Here's a full array of disaggregated results from the fourth grade NAEP reading results released yesterday. Most gaps stayed the same, but the race gap widened while the lunch gap narrowed slightly.
These graphs are a good illustration of a key feature of statistical significance: a change that is significant with a large sample of students may not be counted as significant for a smaller group. For example, the four-point improvement for students without disabilities is significant, but the four-point improvement for students with disabilities is not--because fewer students with disabilities were tested.
The three-point movement for students in the free-and-reduced-lunch program is also significant, while the three-point changes for male, female, and white students are [not]. The two points for students not in the lunch program are not significant growth. Unsurprisingly, raising black student results one point is not significant either.