Thank you for following the work of the AEI-Brookings Election Reform Project. We’ll continue looking at the issues of election reform at AEI and Brookings. For new work on congressional redistricting, please visit

Viewpoint: A Dose of Honesty for Florida 13
Matt Weil, Research Assistant, AEI-Brookings Election Reform Project

January 10, 2007

Eighteen thousand votes were lost. Voting machines refused to record votes. Those are two of the claims in a contested seat for the U.S. House of Representatives. The November 2006 election did not result in a total system meltdown but it did underscore that there are numerous weak points in our voting systems nationwide.

The recent major overhaul to our more than two hundred year-old system came in the form of the Help America Vote Act of 2002. It was not a perfect piece of legislation, but it certainly helped many jurisdictions modernize their voting processes. Still, the most important outcome of the passage of HAVA was not the improvement of specific technology or other statewide mandates. Rather, HAVA was the first real public discussion about voting administration in United States history.

American democracy works best when the parties debate with the greater good in mind instead of short-term self-interests. Compromise must be the goal on an issue as basic as the framework for picking the representatives of the people if our system is to have any legitimacy. Unfortunately, forget about compromise because the parties are not even willing to confront the real issues.

The best example of this stalemate is the race in the Florida 13th Congressional District. The certified results show that Vern Buchanan (R) won the election with a 369-vote margin over Christine Jennings (D). However, an election is only legitimate when the loser accepts a set of results that are widely viewed as fair and complete, and that has not happened in this race. The House of Representatives seated now-Rep. Buchanan last Thursday, but Jennings has not dropped her challenge.

Jennings’ problem with the race lies with Sarasota County’s unusually high “undervote rate.” An undervote happens when no vote is recorded for a specific race. Of the 238,249 votes cast in the race district-wide, 124,119 votes, slightly more than half, came from Sarasota; of these, 18,412 were undervotes. That equates to an unbelievable 12.92% undervote rate. The undervote rate for the congressional seat in the rest of the district was 2.39%. To put those rates in perspective, in Sarasota County, the undervote rate for U.S. Senate was 1.14%, for Governor was 1.28%, and for Commissioner of Agriculture was 5.19%. These results contradict the reasonable expectation that undervote rates rise as voters move down the ballot and know less about lower profile contests, thus choosing not to vote.

Adding an additional layer of analysis into the election results yields another surprising statistic. Electronic voting machines, including the ES&S iVotronic used in Sarasota, alert voters to potential undervotes and offer them the opportunity to fix their votes before casting them. Paper absentee ballots do not have such a built-in failsafe. Thus, one would expect the undervote rate to be higher in absentee voting than in electronic voting. Yet, in this race, the opposite is true. The absentee undervote rate was just 2.50% and the Election Day undervote rate was 13.92%! There is something fishy going on in Sarasota.

Despite the startlingly high undervote rate, neither party has addressed the most likely explanation for the problem. Republicans have argued that the large amount of undervotes is due to discontent over a race that turned ugly. The race for the Florida 13th was the most expensive House race in the country with a combined $8 million spent. The Republican extension of this claim is absurd. Apparently, the discontent over the negative campaign was five times as great in Sarasota County as it was in the rest of the district—which got the same TV ads, the same mailings, the same news reports. Certainly some of the undervote could be because of a negative race, but I would imagine that the Republican spokespeople have to hold their noses while making such an obviously ridiculous argument.

The Democrats are no better. For weeks after the election, they spent time and money gathering information about voting machine malfunctions in which votes for Jennings were cast but not recorded. There is little hard evidence to support the Democrats’ claim. A statistical analysis done on the 18,000 undervotes by MIT professor Charles Stewart does show that Jennings would have overcome the 369-vote deficit if people who intended to vote for her had done so because she won almost 53% of the vote in Sarasota. In his analysis of the undervote, though, he concludes that there is “a substantial possibility that the exaggerated undervote rates in Sarasota County were not solely due to voter confusion, but also caused by factors related to machine malfunction.” He is a consultant to the Jennings campaign as it fights in court for access to the computer code, which runs the voting machine software. In late December, a judge refused to order ES&S to give the Jennings campaign access to the computer code ruling that the “[p]laintiffs have presented no evidence to demonstrate that the parallel testing was flawed and/or the results not valid. ...The testimony of Plaintiffs’ experts was nothing more than conjecture and not supported by credible evidence.”

