Columbus Elementary School seems to have everything going for it. Everything, that is, except good test scores.
The school spends almost $8,000 per pupil, far more than the national average of $6,300, to pay for extra training for teachers, equipment and books galore. Its campus is new, designed as a cozy village of airy, ochre-colored cottages.
Mental health counseling, social workers and tutoring before and after school are available. Perhaps most important, the school is richly endowed with an asset considered essential to academic success: involved parents.
Parents helped raise $1.2 million, an unheard-of sum for a public elementary school, to cover "extras" such as a science lab and day-care facilities for the campus that opened in 1997 to replace buildings damaged in the Loma Prieta earthquake.
But the school also stands out for something else: its staunch commitment to "discovery learning," a decade-old philosophy that says students learn more deeply when they figure things out for themselves through trial and error and individual projects.
It is an approach rapidly falling out of favor among school reformers, who argue that highly structured lessons are more likely to turn around a situation in which nearly two-thirds of the state's fourth-graders are poor readers.
If there is any place that so-called progressive methods should work, it is at Columbus, given its extra resources.
However, in the two years since the school reopened, its test scores have dramatically worsened. Columbus Principal Rebecca Wheat and her teachers worry that they are running out of time. If scores don't improve, the state could eventually take over the school and replace them. More immediately, raising money to continue the comprehensive array of services will be difficult if test scores remain low.
Scores for grades two and three range from the 23rd to the 35th percentile in reading, math, language usage and spelling, as measured against a national norm. That means the average third-grader at the school reads worse than 77% of students nationally.
The scores for grades four and five are somewhat better, but have fallen across the board by as much as 14 percentile points.
"These are very, very high-stakes tests," Wheat said.
She believes that the school's teachers do pay enough attention to teaching basic skills. "Our decisions have been very thoughtfully made," Wheat insisted. "But I do think the pressure for test scores does put pressure on people to do more direct teaching."
At Columbus, the debate over how to teach is complicated by the wide range of students it serves. Located in one of the poorest neighborhoods of west Berkeley's flatlands, Columbus also draws from affluent areas in the hills overlooking the campus of the University of California.
The income gap, which tends to fall along racial lines, shows up in the test scores. Last year, scores for the 18% of students who are white were generally twice those of African American students, who make up 31% of the students. White students also scored far higher than the 40% of the students whose first language is not English. Moreover, that gap grows larger as students get older.
A Safety Net of Support
The school's efforts to make all students and their families feel comfortable--regardless of the language they speak or what they do for a living--is striking.
The school has a language "immersion" program that, deliberately, has half native English speakers and half Spanish speakers and is taught entirely in Spanish. All children first learn to read in Spanish, but the Spanish speakers also pick up English from their classmates. Gradually, teachers introduce more formal English instruction with the goal of all children winding up fully bilingual.
Teachers and office workers speak Spanish. The school has tried to attract poorer families to after-school enrichment activities such as stick drumming and yoga by subsidizing fees. Last year the school hired seven "parent advocates" to help parents obtain a wide range of services: from child care to dental services and even emergency surgeries.
The idea, said Alison T. Jones, the coordinator of services for students at risk of failing academically, is to "provide a safety net of tutoring and family support so that when the kids are together in the classroom, those differences are not so apparent."
But teachers at Columbus are philosophically opposed to separating students by reading level or ability. "If you put all the low kids together, it's deadly. The level of discussion is not high and the motivation is not there," said Mary Burmeister, a teacher who was conducting a reading lesson for a dozen fourth-graders one sunny day in May.
To engage the students, she put the title of the day's story on the board--"The Friends of Kwan Ming"--and told them to write down a prediction of what it was about. "Spelling doesn't count," she told them.
Then she had them share their answers with a partner and then with the whole group. Finally, she read the story to them, because although some of the students were capable of reading novels by the author Roald Dahl on their own, others were learning-disabled and not reading at all, and still others were not fluent in English.
Columbus uses federal funds to provide tutoring before school for those who are falling behind. But those who need it don't always show up, Jones said.
"Mondays are difficult," she said one morning, observing a remedial session with two students instead of the eight who were scheduled. Sometimes, families are contacted by the parent advocates, who may help arrange transportation.
The school also uses federal funds and a subsidy from the Berkeley Unified School District to pay the salaries of two teachers who are trained in Reading Recovery, a controversial program that has shown mixed results and is very costly.
"Even if Reading Recovery doesn't help their reading that much, they've got this one-on-one attention and it's really helped [students'] self-esteem," Jones said.
What the school does not do is systematically teach the sounds of letters. That is now the approach to reading instruction sanctioned in state law covering textbook purchases and professional development.
