TCS Daily

People Vote, Machines Don't

By Geralyn M. Miller - September 29, 2004 12:00 AM

The controversy generated by the close outcome of the 2000 presidential election prompted Congress into passage of the Help America Vote Act (HAVA). This historic measure authorizes $3.8 billion and change to "fix" an ailing voting system. That money is being spent on a variety of changes including the creation of statewide voter registration databases, paid television advertising promoting voting in America, and updating voting equipment. Yet, exactly what it is about the system that needs fixing remains elusive to many. The conventional wisdom that emerged among scholars, media commentators, the working press and politicians centered on the punch card voting technology as the culprit. A careful review of the evidence to that effect, though, would have left that wisdom dangling every bit as much as the chads left by inadequate punches.

There could be serious problems with some of the assumptions Congress operated on when it crafted the Help America Vote Act. Scholarly researchers across the country have derived a figure for "voting error", referred to as "residual vote", by simply subtracting the number of votes cast for a particular office from the actual number of ballots cast, or the number of people voting in the election. This assumes that everyone who steps into the voting booth actually tries to cast a vote for each office on the ballot, but that the vote, for some reason, wasn't counted. Commentary by various pundits and interest group spokespersons seems to reject the idea that voters may simply not wish to cast a vote for either of the two particular individuals running for election in some offices. There has been a failure by scholars and election administrators to adequately distinguish between those voters who decline to mark the ballot for a given office and the far fewer number who may have tried to but failed for some reason to have that vote recorded.

It's conceivable that technology can, under certain circumstances, contribute to a larger than expected residual vote. An example would be when equipment malfunctions occur. But ultimately machines are inanimate objects that are highly dependent on humans for their performance as tools. For example, in the 2000 election, the high residual vote in Illinois' Cook County, estimated at approximately six percent, was apparently attributable to a misalignment of punch card voting trays that could have been prevented with a modest effort at quality control by manufacturers and local election officials. Elsewhere in Illinois such problems were not to be found. That is an illustration not of a systematic problem of a particular technology but of the way in which humans interact, variably, with technology.

The closeness of the 2000 election in Florida may have taught us precisely the wrong lesson. The machine tabulations varied little from one county to another. It was only when humans began interpreting ballot markings -- or even the lack of markings -- that it seemed as though the punch card systems were the cause of the problem. It was the humans at one Florida county local election board who decided to utilize the infamous "butterfly ballot", intending to better allow for voter choices among the many parties eligible to offer Presidential candidates. While closely split on whether to stop the Florida re-count and effectively make George W. Bush the President, U.S. Supreme Court members fell just short of unanimity in finding that the re-count was proceeding in the absence of any standards for humans to determine a valid vote if the election technology did not tabulate one.

One need only think back to the days in which machine politics flourished to understand the power of people versus voting technology. Old-time political bosses sent their armies of precinct workers door-to-door to deliver legendary majorities for offices the party organization needed, caring little about the other offices on the ballot. Come to think of it, reformers have inveighed against straight-ticket voting mechanisms favored by machine partisans that may have reduced "residual vote." As experience in East St. Louis, Illinois suggests, it's possible that high residual vote levels for specific offices are at least partly the result of failure by local political organizations to include those offices on sample ballots in order to focus voter attention on offices that matter to them. This gives modern meaning to the old cry of local political machines, "to hell with the White House, give us the courthouse."

Attributing voting errors to the equipment fails to take into account that voting is not a frequent and routine part of our lives. When we make our infrequent visits to the polls, we have to remember the steps in a process that we have not gone through in, perhaps a year, two, or maybe, even longer. The mere fact that we have not learned the process through rote exercise breeds error. Imagine if people only used an ATM machine once or twice every two years. Have you watched the confusion levels lately at those self-checkout lanes at the supermarket? When voters in the Florida 2002 primary walked into the voting booth and encountered a new type of voting equipment, they had to learn how to operate that equipment. It didn't matter that the new machinery had been placed there to "cure" perceived problems of the past. Varying literacy levels among the populace, unique local governing systems within voting jurisdictions, and the dynamics of political organizations across those jurisdictions all come into play in the outcome of our elections. This hasn't been taken into consideration in previous voting studies but it should certainly be taken into consideration with policy decisions.

Current polling data suggest that the utility of reforms mandated by the HAVA including replacement of punch card equipment may be sorely tested as we face another election that could turn on statistically insignificant margins. Did the $3.8 appropriation by Congress cure America's electoral problems? My bet is that we are about to discover that it's the people, not the machines that drive the engines of our elections.

The author is a professor in the School of Public and Environmental Affairs at Indiana University-Purdue University Fort Wayne.


TCS Daily Archives