On Area Specific Uncertainty Measures in Small Area Estimation

University Of Maryland, College Park, College Park MD

Investigators

Abstract

This research project will investigate new ways to evaluate small area estimates. There is a growing demand for statistics related to small geographic areas. These small area estimates are used by policymakers to assess the well-being of a nation, develop public policies, and allocate funds in various government programs. Users of these estimates generally are interested in understanding the uncertainty that is relevant to their specific small area as opposed to average uncertainty measures. However, most of the uncertainty measures proposed in the literature for small area estimates are not area specific in the study variable. This project will develop new uncertainty measures and examine their effectiveness in measuring area specific uncertainty in the study variable. A focus of the project will be on the mentoring of students and early-career researchers. The new research results will be incorporated into both a graduate course and a short course on small area estimation. Software will be developed to implement the new techniques and will be made publicly available. This research project will examine the versatility of the adjusted maximum likelihood method, already found useful in certain small area problems, in obtaining reliable uncertainty measures of the estimates that are small area specific in the study variable. The investigator will conduct a design-based simulation and real data analysis in order to understand the robust properties of the proposed uncertainty measures. Data for evaluating the new methodology will be drawn from census and household survey data. The project is supported by the Methodology, Measurement, and Statistics Program and a consortium of federal statistical agencies as part of a joint activity to support research on survey and statistical methodology.