Partial Hand-Count of Ballots Reveals Few Irregularities

In the week since the Oct. 7 recall election, academics and election officials have continued to argue over whether the punch-card voting machines used in Los Angeles and six other California counties disenfranchised substantial numbers of voters.

On Wednesday, Los Angeles County Registrar-Recorder Conny McCormack invited reporters and the public to watch election workers hand-count ballots from 19 of the county’s 1,786 precincts.

The 1% hand count, which is mandated by state law, usually is ignored by everyone except election bureaucrats. This year, however, it drew more interest because of the intense focus on the voting machines. Arguments over the accuracy of the machines led a federal court in September to order the recall election postponed, a ruling that was reversed.

The pink ballot cards were examined by teams of four inspectors seated facing one another and calling out chad numbers as though they were playing an odd game of doubles Battleship. As McCormack had predicted, the count showed little evidence of Florida-style hanging chads, dimpled ballots or other visible irregularities.

Researchers who have challenged the punch-card machines, however, say the hand-count is not definitive.

Those who criticize the machines point to figures showing that 8.7 million Californians cast ballots in the recall election, but only about 8.3 million were recorded as having voted on the recall question.

The 400,000 people who did not vote on the recall were not evenly distributed across the state. Voters in the seven counties that used Florida-style punch cards in the recall were significantly more likely than those using other types of voting systems to leave the recall question blank, with about 1 in 16 ballots registering no vote.

Several researchers around the country say the uneven distribution of the blank ballots indicates that the punch-card machines failed to record some people’s votes. The numbers would not be nearly enough to change the results of the election, but they do indicate continued problems with punch-card machines, those researchers say.

The seven punch-card counties -- Los Angeles, San Diego, Mendocino, Sacramento, Santa Clara, Sierra, and Solano -- account for about 40% of the state’s voters.

In Los Angeles County, which had the most blanks, about 180,000 people -- or 1 of every 11 -- did not vote on the recall question.

By contrast, fewer than 1 in 100 voters in Alameda County -- which uses touch-screen voting technology -- abstained in either race. Los Angeles voters who used touch-screen machines during early voting skipped the recall question even less frequently, with only 149 out of about 45,000 voters abstaining.

“The numbers all point to the punch cards,” said Henry E. Brady, a professor of political science at UC Berkeley. Brady was the main expert witness in the federal court case brought by the American Civil Liberties Union last month against the use of punch-card voting systems in the recall.

McCormack disagrees, saying the missing votes result from substantial numbers of voters choosing to skip the recall question, voting only on candidates to replace Gov. Gray Davis or on the two ballot propositions.

According to county figures, the precincts where most voters skipped the recall question were geographically scattered. Westside voters were most likely to cast recall votes, but pockets of blue -- indicating high blank rates -- dotted a large, color-coded map McCormack had posted on the wall of the counting room.

A loose band stretched across immigrant-heavy areas from East Los Angeles through the San Gabriel Valley, but precincts on the Palos Verdes Peninsula and in Long Beach, Whittier and La Mirada also showed heavy rates of voters not casting ballots on the recall.

By contrast, people who voted on the recall but did not choose a replacement candidate were heavily concentrated in South Los Angeles from mid-city south through Compton, areas with large numbers of black voters. That finding would be consistent with the theory that a substantial number of black Democrats voted against the recall but did not want to vote for the leading Democratic candidate on the replacement ballot, Lt. Gov. Cruz Bustamante.

Michael Alvarez, a professor at Caltech who specializes in voting technology, said it was possible that people missed the recall question because they were confused by the layout of the ballot or the rules of the recall, rather than thwarted by the machinery.

A Times poll taken a week before the election indicated that 17% of voters were confused about which parts of the ballot they could vote on.

Exit polls indicated that about half of those who skipped the recall question did so on purpose, possibly as a form of protest or out of indecision.

McCormack admitted that some voters might have missed ballot questions despite efforts to educate them about the recall ballot and the punch-card machinery in the weeks before the election.

“But any assessment of intention is just speculation at this point,” she said, adding that she knew many people who had planned not to vote on the recall question. “You just can’t know.”