Education’s Big Data: Longitudinal and Cross-Sectional Survey Studies at the National Center for Education Statistics

Introduction: The National Center for Education Statistics (NCES) provides a central and trusted repository of public- or restricted-use datasets that are nationally representative, meaning findings can be generalized to learners across the United States. This paper will (1) highlight NCES data likely of interest to prevention scientists, (2) discuss possible data linkages between NCES data and outside data sources, such as the Civil Rights Data Collection or the American Community Survey, and (3) outline steps to obtain NCES data for researchers’ own analyses. Special attention will be paid to NCES’ K to 12 longitudinal studies: the Early Childhood Longitudinal Studies (ECLS), the Middle Grades Longitudinal Study of 2017-18 (MGLS); and selected studies from the suite of NCES high school longitudinal studies.

Prevention-Focused Data Availability: General background information will be provided on NCES data, highlighting selected constructs of interest to prevention scientists. NCES longitudinal studies consider multiple features of children’s or students’ experiences in multiple contexts, and from multiple reporters. For example, analysts of the ECLS data have conducted research on the antecedents of numerous social, behavioral, and physical health outcomes, while analysts using the high school datasets have identified individual-level factors that predict which educational and occupational pathways a student is likely to follow. In addition to this overview, we will also address common misunderstandings about what can and cannot be done with NCES data.

Data Linkages: Researchers may use NCES datasets to conduct several different types of “big data” linkages to facilitate complex analyses. Three such linkages are (1) cross-cohort linkages to compare individuals from different decades (as exemplified by the ECLS program studies), (2) cross-study linkages between the NCES longitudinal studies to facilitate longer-term longitudinal analyses than would be possible with any one individual dataset (as available through the synthetic K to 12 cohort from three of NCES’ longitudinal studies programs), and (3) administrative data linkages (such as to school- and neighborhood-level data collected by other NCES studies and the U.S. Census Bureau).

Data Access: The differences between NCES public- and restricted-use files, including how to access them and available support (online trainings and study documentation), will be explained.