Ballot-tampering charges fly in S. Texas recount

Incumbent loses lead after untallied votes found on rival's turf

09:46 PM CST on Friday, April 2, 2004

By PETE SLOVER / The Dallas Morning News

SEGUIN, Texas – The lore of South Texas election shenanigans grew this week as teams of lawyers and politicos dashed county to county in a congressional vote-counting dispute reminiscent of Florida 2000 – or even Box 13.

The incumbent, Democratic Rep. Ciro Rodriguez of San Antonio, has seen his lead evaporate after recounts turned up previously untallied votes – akin, his backers say, to the suspect crate of late-discovered ballots that clinched Lyndon B. Johnson's 1948 Senate win.

"We haven't seen anything like this since Box 13," said Mr. Rodriguez's lawyer, Randall "Buck" Wood, referring to the LBJ votes from Jim Wells County. "There is no innocent explanation,"

An 11-county recount this week showed that primary foe Henry Cuellar of Laredo erased a 145-vote deficit and moved ahead by 203 votes, nearly all of them on solid Cuellar political turf, the border counties of Webb and Zapata.

Mr. Cuellar's staff attributed the vote swing to human fatigue and mechanical woes similar to those that plagued the chad-pocked presidential recount four years ago in Florida.

The challenger, already calling himself the winner, panned as a cliche the characterization of his home turf as an extended, corrupt political ward.

"I resent the fact that people from outside the border come and paint everybody with a taint, saying there's fraud and illegal activity," Mr. Cuellar said. "It's unfair for all those honest people who were counting the votes."

Experts said that whatever explanation is ultimately deemed true – probably in court – the conflict serves as a reminder of the fragility of the voting process, that a perfect election is only as plausible as a flawless machine, or a faultless human, to run it.

"We like to assume that we could find a perfect voting system, that we could find a perfect voting," said Lorrie Faith Cranor, an assistant professor of computer science and an election machinery expert at Carnegie Mellon University in Pittsburgh. "The idea is, 'How much can we minimize the error?' We can certainly get it a lot better than we are currently."

Optical scanning

All but two of the 11 counties in the sprawling 28th Congressional District use optical scanning in which voters pencil in a spot to mark their preference – the technology used in 148 of 254 Texas counties.

The initial tally in the March 9 primary showed Mr. Rodriguez ahead by 145 votes out of more than 48,000 cast. In the recount of counties other than Webb and Zapata, both men gained votes, and Mr. Rodriguez boosted his edge by one vote, to 146.

Here's what has raised most of the questions:

• In Zapata County, the machine that reads the ballots had broken down on election night. Ultimately, a crew of inexperienced high school poll workers counted the ballots the next afternoon. The recount showed both men gaining votes from ballots that had been originally read as blank in their race, in roughly the same proportion as the initial vote count dominated by Mr. Cuellar.

Mr. Cuellar, Texas' former secretary of state, picked up 180 more votes than Mr. Rodriguez – enough to nudge him in front by 25 votes.

• In Webb County, which includes Mr. Cuellar's hometown of Laredo, a puzzling outcome: The total number of ballots increased by 115 – all of which went to Mr. Cuellar. He also gained 62 ballots that previously had been counted by machine as blank, while Mr. Rodriguez picked up no such votes.

Mr. Rodriguez's lawyers said there could be no explanation other than fraud and tampering, noting the Webb County disparity and that the new Cuellar votes all came out of a segregated pool of early votes. They plan to ask a judge as early as next week to toss the election results.

Several university experts said that mechanical failings and corrections usually harm or benefit both sides randomly – in the same proportion as the overall vote – and that statistically, Mr. Rodriguez should have gotten at least some of the late-added votes since he won one-sixth of the initial Webb County tally.

Mr. Cuellar's attorneys attributed the new votes to the recount team's separation of "clumped" ballots that were stuck together and not counted by the optical scanning machine on election night.

Mr. Rodriguez, who has not conceded, responded that the recount added three-quarters of 1 percent more ballots to the initial total, far in excess of expected "clumping," and that the multiple passes performed on election night would have detected those problems.

"Something is happening, and it's not correct and it's not appropriate," he said.

An expert familiar with the brand and model of voting machines used in Webb County agreed with the Rodriguez camp, saying that the expected number of miscounts for all reasons – including clumping – would be a fraction of that reported in Webb County.

'Shocked to see'

"I know how the anti-clumping mechanism of that machine works. It's pretty good," said Douglas Jones, an associate computer science professor at the University of Iowa and a member of the state board that certifies voting machines there. "I would be shocked to see clumping like that."

He said clumping could increase if poll workers failed between elections to change a wide rubber belt used to separate ballots. "They'd have to go for a long time before it got that bad," he said.

Punch-card technology, soon to be illegal under federal laws passed after the Florida recount – proved among the most reliable in the congressional recount: Hays County, the only venue using the punch card, reported just a one-vote change.

The only county with a perfect recount, matching election night returns: tiny McMullen County, where 108 voters circled their choices on a paper ballot for counting by hand. That method will still be permitted under the new federal law.