The replicate weights for the SD/LEP students were raked using the same procedures discussed in Raking of Weights. Control totals for each replicate were calculated based on the totals for the replicate weights. This maintained the magnitude of the variance leading into the raking procedure. The levels of the raking dimensions that were used for the replicates were the same collapsed levels as shown in Raking Dimensions for Full Sample Student Weights. Let be the raking factor for replicate r from the raking cell k" (resulting combination of raking variables) for student j from school i. The final replicate-specific student weights were then obtained by applying the student raking procedures to each set of replicate student weights. The final replicate r student weight for student j in school i in raking cell k" was calculated as
It should be mentioned that some trimming was conducted after the raking procedure. The full sample trimming factors from the second trimming procedure were applied to each raked replicate weight.