Introduction

Beginning on election day, November 7, 2000, Palm Beach County, Florida, attracted national and ultimately international attention because thousands of voters in the county complained that they had a difficult time understanding the now infamous butterfly ballot and that they consequently cast invalid or erroneous presidential votes.² Eleven groups of Palm Beach County voters filed independent lawsuits seeking relief because, they claimed, they and others had made mistakes in their votes for President because of the misleading and confusing format of the ballot.³Within four days of the election, more than a dozen scholars had posted to the Internet analyses of what appeared to be an exceptional electoral outcome in Palm Beach County. ⁴ Early analytical work drew attention to an apparent excess in the number of votes received in the county by 2000 Reform Party presidential candidate Patrick Buchanan who, in the initial and uncertified count of Palm Beach County ballots, received 3,407 votes. The technical approaches used in the early analytical work varied greatly, but most analysts agreed that 3,407 was many times more than the number of votes to be expected for Buchanan, with 2,058 being a conservative minimum for the number of votes that were counted for Buchanan but that were irregular.⁵

The number of votes--at least 2,000--that appear to have been given erroneously to Buchanan was larger than the final, certified vote difference that separated Gore from his Republican challenger, Texas Governor George Bush. Final election results gave Gore 2,912,253 and Bush 2,912,790 votes in Florida, a difference of only 537 votes.⁶ Because Florida was a pivotal state in the overall presidential election, the Palm Beach County butterfly ballot may have itself been pivotal.

What characterizes a butterfly ballot is the presence of two columns of candidate names that together sandwich a single column of punch holes. Figure 1 contains a picture of the Palm Beach County presidential ballot. The alleged source of confusion which some Gore supporters claim led them to vote for Buchanan concerns the second valid punch hole (#4). This hole corresponds to Buchanan, the first candidate on the right hand side of the ballot. But, from the perspective of the left hand side of the ballot, the second candidate listed is Gore. An individual who scanned down the left hand column of his or her presidential ballot and only later considered the right hand column could mistakenly assume that the punch holes correspond, respectively, to Bush, Gore, Browne, and so forth.

The question of whether the Palm Beach County butterfly ballot caused systematic voting errors is ultimately a question about ballot design.⁷ As of this paper's writing, published research on ballot design is not extensive. Sinclair et al. (2000) report experimental evidence showing that a double-column ballot format like the one used in Palm Beach County can be more confusing and cause more voter errors than a single-column ballot. Hamilton and Ladd (1996) use a collection of North Carolina counties to show that the extent to which ballot format made straight-party voting easy had measurable though small effects on voting outcomes. Finally, Darcy (1986) argues that position effects on ballots--namely, which candidate is listed first--can have small consequences for candidate vote shares but that these effects do not matter much for races in which party labels are very important. In general, very little is known about optimal ballot design, the relationship between ballot format and error rates, and so forth.

Our paper's analysis of the Buchanan vote share in Palm Beach County combines a strategy of triangulation with the elimination of plausible alternative explanations. We show that different models using different types of data all point to the same conclusion regarding the Buchanan vote in Palm Beach County. We begin, in Section 1, by examining a collection of 3,015 counties from across the United States to show that the Buchanan vote share in the Palm Beach County was highly unusual and excessively large. We also show that Palm Beach County, which did not use the butterfly ballot for the previous presidential election, did not have an unusual level of Reform support in 1996.

Section 2 describes the results of a natural experiment: Buchanan's vote share among Palm Beach County absentee voters, who did not use a butterfly ballot, was much lower than his vote share among election day voters, who did use the butterfly ballot. In Section 3 we examine precinct-level election day returns within Palm Beach County and find that Buchanan votes tended to come from the most liberal parts of the county. This pattern, which is absent in voting returns for other Reform candidates, is consistent with the claim that Buchanan's votes came from mistaken Gore supporters. Finally, Section 4 concludes.