Language Models Learn Metadata: Political Stance Detection Case Study
This work addresses stance detection for social science applications, but it is incremental as it focuses on improving metadata integration in an existing task.
The paper tackled the problem of optimally incorporating metadata into political stance detection, showing that a simple baseline using only party membership surpasses the state-of-the-art and that prepending metadata to speeches outperforms all baselines.
Stance detection is a crucial NLP task with numerous applications in social science, from analyzing online discussions to assessing political campaigns. This paper investigates the optimal way to incorporate metadata into a political stance detection task. We demonstrate that previous methods combining metadata with language-based data for political stance detection have not fully utilized the metadata information; our simple baseline, using only party membership information, surpasses the current state-of-the-art. We then show that prepending metadata (e.g., party and policy) to political speeches performs best, outperforming all baselines, indicating that complex metadata inclusion systems may not learn the task optimally.