Voting Matters--Securing our elections Mar 07

The National Institute of Standards and Technology (NIST) produced a white paper on electronic voting machine vulnerability and made recommendations about accrediting voting machines. Titled Requiring Software Independence in VVSG 2007: STS Recommendations for the TGDC, November 2006 it can be found at http://vote.nist.gov/
The Voluntary Voting Systems Guidelines (VVSG) Version I is available on this site: http://www.eac.gov/vvsg_intro.htm

Why do we care about the VVSG? Well, under the Help America Vote Act (HAVA) voting machines have to be accredited and tested. NIST has studied existing laboratories that the manufacturers have been using. NIST determined that the current proprietary testing is inadequate and needs independent oversight and national test standards.

The concept of software independence in voting systems is a great idea from NIST. The idea is that a detection method (such as audit) independent of the software used to tally the vote will detect errors. NIST was thinking of electronic tools to produce an independent voter verified audit trail. Interesting, but can we trust the manufacturers?

Representative Holt has a bill to reform HAVA that is under consideration in the House. The bill, HR-811, has a several strong points. It requires voter verifiable paper trails and a mandatory audit program. While a lot of observers disagree on the audit techniques, it does require the single most important technique to detect fraud: audits. Some feel that the audit requirements are too detailed and should be described more broadly, but they are better than what we have now.

Critics have identified weaknesses in Holt’s bill as well. One is that it doesn’t require that citizens have full access to voting records or to observe the count. Some states don’t allow voter or candidate access, hence the count is done in secret and ballots cannot be independently examined. Another criticism is the composition of state audit boards does not require mathematicians or statisticians as members. Several loopholes and exceptions to audit requirements are also in the bill that may be abused. Many of the other criticisms of the Holt bill really are arguments about whether it goes far enough. http://thomas.loc.gov/cgi-bin/bdquery/z?d110:h.r.00811:

While the League has not set a position on the Holt bill, it adopted a resolution at the last national convention endorsing voter verified paper ballots or an auditable record. So when discussing the Holt bill, we can confidently say that the League endorses both a paper record of the vote and an audit protocol to ensure fair elections. Many of the other provisions of the Holt bill are subject to disagreement, but if we stick to the two items above, we will have strong influence on our representatives.

As an interesting test case, New Mexico switched from direct recording electronic voting machines in 2004 to paper based optical scan systems in 2006. In Native American and Hispanic areas the undervote rate dropped by 69-85%. As I mentioned in previous columns one of the major problems with DREs is that they have little or no undervote protection, whereas optical scan readers can detect and reject a ballot immediately if an under or over vote is detected. It should make us all wonder what the 2004 elections would have been like if they were all conducted on paper ballots. For a little light reading, here is the link: http://www.votersunite.org/info/NM_UVbyBallotTypeandEthnicity.pdf