|
SBE 2020: Submission Detail
| ID Number: |
296 |
| Title: |
Total Survey Error, Data Quality, and Statistical Error: Recommendations to the National Science Foundations Social, Behavioral, and Economic Sciences Directorate for 2020 Planning |
| Lead Author: |
Jans, Matthew E |
| Abstract: |
Social, behavioral, and economic research funded by NSF SBE is often interdisciplinary in nature (i.e., using multiple research methodologies). All research methodologies have strengths and weaknesses; one way to express these is through statistical error (e.g., sampling error, nonresponse error, measurement error). The Total Survey Error perspective on measurement and statistical estimation (Groves et al., 2009) accounts for most potential error sources in statistical estimates. We propose that this framework - although it provides a general road map to measurement and reporting of statistical findings - needs to incorporate other statistical and psychometric fields. This white paper briefly addresses the primary strengths and weaknesses of each approach to error and offers suggestions for evaluating a proposals approach to error. We recommend that NSF-SBE strengthen statistical error research by 1) requiring more documentation of and research into data quality and statistical error on substantive proposals, and 2) prioritizing data quality and statistical error research as a fundable research aim itself. Such research is essential because it builds on the basic components of statistical inference used by quantitative researchers, encourages stronger ties between substantive and methodological research, and provides a link among different facets of methodological and statistical research. |
| PDF: |
Jans_Matthew_296.pdf |
SBE 2020 Home
|
