samkanej/Dual-Causality-Models

This repo walks through the statistical models I created to test dual causality factors between regime type and armed conflict in Africa during the course of my undergraduate studies. # Dual Causality Models This repo walks through the statistical models I created to test dual causality factors between regime type and armed conflict in Africa during the course of my undergraduate studies. **Files:** The following files are included in this repo: * An Rmd file containing the code used to clean and create datasets, write and execute statistical models, and output effects plots; * The five original datasets that I used to create a bespoke dataframe for the purposes of my analysis; * Word documents with a sample and the complete text of the final paper I submitted for my undergraduate senior seminar in African conflict; and * A PowerPoint presentation sumamrizing the research project's methodology and key findings. ## Background In the spring of 2020, I completed a senior seminar in African conflict as part of my major in government. Throughout the semester, I completed a mixed-methods research project, applying my skills in statistical modeling and R to analyze the dual causality between African regime type and armed conflict. Dr. Laura Seay, my advisor and ultimate guide through my undergraduate studies, taught this course and offered tremendous support as I ventured into the quantitative realm – a generally untapped option within the Colby government department. ## Abstract This paper examines the relationship between African states’ regime types and intrastate conflict onset, severity, duration and outcome. A robust body of literature on regime type and conflict focuses largely on global trends of a one-way directionality, using regime type to explain conflict. I expect a causal mechanism to operate in both directions between the two: civil wars can affect the kind of regime that emerges post-conflict and an African state’s regime type can influence the onset, severity, duration, and outcome of an internal conflict. I rely on logistic regression, multiple linear regression, and ordinary least squares regression models to test these association …