The National Assessment of Educational Progress (NAEP) recently expanded its national survey to include state representative samples on a trial basis. In 1992, the NAEP undertook studied non-response in the trial assessment through two simulation projects. The first looked at state assessments that were similar to NAEP tests for states with low NAEP response rates. The second project took states with 100 percent response rates and simulated levels and patterns of non-response of states with low NAEP response rates. In the first case, the simulation using figures from Illinois suggests that it might be useful to explore the use of substitution further for other states. Results suggest that one could simulate the patterns of non-response based on states with less than 100 percent response rates to try to develop a sense of when non-response becomes so large it has an impact on the estimated results. Two tables present analysis results. (SLD)