The 1990 National Assessment of Educational Progress (NAEP) will mark the first time that state-level comparisons will be possible with NAEP data. The 1990 data will contain, as a trial component, state-level assessments of eighth graders in mathematics. Mathematics and reading comparisons for fourth graders will be added in 1992. Statistical issues associated with these innovations include those related to the following: (1) estimation of within state population characteristics, including estimating variability and accounting for error of measurement; (2) comparison of the data from trial state assessment with data from the national assessment, including evaluating the effect of type of administration and equating the observed administration results to the national results; and (3) between state comparisons. Scaling models should be used to produce estimates of the distribution of abilities within the subject areas, the models should distinguish between content areas, and the models require special technology to take the amount of available information into account. Sample structure must also be addressed. It was noted that the fact that the states differ in terms of demographic, socioeconomic, and other indicators is important, and that these differences may affect the comparability of achievement results between states. (TJH)