Skip Navigation

Table of Contents  |  Search Technical Documentation  |  References

School Sample Selection


Self-Weighting at the Student Level

Computation of Measures of Size

Sample Selection

New-School Sample Selection

The general target for each jurisdiction is 6,510 sampled students for each grade. This target is a product of 105 schools, with 31 students per subject per school, and two subjects—mathematics and reading. These figures were the results of the following assumptions and calculations. As specified by ETS in consultation with NCES, the sample size requirement for each participating state and jurisdiction was approximately 2,500 assessed students per grade and subject. To accommodate the inclusion of pilot tests for grades 4 and 8 in the same sessions as the operational assessments, and to allow for student attrition, the specified student sample size was increased by a factor of 48/47. Thus, instead of taking the traditional 60 students per school as in the past, the design called for a sample of 60 X 48/47 = 61.3 students per school. Rounding to the next highest integer resulted in a target sample size of 62 students per school, which corresponds roughly to two assessment sessions for most schools. Then, to provide a 5 percent allowance for ineligible schools on the sampling frame, the traditional sample of 100 schools per jurisdiction was increased to 105. Hence, the target sample size of 6,510 was obtained by multiplying the school target by the target number of students within each school.

Last updated 18 August 2008 (KL)

Printer-friendly Version