Table of Contents  |  Search Technical Documentation  |  References

NAEP Sample Design → NAEP 2002 Sample Design → 2002 National Main Assessment Sample Design → School Sample Selection for the 2002 National Main Assessment → School Selection for Fourth- and Eighth-Grade Public Schools

NAEP Technical DocumentationSchool Selection for Fourth- and Eighth-Grade Public Schools

			Targets for Public Schools, Fourth and Eighth Grades Targets for NAEP State Nonparticipating Jurisdictions

As stated above, in general, the fourth- and eighth-grade public school samples for the national study were identical to the state assessment samples. The samples for the state assessment were all used for national NAEP as well. There were several states and jurisdictions that chose not to participate in the state assessment, so that no recruitment was done of the State NAEP samples in these cases. These jurisdictions were as follows:

• Alaska,
• the Bureau of Indian Affairs (BIA) schools,
• Colorado,
• Iowa (for eighth grade only),
• New Hampshire,
• New Jersey, and
• South Dakota.

For these jurisdictions, subsamples were drawn from the state samples to use for the national sample.

The process for drawing subsamples was to compute appropriate school probabilities of selection using the national targets for the jurisdictions (as if no State NAEP samples were drawn). Then these probabilities were divided by the initial State NAEP probabilities of selection for the schools to obtain conditional probabilities of selection for subsampling (in all cases these measures of size were less than or equal to 1). School samples were drawn using the conditional probabilities. The resultant unconditional probabilities of selection for the subsample of schools were equal to the appropriate values for a stand-alone national sample.¹

The targets were developed before it was known exactly which states and jurisdictions would participate in the state program. Also, it was not known who would be administering the state assessment program assessments, which impacted the design. Therefore, targets were established for all states. The goal was 32,000 assessed students for each grade. The targets allowed for school and student nonresponse at rates observed in NAEP 2000. After computing initial targets for all states, the national student sample size targets for the two largest states California and Texas were reduced to half of the State NAEP targets, or to 3,150 in each case.² The reduced California and Texas targets were subtracted from the national targets allowing for school and student nonresponse, and the remaining national targets were allocated to the 48 states, District of Columbia, and to BIA and DoD schools proportional to estimated grade enrollment.

¹ NAEP sampling staff did this for the new school sample as well.
² It should be noted that both California and Texas participated in the state assessment, so that these targets are only relevant in the computation of targets for the nonparticipating states. This reduced target of half of the state sample was established to limit the national subsample to at most half of the state sample in each state because of possible issues in equating subsamples.

Printer-friendly Version