Haoran Shi
Doctoral researcher · London School of Economics
I am a doctoral researcher in the Department of Psychological and Behavioural Science at the London School of Economics, where I am completing my PhD. My research asks what sustains people's commitment to political parties, when that commitment shifts, and how it shows in public talk. Across four projects I combine a scoping review, machine learning on British panel surveys, and large-scale text measurement of US news and Reddit discussion, with language-model coding checked against human coders. I have taught social psychology, experimental design and statistics at LSE. This site collects my papers, code, data and other finished work.
Updated October 2026
Selected works
04 ENTRIESBranding at the water’s edge: Perceived elite polarization of Ukraine aid, 2022–2024
Measures how far apart Republican and Democratic positions on Ukraine aid appeared in US news coverage, 2022–2024, using sentence-level language-model coding validated against human coders.
Commenting at the water’s edge: Party brands and expressed partisanship on Ukraine aid, 2022–2024
Examines whether months in which the parties' Ukraine-aid positions were more clearly differentiated in the news coincided with more recognizably partisan comments on r/politics. Observational design; it measures comments, not individual users.
Predictors of partisan strength and in-party affect: a scoping review
A scoping review of what predicts how strongly people identify with a party and how warmly they feel toward it.
Stable states, dynamic changes: Machine learning reveals how identity anchors and issues drive British partisan dynamics
Explainable machine-learning models on British Election Study panels from the Brexit and COVID-19 periods, separating what anchors levels of partisan strength from what predicts changes in it.