|
Activity Number:
|
255
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 1, 2016 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Survey Research Methods Section
|
|
Abstract #320636
|
View Presentation
|
|
Title:
|
Parametric Bootstrap Mean Square Error Estimates for Different Small Areas in the Annual Survey of Public Employment and Payroll
|
|
Author(s):
|
Bac Tran*
|
|
Companies:
|
U.S. Census Bureau
|
|
Keywords:
|
governmental units ;
empirical best prediction ;
small area estimation ;
parametric bootstrap
|
|
Abstract:
|
We illustrate a small area estimation methodology to estimate employment by combining the U.S. Census Bureau's Annual Survey of Public Employment and Payroll (ASPEP) with the previous census records using an empirical best prediction (EBP) methodology. The employment data are usually subject to skewness and heteroscedasticity and thus the well-known EBP methodology based on unit level linear mixed normal model does not fit well. In order to get around the problem, we apply a unit level linear mixed normal model on the log-transformed employment. In this paper we discuss the performance of the parametric bootstrap to estimate mean squared error for different small areas and compare its performance between the unit-level and area-level.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2016 program
|
