Optical Scanners Topped Pregnant Chads as Most Flawed in Florida

When Democratic candidate Al Gore challenged the results of the 2000 presidential election in Florida, attention focused on the state’s problem-plagued punch-card ballots with their hanging and pregnant chads.

But another voting system was even less reliable than the punch cards, the Orlando Sentinel found: an optical scanning system used in 15 of Florida’s 67 counties.

That system, in which ovals on paper ballots are filled in by pencil and scanned at a central county office, resulted in 5.7% of all ballots being rejected--compared with 3.9% in the counties that used punch cards.

The Sentinel examined all 15,596 of the ballots rejected in those 15 counties and found that thousands of potentially valid votes were lost because ballot designs were confusing, counting methods were inconsistent, and--in some cases--election officials never looked at ballots that were rejected by machines.

On more than 1,700 ballots, the voter’s choice for president was clear, but the ballot was rejected nonetheless.

In an additional 5,000 cases, voters’ mistakes--apparently caused by confusing ballot design--made it impossible to determine the voters’ intent.

A Vast Array of Ballot Errors

The ballots, which were easy to examine with the naked eye, revealed a wide variety of problems:

* Election officials threw out 962 ballots on which voters filled in the oval for Gore or Republican candidate George W. Bush but also wrote the same candidate’s name in a space labeled “Write-in candidate.” Many voters, it appeared, mistakenly took that as an instruction to confirm their choice. Some counties examined those ballots by hand and counted the write-in votes as valid, but others did not.

* In more than 100 cases, clear votes for Gore or Bush went uncounted because voters used pens or markers that could not be detected by voting machines. Their ballots were never examined by canvassing boards, even though their votes were clearly visible to the eye.

* Because a presidential vote is actually a vote for that candidate’s electors, most counties’ ballots instructed voters: “Vote for group.” Many voters took that literally, filling in ovals for several candidates. “You have to ask yourself, what in the world were they thinking?” said Franklin County Supervisor Doris Gibbs, examining a ballot in which the ovals for all 10 presidential candidates were filled in.

* In 244 cases, ballots were rejected because voters tried to erase mistakes. Most counties’ ballots said nothing about erasing, but some told voters to ask for new ballots if they made an error.

* On the other hand, in Bradford County, voters who made mistakes were given white stickers to place over their errors. One Bush voter covered two errors with stickers; the counting machine threw that ballot out.

* Some voters made no mistakes at all, but their votes still were thrown out. In Charlotte County, 14 absentee ballots were discarded because a counting machine mistakenly counted a crease--made when the ballots were folded for mailing--as a second vote.

* In 14 of the 15 counties, the ballots split the presidential candidates into two columns. In those counties, 4,268 voters mistakenly chose Bush or Gore in column one and another presidential candidate in column two.

In 13 of the counties, the second column was headed by two minor-party candidates, Monica Moorehead of the Workers’ World Party and Howard Phillips of the Constitutional Party. More than 2,416 voters picked Gore in the first column and Moorehead or Phillips in the second column, disqualifying their ballots. An additional 1,852 voters chose Bush in the first column and Moorehead or Phillips in the second.

The 15,596 ballots, examined in a joint project by the Orlando Sentinel, the South Florida Sun-Sentinel and the Chicago Tribune, were identified by election officials as overvotes or undervotes--meaning counting machines either detected multiple votes for president or no votes at all.

Most of the 15 counties using this optical scanning system are small and rural, and together they represent just 4.6% of all the ballots cast in Florida’s Nov. 7 election.

But because of their high rate of problems, they accounted for 8.6% of all rejected ballots in the state.

Other Scanners Let Voters Correct Errors

The 15 counties all used a voting system that collected paper ballots at each polling place but tabulated them by machine at a single county office.

Other counties used similar paper ballots that are fed into an optical scanner at each polling place, with the voter present. If the ballot is mismarked, the machine spits it out so the voter can correct mistakes. In that system, the reject rate is normally less than 1%.

But many of the smaller counties say they cannot afford the precinct counters, which cost several thousand dollars each.

The distinction between precinct and centrally tabulated voting systems--an obscure detail before Florida’s election debacle--is likely to become an important issue. Gov. Jeb Bush has said that he wants all Floridians to cast their votes on paper ballots by the 2002 election, a change that could cost $35 million in equipment upgrades.

If canvassing boards had performed a visual recount of these 15,596 ballots in November, Gore could have seen a net gain of 366 votes, the study found. That number counts only votes that could be clearly and unambiguously discerned by eye.

However, Gore never asked for a recount in these 15 counties, most of which are heavily Republican.

Bush won Florida--and, as a result, the presidential election--by an official margin of 537 votes.

Ballot reviews in other areas have turned up varying results. The Palm Beach Post examined 10,600 rejected punch-card ballots in Miami-Dade County and found an additional six votes for Bush.

A group of media organizations, including the Sentinel and The Times, is reviewing all 180,000 ballots that were rejected in Florida, a project that will require several months.