Understanding Misogynoir: A Study of Annotators' Perspectives

WebSci '23: Proceedings of the 15th ACM Web Science Conference 2023(2023)

引用 0|浏览9
暂无评分
摘要
"Misogynoir" is the anti-Black racist misogyny experienced by Black women, which is characterised by components of both racism and sexism. Misogynoir is challenging to detect due to its inherent subjectivity and its intersectional nature, and people's opinions and interpretations of such hate might vary, which adds to the challenges of understanding it. In this paper, we explored how and some potential why's different annotator characteristics influence how they interpret and annotate a dataset for potential cases of Misogynoir and Allyship. We sampled tweets containing public responses to self-reported misogynoir cases by four prominent Black women in technology, designed an online annotation task study, and recruited annotators of diverse ethnicities and genders from the Prolific crowdsourcing platform. We found that participants' sources of evidence in judging and interpreting content for potential cases of Misogynoir and Allyship, even in circumstances where they all agree on a prospective label, vary across different factors, such as different ethnicity, lived experiences and gender. In addition, we present a variety of plausible interpretations influenced by the various annotators' characteristics. This study demonstrates the relevance of different annotator perspectives and content comprehension in hate speech and the need for further efforts to understand intersectional hate better.
更多
查看译文
关键词
intersectional hate,misogynoir,hate speech detection,datasets annotation,crowdsourcing
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要