Commissioner, National Center for Education Statistics
National Assessment of Educational Progress
Achievement Gaps: How Hispanic and White Students in Public Schools Perform in Mathematics and Reading on the National Assessment of Educational Progress
June 23, 2011
Good afternoon. Today we’re releasing our report on achievement gaps—the differences in performance of Hispanic and White students as they have changed over time, using national and state data for mathematics and reading from the National Assessment of Educational Progress, or NAEP. As this report will show, scores for both Hispanic and White students have improved over time, but for the most part the gaps remain significant, at both the state and national levels.
While many reports show that, on average, White students have higher scores than Hispanic students, today’s report is the first to focus on the Hispanic-White achievement gap at the state level. This report provides an accessible summary of gap data for every state—how the gaps stood in 2009 and how they’ve changed over time, based on the common yardstick of the NAEP assessments.
The report describes the relative populations of both Hispanic and White students for each state, how the scores for Hispanic and White students in each state compare to the national averages for those students, and how the size of a state’s gap compares to the national gap. At the national level, we also have separate comparisons of low- and higher-income students and for English Language Learners compared with students who are proficient in English.
In some states, such as California and New Mexico, Hispanic students, at least at grade 4, constitute half of the student population, while in other states the percentage of Hispanic students is less than 5 percent. By showing the size of the gaps for each state, and trends in the size of the gaps over time, this report gives a more complete portrayal of student performance than can be obtained from national results alone.
We have national and state results in this report, for both reading and mathematics, at grades 4 and 8. Our most recent data are for 2009, in all cases. At grade 4, our state results in both mathematics and reading go back to 1992. At grade 8, the timeline is a bit more complicated. The grade 8 state mathematics assessment was first administered in 1990, while the grade 8 state reading assessment did not begin until 1998. The state NAEP assessments began as trial assessments and were phased in over a period of years. Participation in state NAEP was voluntary in the first assessments, so we can’t make comparisons back to the earliest years for all the states.
Nationally, we have mathematics results since 1990 for both grades, while reading results go back to 1992. In all of these cases, we can compare state results to the national results.
Between 43 and 47 states had sufficient Hispanic and White student populations in 2009 (depending on the subject and grade) to allow for comparisons. Our state-level results actually include 52 jurisdictions, because we assess students in the District of Columbia and the Department of Defense Schools along with the 50 states. In earlier assessments the Hispanic student sample in many states was too small to allow for the reporting of reliable results, so comparisons to the earliest assessments are not always possible, even when states did participate in the assessment.
Nationwide, 21 percent of fourth-graders were Hispanic in 2009. However, this percentage varies widely across the country. The states with a 4th-grade Hispanic population ranging from zero to 5 percent of students are concentrated in the far north, near the Canadian border and along a sort of Mississippi/Ohio River corridor stretching from Louisiana to Ohio. Those states with a 6 to 10 percent Hispanic fourth-grade population concentration are largely located along the East Coast and in the Midwest. In six states—Idaho, Utah, Nebraska, Kansas, Delaware, and the District of Columbia—the percentage was between 11 and 15 percent. In 10 states, counting the Department of Defense school system, the Hispanic fourth-grade population was 16 to 25 percent of 4th-grade students. These included Washington and Oregon in the Northwest, and such large states as Florida, Illinois, and New York. The six states with a grade 4 population that is at least 26 percent are all located in the Southwest. In California and Texas, the two largest states, the percentages were 51 and 49 percent respectively.
2009 Hispanic--White Achievement Gap
The gaps between average NAEP scores at the national level for mathematics and reading at grades 4 and 8 in 2009 ranged from 21 to 26 points. The score gaps between Hispanic students who were English Language Learners (ELLs) and Hispanic students who are not ELL students ranged from 19 to 39 points, comparable to or larger than the overall Hispanic-White gaps. About 35 percent of Hispanic students were ELL students at grade 4 in 2009, and about 20 percent were ELL students at grade 8.
A third type of score gap was obtained by subtracting the score for non-ELL Hispanic students from the score for White students, who are almost entirely non-ELL. Here the gaps ranged from 14 to 19 points, smaller than the overall gaps.
Comparing the sizes of these three types of gaps indicates the extent to which the overall Hispanic-White gap is influenced by the relatively low performance, on average, of Hispanic ELL students, coupled with the size of the Hispanic ELL population. For example, at grade 8 the gaps between non-ELL Hispanic students and ELL Hispanic students was 34 points in mathematics, and 39 points in reading, indicating the relationship between ELL status and relatively low performance on these assessments. For both reading and mathematics, the gap between non-ELL Hispanic and ELL Hispanic students is larger at grade 8 than at grade 4.
National Trends in Gaps and Scores
Now I’ll describe national trends over time for both average scores and score gaps for Hispanic and White students. We have overall results, along with results by eligibility for the National School Lunch Program, which is a measure of student family income. (Students who are eligible for program come from low-income families, while those who are not eligible come from higher income families. For both mathematics and reading, the overall results go back to the early 90s, while the school lunch eligibility results only go back to 2003.)
