National probability samples of schools and students are selected to represent the United States. The numbers of schools and students vary from cycle to cycle depending on the number of subjects and items assessed. This national sample has sufficient schools and students to yield data for public and nonpublic schools and four regions of the country (northeast, southeast, central, and west), as well as sex, race, degree of urbanization of school location, parent education, and participation in the National School Lunch Program (NSLP).
The sample of schools is selected from geographic sampling units. There are 22 geographic sampling units that are always selected. These are large metropolitan areas such as New York City, Los Angeles, and Chicago, and between them they include about 40% of the students in the country. Without this group, the sample would not be representative of the United States.
An additional 72 geographic sampling units are selected randomly to represent the rest of the United States. This second group includes smaller cities and rural areas, and the choice of areas varies with each assessment cycle. This design allows the centrally administered assessments to be scheduled efficiently.
Students are selected randomly within schools. From 30 to 150 students are selected in each school depending on the size of the school and the number of subjects to be assessed. Some of the students who are randomly selected are classified as students with disabilities (SD) or as limited-English proficient (LEP) students. NAEP's goal is to assess all students in the sample, and will assess these students if at all possible. To read more, see the full description of accommodations issues.
The number of schools ranges from 900 to 2,500 and the number of students from 40,000 to 150,000. Again, the variation depends on the number of subjects to be assessed and the number of assessment booklets developed for each subject. Generally, the requirement is for each assessment booklet to be answered by 2,000 students.