On July 12, 2001, the American Association for the Advancement of Science sponsored a briefing on the Caltech-MIT voting technology project, which was released to the general public on July 16, 2001.  Members of the research team included: Stephen Ansolabehere (MIT) Professor of Political Science; Thomas Palfrey, Ph.D. (Caltech) Professor of Economics and Political Science; Jehoshhua Bruck, Ph.D. (Caltech) Professor of Computational and Neural Systems; Ted Selker, Ph.D. (MIT) Associate Professor of the MIT Media Lab.  Rep. Vernon Ehlers (R-MI) introduced the project team.

Rep. Ehlers – spoke about the Science Committee s bill (HR 2275) that he co-sponsors with Rep. Jim Barcia (D-MI).  Rep. Ehlers indicated the full Science Committee would hold a markup on this bill during the week of July 18 and that he anticipates the bill will eventually pass the House.  He stated that he also expects the House Administration Committee, of which he is a member, to move legislation but he doesn t think dollars will be forthcoming.

Stephen Ansolabehere - opened his talk with a slide showing that 4 to 6 million votes were lost during the last election.  That total was made up of 1.5 million votes lost due to faulty machines, 3 million votes due to registration problems (this number reported to the Census Bureau), and up to 1 million lost votes as the result of long lines and crowded polling stations.

In addressing the issue of voting equipment, Prof. Ansolabehere stated that counties surprisingly are doing their  own thing  when it comes to purchasing equipment.  There is not a lot of interaction among counties about voting equipment and there are not many opportunities for them to share information concerning performance of equipment.  He indicated that the Project felt that this was an opportunity for the Federal Gov t to get involved and help counties with information sharing.

Thomas Palfrey - focused on the economics of the voting industry.  The industry can be divided into three parts: equipment manufacturers, election administrators, and voters.  There are relatively few equipment manufactures and the total market size is $150M - $200M, smaller than the domestic lawn mower market.  The largest company has 60% market share and the top three companies have a combined market share of 80%.  It s a relatively stable and old style industry where manufacturers market their equipment directly to election administrators.  He said that due to some of the problems brought to light by the last election several new entrants have entered the market with innovative IT voting machines.

 In looking at election administrators, Palfrey said  they found a highly desegregated and fragmented structure.  There are 3,100 counties in the United States and very little information gets shared.  During the Project this was the most difficult area to get information on because there is no one that represents the interests of the administrators.  There are several organizations but none are able to provide an administrator with information on performance of machines.

Prof. Palfrey offered several solutions.  In the short term, he suggested phasing out dominated technologies (punch cards, lever machines, underperforming DRE s, and centrally counted optical scanners), providing provisional ballots, having better poll site registration data, training for poll workers, and more poll workers.  Long term solutions suggested were providing federal funds for upgrades, uniformity of recounting procedures, field testing of new equipment, and integration of registration databases at pollsites.

They recommended that a federal agency be charged with providing an information clearing house and data bank for states and localities to access, as well as, carry out research on voting technologies.  Also needed is a system for certification and testing of machines that allows for flexibility in design.  He noted that companies are reluctant to provide upgrades to their software because it takes over a year for the companies to have that software certified.  So the companies have a disincentive for software upgrades.

Jehoshua Bruck - discussed a new paradigm in the way the voting process is viewed.  He talked about dividing the system into two parts: (1) vote generation and (2) vote management.  Vote generation was described as the presentation of the candidates and issues to the voters and the actual casting of the vote.  Vote management takes the casted vote and uses it as an input where the eligibility of the voter is then verified, recorded and added.  In viewing the voting process in this manner then vote generation can take place anywhere, whether it is at kiosk, home computer, etc.

Ted Selker - talked about several problems (other than hanging chads in Florida) that were encountered in the last election.  Most of these problems involved difficulties voters had in interfacing with the equipment.  These included confusion with how the machine operated, alignment problems with ballots, audio difficulty (pause in the wrong place), and mechanical difficulty (buttons not working or hard to operate).

During the Q&A session questions were asked about the recently released FEC Standards.  The Project team responded that they had not had a chance to thoroughly look at them but they felt that these new standards were updates of the previous standards.  The standards were described as durability standards dealing with drop and freeze tests of the voting machines.  They still did not adequately deal with issues such as software security, useability, and open source codes for software.