ACLU of Florida analysis of racial impact of Miami DRE's

From: Dennis <dpaull_at_svpal_dot_org>
Date: Sun Nov 16 2003 - 17:50:07 CST
Hi all,

Here are several articles about studies of the 2002 Florida election.

In particular, there were a number of folks that possibly left the voting
both without clicking the final 'CAST YOUR BALLOT' button. Poll workers
were supposed to check for that, but either didn't at all or missed quite
a few. I can believe that they were busy doing other tasks.

In any case, this was quite a problem. Anyone know of a way to solve it?

Of course if there is a separate optical scan ballot reader that the
voters cast their printed ballot into, this is not a problem.


Date: Sun, 16 Nov 2003 15:16:49 EST
Subject: [Votingtech] ACLU of Florida analysis of racial impact of Miami DRE's
List-Id: <>
List-Unsubscribe: <>,
List-Archive: <>
List-Post: <>
List-Help: <>
List-Subscribe: <>,


I thought people would be interesting to know about this study done of major problems with performance of Miami's new DRE's in the hotly contested gubernatorial primary in September 2002 -- the one in which Janet Reno lost by a small percentage to Bill McBride on the Democratic side. There was a shockingly high rate of no-vote-registered in the top-level election. According to the ACLU of Florida, these problems disproportionately affected people of color and low-income people.

Martha Mahoney, a law professor at the University of Miami and one of the speakers at our Claim Democracy conference next weekend alerted me to the study, and I've included her email along with the ACLU of Florida news release. She's copied on this email too, if people want to ask her any questions.

I draw from this a particularly important lesson: under our current election administration regime, with a limited federal commitment to improving voting and much decentralized decision-making, the current vendors cannot be trusted to produce the best-possible products, and counties can't be trusted to buy the best of those possible products.

Imagine an alternative reality, one in which there were strong federal protections of the right-to-vote (grounded in a constitutional amendment that protects that right, which our Constitution currently lacks): in this reality, the federal government would have used its immense resources to invest in developing voting technologies that were truly cutting edge, with open source software, public interest goals incorporated without resistance, capacities to upgrade easily with new advances in technology, etc. Counties and states would have needed to use this equipment or something that met the same high standards as this equipment.

But no, we're stuck with the current vendors -- relatively small for-profit companies cutting corners to make a buck, fighting easy-to-do capacities like being able to handle ranked-choice elections and producing machines that aren't nearly as good as they should be. And counties and states are jumping to buy them because they don't know if the money's going to be there in the future. If that money isn't there in the future -- which  is all too possible, given our shoddy federal commitment to the right to vote - they then will be stuck with that inferior equipment for years to come.

Call it democracy in America -- on the cheap. Let's keep that in mind the next time our leaders lecture other nations about how important we think democracy is. And I sure wish the debate over voter-verified paper trails focused its finger on the real problem, which is our acceptance of a failed election administration regime.

- Rob Richie, Center for Voting and Democracy

[ From Martha Mahoney]

Much recent discussion about voting technology and paper trails seems to be framed as a tension between computer security to protect against fraud and an interest in new technology that is expected to be of particular benefit to racial minorities, language minorities, and disabled voters.  This framework fails to reckon with the experience of voters and civil rights activists in Miami-Dade County, a flashpoint for voting problems.  I want to make sure that the civil rights questions that have emerged in Miami are addressed in this debate.

The iVotronic direct recording electronic (DRE) touchscreen system used in Miami has been certified by Florida and other states and adopted in other jurisdictions as well as Broward and Miami-Dade County.  In September 2002, the new system collapsed during the first election in Miami in which it was used.  The Florida ACLU studied the racial impact of voting problems in that disastrous primary.  See The 198 precincts across the county that reported problems with the machines had an average African-American voter registration of 27%, much higher than the county average of 19%.  The 31 precincts that reported most problems had an average African-American registration rate of more than 50%.  As in the 2000 elections, the impact of voting problems fell most heavily on African-American voters. 

In those 31 problem precincts, ACLU researchers counted the number of people who signed in to vote and compared those numbers to the votes that were recorded on the DREs.  They found a lost vote rate of 8.2% -- 18,752 voters signed the rolls to vote, but only 17,208 votes were recorded; 1,544 votes were lost. The Miami Herald reported that the rate of over- and under- votes in the same precincts on the infamous Votomatic punch card ballots had been 6.75% in 2000.  Andrea Robinson, ACLU: Black Voters Disenfranchised, Miami Herald, Oct. 22, 2002, at 3B.  In other words, in those precincts, the percentage of lost votes was actually higher in 2002 on the new touchscreen machines.

