Students Fall Short in New State Tests : Education: Measurement of skills against rigorous standards produces a highly unflattering picture of public schools. Results are poorest in math, best in writing.

State education leaders today released the first results of a revolutionary new system for assessing how well California students are learning that gives a highly unflattering snapshot of public schools.

“This is not pretty,” acting state schools chief William D. Dawson acknowledged in summing up the performance of 1 million students who took the pioneering California Learning Assessment System tests in reading, writing and mathematics.

For the first time, California students were measured not against each other but according to rigorous performance standards set by a task force that included educators, school board members, business leaders, parents and testing experts. The tests were given last spring in grades four, eight and 10.

Math results were especially grim. Statewide, at least one-third of tested students in each grade demonstrated little or no understanding of basic math concepts. Only 7% of fourth-graders showed a substantial grasp of mathematical concepts; eighth- and 10th-graders fared only slightly better.

More than 80% of students in each grade in the Los Angeles Unified School District showed only “limited mathematical thinking. . . . While responses are sometimes correct, student work often falls short of providing workable solutions.”

Reading performance across California was closer to the highest standard but nothing for school officials to brag about. About 20% of the middle school students, 23% of the elementary pupils and 30% of the high school sophomores demonstrated only a “superficial understanding” of what they read.

Writing scores were somewhat better. Although very few students attained the highest standard, at least 40% in each grade tested were able to communicate in a clear, adequately organized way. But the majority wrote incoherently and made frequent errors in spelling, grammar and punctuation.

“We have chosen to set a very high standard . . . so I was not particularly surprised” by the results, Dawson said. “But I was certainly reminded . . . of how far we have to go to yield student progress that we are going to be satisfied with.”

The standards were written last summer by a 57-member task force and are based on state-mandated curriculum guidelines that tell every public school what California students should be studying. The standards, which were approved by the State Board of Education, aim to measure how much of the curriculum is absorbed by students.

For each subject, the standards describe on a scale of 1 to 6 the types of skills students should pick up. In writing, for example, a student meeting the sixth, or highest, standard can write in a “confident, purposeful, coherent and clearly focused” way. The writing of a student at level 1 is “brief, incoherent, disorganized and undeveloped.”

In addition to the standards that compare students in Los Angeles and Santa Ana on the same scale as students in Fresno and Eureka, CLAS represents a new approach to testing that is growing in favor.

Gone are multiple-choice exams that force students to pick a single correct answer--and let them guess with a fair chance of success. The new tests require students to write at least some of their answers, respond to open-ended questions about their reactions to literature or explain how they arrived at solutions to math problems.

Students are not scored solely on whether they arrive at a single correct answer. Understanding basic principles and displaying the ability to solve problems also count.

The goals of the new system, which is often called “performance-based” testing or “authentic assessment,” are to improve teaching and raise the achievement of the 5.2 million students in California’s increasingly diverse public schools.

“By setting high standards, we learn that nobody is doing very well, and that is painful,” said Lauren Resnick, a director of the Washington-based New Standards Project, which is helping 19 states develop criteria for student achievement. “It’s also the necessary first step to improvement.”

The results, released earlier to school district officials, have already begun to cause consternation. School districts, primarily in affluent suburbs, that have traditionally scored at the top of standardized tests may still compare well to poorer districts, but the tests find that even they can stand much improvement.

“As we look carefully at the data and what it actually says, we don’t think it really means anything,” said Barbara Smith, assistant superintendent of the Capistrano Unified School District in southern Orange County. The district fared well compared to other districts in the state, but Smith argued that the paucity of high-end scores says more about the test than it does about students’ achievement.

“There must be a mistake,” said Tom Jacobson, principal at Corona del Mar High School in Newport Beach, where 97% of eighth-graders scored in the lowest two levels in math. He said 98% of the students go to college and the school’s past test scores typically were in the 95th percentile statewide.

Michael Kirst, a Stanford University education professor and a director of Policy Analysis for California Education, a Northern California think tank, said the new system’s “main impact is not on low-scoring schools that knew they were in trouble but on those districts that were used to being on the top, and now suddenly they have a lot of kids (in the lower and middle range).”

“This will raise a lot of questions with parents and with school boards,” Kirst said.

The reactions are similar to those in the handful of other states that have tested students against high standards, especially Kentucky. Outcry from parents in usually high-scoring Louisville schools slowed acceptance of the system, according to Monty Neill of the Massachusetts-based FairTest, which lobbies for better assessment methods. But within a year, “the message got across,” and Kentucky is continuing to use statewide standards of achievement, he said.

On California’s new tests, traditionally low-scoring districts continued to fare poorly. The Compton Unified School District, which the state took over last year because of looming budget deficits and poor academic performance, finished in the bottom five of Los Angeles County school districts in all three subjects and all three grade levels.

As in other tests, the new assessment system showed a strong correlation between family background--especially in regard to wealth and parents’ education--and students’ academic performance.

To nobody’s surprise, students in affluent Marin County public schools did better than just about any others in the state, while those in Imperial County, with large numbers of poor and recent immigrants, came out near the bottom.

Yet even most Marin County students fell short of reaching the highest levels of the test’s standards in all three subjects. In math, 57% of Marin County fourth-graders scored in the lowest two levels, showing little grasp of basic concepts.

Scores also varied widely by ethnicity. On the 10th-grade writing test, for example, 21% of Asian-heritage students and 16% of white students met the highest performance standards. That compared to 6% of black students and 5% of Latino students who achieved at the highest levels.

