1 million votes saved by election upgrades

ASSOCIATED PRESS

Improvements to voting machines and election administration saved 1 million votes that otherwise likely would have gone uncounted in the 2004 elections, with states and counties that made the most comprehensive upgrades recovering the most votes, an academic analysis says.

The report, released yesterday by the Caltech/MIT Voting Technology Project, looked at a key measure of election integrity — residual votes, or ballots cast during an election on which voters failed to mark a choice or machines did not record it.

It was one of the fundamental problems of the 2000 voting stalemate and a focus of subsequent reforms.

In every election, some voters intentionally leave the presidential contest blank, perhaps as a protest or because they are not interested. But analysts estimate that should be, on average, roughly one-half of 1 percent of all ballots, and anything above that is likely caused by flaws with design, equipment or voter education.

In 2000, the national residual vote was 1.9 percent of ballots cast for president. The report found a significant improvement this year, with the residual vote falling to 1.1 percent. The analysis examined 37 states and the District of Columbia; figures were not available elsewhere.

Basing the calculations on that change, the report concluded that of the slightly more than 122 million Americans who voted for president in November, roughly 1,030,000 voters were likely to have mistakenly left their ballots blank.

The report “provides some corrective to what’s getting to be a common interpretation of this election, which is that the election was deeply flawed and things were worse in 2004 than in 2000,” said Charles Stewart III, the report’s author and a Massachusetts Institute of Technology political science professor.

“By this measure — which is a measure that people were trying to improve on — things were significantly better in 2004 than in 2000,” he said.

Among other findings, the report concluded:

• The greatest improvements in residual votes occurred where two things happened together: states undertook overall improvements in administration and voter education, and counties upgraded voting machines. Places where punch-card machines were replaced showed particularly strong gains.

• Florida, the scene of the 2000 postelection stalemate, and Georgia had the biggest drop in residual votes. Florida went from 2.9 percent to 0.4 percent; Georgia went from 3.5 percent to 0.4 percent. Both underwent comprehensive reform, with Georgia putting in electronic-voting machines statewide, Florida scrapping punch cards and both states launching ambitious voter-education campaigns.

• The five states that worsened included three that made little, if any, changes to their voting machines — Connecticut, Iowa and Nebraska, the report said. California and Indiana also saw a rise in residual votes.

The voting problems of 2000 spurred a federal law that has distributed $2.2 billion to states to upgrade equipment, centralize voting registration and improve election practices.

The report says other factors might influence the number of residual votes, including the presidential candidates and how energized the electorate may be, along with the awareness among election officials that they were under scrutiny.

Story Continues →