Print Friendly and PDF
only search openDemocracy.net

Democracy needs good design

About the author
Louise Ferguson is a research associate with the Work Foundation’s iSociety project, and co-authored its 2003 report Getting By, Not Getting On: Technology in the UK Workplace. She is a leader of the Usability Professionals’ Association International Voting and Usability Project, and has written widely about technology, user experience and voting design.

In the debate surrounding new voting technologies acquired by many individual states and counties across the United States during 2003-04, most commentators have missed the real reason why the country teetered on the edge of a voting meltdown in the presidential election of 2000: bad design resulting in poor usability

In Palm Beach County, Florida, the local supervisor of elections Theresa LaPore had redesigned the so-called “butterfly ballot”, a style of ballot-paper used in a number of American states. She increased the print size, and consequently needed to change the format and layout of the ballot. Voters found it confusing and made many mistakes in casting their vote; in the count, some 19,000 ballots were rejected and thousands of elderly Jewish voters who thought they had voted for Al Gore found they had cast their votes for rival candidate Pat Buchanan. Bush won Florida by 537 votes, and a wide range of problems with voting began to emerge.

"I was trying to make the print bigger so elderly people in Palm Beach County can read it”, said LaPore at the time. “We sent out sample ballots to all registered voters, and no one said a word.” LaPore created her ballot with the best of intentions, a wish to help the many elderly voters in her county; but good intentions alone do not lead to good design.

Theresa LaPore was also mistaken if she thought voters could tell a badly designed ballot just by looking at it. People need to use something to find the problems in it – and even then, problems won’t get reported unless someone is recording the event. To make matters worse, unless we know how someone intended to vote, how do we know whether they were successful? Many Palm Beach County voters “successfully”, but unintentionally, voted for Pat Buchanan.

So just how difficult is it to vote? What seems obvious to a person who knows the answer is frequently not at all obvious to someone new to the process – even if, as in the United Kingdom, it appears as simple as putting an “x” next to a name on a ballot-paper. Moreover, voting has some peculiarities: we vote infrequently, the process is different each time, and there is no individual feedback to tell us whether we voted as intended.

From Florida to London

In the last decade, British voters have had to deal with a range of new voting methods. At the time of the 2004 mayoral elections in London, which used the supplementary vote system, the Guardian published comments from several high-profile voters:

“There were no instructions on the paper at all”, said Channel 4 news presenter Jon Snow. “It simply said ‘what is your first and second choice for mayor?’. I said my first choice was ‘x’ and my second was the same guy, so I voted for him twice. But…you’re not allowed to vote for the same person twice.”

And so Snow, undoubtedly a bright person, unwittingly fell victim to a usability problem: he lost his second vote because the ballot-paper was poorly designed.

He wasn’t alone. In both 2000 and 2004, the Electoral Reform Society criticised the ballot-papers used for the London mayoral race. The design made it easy to do one of two things: cast two votes for the same candidate (invalidating the second-choice vote), or cast both votes in the same column (invalidating the entire ballot-paper). Because of inadequate statistics collection by London Elects, there is no clear record of how many voters erroneously cast both their votes for the same candidate.

Poor design can easily prevent 20% or more voters from being able to complete a task successfully – and elections are often decided on margins as slim as a few percentage points. But the need for good design clearly extends beyond the ballot paper, it can be a crucial element in the entire electoral process. Elections involve voter registration, finding out about candidates and procedures, casting votes and getting results. The whole process – from voting forms, websites, signage, election information to instructions on how to participate – needs to be considered; and other players in the voting “ecosystem”, such as electoral officials, also need to be included in the design approach.

The flaws in technology

Non-designers tend to assume that introducing new voting technologies might instantly produce a simple and easy-to-use system. True, innovations like electronic voting kiosks may bring benefits, such as enabling disabled users to cast their votes independently for the first time, or preventing overvoting (casting more votes than allowed); but there is no guarantee that such benefits will be designed at all.

Furthermore, new technologies can introduce their own problems: a voting kiosk may have screens that are difficult to navigate, use obscure language, or fonts or colours that are difficult to read; it may have a ballot layout that is confusing, software that makes it difficult or impossible to correct mistakes or review the ballot, or low audio quality for voters with low vision.

The events of 2000 in the United States provoked a drive to update voting technologies across the country. But in Florida, new touchscreen machines reportedly fail to record votes eight times more often than systems using optical scanning to count ballot-papers. Lack of research often means that such problems only surface after expensive systems have been purchased and put in place.

In the United Kingdom, electoral pilots using the internet, voting kiosks, digital television, telephone and text messaging have been introduced in recent years. Here too, badly designed new voting technologies create problems. In its 2003 report, “Polls Apart: developing inclusive e-democracy”, the disability charity Scope stated: “We do not believe any kiosks represent an improvement to the traditional ‘pen and paper’ method.“ Other technologies suffered from similar problems: “…across all the channels in all the pilots [it was felt that] the usability could be improved. It sounds obvious but making something as simple as possible will make it more accessible to more people.”

To ensure good design, usability needs to be built in from the beginning, it cannot be added on as an afterthought. Clayton Lewis’s rejection of the ‘peanut butter theory’ of usability is relevant here: “usability is not a spread that can be smeared over any design, however dreadful, with good results if the spread is thick enough.” To achieve usability, the needs of the people who have to use the system are the appropriate starting-point, not the requirements of the technology.

Elections are for people

This approach is known as user-centred or human-centred design. It is now being used in the commercial world, and is supported by a large body of research and international standards. It involves studying the problems people experience with existing or similar systems before proposing new designs; evaluating proposed designs by trying them out with real users; and using the evaluation results to improve the proposed design.

This process is quite different from market research. Opinion surveys and focus groups ask people what they think, but it’s necessary to see what they do, how things are used, to understand what the problems are and how design can improve matters. Nor should the process be confused with quality inspection: this has a different purpose (for example, checking that candidates’ names are correctly spelt) and would not involve real users.

Voting design and usability are finally on the American election agenda. The Chicago-based Design for Democracy project, for example, has worked with local administrations to improve the design of election materials. In April 2004, the US National Institute of Standards and Technology recommended to Congress that usability standards for voting be introduced. The US Elections Assistance Commission has announced the appointment of a technical committee to establish usability and other standards for voting.

The United Kingdom and other democratic countries have not yet suffered a high-profile, Florida-style, voting design disaster. This may help explain their relative lack of interest in the subject. But such a design disaster could change the results of an election and arouse widespread distrust in the electoral process. It’s not sensible to continue in an amateur fashion – leaving design and usability to chance, to untrained people, or to the whim of system vendors. Better, more usable ballots, forms, instructions, systems and processes can be designed that can help to deepen trust in the integrity of the democratic process. All that is missing is the political will.


We encourage anyone to comment, please consult the
oD commenting guidelines if you have any questions.