|
Activity Number:
|
414
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Wednesday, August 1, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #309054 |
|
Title:
|
Using Interviewer Observations To Improve Nonresponse Adjustments: NES 2004
|
|
Author(s):
|
Andy Peytchev*+ and Kristen Olson
|
|
Companies:
|
RTI International and University of Michigan
|
|
Address:
|
3040 Cornwallis Road, Research Triangle Park, NC, 27709-2194,
|
|
Keywords:
|
nonresponse ; weighting adjustment ; propensity scores
|
|
Abstract:
|
One widely used method for adjusting for the effects of unit nonresponse is weighting. Propensity models and weighting class adjustments commonly employ available demographic characteristics. However, interviewers can make observations about the sample member, the housing unit, and the doorstep interaction that could be related not just to the propensity to respond, but also to statistics of interest. Using data from the 2004 American National Election Study we demonstrate whether the use of such interviewer-collected auxiliary data is useful for weighting adjustment beyond the current poststratification weights. In addition we demonstrate how adjustments that use interviewer observations affect only some weighted estimates. We conclude with a discussion of the need to collect auxiliary variables that are associated with key statistics of interest, not only response propensity.
|
