The Great Divergence: Growing Disparities Between the Nation’s Highest and Lowest Achievers in NAEP Mathematics and Reading Between 2009 and 2019

The Great Divergence blog post image

Grady Wilburn, National Center for Education Statistics

Brian Cramer, Optimal Solutions Group

Ebony Walton, National Center for Education Statistics

With the release of the 2019 Nation’s Report Cards for mathematics and reading in grades 4 and 8, we have compiled an overview of key results spanning the last two decades of NAEP. The dominant theme that emerges is the appearance of a growing divergence in achievement between the highest and lowest achieving students. This divergence is seen for the nation as a whole, across states, and for student groups by race/ethnicity and socioeconomic status.

This post compiles results from the first year available for all states, 2003, through 2019. Results are divided roughly into decades. The first period is between 2003 and 2009 (the “2000s”) and the second period is between 2009 and 2019 (the “2010s”). All results were produced by the online NAEP analysis tool, the NAEP Data Explorer.

More Gains on NAEP in the 2000s Than the 2010s

Between 2003 and 2019, the nation made gains in both mathematics and reading at both grades with the exception of reading grade 8, where the nation made no gain. However, most of the gains occurred between 2003 and 2009. In mathematics, the nation made gains of five points at grade 4 and six points at grade 8 between 2003 and 2009—but the 2019 score at grade 4 is up just one point from 2009 and the 2019 grade 8 score is not measurably different from 2009 (see figure 1a).

Figure 1a. Mathematics Score Changes for the Nation (public)
Grade2003-192003-092009-19
green triangleSignificant increase (p <.05) blue diamondNo significant change (p <.05)
Grade 4 green triangle(6) green triangle(5) green triangle(1)
Grade 8 green triangle(5) green triangle(6) blue diamond     
Figure 1b. Reading Score Changes for the Nation (public)
Grade2003-192003-092009-19
green triangleSignificant increase (p <.05) blue diamondNo significant change (p <.05)
Grade 4 green triangle(3) green triangle(3) blue diamond
Grade 8 blue diamond      green triangle(1) blue diamond

Forty-four states made gains in each grade between 2003 and 2009, while only 13 states made gains at grade 4 and 5 at grade 8 between 2009 and 2019 (see figure 2a).

In reading at grade 4, the nation gained three points between 2003 and 2009 but made no gains between 2009 and 2019; twenty-two states made gains between 2003 and 2009, compared with only 9 states between 2009 and 2019 (see figures 1b and 2b).

In reading at grade 8, the nation gained one point between 2003 and 2009 but showed no gain between 2009 and 2019; eight states made gains between 2003 and 2009, compared with only seven states between 2009 and 2019.

Figure 2a. Number of States with Mathematics Score Increases
Grade2003-20092009-19
Grade 44413
Grade 8445
Figure 2b. Number of States with Reading Score Increases
Grade2003-20092009-19
Grade 4229
Grade 887

Divergence Between Highest and Lowest Achievers

The stagnation in average scores since 2009 masks a divergence in progress at the highest and lowest levels of achievement. If we look at changes over time in scores at the 10th, 25th, 50th, 75th, and 90th percentiles, we can see whether progress has been even across the performance distribution of students or if there is more progress at the highest levels of achievement or the lowest.

What we observe in the 2000s is progress across the performance distribution in mathematics and at the lowest achievement levels in reading. For example, scores in grade 4 mathematics at the 10th percentile in 2009 were five points higher than in 2003—and so were scores at the 90th percentile (see figure 3a). At grade 8, scores in 2009 were six points higher at both the 10th and 90th percentiles than in 2003.

In reading, the score at the 10th percentile at grade 4 rose seven points between 2003 and 2009 and two points at the 10th percentile at grade 8. Scores at the 90th percentile at both grade levels were not measurably different between 2003 and 2009 (see figure 3b). This suggests progress in achievement among the lowest achieving students, but not among the highest achieving.

