We need to go deeper: measuring electoral violence using convolutional neural networks and social media

Abstract

Electoral violence is conceived of as violence that occurs contemporaneously with elections, and as violence that would not have occurred in the absence of an election. While measuring the temporal aspect of this phenomenon is straightforward, measuring whether occurrences of violence are truly related to elections is more difficult. Using machine learning, we measure electoral violence across three elections using disaggregated reporting in social media. We demonstrate that our methodology is more than 30 percent more accurate in measuring electoral violence than previously utilized models. Additionally, we show that our measures of electoral violence conform to theoretical expectations of this conflict more so than those that exist in event datasets commonly utilized to measure electoral violence including ACLED, ICEWS, and SCAD. Finally, we demonstrate the validity of our data by developing a qualitative coding ontology.

Electoral violence: An introduction

“…However, we need to expand inquiry into other arenas of violence. For example, social media forms a space for threats and intimidation during election periods (Muchlinski et al, 2019). Additionally, studies on the gendered impacts of electoral violence show how female voters and candidates often face violence in the private space of their home, away from the public limelight (Bardall, 2011; Bjarnegård, 2018).…”

The electoral success of social media losers: a study on the usage and influence of Twitter in times of elections in Paraguay

“…Machine learning and natural language processing were employed to monitor Twitter, looking for the existence of violent incidents (e.g., physical attacks on opposition politicians, violence between opposition and incumbent parties, candidate assassination) during the 2015 parliamentary election in Venezuela (Yang et al 2016). Also, in the context of the Venezuelan parliamentary election of 2015, Muchlinski et al proposed a neural networks approach to estimate forms of political violence from tweets (Muchlinski et al 2021). But researchers did not only introduce approaches to monitor electoral incidents through social media but also proposed mechanisms to report them.…”

Twitter as research data

“…The sample representativeness of the general population of Twitter data is still debated, but it appears to be highly contingent on the research question. For certain topics, research has shown that Twitter is a credible source to reproduce real-world outcomes such as protest and violence (Muchlinski et al, 2021; Sobolev et al, 2020; Steinert-Threlkeld, 2018; Zhang & Pan, 2019). However, for other research questions, such as studying public opinion across the United States, we align with other scholars who express caution about extrapolating Twitter users’ opinions to other Americans, such as about people’s happiness level (Jensen, 2017).…”