Founded in 2009 to develop tech-focused election tools, Clear Ballot is a leader in election technology innovation. Their browser-based software, used in conjunction with commercially available hardware, scales to election jurisdictions of all sizes and is already used in nine states.
To trial Clear Ballot’s latest electronic kiosk voting systems, Perkins hosted a mock election on its Watertown campus. Over the course of four days, the organization leveraged its deep ties to the blindness and broader disability community to test the system with 44 users, both with and without disabilities.
Working in accordance with mandated Common Industry Format (CIF) voter machine testing protocols, Perkins also provided the testing facility itself, directions for using the machines and recruited and trained more than 24 poll workers to conduct question-and-answer sessions with users to gain a better understanding of the challenges they faced during the mock election. As a final deliverable, Perkins Access created a comprehensive usability testing report that described the research method, experimental design, testing procedure, a summary of the test results and recommendations for improving the experience for voters of various abilities.
On the product side, voters used a variety of methods to mark their ballots, including touch screens (with and without pointing devices), a keypad and a sip-and-puff mechanism. Additionally, voters took advantage of different display settings, which included larger text size and various color and contrast options.
Perkins Access experts provided Clear Ballot with a written report detailing their findings, which the company used to create a more accessible product and keep in compliance with federal regulations.
- The accessible voting system was found to be usable for a broad spectrum of voters with diverse abilities
- 95.65% of voters in the test group were able to successfully mark and cast their ballots
- The mean accuracy of all voters in completing the 20 voting tasks as directed was 91.32%
- The average time on task for completing the ballot was 16.5 minutes, a significant decrease from the average time of 20.7 minutes in the 2016 Clear Ballot usability test
- The average amount of assists requested during the voting process was 2.48, a significant decrease from the average amount of assists (4.54) in the 2016 Clear Ballot usability test
- Poll workers served a critical role in assisting voters with disabilities with the voting systems, including helping with various system functions, printing and inserting ballots into ballot boxes and the scanning device
“We have the devices, we can get them somewhere for the study. But Perkins understands the etiquette of working with disabled people, they have connections to the disabled communities, a great campus, great facilities and underlying infrastructure. A lot of the things that would be difficult to impart to researchers who have never worked with disabled communities before, we don’t have to sweat that at all.”
– Ed Smith, Vice President of Product Development, Clear Ballot