This year Yu Xie, Professor of Sociology and Statistics at the University of Michigan, presented the lectures. The lunchtime presentation was entitled 'Population Heterogeneity and Causal Inference.'
Abstract: Yu Xie argues that the very objective of social science is not to discover abstract and universal laws but to understand population variability. He calls this "Variability Principle." Causal inference with observational data in social science is impossible without strong assumptions. There are two potential sources of bias. The first is bias in unobserved pretreatment factors affecting the outcome additively. The second is bias due to heterogeneity in treatment effects. The first potential source of bias is usually handled with either collection of new data or unique design features (such as fixed effects models). Our understanding of the second source of bias is so far inadequate. In this presentation, Yu Xie discusses a simple scenario of "composition bias," which is a form of selection bias, under the classic assumption of ignorability. Both simulation and empirical examples are given.
The Clifford C. Clogg Memorial Lecture entitled, 'The Three Principles in Social Science,' was presented in the afternoon.
Abstract: In this presentation, Yu Xie argues that the very objective of social science is not to discover abstract and universal laws but to understand individual and contextual variabilities in populations. Thus, a fruitful quantitative paradigm for social science should not imitate physical science to be "exact." Rather, it should borrow Darwin’s evolutionary biology and call for empirical research that yields valid and informed understanding of actual social phenomena, social processes, and human behaviors. Three basic principles for social science research are proposed. First, variability is the very essence of social science. Second, social grouping reduces such variability. Third, patterns of population variability may vary with social context, which is often defined by time and space. For illustration, Yu Xie discusses implications of regression analysis with survey data—the predominant mode of quantitative inquiry in social science.
The lecture was followed by a reception in the Willaman Gateway (Life Sciences Building).