The 2003 national main assessment in mathematics and reading was designed to achieve a nationally representative sample of public and private school students in the fourth and eighth grades. The sample was designed to cover the 50 states and the District of Columbia, including Bureau of Indian Affairs (BIA) and DDESS. Although data were collected from the territories and DoDDS as part of the state assessments, these schools were not considered part of the national population. The target population included all students in public and private schools who were enrolled in fourth or eighth grade at the time of assessment. The national sample consisted of a combination of the state samples of public schools for 4th and 8th grades and nationally-representative supplemental samples of fourth- and eighth-grade private schools.
The selected samples were based on a two-stage sample design that involved selection of schools within strata and selection of students within schools. The first-stage samples of schools were selected with probability proportional to a measure of size based on the estimated grade-specific enrollment in the schools.
For fourth and eighth grades, the base public school samples were the NAEP state assessment public school samples for each jurisdiction. Likewise, the NAEP state student samples and assessments became the NAEP national main student samples and assessments as well.
Since the NAEP state samples only cover public schools for fourth and eighth grades, there was a need to supplement the NAEP state samples with a nationally representative private school sample for fourth and eighth grades. In addition, to account for the possibility of jurisdictions dropping out of the state assessment, it was necessary to draw student subsamples from the NAEP state school samples. The student subsamples would then represent those jurisdictions in the national main sample in the event that the jurisdictions dropped out of the state assessment.
The target sample sizes for public school assessed students in grades 4 and 8 were established at the jurisdiction level and not at the national level, since the samples had to satisfy the requirements for both the nation and the states. For each participating jurisdiction, a target of approximately 2,500 assessed students was established by grade and subject as meeting those goals.
The private school samples for each grade were explicitly divided into five strata, and implicitly stratified within the explicit strata. A target sample size of 12,600 assessed private school students was established for both grades 4 and 8.
Finally, as was done for the grades 4 and 8 state NAEP samples, substitutes were preselected for each private school sample.