From Personalization to Polarization: How Instagram's Algorithms Shape Pathways to Gender-Based Hate

Abstract

As Gen Z men trend more conservative and Gen Z women more liberal, scholars have identified social media as a key site of ideological polarization, especially around gender. While much cultural attention has been given to the “Alt-Right Pipeline” for exposing young men to misogynistic content, little empirical research has examined whether algorithms actually systematically promote Gender-Based Hate or whether similar patterns exist for women. This study addresses that gap through a sock puppet audit of Instagram’s recommendation algorithms. Eighteen sock puppet accounts, assigned to different content-preference treatments, were used to collect recommended posts over time. Posts were qualitatively coded using a deductive approach to assess three dimensions of diversity based on Michiels et al.'s (2022) definition of filter bubbles. To capture hostility toward any gender, this study introduces the concept of Gender-Based Hate (GBH). Findings offer insight into how algorithmic personalization may reinforce gender-hostile ideologies among young users.