Prediction in Social Science: A Tool to Study Inequality in Populations
Biography: Ian Lundberg is a Postdoctoral Scholar in the Department of Sociology and California Center for Population Research at UCLA. His research develops statistical and machine learning methods to answer new questions about inequality in America. Past work is published or forthcoming in PNAS, the American Sociological Review, Demography, the Journal of Policy Analysis and Management, Sociological Methodology, Sociological Methods and Research, and Socius. This academic year, Ian is working on an NSF-funded postdoctoral project developing computational methods to study income mobility. In 2022, he will begin as an Assistant Professor in the Department of Information Science at Cornell University. You can read more at ianlundberg.org.
Abstract: Predictive algorithms could transform methodology in social science, yet the mapping between prediction and scientific knowledge is not always clear. This talk will address three uses of prediction: (1) predicting outcomes for individual people, (2) predicting unobserved factual outcomes to describe populations, and (3) predicting counterfactual outcomes for causal claims. I will argue that prediction of individual-level outcomes is often difficult in social science, yet predictive algorithms which are imperfect for individuals (1) can nonetheless be useful in support of population-level claims (2 and 3). This framework for the use of prediction is well-suited to the integration of perspectives from social science (defining the population-level quantity to be estimated) and data science (building a predictive model to estimate that quantity).