Partisan habitual voters in South Korea: Employing Random Forests to understand Korean voters' electoral choices

Abstract Objective This study investigates whether recent political events in South Korea, such as a presidential impeachment and changes in leadership, have caused a realignment of voters' political affiliations. We aim to determine whether political party allegiance has transitioned into a pattern of habitual voting in the last ten years. Methods Employing a Random Forest model based on machine learning approaches, we examined polling data from the last ten years of Korean presidential elections. Instead of relying on hypothesis‐led methods, our analysis emphasizes the intrinsic characteristics of the data to uncover crucial elements affecting voter behavior in each election. Results Our results reveal a considerable consistency in voting patterns over the past decade, implying that allegiance to political parties has turned into a habitual behavior. Significant factors affecting voter choices in the 2022 election were also recognized as indicators of election outcomes five and ten years prior. Conclusion The study reveals that voter behavior in South Korea remains stable despite notable political changes, primarily influenced by habitual elements rather than substantial changes in party loyalty. This suggests that political allegiance has become ingrained as a habitual behavior for Korean voters.