Within ethnic groups, the test scores reflected the educational achievement of the students’ parents. Among black 10th-graders, 38% of those whose parents hold graduate school degrees scored at the high end of the writing test, compared to 22% when parents were just high school graduates. For white 10th-graders, 61% whose parents have graduate school degrees scored high in writing compared to 30% whose parents did not graduate high school.

Across the state, the new testing system has already stirred controversy among some conservatives over the style and content of the tests. Others have criticized the scoring system, snags in getting started and the comparatively high cost of the tests.

The tests were scored by 2,000 specially trained teachers at 35 sites around the state. To save money, only a sampling of tests were scored. Delays in getting test materials and directions to teachers and in releasing results have set off complaints that schools had little time to prepare last year and little opportunity to change course before this spring’s exams.

Despite the glitches and budget limitations, assessment experts consider the new generation of tests far superior to anything that came before.

“People have very high expectations for the (new) system, but they need to recognize that it is going through a development process that will take time” to perfect, said Eva Baker, a director of the U.S. Center for Research on Evaluation, Standards and Testing at UCLA. “Our alternative is to go back to what we were doing, and that was not working.”

For many years, California used standardized achievement tests to measure the progress of its students. Under former state schools chief Bill Honig, who used the tests to spur public schools to do better, the annual results were widely publicized during the 1980s.

But except for a writing segment, the tests relied heavily on easily scored multiple-choice questions. They were designed to help gauge academic progress by the state’s public school students as a whole; they also could be used to assess a given district or individual school, but they were never intended to provide reliable measures of individual students.

In addition, the old tests given under the California Assessment Program (CAP) were graded on a curve and ranked schools by percentile.

The old system was dropped when then-Gov. George Deukmejian vetoed funding in 1990. His successor, Gov. Pete Wilson, a year later authorized money for the development of a new system, but insisted that it yield results for individual students as well as averages for the state, districts and schools.

With a second round of testing scheduled for this spring, individual results will be reported for students in grades four and eight but not for 10th- or fifth-graders, who will take the first history and science exams.

Because of the cost, individual scoring and tests in additional subjects will be phased in. By the 1997-98 school year, plans call for individual scores to be available on all tests--reading, writing and math in grade four, history and science in grade four, and in all five subjects for grade 10.

Testing will eventually also be conducted in Spanish, the language of the state’s largest group of students who are not fluent in English. Currently, students who have attended school in this country for less than 30 months and those who are enrolled in classes taught primarily in their native languages are not required to take the tests.

About 15% of pupils in the tested grades, including some special-education students, were not required to take the test last spring, education officials estimated.

Finally, when fully in place, the assessment process will include the evaluation of a portfolio of classroom work by each student. The portfolio will help pinpoint areas of progress and weakness. Parents will be able to go over portfolio work with teachers and get their child’s results on the statewide test, but will not be able to see the actual test materials, which are kept confidential.

State fiscal analysts expect the full system to cost at least $55 million a year; the old testing system cost $12 million a year. As with other state programs, the assessment system must rely year to year on funding from the governor and Legislature, injecting an element of uncertainty and sensitizing test advocates to political pressures.

Even before publication of the first results, education officials yielded to pressure from conservatives and deleted two test items based on writings by award-winning author Alice Walker. Civil rights groups protested and threatened to sue, and hearings on the issue are scheduled in Sacramento for Friday and next week.

The state legislative analyst’s office also cautioned recently that the new testing system may be “trying to accomplish too many diverse goals with the amount of available funding” and may not produce reliable individual scores this spring.

Despite the cost, the state’s business establishment is “advocating very strongly for this,” said Jere Jacobs, a Pacific Telesis Group executive and a member of the California Business Roundtable. “This gives us the standards we have been talking about for a long time. . . . It’s the cornerstone of a very rational accountability system.”

By setting high, clearly defined standards closely tied to the state’s new curriculum guides and textbooks, educators expect to be able to more clearly assess what improvements are needed. They expect the standards to improve teacher training and to spur schools to more closely follow the curriculum.

After this benchmark-setting year, schools will judged not by where they started but by the progress they show, schools chief Dawson promised.

“We can either face the reality of what our kids can and cannot do . . . or we can stick our heads back in the sand,” Dawson said.

Stanford’s Kirst said many schools, after several years of lean budgets, cannot afford to buy the needed textbooks and equipment or train staff members in the new methods and materials the updated curricula reflect.

“People have been enthusiastic, but they don’t have the resources. The question to the state is, ‘What are you going to do to help the schools?’ ” Kirst said.

Times staff writer Jodi Wilgoren contributed to this story.

* MORE ON TEST SCORES: Additional stories. B1

Results for local schools and districts. B2

Test of Knowledge

Here are some sample problems from the California Learning Assessment System:

* Fourth-Grade Writing: Write a letter to the Student Council giving ideas how to improve lunch time at your school. . . . Do your best to convince the Student Council that your suggestions are good ones.

* Fourth-Grade Math: Maria wants to buy a 75-cent snack from a vending machine. The machine takes only nickels, dimes and quarters. Maria has seven nickels, five dimes and two quarters. Show all the different ways she could pay for the snack (you may use words, diagrams or charts). Which of your ways uses the fewest number of coins? Explain why this is true.

Answer: Nine ways with exact change. Two quarters, two dimes and one nickel uses fewest coins.

* 10th-Grade Math: A game involves two cubes with sides numbered from 1 to 6. After throwing the two cubes, the smaller number is subtracted from the larger number to find the difference. If a player throws the cubes many times, what difference will probably occur most often? Provide a diagram and written explanation that you could use to explain to a friend.

Answer: 1. Supporting work should show that problem involves 36 possible outcomes.