Because large scale surveys like NAEP use complex sampling designs, resampling methods such as jackknife or bootstrap must be used to estimate the precision of student achievement. This project investigated the use of hierarchical linear models to estimate standard errors for NAEP proficiency scores.
