In the first randomized-assignment study in which English-language learners were followed for as long as five years, researchers have found that Spanish-speaking children learn to read English equally well regardless of whether they are taught primarily in English or in both English and their native language.
The findings from the Center for Research and Reform in Education at Johns Hopkins University may take some fuel out of the fire in the national debate over which is best for teaching children from immigrant families to read: English immersion or bilingual education.
“People have been fighting for years and years about the language of instruction, thinking that it was either terribly important to teach in English the whole time or terribly important to teach in Spanish and then English. Both groups were wrong,” said Robert E. Slavin, the director of the center, in Baltimore, and one of the researchers.
The study compares the academic performance in reading of several hundred ELLs who entered school in kindergarten and were randomly assigned to either “structured English immersion”—in which teachers give daily reading lessons only in English, except for occasional Spanish explanations—and transitional bilingual education, in which students are taught reading initially only in Spanish, with a transition to English starting as early as 1st grade and completed by 3rd grade.
The researchers followed three cohorts of ELLs who entered kindergarten in 2004, 2005, and 2006. Only the group that entered kindergarten in 2004 was followed for a full five years, through 4th grade.
For both educational methods, teachers used Success For All, a reading program that Mr. Slavin developed. Success For All is essentially the same whether in English or Spanish.
The researchers plan to present their findings later this spring at the annual meeting of the American Educational Research Association and have submitted the study for publication in a journal, which they declined to name until acceptance is confirmed. The U.S. Department of Education’s Institute of Education Sciences underwrote the study with a $5.4 million grant.
No Extremes
Russell Gersten, a researcher who has tended to favor English immersion, and Claude Goldenberg, a researcher who has tended to back bilingual education, both characterized the study as highly important, especially because they believe it’s the first randomized-assignment study comparing the methods since the 1970s.
Mr. Gersten, a professor emeritus of education at the University of Oregon, interpreted the John Hopkins findings as showing that ELLs could make solid progress in English reading via either approach, and get much closer to, but not reach, parity with native speakers by 4th grade.
The study found that students in bilingual education had an edge in Spanish reading skills over students in English immersion in the early grades, while the reverse was true for English reading skills. But differences evened out by the 4th grade, with students scoring about the same in Spanish and in English, the researchers reported.
Thus, as Mr. Gersten sees it, the study doesn’t support the views of some bilingual education advocates that ELLs taught in their native languages transfer all those skills to English easily and then make phenomenal progress in English. Neither does the study show, he said, that students in English-immersion programs get a huge, long-lasting boost for having been taught only in English, as some advocates of that method contend.
The study could help neutralize some of the acrimony of the bilingual education debate, according to both Mr. Goldenberg, an education professor at Stanford University, and Mr. Gersten.
“It doesn’t support either extreme position in the bilingual education debate, which I think most responsible people have retreated from,” Mr. Goldenberg said.
He said he’d like to see more analysis of the data and a clearer explanation of its methodology. The researchers, for example, didn’t break out data for each of the three cohorts of students they followed, which Mr. Goldenberg contends obscures the picture of the academic progress of each group, consequently not taking full advantage of the longitudinal aspect of the research.
Programs Dropped
The study involved six elementary schools, one each in California, Colorado, Illinois, Minnesota, New Mexico, and Texas. Interestingly, two of the six elementary schools in the study have dropped their bilingual education programs.
Jennifer Geissler, the literacy coach for Pioneer Charter School in Denver, said her school switched to English immersion for all grades during the 2007-08 school year, though it was part of the study through the end of last school year.
Because of a federal consent decree, regular public schools in Denver are required to provide bilingual education. Ms. Geissler said her school, where almost all students are Hispanic, wanted to distinguish itself in its educational approach from nearby schools.
Parents seemed to prefer an English-immersion approach, she said.
Aware of the debate about the two methods, Ms. Geissler said that some educators might suggest that her school is doing Spanish-speaking students a disservice by not putting more emphasis on their native language. Still, she added, “we believed from the beginning it was about good instruction, using effective strategies, and using data to analyze kids and what they need.”
Also this month, Education and Finance Policy published a study by an economist from the University of Kentucky comparing bilingual education and English-immersion programs in California that had similar but less conclusive findings than the Johns Hopkins study. The study didn’t use randomized assignment but it had a very large sample of students.
Christopher Jepsen, the economist at the University of Kentucky who conducted the study, examined the scores on California’s English-language proficiency test in 2003 and 2004 for nearly half a million English-learners in grades 1-5. He found that such students in 1st and 2nd grade participating in bilingual education programs scored substantially lower in English proficiency than such students in English-immersion programs. But in grades 3-5, the English-proficiency scores were very similar, regardless of whether students were in bilingual education or English immersion.
He said he couldn’t be as conclusive as the Johns Hopkins researchers because he didn’t follow any of the same students over time, as they had. But based on what he found, Mr. Jepsen said, he doesn’t think “we should be that worried” about whether students receive bilingual education or English immersion.
