“We get to do work that is actually making its way into the real world. It’s so important,” says Sarah Morin. She’s referring to the research that she and Grant McClearn, both juniors studying computer science, do with Dr. Poorvi Vora (Computer Science). Both of them have been working with Dr. Vora on election auditing for more than a year.
“The goal of an audit is to confirm that the reported winner really did, in fact, win,” explains Grant. “We do this by drawing a series of samples from the ballots. The audit stops and confirms the reported outcome when there is convincing statistical evidence that the reported winner did, in fact, win.”
The number of ballots that the auditors have to draw for an election depends very much on the margin of the winner’s victory rather than the total number of ballots cast; the bigger the margin of victory, the fewer the number of ballots they need to sample. Grant and Sarah had a taste of this last November, when they helped Dr. Vora and her collaborators pilot a new approach to statistical election audits in Mercer County, PA. The state plans to deploy statistical election audits with rigorous error guarantees, known as risk-limiting audits, all over the state in 2020.
Leading up to the pilot, they each had their own research roles. Grant worked primarily on trying to increase the efficiency of the audits, that is, minimizing the number of ballots that elections officials have to draw from ballots cast in order to confirm their accuracy. Sarah converted into Python code the statistical methods that Dr. Vora generates; she then ran the methods on various elections data to determine whether the methods worked in an experimental setting.
Sarah and Grant both believe they’re learning a lot from their research. They mention learning how to present their work and how to write abstracts for conferences, for example. But the research is “first and foremost, a lot of statistics,” Grants says. Sarah seconds that with a hearty, “Oh, yeah.”
They also appreciate getting to hear from other people who are leaders in the field—people with whom Dr. Vora works directly—as well as the chance to talk about their research with others.
“It is also fun to have your work make sense to people outside of your field. It’s pretty easy for us to explain how an audit works, and it’s nice to be able to share it with people,” Sarah states.