Delaware deserves credit for lowering its n-size from 30 to 15.
Lowering the n-size will ensure more schools are paying attention to more subgroups of students, and the state provides compelling data on how many more students will be included as a result of this change. Because the lower n-size will also add extra volatility, the state may want to consider running statistical tests to monitor year-to-year swings to see if they reflect real changes, and if not, including multiple years of data in its calculations.
Delaware will not be incorporating subgroup scores into the state’s school-rating system.
While not ideal, this could be acceptable if Delaware outlined clear, rigorous criteria to identify “targeted support” schools with low-performing subgroups. But, although the state plans to identify some schools under this definition, it’s unclear if the state is planning to identify 5 percent of each subgroup (i.e., the 5 percent of schools with the lowest-performing subgroup of black students, the 5 percent of schools with the lowest-performing subgroup of students with disabilities, etc.) or if the state is envisioning one composite group based on all subgroups.
Delaware has given deep consideration to its methodology for measuring the performance of English learners.
However, it weakens that approach somewhat in its school accountability system. It proposes to delay the full inclusion of newly arrived English learners in its accountability systems until they have been in Delaware schools for four consecutive years, which excludes students who are most at risk of being left behind.