English-language learners do better academically over the long term if they participate in special programs to learn English at the start of their school careers, rather than attend only mainstream classes, according to one of the largest longitudinal studies of such students ever conducted.
That conclusion comes from a study of English-language learners released last month by Wayne P. Thomas and Virginia P. Collier, researchers at George Mason University in Fairfax, Va.
“We now know that very unhappy things happen when you just submerse English-language learners in a regular classroom—when the teacher has no special training and no special things are done for them,” Mr. Thomas said.
The authors say the study also confirms what they found in earlier research: Students who take bilingual education classes do much better on standardized tests after entering mainstream classes than students who take English-only classes.
The study reports on student records from 1982 to 2000 provided by five school districts, including the 208,000-student Houston district, and is part of an ongoing, federally financed study of programs for English-language learners in 16 school districts. (“Learning Gap Linked to LEP Instruction,” April 25, 2001.)
Mr. Thomas and Ms. Collier stressed in an interview that some bilingual education programs are much more effective than others, something they say is often lost in the national debate about how best to teach students English.
For instance, they’ve found that transitional bilingual education, in which students are taught some subjects in their native languages with the expectation that they will move as quickly as possible into mainstream classes, is only slightly more effective than English-only instruction.
“The issue is not just bilingual versus English-only,” said Mr. Thomas. “How effective you are depends on what kind of bilingual and English-only programs you’re talking about.”
The study found that long-term bilingual education programs that develop strong literacy both in students’ native languages and in English—in contrast with short-term programs that emphasize learning English as quickly as possible—are the most effective kinds of programs. In fact, the study says they’re the only kinds of programs that fully close the achievement gap between English-language learners and native English-speakers over the long term.
The researchers give “90-10" two-way bilingual programs, begun in 1996 in Houston schools, as an example. Those programs strive to teach both native speakers of English and Spanish attending the same classes academic content in both languages. The students initially receive 90 percent of instruction in Spanish and 10 percent in English.
The amount of English used for instruction increases with each grade. English- language learners in such classes scored at the 51st percentile in reading on standardized tests at the end of 5th grade.
In contrast, Houston students who participated in “ESL content” programs, in which teachers use English-as-a- second-language techniques to teach core academic courses, performed at the 32nd percentile on standardized tests in reading in the 11th grade.
And English-language learners in Houston public schools whose parents had chosen to place them only in mainstream classes scored on average only at the 12th percentile on standardized reading tests in 11th grade.
Many of the findings of the researchers are longitudinal, following the same students over time.
But the particular comparisons regarding Houston’s programs are not based on the achievement of the same students over time. Rather, they are based on different students enrolled in the same kinds of programs.
Kenji Hakuta, an education professor at Stanford University, said the study was important for the field because of its scale and ability to follow students over a long period of time. The much more common short-term studies in the field “are limited as to what they tell us,” he said.
Mr. Hakuta pointed out that Mr. Thomas and Ms. Collier used a research approach that differs from the methodology of some other large-scale studies, in that they deeply involved district personnel as observers and collectors of data. Such an approach provides them with excellent access to student information, he said, but also causes the study to lose some objectivity.