Caution Urged in Interpreting Calif. Scores
After numerous scoring glitches and weeks of delay, California late last month released detailed results of student achievement tests.
An especially intense spotlight had been trained on the scores of students whose native language is not English--largely because of Proposition 227, the ballot initiative passed last year that curtailed bilingual education in the state's classrooms. ("Scoring Glitch Clouds Impact of Prop. 227," July 14, 1999.)
Statewide, the test results show a glaring achievement gap between limited-English-proficient students and those who speak English fluently. Far fewer LEP students performed at or above the national average on the standardized test.
The results also show, on the whole, that LEP students' scores are higher than they were last year. But the increases by LEP students are comparable to the gains made by all California students, state officials said.
"I honestly think it's hard to glean any real conclusions about Prop 227 from this first-year score," state schools Superintendent Delaine Eastin said. LEP students "are not moving up dramatically higher. It's really too early to celebrate or throw stones."
Many testing experts and educators urged caution in interpreting results from the Stanford Achievement Test-9th Edition, which is expected to generate data for use in the state's accountability reforms.
The test publisher, Harcourt Educational Measurement, in the past few weeks has corrected scoring problems, including a computer error that resulted in overstating achievement by LEP students throughout the state and triggered a delay in releasing detailed scores.
But concerns remain about the nearly 420,000 students whose test records were sent to Harcourt without an indication of their language proficiency. It is unclear how many of those students are LEP.
Penalty for Glitches?
This week, the state school board will discuss the testing program and possible financial penalties against the test publisher. Harcourt Educational Measurement has apologized for the inaccuracies and pledged to improve its system. With a combined price tag of $22.2 million this year, California schools make up the company's largest testing customer.
Meanwhile, the president of Harcourt Educational Measurement, Joanne Lenke, resigned last week, effective Aug. 9. A spokesman for Harcourt Inc., the testing division's parent company, said the move was unrelated to California's testing program. He added that Ms. Lenke would remain a consultant to the company.
The Stanford-9, which all California students were required to take for the first time last year, tests students in grades 2-11 in reading, language, and mathematics. Students in grades 2-8 are tested in spelling, and those in grades 9-11 are tested in science and social science. The test is offered only in English.
Proposition 227 supporters said they viewed the results as a victory for the new law, which calls for teaching LEP students almost exclusively in English. The law, in most cases, requires schools to teach such children in sheltered-English-immersion classes until they acquire a working knowledge of English, which the measure's backers say should take about a year. Students are then to move into mainstream classes.
"There are probably few educational reforms which have had such rapid and substantial impact," said Ron K. Unz, a wealthy California software entrepreneur and the architect of the state initiative. Mr. Unz released an analysis of test data that showed LEP students in grades 2-6 posted gains more than double those for all California students in the same grades. But several testing experts said the analysis was flawed and invalid.
Kenji Hakuta, a Stanford University education professor and language expert, said his own analysis showed that LEP students made gains in districts that have maintained bilingual programs, in districts that never had bilingual programs, and in districts that have moved to English immersion in light of Proposition 227.
Gauging 227's Impact
Some school officials and others argue that it's inappropriate to test students in a language they are still learning. And they point out that while nearly a quarter of California's students are LEP, such students make up less than 2 percent of the national norming sample used in the Stanford-9 test.
Mr. Hakuta said that while Proposition 227 clearly has had an impact on classrooms, it's nearly impossible to tell its effect on test scores when so many other programs, such as class-size reduction, are going on in the state.
One district that Proposition 227 backers say has strictly followed the new law is the 21,500-student Oceanside Unified School District, which has eliminated its bilingual programs. ("Calif.'s Year on the Bilingual Battleground," June 2, 1999.)
Last month, a group that includes parents of LEP children responded to the district's handling of Proposition 227 by filing a federal civil rights complaint.
In June, before the state had officially released test results, the closely watched district released scores that showed remarkable gains for its LEP students. But when the state released corrected scores last month, the district appeared to post much smaller gains.
Vol. 18, Issue 43, Page 12Published in Print: August 4, 1999, as Caution Urged in Interpreting Calif. Scores