Opinion How to Fix an Election by Sean Savage Editor’s note: this opinion piece was originally submitted shortly after the 2002 elections, but it is just as relevant today. About once a month I join my fellow native Floridians in a massive group cringe as the latest piece of staggeringly shameful Sunshine State news rockets its way around the planet. The latest election embarrassment hit me harder than most Florida fiascos because Human Computer Interaction professionals and journalists were to blame, and I'm an HCI geek and an ex-reporter. We didn't learn a thing when Florida made itself the butt of barroom jokes from Stockholm to Singapore by ruining the 2000 presidential election. We made the same mistakes last week. Let's reflect on this for a moment, before we botch another election. The media overlooked the core problem behind the gubernatorial election screwup, just as they did in the stories about the presidential-campaign butterfly-ballot screwup. This oversight will not recur if Human Computer Interaction professionals do their job, if they explain to the media and the public the importance of involving users in technology design. Here's the core problem: the vote-handling system in question doesn't work; it fails because it was not designed for the people who use it. Rather than dealing with this, most news stories focus on whether voters and poll workers were trained long enough, whether laws were broken in the handling of votes, how results were analyzed, whether there's a conspiracy afoot to steal the election, and so on. These latter questions are important but they're secondary to the core problem; whether or not you have a conspiracy on your hands, you still have a broken ballot system. The New York Times editorial page echoed most news outlets Sunday in its analysis: "...it appears that most of the problems were caused by improperly trained workers and by voter confusion." This is like saying the World Trade Center fell because the weather got really hot for a few hours in those middle floors. Dade County may have dropped the ball in training poll workers. But when people are expected to undergo 12 hours of training before they can operate a simple ballot machine, something is horribly wrong. Reporters, like the rest of us, expect new technologies to be complicated and difficult to use. After decades of wrestling with the blinking "12:00" on the VCR, who can blame them for forgetting the whole point in designing computerized ballot systems: to make them easier to use and less error-prone than their predecessors? Why were the ballot devices so confusing? Because the designers failed to bring the voters and poll workers into the design process. Most people don't understand that the years of intense training and hard work that turn a person into a talented, capable engineer simultaneously convert that person into a special sort of creature. Such creatures can build fantastic devices that give us wonderful powers, but in mastering these technical intricacies they lose the ability to see the world through the eyes of a regular person. So if these creatures develop a tool completely on their own, if they presume to understand the regular people who will use the tool, those regular people will invariably end up with a tool that confuses the hell out of them. At best, regular folks will let that annoying "12:00" keep blinking because setting that clock is such a hassle. At worst they'll press the wrong button and crash an airplane or bungle an election. Human Computer Interaction workers understand all this because they've watched this typical cycle of failure unfold countless times to produce countless unusable products. HCI workers study a tool's intended users, and they involve those users with engineers during the design process to ensure that the tools will work for their intended audience. They tailor tools to users, so that nobody has to take a 12 hour class before using a simple voting booth. Most reporters have never heard of a Human Computer Interaction professional, so who will a reporter call regarding the design and implementation of new voting technologies? At best, she'll call the engineers who built the ballot machines and who, in the case of a fiasco like this, are out of touch with their audience and honestly view the problem as a lack of training, a lack of tailoring the users to the tool. So the media never understand this: tools that work are tailored to their users. From the ground up. So the public, and the civil servants who purchase voting machines, never recognize the real problem. So the problem worsens and votes go uncounted. Journalists need to learn what Human Computer Interaction professionals do; they need to consider the HCI perspective when covering stories about failing systems and confusing tools, so the public and the government can solve these problems. Before this can happen, Human Computer Interaction people need to educate media people about their profession. Let's fix this problem. Spread the word. SIGCHI Bulletin May/June 2003 13

/lp/association-for-computing-machinery/how-to-fix-an-election-EZshbivoL7