Who’s failing – the student or the test?

It's hard to get jazzed about four-digit subtraction. As noon approaches on this freezing January day, the kids in teacher Gianna Amsberry's third-grade math class have been shut inside all morning. They're ricocheting off the furniture, trying to impress their crushes, seizing any opportunity to think about something besides carrying and borrowing.

Except for one boy, bent over his worksheet, his raised hand waving hard enough to stop traffic: "I know! I know!"

He does know. He's known the answer to the past five problems. He finished his whole assignment in the time it took kids around him to do the first third. Now, at the board, another student in an orange stocking cap is stuck on 3,516 minus 1,682. At his desk, the boy tries not to shout out the answer, but it's hard.

"What's 11 take away 8?" Ms. Amsberry prompts.

"Four?" tries the kid at the board.

"No!" whispers the boy.

"Three?"

"This is getting better," the boy says, grinning.

This is Bill Clinton Hadam, a 9-year-old who, two years ago, arrived in Atlanta from a Tanzanian refugee camp without a word of English. He came to the International Community School here last year unable to read, and at first could hardly get through a day without bursting into frustrated tears. Now, halfway through third grade, Bill has lots of friends. He's a diligent student, a natural athlete, a sensitive soul, and funny. He's also an obedient son and a protective older brother.

In Georgia's eyes, and those of the US government, he's something else: a failure.

What does it mean, in America's public schools, for a student to fail? What does it mean for a school to fail its students?

The current national framework for answering these questions - the 2002 No Child Left Behind act (NCLB) - could not be more controversial. A 2007 effort to overhaul it ended in congressional deadlock. Calling it "one of the emptiest slogans in the history of politics," President Obama has pledged to rewrite the law.

Meantime, critics and proponents agree on this: Finding a good way to describe and evaluate student success is key to the future of public education.

This year, Bill is studying a new Georgia curriculum that's based on Japanese math standards for his age-group. To pass, he must master place value and basic multiplication, division, fractions, and decimals. He must learn to calculate areas and perimeters, construct tables and bar graphs, use geometric language like "diameter" and "isosceles," and grasp the concepts of mathematical formula and proof. He must demonstrate all this in English, on an NCLB-mandated state test that features paragraph-long word problems and language like "associative property of multiplication."

In April 2007, six months after leaving Mkugwa refugee camp, Bill got 774 on Georgia's first-grade math exam. The minimum passing score is 800. Last spring, after a year at ICS, he scored 784 in math and 760 in reading. Though he'll probably come closer this April, Amsberry worries he's not caught up enough in English to top 800 on the third-grade test.

But Bill's English is good enough that he can make himself understood to anyone. He loves reading, especially Dr. Seuss. He recites his times tables with pride. In math, he's tentative with new material, but once he's learned a concept, he blazes through assignments.

He says he prefers the challenge of ICS to Indian Creek Elementary, the public school he attended his first year in America. "Because," he explains, "we have to do homework first. At Indian Creek we didn't get homework that much, and we don't need to practice, like, multiplication, times tables, all of those kind of stuffs. [At ICS] we always get homework, every single day."

It's frustrating, says Amsberry: Bill will spend 16 days of this 180-day school year filling in Scantron bubbles. But the tests are notoriously unreliable indicators: Kids have bad days, questions have cultural biases, English language learners get stuck on unfamiliar words ("Gretchen," "hen-house" ) and miss problems they know how to solve.

Standardized tests are such blunt instruments, Amsberry says, they're useless to describe the progress a kid like Bill has made. "Tests obviously aren't going to show that once you have a conversation with him about a problem, give examples, and explain what they're asking for, he can get it from there."

This has real consequences for Bill. If he fails the reading portion of Georgia's test this spring, he may not go on to fourth grade next year. If he fails math, he'll wind up in a remedial class again. ICS was designed to nurture refugee kids like him. But there's no such middle or high school to receive him if he's behind when he graduates from sixth grade.

It also has consequences for ICS and the state, which can be sanctioned on the basis of Bill's scores. Last year, when Georgia rolled out a new math exam, ICS failed to make "adequate yearly progress" - NCLB's measure of success - when 38 of 56 fourth-graders scored below the state standard.

The school received accolades last spring: recognition as a national Title 1 school for narrowing the achievement gap between rich and poor students, approval to offer the prestigious International Baccalaureate program of study, and Page 1 coverage in The New York Times. But to succeed under NCLB, it must turn its scores around this year - or face costly changes and possible closing.

ICS teachers and administrators say last year's scores, aberrantly low in the school's seven-year history, were largely due to problems with English comprehension. But politicians and educators say an even greater problem for schools nationwide is the sway these tests hold in determining whether a student and her school are failing.

No serious educator argues with NCLB's central premise: Every child deserves a world-class education, and the nation should hold itself and its institutions accountable if they fail to provide one. It's crucial to assess kids, they say - not just yearly, but daily, to know what to teach them tomorrow. Most educators dispute the wisdom of using standardized tests, or any single indicator of student progress, to decide a kid's - or a school's - future.

"How can one exam possibly be the factor in whether they're going to go to fourth grade? That's a joke to me," says ICS assistant principal Tahisha Edwards. "And I think it's an unfair factor as well that the CRCT" - Georgia's Criterion-Referenced Competency Test - "is used to decide whether we [as a school] are a success or failure. It doesn't say anything about what we do - or any school does."Nor does it say that Bill Clinton feels like such a success in math, he's champing at the bit to get to harder material. It doesn't say he can now read and write English as well as his Congolese father. It doesn't say these things are possible because ICS's founders, staff, and parents conceived and built a school rich with volunteer tutors and cultural celebrations, that first fed and comforted Bill, then offered him an array of services to help him learn in a new language and environment.

It does, however, say some things about weak spots. This year, the school is making a big push in math: training teachers, seeking the help of an instructional coach, piloting a new textbook in its remedial classes, and assessing student progress with benchmark exams at the start and end of each semester. The benchmark tests third graders took in August and December, miniversions of the ones they'll face this spring, showed some patterns, says Ms. Edwards. The kids made gains in measurement. They still have trouble with word problems and computation, but their tests were revealing. Take computation, your basic 120 x 95 =___. Kids would often get such a problem right - but then, accustomed to practicing on worksheets of 20 identical problems, they'd continue down the page, multiplying on problems that asked them to subtract or divide. They were also easily taken in by "close distractors" - answers designed to mix them up, like 52 where 25 was correct.

In word problems, language seemed key. One problem featured the commutative property of multiplication: the principle behind 5x3 = 3x5. Simple. But the name? "Everyday Mathematics" - the University of Chicago textbook series that ICS has used since its founding - refers to this as the "turnaround rule." Anne Garbarino, who teaches an ICS third-grade section down the hall from Bill's class, says native English speakers solid on the "turnaround rule" missed the question because they couldn't name the property behind it.

In the end, Ms. Edwards says, it appears most kids who scored below 50 percent on the exam did so not because they didn't know math, but because they weren't savvy test takers. And they need to be in order not to get left behind under "No Child."

"And that concerns me," she says, "because they're 8. Can't they just be kids?"

Thursday:How best to teach kids who are learning English as a second, third, or fourth language even as they learn to read and and multiply.