Under the accountability structure put in place by the No Child Left Behind Act of 2001 (NCLB), states are required to bring all of their students to proficiency in math and reading by 2014. Schools and districts are held accountable for the performance of students in particular “subgroups” as determined by race, socio-economic status, and participation in certain programs like special education. Because states aren’t going to achieve this goal, the Department of Education announced in November 2011 that states could apply to have the requirements waived if they proposed – and the Department accepted – alternative student performance plans.
Many have argued that the waivers are turning the original intent of NCLB on its head. States have submitted extremely complicated, opaque, and watered-down accountability measures. Look no further than the myriad ways states propose to measure schools by how student subgroups perform.
Colorado, Florida, and New Mexico are three interesting cases. Despite the Department’s guidance requiring states to include data from individual subgroups in their plans, Florida and New Mexico chose to forgo subgroups defined in NCLB and instead create one “low-performers” subgroup. Colorado, on the other hand, hews to NCLB subgroups and the Department’s guidance, but still makes key changes. Below we discuss how each state’s approved plan treats student subgroups.
Colorado uses subgroup performance in three components of its performance framework: what the state calls growth; growth (or achievement) gaps; and postsecondary- and workforce-readiness. The framework accounts both for growth that English Language Learners make in traditional categories, as well as improvement in their English skills as measured by the Colorado English Language Acquisition Proficiency Assessment (CELApro). The state also disaggregates graduation rates and scores on state math and reading tests by subgroup to determine the degree to which schools are effectively reducing the achievement gap. In total, subgroup performance makes up 37.5 percent in a school’s grade in Colorado. So, while Colorado may not give the same focus to subgroups as they received under NCLB, the state still provides substantial weight to the performance of traditionally high-needs students.
Florida more or less scraps the NCLB subgroup framework and focuses instead on the 25 percent of students showing the least improvement from year to year, as well as students deemed “at-risk.” At-risk students are those who enter high school below grade level based on their eighth grade Florida Comprehensive Assessment Test scores in reading and math. Although in theory this measure likely captures the performance of subgroups, it has the potential to mask the degree to which individual subgroups are struggling or succeeding.
The state uses test outcomes of the lowest-performing 25 percent of students in its measure. But the state further complicates the issue by choosing to focus on the graduation rates of at-risk students instead – a likely overlapping but different population of students. In total, at-risk and low-performing students make up about 18.75 percent of a school’s grade in Florida’s model. But the metric also includes the performance of “acceleration” students, accounting for 18.75 percent of a school’s grade, too. (Acceleration refers to the number of students who took assessments higher than their grade-level and the scores on those tests—the top performers at schools.) By including the performance of accelerated students in a school’s grade and giving it the same weight, the state essentially negates the impact of struggling students on those grades.
Similar to Florida, New Mexico does not include disaggregated subgroup data for student growth or graduation rates. Instead, the state accountability metric separately accounts for the performance on state math and reading tests of the top 75 percent and lowest 25 percent of students. And rather than accounting specifically for the graduation rates of low-performing students, New Mexico includes graduation rates both for students who graduate in four years and those who graduate in five years. In total, low-performing and high-performing students each account for 15 percent of a school’s grade, and five-year graduation rates make up four percent of the grade. New Mexico, too, seems to allow the performance of its top performers to overshadow that of its lowest-performers in its metric.
Why did the Department grant the waivers to these states even though some of them stopped disaggregating subgroups—the explicit goal of NCLB? Certainly, the Department needed to give states a way out of the current system. But did they throw the proverbial accountability “baby” out with the bathwater? Or did the sheer complexity of the state proposals, some over 500 pages, obscure the reality of the new accountability measures by making them appear more rigorous?
It remains to be seen if these new metrics will dramatically change states’ accountability outcomes. However, by allowing states to develop their own accountability systems, the Department likely enabled states to minimize the impact of low-income, limited English proficiency, and minority students on their school accountability measures. This could ultimately give schools an opportunity to ignore the needs of these students, defeating the purpose of federal funds specifically targeted to them.