I’ll begin with mathematics, and discuss the overall gaps between Hispanic and White students, the gaps between Hispanic and White students eligible for free or reduced-price lunches under the National School Lunch Program, and between those not eligible for the program. Students who are eligible for the program come from low-income families, while those who are not eligible come from higher income families. For both mathematics and reading, the overall results go back to the early 90s, while the school lunch eligibility results only go back to 2003.
For grade 4 mathematics, none of the three gaps narrowed. At the same time, scores for all groups increased.
We see almost the same pattern for grade 8 mathematics, with one exception. The gap between eligible Hispanic and White students narrowed. Scores for both groups increased, but the increase for Hispanic students was larger.
For reading at grade 4, we see exactly the same pattern as at grade 8 mathematics—increases for all groups, and a narrowing of the gap for eligible Hispanic and White students, due to a larger increase in the score for Hispanic students.
In grade 8 reading, there is a similar pattern—a narrowing of the gap only for eligible students. This time, however, the narrowing was due to an increase in scores for Hispanic students coupled with no significant change in the average scores for White students. For all other groups, scores increased.
The pattern is the same for mathematics and reading, whether grade 4 or grade 8—higher scores in almost all cases, and gap narrowing only for students from lower-income families. As these findings suggest, improved performance for Hispanic students does not necessarily result in narrower gaps, if scores for White students improve as well.
An example of the decline in the gap between White and Hispanic students eligible for free or reduced price lunches can be found in grade 8 mathematics, where the gap in scores declined from 17 points in 2003 to 13 points in 2009. Scores were higher for both groups, but the increase for Hispanic students was larger. In grade 8 reading, we also see a narrowing of the gap, from 17 points to 15 points. Again, scores for both groups increased.
In 2009, 72 percent of Hispanics, compared to 24 percent of White students, were eligible for the National School Lunch Program, based on data from the grade 8 reading assessment. For both groups, eligible students scored lower than not eligible students. Hispanic students are more likely than White students to be eligible, which contributes to larger overall gaps.
Reading gaps vary for 8th-graders by English language learner status
Now we’re going to look at the same score gaps in a different way, showing how they have changed over time. I’ll discuss the results for grade 8 reading here, but please note that this information for grade 4 reading and grades 4 and 8 mathematics are in the full report.
First I’ll compare the reading scores of all grade 8 White students with the scores of Hispanic students who were not ELL from 1998 to 2009. The gap fell from 22 points in 1998 to 15 points in 2009. Scores for both groups increased, but the increase for non-ELL Hispanic students was larger. As we saw earlier, the overall national grade 8 reading gap was 24 points.
A similar comparison between non-ELL Hispanic students and ELL Hispanic students shows that the gap increased from 32 points in 1998 to 39 points in 2009. In fact, the 2009 gap was larger than any previous gap. The average score for non-ELL Hispanic rose 10 points from 1998 to 2009, while the score for ELL Hispanic students did not change significantly.
In 2009, 6 percent of all grade 8 students were identified as ELL by their schools. Twenty percent of the Hispanic students at grade 8 who were assessed in 2009 were ELL students, while the percentage of White ELL students was less than one half of one percent. In 1998, 2 percent of students nationally were ELL. Fifteen percent of Hispanic students were ELL in 1998, while the percentage of White students who were ELL rounded to zero.
2009 State Gaps Compared to the Nation
Now we’re going to look at the Hispanic-White gap at the state level, comparing the size of the state gaps to the national gap.
In mathematics at grade 4, the national gap was 21 points. Eleven states had a gap that was smaller than the 21-point national gap, while in 30 states the difference between the national and state gap was not statistically significant. In the remaining six states where a comparison could be made, the gap was larger.
In grade 8 mathematics, the gap was 26 points nationally. In 15 states, the gap was smaller. In 24, it was not significantly different. In six states, the gap was larger.
In grade 4 reading, the national gap was 25 points. Thirteen states had a smaller gap, while in 29 states there was no significant difference, and in 6 the gap was larger.
In a few states with small percentages of Hispanic students, the difference in scores between Hispanic and White students was not statistically significant for a given grade and subject, and thus there was no measurable gap. This was true for Missouri in grade 4 mathematics, Hawaii and Missouri in grade 8 mathematics, Mississippi and Ohio in grade 4 reading, and Kentucky in grade 8 reading.
States with largest Hispanic population compared to Nation in grade 4 mathematics in 2009
The five states with the largest Hispanic population, listed in alphabetical order, are California, Florida, Illinois, New York, and Texas. Together, these states have about 70 percent of the nation’s Hispanic student population. For illustration, I’ll discuss results for grade 4 mathematics and grade 8 reading. Results for grade 8 mathematics and grade 4 reading are in the full report.
In mathematics at grade 4, California’s gap was larger than the nation’s. Hispanic students in California had an average score that was lower than the score for Hispanic students nationally while White students in California had a score that was comparable to that of their peers nationally. In 2009, 51 percent of California’s fourth-graders were Hispanic.
