Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Analyzing political stances on Twitter in the lead-up to the 2024 U.S. election

Hazem Ibrahim, Farhan Khan, Hend Alabdouli, Maryam Almatrooshi, Tran Nguyen, Talal Rahwan, Yasir Zaki

TL;DR

The study addresses polarization and ideological dynamics on Twitter around the 2024 U.S. election by applying a tri-LLM ensemble to classify ideological stance into five categories for 1,235 candidate tweets and 63,322 replies, with human validation. It reveals asymmetric framing between candidates and constituents, with Republicans more often criticizing Democrats and replies showing divergent engagement patterns. Event-driven shifts are quantified using a regression-discontinuity design around major political events, highlighting rapid changes in AR/PR versus PD/AD activity. The work offers actionable insights for policymakers and platforms seeking to mitigate polarization and promote healthier online discourse during elections.

Abstract

Social media platforms play a pivotal role in shaping public opinion and amplifying political discourse, particularly during elections. However, the same dynamics that foster democratic engagement can also exacerbate polarization. To better understand these challenges, here, we investigate the ideological positioning of tweets related to the 2024 U.S. Presidential Election. To this end, we analyze 1,235 tweets from key political figures and 63,322 replies, and classify ideological stances into Pro-Democrat, Anti-Republican, Pro-Republican, Anti-Democrat, and Neutral categories. Using a classification pipeline involving three large language models (LLMs)-GPT-4o, Gemini-Pro, and Claude-Opus-and validated by human annotators, we explore how ideological alignment varies between candidates and constituents. We find that Republican candidates author significantly more tweets in criticism of the Democratic party and its candidates than vice versa, but this relationship does not hold for replies to candidate tweets. Furthermore, we highlight shifts in public discourse observed during key political events. By shedding light on the ideological dynamics of online political interactions, these results provide insights for policymakers and platforms seeking to address polarization and foster healthier political dialogue.

Analyzing political stances on Twitter in the lead-up to the 2024 U.S. election

TL;DR

The study addresses polarization and ideological dynamics on Twitter around the 2024 U.S. election by applying a tri-LLM ensemble to classify ideological stance into five categories for 1,235 candidate tweets and 63,322 replies, with human validation. It reveals asymmetric framing between candidates and constituents, with Republicans more often criticizing Democrats and replies showing divergent engagement patterns. Event-driven shifts are quantified using a regression-discontinuity design around major political events, highlighting rapid changes in AR/PR versus PD/AD activity. The work offers actionable insights for policymakers and platforms seeking to mitigate polarization and promote healthier online discourse during elections.

Abstract

Social media platforms play a pivotal role in shaping public opinion and amplifying political discourse, particularly during elections. However, the same dynamics that foster democratic engagement can also exacerbate polarization. To better understand these challenges, here, we investigate the ideological positioning of tweets related to the 2024 U.S. Presidential Election. To this end, we analyze 1,235 tweets from key political figures and 63,322 replies, and classify ideological stances into Pro-Democrat, Anti-Republican, Pro-Republican, Anti-Democrat, and Neutral categories. Using a classification pipeline involving three large language models (LLMs)-GPT-4o, Gemini-Pro, and Claude-Opus-and validated by human annotators, we explore how ideological alignment varies between candidates and constituents. We find that Republican candidates author significantly more tweets in criticism of the Democratic party and its candidates than vice versa, but this relationship does not hold for replies to candidate tweets. Furthermore, we highlight shifts in public discourse observed during key political events. By shedding light on the ideological dynamics of online political interactions, these results provide insights for policymakers and platforms seeking to address polarization and foster healthier political dialogue.

Paper Structure

This paper contains 4 sections, 2 figures, 1 table.

Table of Contents

  1. Introduction
  2. Methodology
  3. Results
  4. Limitations and Future Work

Figures (2)

  • Figure 1: The proportion of candidate tweets (A) and candidate tweet replies (B) classified as Pro-Democrat, Anti-Republican, Neutral, Anti-Democrat, or Pro-Republican (*: $p < 0.05$, **: $p < 0.01$, ***: $p < 0.001$). The average number of replies of a certain ideological stance made to (C) Democrat candidates and (D) Republican candidates per candidate tweet, as well as the average engagement metrics (likes, retweets, views, replies) such replies receive.
  • Figure 2: RDiT analysis of the ideological stance of tweets in the two weeks surrounding major political events. (A) First presidential debate between Biden and Trump; (B) Supreme Court ruling that a president has absolute immunity from criminal prosecution for core constitutional powers; (C) Trump's attempted assassination during a rally in Pennsylvania.