This talk presents research that used machine learning and natural language processing to analyse the content of newspaper articles in order to detect gender bias in the coverage of politicians. The content of newspaper articles was analysed to highlight differences in the coverage of male and female politicians. These differences were then examined to identify evidence of gender bias.
The corpus analysed in this research contained Irish newspaper articles featuring male and female politicians over a 15-year period along with coverage of candidates in the 2011 Irish presidential election. The corpus was processed using various natural language processing techniques. Machine learning algorithms were then used to identify differences in the coverage of male and female politicians in text classification tasks and the context in which some of these differences occurred were analysed for evidence of gender bias.
This research uncovered evidence of gender bias including stereotypical portrayals of politicians and differences in how male and female politicians were associated with policy issues. It also presents an approach to automatically detecting patterns in text that facilitates large-scale analyses of corpora.
Susan Leavy is an adjunct assistant professor in the School of Computer Science and Statistics at Trinity College Dublin. She just completed her PhD, which used techniques from artificial intelligence including machine learning and natural language processing to analyse the representation of female politicians in the media. She holds an MSc in Artificial Intelligence from Edinburgh University, an MPhil in Gender and Women's Studies from Trinity College Dublin and has a decade of international experience in technology in investment banking.
This is a joint meeting of the LIP, RiGLS and UCREL research groups.
Contact: Veronika Koller, v.koller@lancaster.ac.uk
Extra session co-organised with RiGLS/LIP