Identity politics thrives in contemporary politics, and social media have become a crucial battleground. One reason is arguably the emotionality of identity politics that gets amplified by social media algorithms. Especially short video social media strongly convey human emotion through text as well as facial displays and tone of voice. It is therefore crucial to understand how identity politics takes shape emotionally on these platforms.

Due to experiences with marginalization, self-stereotyping and strategy, we expect differences between members of marginalized and dominant groups, arguing for group equality or dominant group hegemony. Therefore, the aim of this paper is to investigate to what extent dominant and marginalized group members show different emotions when discussing identity politics on social media.

We will study this in an automated visual content analysis of TikTok videos regarding a gendered issue. Although it cannot measure gender exactly, automated classifiers will approximate speakers’ genders from their physical appearance, as well as deduct emotions from their facial expressions, tone of voice, and verbal arguments. In addition, a supervised machine-learning model will distinguish whether speakers argue for gender equality or male hegemony. In combination, this allows us to map group emotions in identity politics.

As such, this paper combines essential theoretical contribution with methodological innovation to take a novel perspective on group communication in identity politics.