Questions Related to Analysis and Results
Questions Related to Demographic Data
View frequently asked questions about the Black-White achievement gap report.
The report uses public school data from the National Assessment of Educational Progress (NAEP) grades 4 and 8 mathematics and reading assessments to produce the following:
Only public school data are used because NAEP only collects public school data at the state level.
The Nation’s Report Card provides the most current NAEP data (2009) with comparisons going back as far as 1990. It reports only national trends in Hispanic-White gaps, while this report gives a comprehensive picture of gap changes over time, for all jurisdictions and within subgroups. In addition, the report presents new analyses on the Hispanic-White achievement gap broken down by gender, NSLP eligibility, and ELL status for both grades 4 and 8.
The following table shows how many states were included at each grade and subject.
2009 NAEP Assessment
Some states, like West Virginia, did not have enough Hispanic students for NCES to obtain a reliable sample. The District of Columbia, at grade 8, did not have enough White students. In addition, some states that had reliable data for 2009 did not participate in earlier NAEP assessments. In some cases, states did participate but the NAEP sample was not large enough to produce reliable results for Hispanic students. However, all available data for a state appear in the report, even if no comparisons of Hispanic and White student performance could be made.
No TUDA districts were included. The TUDA assessments are relatively new and many of the districts have not participated in enough assessments to generate meaningful trend data.
The trend begins in 1990 for mathematics and 1992 for reading for nationally representative results. For state results, grade 8 mathematics was first assessed in 1990, grade 4 reading and grade 4 mathematics were both first assessed in 1992, and grade 8 reading in 1998.
The analysis of NAEP data contained in this report should not be seen to imply causal relations. Simple cross-tabulations of a variable with measures of educational achievement, like the ones presented here, cannot be considered as evidence that differences in the variable cause differences in educational achievement. There are many possible reasons why the performance of one group of students will differ from that of another. Inferences related to student group performance should take into consideration the many socioeconomic and educational factors that may also be associated with performance.
All results given here are in terms of average scores, which reflect a wide range of student performance. Some Hispanic students could score above the average for some White students and some White students could score below the average for Hispanic students. All statistical tests are performed using unrounded numbers. Therefore, occasionally the lower scale score plus the gap might not equal the higher scale score shown in this report’s graphics.
Ranking states by gap size is not very useful because, for most comparisons, the differences in gap size are not statistically significant. As tables in the report show, in 2009 the sizes of the gaps in most states were not statistically significant from the national gap, regardless of subject or grade.
For both grade levels, there was generally a larger percentage of White than Hispanic students who participated in the 2009 assessments at the national level. For example, the 2009 mathematics assessment student sample was 54 percent White and 22 percent Hispanic at grade 4 and 56 percent White and 21 percent Hispanic at grade 8. In 2009, at the state level for grade 4 the largest percentage of Hispanic students participating in NAEP was 51 percent in California, while the smallest percentage of Hispanic students participating in NAEP was 1 percent in Maine and Vermont.
Yes. In 2009, California, the District of Columbia, New Mexico, and Texas have a larger percentage of Hispanic students participating in NAEP than White students at both grades. Arizona has a larger percentage of Hispanic students participating in NAEP than White students at grade 4.
Achievement Gaps provides demographic information on:
Yes. The report also includes data on the Hispanic-White gaps by gender and NSLP eligibility, as well as ELL status at the national level only.
Yes. The data needed to calculate gaps are available through the NAEP Data Explorer, but not the differences in the gaps.
Several reports by secondary researchers have used different techniques and data sources to discuss gaps in student achievement. The methodology in this report is the same used for NAEP Report Cards and the Black-White Achievement Gaps report. This Achievement Gaps report defines gaps as the difference in scale scores between two student groups. It does not contain achievement level data, nor does it use sources of data other than NAEP.
When NCES began to administer Main NAEP at the state level, the first assessments were given on a trial basis, administered one grade at a time to lessen the cost and the administrative burden on the states. In addition, states participated on a voluntary basis. The state NAEP mathematics assessment was first administered in 1990, at grade 8 only, followed by the first state NAEP reading assessment in 1992, at grade 4 only. Grade 4 mathematics was added to the state assessment in 1992, but grade 8 reading was not administered at the state level until 1998.
Nationally, scores for Hispanic students have increased in 2009 since the early 1990s in both subjects and at both grades, but scores for White students have increased as well.
The gap narrowed for NSLP-eligible students at grade 8 in mathematics and for NSLP-eligible students at grades 4 and 8 in reading. The gap also narrowed between White and non-ELL Hispanic students at both grades in mathematics and reading.
However, since the early 1990s, the Hispanic-White achievement gap for public school students has not narrowed for fourth- or eighth-graders in either reading or mathematics, both nationally and for almost every state.