Online Program Home
My Program

Abstract Details

Activity Number: 313 - Statistical Models in Survey Sampling and Analysis
Type: Contributed
Date/Time: Tuesday, July 31, 2018 : 8:30 AM to 10:20 AM
Sponsor: Survey Research Methods Section
Abstract #329531 Presentation
Title: Applications of the Parametric Approach to Estimation of Totals and Means for Complex Survey Data in the Presence of Full Response
Author(s): Ismael Flores Cervantes*
Companies: Westat
Keywords: Efficient estimation of totals; model-assisted estimation; design-based estimation; complex survey sample data; calibration; variable selection
Abstract:

The Parametric Approach (PA) to survey sampling is a recently developed framework for the study of estimation problems in survey sampling with or without nonresponse. This approach provides tools for the development of efficient and unbiased estimators. The PA methodology uses an explicit working model for the finite population and estimates the parameters of the distribution while taking into account the sample design. It incorporates most of the classical statistics modeling techniques for improving the efficiency or reducing nonresponse bias of survey estimates. The functional of a PA estimator is determined through statistical tests of parameters or model goodness of fit for the observed sample. Three examples that show the strengths of the PA for estimation in the presence of full response are discussed. In the first example, the PA algorithm determines the functional form of the estimators based on the observed data. In the second, the PA is used as a method for variable selection for control totals of calibrated estimators. The last example shows how the PA can automatically produce new forms of estimators or PA estimators based on the available auxiliary variables.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2018 program