|
Activity Number:
|
37
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 6, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #305689 |
|
Title:
|
A Bridge between the Greg and the Linear Regression Estimators
|
|
Author(s):
|
Sarjinder Singh*+ and Raghunath Arnab
|
|
Companies:
|
St. Cloud State University and University of Botswana
|
|
Address:
|
Department of Statistics, St Cloud, MN, 56301,
|
|
Keywords:
|
Greg ; linear regression estimator ; estimation of total and variance
|
|
Abstract:
|
We discovered there is a choice of weights that builds a bridge between the GREG proposed by Deville and Sarndal (1992) and the linear regression estimator due to Hansen, Hurwitz, and Madow (1953) while using one auxiliary variable. It gives the same result as given in Singh (2003, 2004) and Stearns and Singh (2005) for unequal probability sampling by using two calibration constraints in the presence of one auxiliary variable. Thus, these approaches can be considered as alternative to each other while considering use of one auxiliary variable. The bridge developed here reconfirms that the sum of the calibrated weights should be equal to the sum of the design weights in a given sample. The approach by Singh (2003, 2004) and Stearns and Singh (2005) seems simple while using multiauxiliary information.
|
