A statistical analysis of gender representation in Swedish political reporting

Sammanfattning: This thesis examines gender representation in Swedish political media using text data from Dagens Nyheter (DN), Sweden's largest daily newspaper. Descriptive statistics and regression analysis provide evidence that there are discrepancies in the amount of coverage that men and women receive; both in terms of number of articles, and article length. The existence of gendered language is studied using topic analysis and lasso-logistic regression. The results from the topic analysis show that professional titles appear more frequently in articles about men, and words related to family appear more frequently in articles about women. However, the evidence presented here does not suggest that articles about women in party leadership positions are shorter than articles about their male counterparts.