School Loses in API Numbers Game

Principal Ben Carpenter of Ball Junior High School was expecting to be happy about his school’s score on the Academic Performance Index, the new state system for ranking and evaluating schools.

After all, both Latinos and whites, the two statistically significant groups of students at the Anaheim school, had not only achieved the growth targets set for them by the state but also exceeded them. The school’s Stanford 9 scores, on which the API is based, improved in all areas except reading.

So Carpenter was shocked when state officials announced earlier this month that Ball’s overall API had decreased by 5 points instead of hitting its 12-point growth target. This makes the school ineligible for a massive pot of financial rewards being handed out to high-performing schools.

“Looking at this, it makes no sense,” Carpenter said. “The teachers were taken aback. We had all worked so hard . . . and we haven’t gotten an explanation for why this happened.”

State officials said there was an easy answer: The demographics at Ball have changed dramatically since 1999. The number of Latino students tested increased from 255 to 357, while the number of white students decreased from 157 to 133. Typically, Latino students, who might not be fluent in English, do not perform as well on the Stanford 9 as white students. Subsequently, their API scores also are lower. This is the case at Ball, where Latino students’ API scores as a group were about 200 points lower than the white students’ average.

Both groups exceeded their state-set benchmarks. The benchmark for Latino students was 483; students exceeded that by 8 points, scoring 491. White students scored 697, exceeding their benchmark by 42 points. But because state officials calculated the school’s API target score assuming the school would have roughly the same demographics this year as in 1999, the school did not meet its overall growth target even though scores for both subgroups improved dramatically.

“You go out and play a football game and outscore your opponent and then they tell you you lost,” Carpenter said. “It’s frustrating.”

In an attempt to improve overall scores at Ball, the school will be taking part in a carrot-and-stick state program that gives schools extra money for enrichment programs. But if scores do not improve, the state can take over the school, fire the principal or even close the campus.

Carpenter said he hopes the state program will help. But he reassured his teachers that they were doing everything right.

State officials were unsympathetic to Carpenter’s complaint that the system is unfair.

Pat McCabe, who works in the state Department of Education’s testing and evaluation division, said schools where the demographics had changed dramatically could have opted out of the API scoring this year. But such requests had to be made before API scores were released. “You can’t do it after the fact,” McCabe said. “Otherwise, people will say, ‘Oh, my! My API went down. I’m not the same’ ” demographically.

Six other Orange County schools--South and Sycamore junior highs in Anaheim, College Park Elementary School in Newport-Mesa Unified, Robert M. Pyles Elementary in the Magnolia School District, Lincoln Elementary in Santa Ana Unified and C.E. Utt Middle School in Tustin Unified--also saw their subgroups improve but their overall API scores go down.

David Marsh, a USC education professor who is an expert on testing, said the state should create an appeals process for schools in that situation. “We’re really early in the use of the API in California,” he said. “There are going to be special circumstances where a school has done noble work and deserves to be recognized for it.”

Carolyn Houston, principal at South Junior High, said she was more puzzled than frustrated or angry about the results. South is 78% Latino, and Latinos make up the only statistically significant group, meaning it accounts for at least 20% of the student body. Scores for Latino students surpassed their API target, she said. “I don’t think we totally understand the mathematical formula that is being used,” she said. “I think we were surprised because we put a lot of effort and time into this.”

Susan Harris, principal of Pyles Elementary, said she was puzzled that her school’s API score had gone down but had decided to concentrate on doing better next year.

“I can’t figure it out,” she said, noting that 81% of the students at the school are Latino, making Latinos the only statistically significant group. Latino students exceeded their growth target of 17 points, improving 18 points. But the school as a whole fell three points shy of its growth target of 17. “You can look at it 100 different ways,” Harris said. “But we’ll do better next year.”

Michelle Fulkerson, PTA president at College Park Elementary in Costa Mesa, said she was upset that a standardized test counted for so much. “I honestly think the test as a whole is unfair,” she said. “I don’t think it shows the true colors of the school.”