The Use of Machine Learning Methods in Political Science: An In-Depth Literature Review

In the past decade, applying machine learning methods in political science has grown in popularity. The increase in data volume and sources motivated researchers to turn to these data-driven methods as an alternative to classical statistics. Several review papers have proposed theoretical typologies for applying machine learning in social sciences. We present an overview of how and why machine learning methods are actually implemented in the field of political science. The aim of this study is to conduct an empirical analysis of the political science literature that uses machine learning as a research method. We applied the PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) framework for systematic review studies to systematically select 339 articles (1990–2022) from Web of Science and Scopus, evaluated their relevance based on a set of inclusion criteria, and created a database with the key characteristics. Overall, we observed that political scientists have embraced machine learning as empirical method based on the increased use in the past 10 years. We found that the use of machine learning is the most prevalent in political communication and conflict and peace studies, and that topic modeling, support vector machine, and random forest are the most used methods. Our results indicate that reporting on optimizing machine learning models through hyperparameter tuning could be more transparent, and researchers should conduct their own benchmarking when choosing the most suitable model.