ADVANCED QUANTITATIVE METHODS IN POLITICS AND INTERNATIONAL RELATIONS

The aims of this course are to provide an in-depth understanding of advanced quantitative methods and Machine Learning (ML) applications within the context of political science. Students will gain proficiency in the simple and multiple regression framework, understanding inference in regression models, goodness of fit, and interpretation of regression coefficients. Issues such as multicollinearity and interaction effects will be addressed, along with regression models for categorical dependent variables like logistic regression and probit models. Advanced regression techniques, including Poisson and negative binomial regressions, as well as an introduction to the multilevel regression framework, will be covered. Additionally, the course will delve into causal inference topics, such as instrumental variables (IV), difference-in-differences (DiD), and regression discontinuity designs (RDD). Practical applications of Machine Learning in political science will also be explored, encompassing supervised learning (e.g., classification, regression, random forests) and unsupervised learning (e.g., clustering, dimensionality reduction), with concrete examples of ML applications in policy planning analysis, as well as international relations.

It is preferable to have a basic knowledge of regression models in quantitative analysis and basic data analysis tools. However, a recap will be conducted at the beginning of the course to ensure that all students have the opportunity to learn and use the most recent data analysis techniques.

By the end of this course, students will have developed a sound understanding of advanced quantitative methods and machine learning applications in the field of political science, particularly in policy impact analysis and evaluation. They will be able to competently assess the effectiveness and implications of various policies through sophisticated regression techniques and causal inference methods. The course will enable students to address issues such as multicollinearity and interaction effects, enabling them to draw meaningful insights from complex data sets. Particular emphasis will be placed on the practical application of machine learning, ensuring that students can implement these tools to predict outcomes and analyze the effects of policy decisions. Through practical experience, students will learn how to evaluate and validate models, applying these techniques to real-world scenarios such as election forecasting and international relations. This comprehensive training will prepare students to use advanced quantitative methods and machine learning to conduct in-depth and impactful policy assessments, ultimately contributing to more informed decision-making in the political sphere.

Course objectives and structure. Overview of advanced quantitative methods and Machine Learning (ML). Regression analysis I: recap of simple and multiple regression framework. Inference in the regression model. Goodness of fit and interpretation of regression coefficients. Multicollinearity and interaction effects. Regression analysis II: regression models with categorical dependent variables. Logistic regression, probit models, multinomial logistic regression. Advanced regression techniques: Poisson and negative binomial regressions. Introduction to the multilevel regression framework. Causal Inference: Causality and correlation. Instrumental variables (IV). Difference-in-differences (DiD). Average treatment effects (AVEs) and regression discontinuity designs (RDD) Machine Learning Applications. Introduction to machine learning in political science. Supervised learning (e.g., classification, regression, random forests), Unsupervised learning (e.g., clustering, dimensionality reduction). Model evaluation and validation. Examples of machine learning applications in predicting election outcomes, policy analysis, and international conflict

Teaching and discussion materials provided by the teachers

This course employs a practical, application-based approach to learning, emphasizing real-world cases and existing literature accompained by theoretical knowledge. Students will engage with case studies from various domains within political science, which will serve as the foundation for understanding advanced quantitative methods and machine learning applications. By analyzing specific instances, students will understand how theoretical concepts are applied in practice. Reviewing and discussing existing scientific studies and report will help students understand the development and application of quantitative methods and machine learning in political science research. Each topic will be introduced through practical applications, showcasing how theory is employed to solve real-world problems, ensuring students see the relevance and utility of the concepts being taught. Students will actively participate in the discussion and research implementation, by applying the concepts learned to real datasets, reinforcing theoretical knowledge through practical experience using the software R. Lectures will be interactive, encouraging discussions and critical evaluations of practical applications. These discussions, based on case studies and research papers, will facilitate deeper understanding through collaborative learning. Concepts will be introduced incrementally, starting from basic principles and advancing to more complex techniques. Each new topic will build on the previous one, ensuring a comprehensive understanding of advanced quantitative methods and machine learning.

The final assessment for this course is designed to comprehensively evaluate students' understanding and application of the concepts and techniques covered throughout the term. The exam will consist of two main components: the submission of a research paper and an oral discussion. Students will be required to write a research paper of 3000-5000 words. The paper should present a clear and relevant research question, supported by a thorough review of the relevant literature. The methodology section should detail the quantitative and/or machine learning methods used, including a discussion of the hypotheses and the theoretical model. In the data analysis section, students will present and interpret their results, evaluating the goodness of fit of their model and discussing any issues such as multicollinearity and interaction effects. The discussion should interpret the results in relation to the research question, highlighting the political or social implications of the findings and acknowledging any limitations of the study. Finally, the conclusion should summarize the key findings and propose directions for future research or practical applications. In addition to the written paper, students will participate in an oral discussion, scheduled shortly after the paper submission deadline. During this discussion, students will present their research findings and respond to questions. This oral component aims to assess the students' ability to articulate their research process and findings clearly and to engage in a critical dialogue about their work. The combination of a written paper and an oral discussion ensures a well-rounded evaluation of both written communication skills and the ability to think on one’s feet and defend one's research conclusions. This approach provides a comprehensive assessment of the students' mastery of the advanced quantitative methods and machine learning applications covered in the course.

The identification of the topic for the final paper to be presented in the examination, can also be the start of the study and research for the dissertation.