The Atlanta Journal-Constitution
Voting processes should be both secure and usable
It would be hard to find a time in our history where the design and performance of voting systems and procedures across our state and around our country have been under as much scrutiny as they are right now. In many ways, advancing technology has made the simple act of casting a secure vote even more complex. We are researchers who have spent much of our careers studying the usability of complex systems, including voting systems, by broad populations in difficult or diverse settings. For a system to rank high in usability it has to be more than just useful — it also must be efficient, easy and pleasant.
In June, the system for voting deployed during the Georgia primary election ranked extremely low in usability, which became very obvious with the resulting long wait times dampening participation.
While a robust elections system requires accurate, secure and fraud-free voting, a truly robust democracy also demands broad participation and representation. An election system with low usability deters all but the most resourced voter, while more user-friendly elections help ensure participation and representation across the voting public. That the Georgia primary voting system was not usable is not saying that it did not work but, instead, that it was difficult to use. For those who confronted long lines but didn’t have three hours available to wait, this lack of usability rendered the voting system, to them, useless.
At Georgia Tech, we responded to the usability challenge of the voting system in June by quickly assembling a team of students and faculty, partnering with Fulton County and launching a project to help ensure a safe, secure and highly usable election system for Fulton voters. Our work included research and development of platforms for voter wait time analysis, polling place layout optimization and COVID-19 safety measures.
We also surveyed in-person voters across a collection of Fulton polling locations during the early voting period for the general election in November and the runoff election in January. We asked questions about the polling system’s usability, perceived COVID-19 safety and overall voter satisfaction.
An analysis of the data reveals one clear overarching theme: Voters found arena polling locations more usable than other polling locations, including libraries, schools and churches. Voters who went to State Farm Arena, Mercedes-Benz Stadium or McCamish Pavilion were 10% more likely to have strong confidence in the integrity and anonymity of the election systems compared to those voters who did not go to stadiums. They were also 10% more likely to strongly agree that their voting experience was easy. And they were 20% more likely to report their experience as COVID-19-safe. Put simply, and with statistical significance, our survey participants reported stadiums as more usable.
State Farm Arena claims to be the nation’s first stadium to offer itself as a polling location, though many other arenas across the nation followed suit. All of us should be proud of the way our professional sports teams stepped up. The organizations that operate sports arenas are filled with professionals focused on customer experience and usability. So should sporting arenas continue to serve as large-scale early voting locations? Maybe. After all, stadiums often are built with considerable public support and taxpayer investment, and offering themselves as polling places is a great way for them to give back to their communities.
The usability concepts that are common in stadiums can also be applied to our more typical polling locations, such as churches, libraries and schools. More userfriendly election systems don’t require sporting venues. Instead, what is required is an investment in useful, efficient, easy and pleasant voting experiences. This investment will only happen with broad political support.
This month the Georgia Legislature is debating a range of elections bills that would have the effect of limiting the usability of our voting systems. One provision would prohibit the distribution of food or water to people waiting to vote, sure to make the experience less pleasant. Another would limit Sunday early voting, making access less easy. None of these restrictions makes our voting systems more usable. Instead, they mostly take us in the opposite direction, and in doing so are likely to diminish participation and representation at the polls.
If usability matters when we watch football and basketball, shouldn’t it matter just as much, if not more, as we perform the central action of our democracy and vote?
Michael L. Best is professor of international affairs and interactive computing at Georgia Tech and has worked globally on election systems monitoring for more than 15 years. Ellen Zegura is Regents professor in Georgia Tech’s School of Computer Science. Richard DeMillo is chair of the School of Cybersecurity and Privacy at Georgia Tech. Both have worked on election technologies and security for more than a decade.