A trimming adjustment procedure is done to detect and trim extremely large student weights. There are several possible causes for such large weights:
underestimation of schools' grade enrollments leading to inappropriately low probabilities of selection for those schools;
presence of large schools (such as high schools) in primary sampling units with small selection probabilities;1
selection of very small schools with very low probabilities to minimize the number of very small schools selected for sample;
high levels of nonresponse coupled with low to moderate probabilities of selection; and
compounding of nonresponse adjustments at various levels.
Students with notably large weights have an unusually large impact on estimates such as weighted means. An occasional unusually large weight is likely to produce large sampling variances of the statistics of interest, especially when the large weights are associated with students with atypical performance characteristics.
To ameliorate this problem, a student trimming adjustment procedure is implemented. The student trimming adjustment works similarly to the school trimming adjustment. Student weights (through the student nonresponse adjustment) are trimmed if their school contributes more than a specified proportion to the variance on the estimated number of students within a given domain. School-level estimates of students are calculated by summing the weights of students (assessed and excluded) in the school.
1 The maximum permissible within-school sampling rate, determined by the maximum sample size allowed per school, could be smaller than the desired overall within-PSU sampling rate for students in such cases.