|
Activity Number:
|
16
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 11, 2002 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods*
|
| Abstract - #301817 |
|
Title:
|
Weighting Adjustments for Unit Nonresponse
|
|
Author(s):
|
Sonya Vartivarian*+ and Roderick Little
|
|
Affiliation(s):
|
University of Michigan and University of Michigan
|
|
Address:
|
1875 Lindsay Lane, Ann Arbor, Michigan, 48104, USA
|
|
Keywords:
|
sampling weights ; survey inference ; unit nonresponse adjustment
|
|
Abstract:
|
Weighting is a common form of unit nonresponse adjustment in sample surveys subject to unit nonresponse. Weighting adjustments involve weights that are inversely proportional to the probability of selection and response. A common approach computes the response weight in an adjustment cell as the inverse of the ratio of the sum of the sampling weights of respondents in a cell divided by the sum of the sampling weights of respondents and nonrespondents in that cell. In Little and Vartivarian (2001), we show that a better approach is to use weighting adjustments that involve an unweighted response rate where the model for nonresponse includes both adjustment cell and survey design variables. When the number of cells thus created is too large, a coarsening method such as response propensity stratification can be applied to reduce the number of adjustment cells. In this paper, we consider the efficiency and robustness of weighting adjustments based on the joint classification of the sample by two key potential stratifiers: the response propensity and the predictive mean based on a regression of the survey outcomes on covariates.
|
