Making Every Ballot Count

Bryan Pfaffenberger
Photo by Tom Cogill

The current controversy surrounding electronic “touch-screen” voting machines boils down to this: If the machine malfunctions, or—worse—if a hacker subverts the count, can election officials restore the votes that were lost?

This dilemma has led to a national movement to reintroduce a paper trail to the voting process. The technology most recently advanced as a solution, the optical-scan system, employs hand-marked paper ballots that resemble standardized test scoring sheets. That way, if there’s a glitch, the ballots can be counted by hand.

Viewed historically, this back-to-paper movement is ironic. “The first voting machines were introduced in the 1890s specifically to remove paper from the voting process,” notes Bryan Pfaffenberger, a historian of technology in the School of Engineering and Applied Science’s Department of Science, Technology, and Society. With the aid of a $27,000 Scholar’s Award from the National Science Foundation, Pfaffenberger is studying the neglected history of the voting machine. 

In the 1880s, partisan manipulation of party-supplied paper ballots led to the rapid adoption of the Australian ballot system, which provides government-printed paper ballots and booths for marking them in private. But this system soon proved vulnerable to a range of new vote-stealing exploits. In response, Jacob H. Myers, an inventor in upstate New York, developed the first successful mechanical lever voting machine. It was first used in 1892 and by 1960, machines based on Myers’ design counted two-thirds of the votes cast in the United States. 

“Early voting machines were developed precisely because the Australian system failed to reform America’s elections—and also because the government-printed ballots could be designed to keep illiterate voters away from the polling place,” says Pfaffenberger. “The lever voting machine was an American voting system because it harnessed American inventive genius to preserve the American ideal of universal suffrage.”  

Although lever machines are vulnerable to certain kinds of fraud, exploits are easily prevented by sound election procedures.  Perhaps the most significant benefit of lever machines is that they are immune to systemic exploitation, which could affect hundreds of thousands of machines. In sharp contrast to the way Americans talk about voting machines today, users of the lever machines expressed misgivings only occasionally. The lever voting machine—though lacking an independent audit trail—had done something today’s voting technologies have been unable do: it won the confidence of American voters and election officials.

“Lever voting machines were one of the most remarkable technological achievements of their age,” Pfaffenberger notes, “but they were expensive to make and difficult to store and transport.” To cope with the complexities of American elections, lever voting machines employed dozens of intricate interlocking systems and as many as 28,000 moving parts. In the mid-1960s, the first punch-card voting systems appeared, weighing a few pounds and costing a fraction of their big, mechanical brethren. As a result, lever voting machine usage declined and by the mid-1980s the remaining manufacturers declared bankruptcy. Few voters use the machines today, except in the state of their invention, New York, where voters and election officials alike swear by their accuracy.

For Pfaffenberger, the history lesson is that Americans have forgotten the perils of paper ballots. “It’s a sad comment on today’s computer-based voting technologies that they can’t achieve what an 1890s invention was able to do—namely, convince nearly all voters that their votes had been counted,” he says.