Workshop Program

STAR-Vote is a collaboration between a number of academics and the Travis County (Austin), Texas elections office, which currently uses a DRE voting system and previously used an optical scan voting system. STAR-Vote represents a rare opportunity for a variety of sophisticated technologies, such as end-to-end cryptography and risk limiting audits, to be designed into a new voting system, from scratch, with a variety of real world constraints, such as election-day vote centers that must support thousands of ballot styles and run all day in the event of a power failure. This paper describes the current design of STAR-Vote which is now largely settled and whose development will soon begin.

Proposals for a secure voting technology can involve new mechanisms or procedures designed to provide greater ballot secrecy or verifiability. These mechanisms may be justified on the technical level, but researchers and voting officials must also consider how voters will understand these technical details, and how understanding may affect interaction with the voting systems. In the context of verifiable voting, there is an additional impetus for this consideration as voters are provided with an additional choice; whether or not to verify their ballot. It is possible that differences in voter understanding of the voting technology or verification mechanism may drive differences in voter behaviour; particularly at the point of verification. In the event that voter understanding partially explains voter decisions to verify their ballot, then variance in voter understanding will lead to predictable differences in the way voters interact with the voting technology.

This paper describes an experiment designed to test voters’ understanding of the ‘split ballot’, a particular mechanism at the heart of the secure voting system Prˆet `a Voter, used to provide both vote secrecy and voter verifiability. We used a controlled laboratory experiment in which voter behaviour in the experiment is dependent on their understanding of the secrecy mechanism for ballots. We found that a two-thirds majority of the participants expressed a confident comprehension of the secrecy of their ballot; indicating an appropriate level of understanding. Among the remaining third of participants, most exhibited a behaviour indicating a comprehension of the security mechanism, but were less confident in their understanding. A small number did not comprehend the system. We discuss the implications of this finding for the deployment of such voting systems.

When the Australian secret ballot was introduced in the 1850s, it not only provided privacy for those voters who wanted it, but it also effectively eliminated coercion by allowing no viable means for voters to prove their votes to third parties. In an environment where the privacy of voters is enforced by independent observers, coerced voters could freely express their true preferences while making their selections.

In contrast, modern technologies render the traditional poll-site protections largely ineffective, and the limited remaining options for preserving these protections will almost certainly disappear in the not-too-distant future. Today, in-person voters routinely carry video recording equipment and other technologies that facilitate coercion into polls, and although not yet ubiquitous, inexpensive and unobtrusive wearable video recording devices are readily available. In view of these realities, it is appropriate to re-examine the efforts and countermeasures currently employed and explore what defenses are possible and reasonable against various forms of voter coercion.

In his State of the Union address, President Obama referred to long lines faced at the polls on November 6, 2012, and said, “we have to fix that.” Although it seems to have received relatively little national attention, Richland County, South Carolina, with more than 12% of its votes cast after polls were officially closed, was probably among the very worst counties in the nation for lines and wait times.

In this paper, we analyze the data from the DREs used for voting in South Carolina, and we compare the voting process in Richland County with that in Greenville County, where there were more total votes and more votes per DRE voting terminal, but where there were fewer than one-half of one percent of the votes cast after closing time.