CLAS Safeguards Often Mean No Results

Plagued by administrative glitches and statistical problems in processing the first round of results from the California Learning Assessment System, the state Department of Education this year instituted a host of safeguards to prevent it from releasing unreliable data.

But the new oversight system left hundreds of schools without any results at all because some students chose not to take the controversial tests. And like last year, others received results for the 1994 tests based on very small numbers of exams.

“When you have such a low number, I think it’s problematic,” said Al Mijares, superintendent of Santa Ana Unified, Orange County’s largest school district. “I don’t think you can draw inferences on the basis of what we received. The whole CLAS thing was, I believe, so mishandled that to now try to make sense of it all and give us information that we can depend on is irresponsible.”

Even in districts where there was widespread participation in the exam, producing statistically solid results, some educators complained that CLAS scores are useless because Gov. Pete Wilson effectively killed the program last fall by vetoing its funding.

Capistrano Unified Supt. James A. Fleming pronounced the results released Tuesday “less than meaningless.”

“I don’t know what value there is, to be honest with you, in beating a dead horse,” agreed Fleming’s colleague, Peter Hartman, superintendent of Saddleback Valley Unified. “It’s like a second eulogy or something.”

Controversy erupted last spring after The Times revealed the state violated its own guidelines and released data that did not meet internal reliability standards to hundreds of schools. The main problem then was that each school’s scores were based on just a sample of student work, and in many cases the sample was too small to maintain statistical validity.

This year, the state scored every student’s exam, following the recommendation of an expert panel appointed in the aftermath of The Times report.

Still, results of the 1994 tests are compromised by the large number of students who “opted out” of CLAS, mainly to protest the content of some of the reading excerpts and writing exercises or the testing approach.

State officials decided not to publish results for any school where more than 25% of the eligible students chose not to take the test or where “opt-outs” significantly changed the demographic profile of the tested group.

But some districts, such as Santa Ana, failed to return the state’s official opt-out forms, so scores were posted even though most students did not take the exams.

Gerry Shelton, a consultant to the Department of Education, expressed confidence in the reliability of this year’s results.

“We did a lot of good work over this last year. The state-of-the-art in terms of calculating standard errors and assessment accuracy probably has gone ahead three or four years because of this,” Shelton said.

In Orange County, more than 100 schools did not receive scores on at least one section of the reading, writing and math exams, administered to fourth-, eighth- and 10th-graders last spring. That includes all 28 schools in the Placentia-Yorba Linda Unified School District, where administrators expect to receive results from the state next month.

A Times computer analysis shows that hundreds of schools statewide, like those in Santa Ana, received scores based on very few tests.

“They’re so incomplete, I can’t even respond to them,” said Robert French of Orange Unified, where two-thirds of the elementary and high schools, and more than 80% of the middle schools, were given zeros instead of scores.