|
Activity Number:
|
185
|
|
Type:
|
Roundtables
|
|
Date/Time:
|
Monday, August 3, 2009 : 12:30 PM to 1:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #303256 |
|
Title:
|
Fitting Models and Estimating Model Parameters Using Data from Complex Surveys
|
|
Author(s):
|
Jean Opsomer*+ and Jay Breidt*+
|
|
Companies:
|
Colorado State University and Colorado State University
|
|
Address:
|
Statistics Department, Fort Collins, CO, 80523-1877, 102 Statistics, Fort Collins, CO, 80523-1877,
|
|
Keywords:
|
design-based estimation ; analytic inference ; model-based estimation ; variance estimation
|
|
Abstract:
|
Analysts in many disciplines often fit models on survey data, which are provided to them by the data collection organization as a weighted data set. The data set might also be accompanied by replicate weights or generalized variance functions, to be used in statistical inference. A major issue facing analysts is how to properly account for the sampling aspects of the data in their model estimation and inference. A number of different approaches are described in the statistical literature, but it is fair to say that none are currently fully accepted outside of the survey statistics community. We review the major approaches and discuss ways in which survey-appropriate estimation and inference can be better integrated into common statistical practice.
|
