This is the central tenet of the Forum on Educational Accountability, which was convened by the National Educational Association and FairTest and joined by more than 100 organizations. In its report on assessment, which provided the substance of a letter sent to Congress in early August, the FEA calls for the federal accountability system to be based on:

Coherent and comprehensive assessment systems [that] provide evidence of student
and school performance in relation to rich and challenging educational goals,
using multiple indicators of student learning from a variety of sources at multiple
points in time.

Such multiple indicators would include “teacher observations; tests that include
multiple-choice, short and longer constructed response items; essays; tasks and
projects; laboratory work; presentations; and portfolios.”

            Certainly schools should have such assessment systems, and in fact they do—they are called classroom grades, which can be based on all those assessment tools and more, depending on the professional judgment of the educators in a given classroom or school.
            But one of the main points of a federal accountability mechanism is to provide a quick check on whether those classroom grades have any grounding in reality—that is, to see whether students who receive As and Bs can at least read and do math at the minimum levels established by their state.
            After all, plenty of students have passed their classes without mastering the skills and knowledge that most people think are appropriate for their age. Powerful evidence exists that this is a much more common phenomenon for children of poverty and children of color than for their more privileged peers, in part because of teachers’ lower expectations. That is why national policy makers from both parties thought it necessary to have some kind of accountability system that would allow parents and community members to have an outside judgment of how students and schools are doing.
            For all the hysteria around it, the federal accountability system of No Child Left Behind is relatively straightforward and easy to explain: All kids in third through eighth grade and once in high school in a given state take the same reading and math tests. Their states report individual results to parents to let them know if their children are meeting state standards and report group results to the public so that parents and community members can see whether schools are doing the jobs we expect them to do. Each state sets goals for schools to meet and any school that doesn’t meet the goals must develop a plan for how it will improve the following year.
            Imagine the complexity of an accountability system based not on standard tests taken by all students in a state but rather on teacher observations, presentations, and portfolios. How could parents and members of the public be sure that all teachers had the same standards for observing student performance and judging portfolios?  Would a science project in one classroom be judged on the same standards as a science project in another classroom in another school in another city? Would some teachers hold their students to much higher—or lower—standards than others? To set a single standard for performance would be almost impossible.
            Besides, it would be hard to imagine a more cumbersome federal intrusion on classrooms. Take, for example, the call by the Forum on Educational Accountability to use formative assessments as part of the federal accountability system. Formative assessments are given to find out where a student’s knowledge is on a certain topic.If a student hasn’t mastered enough of the class content, you actually want him to perform badly on a formative assessment in order to signal his teacher that he needs more help. To include formative assessments in any kind of accountability—including grades—would not only subvert the entire point of a formative assessment but bring the long arm of Washington into almost every reading and math lesson in the country.
            Certainly most of the state tests now in place leave a lot to be wished for in terms of quality and rigor. The entire education and policy community should be calling for better quality tests. But to muck up the federal accountability system with all kinds of other imperfect measures wouldn’t solve the problem of poor test quality—it would simply compound it.