Table of Contents  |  Search Technical Documentation  |  References

NAEP Weighting Procedures → 2002 Weighting Procedures and Variance Estimation → State Assessment Weights → State NAEP School and Student Nonresponse Adjustments → Computation of Student Nonresponse Adjustments

NAEP Technical DocumentationComputation of Student Nonresponse Adjustments

			Development of Student Nonresponse Cells

Student nonresponse in the NAEP 2002 state assessment varied widely across jurisdictions. Student nonresponse was adjusted for through weighting cell adjustment. It should be noted that ineligible students received student nonresponse adjustments equal to 1, and were not included in any aggregations for the computation of student nonresponse adjustments (i.e, the student nonresponse adjustments allowed the assessed students to represent the "to be assessed" students, which included assessed and eligible absent and refusing students).

The following nesting cell structure defined the nonresponse weighting cells for each participating jurisdiction:

• Students with disabilities/limited English proficient (SD/LEP) status (SD and/or LEP, or not);

• school nonresponse cell;

• age of student (classified as "older" student and "normal age or younger" student);¹

• subsidized lunch status (student is eligible for free or reduced-price lunch, or not);² and

• gender of student.

Cells based on this nesting structure that were too small to allow for stable nonresponse adjustment were collapsed with other cells. Details of this process are found in Development of Student Nonresponse Cells. The final student nonresponse weighting cells within each jurisdiction j are designated as STNRADJ_d, d = 1,...,D, where D is the total number of weighting cells designated for the jurisdiction. The student nonresponse adjustments were computed as follows for each nonresponse weighting cell:

where

• S_d is the set of all eligible sampled students in cell d;

• R_d is the set of all assessed students within S_d; and

• STUWGT_sk is the student base weight.

Estimates and standard errors were calculated using both these weights and replicate weights, for each participating jurisdiction, as well as comparisons to the corresponding totals for the full student sample. As can be seen in Weighted Enrollment Checks from the State NAEP 2002 Responding Student Samples, the student nonresponse adjustments have the effect of allowing the responding students to represent the full set of students. The replicate adjustments (Replicate Student Nonresponse Adjustments) are designed to produce variance estimates that reflect the component of variability added by the nonresponse adjustment process.

¹Normal age students for fourth grade are those whose birth months are in the range October 1991 through September 1992, and for eighth grade are those whose birth months are in the range October 1987 through September 1988. For each grade, 'older' children have earlier birth months and 'younger' children have later birth months.

² Missing subsidized lunch status was also a third category.

Printer-friendly Version