Instead, Wheat said, teachers are being trained to address the sounds of letters as the need arises, while students are puzzling through books. That way, she said, teachers can individualize instruction.
Doug Carnine, a professor at the University of Oregon, said that approach might be widening the gap between low- and middle-income children rather than narrowing it.
More affluent or advanced students, he said, might not need the practice and reinforcement characteristic of teacher-directed, phonics-based programs. But weak readers do. "What does it mean for a school . . . to sanction an instructional approach that's probably more suited to the most advantaged?" he asked.
Columbus' teachers are undeterred by such talk.
Ann Gilbert, a fifth-grade teacher, watched as her students worked in pairs measuring the angles of various geometric shapes. The point of the exercise was to discover that the size of the angles in five-sided shapes always add up to the same.
But she wasn't telling her students that. Even when they came up to her with questions, she didn't clue them in. Such a lesson probably won't pay off in higher test scores immediately, she said. "But it will by the end of high school because they'll really know it," she said.
A science lesson that began on the school's oval central courtyard and playground followed a similarly indirect path. Students had spent an earlier session making houses out of cardboard and painting them black, yellow or white.
On this day, they were supposed to put a thermometer in each house and record the change in temperature. The point was to discover that the black houses got the hottest and the white houses remained the coolest.
But the data recorded by students showed that the yellow paper houses absorbed less heat than the white ones--the wrong conclusion. Rather than correct the mistake, or have them repeat the experiment, teacher Nancy Bynes tried to start a discussion about the results.
"The yellow and white results are a mystery to me. I want to hear your ideas," she said.
When her prompts failed to get a discussion started, Bynes told the students to write about what they had seen. Some produced a few partial sentences. A few filled a page. One girl wandered around the room until Bynes sat her down, wrote out several sentences and left blanks for her to fill in.
"It shows you we need the services we're getting," Bynes said. "We have a lot of work to do, a lot of work to do."
Alfie Kohn, whose new book, "The Schools Our Children Deserve," attacks the nationwide movement to raise academic standards through testing, said Columbus should not be judged on its test scores.
"Tests are not intended to look at how much kids understand and how well teachers are helping kids understand," he said. "The tests are designed for the purpose of artificially spreading out scores so they can rank kids against one another."
Parents Stay Committed
State Secretary of Education Gary K. Hart has heard the anti-testing argument many times. He agreed that test scores alone do not provide a full picture of a school. Still, he said, they provide a key element of Gov. Gray Davis' program to make schools accountable for producing results.
He said the state is not trying to dictate how teachers teach. But, he said, "if we're doing all these discovery-learning type things and people are feeling good about what they're doing but they can't show much in the way of . . . basic skills, it's got to raise concern."
So far, parents have not begun to abandon Columbus.
"I'm not even slightly fazed by [dropping test scores]," said Maureen Katz, a psychoanalyst who is a parent of a first-grader at the school. "For me, what's important is the experience my child is having in the school."
Last year, her daughter, Ally, painted portraits imitating Matisse and worked on a mural in the style of Diego Rivera. But she also learned to read in English and Spanish and, over the summer, she has read five books in each language.
Moreover, Katz said, despite its test scores the school remains popular in the district, which has a process by which parents can request where they want their children to go to school. If the state were to try to punish the school for its low test scores, she said, parents would organize in protest.
Still, parents' expectations are high. Jesus Mena, a former parent at the school, said parents will expect to see scores going up by next year. This year, he said, parents want to get a clearer picture of what students are expected to know so they can keep track of whether the school is meeting its goals.
"We are the recipients of these services and we want to be able to say, 'These things are not working,' " he said.
Wheat says the school is working on an improvement plan. It will receive a $75,000 grant this year to add three hours of after-school instruction to reinforce each day's lessons in reading, math and science and provide a place for children to do homework.
Beyond that, she said, she will convene meetings to ensure that teachers are meeting the state's academic standards for what children should learn.
Finally, the school will apply for a state grant of $50,000 to pay for outside experts to help it figure out a strategy for turning things around.
Like other schools seeking that money, Columbus will have to submit an improvement plan to the state. If the plan is approved, the school could receive an extra $200 per pupil per year. The catch is that test scores must go up by 5% a year or the state can pull the rug from under the school and its leaders and take it over.
Wheat is confident that scores will go up. But, she said, the payoff from the school's comprehensive mix of services won't be seen for many years.
The true test will be how many of her students eventually complete high school, go on to college and stay out of jail. "Test scores are certainly part of the whole picture," she said. "It's part of it. But it's not the whole thing."