Investigating the factors affecting femicide using poisson and negative binomial models: The case of Türkiye

Femicide is one of the most significant problems occurring both in Turkey and worldwide. Femicide is the killing of a woman simply because she is a woman. It is not merely a criminal case but also a significant social service issue. The rate of femicide in Turkey is considerable, and many factors influence it. This study aims to identify the factors affecting femicide. In this context, the number of femicides is considered the dependent variable, while unemployment rate, divorce rate, employment rate, crude birth rate, and education indices at the pre-school, primary, secondary, and higher education levels in the context of gender equality are included in the model as independent variables. Using these variables, the number of femicides is modeled using Poisson regression and negative binomial regression analysis methods. The study aims to identify the social, economic, and demographic factors affecting femicide and to contribute to policy development processes accordingly. Data for 2023 and 2024 were obtained from the Turkish Statistical Institute (TÜİK). Within the scope of Poisson regression and negative binomial regression models, information criteria were used to determine the appropriate link function. Using the most suitable link function, model variables were estimated using Poisson regression analysis and the negative binomial regression method. When the results were examined, it was observed that for 2023, the divorce rate and crude birth rate had a positive and significant effect on the number of femicides; other variables (unemployment rate, employment rate, education level according to gender equality) had no effect. Looking at 2024, while the divorce rate had a positive effect on the number of femicides, again, other variables (unemployment rate, employment rate, education level according to gender equality) had no effect.