Using methodology approved by the U.S. Department of Education, Virginia set new academic achievement targets for schools that do little, if anything, to close the achievement gap for groups of at-risk students.

That move, in July, set off a firestorm of controversy within the state, sparking numerous newspaper op-eds, letters to and from state officials, and even a Facebook page in protest over the low expectations set for at-risk students.

Now, the state and the Education Department—which approved the methodology as part of a waiver for Virginia from No Child Left Behind—are working on a do-over.

In a letter sent today to Virginia schools chief Patricia I. Wright, federal officials have outlined an agreement in which the state will redo its annual measurable objectives, or AMOs, so that students who are furthest behind make the greatest progress. The state board of education will approve new AMOs at its Sept. 27 meeting, the letter says, and submit them for federal approval.

“Ambitious but achievable AMOs that require significant closing of achievement gaps between subgroups of students are a critical element of those plans ...,” writes Deborah S. Delisle, the department’s chief K-12 official.

The problem, both sides maintain, was that Virginia recently adopted new, more rigorous tests and the results weren’t available at the time of the waiver application, so no one could know how the AMOs would play out using the new methodology.

But the Education Department went ahead and approved the methodology anyway—without any conditions—when it awarded Virginia a waiver in June.

The result: Virginia publicized new AMOs that might have been realistic but weren’t very ambitious.

As one example, according to the press release from July, after the test results had become available, the gap in math proficiency rates between black and white students on the 2011-12 state tests was 23 percentage points. By the 2017-18 school year, or five years from now, the goal is only to narrow the gap by 2 percentage points. In the end, Virginia aspires to have 57 percent of its black students proficient in math by 2017-18, compared with 78 percent of white students.

The original methodology behind these numbers is complex, and you can read it on Page 51 of Virginia’s approved application. Basically, it focuses on closing gaps between students and subgroups in low-performing schools versus high-performing schools. It does not focus on closing gaps between at-risk student groups, such as low-income students, and their more-well-off peers. This was one of the hallmarks of NCLB, and something the department says is still critically important.

The new methodology, said Virginia education department spokesman Charles Pyle, will set the benchmarks so the lowest-performing students will be expected to post bigger gains to close the gaps.

The federal department knew there were potential problems with Virginia’s AMO process. The peer reviewers who judged the state’s original application found big problems with that particular section of the application, noting a “lack of clarity” on how academic targets were set. The department followed up in a letter citing a “significant concern” with the “lack of ambitious but achievable” AMOs.

Certainly, Virginia revised its application, obviously well enough to garner federal approval. But the state did not provide any modeling on what the new AMOs might look like, no doubt because baseline results weren’t available.

But those who’ve studied the waiver applications note that the approved methodology, by itself, may have been flawed. Virginia wants to close gaps between the schools at the 90th percentile for achievement versus those in the 20th percentile. So if high-performing schools start doing worse, and low-performing schools stay even, then the gap actually improves— but not because low-performing schools are doing a better job. Should the department have seen this coming?

In addition, the original methodology meant different targets are set for different groups of at-risk kids. So the expectation is that black students will perform worse than white students. This made many people hopping mad.

This controversy aside, it’s important to point out that Virginia’s application illustrates another potential problem with these waivers: The academic goals or AMOs states set sometimes don’t have a lot to do with a state’s accountability system, which is the real driver of school improvement. In Virginia, for example, what matters most is whether a school meets the state standards for accreditation.

“AMOs were never meant as aspirational goals for different subgroups,” Virginia’s spokesman told me.

This also illustrates how hypervigilant the feds are going to have to be as they monitor diverse, complex accountability systems in the 33 states plus the District of Columbia that have won waivers. If it had not been for all the publicity, would the department have so quickly spotted the problem?

The Virginia situation also raises important philosophical and policy questions: How much gap closing is realistic? When is it OK to set different academic targets for different groups of students? And how should the achievement gap be measured?