Director Mathematical Statistics Research Center U.S. Bureau of Labor Statistics; BLS

BLS publishes gold-standard data on labor market activity, working conditions, price changes, and productivity in the U.S. economy to support public and private decision making. These data are used by the U.S. Congress and state legislators to make policy decisions, by businesses to make decisions about new site locations and wages, and by millions of citizens to make critical career and financial decisions.

The Office of Survey Methods Research within BLS conducts research to develop behavioral and statistical methods to improve BLS surveys. The work culture at BLS and in OSMR is one where our associates strongly believe in the importance of our vital mission to publish statistics of the highest quality for the American public, while maintaining a healthy work-life balance.

In this role as a Supervisory Research Mathematical Statistician, you will lead a team and partner with them to conduct research to devise novel statistical methodologies. You will develop and lead a research agenda in areas of mathematical statistics that advance the state of practice for the development or analysis of BLS data collected from survey and census instruments. You will present your findings at research conferences and play a leadership role by partnering with researchers and leaders in both government and academia.

You will serve as an expert methodologist who consults with BLS survey programs to develop and implement statistical methodologies that improve the quality of estimation and prediction for published statistics.

Our researchers possess expertise in survey design and the analysis of survey data, including non-parametric estimation methods, machine learning approaches and Bayesian hierarchical probability modeling. Our methods are used to conduct unbiased inference about an underlying population estimated on data acquired from a survey of that population and to develop accurate uncertainty quantification under dependence induced by the survey design used to collect the data.

We are exploring various approaches to encode formal privacy protection into data products released to the public. Our data are typically time- and spatially-indexed and our statisticians express expertise in accounting for these sources of dependence.