Douglas Guilbeault: "Online Gender Stereotypes are Stronger in Images than Text"

Hybrid: In Person, 402 Social Sciences Building & via Zoom

Advancements in natural language processing have spurred the proliferation of studies examining gender stereotypes in online texts, including news and social media. Yet, while these studies suggest a reduction of gender bias in recent years, research indicates that progress toward gender equality has slowed or stalled in vital areas of social life, from hiring practices to household management. Textual measures of online stereotypes are at risk of underestimating the gender gap, which may be more salient in online images that visualize the demographics of people. In this talk, I show that online gender stereotypes are more prevalent in images than texts using a novel dataset comprising over one million images from Google, Wikipedia, and IMDb, mapped to over 3,400 distinct social categories, including occupations (e.g., “doctor”) as well as generic social roles (e.g., “friend”) and lifestyles (e.g., “vegan”); stereotypes in these images are then compared to stereotypes measured by word embedding models trained on billions of words from online texts. To characterize the empirical consequences of this finding, I use an online experiment to show that googling for visual rather than textual descriptions of occupations amplifies people’s implicit bias toward associating men with science and women with liberal arts, a stereotype linked to pervasive inequalities in academia and industry. I conclude by showing how text and images can differ in the kind of stereotypes they encode; for example, I show that gendered ageism, whereby women are pressured to appear younger than men, is particularly pervasive in online images. Implications for algorithmic bias are discussed.


Douglas Guilbeault is an Assistant Professor in the Management of Organizations Group at the Haas School of Business. He studies how communication networks underlie the creation and diffusion of cultural content, such as linguistic categories and social norms. This investigation extends to how communication dynamics are shaped by various sources of influence, such as organizational culture, partisan identities, and algorithmic bias over social media. His work on these topics has appeared in a number of top journals, including Nature Communications, The Proceedings of the National Academy of the Sciences, and Management Science, as well as in popular news outlets, such as The Atlantic, Wired, and The Harvard Business Review. Guilbeault’s work has received top research awards from The International Conference on Computational Social Science, The Cognitive Science Society, and The International Communication Association. Guilbeault is a co-director of the Berkeley-Stanford Computational Culture Lab, and he is a faculty affiliate of the Berkeley Institute for Data Science. He received his Ph.D. from the University of Pennsylvania, where he trained in the Network Dynamics Group led by Damon Centola. He is a recent recipient of Stanford’s Art of Science Award for his visual piece, “Changing Views in Data Science over Fifty Years.”