Introduction to JEBS Special Issue on NAEP Linked Aggregate Scores

Abstract

The Stanford Education Data Archive (SEDA) was created by Sean Reardon, Andrew Ho, Demetra Kalogrides, and their colleagues using annual state summative test score data retrieved from the EDFacts Restricted-Use Files and publicly available NAEP data from the National Center for Education Statistics. SEDA provides test score data on a common scale across all states for mathematics and reading language arts for students in Grades 3 through 8 for almost all schools, districts, and counties in the United States. An online tool (edopportu nity.org) allows users to visually compare schools and districts from anywhere in the country. Data also include various covariates at each of these levels, and all the data can be downloaded for free for analysis. These data have the potential to be a very valuable resource for researchers, educators, policy makers, and possibly even the general public. The catch is that there is no common standardized test administered to students in Grades 3 through 8 in all schools and school districts in all states. NAEP is only administered in a relatively small sample of schools in each state and only to students in Grades 4 and 8 and only every other year. The school data in SEDA are derived from the annual tests administered by each state in accordance with federal regulations. Reardon, Ho, Kalogrides, and colleagues start with aggregate data of the numbers of students in each school or district meeting various performance levels on their state standardized tests. State tests are on different scales and test somewhat different content. They also use different cutoffs for performance levels that are not common across states. Reardon, Ho, Kalogrides, and colleagues convert these frequencies to means and standard deviations for the scores in each school or district using the Heteroskedastic Ordered Probit model that was developed into a series of papers in JEBS (Lockwood et al., 2018; Reardon et al., 2017; Shear & Reardon, 2021). They then link these means and standard deviations to the NAEP scale using methods described in Reardon et al. (2021). Reardon, Ho, Kalogrides, and colleagues stitched together a collection of methods to create a national data source of Journal of Educational and Behavioral Statistics 2021, Vol. 46, No. 2, pp. 135–137 DOI: 10.3102/10769986211001480 Article reuse guidelines: sagepub.com/journals-permissions © 2021 AERA. https://journals.sagepub.com/home/jeb