|
Activity Number:
|
135
- Nonresponse Adjustment and Weighting
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 30, 2018 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Survey Research Methods Section
|
|
Abstract #329101
|
|
|
Title:
|
Evaluation of Nonresponse Adjustment Options on the National Health and Nutrition Examination Survey
|
|
Author(s):
|
William Cecere* and Minsun Riddles and Te-Ching Chen
|
|
Companies:
|
Westat and Westat and National Center for Health Statistics
|
|
Keywords:
|
survey weights;
logistic regression;
nonresponse adjustment;
classification trees
|
|
Abstract:
|
A critical component of the survey weighting process is how best to account for unit-level nonresponse. The critical question is how to adjust the survey weights to effectively reduce nonresponse bias while minimizing weight variance. For this research, we use the National Health And Nutrition Examination Survey (NHANES) survey respondents as the population and generate several nonresponse mechanisms that favor each method of nonresponse adjustment in turn. The generated data are then used to compare methods including logistic regression and classification tree options across a variety of software through simulation. We extend the results of Lohr, Hsu, and Montaquila (2015) to test a new tree-building method that applies an objective function based on reducing the variance of selected outcomes.
|
Authors who are presenting talks have a * after their name.
