Associate Professor
Alexander Volfovsky's research concentrates on developing theory and methodological tools for computational social science applications, with particular interests in causal inference, high dimensional data, and network analysis. He is interested in assessing fundamental assumptions such as exchangeability and stochastic equivalence that underly many network models and to this end has developed testing and estimation procedures for complex dependence structures among actors in a network. He has recently been working on tools for causal inference and missing data problems in complex observational studies. Volfovsky has two distinct but complimentary threads in this direction: (1) nearly exact matching in high dimensional covariate spaces and (2) settings where networks lead to a breakdown of traditional approaches.