Ballot blunders hit poor districts hard, study finds

The Washington Post and the new york times

WASHINGTON — The votes of poor people and minorities were more than three times more likely to go uncounted in the 2000 presidential election than the votes of more affluent people, a congressional study has found.

But modern voting machines can significantly reduce the disparity, according to a national study of the 2000 elections.

The report, prepared by the minority staff of the House Government Reform Committee and without the participation of Republicans, is the first comprehensive nationwide examination of voting patterns and error rates. It shows that what happened in Florida last year was not an isolated phenomenon.

The study, scheduled for release today, was prompted in part by Vice President Al Gore’s loss to Gov. George W. Bush in an election that ultimately came down to several hundred disputed ballots in Florida, whose 25 Electoral College votes went to Bush after the Supreme Court halted a recount there. Bush lost the popular vote but narrowly won the electoral vote.

The study focused on 40 congressional districts in 20 states. Overall it found that 4 percent of all ballots cast in the low-income districts were not tallied for the presidential race, compared with 1.2 percent in the higher-income districts. In two low-income districts, roughly one in 12 ballots — 7.9 percent — were not counted in the presidential race, while the lowest error in one of the more affluent districts was just 0.4 percent.

Most of the districts with low rates of uncounted ballots were the areas of greater affluence and with smaller percentages of minorities. But not in all cases. The 7th District in Alabama, for example, where 68 percent of the population is minority and 31 percent have incomes below the poverty line, had the lowest undercount rate, with just 0.3 percent. The district, stretching from Birmingham to Montgomery along the western part of the state, uses optical scan machines and counts ballots at the precinct, a system that allows voters to correct mistakes before they leave the polling place.

Rep. Henry Waxman of California, the ranking Democrat on the panel, called the disparities "an outrage" and said that more often than not, precincts where poor people lived had older voting equipment.

"I think a lot of people thought the problem was a Florida problem and not a problem all around the country," Waxman said. "This report shows it’s a national issue and we need the federal government to step in."

Overall the study found that the disparities between undercounts in low-income and in more affluent districts were reduced with better technology. For example, where punch card machines were used, the error rate in the low-income districts was 7.7 percent compared with 2 percent in the more affluent districts, a gap of 5.7 percentage points.

When optical scan machines were used and counted in the precincts, the error rate was 1.1 percent in the low-income districts and 0.5 percent in the affluent districts, a difference of 0.6 percentage points.

The study says some analysts estimate that 1.9 percent of all ballots cast in the 2000 presidential race were not counted.

No doubt some were not counted because voters simply chose not to vote for a presidential candidate, or voted for two candidates, the study says. "More often, however," it adds, "the ballots were discarded because the voting machine failed to accurately record the intention of the voter."

The 1.9 percent national no-count rate is equivalent to almost 2 million votes, the study notes, and "in a close election, these discarded ballots could mean the difference between victory and defeat."