Your vote is at risk in some states: here's why

By THOMAS HARGROVE and MICHAEL COLLINS
Scripps Howard News Service
July 07, 2004

- Call them the dirty dozen of democracy.

Election officials in 12 states did not report how many ballots were cast when they certified 26,349,619 votes for president four years ago, making it impossible to know how many votes were lost because of inaccurate counting machines or other tabulation errors.

Most are still unprepared to check for missing votes this November, increasing the odds that America will face another uncertain presidential election. Experts warn that the mistakes painfully discovered in Florida in 2000 could be repeated.

"This is really embarrassing," said Deforest Soaries Jr., chairman of the new U.S. Election Assistance Commission created by Congress to fix Florida-like voting problems. "How can we ever measure the error rate without having the global vote numbers?"

Eleven states did not release official ballot counts in the 2000 presidential race - Alabama, Arkansas, Delaware, Maine, Mississippi, Missouri, Oklahoma, Pennsylvania, Tennessee, Texas and Wisconsin. Indiana counted the number of ballots cast, but published the information more than three months after results were certified and well after winning candidates had assumed office.

Texas political activist Madeleine Hervey said the failure to collect that data shows a lack of respect that many election officers have for voters. "They don't give a fig, a flying fig," she said.

Hervey believes her ballot was among an estimated 41,000 votes lost during the 1998 general election in Dallas County because of a computer glitch in the county's electronic voting machines. "My vote got flushed," she said.

Thirty-eight states and the District of Columbia reported complete election figures four years ago. There were 80,644,664 ballots cast, but only 79,039,401 presidential votes counted. Much of the 1,605,263-vote difference was caused by inaccurate tabulating equipment, state and local election officials agree.

Florida drew the world's spotlight in 2000 when 178,145 ballots didn't register a vote for president, about 2.9 percent of all ballots cast. Election experts suggest that the undervote - the difference between the numbers of ballots cast and votes counted - should be investigated whenever it exceeds 2 percent in major elections.

Yet for all of the flaws uncovered in Florida, Sunshine State officials counted the number of ballots cast so that the size of the problem was immediately obvious. Election activists increasingly are worried over what happened to a quarter of the nation's vote in those states that didn't count ballots.

"You'll never know if a patient has a fever if you don't even have a thermometer," said Doug Chapin, director of electionline.org, a nonprofit clearinghouse of election-reform information.

Scripps Howard News Service contacted top election officials in all 12 states, asking for a ballot count from their election archives in an attempt to estimate possible tabulation errors. Only Delaware and Indiana provided a complete accounting for the 2000 election, although officials in three other states said they still hope to eventually provide the data.

"This information is something we've always had internally, but we've never made this into a public record," said Delaware Commissioner of Elections Frank Calio.

Scripps Howard compared the ballots cast against the votes recorded for president four years ago and gave Calio a report showing an above-average undervote in the state's most urban area. New Castle County reported 212,995 presidential votes counted out of 220,871 ballots cast. The apparent difference of 7,876 votes represents an undervote of 3.6 percent, which is worse than Florida's and almost twice the national average.

"I find it difficult to understand the huge undervote in New Castle," Calio said. "That research was eye-opening. We've never had any reports of anything going wrong with our machines. We need to find out what the problem is. We're not going to whitewash this."

Several other states' chief elections officers readily agreed their vote accounting procedures in the past were inadequate.

"We are aware of this now and I am trying to do something about it," said Texas Secretary of State Geoffrey Connor. Members of his staff began "talking about this internally" in April when a Scripps Howard reporter asked why Texas does not count its ballots, Connor said.

Ten states - Arkansas, Delaware, Indiana, Maine, Mississippi, Missouri, Oklahoma, Tennessee, Texas and Wisconsin - promised they would try to issue county-by-county ballot counts in November. But most of them have yet to make a successful accounting despite attempts to do so in recent elections.

One state that did present a successful ballot count for the 2002 general election was Maine, where Secretary of State Dan Gwadosky convinced state lawmakers to approve reforms mandating a ballot count in all voting jurisdictions.

