Speakers
Description
While the online behaviour of organised hate groups has been extensively studied, only recently has attention focused on the behaviour of individuals that produce hate speech on mainstream platforms. One of the most relevant targets of online hating is women. To identify and analyse producers of derogatory, offensive, insulting, denigrating content towards women, we will focus on content shared on Twitter, given the relatively easy access to networked data based on friend/follower relations. We created an extensive corpus of tweets, gathering Twitter contents containing keywords related to a selection of politically-active women, feminists, female entrepreneurs, journalists, and influencers. A misogynistic tailored lexical dictionary was applied to classify the tweets and derive an index of misogyny at the producer level to select a subset of seed users that are more likely to share misogynistic contents. Trainees, who are engaged in an internship program, analysed in detail these seed users, considering their tweets and other metrics such as the number of contents shared and the number of friends/followers. We retrieved the friend/follower networks of the selected seed producers, their features, and textual contents to classify the nodes in the network and to identify which nodes are more likely to share misogynistic contents. The network relational information provided by the intrinsic nature of social network platforms can be included in the classification task. The relevance of relational data is explained in terms of homophily: knowing that members of a particular community are prone to creating abusive content, and knowing that a given author is connected to this community, allows us to use the information beyond linguistic signals. Finally, we investigated the resulting classified network to recover the structure of the misogynistic network.
Project funded by EU Next Generation, MUR-Fondo Promozione e Sviluppo-DM 737/2021
| Topics | • Textual data analysis and network methods |
|---|---|
| Keywords | Network analysis, Misogynistic networks, Twitter, Social Media analysis; Misogynistic lexical dictionary, Network data, Misogynistic language |
