TikStance: TikTok Political Stance Data
TikStance links TikTok videos, metadata, and parent-linked comment trees for stance analysis toward Trump, Biden, and Harris.
TL;DR — TikStance is a multimodal, context-aware TikTok dataset for political stance detection. It contains 161 videos and 13,876 comments linked as parent-linked comment trees, covering Donald Trump, Joe Biden, and Kamala Harris during the 2024 U.S. election cycle. Items are labeled for video-to-target and comment-to-target stance using Favor, Against, and None. Final Krippendorff's alpha values are 0.743, 0.723, and 0.722 for the Trump, Biden, and Harris subsets, respectively.
Core contribution
TikStance is presented as a dataset contribution for political stance detection on TikTok. The abstract states the motivation directly: political discourse has moved to short-video platforms, while computational analysis is constrained by the "scarcity of datasets that jointly preserve audiovisual information and hierarchical conversations."
The dataset's core contribution is the combination of host videos, video metadata, and parent-linked comment trees. Evidence: "Each discussion unit links a host video and its metadata to a parent-linked comment tree." This makes the corpus relevant for stance analysis that uses more than isolated text comments.
The abstract does not claim a new model or benchmark performance. It says TikStance "supports research in multimodal stance detection, political communication, computational social science, and context-aware natural language processing."
Full analysis, extracted claims, numerical results, entity graph, FAQ, related work, applications, and BibTeX are available via x402 micropayment.