THE FRAME GAME
In the aftermath of the 2000 election, there were many calls for electronic voting. After all, many of the problems producing the 2000 election debacle arose due to the use of paper-based voting systems. There were punch card ballots that may or may not have contained a vote, depending on your view of how pregnant a “chad” could be before it gave birth to an actual vote. There were optical scan ballots where a stray mark next to a second candidate's name created an overvote, or a voter putting an “X” in the oval next to a candidate's name instead of filling the oval completely created an undervote. Watching the news in the days following the 2000 election, viewers could be forgiven for thinking that Florida's county canvassing boards were being asked to be mind readers, divining voter intent from the well-handled ballots. With the spectacle of election workers' holding the ballots up to the light to see through the chad holes, or staring and trying to determine if the line on the ballot was a vote or a stray mark, it often seemed like an Ouija board would be an equally appropriate tactic for accomplishing this goal.
These problems with paper-based voting systems in the 2000 presidential election led to many studies and reports, predominantly focused on “residuals,” over- and undervotes, and spoiled ballots. There has developed a healthy academic cottage industry of studies on these questions, especially in the political science field (for a summary of the studies to date, see Alvarez, Ansolabehere, and Stewart 2005). This literature has established, especially with the additional data presented by the 2004 presidential election (Stewart 2005; 2006), a number of conclusions, some commonly accepted, others more controversial.
Among the more commonly accepted results is that punch card voting systems, especially the prescored punch card varieties (going by various trade names, including Votomatic or Pollstar), typically have some of the highest residual vote rates of any voting system used in recent elections in the United States. Although some studies found that electronic voting machines had high residual vote rates before 2000, studies using more recent data have tended to show dramatically lower residual vote rates. Scholars argue about the reasons why the more recent data indicate lower residual vote rates for electronic voting systems, debating whether it is due to new technology, heightened awareness of potential errors by voters and election officials, or better training of poll workers (Stewart 2005).