L.A. SCHOOL REFORM GETS MIXED REVIEWS : SOME ETHNIC GROUPS NOT MAKING MAJOR GAINS UNDER LEARN.

Byline: Kimberly Kindy Daily News Staff Writer

An analysis of the LEARN program provides mixed news about whether the reforms are reaching the classroom, along with troubling signs that academic performance of English-speaking Latino students is falling at some LEARN schools.

Board of Education member David Tokofsky said the analysis indicates that the Los Angeles Educational Alliance for Restructuring Now program needs better tuning to meet the needs of students in the predominantly Latino Los Angeles Unified School District.

``It shows we are not reaching Latino students at the rate we are reaching other students,'' Tokofsky said. ``It shows they need to make changes and improve the program for these kids.''

The report, conducted by the Evaluation and Training Institute, tracked standardized test scores at schools that joined the LEARN program in 1993-94, the first year of the program, and compared them to scores at non-LEARN schools.

The results indicate that, overall, the LEARN schools improved at a higher rate than at non-LEARN schools. But at individual schools - and among students from particular ethnic groups - the rate of improvement varied significantly.

The study also analyzed the rate at which students transitioned from bilingual to English-only classrooms, at both LEARN and non-LEARN schools, and the overall attendance rate at LEARN and non-LEARN schools.

In each of those categories there was no appreciable difference between LEARN and non-LEARN schools.

The analysis of test scores focused primarily on changes at 27 of the 33 schools that joined LEARN in 1993-1994.

Half of those schools reported improvement of 1 percent to 30 percent in scores on the Comprehensive Test of Basic Skills.

Those results were described as ``remarkable'' by the president of Evaluation and Training Institute, which has evaluated the effectiveness of over 500 academic programs.

``We have never seen signs of changes this early in a program,'' said Clare Rose, who also served as the senior project director of the study. ``This is a first in our history.''

Superintendent Sid Thompson, however, acknowledged that while half the schools showed improvement, the other half either dropped or showed no gains.

``It's a case of whether the glass is half-full or half-empty,'' he said.

Also of concern were the demographic breakdowns, which showed the greatest improvements among pupils of African-American and Asian descent, and the least among white students and English-speaking Latinos.

For example, among the LEARN schools reading scores were unchanged, language-arts scores dropped 1 percent and math scores dropped 2 percent among English-speaking Latino pupils.

Among white students at the LEARN schools, language-arts scores were unchanged and reading and math scores were each up 1 percent.

Among African-American students at LEARN schools, reading scores were up 6 percent, language arts scores were up 8 percent, and math scores were up 9 percent.

Among Asian students at LEARN schools, reading scores were up 3 percent, math scores were up 5 percent, and language arts scores were up 8 percent.

Thompson said he has asked ETI to conduct further study into why that's the case.

``We need to analyze it,'' he said. ``The Hispanic scores - it looks flat or down. We need an answer for these Hispanic youngsters.''

Some district officials believe that English-speaking Latino students are not scoring high on the tests because many of them have just transitioned out of Spanish-speaking classrooms and still face a language barrier.

That assumption is buttressed by the ETI study, which showed that Spanish-speaking Latino students showed improvement on the Spanish-language version of the tests.

Another factor noted Thursday is that ETI believes the first-year LEARN schools fared better not only because they have been in the program longer, but also because they have fewer Latino students.

``The phase I schools are not representative . . . They have a smaller Latino population, smaller than the district average,'' said program manager Michael Butler.

Elisa Hernandez, an attorney for the Mexican American Legal Defense and Educational Fund, suggested that the results would rekindle debate over whether the school district, whose student population is 68 percent Latino, should have a Latino superintendent.

Thompson has announced that he will retire when his contract runs out next year, and the school board has begun a search for his replacement.

``They need someone who is sensitive to the needs of Latino students,'' Hernandez said. ``For them to not take into account the benefits of a Latino candidate would be a mistake.''

Leticia Quezada, chairwoman of a citizens committee that is advising the board on the superintendent search, agreed that the district should take the Latino students into account when selecting a new superintendent.

``I have asked that there be criteria requiring that the next superintendent be bilingual,'' she said. ``We've passed the threshold and have arrived at the point where we can say the leader ought to be bilingual.''

Nearly 40 percent of the district's kindergarten through 12th grade campuses have joined the LEARN program, and all schools are expected to be folded into the reform by 1999.

Under LEARN, decision-making authority is transferred from the central office to the local school sites. School principals and elected teacher leaders are given extensive management training.

ETI is working on a case study, expected to be completed in August, of 19 LEARN schools to pinpoint the common factors at schools that have shown the most improvement under LEARN.

Thompson said he wants to learn why some schools are experiencing steady increases or jumps of 10 percentage points or higher in test scores.

``Two or three percent doesn't really mean much unless there is steady growth shown over time,'' he said.

The district also faces a problem in finding an appropriate testing or analysis mechanism to assess student performance.

Testing experts have advised the district to do away with the multiple-choice CTBS test because its information for comparison of how students perform at the national level is almost 10 years old and because it doesn't test a student's ability to apply knowledge, as an essay test would.

The district is planning to implement a new test, which means that drawing comparisons to previous years will be difficult, if not impossible.

``We're going to have a problem with using different tests, and people are probably going to ask, `Are we are cutting it out so we can blow smoke over the issue?' '' Thompson said.

He said he will ask his senior staff to consider whether changing tests would cripple the district's ability to review LEARN's progress.