Go to UPA home page The U P A voice

August 2008 Contents

UPA Job Bank

UPA 2008


Ballot Design and Usability

Whitney Quesenbery

Whitney is a past-president of UPA, and leads the Usability in Civic Life project. In civilian life, she works on user research, usability and IA projects for organizations like the UK's Open University and the US National Cancer Institute.

On July 21 the Brennan Center for Justice released a report about the impact of poor ballot designs and unclear instructions on voters. I am listed as one of the authors. Other members from the UPA Usability in Civic Life Project also worked on this project.

In addition to the design problems that are its focus, the report also discusses the importance of usability testing as a final check on ballot layout and instructions text. Many of the problems in the report would likely have been caught with even an informal test. The report highlights a usability testing kit for local election officials, the LEO Usability Testing Kit. As part of briefings for election officials on the findings of the report, we are providing an introduction to usability testing and practice in conducting simple usability tests.

For a background on the Better Ballots project, please see the press release immediately below.

Dana Chisnell, Caroline Jarrett, Ginny Redish, Josephine Scott and Sarah Swierenga particpate in this project as members of the Ballot Design Task Force. You can keep up with our activities on the Ballot Usability and Accessiblity blog.


Bad Ballot Design Results in Staggering Numbers of Lost Votes

– 07/21/08

Thousands of Voters Disenfranchised in Recent Elections, Particularly Elderly, Low-Income & New Voters

Brennan Center Study Shows 13 Ballot Problems that Bedevil Elections Across the Country – Voting Experts Offer Guidelines For Improved Ballot Design in Time for ‘08 Election

New York – Today voting technology experts at the Brennan Center for Justice at NYU School of Law released an extensive analysis of election ballots from across the country that shows that hundreds of thousands of voters were disenfranchised in recent elections as a result of badly designed ballots and confusing voting instructions.

Better Ballots, a product of the Center's national Task Force of design and usability experts and election officials, cites 13 frequent problems that continue to plague elections even after the infamous "butterfly ballot" in Florida's Palm Beach which resulted in 30,000 lost votes in the 2000 election.

The report recommends specific guidelines for county, state and federal election officials that will help avoid further voting blunders in 2008—and beyond.

Significantly, the report shows that ballot design problems disenfranchise disproportionate numbers of elderly, low-income and new voters.

"Design flaws resulted in hundreds of thousands of lost votes in recent elections," said Lawrence Norden, Counsel at the Brennan Center for Justice, "and the same error-causing designs still plague ballots across the country."

"The good news is that states and counties can take steps to improve ballots now, well ahead of this year's general election. We implore them to do so," he continued.

"Elderly citizens, especially, too often have their goal of casting a vote thwarted by jumbled design, and sheer incomprehensibility. This is one election problem with no villains. We can fix ballot design issues over the summer, so that in November every vote cast is a vote that will be accurately counted," said Michael Waldman, Brennan Center Executive Director. "We had three million newly registered voters just in the first three months of 2008."

The 13 design flaws identified in Better Ballots fall into three general categories- ballot layout, text formatting and voting instructions- all illustrated in an interactive demonstration available on the Brennan Center website www.brennancenter.org/ballot. Typical flaws include:

  • Listing candidates for the same office on two or more pages or columns;
  • Listing multiple contests on the same screen, in electronic touch-screen systems;
  • Placing response options on both sides of a candidate's name;
  • Providing instructions that are full of legal or election jargon, and difficult for many voters to understand.

The report examines ballots from states including Florida, Ohio, Wisconsin, Illinois, California, Tennessee and Kansas, and draws on results from the 2000, 2002 and 2006 elections in both state and federal contests.

In each case, the Brennan Center study provides quantitative data that illustrates the cause of uncounted votes. Better Ballots makes clear that all voters are at risk for lost or misrecorded votes, all voting technologies are vulnerable to the effects of a poorly designed ballot, and, that whether at the top or the bottom of the ballot, all political contests can be subject to unrepresentative vote tallies due to ballot flaws.

In the most egregious and well-known case, the "butterfly ballot" used in Palm Beach County, Florida during the 2000 presidential election, the presidential race was split into two columns, which, as the Brennan Center report explains, likely caused more than 2,000 Democratic voters to mistakenly vote for Pat Buchanan and threw out an additional 20,000 votes due to double-voting—in a race that was decided by fewer than 600 votes.

Better Ballots documents several similar instances from around the country:

  • 2002 in Kewaunee County, Wisconsin: ballot listed candidates for Governor in two separate columns resulting in one in ten votes thrown out;
  • 2006 race in Sarasota County, Florida: separate contests for Congress and Governor were presented on the same electronic touch screen and led to more than 14,000 invalid votes in a race for the U.S. House of Representatives that had a victory margin of 369 votes;
  • 2008 Democratic presidential primary in Los Angeles County, California: over 12,000 non-partisan votes weren't counted due to confusing instructions over how to indicate party affiliation.

"The ballot problems documented in this report have simple, concrete solutions that will make the voting process more reliable, trustworthy and secure," Norden said.

Better Ballots offers several recommendations to county, state and federal government officials for how to improve ballot design, with emphasis on local measures that can be taken in time to affect the November 2008 election. They include:

  1. Develop a checklist of design best practices for ballot designers and voting systems.
  2. Conduct usability testing of ballots before finalizing the design and instructions.
  3. Publicize sample ballots that look like actual ballots voters will use on election day.
  4. Make necessary changes based on usability testing and public sample feedback.
  5. Require counties to report the number of overvotes, undervotes, and spoiled ballots.
  6. Review and amend election laws that prevent counties from employing the most usable ballot designs.

"Ironically, eight years after the 2000 election, and billions spent on new voting machines and voter registration databases, the basic problem of poor ballot design has not been effectively addressed," stated Norden. "Year in and year out, we see the same mistakes in ballot design, with the same results: voters disenfranchised by confusing ballot design and instructions. Too frequently, these events raise serious questions about whether the intended choice of the voters was declared the winner of the contest. With better ballot design, 2008 could be the year we put an end to that," Norden concluded.


Usability Professionals' Association
promoting usability concepts and techniques worldwide
Phone + 1.630.980.4997         office@upassoc.org

Contact the Voice