Ohio: Punch cards may hurt Blacks

• Originally published in THE COLUMBUS DISPATCH on October 17, 2004 [more ]
  

Copyright © 2004, The Columbus Dispatch

By: Darrel Rowland

Presidential votes from Ohio's predominantly black neighborhoods will go uncounted at nearly three times the rate of those from the rest of the state on Nov. 2 if the 2000 election is any guide.

The prime suspect for the disparity? Punch-card ballots.

The Dispatch conducted the first precinct-by-precinct computer analysis that combined Ohio voting results from 2000 with racial data from that year's U.S. Census. The objective was to examine the state's 94,569 residual, or uncounted, presidential votes.

In areas with the highest population of blacks, the rate of ballots with no votes counted for president was about 5 percent, the analysis shows. For the rest of Ohio precincts it was less than 2 percent.

The disparity is even more glaring if you hone in on the parts of Ohio that had the highest 10 percent of uncounted ballots. The odds of a predominantly white precinct making that list were 2 out of 33. But for predominantly black precincts, it was 2 out of 3.

Every one of those black precincts used punch-card ballots in 2000 -- and plans to again this year.

A precinct is considered "predominantly" black or white if either race makes up at least 90 percent of the voting-age population.

Elections officials have stepped up voter-education efforts -- especially in counties using punch-card ballots -- because the stakes are so high this year.

If the presidential race in Ohio winds up as close as many predict, the uncounted votes could make a difference in whether President Bush or Sen. John Kerry wins the state -- and possibly the White House. That's because the precincts with the highest rate of uncounted votes are heavily Democratic.

In 2000, Vice President Al Gore carried those precincts by a whopping 66 percent to Bush's 31 percent.

Gore lost statewide by 3.6 percentage points. (Bush's victory in Ohio four years ago cannot be attributed to uncounted votes because his margin was about 167,000, well above the 94,569 untabulated votes.)

The public-education effort will be especially crucial for about 1 million newly registered Ohio voters -- including more than 700,000 who could be using punch-card ballots for the first time Nov. 2.

Why the racial disparity in uncounted votes?

"Those variations are strikingly associated with poverty and lower education," said Herb Asher, an Ohio State University political-science professor who began studying punch-card voting more than two decades ago.

His comments are echoed by other experts.

The rate of uncounted votes is also high in Ohio's Appalachian counties, which are associated with higher poverty and lower educational levels, Asher noted.

Because voters use secret ballots, there is no way of telling for sure how many intentionally did not cast a vote for a presidential candidate. But studies and exit polls have indicated that the percentage of uncounted votes is far higher than the rate of voters who skipped the presidential race on purpose.

Punch-card ballots are especially prone to error because so many things can go wrong, as Americans discovered during the Florida recount in 2000. If the ballot is not laid out or aligned properly, or voters do not correctly insert the card into the voting device, the wrong holes can be punched.

The punch cards are subject to over-votes, when more than one hole is punched for a race. They also are subject to undervotes, when a clear punch is not made in the card, sometimes resulting in the infamous "hanging chads." In both cases, no vote is counted.

Ohio Secretary of State J. Kenneth Blackwell acknowledges he finds himself in the awkward position of administering a voting system that he once said "invites a Florida-like calamity."

But with increased poll-worker training and the extensive public-education campaign -- especially with minority groups -- he now says Ohio's voting setup "isn't so dysfunctional that no one should trust the outcome."

George Forbes, former Cleveland City Council president who is now head of the city's branch of the National Association for the Advancement of Colored People, said black leaders in Cuyahoga County are working with elections officials in an attempt to head off problems on Election Day.

"It's not going to do us any good to go in the day after the election and say we got cheated," he said.

Guides to using punch-card ballots are being mailed to all registered voters, Forbes said, and every precinct in Ohio's largest county will have a coordinator with a direct line to the county elections board for troubleshooting during the vote.

"I still anticipate problems," Forbes said. "I think that it's going to be an overload of people at the polls," which might make some voters reluctant to ask for assistance.

In 2002, the American Civil Liberties Union sued the state over alleged racial disparities stemming from punch-card voting in Hamilton, Montgomery and Summit counties. A federal judge's decision won't come until after the election.