On the new system, disenfranchisement was more difficult to prove:  the DREs without paper records do not provide evidence about what happened between signing in and counting votes.  We cannot prove that the people who signed in actually voted.  Voters sign the rolls late in the process of moving toward the machines-yet there is no way to prove that they did not leave without voting.  Voters may have failed to push the "VOTE" button.  Pollworkers are trained to push the "VOTE" button if voters do not, but it is possible that pollworkers also failed to push it.  It is also possible that the machines failed to record the votes properly.  In short, no one can prove that these "lost" votes were ever cast or, if cast, why they were not recorded.  The repeated experiences of uncounted votes have had serious consequences.  The NAACP finds that people are less interested in registering and voting because the failures of 2000 and 2002. 

Voter-verified paper trails would help address the problem of lost votes-it would be easier to identify when a vote had been finally recorded, and the voter could have verified the official record.  (Lost votes are not unique to Miami-Dade or to the iVotronic system.  See, e.g., George Bennett, 1 Race, 2 Candidates, 78 Votes for Neither?, Palm Beach Post, Mar. 29, 2002, at 1A, describing an election held on a Sequoia touchscreen system.)  Additional problems with the iVotronic system further revealed the imperfect nature of this new technology and the need for further change and improvement to meet civil rights concerns.    

Improved access for language minorities is one of the theoretical advantages of DREs, but the iVotronics had difficulty with the software for trilingual ballots.  Trilingual programming (English/Spanish/Creole) required very long boot-up times for the machines, creating huge problems in the September 2002 primary.  In November, the elections department decided to turn on the machines the night before the election and have police guard them overnight.  Elections officials did not always produce zero tapes in the morning on every machine to verify that no ballots had been added overnight.  That practice could be improved, although obviously it will remain subject to human error.  The elections department reports that a newly certified software system now makes boot-up time for trilingual ballots much faster.  They still plan to turn the machines on overnight and to hire security guards in locations that do not have adequately secure locations for the machines.  The additional financial cost and policy considerations should weighed in the discussion about voting technology.  .

Imperfections in the touchscreen system also affected blind voters.  The DREs are supposed to provide a secret ballot for visually impaired voters using an audio menu.   In January 2003, the County Manager published his Post-Election Analysis and Recommendations reviewing the November election.  The  iVotronics were reported to have a poor audio menu.  "Success rates for the use of these devices by the sight-impaired were not acceptable." Analysis&Recommendations.pdf at 10.  The report proposed installing a second set of headsets and a visible screen, so that assistants can help blind people vote.  Id. at 10, 21.  Those proposals would have taken away both privacy and independence for blind voters, but the machines have not been changed since last year's elections.

he experience in Miami proves that state certification does not insure against flaws in the operation of touchscreen computers. The arguments for and against voter-verified paper records have focused more on fraud than on new forms of disenfranchisement.  The idea that significant numbers of votes could disappear in minority communities seems largely absent from the discussion.  I take seriously the concerns that have been raised about difficulties that may be created by adding individual paper records to existing touchscreen systems.  On balance, though, I have come to believe that it will be necessary both to provide voter verified records and to make sure that verification methods provide equal privacy and independence for sight-impaired voters, whether this is done by adding paper plus additional audio verification to DREs or by adopting other voting systems that do not conceal the evidence of their failures.   

I think that most of the nation learned lessons that were far too limited from the 2000 election in Florida.  The public seems to have learned that punch-card ballots were unreliable and that recounts are nightmares to be avoided, when the better lesson would be that it is vital to adopt and maintain voting technology to ensure that no attempts to vote are lost and that every vote cast is counted accurately.  To increase political participation, it should be a priority to rebuild the faith in the voting system of the people who have suffered the worst impact of election failures.  Voters elsewhere need not experience the frustration that has happened to voters in Florida--but to avoid it, the Florida experience will have to be addressed.    

Note:  The Miami-Dade Election Reform Coalition, which is advocating the adoption of voter-verified paper records, includes the NAACP, ACLU, Southwest Voter Registration and Education Project, Fanm Ayisyen Nan Miyami (Haitian Women of Miami), Brothers of the Same Mind (an organization of and for ex-offenders), and other community groups as well as diverse individual activists from around the county.  I've been a citizen member of the Coalition since our first meeting last September, but this letter represents my personal position on the issues.  See

I look forward to hearing what you think -

Best wishes,

Marnie Mahoney

Martha R. Mahoney
Professor of Law
University of Miami School of Law
Box 248087
Coral Gables FL 33124-8087


>From ACLU of Florida

Analysis of September 10th Voting Fiasco in Miami Dade Demonstrates Disproportionate Impact on Racial Minorities, ACLU Says 

October 21, 2002

Miami, FL – African-American voters were disproportionately disenfranchised by Miami-Dade County's chaotic September 10th election, according to a report released today by the American Civil Liberties Union (ACLU) of Florida.  An examination of 31 problem precincts revealed at least 1,544 lost votes, approximately half of which were from African-Americans, and a "lost vote" rate of 8.2% of all voters who went to the polls and signed in to vote.