Figure 3a. Mathematics Score Changes for the Highest and Lowest Achievers
Grade2003-092009-19
10th90th10th90th
green triangleSignificant increase (p <.05) blue diamondNo significant change (p <.05) red triangleSignificant decrease (p <.05)
Grade 4 green triangle(5) green triangle(5) red triangle(3) green triangle(5)
Grade 8 green triangle(6) green triangle(6) red triangle(5) green triangle(5)
Figure 3b. Reading Score Changes for the Highest and Lowest Achievers
Grade2003-092009-19
10th90th10th90th
green triangleSignificant increase (p <.05) blue diamondNo significant change (p <.05) red triangleSignificant decrease (p <.05)
Grade 4 green triangle(7) blue diamond red triangle(6) green triangle(3)
Grade 8 green triangle(2) blue diamond red triangle(6) green triangle(4)

The story is different between 2009 and 2019, where we see a divergence in growth between the lowest and highest achieving in both subjects and grades. The score at the 10th percentile dropped between 2009 and 2019 in both subjects at both grade levels. On the other hand, the score at the 90th percentile rose in both subjects and grades. The 90th percentile rose by five points in mathematics in each grade. In reading, the score at the 90th percentile rose three points at grade 4 and four points at grade 8.

This phenomenon appears across major student reporting groups by race/ethnicity and socioeconomic status. Black students, Hispanic students, and White students, as well as students eligible for the National School Lunch Program (NSLP) all show a growing divergence in achievement between the group’s lowest and highest achieving students. A period of gains at the 10th percentile over the 2003 to 2009 period (except for Black and Hispanic students in grade 8 reading) is followed by losses or stagnation between 2009 and 2019 (see figures 4a-4d). At the 90th percentile, there are gains or stagnation over the 2003 to 2009 period followed by more gains between 2009 and 2019.

Figure 4a. Mathematics Score Changes for the Highest and Lowest Achievers, by Race/Ethnicity - Grade 4
Race/Ethnicity2003-092009-19
10th90th10th90th
green triangleSignificant increase (p <.05) blue diamondNo significant change (p <.05) red triangleSignificant decrease (p <.05)
White green triangle green triangle red triangle green triangle
Black green triangle green triangle red triangle green triangle
Hispanic green triangle green triangle blue diamond green triangle
NSLP eligible green triangle green triangle red triangle green triangle
Figure 4b. Mathematics Score Changes for the Highest and Lowest Achievers, by Race/Ethnicity - Grade 8
Race/Ethnicity2003-092009-19
10th90th10th90th
green triangleSignificant increase (p <.05) blue diamondNo significant change (p <.05) red triangleSignificant decrease (p <.05)
White green triangle green triangle red triangle green triangle
Black green triangle green triangle blue diamond green triangle
Hispanic green triangle green triangle blue diamond green triangle
NSLP eligible green triangle green triangle blue diamond green triangle
Figure 4c. Reading Score Changes for the Highest and Lowest Achievers, by Race/Ethnicity - Grade 4
Race/Ethnicity2003-092009-19
10th90th10th90th
green triangleSignificant increase (p <.05) blue diamondNo significant change (p <.05) red triangleSignificant decrease (p <.05)
White green triangle blue diamond red triangle green triangle
Black green triangle green triangle red triangle green triangle
Hispanic green triangle blue diamond blue diamond green triangle
NSLP eligible green triangle green triangle red triangle green triangle
Figure 4d. Reading Score Changes for the Highest and Lowest Achievers, by Race/Ethnicity - Grade 8
Race/Ethnicity2003-092009-19
10th90th10th90th
green triangleSignificant increase (p <.05) blue diamondNo significant change (p <.05) red triangleSignificant decrease (p <.05)
White green triangle blue diamond red triangle green triangle
Black blue diamond blue diamond red triangle green triangle
Hispanic blue diamond blue diamond blue diamond green triangle
NSLP eligible green triangle blue diamond red triangle green triangle

Data Available in the Nation's Report Card and NAEP Data Explorer

More results are available in the Nation’s Report Card, as well as the NAEP Data Explorer, a powerful online tool for analysis of NAEP data. NAEP is not designed for causal analysis that could determine the cause of the divergence. However, further analysis of NAEP could help to pinpoint when the divergence began and identify the jurisdictions where the divergence is most prevalent or perhaps not occurring.



Leave a Comment

What did you think about this blog post? Share your thoughts and/or questions with us by leaving a comment below. Click here to read our comment policy.