What's the right balance between pushing schools to reach a goal and giving them credit for making progress, even if they fall short?
That's a key question lawmakers are considering as they prepare to reauthorize the federal No Child Left Behind Act (NCLB), set up five years ago with the hope of closing achievement gaps in public schools.
The call has been coming from all corners to incorporate "growth models" into the way schools are judged. Rather than simply comparing this year's group of third-grade or eighth-grade scores with last year's, advocates say, states should also track the progress of students individually, and see how much they learn over the course of a year.
Under the current system, a school in Milwaukee with only 2 percent of its students scoring "proficient" in reading and math progressed to about 42 percent proficient in a few years. But because that number was still below state targets under NCLB, it received the same label as schools that may have made no progress at all, according to the recent report by the bipartisan Commission on No Child Left Behind.
Expand credit for improvements
Everyone from civil rights groups to teachers' unions to the secretary of Education herself has acknowledged that giving schools some credit for their students' improvements makes sense. For instance, they say it would be more fair to schools with a high turnover of students (common in high-poverty areas), because they could show the learning outcomes of students actually taught there that year, versus newcomers, who may have low scores reflecting their previous experience.
"If we [want] to improve the effectiveness of schools, how are we going to do that if that's not what we're measuring?" says Paul Barton, an educational consultant and former director of the Educational Testing Service Policy Information Center. By simply looking at whether a certain percentage of students score proficient on an end-of-year standardized test, he says, states are measuring what students have learned in their entire life up until that point – which is affected by many factors other than what they are taught in school.
The concept resonates with the public: According to a Phi Delta Kappa/Gallup Poll last year, 81 percent of Americans say they favor looking at the improvement students make during the year, rather than the percentage passing a test at the end of the year.
The mechanism that would be revised is known as AYP – the requirement that schools make "adequate yearly progress" toward the ultimate goal of all students scoring proficient on state tests in reading and math by 2014.
AYP is "a blunt instrument," said Tommy Thompson, co-chairman of the bipartisan commission, whose report includes 75 recommendations for improving NCLB.
If a school does not meet AYP for all subgroups of students (such as ethnic groups, low-income students, and English-language learners) for two years in a row, it must provide tutoring options and allow students to transfer out. Eventually, schools can be taken over by the state or forced to restructure.
Many school administrators have expressed frustration that their school is perceived as failing if even one subgroup falls short by a few points. That's why the bipartisan commission also suggested that schools should only be judged as not meeting AYP if the same subgroups fell behind the goal for two years in a row.
Test scores vs. tracking students
The current law relies heavily on a testing system designed for a different purpose – to sort and track students based on whether they were performing at a level that was average for their grade.
"When we came along with sanctions and accountability, we simply took the tests we had. Nobody started over from scratch," Mr. Barton says. But accountability for closing gaps is a different goal and requires a different kind of testing, he and many others insist.
Counting student progress has precedent in a few states, such as North Carolina. Since 1997, schools there have been measured by a combination of proficiency and improvement.
NCLB "made a complicated system even more complicated," says Lou Fabrizio, director of accountability services with the North Carolina Department of Public Instruction. "Schools would be designated as meeting a growth component under our program, but they would potentially be listed as not making AYP, and that [causes] a little bit of confusion," he says.
But recently he's been glad to add yet another layer of complexity. North Carolina is one of five states that the US Department of Education is allowing to experiment with growth models in AYP calculations.
After one year, "it did not help as much as we would have hoped," Mr. Fabrizio says, but it's too soon to judge how many more schools might be able to show progress under such a system. (Last year, 45 percent of the state's schools met AYP.)
The US House Committee on Education and Labor held a hearing last month on growth models and other assessment issues. A committee staff member says the research in support of growth models is almost overwhelming, and it's no longer a matter of whether to use them, but of how. The committee aims to prepare a reauthorization bill by this summer, though a full vote by Congress on such bills could be delayed until after the 2008 elections.
One point of contention is whether all students should be expected to reach proficiency by 2014. Some critics say that goal is artificial.
"We need to start in reality instead of a fantasy," says Monty Neill, executive director of FairTest in Cambridge, Mass., one of more than 100 groups that have signed a Joint Organizational Statement on NCLB. Schools serving high-needs populations should be expected to aim for a growth rate based on what's already being achieved by schools doing well with similar populations, he says.
But US Secretary of Education Margaret Spellings calls the 2014 goal one of her "bright line principles."
The DOE put out a proposal earlier this year that noted, "a growth model is a tool to achieve proficiency by 2014, not a loophole to avoid it."