|
Activity Number:
|
70
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, July 29, 2007 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #308837 |
|
Title:
|
Small-Area Estimation for a Survey of High School Students in Iowa
|
|
Author(s):
|
Lu Lu*+ and Michael D. Larsen
|
|
Companies:
|
Iowa State University and Iowa State University
|
|
Address:
|
CSSM, 208A Snedecor, Ames, IA, 50011,
|
|
Keywords:
|
Hierarchical Bayesian Analysis ; Two-Way Stratification ; Sparse Design ; Auxiliary Information ; Generalized Linear Models
|
|
Abstract:
|
Iowa's State Board of Education conducted a stratified multi-stage sample survey to study the availability of employment preparation courses and the degree to which students in Iowa's public high schools enroll in those courses. Given the budget restrictions, the design takes a small sample of PSUs from a two-way stratification of districts, which causes high variability in direct survey estimates. A hierarchical Bayesian (HB) analysis, which borrows strength across strata with similar characteristics, is suggested to increase stability. When the number of strata increases such that the design can not assure that sample is allocated to every stratum, a design which is balanced in terms of both stratification factors is necessary and a HB predictive approach is highly recommended. The involvement of relevant auxiliary variables also should improve the estimation.
|
