Educators distrust 'small' SABE sampling

SCORES: Spanish-language test must overcome validity challenges like those faced by the Stanford 9 _ but for different reasons.

This is a tale of two tests taken by California students.

Results of both have come into question _ but the basis of the questioning is different.

The Stanford 9 tests were required for all California students in grades 2-11 in 1998-99, the second year the tests have been given. Reports of those test scores are running three weeks behind schedule because of computer-programming errors by the publisher, Harcourt Educational Measurement.

The Spanish Assessment of Basic Education, or SABE/2, test scores were released on time, but the validity of the results are in dispute because so few students took the tests.

Only 7,100 Orange County students took the SABE, compared with 425,000 eligible to take the Stanford 9.

“It’s more for individual (student) information for parents and teachers,” said Gerry Shelton, test administrator with the state Department of Education. “The usefulness of group scores is highly questionable. “

The SABE tests in reading, math, language and spelling were required for the first time this year for Spanish-speaking, limited-English students who have attended California schools 12 months or less. Also eligible were students who have been taught part-time in Spanish this past year.

Only five of Orange County’s 27 districts _ Anaheim High, Capistrano Unified, Orange Unified, Placentia-Yorba Linda Unified and Santa Ana Unified _ tested more than 300 students.

At least 10 scores per grade were needed to calculate a group average. Only a handful of schools, mostly in Santa Ana, had enough SABE test takers to report grade-by-grade results.

Santa Ana Unified officials scrutinized the scores to assess the quality of their bilingual programs. Only 400 of the district’s 3,400 SABE test takers had been in California schools for less than a year. The others were products of the district’s _ or other districts’ schools’ _ bilingual programs.

More than half the Santa Ana students who took the test were in grades 2-5, and the majority of them outperformed the national average in reading and spelling, said Linda Delgiudice, Santa Ana’s director of evaluation.

“What this tells me is that, in the early years, the children are getting stronger in Spanish,” she said.

Both the Stanford 9 and SABE tests report scores on a curve, with the 50th percentile representing average.

But it is impossible to compare results on the SABE and Stanford 9 scores, experts said. Besides being written in different languages, the tests ask different questions and the averages are based on different students.

The Stanford 9 averages were derived from a national sample of 250,000 students who took the test in 1995. The SABE norms were based on scores of 9,000 native Spanish-speaking students from 12 states who took the test in 1991.

Testing experts said the original SABE students probably represent more low-income and disadvantaged students than the group who took the Stanford 9.

SABE scores for some students were not posted on the state’s Internet site, a discrepancy that raised questions about the accuracy of the results similar to those that have dogged the Stanford 9 report.

At Gates Elementary in Lake Forest, 273 students in the school’s dual-language Spanish-English immersion program took the test. But the state Department of Education’s Internet site reported results for only five students.

“I’m trying to figure out what they’re reporting here,” Gloria Roelen, Saddleback Valley Unified School District’s bilingual-education coordinator, said of the Internet report.

Answers were not immediately available from state officials or representatives of CTB/McGraw-Hill. But they said they have received no indications of the programming errors that have plagued the Stanford 9.

“We haven’t seen anything of that magnitude with the SABE/2, but that doesn’t mean there haven’t been some glitches,” said Gerry Shelton, state administrator for testing.

“We haven’t heard from a single school or the state in terms of any problem,” said Michael Kean, CTB/McGraw-Hill vice president for public and governmental affairs.

The state Assembly Committee on Education on Wednesday approved a bill designed to avoid the problems that have stymied the Stanford 9. It would give the state until July 15 _ instead of June 30 _ to publish scores on the Internet, allowing more time to uncover the kind of errors that marred this year’s release.

The bill, sponsored by Sen. Dede Alpert, D-San Diego, would also create a single contract between the state and Harcourt, which has contracts with each of California’s 1,000 school districts.

Register news services contributed to this report. John Gittelsohn can be reached at [email protected]

Comments are closed.