Explore Harvard's Nieman network Nieman Fellowships Nieman Lab Nieman Reports Nieman Storyboard

Ask your election officials: Are you following e-voting issues in the news?

ASK THIS | May 129, 2006

A blogger who has been obsessively tracking the growing number of electronic voting horror stories around the country -- stories largely ignored by the national media -- offers an annotated list of questions to help election officials learn from others’ mistakes.


By Brad Friedman
Brad@BradBlog.com

Electronic voting has suffered from continuing and epidemic failures across the country in the primary elections so far in 2006. And the private companies receiving billions of federal dollars from the Help America Vote Act as local governments “upgrade” their election systems have time and again proven their inability to properly service election officials and fulfill contractual obligations.

Even this early in the election season, machines from all of the companies have proven to be disastrous for the few states that have already had primaries or have them scheduled just around the bend.

Note: Click here for other items on Voting Security from NiemanWatchdog.org

Q: Since Electronic Systems and Software (ES&S), the largest voting machine company in America, has failed in at least half a dozen states so far -- including Texas, Oregon, West Virginia, Pennsylvania, Indiana, Missouri and Arkansas -- to provide ballots, programming or even machines in some places in time for the start of elections (and as specified by their contracts), what specific plans do you have in place to provide a back-up voting system if they, or one of the other vendors, fail to supply ballots, machinery and/or programming in time for your election this year?

Monumental, yet scandalously under-reported failures by ES&S this year alone have led to investigations, official hearings, election contests, legal actions or threats thereof in all of the states named above, and several more are likely to take similar actions soon.

See, from Indiana:

  • Counties dealing with ballot problems, Indianapolis Star, April 7, 2006
  • State probes voting machine vendors, Indianapolis Star, April 13, 2006
  • Voting machines pulled, Indianapolis Star, April 14, 2006.
  • Five questions for: Doris Anne Sadler (Marion County Clerk), Indianapolis Star, April 30, 2006: “Q. Which is harder to manage, your two children or ES&S? A. Oh, ES&S, definitely. My children are really very easy. In fact, at times I think my children would have done a better job with the voting machines. And they're (ages) 7 and 4.”
  • Voting probed in southern Indiana, Associated Press, May 4, 2006: “State officials will investigate problems with voting systems in four Indiana counties where determining winners in Tuesday’s primary became so cumbersome that workers in one county gave up tabulating returns.”
  • State to probe computer problems, Louisville Courier-Journal, May 4, 2006: “Computer problems in four Southern Indiana counties made determining winners in Tuesday's primary chaotic, but state officials said yesterday that voting went smoothly elsewhere. Secretary of State Todd Rokita plans to send his chief counsel to Clark, Harrison, Jackson and Washington counties this week to investigate problems with voting systems sold and maintained by Nebraska-based Election Systems & Software, his office said yesterday.”

From Texas:

  • Bexar's early voters to put pen to paper, San Antonio Express-News, April 27, 2006: “Bexar County's electronic voting machines — for which it paid $8 million — will be unplugged when early voting starts Monday in local elections because the equipment supplier failed to deliver the necessary software, according to county officials. The county Elections Department instead will use emergency paper ballots and may have to do so again on election day, May 13.”
  • Early voters, be prepared to write, Wichita Falls (Tex.) Times Record News, April 28, 2006: “ES&S's lack of election readiness prompted a letter from Secretary of State Roger Williams. In the letter, Williams threatened to withhold state money, decertify ES&S as a Texas vendor and find the company in breach of contract if it did not increase its election efforts in Texas, according to Trey Trainor, general counsel for the secretary of state.”

From West Virginia: 

  • Election test delayed, Charleston (W.Va.) Gazette, April 29, 2006: "A deadline to test electronic voting equipment in West Virginia was pushed back to the day before the May 9 primary election because of various problems or delays with the new machines....Some counties had a software glitch in their new electronic systems and others have not received the necessary software to carry out the election electronically. Because of the delays, counties were bearing down on a required deadline to have equipment tested a week before the election." 

