AI Feedback Enhances Community-Based Content Moderation through Engagement with Counterarguments

This project includes two experiments: (1) the note-writing experiment, in which participants were asked to write a note providing additional context to a social media post and then revise that note after receiving feedback from an AI agent; and (2) the note-evaluation experiment, where participants were shown notes created in the previous experiment and asked to rate each note based on its helpfulness and its perceived partisan leaning (i.e., whether it favors Republicans or Democrats). Detailed information about both experiments is provided in the main paper. This repository includes the raw data files from the experiments, along with the cleaned datasets for both participants and notes.