State Releases Rankings That Show Improvement in Schools

Seventeen percent of California’s public schools met a state target for academic accomplishment last year, a 5-point improvement over 1999, according to the state’s second statewide ranking, released Tuesday.

Delaine Eastin, state superintendent of public instruction, called the gain “important and substantial.”

Still, few schools advanced or slipped significantly in their positions against one another in the rankings, primarily because most schools raised their test scores in 2000.

Los Angeles Unified School District managed to pull 12 of its elementary schools out of the basement. Still, 32% of the district’s elementary schools--and one-third of the district’s 561 schools overall--ranked in the lowest 10% band.

Although state and local school officials cheered the fact that schools are moving in the right direction, they continue to raise concerns about California’s efforts to create a valid accountability system.

“We’re just beginning the era of confusion of what is the API,” said Barbara Miller, research director with EdSource, a nonprofit organization in Palo Alto.

The Academic Performance Index, for now, is based solely on the Stanford 9, a standardized basic skills test that is given each spring to about 4.3 million children in grades 2 through 11.

But in the next few years, the state Department of Education plans to introduce other measures aimed at assessing how well students are learning what California wants them to. That raises the specter of a constantly shifting target that will complicate comparisons from year to year.

As it is, the state has chosen to release API data in two chunks. The first comes in the fall, when the API index is released and schools learn whether they met their goals. The second comes in January, when rankings on a scale of 1 to 10 are unveiled--with 10 being best--along with performance targets and rankings of schools with similar demographic traits.

Those releases, along with the summer publication of Stanford 9 scores, heap number upon number, making it tough for anyone to grasp the meaning of each piece.

Even Critics Say State’s on Right Track

“The effectiveness of this would be aided by as clean and compact a release as possible,” said David Rogosa, a statistician and education professor at Stanford. Still, Rogosa, who in the past has questioned the trustworthiness of some of the numbers, said the state appears to be on the right course.

The state’s two top education legislators have vowed to tweak any elements of the testing and accountability program that don’t stand up to scrutiny. The Stanford 9 testing program, which grades students against a national sample, is up for reauthorization this year.

“We are early in our experience,” acknowledged Kerry Mazzoni, Gov. Gray Davis’ education secretary. “Over time, there may be things we need to take another look at.”

The API, started last year as the cornerstone of Davis’ effort to improve the lackluster performance of schools, is intended to measure academic improvement. Each of about 7,000 schools receives a score ranging from 200 to 1000. The state set 800 as a target last year, and only 12% reached that level based on their 1999 test scores.

Improvement, rather than absolute standing, is the goal. Schools that reach or exceed their targets are eligible for financial rewards, ranging from modest per-pupil amounts for the schools to hefty $25,000 bonuses for teachers in low-ranking schools that show huge gains.

Many educators praise the API for keenly focusing attention on the classroom. Although some teachers say they have felt pressure to “teach to the test” and even to cheat, others say the system has spotlighted the importance of teaching California’s rigorous content standards.

Long Beach Unified School District has found that students do better when teachers aim higher.

“I feel that everything has been for the best,” said Kathy Stadtlander, a second-grade teacher at Long Beach’s Whittier Elementary School, which rose to a ranking of 3 from a 1 last year. “Everybody should have accountability.”

Rogosa says his analysis indicates that the statewide rankings appear to be fairly reliable. In a forthcoming report, he says the chances of schools obtaining rewards in error or failing to get rewards that were merited are reasonably small.

One key to that, he said, is that the state requires schools to bring up the results of sizable subgroups, such as poor children and ethnic groups.

He is less sanguine about the similar-schools rankings, which weigh a school’s performance against that of 100 other schools with comparable ethnic and socioeconomic traits.

The data used to compile the lists of similar schools are imprecise, Rogosa and others acknowledged, in many cases because the information comes from the children. Last year, the state had to revise the similar-schools rankings because some schools provided faulty data.

State officials say the numbers this year are much sounder. Thus, a school with a low statewide ranking of 3 could take heart from a similar-schools comparison of 8. That would indicate that it was doing better than most other schools with similar challenges.

Eastin said the state has been hampered by the lack of a computer system that could keep tabs on everything from parental education levels to graduation and attendance rates.

Rogosa said schools could still glean valuable information from the similar-schools rankings. He found wide variation among bands of similar schools, with typical spreads of 200 to 300 points. Principals, he said, could visit schools that have exceeded expectations to learn what they are doing differently in the classroom.

The similar-schools rankings indicate that “you’re not preordained to do poorly,” Rogosa said.

Even seemingly high-performing schools in California have a long way to go, Rogosa noted. A school that hits the statewide target of 800 still has nearly a quarter of its enrollment below the national average.

The index peak of 1000, which no school has hit, might imply perfection. But, to reach that level, each student at a school would have to score at the 80th percentile in each of four subjects tested. Put another way, each student would have to get a little more than four of every five questions correct.

Scores Are Used to Reward Schools

Predictably, whether schools preferred to accentuate the statewide or the similar-schools ranking hinged on where they fell in each.

Thomas Jefferson Elementary in Burbank, for example, had a statewide rank of 8 but a similar-schools comparison of 3.

“I don’t really know what that [similar-schools ranking] means, and it really doesn’t mean that much to us, because I don’t really know what the other schools are like,” said Principal Lynn Perske.

The state’s listings also set new growth targets that schools must meet to qualify for the monetary awards.

The first rewards for schools based on their gains last year are expected to go out in early February. Teachers and other staff members should begin receiving checks in April or May.

Schools that perform poorly on the index may qualify for additional state and federal funds. Under state law, if they still fail to boost student scores, they could in extreme cases risk being taken over by the state.

The public may view the latest API numbers on the Web at https://api.cde.ca.gov.

*

Times staff writers Jessica Garrison and Elaine Gale in Orange County and Massie Ritsch in the San Fernando Valley contributed to this story. Richard O’Reilly, Times director of computer analysis, and Sandra Poindexter, data analyst, also contributed.

(BEGIN TEXT OF INFOBOX / INFOGRAPHIC)

L.A. County Rankings

Following are the percentages of schools in Los Angeles County that fell into each ranking of API test results. The statewide ranking is on a scale of 1 to 10, with 1 representing the bottom 10% of schools and 10 the top 10%.

*--*

Los Angeles Unified Rest of L.A. County State % in rank % in rank % in rank % in rank Rank for 1999 for 2000 for 1999 for 2000 10 2.8% 3.1% 10.5% 10.3% 9 3.4% 2.7% 8.9% 9.2% 8 3.4% 3.8% 7.5% 7.7% 7 4.1% 4.7% 7.7% 7.5% 6 5.1% 4.5% 9.0% 8.9% 5 8.7% 8.2% 10.0% 8.8% 4 6.5% 8.3% 11.8% 12.6% 3 12.9% 12.7% 12.7% 13.4% 2 18.9% 18.8% 11.4% 11.8% 1 34.2% 33.2% 10.4% 9.7% Median rank 3.25 3.29 5.22 5.22

*--*

*

Source: California Dept. of Education, API 2000 rankings

Researched by RICHARD O’REILLY, Times director of computer analysis, and SANDRA POINDEXTER, Times data analyst