Explaining the Role and Impact of Gender within Online Political Discourse: Insights from Explainable Machine Learning

Politicians on social media shape public understanding of issues and have the potential to influence voting decisions. Social media platforms often reflect existing inequalities such as gender differences. This study examines the role of gender in online political discourse using a combined method approach. We focus on posts on X of fourteen German federal politicians from 2021 to 2024 to identify gender-based differences in language and content. The key contribution lies in developing a machine-learning classification model. We use explainable artificial intelligence to examine the linguistic and content-based features of gendered communication. Findings reveal that women are less visible in public political discourse and receive significantly fewer public reactions (likes, replies, reposts) than men. Female politicians are more likely to discuss a broader range of topics such as social issues, foreign policy, humanitarian aid, and use emojis, whereas male politicians tend to frame topics such as the COVID-19 pandemic, surveillance, and digitization.