From Arkansas:

  • Central counties' early ballots going paper as screens fail test, Arkansas Democrat-Gazette, May 5, 2006: “Election officials scrambled Thursday to ready paper ballots after learning that touch-screen voting machines in Pulaski County and several other central Arkansas counties won't be ready for the start of early voting just three days away. In Pulaski County, the state's largest election jurisdiction, programming for iVotronic machines supplied by Election Systems & Software proved flawed in late-afternoon tests.” 

Q: Since all of the major voting machine companies have had machines simply fail to work, for a variety of reasons, during elections in 2006, what specific back-up plans will you have at the polling places to ensure that citizens will still be able to vote if the electronic voting machines fail to operate?

Machines have failed to start at polling places in 2006 in almost every state which has held an election so far this year.

Q: Since thousands of memory cards on ES&S systems have recently been found to have completely failed in pre-election tests in several states, have you tested all of the memory cards for your systems, no matter who the vendor, to determine if they are working or not? And do you have back-up plans should those cards fail on Election Day?

Memory cards store vote tabulation and ballot definition data. If they fail to work, machines may refuse to start up and/or vote totals may not be stored correctly or at all. In recent pre-election tests in Summit County, Ohio, more than 30% of cards tested completely failed. In North Carolina, after Election Integrity Advocates notified the state about the problems in Ohio, more than 1000 memory cards were found to have failed. Neither the Elections Assistance Commission (EAC) nor the vendors have notified jurisdictions appropriately of this recently emerging, and very serious, concern.

See:

Q: So far, memory cards have been lost after elections in at least two states this year. Since we know that vote totals can be corrupted on those memory cards after an election, what specific plans do you have in place to ensure the proper chain of custody for such media after the close of polls on Election Day?

Diebold memory cards went missing in Cuyahoga County after the May 2, 2006 Ohio primary, and more than 400 disappeared from Sequoia machines in and outside of Cook County after the March 21, 2006 primary election in Illinois.

See, from Ohio:

  • People, machines account for glitches in Cuyahoga voting, Cleveland Plain Dealer, May 4, 2006: “A day after the polls closed, Cuyahoga County elections officials still couldn't say who won Tuesday's elections…. Problems with both types of the county's new electronic voting machines were to blame. Human error added to the crisis, with one in five elections workers failing to report to work Tuesday. And many who did show up struggled with setting up the new machines. With the touch-screen machines that voters found in their precincts, the problem was with memory cards, the brains of the machines. Elections workers lost cards for about 50 machines. Workers actually lost 70 of the memory cards, they realized Wednesday, but they found 20 of them by evening.” 
  • If the election was held Tuesday . . . Why are we still counting? Cleveland Plain Dealer, May 4, 2006: “1. Testing was too little, too late; 2. Ballots had printing problems; 3. Polling places were not ready; 4. Memory cards were lost.”
  • Elections board opens probe into voting fiasco, Cleveland Plain Dealer, May 9, 2006: "Blame was widely spread, with board members especially critical of Diebold Elections Inc., manufacturer of the touch-screen machines used at the polls and optical-scan machines used to count absentee ballots."

From Illinois:

  • New machines, poor training slowed count; Precincts uncounted even after Wednesday, Chicago Tribune, March 23, 2006: ”As ballot counting stretched into a second night, judges and election officials on Wednesday blamed confusion in the primary election on vast numbers of poll workers who had not been trained for Cook County's complicated new electronic voting system…  At noon Wednesday, Chicago was missing 252 memory cartridges, 93 from machines that scanned in paper ballots and 159 from touch screens. County officials couldn't find 162 memory cartridges from suburban precincts--68 from optical-scanning machines and 94 for touch-screen balloting. The problems led midday to the sight of Cook County Director of Elections Clem Balanoff--his tie loosened and eyes bleary--rifling through blue duffel bags at a county warehouse for precinct returns for uncounted votes on missing memory cartridges. 

Q: In making your decision about which voting system to go with in your state/county, have you factored in maintenance costs for these systems? If so, how much are those maintenance costs estimated to be, and where will the money come to pay for them?

