The Center for Behavioral Science Methods (CBSM) leads the application of social science and survey methods to improve the design of questionnaires, data collection instruments, and information products. CBSM also leads general methodological research to improve data quality from surveys and censuses, develops new methods for design and evaluation, and applies these methods to emerging data products.

Staff use both qualitative and quantitative methods drawn from statistics, psychology, sociology, linguistics, and anthropology to identify and reduce measurement error in existing surveys and to develop questions for new topic areas. Typical research methods include one-on-one cognitive testing, focus groups, debriefings, expert review, behavior coding and split ballot field experiments. CBSM staff also adapt surveys for alternative modes of data collection (e.g., web, mobile devices, mail surveys, interviewer-administered modes), and conduct research to ensure comparability of measurement across languages and cultural groups. In addition, CBSM extends these methods to evaluate the clarity and effectiveness of information products disseminated through the web, apps, and print reports, and messages designed to inform the public about research and encourage participation. Most recent extensions include applying established methods to evaluate the quality of emerging data sources such as administrative records and paradata, and employing new methods such as machine learning to evaluate and supplement conventional survey data.

Staff regularly collaborate with counterparts in other federal government agencies, academia, and private research organizations, and exchange state-of-the-art findings from the literature and cross-cutting developments in the research community. Research findings are disseminated at professional meetings, seminars and workshops, and through working papers, journal publications and book chapters.