NSCG Sampling Issues When Using an ACS-Based Sampling Frame

• Activity Number: 453
• Type: Topic Contributed
• Date/Time: Wednesday, August 3, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Section on Survey Research Methods
• Author(s): David Hall*+ and Stephen Cohen and John Finamore and Flora Lan
• Companies: U.S. Census Bureau and National Science Foundation and National Science Foundation and National Science Foundation
• Keywords: NSCG ; ACS ; Sampling Frame

Abstract

The National Survey of College Graduates (NSCG) is a source of detailed statistics on the nation's science and engineering labor force. The Census Bureau conducts the NSCG on behalf of the National Science Foundation (NSF). Historically, the NSCG selected its sample once a decade from the decennial census long form respondents. In 2010, the Census discontinued the long form, so the NSF switched to using the American Community Survey (ACS) as a sampling frame for the NSCG. This switch brought both opportunities and challenges. The ongoing data collection methods of the ACS offered the potential for improved coverage of the NSCG, and the inclusion of a new field of degree question on the ACS allowed for a more efficient sampling effort. Meanwhile, the ACS also introduced increased weight variation and complexity. This paper provides insight into our efforts to address the challenges of an ACS-based sampling frame by discussing the 2010 NSCG sample design issues, including the determination of an appropriate sample selection technique, the incorporation of an iterative sample allocation approach, and the identification and resolution of obstacles encountered during sample selection.