|
Activity Number:
|
64
- Nonparametric Modeling of Survey Data
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Sunday, July 30, 2017 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Survey Research Methods Section
|
|
Abstract #323003
|
View Presentation
|
|
Title:
|
Model-Assisted Estimation with Regression Trees
|
|
Author(s):
|
Kelly McConville* and Daniell Toth
|
|
Companies:
|
Swarthmore College and US Bureau of Labor Statistics
|
|
Keywords:
|
model-assisted estimation ;
regression trees ;
complex surveys
|
|
Abstract:
|
Auxiliary information can increase the efficiency of survey estimators when the estimator accurately captures the relationship between the variable of interest and the auxiliary variables. Under a model-assisted framework, we present a regression tree estimator for a finite population mean. Regression trees can capture important interactive effects missed by linear regression and do not suffer from multicollinearity issues when the auxiliary variables are highly collinear. We establish consistency of the model-assisted regression tree estimator and compare its performance to other survey estimators using the US Bureau of Labor Statistics Job Openings and Labor Turnover Survey.
|
Authors who are presenting talks have a * after their name.
