Overview

Survey Methods Research

The Survey Methods Research staff evaluates the effectiveness and soundness of the survey methods currently used by BLS programs, investigates alternative methods to determine their appropriateness for BLS programs, and develops new methods for improving the efficiency and quality of BLS programs. The SMR staff also conducts research on cross-program issues, consults with program offices on an ongoing basis, and supports ongoing improvement activities for the major statistical programs.

The program is organized into two parts: the Behavioral Science Research Center and the Mathematical Statistics Research Center. Research conducted by the behavioral scientists concentrates on the measurement and reduction of nonsampling error through, for example, questionnaire design studies, investigations into respondent-interviewer interactions, usability studies of computer assisted data collection systems, and the development of response-level data quality measures. Research conducted by the mathematical statisticians concentrates on the measurement and reduction of sampling and nonsampling error through, for example, the development of more efficient sample design, evaluations of alternative estimators, modeling time series for the purpose of seasonal adjustment, model-based estimation, and investigating alternative methods for handling missing data. In support of the BLS Internet site, dissemination-related research activities conducted by this staff address the areas of human-computer interaction, information seeking and retrieval, knowledge management, and data that describe other data (i.e., metadata).

The SMR Staff serves as a resource for unique problems in mathematical statistics, usability research and nonsampling error.

Last Modified Date: January 06, 2006