San Jose Mercury News

Monday, July 13, 1998

In Any Language, STAR Results Are Inconclusive

Are bilingual students better?

Maybe. Maybe not.

Students who have mastered English as a second language outscored native-English speakers in most subjects at most grade levels -- sometimes by wide margins -- on the STAR (State Testing and Reporting) exam.

But we don't yet have the data that would let us compare apples to apples and oranges to naranjas.

Analysts are still crunching the demographic data that will help us really understand which students and schools are succeeding, and which are not. Here are some tips for understanding the numbers we've got now.

This spring, after four years without a statewide exam, students in grades 2 through 11 took the same multiple-choice test, comparing their achievement to a national sample of students. Gov. Wilson insisted on testing all students, including those who aren't fluent in English. The Oakland and Berkeley school districts sued, and got a court order blocking the state from releasing scores for students classified as Limited English Proficient (LEP). School districts can release all scores, and some have done so.

Only two percent of students in the national comparison group were limited in English compared to 22 percent of test-takers here. So it doesn't make sense to evaluate the effectiveness of a school or a district by lumping together the scores of students who understood the test with the low scores of students who didn't. However, over time, STAR will show whether LEP students are making progress.

A major apples-to-naranjas problem stems from the way we decide who does or doesn't know English. We assume that if a student tests below the 35th percentile and comes from a home in which someone speaks a language other than English, that the student is not proficient in English. By definition, a LEP student tests poorly in English, and a Fluent English Proficient (FEP) student tests above the 35th percentile. Some districts won't reclassify LEP students as proficient until they reach the district average, which may be considerably higher.

In Santa Clara County, students whose home language is not English do better, once they achieve English proficiency, than students from English-only homes.

For example, English-only second graders score at the 59th percentile in both reading and math; students reclassified as Fluent English Proficient score at the 75th percentile in reading, the 81st in math.

Despite a misleading Chronicle story, which labeled these students ``bilingual education graduates,'' we don't know whether they were taught in their native language or mainstreamed in English. Nor do we know whether they're immigrants or born in the United States, how many years of schooling they've had here, their ethnicity, their parents' income or education. We know nada.

Except that we know we're comparing English-only students, who represent the full range of achievement levels, with students who represent the 35th to 99th percentile in language skills. Students who score lower than the 35th percentile aren't considered fluent in English.

Let's say that STAR shows students who've moved from limited to fluent English proficiency score at the 40th percentile in reading at the end of fifth grade in Metropolis Unified; in Smalltown Elementary, they test at the 60th percentile. Which district is doing a better job of educating students who start with low English proficiency?

The answer is: Not enough information provided.

Maybe Metropolis has moved 90 percent of LEP students to English proficiency, while Smalltown's FEP students represent only 10 percent of students who started with limited English. Maybe one district serves kids from poor, poorly educated, transient families, while another has easier-to-educate immigrant students.

If Smalltown argues that their LEP students eventually catch up, we can look at eighth grade scores, to see how many students will be prepared to take a strong academic course load in high school.

Looking at 11th grade, the last year students are tested, creates another selection problem: By then the least successful students have dropped out.

Despite the drop-out factor, STAR scores for all groups went down significantly in high school. We don't know why.

The scores of the LEP group decline from grade to grade, which makes sense. Increasingly, this group is composed of slower learners who never reached proficiency.

The group of students who've moved from limited to fluent English proficiency starts with very high test scores in second grade, since the fastest learners are the first to qualify, and goes down in later grades as less-rapid learners join the group.

STAR counts another group of students who achieved English proficiency before they started school, but come from homes where they hear a different language. In this county, they post consistently above-average scores from second through 11th grade, except for a decline to average levels in reading starting in ninth grade. They score higher than English-only students in every subject and in every grade.

These students may benefit cognitively from knowing two languages, though we don't really know how well they know their parents' language. They may have the immigrant work ethic. But don't forget that this also is a selected group that excludes slow learners; kids had to be performing adequately in order to get the Fluent English Proficient label.

By contrast, students from English-speaking homes are assumed to be proficient in English, even though some of them start school with minimal vocabulary, speaking skills or exposure to the printed word.

In short, if you see a conclusion, don't jump.

Joanne Jacobs is a member of the Mercury News editorial board. Her column appears on Mondays and Thursdays. You may reach her at 750 Ridder Park Dr., San Jose, CA 95190, by fax at 408-271-3792, or e-mail to .