« School Districts That Offer Dual-Language Classes Through High School | Main | New York Offers Curriculum Advice »

Arizona Spells Out "Research-Based" Models for English Immersion


Arizona policymakers are using a buzz phrase popular in education circles in saying the models for structured English immersion that school districts must implement this fall are "research-based." But a document released by the Arizona Department of Education citing research to back those models shows that, for some aspects of the models, the research base is scant. The document acknowledges that high-quality research in general about instruction for English-language learners is limited.

The state is calling for English-language learners, in their first year in U.S. schools, to receive four hours of instruction each day in a separate block of time for English-language development. But the document cites only one study that shows it's more effective for English-language learners to have a fixed block of time each day in which they focus on developing their English-language skills than not to have it. What's more, the document doesn't cite any research that tells how long that block of time should be. The document also doesn't cite any research telling why it makes sense for the state to favor English immersion over bilingual methods, which the state has done since Arizona voters passed a ballot measure in 2000 that curtailed bilingual education.

In a telephone interview last week, Tom Horne, the Arizona superintendent of public instruction, told me that a study by Joseph M. Guzman published in Education Next in the fall of 2002 supports the use of structured English immersion with ELLs. (One expert in the field says that study's use of data doesn't provide valid inferences on whether bilingual education or English immersion is better. See here.)

Arizona's requirement that school districts provide four hours of English-language development each day is contained in a law passed in the 2005-2006 school year, but most school districts didn't implement that law, according to Mr. Horne. He said some are providing ELLs with "as little as a half-hour" of such instruction per day.

But now, he said, a task force for ELLs has spelled out what those four hours of English instruction should look like. A June 21 memo to school district superintendents says ELLs at the two lowest levels of English proficiency should receive 45 minutes of oral English and conversation; an hour each of grammar, reading, and vocabulary; and 15 minutes of pre-writing instruction.

The Arizona Republic ran an article on July 14 telling more about the implementation of the policy.


One concern I have is the 4-hours of ELD. Is this a separate 4 hour block of time, or can time spent in other classroom such as "Reading" be counted?

Another concern: the definition of what ELD is quite restrictive: phonology, morphology,syntax, etc. It seems to me that in order for these students to learn more effectively, the content should be more meaning oriented and made comprehensible. The so-called model seems to be heavely focused on grammar and form.
Thank you,
Gino Martell

It's too bad Tom Horne is still relying on the Guzman study, published in a conservative "think tank" magazine rather than a research journal. The advantages Guzman reported for English learners taught through English-only approaches were modest, but it is important to point out that the conclusions of the report are incorrect, following from a significant flaw in the research design. In Guzman’s report, the group of students identified as having been taught through ESL was defined as those students who indicated in a survey that they had been taught using special instruction in English for non-English speakers but had not received "foreign language instruction in a non-language topic." Guzman identified bilingual education students as those who indicated that they had received "foreign language instruction in a non-language topic" but not special instruction in English for non-English speakers. However, as is well known among educators, a correctly designed bilingual education program provides both native language support and special instruction in English for non-English speakers, not one or the other. Thus, Guzman’s group of “bilingual education” participants was ill-defined. In a brief footnote, Guzman indicates that including students who reported both treatments (as one would expect in a correctly designed bilingual education program) in the statistical analysis increased the effect of bilingual instruction on years of education completed, but no details were given.

But in search of good research on whether immersion is best, one should turn to syntheses of research -- and we have seen five independently conducted syntheses prepared in the last couple of years, all favoring bilingual education. Even Christine Rossell, who has long argued against the approach, makes modest but clear favorable comments about bilingual education in her recent "meta-murky" paper.

Tom Horne is cherry-picking data, the way the tabacco industry once did to try to mount arguments that smoking does not cause cancer.

But I'm afraid Horne's political war on kids who want to learn English and succeed in school is only just beginning. His Task Force is crafting measures which will make Horne's implementation of Prop 203 look like a cake walk.

Comments are now closed for this post.

Follow This Blog


Most Viewed on Education Week



Recent Comments