Broken Ballots: Will Your Votes Count? by Douglas W. Jones and Barbara Simons (review)

In Broken Ballots, two respected computer scientists and election-reform activists—Doug Jones of the University of Iowa and Barbara Simons, formerly of IBM Research and a past president of the Association for Computing Machinery—join forces to recount the history of American voting technologies, from Samuel Jollie’s glass ballot box of 1856 to today’s computerized systems. They take their inspiration from long-ignored election administration expert Joseph Harris, who asked, in 1934, why the United States had thus far failed to develop an election system capable of producing clear, unambiguous results. In spite of more than seventy-five years of technological progress since Harris’s day, much the same can be said right now. Why?

According to the authors, the answer is best appreciated by telling the voting machine’s story, from its first use in the early 1890s through the electronic voting machine battles of the early twenty-first century. Throughout, they find a persistent failure to evaluate election technologies according to a set of “bottom-line attributes,” which include accessibility, accuracy, auditability, reliability, security, transparency, uniformity, and usability (pp. 348–53). More recently, they find an unwillingness to profit from the expertise of computer-science professionals, leaving voters to suffer at the hands of officials who uncritically accept vendors’ exaggerated claims, regulators who operate with flawed assumptions, and judges who cannot grasp elementary statistics. In their concluding chapter, they recommend specific reforms that follow, they claim, from the stories they tell. Among these reforms are many that are long overdue, including developing mandatory standards for voting machine and voter-registration database accuracy, security, and reliability; restoring voting rights to felons who have served their sentences; prosecuting efforts to deprive citizens of their voting rights; improving the design of ballots; and avoiding insecure ballot submission methods, including internet, e-mail, and telephone voting.

This book’s chief strength is its willingness to examine the voting machine’s history in light of the basic, “bottom-line” concepts taught in the upper-division curriculum of any good computer-science department—but that is its chief weakness as well. For one thing, the authors’ bottom-line attributes can and do work at cross-purposes. To ensure auditability, for example, the authors call for laws mandating the retention of ballots for periodic, postelection accuracy checks. But many local election jurisdictions cannot establish chain-of-custody and other security measures capable of securing these ballots from fraudulent postelection administration. If transparency is taken to the extreme advocated by some computer scientists, voting machine vendors would be forced to divulge their code to the public—but, critics say, doing so would raise the risk of undetectable hacks. Additionally, the authors’ list of bottom-line attributes does not include some that clearly deserve mention, including efficiency—specifically, how long it takes to cast a ballot. In recent elections, the greatest risk to citizens’ voting rights has been posed by lengthy lines rather than malfunctioning machines. Yet, efficiency comes at the cost of accessibility and usability, as undergraduate programming students quickly learn.

But the book’s gravest failure is its unwillingness to consider a topic that is not often discussed in computer-science curricula: politics. As is commonly the outcome when scientists write histories of technologies relevant to their fields, Broken Ballots offers a classically modernist narrative: voting machines reflect inventors’ efforts to employ rational, logico-deductive reasoning in order to protect the public from election administration failures and frauds, but their efforts have fallen short of the mark. Voting machine inventors and vendors can and must do better, the authors say, by following their suggested reforms. Still, had the authors gotten the history right, they might have realized that America’s election problems will not be solved by developing better voting machines alone. The voting machine’s story has yet to be told, but it amounted, initially, to an effort to build a technological infrastructure capable of keeping partisanship...