Schools that report low achievement for English-language learners also report low test scores for white and African-American students, and share characteristics associated with poor performance on standardized tests, according to a study released today by the Pew Hispanic Center.
Those characteristics include high pupil-teacher ratios, large enrollments, and high levels of students who are eligible for free and reduced-priced lunches, a reflection of families鈥� socioeconomic backgrounds.
But the study by Richard Fry, a senior research associate at the Washington-based center, also found that when English-learners are not isolated in low-achieving schools, the gap between their test scores and those of other students narrows significantly.
That finding indicates that the gap between English-learners and other students 鈥渋sn鈥檛 so much because of the characteristics of the student and family, but also because of the school they are attending,鈥� Mr. Fry said in an interview.
The study, analyzes students鈥� test scores in mathematics drawn from three U.S. Department of Education databases. It focuses on the five states with the most English-language learners: Arizona, California, Florida, New York, and Texas.
鈥淭he Pew results make clear that linguistic isolation goes hand in glove with economic, ethnic, and racial isolation, which are in turn tied to worse school outcomes,鈥� Michael Fix, the vice president and director of studies at the Washington-based Migration Policy Institute, said in an e-mail message.
鈥淔or me, the results also reinforce the fact that one critical test of the success of the No Child Left Behind Act will be whether we see any improvement in student outcomes in these low-income, high-ELL schools,鈥� he said.
Mr. Fix also said the Pew results are consistent with findings by the Urban Institute that elementary schools with large ELL populations are more urban, have larger classes, and an enrollment that is more heavily minority than that of other schools.
鈥淭he teachers in these high-ELL elementary schools ... had less experience, less academic preparation, and were less likely to be certified than their counterparts in other schools,鈥� he said.
Patterns Emerge
The study illustrates鈥攂ut does not attempt to explain鈥攕everal patterns that emerge from the test-score data based on the mix of students. For example, the researchers found that English-learners who attend schools with even a small number of white students have higher math-test scores than do ELLs who are almost entirely isolated from such white students.
And black students and white students who attend schools with even a small number of English-learners do worse on standardized math tests than black and white students who are almost entirely isolated from ELL students.
The study鈥檚 results are based on test scores of ELLs of all different racial and ethnic backgrounds鈥攏ot just Hispanic鈥攃ompared with the test scores of non-Hispanic white and non-Hispanic black students. Mr. Fry said that U.S. Census Bureau data shows that the overlap between ELLs and non-Hispanic whites and non-Hispanic blacks is minimal.
In comparing the performance of each group of students, Mr. Fry used the minimum thresholds set by the states for reporting those test scores for student groups under the federal No Child Left Behind law. For example, in Arizona and Florida, schools report scores for ELLs, white students, or other student groups if they have 10 test-takers per grade in that group. California sets the threshold at 11 test-takers. New York and Texas set it at five test-takers.
鈥淯sing this very low threshold, you still get pretty strong differences,鈥� Mr. Fry said.
For example, the study found that 30 percent of ELLs in 8th grade in Florida score at or above the proficient level in math if they attend a middle school that has a minimum threshold number of white students. But at Florida middle schools that don鈥檛 meet the minimum threshold for white students, only about 10 percent of ELL 8th graders score at or above the proficient level in math.
Mr. Fry said he found it interesting to learn from his analysis that white and black students attending schools with a minimum threshold of English-learners have lower math scores on average than do those groups of students who attend schools with too few English-learners to meet their states鈥� thresholds.
