Sampling US Schools - Frame and Design Issues
Krotki, K., & Bland, C. R. (2005, May). Sampling US Schools - Frame and Design Issues. Presented at American Association for Public Opinion Research 60th Annual Conference, Miami Beach, FL.
This research addresses various issues related to the building of a sample frame of schools in the US and then using this frame for sample selection. The frame topics include coverage, accuracy of data, and availability of supplementary information. The frame is based on the combination of two data sources: the Common Core of Data (CCD) and the Private School Survey (PSS). The resulting frame is remarkably versatile and rich allowing the sampler considerable leeway in creating a design that suits the project's budget and analysis objectives.
The sampling topics include stratification, multistage sampling, oversampling, and the specific challenge of simultaneously sampling two or more independent samples from the same frame. The last topic addresses the challenge of selecting two simultaneous samples in such a way as to reduce overall field costs compared to the cost of implementing the samples independently. The goal is to select the two samples in pairs or cluster such that, one, each cluster consists of units from the two samples and, two, the units are geographically compact. Several methods are described. One group of methods involves a multistage selection of units starting with appropriately-sized primary sampling units (PSUs). Another general approach is to resort to list sampling and arranging dynamically for the geographic linkage of units from the two frames.
Specific application of the methods are discussed in the context of the 2006 US component of the PISA (Program for International Student Assessment)/PIRLS (Progress in International Reading Literacy Study) survey programs.