Abstract:
|
Faced with declining response and increased risk for nonresponse bias, surveys have increasingly relied on post-survey adjustments to ensure the accuracy of estimates. In particular, researchers are exploring the potential of paradata, including interviewer observations in face-to-face surveys, to correct for nonresponse bias. In 2013, the National Health Interview Survey (NHIS) added 15 interviewer observations collected during the first personal visit attempt. Examples include the physical condition of the sample unit and evidence of smokers. Using 2013 data, we assess the utility of these observations for nonresponse adjustments. Of specific interest is whether adjustments based on these measures represent improvements over current procedures. Using a set of key survey estimates, we address the following questions: Are substantial shifts in estimates observed when the observational measures are used in the adjustments? What differences are observed in the variance of survey estimates by adjustment method? We discuss the implications of our results for NHIS weighting procedures and the utility of observational measures for post-survey adjustments more generally.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.