Proceedings of the Annual Meeting of the Cognitive Science Society

For many categories of people, men are perceived as the more default or typical members whereas women are perceived as more atypical. This bias can lead to an asymmetry in the existence and frequency of categories marked by gendered language. Here we explore the extent to which this asymmetry exists in two institutional category systems: the Library of Congress Subject Headings (LCSH) and English Wikipedia. We find that the LCSH exhibits more bias towards women than Wikipedia, and that in the LCSH this bias has not changed in the last 30 years, whereas Wikipedia shows a noticeable increase in gender balanced categories during the early 2010s. These findings suggest that more can be done to reduce gender bias in the LCSH and demonstrate how principles of typicality and categorization play out in real-world settings.