This study examined two major questions: do national and state assessments provide consistent information on the performance of state education systems, and what accounts for discrepancies between national and state assessment results if they are found? Data came from national and state assessments in grade 4 and grade 8 mathematics from 1992 to 1996 in Maine and Kentucky from the National Assessment of Educational Progress (NAEP), the Kentucky Instructional Results Information System (KIRIS), and the Maine Educational Assessment (MEA). NAEP and state assessments reported inconsistent results on the performance level of students in Maine and Kentucky across grades and years. Both MEA and KIRIS appear to have more rigorous performance standards, which reduces the percentage of students identified as performing at Proficient/Advanced level. These discrepancies may be understood in light of the differences between the NAEP and state assessments in their definitions of performance standards and the methods of standard setting. The size of achievement gaps between different groups of students appeared somewhat smaller on state assessments than on the NAEP, and the sizes of achievement gains from the states ownassessments were considerably greater than that of the NAEP. These findings raise cautions in using either national or state assessments alone to evaluate the performance of particular state education systems. Policymakers and educators should become more aware of the unique features and limitations of the current national and state assessments. (Contains 3 figures, 17 tables, and 22 references.) (SLD)