Why do we need voter-verified paper ballots (VVPBs) in 2004?

Numerous elections around the country have already been compromised by paperless DRE (direct recording equipment) voting machines lacking any provisions for performing a recount.

ü In a single-race election in Broward County (FL) in January 2004, 134 voters apparently failed to select any candidate. The margin of victory was 12 votes out of more than 10,000. Florida law requires a recount when the winning margin is less than 0.25%, but no recount was possible.

ü After a 2003 Washington, D.C. election, testing showed that approximately one out of every 100 votes cast for one candidate was not being recorded, a subtle error which was detected only by chance. That candidate lost by a margin of about 1% of the vote.

ü So many voting machine problems occurred in Hinds County (MS) in November 2003 that the Mississippi Senate had to declare the election invalid and schedule a new one in February 2004.

In other elections, problems are suspected but undocumentable for lack of an audit trail. In these and dozens of other recent cases across the country, no recount has been possible because there is no voter-verified paper ballot available from which the will of the voter could be determined.

While voter-verified paper ballots will not solve all of the problems with electronic voting machines, they are essential to represent the intent of the voter for the purposes of recounts and audits. We need voting systems that are robust, secure, transparent, and able to detect errors when they do occur. Paperless DRE voting machines meet NONE of these requirements.

What is a paperless DRE voting machine?

A paperless DRE voting machine both records and tabulates votes. Voter input is accepted through a touch screen or audio interface system and recorded internally. At the end of the day the machine produces vote totals. Although the machines may be connected to printers that produce these reports, they do not produce a paper ballot that can be verified by the voter.

Doesn't HAVA require these voting systems to produce a permanent paper record with a manual audit capacity for the system?

Senator John Ensign (R-Nevada), who wrote this section of HAVA, says that a VVPB was the intention, but voting machine vendors and some election officials assert that this requirement is met by the end-of-day totals and "ballot images". However, if errors are made in recording or storing votes, as in the examples above, the end-of-day printouts merely reproduce those errors. "Ballot images" produced at the end of the day are NOT ballots, because the voters never saw them.

Doesn't HAVA require that voting systems provide an opportunity for the voter to review and correct his vote?

Paperless DREs allow voters to review and correct their votes BEFORE they are recorded. But they provide no way for the voter (or anyone else) to verify that they were actually recorded as intended. Without a VVPB, no subsequent check can determine whether a vote was recorded correctly or incorrectly. In some cases it is clear that votes were not recorded at all, but even in these cases, there is no way to determine the intent of the voter.

Why isn't it enough that these systems have been tested and certified?

The failures cited above occurred on machines that had been certified by the state and had passed a standard battery of "logic and accuracy" tests. Other types of problems have been traced to commercial off-the-shelf portions of the software that were not tested at all. In several states, the certified software was modified or "patched" by the vendor AFTER certification but before the election. Existing tests and procedures have thus failed to prevent major, outcome-altering equipment problems.

Computer scientists know that it is impossible to create flawless software. For this reason, high-performance software systems are designed with many layers of redundancy that either prevent failure or make recovery possible. The type of redundancy in DREs (recording the vote in several places within the machine) is insufficient because a failure between the point at which the voter indicates that his selections are complete and the point at which his vote gets recorded cannot be recovered.

HAVA calls for the development of technical guidelines and the use of accredited laboratories for certifying machines, with the National Institute of Standards and Technology (NIST) playing a major role. As of spring 2004, the required committee has not been formed and the work has not been funded. The current certification processes are inadequate. For example, David Jefferson of the California Voting System Procedures Panel said in April, "If we had known then what we know now, none of these machines would have been certified." The use of voting machines with so many documented problems must be halted until comprehensive standards have been put in place.

Don't we have to buy these systems in order to meet HAVA requirements?

HAVA does not specify equipment or technologies to be used. It specifies intended outcomes. No jurisdiction is required to buy new equipment if its existing voting system can be upgraded to meet these requirements. And in fact, Section 301(c)(2) of HAVA explicitly forbids the interpretation of the Act in a way that would make it impossible for a paper ballot voting system to meet its requirements.

Section 301 of HAVA requires that any voting system used shall allow individuals to vote privately and independently even if they have physical disabilities (including visual impairment) or limited English (including speakers of languages with no written form). It also requires that voters be notified when they have selected more than one candidate for a single office and allowed to correct their ballot. There are other ways to meet these requirements besides paperless DREs, including ballot printers with audio interfaces for the visually or linguistically limited that produce printed ballots to be tabulated by ordinary optical scan systems. The paper ballot, whether marked by the voter or the printer, can then be retained for recount and audit purposes. Additional technology is likely to be in place in time to meet the 2006 HAVA deadline.

But don't paper ballots disadvantage the visually impaired and other non-readers?

A December 2003 opinion by the U.S. Department of Justice says that systems using voter-verified paper ballots are permissible as long as all voters are afforded equivalent opportunities to verify their ballots. A ballot scanner with an audio interface can meet this requirement for the visually impaired or linguistically limited. Optical scanners already provide information to sighted voters about over-voting, giving them an opportunity to make corrections on the recorded ballot before it is counted.

Isn't it too late to change anything now in time for the November election?

IT IS NOT TOO LATE to halt the rush to purchase paperless DREs for the upcoming election.

IT IS NOT TOO LATE to decertify machines that have already been purchased, even if this lengthens the process of reporting results in November. Integrity, not speed, is the most important attribute for any voting system. Several states have already decided not to allow paperless DREs to be used in the November elections.

IT IS NOT TOO LATE to educate all voters to insist on paper ballots in November. Most states still have the capability to process paper ballots for absentee voting.

And IT IS NOT TOO LATE FOR THE LWVUS TO WITHDRAW ITS ILL-ADVISED OPPOSITION TO VOTER-VERIFIED PAPER BALLOTS. Instead of supporting the rush to purchase poorly designed proprietary voting systems, LWVUS must take the lead in ensuring that all voting systems meet high standards of integrity, transparency, and accountability for all voters.