Frequently Asked Questions About the Black-White Achievement Report

There is little information on the relationship between school composition and achievement gaps. Because of the growing concern with resegregation in schools (Frankenberg, Lee, & Orfield, 2003; Orfield, Kucsera, & Siegel-Hawley, 2012), there is a need for additional information about this relationship.

Black student density is defined as the percentage of students in a school who are Black. For this study the percentages were calculated using the number of students in the school (regardless of race/ethnicity) in the denominator and the number of Black students in the school in the numerator.

The term density has been used in previous NCES reports (e.g., NCES, 2012) and in peer-reviewed articles (e.g., Madyun & Lee, 2010; Gieling, Vollebergh, & van Dorsselaer, 2010) to indicate demographic composition within schools.

This report begins with a description of how Black student density differs for Black and White students, and identifies the U.S. regions and locale types where the highest Black student density schools are located.

The report then highlights results from two separate sets of analyses. The first analysis explores the relationship of Black student density with achievement and achievement gaps. This analysis explores whether the identified relationship between density and student achievement persists when accounting for student demographic characteristics such as socioeconomic status (SES) and other school characteristics that may be associated with achievement. It also explores results separately for males and females.

The final section of the report builds on prior research that explored whether achievement gaps might be found more within schools or between schools. Whether gaps are larger between schools or within schools can inform the actions of those who are concerned with improving student achievement and reducing the achievement gap.

For technical aspects of reporting student achievement, see Analysis and Scaling for NAEP.

The descriptions of the extent of Black student high- and low-density schools and their location relied on two data sources:

1. NAEP 2011 Mathematics Grade 8 Assessment

2. Common Core of Data (CCD) 2010–11

The CCD data were limited to schools that offer eighth grade and whose Type was identified as a “regular” school.

These analyses were based on the NAEP 2011 Mathematics Grade 8 Assessment student, teacher, and school information for Black and White students. Additional school-level information was added onto the NAEP data from the CCD.

No. These analyses focused on the average achievement of Black and White students only. Regression analysis was used to adjust the means to control for SES, student, teacher, and school factors, but the regression model was estimated using only Black and White students.

The student characteristics included measures from the NAEP student questionnaire: race/ethnicity, disability status, gender, and SES measures. Teacher characteristics included measures from the NAEP teacher questionnaire such as teacher qualifications and teacher instructional strategies. School factors included SES measures aggregated from student SES measures as well as other school factors from the NAEP school questionnaire and from the CCD. See the methodology companion for a full list of the variables used in the model.

Regression analysis was used to calculate average achievement for Black and White students in each Black student density category to account for differences in student, teacher, and school characteristics. When the analysis accounts for these differences, the averages reported represent the average achievement for students as if they all had the same student, teacher, and school characteristics. The methodology companion provides a complete description of the model used, how means were calculated, and how significance testing was conducted.

This analysis was based on the NAEP 2011 Mathematics Grade 8 Assessment student, teacher, and school information for all students in the sample.

These are differences in the achievement of Black and White students. It is important to note that this analysis does not identify what causes these differences, but rather, it describes whether the differences in achievement between Black and White students are observed between or within schools. Each portion—between or within—can be addressed by different sorts of policies.

NAEP has two major goals: to compare student achievement in states and other jurisdictions and to track changes in achievement of fourth-, eighth-, and twelfth-graders over time in mathematics, reading, writing, science, and other content domains. To meet these dual goals, NAEP selects nationally representative samples of students who participate in either the main NAEP assessments or the long-term trend NAEP assessments.

For technical aspects of reporting student achievement, see Analysis and Scaling for NAEP.

The relationship between Black student density and achievement is a key variable in the statistical analysis: the stronger the relationship between density and achievement, the larger the portion of the gap attributed to between-school differences.

In the extreme, if differences in the school density of Black students are not related to the achievement gap, then differences in achievement cannot be found between schools and are only found within schools.

No. These analyses decompose the whole achievement gap, NOT just the part that cannot be attributed to SES or other factors.

These analyses seek to examine how policies can best address the entire Black-White achievement gap, regardless of whether it is explained by SES or not. In other words, if the analyses were to look at only the portion of the gap not explained by SES or other factors, the results would not be able to help inform how to close that part of the gap that is explained by SES.

The absence of student and school control variables from these analyses may cause confusion as it is a departure from the analysis presented in the first chapter of the report on the relationship between achievement and density, where SES and other factors were accounted for in the model. The research in that previous chapter sought to investigate whether density was correlated with the achievement gap. In such an investigation, one would want to control for SES and other potentially confounding factors to examine potential relationships among density, the variable of interest, and the achievement gap. The decomposition analysis is different in that it is descriptive and does not seek to determine a correlational relationship. Specifically, the decomposition analysis is a description of where the achievement gaps are occurring so that policies might be optimally directed (e.g., focused on the distribution of resources within, rather than between, schools).

We know that prior research shows that SES is related to the achievement gap, so it is likely that part of the 31-point gap nationally is due to SES differences. The analyses in the report, however, are focused on providing information for policies that address the gap, not uncovering the causes. SES differences are not policy levers—in other words, education policy cannot directly change the SES of a student’s family. However, education policy can redistribute resources (as an example) such that even the portion of the gap attributable to SES differences is addressed.

Different researchers have used different formulae to conduct this decomposition. Reardon (2008) reconciled these different approaches by identifying a portion that all of the different approaches attributed to within-school differences, a portion that all attributed to between-school differences, and remainder that was classified by some approaches as between school and others as within school. Reardon labeled the third portion as “ambiguous,” but in this report we use the label “indeterminate.” Theoretically, policies focused either on the within-school or between-school differences might help close the indeterminate portion of the achievement gap.