"The 2002 general election was the first time we've done this calculation," Gwadosky said. "But our voter turnout tends to spike dramatically on presidential elections. So this November will be our first real test."

According to Gwadosky's first attempt at tracking lost votes, the state issued 511,609 ballots in 2002 and counted 505,190 votes for governor, an undervote of just 1.25 percent. But that apparent undervote was incorrectly reduced by an error from the town of Lisbon, where 1,715 more votes were counted than reported ballots cast.

"Because the total-ballots-cast figure was a new part of the process ... some errors occurred in those tabulations at the local level," said state spokesman Doug Dunbar. "There was no indication that vote totals for candidate races or ballot questions in Lisbon, or elsewhere, were incorrect."

Mississippi officials are still preparing a report comparing the number of votes counted against the number of ballots cast in the 2002 general election, even though the Mississippi legislature mandated that counties begin reporting the data that year. "We're still trying to finalize everything," said state spokesman David Blount.

Mississippi Secretary of State Eric Clark said he is hopeful he can count the ballots more quickly in November. "At least, I hope we will," Clark said. "A lot depends on local election officials in our 82 counties, and how seriously they take this."

Tennessee officials were able to complete a review of ballots cast in the 2000 presidential election, but were unwilling to release the findings. "We didn't get good information," conceded Tennessee Elections Coordinator Brook Thompson. "We know some of the county numbers were not correct."

Thompson said the findings from counties with credible information were disturbing. "Punch cards were the worst," he said. "We had something like 96.5 percent of the vote count, or that there was a -off of about 3.5 percent, for punch cards."

At Scripps Howard's request, Missouri officials for the first time tried to tally ballots cast in the last presidential race and were able to compile reports from 77 of the Show Me State's 116 election jurisdictions. "We will request complete information from all counties in November," said state spokesman Spence Jackson.

Missouri officials said there were at least 37,149 more ballots cast than votes were counted for president four years ago, with at least 10 counties reporting a rate of undervoting worse than Florida's. The worst was an undervote of more than 7 percent of the ballots cast in St. Clair County.

Arkansas Secretary of State Charlie Daniels said state lawmakers in 2001 required an accounting of all votes that did not tabulate correctly. But so far, the state has not prepared such an accounting.

"I think so," Daniels said when asked if the state will report spoiled ballots in November. "It's always good to have all the data on hand to make decisions. How does each county stack up? Are the polling workers properly trained?"

Both Wisconsin and Oklahoma officials said they could count the number of ballots cast, but they had never been asked to do so in past elections. "It wouldn't bother us at all to start counting this," said Oklahoma State Elections Board chief Michael Clingman.

"Wisconsin actually does collect the number of voters who come to the polling place," said Wisconsin State Elections Board Executive Director Kevin Kennedy. "The real problem is that by the time we get those numbers reported from our 1,850 municipalities, nobody pays any attention to the information."

Wisconsin had yet to provide complete data from either the 2000 or 2002 general elections.

Indiana Secretary of State Todd Rokita said it was "an anomaly" that Indiana was slow to report ballots cast in 2000, and promised to report the information on election night this year. But Rokita said his office "has never completed an official tally" of undervotes or made a formal comparison between votes counted and ballots cast.

"That's because, in Indiana, the county clerks have a lot of the responsibilities for putting on elections," Rokita said. "To get that data and then to come in and force a recount or a special election based on that, well, we have no legal jurisdiction to do that."

Officials for Alabama and Pennsylvania said they are not in a position to make a complete accounting of ballots cast come November.

"There is no provision in the Election Code that calls for counties to report to us the actual number of ballots cast," said Brian McDonald, spokesman for Pennsylvania Secretary of the Commonwealth Pedro Cortes. "We can't enforce a law that doesn't exist."

Alabama Secretary of State Nancy Worley said she will ask the Alabama legislature to require counties to count ballots in future elections. "This is the first time this question has been raised to me," Worley said. "I didn't realize we were one of just 12 states that doesn't report this."