On Auditing Elections When Precincts Have Different Sizes

Abstract

We address the problem of auditing an election when precincts may have different sizes, and suggest methods for picking a sample of precincts to audit that takes precinct size into account. one method yields optimal auditing strategies together with an exact measure of its effectiveness (probability of detecting corruption of a given size). We restrict attention to basic auditing strategies, in which each precinct P is audited independently with some probability p determined by the auditor. The auditing probability for a precinct will depend on the size of the precinct, with larger precincts audited more frequently; when all precincts have the same size they will have the same probability of being audited.