The most likely reason for such a high undervote rate on Election Day is poor ballot design. In the Florida 13th, the U.S. Senate race was listed first and had six declared candidates plus a spot for a write-in. The U.S. House race would not completely fit on the first screen, so it was moved to the top of second screen above the gubernatorial race. Those voters intending to vote straight ticket, or who just moved quickly through the ballot, were more likely to miss the race at the top of the second screen and move to the more prominently displayed gubernatorial race just below it.

Voter confusion is nothing new to Florida voters. No one will ever forget the butterfly ballot used in Palm Beach County in 2000, which many people believe cost Al Gore thousands of votes in Florida and thus the presidency. The reasons, however, that neither party wants to accept that ballot design is the most likely problem are purely partisan. For the Democrats, it means that there is no real recourse. A machine malfunction might void an election, but carelessness on the part of the voters hardly seems like a reason to overturn the results. The Democrats know that voter confusion alone yields no remedy. Republicans are unwilling to concede that the undervote was anything but a protest vote because they would be tacitly admitting that voters intended to install Jennings in the seat, not Buchanan.

Christine Jennings’ challenge to the election now sits in the House Administration Committee, though the committee is unlikely to address the contest until the matter is completely adjudicated in the Florida courts. The House should not overturn the election. It would set a bad precedent to overturn elections for poor ballot design, which in many states – including Florida – is open to party review before the election. Is it unfair to Ms. Jennings? Yes, it probably is. On the bright side, her loss could result in another giant step forward for election reform. Here’s hoping that 2007 brings us one step closer to bringing voting into the twenty-first century!

Matt Weil can be reached at

Viewpoint is an occasional feature in which members of the AEI-Brookings Election Reform Project staff analyze various election reform issues.
Featured Resources
We reflect on the state of election administration in the United States almost a decade after the 2000 presidential election and suggest how additional changes in technology, election law and administrative practices might further strengthen American elections in the years ahead.
The security risks inherent in internet-based voting today pose significant barriers to the adoption of this method in nationwide elections. The author suggests changes to both internet security and the allocation of election tasks to improve the applicability of the internet in elections.
This report takes a comprehensive look at Ohio's election system, report highlighting both successes and failures and making a range of recommendations.
This report finds that the increase in the no-valid-vote rate in the 2008 presidential race in Florida was due to excessive overvoting statewide, not to the change in voting technologies.
This report explores the concept of open source software as it applies to voting systems, suggesting areas in which it may benefit the market but cautioning against a too-enthusiastic embrace of the technology.
Research Projects
FairVote develops and promotes practical strategies to improve elections at the local, state and national levels.
As part of its broader research focus on elections, campaign ethics, campaign finance, and the legislative process, the Center for American Politics and Citizenship at the University of Maryland is engaged in research projects on voting technology and ballot design specifically.
The Brennan Center for Justice at New York University School of Law is a non-partisan public policy and law institute that focuses on fundamental issues of democracy and justice.
This project aims to evaluate the current state of reliability and uniformity of U.S. voting systems; to establish uniform attributes and quantitative guidelines for performance and reliability of voting systems; and to propose specific uniform guidelines and requirements for reliable voting systems provides daily news updates on election reform issues, as well as deeper analysis of selected topics, including recent reports on voter registration, recount procedures, and the progress in implementing the Help America Vote Act since 2002.
The American Enterprise Institute for Public Policy Research
The Brookings Institution
© Copyright 2016, AEI
and The Brookings Institution