Figure 4

Topical bias. Word clouds for biographies of women (top) and men (bottom), with birth date before 1900 (left) and since 1900 (right). Spaces in bi-grams are replaced with an underscore. Font size is proportional to PMI with each gender. Colors depict the four categories: gender in orange, family in green, relationship in violet, and other in blue. Beside professional and topical areas, words in the gender, relationship, and family categories are more dominant in articles about women born before 1900. Gender-specific differences are much less pronounced in articles about people born since 1900.