State and local students, regardless of English language competency,
improved their scores on the state’s basic skills test at most grade levels in all subjects, but made the most notable gains in lower grades.
Educators greeted this year’s improved scores by limited-English-speakers as evidence of better learning, but said information released yesterday did not make a definitive case for English-only instruction.
Not only did the limited-English-speakers’ scores improve over last year’s on the Stanford 9 test, but so did those of California students as a whole.
Such across-the-board improvement makes it unclear whether the English learners’ gains are attributable to a voter-approved ban on bilingual education, a series of concurrent reforms or simply a familiarity with the test.
Delaine Eastin, state superintendent of public instruction, said the scores are a sign that state education reforms have helped students.
High school scores continue to lag behind national averages in most grades,
though, and both local and state students showed a drop in reading scores in the ninth grade.
Some statewide information was made public early this month, but yesterday’s release was the public’s first look at school-by-school scores on the state tests in reading, math, language arts, spelling, science and social studies.
About 4.26 million second-through 11th-graders, including 341,000 in San Diego County, took the tests.
Locally, highlights included:
In math, the county’s English-proficient students scored well above the national average in all grades. In third grade, the county average was the 68th percentile, meaning students scored better than or equal to 68 percent of the test-takers in the national sample.
Gains were made in almost every grade and almost every subject by the 140,000 students in San Diego city schools, which account for about a third of the county’s public school enrollment.
Countywide, there was a rise of 9 percentile points in second-grade math scores and a 10-point improvement in third-grade math scores.
Spelling scores improved countywide in every grade, yet still lag behind scores in other subjects and remain below the national average in every grade tested.
True to long-standing trends, the highest scoring schools tended to be those in more affluent areas of the county. High family income is one of the strongest predictors for high test scores.
In addition, educators continue to grapple with the achievement gap, a decades-long disparity between the academic performances of some minority students and their white counterparts. Educators say these factors as well as large concentrations of non-English-speakers in some areas make school-to-school comparisons unfair.
San Diego County has far more minority, poor and non-English-speaking students than the national sample to which its scores are compared.
To examine the differences in scores across such demographic divides, the state has calculated average scores for seven groups: all students, those proficient in English, the non-English-proficient, low-income,
non-low-income, male and female.
But Proposition 227 made the language breakdowns the most eagerly anticipated scores.
The passage of the state measure in June 1998 mandated that teaching be done overwhelmingly in English. The polarized debate that characterized the campaign lingered through the inaugural year of its implementation.
Yesterday the discussion shifted from conflicting anecdotes to varying interpretations of the first batch of empirical data for 900,000 test-takers classified as limited-English-proficient.
“They’re moving up, but they’re moving up consistently with the rest of the kids in California,” Eastin said.
But Proposition 227 author Ron Unz said the scores were evidence of the effectiveness of his proposition. He pointed to greater proportional increases in scores among limited-English-speakers than among their English-proficient peers as evidence to buttress his claim.
By Unz’s calculation, the state’s English learners increased their percentile rank by about a fifth in one year.
“I’m not familiar with any other education reform . . . that’s been able to raise (percentile) scores by 20 percent in just seven months,” Unz said.
By state law, the scores were to have been posted on the Internet June 30.
Just before that deadline, test publisher Harcourt Educational Measurement admitted it made a massive scoring error that resulted in falsely inflated average scores for limited-English-proficient students.
Harcourt said it mistakenly mixed the scores of 250,000 fluent English speakers with those of limited-English-proficient students, tainting both groups’ average scores.
Though the state returned the figures to Harcourt without releasing them,
individual districts had publicized their scores and celebrated huge gains by English learners before the error was discovered.
Unz held up Oceanside as proof of the success of Proposition 227. Oceanside has implemented Proposition 227 so aggressively that none of the 22,000 students in the district receives more than a smattering of instruction in a language other than English.
It first appeared that Oceanside’s English learners had in some cases improved their percentile rank by as much as 475 percent. The astonishing apparent increase led Superintendent Ken Noonan to muse that perhaps educators had been cheating English learners in the past by teaching them in their native languages.
As expected, the corrected information showed more moderate increases than initially reported for limited-English-speakers, with gains of about 13 percent to 24 percent.
Meanwhile, the pro-bilingual education coalition Californians Together highlighted scores in neighboring Vista Unified School District, where bilingual education continued virtually untouched. Vista’s limited-English-speakers, too, recorded gains in most grades in reading.
Vista’s limited-English-speakers outscored Oceanside’s in all grades except second in reading.
Even the corrected information is imprecise. Harcourt officials told the state Board of Education last week that it received more than 400,000 answer sheets with no language designation, so almost 10 percent of test-takers are not factored into the new language-specific averages.
The company also botched the scoring of 44 year-round school districts,
including Vista Unified, by incorrectly ranking individual year-round students and schools against a national average.
Despite the problems, Eastin said she remained committed to the continuing use of the Stanford 9, which the state will use to rank schools later this summer and eventually reward improving schools and possibly penalize poor-performing schools — in extreme cases by closing them.
“This is the first of a series of steps we have to take to get our kids up to world-class standards,” she said.
Harcourt faces possible state action for its botched handling of the state testing system, for which it is paid $34 million.
The state Board of Education will consider Aug. 2 whether to withhold $1.1 million as a penalty, and the state could seize Harcourt’s $2.2 million performance bond by Dec. 31.
The corrected scores allowed the first statewide posting of county, district and school scores. The scores are available on the Internet at
While limited-English-speakers showed improvement, their scores remain far below national averages. In the state, in the county and in San Diego city schools, limited-English-proficient students’ reading scores improved markedly in second and third grades, which have benefited from class-size reduction since 1996. Their reading scores inched up by a point or two in other grades.
Isolating a single cause for the test score improvement may be difficult,
since California education has undergone a whirlwind of reform in recent years. Changes have included reduced elementary-grade class sizes, new grade-by-grade content standards, increased investment in textbooks,
libraries and technology, standardized testing and a pronounced slant toward a more back-to-basics teaching strategy for reading and math.
Testing experts say, too, that in the second year of a standardized test the scores tend to rise as test-takers gain familiarity with the format and educators have a better feel for the content of the exam. Harcourt officials said they have consistently seen a second-year rise in test scores in other states that use the Stanford 9.
Almost everyone agreed that the scores were good news for English learners.
Kenji Hakuta, a Stanford professor and a member of Californians Together,
said, “I think the victory is, we’re paying attention to how these kids are doing.”
Staff writers Susan Gembrowski, Maureen Magee and Angelica Pence contributed to this report.