'Detecting latent events, such as regimes in the macroeconomy or collusive behavior in markets, is an intriguing problem. Professor Steigerwald's research team is developing and refining likelihood ratio tests, which depend on functions of Gaussian processes. They have analyzed the power of such tests with an eye toward constructing a new set of tests that offer greater power. Their team also works on the problem of determining the significance of attributes from data that is naturally grouped into clusters, such as children in a classroom or workers in a firm. The focus of this work is to understand how the degree of similarity across clusters affects the behavior of test statistics. In some of these settings, the data under study is close in size to the complete population. They are working to understand how to account for this both in assessing the significance of attributes and of causal variables, the later within the setting of heterogeneous treatment effects. Applied projects for their team include measuring the values of journal subscriptions and assessing the benefits of water recycling'.

See: Douglas Steigerwald | Department of Economics | UC Santa Barbara (ucsb.edu)