Random Manual Risk-Limiting Post-Election Audits

Abstract

This article presents post-election audits designed to limit the risk of certifying an incorrect election outcome to any desired low probability, say α = 0.05. Its contributions are to: provide accurate upper margin error bounds for single or multiple winner election audit sampling; describe Stenger’s numerical method for precisely calculating uniform probability sample sizes and revise Aslam et al.’s weighted by error bound sampling method; provide a threshold for net margin error to trigger an expanded audit; explain why the maximum level of undetectability by observation assumption of less than one (1) for multi-ballot audit units requires that the runner-up or losing candidate(s) be allowed to select one or more additional precincts to be audited; and provide a sample R program for 3 sampling methods for risk-limiting post-election audits.

Keywords: vote miscount, post election audit sampling, risk limiting election audits, expanding election audits, election margin error bounds

JEL Classification: C60