ABSTRACT
Here is a startling fact. In the 2000 presidential election, roughly 2 million people showed up on election day, but did not have a vote counted in the election.1 Some of these people purposely did not vote in the presidential election, but a substantial number-an estimated 1.5 million people-believed that they had voted when in fact they had not.2 Why were their votes not counted? Some may not have followed directions, but many people’s votes were not counted because of problems with the voting machines they used to cast their ballots. Before 2000, only a handful of political scientists and practitioners paid attention to the pros and cons of various voting machine systems. As Stephen Ansolabehere and Charles Stewart III, two political scientists who had studied this issue, write, “The methods used to cast and count ballots are surely one of the most mundane aspects of elections.”3 Few people would have considered the types of voting machines jurisdictions used to be an integral aspect of debates over fairness of elections and the quality of electoral democracy. The 2000 presidential election, particularly in Florida, proved otherwise. The country was captivated as election offi cials in several Florida counties counted and recounted people’s ballots. Yet because of problems with the punch card ballots used in the counties, it was unclear on many ballots what the intentions of the voters actually were. In an election that was ultimately decided by 537 votes, it quickly became apparent that voting machines mattered. In the aftermath of that election emerged a new wave of studies examining the reliability of different systems, most notably the Caltech/MIT Voting Technology Project’s Voting: What Is, What Could Be and the General Accounting Offi ce’s Statistical Analysis of Factors That Affected Uncounted Votes in the 2000 Presidential Election.4
