## I. Overview

Students designated as English language learners (ELL) tend to go to public schools that have low standardized test scores. However, these low levels of assessed proficiency are not solely attributable to poor achievement by ELL students. These same schools report poor achievement by other major student groups as well, and have a set of characteristics associated generally with poor standardized test performance—such as high student-teacher ratios, high student enrollments and high levels of students living in or near poverty. When ELL students are not isolated in these low-achieving schools, their gap in test score results is considerably narrower, according to a Pew Hispanic Center analysis of newly available standardized testing data for public schools in the five states with the largest numbers of ELL students.

About 4 million U.S. public school students received ELL services in the 2003–04 school year, accounting for 8% of all public school enrollment that year (NCES, 2006). Public schools in the states that are the focus of this report (Arizona, California, Florida, New York and Texas) educated about 70% of the nation’s ELL students.

Prior analyses of assessment data uniformly indicate that ELL students are much less likely than other students to score at or above proficient levels in both mathematics and reading/language arts. This report quantifies the extent of ELL concentration in low-achieving public schools and the degree to which this isolation is associated with the large achievement gap in mathematics between ELL students and other major student groups.

The new standardized test data show that in each of the five states examined in this report about 90% of the ELL students who took the state assessment test were educated in public schools that had at least a minimum threshold number of ELL students. ELL students tended to make up either a majority or substantial minority of the student populations of these schools. For example, in the California public schools in which ELL test-takers were concentrated, they constituted 45% of all test-takers. In the other California public schools (where the number of ELL students was below the minimum threshold), ELL test-takers were just 6% of all test-takers.

In all five states investigated and irrespective of grade levels ELL students were much less likely than white students to score at or above the state’s proficient level. However, when ELL students attended public schools with at least a minimum threshold number of white students, the gap between the math proficiency scores of white students and ELL students was considerably narrower, the Pew Hispanic Center analysis has found. This suggests that the lag in test score achievement of ELL students is attributable in part to the characteristics of the public schools they attend.

ELL students perform better on the state’s standardized math assessment test if they attend a public school with at least a minimum threshold number of white students. For example, among eighth-grade ELL students in Florida, about 30% score at or above the proficient level in math if they attend a middle school that has a minimum threshold number of white students. Among Florida ELL eighth-graders at middle schools that do not have a sufficient number of white eighth-grade students, only about 10% scored at or above the proficient level in math.

The relatively poor proficiency levels at public schools with high concentrations of ELL students is underscored by comparing the standardized test scores of white and black students who attend the schools in which ELL students are concentrated with the scores of white and black student who attend other public schools. In California, 75% of white third-grade students who attend public schools without the minimum threshold number of ELL students perform at or above the proficient level on the state’s mathematics assessment test, whereas just 67% of the white California third-graders who attend schools with the minimum threshold number of ELL students score at or above the proficient level.

The average proficiency rate in math for black third-graders who attend California public schools without the minimum threshold number of ELL third-grade students is 46%. In contrast, 34% of black third-grade students who attend California public schools with the minimum threshold number of ELL students score at or above the proficient level on the state’s mathematics assessment test.

Most of this report’s findings are based on analyses using three U.S. Department of Education databases. The analysis of mathematics performance on state-designed assessments across different types of public schools utilizes the new National Longitudinal School-Level State Assessment Score Database. The NLSLSASD maintains state standardized assessment test results for every public school in a state. Because the NLSLSASD is a school-level data set, we can identify for the first time which public schools tested English language learner students and thus measure at the state level the degree of concentration of ELL students in particular schools. Using the NLSLSASD’s standardized testing results by subgroup, the analysis illuminates the potential role of school isolation in student test score performance.^{1}

Previous Pew Hispanic Center analyses of standardized testing data for public schools revealed a large achievement gap between ELL students and other students in math and reading proficiency (Fry, How Far Behind in Math and Reading are English Language Learners?, Pew Hispanic Center, June 6, 2007), and that black and Hispanic students are increasingly isolated from white students in the public schools (Fry, The Changing Racial and Ethnic Composition of U.S. Public Schools, Pew Hispanic Center, Aug. 30, 2007). This report builds on those findings by illustrating that the educational isolation of ELL students is associated with the math proficiency gap between English language learners and other students. It also shows that white and black students who attend the public schools in which ELL students are concentrated are doing worse than their peers who attend public schools with few English language learner students.

Among the report’s other key findings:

- Nationally, the English language learner student population is expected to grow rapidly. The projected number of school-age children of immigrants will increase from 12.3 million in 2005 to 17.9 million in 2020, accounting for all the projected growth in the school-age population.
^{2}A significant portion of these children of immigrants will likely require ELL services. - In the five states with large ELL student populations, the proportion of ELL students scoring at or above the proficient level on the state mathematics test is often below the proportion of black students scoring at or above the proficient level. For example, in Texas 22% of ELL eighth-graders scored at or above the proficient level on the math assessment, compared with 44% of black eighth-graders.
- In both elementary grades and middle school grades in these states, ELL students are much less likely than white students to score at or above the proficient level in mathematics. The measured gaps are in the double-digits. For example, in Florida 45% of ELL third-graders scored at or above the proficient level on the math assessment, compared with 78% of white third-graders, yielding a white-to-ELL gap of 34 percentage points.
- ELL students who took the state mathematics assessment were heavily concentrated in the public schools that had to disclose publicly the English language learner testing results — that is, public schools with a minimum threshold number of ELL students taking the test. White test-takers and black test-takers were much less concentrated in the public schools reporting ELL testing outcomes. For example, in New York more than 90% of the fourth-grade ELL students taking the math test attended the 763 elementary schools that reported their test scores. The New York public schools that reported results for ELL fourth-graders educated less than 20% of white fourth-grade test-takers in the state and slightly more than half of black fourth-grade test-takers.
- In the five states with large ELL student populations, the public schools in which ELL test-takers are concentrated are much more likely to be central city schools.
- The public schools in which ELL test-takers are concentrated have a much higher enrollment, on average, than other public schools in the state.
- The middle schools in which ELL test-takers are concentrated have, on average, significantly higher student-to-teacher ratios than other public schools in the state.
- The public schools in which ELL test-takers are concentrated have, on average, a substantially greater proportion of students qualifying for free or reduced-price school lunches.
- The public schools in which English language learner students are concentrated are significantly more likely to be designated Title I schools. A Title I school has a student body with a large proportion of economically disadvantaged students and receives additional federal funding. For example, in Arizona 92% of the schools that reported test results for ELL students on the third-grade math assessment were eligible for Title I funds. Of the other Arizona elementary schools, half were Title I-eligible.

### A Note on Terminology

The terms “white” and “black” refer to the non-Hispanic components of those students.