Several states have been shocked upon learning that maintenance costs were some 1000% higher than previously expected by elections officials and made clear to them by the electronic voting machine vendors. One of those complaints came recently from Maryland's Republican Governor Bob Ehrlich.

Additional, usually overlooked, "hidden costs" from the vendors include technical support contracts, licensing contracts, ballot programming contracts and ballot printing contracts amongst other fees that have surprised many boards of elections only after they have already committed to purchasing machines from a specific vendor.

See:

  • Feb. 15, 2006 letter from Md. Gov. Ehrlich to the state Board of Elections in which he alleges: "The actual cost [of the statewide voting system], which has been financed by the State Treasurer was $65,564,674 – an almost 78 % increase from the original cost estimate. However, this misjudgment pales in comparison to the 1000% increase for estimates of the annual maintenance costs for this system. The 2001 fiscal note estimated such maintenance would be $858,000. For the upcoming fiscal year, the State Board of Elections requested $9,528,597 for these costs. The cost of Maryland’s Diebold voting machines has skyrocketed as our confidence in the system has plummeted.”

Q: In light of the fact that several electronic voting machine vendors have been found to have installed federally uncertified software on voting machines around the country, what specific actions are you taking to ensure your vendor doesn't install any uncertified (either state, federal or both) changes to your voting systems? Do you know what "version control" is and have you questioned your vendor about it?

The installation of uncertified or incorrect versions of software and software patches onto voting machines is a growing problem. Just a couple of examples include MicroVote having been found to have installed software in Indiana which was neither federal nor state certified. They may be facing a $300,000 fine for each instance. Diebold was decertified in California in 2004 when it was discovered they were doing the same thing (and lying about it.)

See:

  • Indianapolis-Based Voting Machine Maker Faces Tough Questions, WISH-TV, April 18, 2006: The state's lawyer claimed that Microvote sold uncertified voting equipment in at least two counties, that it supplied and permitted the use of unreliable voting equipment and that it marketed software that may not be approved for use during the May primary. For more than two years, I-Team 8 has exposed problems with Indiana's voting machines and the companies that make them. 

Q: Do you plan to audit the paper ballots used with optical-scan systems -- or the "paper trails" created by touch-screen systems -- to determine the accuracy, after the election, of the machine counts?  If not, how can you have any certainty that the machine count is the same as the paper count? And if so, what percentage of those ballots or "paper trails" do you commit to manually counting in order to determine if the machines accurately tabulated the election?

The only way to ensure that the totals reported by electronic voting machines are accurate is with a manual hand-counted audit of randomly selected portions of those ballots or the "paper trails."  The number of ballots selected for hand-count must be large enough to ensure a scientific certainty that the sample counted will reveal most discrepancies found with the machine tabulated totals.

Q: If there is a discrepancy discovered between the machine count and the paper count (or audit) which count will be the official one? Do you have rules in place before the election so such a decision won't have to be made afterwards if there is found to be such a discrepancy?

Stunningly, many election officials I have spoken with, even if they plan to recount ballots or "paper trails" after the election, have no procedure in place to determine which will be the official count in the event of a discrepancy. Waiting until after the election, when the problem is revealed, to determine such procedures is not acceptable and will lead to nothing but trouble.

There is no way to validate the count made by the software itself. So the only verifiable count is the paper, which must be the official "ballot of record" in such a dispute.

Q: For states and counties that use touch-screen machines with "reel to reel" paper trails: Have you run any tests prior to the election to determine if the "voter-verified paper trail" is actually manually hand-countable in the case of an election contest or manual audit? If so, how many people, how much time and how much money have you determined will be required for such an audit?

The so-called "reel to reel" or "toilet paper rolls", akin to the thermal tape used for credit card purchases, is what makes up the "paper trail" in most electronic DRE, touch-screen systems. The paper is one long continuous roll, with very small lettering, and almost impossible to count manually, at least without a very long and labor intensive effort. So far, as far as we know, nobody has ever done a full manual hand-recount of the "paper trails" used with touch-screen voting systems.



The NiemanWatchdog.org website is no longer being updated. Watchdog stories have a new home in Nieman Reports.