Published on Saturday, January 6, 2001 in the Miami Herald
1,700 Dade Voters Mispunched Chads
If the voters' cards had been aligned properly, the study found, Vice President Al Gore would have gained 316 more votes than the president-elect, George W. Bush.
by Andres Viglucci
As many as 1,700 Miami-Dade County voters invalidated their presidential ballots because they mistakenly punched the chad immediately below the one corresponding to their preferred candidate, a California researcher has found.
Those voters penetrated a meaningless chad -- one that didn't correspond to any candidate -- probably because their punch cards were not properly aligned with ballot books in the voting booth, said Anthony Salvanto, a faculty fellow at the University of California at Irvine's political science department.
If the voters' cards had been aligned properly, Salvanto said, Vice President Al Gore would have gained 316 more votes than President-elect George W. Bush.
The findings are the latest evidence that many voters whose ballots showed no presidential preference actually did intend to vote. The thwarted votes found by Salvanto represent more than 15 percent of Miami-Dade's 10,650 undervotes.
Salvanto and Miami-Dade elections officials believe the voters somehow failed to properly align their ballot cards with the ballot books in the voting booth, leading them to inadvertently punch the wrong chads. They surmise voter unfamiliarity with the county's punch-card machines might be the cause. Salvanto suggested alternatively that the ballot slots in some voting machines may have been misaligned.
Salvanto is conducting a national study of voting behavior and had arranged before the presidential election to obtain data for every ballot cast on Nov. 7 in Miami-Dade. The machines that read the ballots can be programmed to provide digital representations indicating how each card was voted. The data comprises only ballots that can be cleanly read by the machines, and thus includes no hanging or dimpled chads.
The pattern Salvanto found favored Gore: 1,012 voters punched No. 7, one below Gore's position at No. 6; 696 voters punched No. 5, one spot below Bush's position at No. 4.
Often the pattern was repeated throughout the length of a voter's ballot, Salvanto said.
On about 400 ballots, voters punched unassigned holes precisely one spot below those for Democratic candidates for U.S. Senate, commissioner of education and insurance commissioner as well -- leading Salvanto to conclude that those voters meant to support a straight Democratic ticket.
``It's particularly telling when they punch other unassigned chads that are one down from the Democratic candidate,'' Salvanto said. ``This happened all the way down their ballot.''
Even if the U.S. Supreme Court had not stopped the recounting of undervotes in Florida, however, it's unlikely these ballots would have helped Gore because they were clearly improperly punched.
But Salvanto, who studies voter behavior in polling booths, said it shows how some voters can become confused by voting systems to the point of botching their ballots.
``The voting machine is not an everyday appliance. That's why some people have advocated moving to more of an ATM-like system, which people see and use every day,'' he said.
``You're taking a piece of cardboard and sticking it into a slot. The vast majority of people do this correctly. But what we do see is that the system, by virtue of the fact that you have to take this flimsy piece of cardboard and put it correctly into the slot, is just flawed enough that it can impact a close election,'' Salvanto said.
His report also bolsters other evidence suggesting that black voters were disproportionately affected by balloting problems.
Each of Miami-Dade's 614 precincts had at least one such mispunched presidential vote, he said. Some had as many as 15, and one had 16. Thirty precincts, 19 of them predominantly black, had more than 10 such mispunches.
Salvanto theorizes that some voters may have been unable to push their ballot cards all the way into the slot on the voting machine, causing the holes on the ballot book to sit over invalid chads one spot below the correct position.
``The card was not put in all the way,'' he said.
But Miami-Dade Elections Supervisor David Leahy, while generally concurring with Salvanto's findings, said the voting machines have a safeguard that makes it impossible to punch through any chads at all unless the ballot card is fully inserted. The safeguard is a plastic shield that blocks the stylus from punching through the card in invalid spots.
Inserting the ballot all the way releases a spring that causes the shield to retract.
``We played with a machine yesterday for half an hour, trying out different things,'' Leahy said. ``But if the ballot is not fully seated, you just hit plastic. It doesn't punch.''
Instead, Leahy hypothesizes, those voters simply laid their ballot cards on top of the ballot book. If voters aligned the bottom of the card with the bottom frame holding the book, they would have the same one-spot-down misalignment, with the arrows corresponding to candidates' names pointing to invalid chads. Thus, a Bush voter would punch No. 5 instead of the right spot, at No. 4.
That scenario would require voters to ignore the fact that the number beside the candidates' names would not match the numbers on the chads by the arrows, Leahy said.
Another possible explanation is mechanical failure -- a misalignment in the voting machine slot or a malfunction in the spring mechanism, for instance. But Leahy discounted the possibility, saying the pieces in the slot fit together so tightly there is ``no room'' for misalignment.
And any mechanical problem would likely have produced more than the relatively few bad votes Salvanto found in every precinct, he said. Each machine was used by an average of 96 voters on Election Day, Leahy said. The average number of presidential votes cast per precinct in the county was 1,065.
Salvanto has not examined any machines and so can't say whether mechanical problems might account for the mispunches, but said the possibility requires investigation. If, for instance, the slots were slightly misaligned or a mechanical problem on a particular machine caused mispunches only intermittently, he said, it might explain why there were relatively few examples in each precinct.
``I would like to know how bad the problem is, if there is one. Does it happen every time a machine is used or just once in a while?'' he said. ``Even if there is a small problem, it's a problem. Every vote is supposed to count.''
Herald staff writer Geoff Dougherty contributed to this report.
Copyright 2001 Miami Herald