Linking Statewide Tests to the National Assessment of Educational Progress: Stability of Results

During the past few years there has been considerable discussion among educational policy makers and measurement specialists regarding the possibility of linking data from different assessments. For example, several states have expressed an interest in linking their statewide assessments to the National Assessment of Educational Progress (NAEP), hoping that through linking, the results of a state’s own assessment can be compared to national results provided by NAEP and possibly even to international test results. To explore this possibility, CRESST researchers Robert Linn and Vonda Kiplinger compared statewide mathematics test data for eighth-grade students from 1990 and 1992 with results of the NAEP Trial State Assessment administrations for the same years. Equipercentile equating procedures were applied and the results analyzed. The authors found that the conditions for test equating were not satisfied in this comparison. “The lack of stability,” write Linn and Kiplinger, “suggests that linking standardized tests to NAEP using equipercentile equating procedures is not sufficiently trustworthy to use for other than rough approximations.” However, the authors note the tests were not designed with the purpose of linking in mind and that much better results might be expected if the tests had been designed in accordance with a common framework. “If linking is an important goal,” say Linn and Kiplinger, “then it would seem wise to assure, at a minimum, that the tests share a common content framework.”