Leveraging Existing Data to Study Educational Inequality: Examples, Lessons, and Recommendations

• In Event: 42.023 - How Educational Research and Assessment Can Be Used to Address Educational Inequities: An International Perspective

Abstract

This talk describes three efforts to study patterns, trends, and causes of educational inequality as measured by disparities in academic performance. The first uses data from 14 nationally representative studies conducted in the U.S. from 1960-2010 to estimate the trend in the “income achievement gap” in the U.S (Reardon, 2011; Reardon & Portilla, 2016). The second uses data from state accountability tests administered in every public elementary and middle school in the U.S. from 2009-2014 (roughly 250 million test scores) to describe the academic performance of students in every school district in the U.S., broken down by race, ethnicity, free-lunch eligibility, and gender. The third uses data from the PIRLS and PISA international assessments to construct comparable measures of the income achievement gap in OECD countries from 2000-2012 (Chmielewski & Reardon, 2016).
In each case, the data available were not ideally suited to the study of inequality. In particular, family income and socioeconomic status were not measured consistently across studies and data sources. In many cases, the tests were not scored comparably in different studies and data sources. To address these issues, my colleagues and I adapted or developed methods to render the measures comparable.
The talk ends with conclusions and recommendations regarding the types of data and methods needed to facilitate the study of educational inequality.

Author

• Sean F. Reardon, Stanford University