NAEP Technical DocumentationOversampling SD/LEP Students for the 2001 National Main Assessment

The procedures for assessing students with disabilities (SD) and limited-English-proficient (LEP) students varied by sample type. SD/LEP students in sample type 3 were offered various accommodations not available to either non-SD/LEP students or SD/LEP students in sample type 2.

As a measure to ensure an adequate sample size of SD/LEP students from both sample types 2 and 3, oversampling procedures were applied to SD/LEP students at all three grades. In this way, comparisons of the effect of offering accommodations to students have enhanced power to detect effects.

The general intent of oversampling within each school was to select SD/LEP students at twice the rate at which non SD/LEP students were sampled (or to include all SD/LEP students if there were not sufficient numbers to permit sampling at twice the rate.) There was no oversampling of schools as part of the procedure.

The procedure was as follows. In each school where oversampling of SD/LEP students was to occur, the initial desired student sample selection was drawn from the full list of eligible students. In addition, in public schools in low minority areas, oversampling of Black and Hispanic students occurred. Among those students not selected for either of the two prior sampling operations for this school, the SD/LEP students were identified. A sample from among these was drawn, using a sampling rate that would achieve the double sampling rate required overall. In most grade 4 cases, this involved selecting all such SD/LEP students in the school. The weighting procedures ensured that the results were not biased as a result of the relative under-representation of SD/LEP students from smaller schools.

Last updated 15 January 2010 (GF)

Printer-friendly Version