Towards a More Engaged Democracy: Using Statistical Methods to Detect Election Fraud in the 2022 Philippine National Elections

• Main Authors: Cardaño, Juan Miguel, Mande, Bryan Patrick, Tionko, Seth William, Asuncion, Aldrich Ellis C, Briones, Jeric C.
• Format: text
• Published: Archīum Ateneo 2024
• Subjects: National Elections; Fraud Detection; Statistical Models; Voting Irregularities; Engaged Citizenry; Civic and Community Engagement; Politics and Social Change; Public Affairs
• Online Access: https://archium.ateneo.edu/mathematics-faculty-pubs/279; https://archium.ateneo.edu/context/mathematics-faculty-pubs/article/1280/viewcontent/ES_ML_2024_S_076.pdf
• Institution: Ateneo De Manila University

This paper aims to show how citizens can further participate in the elections, aside from just voting. Given the accusations of fraud, this study demonstrates how existing election fraud detection methods can be used in the 2022 Philippine National Elections (PNE) context. Specifically, histograms known as 2D vote turnout distributions were first utilized to model the frequency of the winner's percentage of votes based on the voter turnout in each electoral unit, with key parameters introduced to detect fraud. Vote turnout distributions were then simulated using parametric models involving the aforementioned fraud parameters, with the goal of replicating the actual election fingerprints. Using provincial election data for the presidential elections, initial results showed that the presidential election data in the 2022 PNE proved to be inconclusive for some aggregate levels. That is, compared to other countries, fraud cannot be detected on a large scale, but rather on a smaller scale. For example, in one town in the country, the fraud parameters indicated the possibility of fraudulent mechanisms present in the election data. Thus, using statistical tools for electoral fraud detection can empower the citizens to be more vigilant and engaged in their country’s democratic processes.