|
Activity Number:
|
36
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, July 29, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #308996 |
|
Title:
|
Selection of Small and Large Schools in State and County School Surveys
|
|
Author(s):
|
Tonja M. Kyle*+ and Pedro J. Saavedra and James G. Ross
|
|
Companies:
|
Macro International and Macro International and Macro International
|
|
Address:
|
11785 Beltsville Drive, Calverton, MD, 20705,
|
|
Keywords:
|
PPS sampling ; minimum replacement ; stratification ; design effect ; cluster sample ; intra-class correlation
|
|
Abstract:
|
The most common method when conducting state or county school surveys is to first select schools with probabilities proportional to size using enrollment as the measure of size, and then sample the same number of students from every school. Ideally, this would result in the same probability of selection for every student in the frame. One difficulty is that in both state and county surveys there are usually a number of schools with enrollments below the number of students targeted for each participating school. A second difficulty, primarily present in county surveys, is that there are often schools where the calculated probability of selection exceeds 1.0. This paper explores different ways of handling the above problems and their impact on design effect using simulations from publicly available school data. Variables with low and high intra-class correlations are used.
|
