2004 vote count smoother, still some problems

By THOMAS HARGROVE
Scripps Howard News Service
December 22, 2004

- America did a much better job of counting its votes this year.

The nation's $2.2 billion investment in new voting machines and other election reforms reduced the number of ballots that failed to count in last month's presidential election, according to a Scripps Howard News Service study based on final or nearly complete election data provided by 43 states and the District of Columbia.

Nearly 99 percent of the 122 million ballots cast in the Nov. 2 general election successfully recorded a vote for president, a significant improvement over the 98 percent that counted in the race for the White House four years ago.

"It's a start, a very good start," said DeForest Soaries, chairman of the U.S. Election Assistance Commission created by Congress to fix the kinds of election problems found in Florida four years ago. "Have all of our problems been solved? No. But there's much more to celebrate in this election than to mourn, no question about it."

But the study did find wide variations among states in the accuracy of vote counts. Some states were eight times more likely than others to have ballots that didn't tally a choice for president last month, often because of poor ballot design or errors in tabulation.

"We still have a lot of work to do," Soaries said.

The 2000 presidential recount in Florida made many Americans aware for the first time that there is a significant margin of error in how votes get counted. Several states enacted sweeping reforms that included wholesale abolition of punch-card and lever-action voting machines in favor of newer technologies like optically scanned balloting and touch-screen computer terminals.

Most of those states showed dramatic improvement this year, including Florida itself. Only 30,509 of the Sunshine State's reported 7.6 million ballots failed to record a vote for president, down from 178,145 that were silent in the presidential race four years ago.

"That was a wakeup call for the whole nation. We all had a huge challenge because of the ghost of 2000," said Florida Secretary of State Glenda Hood. "Florida took the lead and put an election reform package into place. We pretty much covered everything, while a number of states still have a lot of work to do."

Florida went from 97.1 percent of its ballots recording a presidential vote four years ago to 99.6 percent counting last month.

The state was just slightly behind Georgia, which switched to a statewide touch screen voting system after its hodgepodge of aging voting technologies produced an even worse rate of undervoting than Florida's in 2000. Both Florida and Georgia trail Nevada where only 2,246 ballots - less than 0.3 percent of all ballots cast - failed to register in the presidential race this year.

"We were first? I predicted we'd have the lowest!" said Nevada Secretary of State Dean Heller. "Last December, we decertified all punch-card machines in this state. Nevada's confidence in the validity of this presidential election is the highest in the nation."

Heller's state is the first to have all-electronic voting with a paper printout backup record that critics say is a necessary safeguard for touch-screen machines.

Nevada also benefits from a "none of the above" option if voters dislike their political choices, a practice Heller said further reduces silent ballots. "In most other states, the only protest vote people have is to vote for nobody. Having a real protest vote helps us," he said.

Illinois suffered the worst rate of undervoting four years ago, with only 96.1 percent of its 4.9 million ballots recording a presidential vote. Almost two-thirds of the discrepancy was in Chicago's Cook County where officials blamed poorly manufactured plastic punch card templates used to help voters punch out cardboard chads.

In the Land of Lincoln this year, about 98.6 percent of its 5.4 million ballots counted.

"We had a learning curve to understand this issue," said Daniel White, executive director of the Illinois Board of Elections. "Four years ago, over 90 percent of our counties were using punch cards. Now, more than 50 percent are using optical scan balloting. And we had a solid effort by election authorities to educate voters."

The worst rates of undervoting for president this year occurred in New Mexico with 97.5 percent of ballots counting, Idaho with 97.6 percent counting, and North Carolina with 97.7 percent.

"We don't spend a lot of time on undervote issues," said New Mexico Secretary of State Rebecca Vigil-Giron. "I'm just speculating that some voters are just not concerned with the presidential race."

But Gary Bartlett, executive director of the North Carolina Board of Elections, did not defend the high undervote or suggest voters are ignoring the presidential race. "I was hoping we would improve over what happened in 2000. But this shows a law in our state that needs to be reviewed and probably be changed," Bartlett said.

Both North Carolina and South Carolina historically suffer unusually high undervotes in presidential elections because, by state law, voters who mark the "straight-party-ticket voting" option must also vote separately for president. Every four years, tens of thousands of voters in both states apparently forget to do this.

Full election data is not yet available from South Carolina, but in North Carolina this year 58,223 ballots failed to register a presidential vote.

A few other states with historically high rates of undervoting showed little or no improvement this year, including the key battleground state of Ohio. Some 96,580 ballots in the Buckeye State failed to register a presidential vote this year, up from 93,991 four years ago.

Ohio Secretary of State Ken Blackwell, noting that 68 of Ohio's 88 counties still used punch-card ballots last month, said the results underscore the need for d voting equipment. But Blackwell also suggested his state's 1.7 percent undervote - more than four times higher than Florida's this year - could be caused by the voters themselves.

"Given human nature, when you're talking about 5.8 million people casting a vote, it wouldn't be too far-fetched to think that you have a small percentage of people who would say, 'A pox on both of your houses,' " Blackwell said.

The Scripps Howard study found that there were at least 1,126,562 ballots that failed to register a presidential vote last month, based on complete information from 40 states and partial data from three others. This is a dramatic improvement from the 1,605,263 ballots that didn't register presidential votes in 38 states that reported complete data (including number of ballots cast) in 2000.

There were many causes for the reduced number of undervotes this year, including a few unusual procedures by state officials.

"Wyoming has this funny situation. It has just been a tradition here that we don't report write-ins," Secretary of State Joe Meyer told Scripps Howard in September when asked why only 96.4 percent of the ballots seemed to register a presidential vote.

Since Wyoming did not report write-ins, it ignored Green Party candidate Ralph Nader, who got 2 percent of the vote as a write-in candidate four years ago.

Meyer promised Scripps Howard he would report all write-in votes in future general elections, a promise he kept last month when he reported 480 write-in votes for a variety of candidates. By more completely reporting votes, Meyer helped reduce the apparent undervote in Wyoming to just 0.8 percent.

"It does take more time on Election Night, separating out the write-in votes and tallying them," Meyer said. "But election officers need to make sure that votes are accurately counted and reported. We are more sensitive to these issues now and we'll just report everything we can from an election."

Many other states agreed to increase the amount of information they gather and report, including the number of ballots cast, critical data in determining if mistakes have been made.

Twelve states failed to report the number of ballots cast four years ago when certifying their results for president. Texas Secretary of State Geoffrey Connor said his staff began "talking about this internally" after Scripps Howard called asking state officials to defend the practice. "We are aware of this now and I am trying to do something about it," Connor said in June.

Connor also kept his promise and reported, for the first time, that 7.5 million ballots were cast last month. From this, it's possible to calculate that 93,083 ballots did not register a presidential vote, or about 1.2 percent of ballots cast.

"No, this was not difficult to do," Connor said after the election. "We've learned from this and we've incorporated this into our training. We want this (ballot count) incorporated into the state election law as well."

All of the so-called "dirty dozen" states that failed to report ballot counts four years ago promised to do so this year. Eight states already have provided at least a statewide tally so far, already making the 2004 election the most completely documented election in U.S. history.

But many states remain slow in reporting the number of voters who went to the polls last month. Apparently the last to report will be South Carolina, which also historically suffers a high undervote. A complete ballot count may not be available until March, two months after winning candidates have been sworn into office.

"We are working to change this," said Marci Andino, executive director of the South Carolina Election Commission. She said ballot counts are a vital cross check to ensure the accuracy of elections.

"Hopefully, we'll have a new voter registration and reporting system by 2006," Andino said. "Until then, we have to work in the confines of what we have."