Florida's User Un-friendly Election Ballot

Theresa LePore is certain to get at least a footnote in history.

LePore, the Palm Beach County, Florida, election official who approved that county's ballots, has made what is certain to become the most infamous user interface design error of all time.

By now, you've seen or heard about the Palm Beach ballot. With its single column of punch-holes centered between candidate lists on two facing pages, the ballot is a perfect example of bad interface design. Any voter who, scanning the left-hand page, noted that George W. Bush was the first listed candidate, and who then proceeded to punch the first hole, successfully voted for Bush. Any voter who, scanning the left-hand page, noted that Al Gore was the second listed candidate, and who then proceeded to punch the second hole, voted for Pat Buchanan, whose name appeared on the right-hand page. (Click here to view the ballot.)

There's a painful irony to the thought that the dumbest American president in history may come to power for the dumbest reason imaginable.

The ballot is an example of what user interface design guru Donald Norman calls a failure to use "natural mappings." In his classic book The Psychology of Everyday Things, Norman describes another, less significant, example of failure to use natural mappings, stupidly designed stovetops. There's probably no American alive today who hasn't accidentally turned on the wrong burner on a stovetop because of poorly arranged knobs. On many stoves, a 2x2 rectangular grid of burners is controlled by a 1x4 linear array of knobs.

Says Norman: "The problem of the stove top may seem trivial, but in fact it is a cause of great frustration for many homeowners. Why do stove designers insist on arranging the burners in a rectangular pattern and the controls in a row? We have known for forty years just how bad such an arrangement is. Sometimes the stove comes with clever little diagrams to indicate which control works which burner. ... But the proper natural mapping requires no diagrams, no labels, and no instructions."

If the Palm Beach ballot had been subjected to the usability testing that most commercial software now undergoes (or, at least, should undergo), the problem would have been detected. With only 550,000 registered voters in Palm Beach County, an astonishing 19,000 ballots were eliminated for having been double-punched. Assuming a roughly 60 percent turnout, that represents an astonishing rejection rate of six percent. And that's not even counting the statistically suspicious counted vote of 3,000-plus for Pat Buchanan. Statistical plots comparing the county-by-county percentages of the vote gained by Bush and Buchanan, the two right-most candidates, show Palm Beach County sticking out like a sore thumb in favor of Buchanan. That doesn't prove anything, but it certainly opens reasonable doubts.

A Bush campaign spokesman, Ari Fleischer, has suggested that Palm Beach County is often troubled by the double-punch phenomena, citing 14,872 ballots invalidated for double punching in the 1996 election. There is an easy way to determine if the ballot design is at fault. If the double-punching is primarily in the second and third holes (the Gore and Buchanan holes), any reasonably-minded individual would have to acknowledge most double-punched ballots represented failed attempts to vote for Gore. (It's not terribly likely that voters were undecided between Gore and Buchanan, and the vast majority of voters voted for either Bush or Gore, not Buchanan.)

Imagine: not only may the next president of the United States have lost the popular vote, not only may his electoral college victory be attributable to fewer than 1,000 ballots, but even that microscopic electoral college victory may be attributable only to voter confusion. If I were George W. Bush, I'd demand public inspection of the rejected ballots, to see if there is a pattern of Gore/Buchanan double-voting. If such a pattern was evident, I'd demand a revote in Palm Beach County, with a redesigned ballot.