Association for Public Policy Analysis & Management

American children of the 21^st century face a world of rising income inequality, increased residential income segregation, and other social forces likely to exacerbate the strength of the association between academic performance and family socioeconomic background. The Great Recession exacerbated these trends, with unemployment and household wealth impacts disproportionately affecting households of lower socioeconomic status (Danziger, 2013). Simultaneously, the education system has undergone a broad education standardization movement (the Common Core State Standards), multiple states have enacted school finance reforms, and there has been a six-fold expansion in the share of students attending charter schools – all reforms pursued with the hopes of reducing educational inequality, at least rhetorically.

The goals of this study are twofold. First, we describe changes in measures of the association between family socioeconomic status and student academic achievement over the last decade at the national, state, and district levels. Second, we perform analyses exploring the extent to which the observed changes are consistent with being caused by any of the major state-level policy changes or documented social trends.

The data are drawn from the Stanford Education Data Archive (SEDA), which includes district-level data on every state accountability test taken in grades 3—8 in the U.S. between 2009 and 2016 (Reardon, Kalogrides, & Ho, 2017). In addition, we examine national, state, and trial urban district (TUDA) subgroup scores on the NAEP exam (2003—2017). These data are supplemented by a measure of district socioeconomic status constructed from the American Community Survey (ACS), as well as measures of state Common Core implementation, school finance reforms, charter school participation, residential segregation, income inequality, and adult employment.

The relationship between family socioeconomic status and educational achievement is captured in slightly different ways for different levels of the education system:

• Nationally, between states, and at the state level, we examine the gap in NAEP scores by student eligibility for free/reduced price lunch.
• Between districts within states, we examine the association between district mean performance on state standardized tests and district socioeconomic status.
• Within districts, we examine the achievement gap between economically disadvantaged and non- economically disadvantaged students on state standardized tests.

Across all levels of the educational system, we find evidence the association between SES and academic performance has strengthened over the past decade. The gap between lunch-eligible and lunch-ineligible students on the national NAEP exam has widened in both math and reading (ELA) since 2005, for both 4^th and 8^th grades. At the state level, the association between district SES and performance increased roughly half a standard deviation, with a larger increase in math than in ELA. This pattern is relatively unchanged after accounting for changes to standardized tests. The gap between economically disadvantaged and non-economically disadvantaged students within school districts also rose consequentially, by 19% of a standard deviation in math and 8% of a standard deviation in ELA.