Hello! I am Gechun Lin. As a PhD student in political science, I am interested in applying computational methods to analyze unstructured qualitative data in political communication for the purpose of studying the interactions among media, political actors, and public perceptions.
My main work focuses on automatic analysis of short texts, such as social media posts and campaign image captions. For instance, in a specific study about similarity measurement for short political texts, I used a deep-learning model GAN-BERT, which takes advantage of contextual representations of texts produced by the language model BERT and is trained further with a GAN framework, to extract the semantic similarity of paired news headlines about same US Supreme Court case decisions. The findings suggest that contentious decisions are associated with less similar media portrayals. Such findings are meaningful. It is commonly known that people are trapped in echo chambers––our media consumption mirrors our political leanings. Different media depictions of court decisions can create contradictory understandings of the Court’s rulings among ideologically diverse readers, which may negatively impact the Court’s ability to build popular consensus on the most controversial policies in a polarized American society.
My education background also encourages me to pursue interdisciplinary studies in law and politics. Before joining the current program, I earned B.A. in Law from Sun Yat-sen University and LL.M and J.S.D. in Law from Washington University in St. Louis. I have commenced several side projects (theoretical and empirical), such as game theory modeling of economic or financial policy-making, quantitative analysis of authoritarian courts, etc.