Indonesian Twitter Dataset for Hostile and Benevolent Misogyny Classification (ID-Misogyny Tweets)
Description
ID-Misogyny Tweets is a dataset of Indonesian Twitter/X posts annotated for hostile and benevolent misogyny. The tweets were collected using 17 predefined Indonesian-language search keywords, split into benevolent-sexism (8) and hostile-sexism (9) categories, and span January 2023 to July 2026. The dataset contains 5,563 tweets labeled into three categories: Hostile Misogyny (1,085), Benevolent Misogyny (444), and Non-Misogyny (4,034). Each tweet was labeled independently by three annotators following criteria grounded in Glick and Fiske’s Ambivalent Sexism Inventory, with the criteria and labels reviewed by an expert in clinical psychology, achieving an inter-annotator agreement of Fleiss’ κ = 0.86. The dataset is provided as a CSV file. Besides the tweet text and its label, it includes an English machine translation, the posting timestamp, the search keyword used, and engagement metrics (retweet, reply, like, quote, view, and bookmark counts) to support further research.
Files
Institutions
- Binus UniversityJakarta, Jakarta