in the US.


That said, there are several more commonly accepted "unfair" election practices that we should be watching for and might be able to prevent. The US could really use a citizen movement to monitor our elections.

1. Removing legitimate voters from the voting rolls. This is done on the state level. Republican states are often intentionally overly zealous in their attempts to remove people from the voting rolls, as they want to remove working class and people of color who are more likely to vote Democrat. This happenned in Florida in 2000, and in Ohio in 2004.

Solution: voting rolls are public. You can get them and document who is being removed. Survey these people to detect any racial/class/partisan bias in who is being targetted by the state.

2. Inaccessible ballots and polling places.

Solution: document where there are long line-ups at the polls, and where polling places failed to open or opened late. Look at the distribution of polling places to ensure that there is no class/race/partisan bias.

3. Election fraud. Electronic or other.

Solution: compare a pre-election poll average, on a state by state basis, to see if the election result is significantly different from the latest prediction. Unless there was a major event in the past couple days, the averaged polls should be predicting the election within 1-2%.

Also, you can compare the exit polls to the results. In 2004, the exit polls showed Kerry winning by a couple percent (Florida, Ohio, and other major swing states). The exit polls were later adjusted as the results came in to have different numbers that were closer to the election result.

There are 4 million people

There are 4 million people who are excluded from federal elections. Most of them (3.7 million) live in Puerto Rico (others live in Guam and American Samoa).

They'd probably have voted 65/35 for Obama. So that adds approximately 1% to the bias. Thus Obama would need 56% of the 16+ population's support to win a 50%+1 majority.

The over representation of small states in the electoral college (getting 2 votes that don't depend on population), which tend to vote Republican, is another factor.