Switch in L.A. School Tests Blurs the Results

No report card will be available this year on efforts to fix the Los Angeles Unified School District because the new student test unveiled last spring cannot be compared to earlier ones.

And because the Legislature on Saturday approved administering a statewide achievement test, next year could mean starting over again as Los Angeles educators confront the results of a third type of test.

Neither of those glitches was anticipated by school board members in January, when they voted to switch from year 14 of the Comprehensive Test of Basic Skills to year one of the Stanford 9.

District officials--like those in many other California districts--were seeking a more modern and accurate exam. For example, in math the Stanford 9 relies on word problems, in contrast to the old test’s extensive use of straight mathematical computation.

To an individual parent, the inability to compare the new test to the old one might seem trivial. But on a districtwide basis it is the only real way to monitor the value of various reforms.

In Los Angeles, which is desperate to document improvement in student achievement, key reform efforts range from the 4-year-old charter schools program to a decade-long attempt to bolster 10 inner-city campuses to the 5-year-old LEARN program, which is rooted in the belief that campuses with more autonomy will perform better.

LEARN President Mike Roos said his group had hoped for some consistency in the district’s tests, especially for the 87 schools that have been involved in the reform four years or longer. But, he said, the two tests appear to be “apples and oranges.”

When the Board of Education decided to switch to the Stanford 9 test, it was counting on an accompanying study to provide an accurate comparison with the old test.

“What is the value of a test if you can’t . . . use it as a reference as you move on?” asked board President Julie Korenstein.

The ideal solution would have been to convert the Stanford 9 results into CTBS equivalents, thus giving the district 15 years of comparison data.

But instead, district officials said last week, they will only be able to express last year’s test results in the Stanford 9 equivalents on a limited basis. The equating study will be released this month, along with the scores.

The district has already been told by test publisher Harcourt Brace that students who scored at the 50th percentile in fourth-grade reading last year on the old test would have reached only the 35th percentile had they taken the new test.

“It’s been a disaster,” said board member David Tokofsky, who was vociferous last year in his doubts about the ability to compare the two tests. “Nothing that we intended to have happen for the good of student achievement and measurability has occurred.”

The equating study is only being done districtwide, not on a school-by-school basis, even though new Supt. Ruben Zacarias has focused his attention on the performance of individual schools.

*

At a forum with 1,000 administrators Friday, Zacarias reiterated his decision to target the 100 lowest-performing schools--as determined by the past three years of scores--a job he conceded is complicated by the data interruption.

In retrospect, Zacarias said he wished the district had not changed tests, particularly when it was clear even last year that the state was moving toward a California-wide assessment.

“I wasn’t involved in that decision,” he said. “Had I been, I’d say I think continuity is important . . . [and] I don’t know that one test is that much better than another.”

Some testing experts outside the district say there would have been ways around the problem--such as giving both old and new tests to a core group of students last spring. A few even said they believed that reasonable comparisons would still be possible.

“Given that educators are supposed to be in the business of learning, it is amazing to me how little attention is given to looking at achievement over a long period of time,” said Robert Calfee, an education professor at Stanford University. “To be honest, though, I’ve never known the LAUSD or any other big school district to really want to do that.”

In the wake of the Legislature’s last-minute resurrection of a statewide testing process, uncertainty reigns over what options California districts will have next year. The State Board of Education will choose a test; if it is not Stanford 9, there might be state funding for the Los Angeles district to give two tests for continuity.

The suggestion of having to administer two rounds of standardized testing brought sighs of frustration from many top district officials last week because it would mean more interruptions of the regular teaching schedule. But Deputy Supt. Ron Prescott, long the district’s top political lobbyist, suggested that test-taking offers another valuable lesson.

“One way you learn about tests is you take them,” Prescott said. “I think these kids maybe need that.”

*

Despite anger by some board members over the mismatch between the two test results, some district officials said they anticipated the loss of direct test comparisons. As a result, said testing coordinator Linda Lownes, schools were not required to set test score goals this year--as they had been for the past two years under an improvement strategy founded by Zacarias’ predecessor, Sid Thompson.

Test scores of individual students have been distributed to all schools and some schools already have passed them on to parents. However, errors in averages for the entire district as well as individual schools have slowed the delivery of those summaries to schools and to the public.

Lownes maintained that the district is “more comfortable beginning with a new baseline of data, because the farther you get away from the original baseline, the more air you get in your results. We were comparing our students back to students who were tested in 1982.”

Though the Comprehensive Test of Basic Skills marketed a modern version of its test last year, at the time of the board’s approval that test had not received the state nod for reimbursement.

As the release of test scores approaches, Lownes and others are warning that the new baseline will not be flattering for Los Angeles students. It will appear to indicate a precipitous drop from an already dismal standardized test performance, they say.

State testing manager Gerry Shelton said such declines are to be expected in the first year or two of an unfamiliar test, just as artificial score inflation occurs when the same test is given for many years.

“It’s a new system, the kids aren’t used to it, they panic a little bit and teachers may not know how to administer it [or] the target may not have been what they were aiming for,” he said. “It’s like moving the target and saying, ‘Oops, you missed.’ ”

*

The challenge for researchers, Shelton said, is determining how much of that drop is real, a statistic usually determinable only after several years.

.

On another testing front, the district is threatening to pull out of a proposed national fourth-grade reading exam championed by President Clinton--which it joined in July amid much fanfare in the nation’s capital--unless the exam is given in Spanish as well as English.

By contrast, the national eighth-grade math test is being given in multiple languages.

Because nearly half of the district’s fourth-graders have learned to read in Spanish instead of English, assessing reading in English would be unfair, officials said.

“If we test everybody in English, our scores are going to be way down on the floor,” said district spokesman Brad Sales. “We feel that that’s an unfair assessment of us as a system.”

Ron Prescott, speaking from Washington, where he was lobbying for changes, expressed optimism that negotiations were favoring the district.

“We have said we’re not going to give any test to kids unless all of them can take it,” Prescott said. “Reading and language are two different things--we have a lot of Spanish readers and it’s inappropriate to say they can’t read because they don’t speak English.”