Fake It Till You Make It: Behind the Scenes of Bot-Driven Popularity

• In Event: Digital Media Campaigning: Innovative Approaches and Questions

Abstract

Can a politician's social media fame influence voters' decisions on public policy, campaign donations, and voting? If yes, bots can artificially inflate social media popularity, and their influence is implausible to fact-check. My research explores the manipulation of social media popularity through programmable bots.
First, using an online survey experiment, I demonstrate that social media engagement metrics (likes, shares, comments, etc.) are indeed a heuristic for perceived electability, with the most pronounced effects on inattentive respondents, i.e., politicians with lots of social media engagement are thought to be most likely to win.
Next, I used the 'Botometer' algorithm to estimate the prevalence of bots among 435 U.S. Congress members. I then proposed a novel theory linking online bot activity to real-world TV news coverage. To establish causality, I leveraged the Twitter API shutdown in January 2023, which significantly hindered bot activity and substantially reduced followers only for politicians with a high level of bot activity. Moreover, I found that fewer bots on the platform impacted these politicians' visibility in broadcast news channels during the post-API shutdown period (February 2023 to August 2023). While visibility declined for all politicians, the drop was more pronounced for politicians with high bot percentages. These effects were only noticeable on mainstream channels like CNN, MSNBC, and FOX News.

Author

• Manu Singh, Columbia University