*New* CSDE Computational Demography Working Group (CDWG) Hosts June Yang on Supervised Topic Modeling with GPT-Assisted Text Annotations: A Study of Ideas around Cohabitation (5/8/2024)

Posted: 5/2/2024 ()

On Wednesday 5/08 from 9:00 AM – 10:00 AM, CDWG will host June Yang to present her research. June Yang is a Ph.D. candidate at UW’s Department of Sociology, a CSDE T32 fellow in Demography and Data Science, and recently joined UW CSDE and eScience Institute as a research scientist. Her current research projects include 1) using social media text data to study attitudes around alternative forms of union formation and childbearing choices, and 2) using complex survey data to study the homeless population in King County, WA. The event will occur in 223 Raitt (the Demography Lab) and on Zoom (register here). Learn more about the talk in the full story.

Title: Supervised Topic Modeling with GPT-assisted Text Annotations: A Study of Ideas around Cohabitation

This study examines values and ideas around cohabitation using text data from the Chinese social media website Zhihu. Scholars have heavily cited the Second Demographic Transition theory to account for rising trends of cohabitation in different countries. The theory argues for the prominent role played by ideational changes such as individualism and the quest for self-actualization in leading to diffused patterns of union formation. China has witnessed a rising cohabitation but paradoxically, according to existing research, marriage is still largely seen as the only cultural ideal in the society. Using text data, this study aims to reveal the complexities of the choice of cohabitation in a patriarchal context. Methodologically, this study utilizes GPT-4 to generate a label from the text data that measures a liberal/non-liberal attitude about cohabitation and further adopts supervised topic modeling to generate latent topics from the text that predict the label.