It’s hard to tell what the theory of action behind the latest ESEA bill is. As far as I can tell, the only accountability–such as it is–is targeted at the bottom 5 percent of schools. Even in those schools the interventions are weak, but the bill has no real ambitions to improve outcomes for students in the remaining 95 percent of schools.
Think I’m exaggerating? Here’s where things stand:

• The bill would not require any state to upgrade their standards. States must simply provide an “assurance” that their own standards are “challenging.”
• The bill would not require any state to improve their accountability system. States must design an accountability system that includes math and reading achievement, graduation rates, and at least one other measure. It seems to have slipped under the radar, but these are the exact same requirements as NCLB! States could, and probably would, continue to do exactly what they have been doing.
• The bill includes some flexibility that states may take advantage of–like adding other measures to accountability systems, adapting tests for students who are above- or below-grade level, or adding student growth to accountability systems–but history suggests states will leave a lot of this flexibility on the table.
• The bill would require states to identify “targeted support” schools with large achievement gaps, but there are no parameters on how states should define this category. It could be 0 schools, or 10 schools, or 1,000 schools. Worse, after the state identifies however many schools it chooses, all the schools would have to do is write an improvement plan of their own devising. Any plan will do.

That leaves the bottom 5 percent of schools. The bill defines this group, known as “comprehensive support” schools, using a very similar definition as the School Improvement Grant (SIG) program and “priority” schools under the NCLB waivers. Based on what we know of those past efforts, comprehensive support schools are likely to come from certain groups of schools. They’re likely to be disproportionately made up of high schools, urban schools, schools with large minority populations, and schools with large groups of English Language Learners (ELLs) and students with disabilities. A few points of evidence on why I think that:
– Why high schools would be disproportionately identified: High schools tend to have lower proficiency rates than elementary or middle schools, so any rank-order list of schools is likely to turn out more high schools. This pattern is common in many states. For just one example, here in D.C. 50 percent of 3rd-8th graders scored proficient in math, compared to just 34 percent in high school. A rank order list based on proficiency rates is going to make high schools look “worse” than middle or elementary schools, even though the neighborhoods and families are the same.
The second reason high schools will be over-identified is a special rule in the bill, which requires that states identify any school with a graduation rate less than 67 percent as a comprehensive support school. That will force more high schools onto the lists, and it will play out very differently depending on the state. Some states have practically no high schools under the 67 percent threshold, while some states, especially Southern ones, will be forced to identify far larger numbers of high schools (See Figure 3 here for a representative map). SIG data confirm this hunch, because high schools were vastly over-represented in the SIG program using nearly the same definition.
– Why urban schools would be disproportionately identified: This depends somewhat on how states create their accountability systems, but urban schools tend to have lower proficiency rates. If states continue to weight those heavily in their accountability systems, as the bill maintains, the SIG data suggests that states will continue to disproportionately identify urban schools as the most in need of improvement.
– Why minority schools would be disproportionately identified: Same as above. While black students represent 16% of the enrollment nationwide, they made up 39-43% of SIG school enrollment, depending on the year. On the other side, while 51 percent of students nationwide are white, SIG schools were only 11-21 percent white.
– Why schools with large populations of ELLs and students with disabilities would be disproportionately identified: States must also add to the “comprehensive support” schools list any school with large achievement gaps (as defined by the state). But as Morgan Polikoff demonstrated in a paper looking at potential lists of low-performing schools in California, this subgroup analysis is likely to identify schools that are large (because they have larger sample sizes), that have more subgroups (that could each potentially fail the test) and that have more ELL students and students with disabilities (who tend to have lower scores).
By focusing on just a small group of largely predictable schools, this bill accepts the status quo for 95 percent of students. Maybe that’s the authors’ intent, but it’s hard to see how it will help improve educational outcomes for the vast majority of kids.