Since the authorization of the use of the National Assessment of Educational Progress (NAEP) through the Trial State Assessment to provide state-based data, debate over the proposed expansion of the NAEP has strengthened. To assist states in analyzing and interpreting 1990 data on eighth-grade mathematics and to assist in planning subsequent NAEP reports, two analytic strategies were used for this report. To highlight relationships between mathematics achievement and variables explaining variation in achievement on the 1990 NAEP, a series of cross-tabulations was prepared for seven states with higher and seven states with lower mathematics proficiency. Secondly, a set of variables was selected for analysis that would indicate differences across dimensions of state policies, curriculum and instruction, and context for education. Analyses indicate that there is a difference between high and low proficiency states on instructional emphasis, with more stress on Algebra in high proficiency states. While no significant differences were found between high and low achieving states in small-group instruction, there were differences in calculator use in instruction. High proficiency states have more experienced mathematics teachers with more mathematics teacher preparation and better educational resources than did low proficiency states. Parental education level was higher for high achievement states, but high and low achieving states did not differ in terms of family stability and mobility. Fifteen tables and one figure present analysis results. There is a six-item list of references. (SLD)