Ohio leaders have started an important conversation about education policy under the Every Student Succeeds Act. One of the central issues is what accountability will look like—including how to hold schools accountable for the outcomes of student subgroups (e.g., pupils who are low-income or African American). Ohio’s accountability system is largely praiseworthy, but policy makers should address one glaring weakness: subgroup accountability policies.
The state currently implements subgroup accountability via the gap-closing measure, also known as “annual measureable objectives.” Briefly speaking, the measure consists of two steps: First, it evaluates a school’s subgroup proficiency rate against a statewide proficiency goal; second, if a subgroup misses the goal, schools may receive credit if that subgroup shows year-to-year improvement in proficiency.
This approach to accountability is deeply flawed. The reasons boil down to three major problems, some of which I’ve discussed before. First, using pure proficiency rate is a poor accountability policy when better measures of achievement—such as Ohio’s performance index—are available. (See Morgan Polikoff’s and Mike Petrilli’s recent letters to the Department of Education for more on this.) Second, year-to-year changes in proficiency could be conflated with changes in student composition. For example, we might notice a jump in subgroup proficiency. But is this an indication of gap closing? Not necessarily: It might be explained by a change in the subgroup’s student composition.
Third, and perhaps most importantly, while reducing the achievement gap remains an important goal, policy makers should not explicitly pit one group of students against another in accountability systems. Unfortunately, this is what the gap-closing component does: It compels schools to disproportionately focus on certain subgroups at the expense of others.
So let’s scrap the gap-closing measure and start over. But how should Ohio proceed?[1]
In my view, state policy makers should create a new report card component dedicated to subgroup performance. It would rely on disaggregated performance index scores (a status measure) and disaggregated value-added scores (a growth or longitudinal measure). Ohio already breaks down value-added scores by three subgroups and would just need to extend those efforts to additional subgroups. The state would need to introduce a subgroup performance index, although that calculation is relatively simple and straightforward. The component could look something like the following (more subgroups could be added, such as gifted students or homeless students, and weights could be altered):[2]
Table 1: Hypothetical subgroup performance report card component
Subgroup | PI Grade | VA Grade | Points Earned for Subgroup |
Race/Ethnicity: Asian | C | D | 1.5 |
Race/Ethnicity: Black | B | C | 2.5 |
Race/Ethnicity: Hispanic | A | D | 2.5 |
Race/Ethnicity: Multiracial | D | D | 1.0 |
Race/Ethnicity: White | B | A | 3.5 |
Students with Disabilities | A | C | 3.0 |
Limited English Proficiency | D | D | 1.0 |
Economically Disadvantaged | C | A | 3.0 |
Composite Subgroup Performance | C | 2.25 | |
* Assigning points in the following way: A = 4; B = 3; C = 2; D = 1; F = 0; equally weighting PI and VA grades and across the various subgroups; rounding the average composite number of points at the half-point interval when making the conversion to a letter grade (e.g., 2.25 rounds to 2.00 = C).
A component such as this should ensure a more technically sound and transparent way of holding schools accountable for subgroup outcomes. The approach would assign responsibility for the achievement and growth of both typically higher- and lower-performing subgroups. It would also send the right message. Here in Ohio, our approach to ratings is streamlined (for the most part, just two key measures) and fair (balancing growth and achievement). Our accountability system needs to work with schools to make certain that all students, no matter their background or starting point, can grow academically and reach their potential.[1] Under ESSA, Ohio will need to implement some type of subgroup accountability measure to identify schools with low-performing subgroups. It may not have to be a standalone report card component or an A–F graded measure as displayed above.
[2] If Ohio goes this route, the state would probably need to disaggregate whichever subgroups are graded on the performance index to also be graded on value added. In other words, Ohio likely could not disaggregate its PI scores for English language learners without disaggregating their VA scores.
