|
Activity Number:
|
293
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #310317 |
|
Title:
|
Evaluating Response Quality of Nonrespondents Using the Imputation Technique
|
|
Author(s):
|
Cong Ye*+ and Ting Yan and Mandi Yu
|
|
Companies:
|
University of Maryland and University of Michigan and University of Michigan
|
|
Address:
|
JPSM, 1218 Lefrak Hall, College Park, MD, 20742,
|
|
Keywords:
|
Measurement Error ; Missing Data ; Imputation ; Response Propensity
|
|
Abstract:
|
Declining response rates in household surveys speak to the increasing difficulty in recruiting sampled persons to respond to a survey request. Survey organizations are spending increasingly more effort on contacting and persuading respondents with a low response propensity in an attempt to increase response rates and reduce nonresponse bias. A major concern, however, is that, if respondents with a low response propensity turned out to be bad reporters, the costly extra recruitment effort could reduce nonresponse error at the expense of increasing measurement error. However, this concern is difficult to address since the measurement error property of nonrespondents is unknown. This paper treats this issue as a missing data problem and imputes for nonrespondents hard to recruit. The preliminary results showed that respondents who are hard to recruit didn't provide worse data.
|
