text-only page produced automatically by LIFT Text
Transcoder Skip all navigation and go to page contentSkip top navigation and go to directorate navigationSkip top navigation and go to page navigation
National Science Foundation HomeNational Science Foundation - Directorate for Social, Behavioral & Economic Sciences (SBE)
Social, Behavioral & Economic Sciences
design element
SBE Home
About SBE
Funding Opportunities
Awards
News
Events
Discoveries
Publications
Advisory Committee
Career Opportunities
See Additional SBE Resources
View SBE Staff
SBE Organizations
SBE Office of Multidisciplinary Activities (SMA )
National Center for Science and Engineering Statistics (NCSE)
Division of Behavioral and Cognitive Sciences (BCS )
Division of Social and Economic Sciences (SES )
Proposals and Awards
Proposal and Award Policies and Procedures Guide
  Introduction
Proposal Preparation and Submission
bullet Grant Proposal Guide
  bullet Grants.gov Application Guide
Award and Administration
bullet Award and Administration Guide
Award Conditions
Other Types of Proposals
Merit Review
NSF Outreach
Policy Office Website
Additional SBE Resources
Exploring What Makes Us Human
Rebuilding the Mosaic Report
Bringing People Into Focus: How Social, Behavioral & Economic Research Addresses National Challenges
"Youth Violence: What We Need to Know" Report to NSF
Social, Behavioral and Economic Research in the Federal Context Report
Expedited Review of Social and Behavioral Research Activities Report
SBE Advisory Committee Web Site (for members only)


SBE 2020: Submission Detail

ID Number: 119
Title: irty Jobs: Fitting a Statistical Model to a Large Dataset with a Large Amount of Missing Data
Lead Author: Cardinale, John A.
Abstract: In this paper, we address the question of how to extract meaningful information from data sets that feature many missing values. Many researchers would argue that data such as these cannot be analyzed. However, here we outline the analysis of a two-level data set with missing values and provide preliminary results to make the case that current technology makes such research endeavors possible. Our data combines a survey of state vocational rehabilitation agencies, carried out by the Government Accountability Office (GAO), with data on individuals who are served by these agencies. The GAO data had a rate of missingness of over 30%. The individual level data, in contrast was more than 95% complete. Modern methods of missing data analysis, specifically multiple imputation, were applied to the combined two-level dataset, to compensate for the high degree of non-response in the GAO survey. We propose that this modern missing data method, while requiring rigorous data cleaning , allows us to gain useful insights. In particular, we found a significant supply-side main effect as well as cross-level interactions between agency capabilities and human capital variables.
PDF: Cardinale_John_119.pdf

SBE 2020 Home

 

Print this page
Back to Top of page