In South Florida precincts, only minor discrepancies found
By ALFONSO CHARDY and DAVID KIDWELL
Miami Herald 27 November 2004
MIAMI - When Broward County officials discovered a computer glitch that miscounted thousands of absentee ballots just two days after the election, fears immediately surfaced about the accuracy of all electronic returns in Florida.
But a check of 18 polling precincts ed at random by Miami Herald reporters in Miami-Dade, Broward and Palm Beach counties revealed no major discrepancy between the number of voters who signed in on Election Day and the number of votes counted by electronic machines.
Herald reporters found minor discrepancies in many precincts, typically a difference of a handful or so between the number of voters who signed in and the number of ballots counted. Election officials gave several explanations, such as voters whose failure to sign in was not detected by poll workers.
Officials maintain there is no evidence to suggest that the vote was compromised, or that Broward's glitch was more than a one-time anomaly.
"We have not noted any discrepancies on this end, and the minor discrepancies that you were talking about are handled at the local level by the supervisors of elections," said Jenny Nash, press secretary for Florida Secretary of State Glenda Hood.
The state's canvassing board certified the election Nov. 14, formally declaring George W. Bush the winner in Florida over John Kerry by nearly 381,000 votes, a tally that helped the GOP maintain its hold on the White House.
Fears about the electronic count arose Nov. 4, when Broward authorities discovered the absentee-ballot glitch.
Officials say the ballot machine software started to subtract votes after the absentee tally hit 32,500. Such ceilings are sometimes set to safeguard against ballot stuffing, but the limit should not have been in place for a bushel of absentee ballots that clearly exceeded the threshold.
After correcting the error, officials discovered thousands of "missing" votes for five proposed state constitutional amendments. As a result, the outcome on Amendment 4, to allow a referendum on introducing slot machines at parimutuels in Miami-Dade and Broward, was reversed, and the amendment was approved. Counts on the other questions had to be adjusted, but the outcomes were not affected.
The glitch was traced to software in the central tabulation machine, which collects data from optical scanners that read individual mail-in ballots.
Broward officials blamed Election Systems & Software, the company that sold it the machines and software, for not fixing a glitch first brought to the firm's attention in 2002. Company spokeswoman Becky Vollmer has said the flaw will be fixed in s.
The error once again triggered concern about the integrity of voting in Florida, which was plagued by widespread Election Day snafus in 2000.
There is no way to recount ballots manually in the current electronic, touch-screen systems used in South Florida.
To check on whether this was a one-time flaw or a sign of larger woes, Herald reporters back-checked the tally by counting about 10,000 voter signatures at 18 randomly ed precincts in South Florida, and then comparing those numbers to the actual vote total.
In three precincts - two in Palm Beach, one in Broward - the numbers were identical. In 15 others across the three counties, the number of signatures differed from the final electronic tally - typically by one to five votes. The largest disparity was 26 in one Broward precinct and 10 in one Miami-Dade precinct.
Seth Kaplan, a spokesman for the Miami-Dade Elections Department, said discrepancies likely stemmed from a few people who signed in but didn't vote, or some who failed to sign in, with that going unnoticed by poll workers.
Broward officials added a third explanation: voters who moved and did not advise the department of an address change, meaning that their names did not appear on the signature register, although their votes did count.
Meantime, Congress' investigative arm, the Government Accountability Office, said last week it will review the Nov. 2 vote count nationwide. House Democrats who pressed for such a review had cited Broward's gambling measure mix-up as one of several voting problems across the United States.