School District Diversity Associated with Reductions in the Achievement Gap

July 29, 2020

The racial academic achievement gap is a persistent and pernicious educational and social challenge, which is complicated by racialized poverty. Despite over three decades of interventions and federal, state, and local policies and initiatives meant to close it, the gap in academic performance between different races or socioeconomic groups persists — staying with students from an early age to high school and beyond, affecting dropout rates, long‐term college graduation and lifetime earnings.

But what if we could help to bridge that gap simply by doing a better job of bringing different groups of students together?

In a new report, we analyzed 2010–2011 statewide standardized test score data for New Jersey, and found evidence that school district diversity can have a dramatic and progressive relationship with the racial achievement gap.

As the most racially diverse state, New Jersey was an ideal sample for our work as it has the potential to act as a microcosm of schools in the United States. It hosts a range of school district “ecologies,” including urban, suburban and rural school districts as well as those along the shore, inland or near large urban centers such as New York City or Philadelphia. The racial makeup of those districts varies from almost complete segregation (PDF) to high diversity, with significant ranges in terms of size and geographic location.

Using data science techniques rarely applied in the educational arena, we looked at standardized test score data for 600,000 students in 2,500 schools in approximately 500 districts and found that the racial achievement gap decreased when diversity increased. Notably, the achievement gap between Black and white third-grade students was more than 60% lower in racially diverse districts when compared to racially homogeneous districts.   Consistent with other research on the effects of peer diversity on academic performance, this result implicates peer racial diversity as a possible pathway to reducing the achievement gap.

Our analysis also revealed powerful insights about the composition of the student populations of New Jersey’s school districts. Our cluster analysis of New Jersey school districts used standardized test scores in language arts and math, school district size and indicators of relative socioeconomic status (SES) and racial bias (how much above or below average the density of specific student populations is within a school district), to reveal eight common profiles.

Three of the eight clusters collectively represented the school districts in New Jersey, which serve students with both above average SES and standardized test scores. Only one of the higher performing clusters had a statistically diverse student population. The remaining five clusters had below average standardized scores with varying degrees of below average SES. School districts with very low student standardized test scores can be identified in clusters with higher levels of segregation.

This modeling shows promise for more sweeping change, as it could be generalized and applied in other states. The strategic adoption of the cluster model could allow for the efficient and effective use of resources and aid in reducing the achievement gap over the majority of school districts. For example, school districts might benefit from sharing, reflecting on, and learning from patterns at the intersection of student achievement and school district structure. Superintendents and district and school administrators could use cluster group identity and classification to share best practices and successful programs with common issues and needs. Savings might result as district similarities can be used to more efficiently apply remedies and facilitate the release of resources that at a geographic level may be less effective. Actions can be tested and then applied differentially to the various cluster groups with positive response, thus increasing the efficiency of focused experimentation across school district clusters.

As we move forward, it is imperative that we use this data to move the needle on equity in education in an area that has long eluded both education leaders and policymakers. Reductions in the achievement gap could correlate with more positive outcomes throughout the course of students’ educational and professional journeys. With progress here, the impact on society will be both significant and essential. By bringing different groups of students together, it is possible that we can take real steps to close the gap that has historically pushed them further apart.

Dr. Stephen José Hanson is a full professor of Psychology, executive member of the Cognitive Science Center, and director of Rutgers Brain Imagining Center (RUBIC) at Rutgers, The State University of New Jersey – Newark. He led this study with ETS staff in the organization’s Policy Evaluation Research Center.

Michael Nettles is senior vice president and the Edmund W. Gordon Chair for Policy Evaluation & Research at ETS. He is a co-author of this study.