Activity Number:
|
75
|
Type:
|
Contributed
|
Date/Time:
|
Monday, August 12, 2002 : 8:30 AM to 10:20 AM
|
Sponsor:
|
Section on Survey Research Methods*
|
Abstract - #300948 |
Title:
|
Sample Allocation and Selection for the National Compensation Survey
|
Author(s):
|
Christopher Guciardo*+ and Lawrence Ernst+ and Chester Ponikowski and Jason Tehonica
|
Affiliation(s):
|
U.S. Bureau of Labor Statistics and U.S. Bureau of Labor Statistics and U.S. Bureau of Labor Statistics and U.S. Bureau of Labor Statistics
|
Address:
|
2 Massachusetts NE, Room 3160, Washington, District of Columbia, 20212, 2 Massachusetts Ave., NE, Room 3160, Washington , District of Columbia, 20212-0001,
|
Keywords:
|
Sample allocation ; NCS ; ECI ; EBS
|
Abstract:
|
Recently, the first of five rotating replacement sample panels was selected for the integrated National Compensation Survey (NCS) Program, which includes the Employment Cost Index (ECI), Employee Benefits Survey, and locality wage surveys. This paper explains, for these surveys, how the sample of establishments was allocated to and selected from industry strata and geographic PSUs, and discusses several interesting issues. For these surveys, we describe how the universe of establishments was divided into two groups: one consisting of larger units that will be in the sample for all five sample panels, and the other consisting of units to be selected independently for each panel. Also, we describe how we forced the ECI sample to be a subsample of the parent NCS sample, despite sampling requirements that would not seem to guarantee this outcome. Finally, we describe how we selected the sample from allocation cells that contain clusters rather than single PSUs, a clustering that was necessitated by the fact that for some industry strata the number of sample units allocated to a cluster is smaller than the number of PSUs in the cluster. Several other allocation issues are also discussed.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2002 program |