|
Abstract:
|
The SDR follows a sample of U.S.doctorates in science, engineering, and health (SEH) fields to monitor the nation's education, supply, and employment of doctorate recipients in SEH fields. The SDR underwent a major sample redesign during the 2015 cycle. As a result, the vast majority of the 120,000 sample members are new to the SDR, making it a tremendous challenge to locate the new sample members. To meet this challenge, the SDR team created a comprehensive adaptive design to periodically identify low performing cases and adjust data collection strategies to ensure data quality. The following measures are computed on a weekly basis: contact and completion rate by key domains, partial R indicators associated with key variables, Mahalanobis distance between respondents and nonrespondents, and cumulative yield by key domains. Based on these statistics, data collection resources are redistributed in order to achieve a more balanced sample. Early evidence shows that the adaptive design system has been moderately effective in boosting response rates for traditionally low performing groups. This paper will report the key outcomes associated with the 2015 SDR adaptive design.
|
