Language-learning surprise

Immersion may pay off for kids acquiring English

Nancy Mitchell, Rocky Mountain News

Monday, September 24, 2007

A study looking at whether poor children in Denver do better in wealthier schools turned up an unexpected - and far more controversial - finding that is raising questions about the use of Spanish in classrooms.

The analysis by the Piton Foundation and the University of Colorado, to be released today, shows that students learning English make dramatically greater gains in wealthier schools than in poorer ones.

A key difference between the schools is that students in poorer schools typically learn English with help in their native language, generally Spanish.

In wealthier schools, students learning English are typically immersed in English.

"It raises more questions than it answers," Brad Jupp, senior academic adviser to Denver Public Schools Superintendent Michael Bennet, said of the study.

He also called the analysis "brand new stuff, cutting-edge work."

"The methods they're using are advanced," Jupp said, "and we're going to have to use these advanced methods to take closer looks at student performance everywhere in the city."

What they learn could affect the federal court order outlining how DPS must instruct its English language learners, who make up 20 to 40 percent of the 72,000 children in its schools.

That order requires schools serving 60 or more native Spanish speakers to have an English language acquisition, or ELA, program, where teachers use Spanish to transition students to English over three years.

Of Denver's 97 elementary schools, 54 have ELA programs, according to the study. And 53 of the 54 are high-poverty schools, where 61 percent or more of the children qualify for federal lunch aid.

"There's something going on in the ELA program that is not working for kids at all," said Alan Gottlieb, who commissioned the study as education program officer for Piton, a Denver nonprofit.

But Gottlieb and CU researcher Ed Wiley - along with DPS' -Jupp - caution against concluding that the use of Spanish is hindering the teaching of English.

"I don't want to say teaching kids in English is the only model," Gottlieb said. "Another possible answer is they have unqualified teachers in the ELA program teaching kids in Spanish and they're not learning anything. The other possibility is that the curriculum is so watered down . . .

"DPS needs to take a hard look, at least, at what's happening in the ELA program in elementary schools . . . because there's something very wrong."

In the analysis, Piton researcher Pam Buckley and Wiley, who is chairman of research and evaluation methodology at CU-Boulder's School of Education, analyzed the records of nearly 37,000 DPS students as they progressed from the third grade to the fifth.

The researchers looked at how the students performed on annual state reading and writing exams, given in English, over that time.

They found that low-income students learning English made virtually no progress on the exams in schools where most children - more than 60 percent - come from low-income families. In virtually all of these schools, children are in ELA programs.

In schools with poverty rates lower than 60 percent, low-income students learning English made greater progress.

And in the wealthiest schools, students learning English almost "catch up" to native English speakers in the poorest schools on state writing exams.

Still, Wiley said, "I'm not ready to conclude one model is good or not good for everybody. As schools differ tremendously, so do kids. The 'why' question isn't something we can answer yet."

The study did not consider differences in the use of native language instruction in teaching English. For example, two schools with higher poverty levels - Bryant-Webster and Fairmont - use an approach called dual language, where Spanish speakers learn English and English speakers learn Spanish.

It also did not look at whether test results differed based on the student's native language. While Spanish is the most common, spoken by nearly 90 percent of English language learners in DPS, Denver students speak more than 115 other languages.

Wiley and Gottlieb said further analysis will be done but noted the original focus of the research was student achievement by income, not language.

Gottlieb, now a vice president with the Public Education and Business Coalition, said the study does support previous research findings that low-income students do better in schools with fewer poor children.

It also found that students from wealthier families who attend high-poverty schools appear to be "entirely unaffected" by the poverty level of the school.

So middle-class families should not worry that sending their kids to poorer schools will "slow down" their own student's rate of learning, the study found.

Income, language and achievement

A study by the Piton Foundation and the University of Colorado analyzed how nearly 37,000 elementary students in Denver Public Schools performed on state tests as they progressed from the third through fifth grade between 2003-04 and 2005-06. Among the findings:

BY INCOME

• Low-income students start out with lower scores on third-grade reading and writing exams than their wealthier classmates, regardless of whether they're attending a poorer or more affluent school.

• Over time, low-income students improve faster on state reading and writing exams if they're enrolled in wealthier schools, where 60 percent or fewer students receive federal lunch aid.

• Achievement gaps between low-income students and their more affluent classmates persisted at all schools, regardless of overall poverty levels.

BY LANGUAGE

• Low-income students learning English make greater gains in wealthier schools. These students gained, on average, less than one point per year on state writing exams in schools where 61 percent or more of their classmates received federal lunch aid. That compares with gains, on average, of 26 points per year in low-poverty schools, where 30 percent or fewer students receive lunch aid.

• By the fifth grade, students learning English in low-poverty schools perform nearly as well as native English speakers in high-poverty schools on state writing exams. Gains are similar, though not as dramatic, in reading.

• To learn more, go to headfirstcolorado.org. The study is being released today in the online education magazine.

mitchelln@RockyMountainNews.com or 303-954-5245