|
Activity Number:
|
464
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #307391 |
|
Title:
|
Using Telephone-Exchange Data To Adjust for Nonresponse: Application in an Establishment Survey
|
|
Author(s):
|
Stephen R. Williams*+ and Ronghua Lu
|
|
Companies:
|
Mathematica Policy Research, Inc.
|
|
Address:
|
600 Alexander Park, Princeton, NJ, 08540,
|
|
Keywords:
|
nonresponse ; analysis weights ; propensity modeling ; establishment telephone surveys
|
|
Abstract:
|
We used logistic regression models to predict probabilities that a unit will be located and, if located, will respond. The inverse of the response propensity can be used to adjust the basic sampling weights. We have applied this method for computing weights in a range of settings, some using telephone-exchange data for household surveys, but know of no application to establishment surveys. This paper describes the application in a recent national survey of educational institutions. The sampling frame for this telephone survey was constructed from several list sources. As expected from such a frame, the survey staff was unable to contact and establish eligibility for a relatively large portion of the sample. With little auxiliary data about both the screened and unscreened sample units this survey offered an important application of propensity modeling using telephone-exchange level data.
|
