Degree Type

Creative Component

Semester of Graduation

Fall 2019

Department

Statistics

First Major Professor

Emily Berg

Degree(s)

Master of Science (MS)

Major(s)

Statistics

Abstract

Regression estimation is a way to use auxiliary information to improve the efficiency of survey estimators. Regression estimation generates a weight such that the weighted sum of auxiliary variables for sampled elements is equal to known population totals. One can view the weight defining the regression estimator as the solution to an optimization problem. The regression weight minimizes the distance to the inverse inclusion probability subject to the calibration restriction. The calibration restriction forces the weighted sum of the auxiliary variable to equal the known population total. Two issues were analyzed. The first is the question of how to choose which covariates to include as control totals. The second is the choice of the metric defining the distance between the regression weights and the inverse inclusion probabilities. These two issues were analyzed through simulation and an application to data from the National Resources Inventory.

Copyright Owner

Agadilov Gani

File Format

PDF

Share

COinS