"Not only are there a significant number of missing votes, but there's also an alarming racial disparity in the errors that occurred during the last election," said JoNel Newman, the Special ACLU Staff Attorney leading the ACLU's voting probe.  A team of volunteer law students from the University of Miami assisted Newman. 

"That the African-American community was disproportionately affected on September 10th is particularly egregious after the well-documented disparities of November 2000.  This pattern of racial discrimination in voting simply cannot be allowed to continue," Newman added.

The study was conducted as part of monitoring of the settlement agreements in NAACP v. Harris, the landmark voting rights case filed after the November 2000 Presidential election.  Among the legal advocacy groups now monitoring compliance with the settlement agreements are the ACLU, Lawyers Committee for Civil Rights Under Law, NAACP Legal Defense & Educational Fund, The Advancement Project, and People for the American Way Foundation.

Through the legal team's public records requests, the ACLU obtained a list of 198 precincts reported to have experienced machine malfunction on September 10th.  Research revealed that those 198 precincts have an average African-American voter registration of 27%, a percentage far higher than the countywide 19% average.  Of those 198 precincts, the legal team focused on 31 for which they had received the most complaints.  The racial disparities in those precincts rise even higher, with an average African-American voter registration rate of 51%.

An in-depth analysis of those 31 precincts revealed that:

§ 18,752 voters signed the rolls to vote, but only 17,208 votes were recorded.  That is, 1,544 votes were lost due to election-day errors that included voters not being able to vote because working machines were not available or the failure of poll workers to cast the ballot as required when voters leave without pressing the red VOTE button.

§ The analysis indicates that voters were disenfranchised at a rate of 8.2% due to "lost votes" resulting from the County's faulty election-day procedures.  (During the November 2000 elections, voters who voted with punch card voting machines were disenfranchised at the rate of 4.37%, which was the spoilage rate attributed to the punch card process.)

§ Because the average African-American voter registration rate is 51% in those precincts, approximately half of the 1,544 lost votes were of African-American voters.

§ If the 8.2% "lost vote" rate is extrapolated to all 198 precincts with reported voting problems, there were 8,500 lost votes.  Those lost votes represent the tip of the iceberg, since they do not account for voters who were disenfranchised because they were not able to sign in at all due to total polling place closure.

§ While the average "lost vote" rate in the 31 targeted precincts is 8.2%, in some precincts that rate is as high as 21.5%.

Dr. Hugh Gladwin, Associate Professor in the Department of Sociology and Anthropology at Florida International University, conducted a statistical analysis of the report for the ACLU.  Dr. Gladwin compared the rate of problems at majority (+50%) black versus majority non-black precincts and determined that there was a large disparity.

"There was a 15% problem rate in non-black precincts versus a 28% problem rate in majority black precincts.  The probability this could have occurred by chance is infinitesimal – .00055," Dr. Gladwin noted.  Dr. Gladwin also controlled for poverty versus race as a predictor of problem precincts.  "Poverty was not a factor that mattered, race was," he added.

Presently, the legal team of NAACP v. Harris has sent Miami-Dade Supervisor of Elections David Leahy a Notice of Material Breach, as required by the settlement agreement before further enforcement takes place.  Mr. Leahy has until October 24th to respond, after which the legal team will review his response and consider alternatives.

The racial disparities revealed by the ACLU echo those of November 2000, detailed extensively by the U.S. Commission on Civil Rights.  The Commission's report showed that while African-American voters made up only 11% of the electorate statewide, about 54% of the state's 180,000 spoiled ballots were cast by African-Americans.  In Miami-Dade County in 2000, African American voters were three times more likely than white voters to have their ballots rejected.  The Commission concluded that, "The disenfranchisement of Florida's voters fell most harshly on the shoulders of black voters."  The full report of the U.S. Commission on Civil Rights can be found online at

According to the Governor's Select Task Force on Election Procedures, Standards and Technology, Miami-Dade County's rate of blank or spoiled ballots under the punch card voting method in the November 2000 elections was 4.37%.  The full report by the Governor's Task Force can be found online at

For a copy of the complete report contact:

Alessandra Soler Meetze, Communications Director,
(305) 576-2337 ext. 16 (office) or (786) 208-7203 (cell)

Click here for more on voting rights

2002 Press Releases

Back to the ACLU of Florida Home Page

ACLU of Florida 4500 Biscayne Blvd. Suite 340
Miami, FL  33137 (305) 576-2336
General ACLU e-mail:


votingtech mailing list

================================= = The content of this message, with the exception of any external = quotations under fair use, are released to the Public Domain ================================= Received on Sun Nov 30 23:17:03 2003

This archive was generated by hypermail 2.1.8 : Sun Nov 30 2003 - 23:17:12 CST