In an interesting development, the City and County of San Francisco Department of Elections has launched an initiative to make voting system usability a priority for their office. They have put up a website about “Usability and Voting Equipment”, and have developed an interesting slideshow training session for Department of Elections staff (which we have converted to an easier-to-view pdf file).

In a letter (dated August 25, 2005), John Arntz (Director of the Department of Elections, City and County of San Francisco) briefly discusses the approach so far that they have taken: “the Department held a pilot program to allow the public to use two voting systems related to our Request for Proposal to purchase a new voting system. We also invited usability professional to attend the event and give their professional opinions concerning various usability aspects of the voting equipment.” The letter concluded by urging the voting system vendors to pay more attention to better study usability and accessibility.

We applaud this effort, and we are always excited when election officials include the public and researchers in their constant efforts to improve elections in the United States.

But we also need to work to do better. We need to work with vendors and election officials to develop collaborations that have the same sort of large sample and controlled experimental research designs that we would expect to see in our own research (or that in peer-reviewed journals). We also need to work with vendors and election officials to convince them that it is in everyone’s best interest when research projects on usability (and other aspects of voting system performance) that reports and data from those studies be made easily available to other researchers and the public.

Examples of such collaborative and public usability research efforts include Thad’s study for Alexandria, Virginia of the Hart Intercivic voting system in November 2002; VTP colleagues Ted Selker and Sharon Cohen’s experimental study of voter verification; Selker, Jonathan Goler and Lorin Wilde’s study of voting interfaces for disabled voters; and the research coming from Paul Herrnson, Ben Bederson, Fred Conrad, Richard Niemi, and Michael Traugott on voting technology and ballot design. At later dates we will have more commentary about this area of research, especially as new studies come out.

Two immediate things come to mind, though, as we review this recent research. First, we need to figure better ways to develop industry, government, and academic collaborations — bridging these different fields can be difficult. Second, we need some methodological development in this area, to help guide future research efforts.