"The (Dispatch) analysis confirms what the ACLU has claimed for a long time: that punch-cards are harmful to all voters but especially to African-Americans," said OSU law professor Daniel Tokaji, who is representing the ACLU in the lawsuit.

"The direct result of the state's use of the hanging-chad punch card is that thousands upon thousands of voters are lost in each election. A disproportionate number of votes lost are those of African-Americans."

In a landmark 1982 study shortly after punch-card ballots became prevalent in Ohio, Asher found that thousands of voters were disenfranchised because of punch-card ballots. He also examined results from numerous urban precincts in the 2000 presidential election. In both instances, Asher found that proportionally more votes went uncounted in black areas.

But John Lott, a resident scholar with the American Enterprise Institute, said studies that focus only on uncounted ballots in the presidential race are giving punch-card ballots a bum rap.

Lott, who testified on behalf of the state in the ACLU lawsuit, examined Ohio races for the U.S. Senate, House and state legislature, as well as president for the 1992, 1996 and 2000 elections.

Lott found that for down-ticket contests, the rate of uncounted punch-card ballots is actually lower than for electronic devices and voting machines, and the racial gap is reversed.

"When you look at it overall, if the issue is nonvoted ballot rate, punch cards actually do fairly well," he said.

Lott's critics say results from races that aren't on the statewide ballot cannot be compared on a statewide basis, but he said he factored out variations from area to area.

This year, 72 percent of Ohio voters live in areas with punchcard ballots -- the second highest in the nation behind Utah and by far the most of any presidential battleground state. Because many other states upgraded their voting systems after the 2000 election, more than a quarter of all the punchcard ballots in the country are now used by Ohio voters.

Promises to modernize the way that Ohioans vote fell by the wayside because of legislative roadblocks and security concerns about electronic voting. In the past four years only one county, Sandusky, has gotten rid of punch-card ballots, which means the portion of Ohioans using them this year is only slightly lower than the 74 percent in 2000.

Electronic-voting devices with voter-verifiable paper audit trails are supposed to be in place by the 2006 elections.

The Dispatch study looked at 2000 voting results obtained from the secretary of state's office for Ohio's 12,000-plus precincts. Census data on racial estimates for each precinct was obtained from Mark Salling, a professor with the Maxine Goodman Levin College of Urban Affairs at Cleveland State University.

Salling's data are generally accepted as accurate by those on both sides of a federal lawsuit in Akron over punch-card voting.

Salling, who was asked to review the Dispatch study, said, "I think there are very strong indications here that African-Americans (and voters in Appalachian counties) suffer more disenfranchisement than do whites from residual ballots that may be attributable to the punch-card technology."

One group that often does not cast a ballot in the presidential election are the Amish, concentrated in Holmes County. Even when that county is eliminated, Ohio still had 39 precincts in 2000 where at least one in 10 presidential votes went uncounted. In all, 93 percent of those precincts used punchcard ballots.

And at least one in 20 presidential votes was not recorded in 635 precincts, which contained about 390,000 registered voters. In those areas, 96 percent used punch-card ballots.

While the racial disparity in residual votes was strongest for voters using punch-card ballots, it also showed up with other voting methods, including electronic, ATM-like devices and voting machines using levers. In areas with optical-scanning devices -- in which voters color in spaces similar to those on standardized tests -- the residual voting rate was actually lower in precincts with heavier black populations than in white areas.

Comparing uncounted votes for each voting method between predominantly black and white precincts is somewhat difficult because every predominantly black precinct is in either Cuyahoga, Franklin, Hamilton, Montgomery or Summit counties -- and all but Franklin use punchcard ballots. Franklin is electronic.

However, the racial disparity for punch-card ballots also holds true when comparing precincts that are 80 percent black with those 80 percent white and when looking side by side at majority white vs. majority black precincts.

Blackwell, Asher and others have predicted the December 2000 ruling by the U.S. Supreme Court settling the contentious presidential race could have ramifications this year.

The justices cited concerns about constitutional "equal protection" rights because of variations in the way Florida conducted its recount. Some have contended that the same reasoning might apply to variances in the way votes are counted.

Asher said the number of uncounted ballots in this year's Ohio's presidential race could easily exceed the margin of victory for the winner.

"One could imagine a scenario where it wouldn't take a race as close as Florida to raise questions about the results," he said.

• Dispatch reporter Mark Niquette contributed to this story.