Miles To Go Before We Sleep

“We have an informational mess at the moment because we have a lot of state-by-state information that just doesn't make sense,” said Mark D. Musick, president of the Southern Regional Education Board, in his keynote address.

NCLB requires schools to test students in grades three through eight each year, and to demonstrate annual yearly progress toward goals that each state sets for itself. The premise sounds straightforward, but it is like a trick question, full of opportunities for failure.

The problems begin with the requirements for the tests. NCLB mandates that they fully address the depth and breadth of the state's academic content standards; be valid, reliable, and of high technical quality; and express student results in terms of the state's standards for student academic achievement.

“The law has very lofty goals for these tests, but it doesn't define valid and reliable,” said Robert Linn of the University of Colorado and the National Center for Research on Evaluation, Standards, and Student Testing (CRESST).

The states are left to set their own parameters for performance. “As you might expect, there is huge variability among states,” said Linn. Some states have lowered previous standards to make it easier to meet NCLB requirements. Most have set easy goals for the early years, and mountainous expectations for themselves later. “Most states have set targets that delay the pain.”

The National Assessment of Educational Progress, or NAEP, which measures fourth and eighth grade reading and math skills, is the only tool that comes close to providing a nationwide measure of progress. Although it did not report complete nationwide data until 2003, it quickly became the benchmark against which performance standards for many states were compared. The results of these comparisons do not suggest progress is being made.

“None of the states have shown rates of growth that would project anything close to 100 percent proficiency in 20 years, much less by 2014,” as NCLB requires, said Linn.

As of November 2003, NAEP results are being reported for all 50 states, which makes NAEP a better benchmark. “With fifty-state data, a different level and kind of discussion is possible, and I think it's a good one,” Musick told the researchers.

But more complete data still won't fix what's wrong with NCLB's approach to testing, said Richard J. Stiggins, the president of the Assessment Training Institute, Inc. The problem is that high-stakes assessments operate on the assumption that maximizing anxiety will maximize learning.

“We have to rethink the relationship between assessment and student motivation,” Stiggins said. “Competent test users rely on continuous classroom assessment in support of learning, plus periodic assessments to verify learning.” The crucial distinction is between assessment OF learning—how much have students learned as of a particular point?—and assessment FOR learning, or, how can we use assessment to help students? Effective education must use both, but NCLB is built upon only one.

The key components of achievement gains are nothing new, he added. Educators know from past experience that classroom assessment accuracy, continuous descriptive feedback, and student involvement in assessment lead to progress.

Linn echoed the thought.

“If you ignore what has been accomplished in the past, you're setting yourself up for failure,” he said.