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 Private Schools

NAEP Technical DocumentationSchool Selection for Private Schools

			Target Fourth-Grade Private School Sample Sizes Target Eighth-Grade Private School Sample Sizes Target Twelfth-Grade Private School Sample Sizes Sampling Parameter Values for Private School Strata

The private school sample was drawn to cover all of the assessment types: the operational assessments,  the field tests, the equating link to the 1998 reading assessment, and the equating link to the 2000 mathematics assessment. The discussion will focus on the operational assessments sampling process in this chapter, but at the school level, it describes the initial school sample for all assessments.

The procedure for setting measures of size to achieve the targets given above are essentially the same as for the 2002 NAEP state assessment, with each private school stratum done separately. Define x_js as the grade enrollment for the school (from the PSS). The initial measure of size MOS_js for school s in each private school stratum j is set as follows as a function of x_js  conditionally (each of the four pairs following the equal sign correspond to a condition on x_js and a functional relationship between MOS_js and x_js which is used to assign MOS_js under that condition: for example, schools with enrollment x_js greater than or equal to 65 had their MOS_js set equal to x_js):

Schools for which x_js was unknown from the PSS frame were assigned a measure of size MOS_js of 25 (the general size of one classroom). Multiple "hits" are allowed in the eighth- and twelfth-grade samples (schools with n hits (n=2, 3,...) receive student sample sizes n times larger). Schools with an E_js  value less than 1 could receive only one hit (indicating a regular student sample size). Schools with E_js  values between 1 and 2 could receive either one hit or two hits, and in general, schools with general E_js  values could receive hits equal to the largest integer smaller than E_js  or the smallest integer greater than E_js. The upper bounds u_j on hits are 1 for fourth-grade private schools, 2 for eighth-grade private schools, and 3 for twelfth-grade private schools.

The measures of size E_js are calculated as follows:

The quantities b_j are the sampling parameters for each stratum j (see Sampling Parameter Values for Private School Strata). These measures E_js are the final measures of size for the fourth- and eighth-grade samples. For the twelfth-grade sample, the constraint to minimize overlap with the Education Longitudinal Study was added, as was done for the twelfth-grade public school sample.

Printer-friendly Version