Robert P. (Bob) Sherman
Professor of Economics and Statistics; Executive Officer of the Social Sciences
Asymptotic Methods; Empirical Processes; Nonparametric and Semiparametric Estimation; Long-Range Dependence; Sampling Bias; Discrete Choice Modeling; Diffusion Modeling
Bob Sherman develops and analyzes methods to determine causal effects in discrete choice models that are suffering from various data deficiencies. In econometrics, for example, a lot of effort is put into modeling and estimating the relationship between a response variable, such as an employee's wages, and measured variables that influence the response, such as education or experience. Econometric models also include an error term that tries to capture how some unmeasured variables may affect the response. The error term is often assumed to have a known probability distribution, an assumption that's usually made for computational convenience and is seldom supported by economic theory. But this assumption is restrictive, and if it turns out to be wrong, then estimators of unknown parameters—such as those that quantify the effect of measured variables on the response—can be unreliable. Sherman's research involves developing semi-parametric estimators of parameters in these models. The estimators are reliable even when the distribution of the error term is unknown. He's especially interested in developing computationally attractive semi-parametric estimators and the large sample theory required to do asymptotic inference with these estimators.
In nearly every scientific discipline, data collected for making inferences about a population of interest can suffer from various deficiencies, rendering conclusions based on standard inference procedures invalid. For example, survey data in the social sciences are often biased due to various types of misreporting, nonresponse, or measurement error. Sherman is also developing ways to detect and correct biases that result from flawed data-generating mechanisms, and developing and analyzing estimation procedures that are valid when bias cannot be removed.
Before joining Caltech, Sherman was a member of the technical staff of Bellcore Statistics and Economics Research Group from 1990 to 1996. He received Caltech's Graduate Student Council (GSC) Teaching Award in 2006.