Dr. Vora awarded NSF grant to study elections auditing

Dr. Poorvi Vora was awarded a one-year, $200,000 NSF RAPID grant to develop risk-limiting tabulation audits. The project, “RAPID: Sequential Sampling in Stages for Statistical Election Audits,” develops algorithms and code for real audits that sample ballots in stages of tens or hundreds or thousands at a time. It is hoped that the developed code will replace some of the existing code used in real audits, which has been developed for sampling a single ballot at a time. Preliminary testing reveals an improvement of factors as large as two and above in the number of ballots required to complete the audit if the election outcome is correct. This improvement is significant because states are just beginning to try out this type of audit, and a large workload will be discouraging for other states. Risk-limiting audits are required by law in Colorado, Rhode Island, Virginia, and Nevada; and about twenty states are exploring performing risk-limiting statewide audits in November this year.