|
Activity Number:
|
347
|
|
Type:
|
Invited
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Memorial
|
| Abstract - #303152 |
|
Title:
|
Horvitz-Thompson Estimation and Unit Nonresponse
|
|
Author(s):
|
William Kalsbeek*+ and Robert Agans and Abigail Panter
|
|
Companies:
|
The University of North Carolina at Chapel Hill and The University of North Carolina at Chapel Hill and The University of North Carolina at Chapel Hill
|
|
Address:
|
730 MLK Jr. Blvd, Suite 107, Chapel Hill, NC, 27599-2400,
|
|
Keywords:
|
Horvitz-Thompson Estimation ; Unit Nonresponse
|
|
Abstract:
|
The underlying principle of Horvitz-Thompson (HT) estimation is that estimates of totals are weighted at the unit level by the inverse of the probability that the unit appears as a respondent in the sample. Without nonresponse, this probability is the sampling probability. With nonresponse, this probability is the product of the probabilities of selection and response. Thus, there is an important theoretical connection between the origins of estimation from probability samples. This paper traces the early developments of HT estimation ignoring nonresponse, reviews how nonresponse affects weighted estimate, considers the importance of total survey design, and summarizes more recent work on efforts to measure response propensities to produce HT estimates with reduced bias due to nonresponse.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2009 program |
