Statistical Innovation Strengthens Understanding of Science Career Trajectories

Two new working papers from the National Center for Science and Engineering Statistics (NCSES) within the National Science Foundation (NSF) demonstrate how innovative statistical methods can expand access to high-quality data while protecting confidentiality, ultimately strengthening research on the education and career trajectories of the U.S. science and engineering workforce. The papers, featuring contributions from Westat staff, are based on post-survey research using data from the Survey of Doctorate Recipients (SDR), a key source of information on the education, careers, and mobility of U.S.-trained doctoral scientists and engineers. The research draws on the 2015, 2017, and 2019 survey cycles.

Public-Use Longitudinal Synthetic Microdata

This working paper describes the development of publicly accessible longitudinal synthetic microdata to improve transparency and reproducibility while safeguarding respondent confidentiality. The experimental statistical products rely on partial data synthesis and are intended for research purposes rather than official statistics. The research and analytic uses of longitudinal SDR data were highlighted in recommendations in multiple Committee on National Statistics (CNSTAT) consensus reports. The authors note: “These requirements and recommendations include obtaining longitudinal survey data on the nature, determinants, and consequences of significant transitions in science and engineering career pathways.” The paper addresses the balancing of 2 main goals: maintaining data confidentiality and retaining data integrity.

Read the Paper

Developing Public-Use Longitudinal Synthetic Microdata With Applications to the Survey of Doctorate Recipients

Westat contributors include Minsun Riddles, PhD; Tom Krenzke, MS; Angela Chen, MS; Lin Li, MS; and Robyn Ferg, PhD.

Unified Approach to Cross-Sectional and Longitudinal Imputation

This working paper introduces a unified imputation approach that preserves reported longitudinal patterns while addressing missing data across survey waves. As the authors note, “Longitudinal surveys offer insights that cross-sectional surveys cannot, such as the ability to track outcomes over time and assess the potential impacts of policy changes.” The paper presents evaluation results for both the 2019 SDR cross-sectional data and the 2015–19 longitudinal file, demonstrating the method’s effectiveness in maintaining data quality for analytic use.

Read the Paper

A Unified Approach to Cross-Sectional and Longitudinal Imputation With Applications to the Survey of Doctorate Recipients

Westat contributors include Minsun Riddles, PhD; Shelley Brock Roth, MS; Laura Alvarez-Rojas, MA; and Véronique Lieber, MS.