|
Activity Number:
|
267
- Using Multiple Sources of Data to Assess and Improve Data Quality
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2020 : 1:00 PM to 2:50 PM
|
|
Sponsor:
|
Government Statistics Section
|
|
Abstract #313384
|
|
|
Title:
|
Using External Sources for Evaluating and Mitigating Nonresponse Bias and Sampling Variability
|
|
Author(s):
|
Tala Fakhouri* and Crescent Martin and Te-Ching Chen and Jay Clark and Minsun Riddles and Joseph Afful and Leyla Mohadjer
|
|
Companies:
|
National Center for Health Statistics and National Center for Health Statistics and CDC/NCHS and Westat and Westat and Peraton and Westat
|
|
Keywords:
|
Total survey error;
sampling deviation;
nonresponse bias;
NHANES
|
|
Abstract:
|
Total survey error (TSE) can be divided into two broad categories: errors of representation and errors of measurement. The objective of this analysis is to investigate two contributing factors of TSE’s errors of representation, namely, sampling error and unit nonresponse error on the 2017?2018 National Health and Nutrition Examination Survey (NHANES). Sampling error was examined using two approaches: (1) the socioeconomic and health characteristics of the 2017?2018 NHANES sampled counties were compared to counties selected in prior cycles using data from the Census Bureau and the Behavioral Risk Factor Surveillance System; and (2) the characteristics of each selected primary sampling unit (PSU) in 2017?2018 was compared to that of all PSUs in the stratum from which the selected PSU was sampled. Unit nonresponse error is composed of bias and variance. For this analysis, we focused on unit nonresponse bias using the Groves and Brick typology: (1) benchmarking to data from the National Health Interview Survey; (2) comparisons to external data from the American Community Survey; (3) studying variation within the respondent set; and (4) comparing alternative post survey adjustments.
|
Authors who are presenting talks have a * after their name.
