Forensic Techniques for Detecting Election Fraud (2024 ESRA Issue)

The idea of free and fair elections is the foundation of a democratic society, where every vote must be counted equally. Elections are essential mechanisms for ensuring accountability, transparency, and representation. Fraud and manipulation corrupt this system, silencing voters and eroding trust. When fraud is uncovered—or even merely suspected—it undermines legitimacy and can spark repression, unrest, or civil conflict. Determining whether electoral outcomes genuinely reflect voter preferences or instead result from manipulation is, therefore, crucial. The credibility of representative democracy depends on the integrity of its elections. Until recently, few quantitative methods existed for assessing electoral fraud beyond in-person observation. Electoral forensics has emerged to fill this gap by analyzing anomalies in results that may signal irregularities. However, scholars continue to debate which forensic tools are the most effective, and no framework currently integrates multiple approaches. This article surveys recent quantitative and machine learning methodologies used in election forensics, with particular attention to their ability to identify at-risk polling stations and estimate the scope of manipulation. We review applications across contexts, evaluate their strengths and limitations, and highlight persistent challenges. In doing so, we provide an updated foundation for future research and practical guidance for election administrators seeking to safeguard elections.