C-SEPLN: Natural Language Processing to detect the gender gap in the media

Speaker: Prof. Maite Taboada, Department of Linguistics, Simon Fraser University

Summary: There is clearly a gender gap in the media. Studies, both qualitative and quantitative, have been clearly demonstrating for years that women are mentioned and quoted less in the news than men, and that when women are mentioned, it is often as victims or caregivers. The speaker's research group has studied this issue in depth, analyzing Canadian media data over the past four years. Using Natural Language Processing (NLP), they analyzed more than 1.5 million articles, and found that the proportion of women cited ranged from 27% to 32%, with peaks on certain topics (COVID-19) or at certain times of the year (International Women's Day). They have also quantified the difference in certain topics, where they have found that women are cited more in news about culture, health, and lifestyle. In contrast, news about politics, business, and sports cite more men. This talk will present the linguistic and computational analysis that led to these conclusions and describe the adaptation of the system, originally in English, to French and Spanish, with implications for PLN in other languages. The system is publicly available as a problem visibility tool (https://gendergaptracker.informedopinions.org/).

Enlaces relacionados