Doctoral Candidate
About
Tyler Chen is a PhD candidate in Political Science and Scientific Computing at the University of Michigan, with a dual master's degree in Statistics. His research studies how political parties and their bases shape each other across very different electoral systems, with a methodological focus on settings where survey data is unreliable, contested, or unavailable.
Substantively, Tyler works across East and Southeast Asia and the United States, conducting research in Chinese, Malay, Indonesian, and Japanese. His interest in these cases stems from the diversity of their electoral systems, where institutional constraints shape the political consequences of population change in distinct ways. Current projects examine why dominant-party regimes fracture in some ways and not others (Malaysia, Japan, Taiwan), how ethnoracial categorization shifts under partisan pressure in U.S. voter files, and how party-state campaigns operate as instruments of bureaucratic coercion. His earlier work on ethnic-urban cleavages in Malaysia's electoral politics appears in Journal of East Asian Studies.
Methodologically, Tyler works in applied Bayesian statistics, ecological inference, and computational text analysis. His central methodological contribution is a hierarchical approximate Bayesian computation estimator that integrates survey microdata with aggregate election returns through a likelihood-free framework, evaluated against classical bounds, MCMC-based Ecological Inference, Multilevel Rrgression with Poststratification, and semiparametric sieve estimators in a large-scale simulation study under varying conditions of segregation, polarization, survey quality, and model misspecification. He has taught both undergraduate and graduate courses in statistical methods.
Tyler co-organizes the Rackham China Reading Group (2023– ) and is a council member of the LSA Graduate Student Advisory Council and the inter-university GETSEA consortium. He also volunteers as a cat foster to the Humane Society of Huron Valley.