In contrast, the gaps in both Florida and New York were smaller than the national gap. In Florida both Hispanic and White students had higher scores than their peers nationally, while in New York only Hispanic students outscored their peers nationally. The percentage of Hispanic fourth-graders in both Florida and New York was 21 percent in 2009.
In Texas, both Hispanic and White students had higher scores than the national averages for these groups, while the gap was comparable to the national gap. In Illinois, the scores of both groups and the gap were comparable to their national counterparts. The percentages of Hispanic fourth-graders in Texas and Illinois were 49 and 16 percent, respectively.
In reading at grade 8, the Florida gap was smaller than the national gap. The average score for Florida’s Hispanic students was higher than the national average for Hispanic students, while the score for White students was comparable to the national average for their peers.
The gaps for the other four states were comparable to the national gap. In Florida, Illinois, and Texas, scores for Hispanic students were higher than the national average for Hispanic students, while in California it was lower. New York was the only state of the five where White students scored above the national average for White students.
In the full report we have complete information for all the states, not just the five I’ve just discussed, for both subjects and both grades. There are individual graphs for each state for each subject and grade that give average scores for Hispanic and White students, as well as score gaps. The graphs contain trend lines, making it possible to compare results between prior assessments and 2009.
Ways Gaps Can Narrow
In viewing these state trend graphs, it is important to note that gaps can narrow or widen in various ways, depending on how the two groups being compared perform over time. For example, we have identified six ways in which gaps can narrow. Some gaps have narrowed due to score increases for Hispanic students either to a greater extent than those for White students, or in the face of unchanged White scores over time. Another possibility is that a gap can narrow even though scores for neither group change significantly. Gaps can also narrow due to declines in scores for White students, while Hispanic scores either increase or remain unchanged. Because scores have generally been increasing for both Hispanic and White students, we don’t see many gap narrowings of this sort.
State Results for Hispanic-White GapsNow we’re going to look at changes in the performance gaps for Hispanic and White students at the state level, from 2007 to 2009, beginning with mathematics.
State Mathematics 2007-2009
In 41 states for which we have reliable data for both 2007 and 2009, scores for neither group changed, nor were there changes in the size of the gaps.
Scores did change in four states, but not enough to affect the size of the gap. In one state, the District of Columbia, scores for both Hispanic and White students increased. In Kentucky, Hispanic students’ scores did not change, while scores for White students increased. In Wyoming and New York, Hispanic students’ scores did not change, while scores for White students decreased.
In two states, the gap did change from 2007 to 2009. In Texas, Hispanic students’ scores decreased, while scores for White students did not change, causing the gap to widen. In Rhode Island, Hispanic students’ scores did not change, while scores for White students increased. Again, the gap widened.
At grade 8, the gap narrowed in three states—Arkansas, Delaware, and Missouri. The gap did not widen for any state. In 40 states, the gap did not change, and we did not have reportable results for both groups of students in 9 states.
State Reading 2007-2009
As with mathematics at grade 4, in reading at this grade, in 41 states for which we have reliable data for both 2007 and 2009, scores for neither group changed, and there were also no changes in the size of the gaps.
In one state, Alaska, scores for Hispanic and White students did not change significantly, but the gap narrowed. This was the only gap change.
In Rhode Island, Hispanic scores did not change, while White students’ scores increased. In Maryland and Florida, Hispanic students’ scores increased, while scores for White students did not. In Wyoming, scores for Hispanic students did not change significantly, while scores for White students decreased.
At grade 8, there were three states where the gap narrowed, Rhode Island, South Carolina, and Wyoming. These were the only gap changes at grade 8.
Summarizing National and State Results
Now we’ll look at a summary of national and state results, highlighting the kinds of information contained in the full report.
National scores increased for both Hispanic and White students in both subjects and grades since the 1990s. However, the gap remained unchanged.
Many states also showed improved scores. In grade 4 mathematics, all 21 states for which we have reportable results for both Hispanic and White students showed higher scores for both groups since 1992. At grade 8, it was 14 of 14. In reading, it was 11 of 21 states at grade 4, while at grade 8 it was one state, Wyoming, out of 22 states.
Making comparisons from 2007 to 2009, gaps narrowed at grade 8 in three states for mathematics— Arkansas, Delaware, and Missouri—and three states for reading— Rhode Island, South Carolina, and Wyoming.
At grade 4, the gap narrowed in reading only, in one state—Alaska. In mathematics, the gap widened in two states—Texas and Rhode Island.
At both grades and for both subjects, non-ELL Hispanic students had higher scores than ELL students and that the gap separating non-ELL Hispanic students from White students was smaller than the overall Hispanic-White gap.
For low-income students, both Hispanic and White, scores improved at both grades and for both subjects and gaps narrowed as well, except for grade 4 mathematics.
This concludes my presentation. There is much more information in the full gap report itself, which is available on the web. In conclusion, I would like to offer my sincere thanks to the students and schools who participated in NAEP, whose cooperation made this report possible.