Unraveling the determinants of domestic violence with decision tree analysis

Purpose This study aims to identify the key determinants of domestic violence in Türkiye by focusing on women's decision-making autonomy, fertility status and their partners' alcohol consumption. The research provides new insights into the complex dynamics of domestic violence using a machine learning approach. Design/methodology/approach The study applies decision tree analysis to nationally representative data from the Turkish National Research on Domestic Violence against Women surveys conducted in 2008 and 2014. The main variables include women's autonomy, education level, partners' alcohol use and the incidence of domestic violence. Findings Results reveal that women's autonomy and fertility are critical determinants of domestic violence. A non-linear relationship exists between autonomy and violence: mothers require higher levels of autonomy to avoid violence, whereas for childless women, low autonomy combined with a husband's frequent alcohol consumption significantly increases the risk of abuse. Originality/value This paper is the first to apply decision tree analysis to domestic violence data from Türkiye, offering a novel, data-driven framework for understanding the interplay between women's autonomy, male alcohol use and the risk of violence. The findings provide valuable implications for policymakers and researchers focused on violence prevention and women's empowerment.