Updated

Punch-card voting may be old-fashioned, but it may not be as detrimental to the vote count as opponents would allege, say some elections experts.

The punch card, with its recognizable "dimpled," "hanging" and "pregnant" chads, has been badly maligned since the 2000 Florida recount dispute, most recently in a decision by a three-judge panel on the 9th Circuit Court of Appeals.

Agreeing with the American Civil Liberties Union (search) that the machines cause too many errors that could disenfranchise tens of thousands of voters in the next scheduled election, the judges ruled Monday to postpone California's Oct. 7 recall election of Democratic Gov. Gray Davis to give counties time to replace the machines.

But elections expert Larry Sabato (search) told Fox News that the judges were remiss in claiming that punch-card machines were inferior to every other type of machine.

"The court was completely wrong in saying punch cards are 'almost undisputed as the most error-prone type of voting machine,'" said Sabato, director of the University of Virginia's Center for Politics (search). "That is absolutely untrue."

According to the California ACLU, punch-card machines, still used in six counties, have an error rate of 2.5 percent, meaning as many as 40,000 votes could be invalidated. In their argument before the court last week, the group argued that because those six counties, which make up 44 percent of the state's voting population, are also 46 percent minority-populated, black and Latino voters will be underrepresented.

"There is something logical about saying if several counties use machines with very high error rates and other counties use perfect machines, that is unequal treatment of voters," said Paul Rothstein (search), a professor at Georgetown Law School.

But other state voting methods that the ACLU finds acceptable are also prone to error, researchers point out. Optical scanners, for example, have an error rate of 2.3 percent, and could invalidate as many as 36,000 votes in the counties where they are used.

Computer touch screens with an error rate of 3 percent could leave 48,000 votes uncounted, and 'Datavote,' a different type of punch-card system used in 20 counties, has an error rate of more than 3 percent — putting as many as 51,200 votes at risk.

Punch cards aren't perfect, but according to research at Cal-Tech and MIT, neither is anything else. Changing to a new ballot system also creates problems of its own.

"Florida recently changed to electronic machines and the error rate was sky high," Rothstein said.

In justifying their ruling, the 9th Circuit judges cite a remark made by California's former Secretary of State Bill Jones, who allegedly said that the punch-card system is "unacceptable." Jones said the word he used was "obsolete."

The judges also claimed punch cards are so unreliable that in 1984, San Joaquin County had to "eliminate the votes of an entire precinct." But San Joaquin officials deny that claim.

The full circuit has said that it may be willing to rehear the case, and the state filed its briefs Wednesday seeking a retrial. The mistaken claims used by the three-judge panel could be an indication as to why some of the court's other judges are willing to re-examine the decision.