[an error occurred while processing this directive]

REVISITING THE U.S. VOTING SYSTEM: A RESEARCH INVENTORY

November 27-28, 2006

Convened by the American Association for the Advancement of Science

Merle S. King

SYNOPSIS OF PROPOSED RESEARCH ISSUES IN ELECTRONIC VOTING

1. DRE systems are perceived as vulnerable to security threats but the analysis is rarely done in the context of competing technologies. Like Churchill’s famous quote on democracy, electronic voting is the worst kind of voting system except for all the others that have been tried. A head-to-head analysis of security and performance anomalies of all contemporary voting technologies would provide election officials and the public with a contextual understanding of the inherent risks of the various voting technologies. This analysis should include all variations of optical scan technologies and tabulation methods, lever machines, punch card, telephone voting, etc.

2. Parallel monitoring of elections is the process of simulating an election as close to its real-world counterpart as possible. What is needed are methodologies for the non-intrusive, forensic analysis of election software and hardware at the precinct level on Election Day, in election mode. Current approaches rely heavily on simulation or disabling certain security features to accomplish the monitoring. Methodologies would be both jurisdiction and vendor specific.

3. Organizations such as the Carter Center are working on the development of a methodology for observing DRE-based elections. Traditional approaches of election observation presume artifacts and timetables that do not apply to DRE-based systems. Methodologies that can be adapted to diverse jurisdictions and the skill levels of observers and produce reliable work products are needed to improve the transparency of election observing.

4. Contemporary voting systems are unified collections of hundreds of programs and subsystems. Current certification procedures can address only a portion of these subsystems and frequently must do so through indirect methods. Election administrators need certification procedures that enable the jurisdiction to verify all software used in election technologies in their systems. Efforts at the NIST National Software Reference Library project need to be supported by methods to implement verification procedures at the individual device level.

5. Has the demand for professional competency in election officials and administrators reached a level where academic preparation for the career is warranted and justified? A model curriculum in election science could guide universities in program development that will eventually produce a generation of election administrators who are adequately prepared to deal with the challenges of election law, election technologies, project management, etc. Has election administration matured to a point where it merits a distinct academic preparation?

6. Most technology communities create user groups – an informal or semi-formal collection of users who share best practices, voice common concerns to the vendor, and can act collectively to influence product refinements and enhancements. The formation of user groups would improve the implementation and maintenance of systems and encourage vendors to improve customer support. User groups could be instrumental as focus groups and data sources for researchers.

7. Case studies are effective teaching tools that enable students to explore the real world issues associated with the classroom theory. Case studies of election equipment vendors, their business models and customer support practices, and election anomalies in general, could provide valuable teaching materials to university faculty as well as continuing education instructors.