Political Content Engagement Model: A large-scale analysis of TikTok political video content features and audience engagement

Abstract

TikTok has emerged as a prominent platform for political information dissemination, where traditional news organizations, political figures, grassroots organizations, and influencers engage audiences on political and civic issues. However, limited research has systematically examined why politically oriented TikTok videos attract engagement. This study introduces the Political Content Engagement Model (PCEM), which explains how political identity, content features, content sources, and topic issues influence engagement. Using a dataset of 578,420 TikTok videos posted by 9722 elite accounts, we use machine learning and topic modeling to analyze how features such as political party references, issue framing, justification, sentiment, civility, and mobilization appeals affect video engagement. Besides, we investigate differences in engagement patterns between liberal- and conservative-leaning TikTok accounts and differentiate between internal and external engagement behaviors. Across both liberal and conservative accounts, civility level and out-party critique consistently emerge as the most powerful predictors of political TikTok video engagement. Our findings contribute to the field of digital political communication by offering insights into TikTok users’ political engagement behavior on TikTok and how different content strategies drive audience interactions.