2025 Theses Doctoral

Posting Politics: Essays on the Supply Side of Social Media

Rutherford, Kylan

This dissertation explores how social media engagement can shape political content on social media by shaping incentives, stimulating content creation, and steering algorithmic curation. Across three papers, I reconcile interdisciplinary findings with the unique attributes of political content.

My first paper presents an analytical framework for understanding how consumer interests and algorithmic sorting influence the types of content produced on social media platforms. Building off of a Downsian framework, I model two producers who adjust the content they create in order to maximize their reach, given the production point of their competitor. Unlike typical Downsian models, social media engagement can come both from preferences being very close to content, or very far, what I term concordant and discordant engagement, respectively. I show that polarization of content production can occur with a sufficient prevalence of discordant engagement, even without polarization in the population or producer preferences. I support this finding through interviews with content creators, including media staffers for Members of Congress.

In my second paper, I investigate how engagement signals affect the production of comments and original posts, in both political and non-political subreddits. I conduct a series of field experiments on Reddit, contrasting both commenting and posting behaviors. I find that Reddit Awards seem to incentivize increased comments for new users, but do little to move veteran Redditors. I find weak evidence for a relationship in the opposite direction for individuals who post, rather than comment. These results suggest that engagement can affect certain types of original content production, including political content. However, posting and commenting are different behaviors that appear to have distinct relationships with engagement and user tenure.

My third paper presents two TikTok experiments designed to highlight how algorithms respond to engagement signals. Again, my aim is to highlight how political content is treated. I conduct an algorithmic audit to show how engagement signals can alter initial recommendations. I find that effect sizes are conditional on the topic of interest, noting that engagement with political content appears to trigger relatively high rate of related recommendations. I support these audit results with a lab experiment, examining how initial engagement signals persist over time. I observe algorithmic behavior over 40 minutes of browsing by treatment-blind users. I find that political recommendations persist for treated accounts, even after significant browsing time. I also present preliminary results for algorithmic effects on user attitudes and experiences.