The first step in computing the replicate weights was to compute replicate student base weights, as follows:
None of the student-level factors were replicated: replication is at the school level alone. The next step was to replicate the student nonresponse adjustments. The student trimming step was not replicated. The replicate pre-poststratification weight was
The final step was to replicate the poststratification adjustments. The replicate poststratification adjustments are
The replicate final weights are
The tables show replicate variance estimates using wij and wij (r) for totals of the following groups, for fourth, eighth, and twelfth grades, respectively:
school type (public, Catholic, religious non-Catholic, private nonreligious),
Note that these totals include all students whether assessed, absent, or excluded, and whatever their assessment (geography or U.S. history).