Small Area Estimation for the National Survey of College Graduates New Cohort

• Activity Number: 571
• Type: Contributed
• Date/Time: Wednesday, August 6, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Survey Research Methods Section
• Author(s): Michael Larsen*+ and Benjamin M. Reist
• Companies: George Washington University and U.S. Census Bureau
• Keywords: American Community Survey ; Borrowing strength ; Domain estimation ; Log-linear model ; Shrinkage ; Survey weight

Abstract

The National Survey of College Graduates (NSCG) has been conducted by the Census Bureau for the National Science Foundation (NSF) since the 1960s. It is the nation's only source of detailed statistics on the science and engineering (S&E) labor force. The NSCG uses a rotating panel design and selects its sample on a biennial basis from the American Community Survey to allow both cross-sectional and longitudinal analysis of education, employment, and demographic characteristics of the S&E labor force. Under this design, the NSCG data is collected and released on a biennial or triennial schedule. The 2010 survey cycle marked the initial use of the ACS as a sampling frame for the NSCG. We examine the feasibility and desirability of using small area estimation methods for estimating characteristics of the S&E labor force in small domains. Small domains are defined crossing levels of two or more variables in the primary analysis domains.