Student sampling for the sampled private schools was identical to that of the State NAEP public school samples. Student sampling for the twelfth-grade public school sample was similar to that of the State NAEP public school samples, except for two features. In the twelfth-grade public school samples, all students were taken if the school had up to 149 twelfth-graders. Otherwise (150 twelfth-graders or more), a sample of 136 twelfth-graders was taken. The second feature was oversampling of minority students in the low-minority stratum schools. If there were any minority students left after the primary sample of 136 students was taken, then the oversample came from this remainder, in such a way that the sampling rate was the minimum of unity and two times the primary sample sampling rate. Thus, all minority students were taken if the primary sample sampling rate P was 50 percent or higher, and a total 2*P of the minority students were taken if the primary sample sampling rate P was less than 50 percent.
For the twelfth-grade public school sample, the student samples were allocated to "A" session (operational reading or writing assessment) or "C" session (link to 1998 NAEP reading study). The general goal was to assign 70 percent of the twelfth-graders to the "A" session, and 30 percent of the twelfth-graders to the "C" session. This was not done through spiraling, but by assigning "sessions" to schools at a classroom level. The smallest schools (enrollment of 34 twelfth-graders or less) had all of their students assigned to either the "A" session (with 70% probability) or to the "C" session (with 30% probability). The remaining schools (enrollment 35 twelfth-graders or more) had some students assigned to "A" sessions and some to "C" sessions, with overall percentages of "A" students equaling 70 percent.
For the private school samples, the school samples provided the students for all of the different assessments: "A," "B," "C," and "D." For fourth-grade private schools, the sampled schools were randomly assigned to "A," "B," "C," or "D," with all sampled students then receiving that assessment. For eighth-grade schools, sampled schools were assigned two sessions (e.g., "AB"), with half the students receiving the first assessment and half the second assessment. For twelfth-grade schools, sampled schools were also assigned two sessions, but only "A" and "C" (the "B" and "D" studies were not done in the twelfth grade).