Guestpost by Anne Hyslop

The ed policy world has finally agreed on something: there is too much testing. Now it may not win me any Twitter followers, but this consensus makes me nervous. Mostly because it makes hasty, extreme solutions to “over-testing” seem tenable, giving them credibility as a logical response because “this is a crisis.” Is it? Teach Plus has shown that, on average, less than 2 percent of class time is spent on mandated testing. While there are outliers, it looks like the excess is coming from the local level, not state tests. And like my colleague Andy Smarick, I see the virtues in our current testing regime, and the consequences in eliminating it without understanding what could be lost.

So I was glad that large urban districts and chief state school officers are working together to tackle issues of assessment quantity, and quality, while maintaining a commitment to annual testing. Same goes for the Center for American Progress’ work on “better, fewer, and fairer tests.” All common sense responses to the over-testing meme. And given growing numbers, especially on the political left, calling for grade-span testing (see: teachers unions, members of Congress, former President Clinton), it is welcome to see a defense of annual testing–with support from Arne Duncan, and even President Obama.

But are they really defending them? On second glance, I’m not so sure. Even the staunchest supporters of grade-span testing, like Randi Weingarten and Linda Darling-Hammond, would support giving students tests each year, just with a caveat: local assessments without consequences, not statewide. As Darling-Hammond, Gene Wilhoit, and Linda Pittenger describe in a new brief, statewide, grade-span testing merely serves to “validate” the results of the annual local tests–while eviscerating most meaningful accountability systems in the process (not a coincidence).

In other words, the right question to ask is not, “do you support annual testing?” but rather, “do you support annual statewide testing?” And despite outward appearances, CCSSO’s and CAP’s support is more tenuous. That’s because both seem ready to embrace district flexibility (read: opt-outs) of state tests, especially in “districts of innovation.” Their new report “Next-Generation Accountability Systems: An Overview of Current State Policies and Practices” includes multiple examples of district opt-out plans, from New York to Kentucky to New Hampshire, and holds them up as models for the future.

“Districts of innovation” is code for districts that are exploring competency-based learning, or project-based learning, or some other (usually) technology-enabled reform to personalize students’ experiences. All good ideas, in theory. But that’s often what they are: just theories. We don’t actually know if they work yet to improve student outcomes. And in order to find out, we must evaluate them. So let’s take the Darling-Hammond approach and use statewide tests as a “validator” of what’s happening at the local level in one of these innovation hotspots.

Located in Danville, Kentucky, Bate Middle School was profiled by NPR’s Anya Kamenetz this year in a piece originally titled “In Kentucky, Students Succeed Without Tests.” Kamenetz paints the picture of an academic renaissance at Bate, which had been slapped with the “needs improvement” label by the state’s accountability system. This renaissance was possible all because Bate chose to forego administering state tests and, instead, tapped into students’ interests with project-based learning and performance-based assessments that were evaluated locally. Except, Bate didn’t get a waiver to skip the standardized tests, as first reported.

And thank goodness for that, because when you examine the latest data from the Kentucky Department of Education, results at Bate are a little more complicated. The school got solid marks–full credit and close to “distinguished”–in the “program review” component, which looks at instruction and curriculum, formative and summative assessments, professional learning, and leadership (all of which should reflect the project-based learning strategy).

But on the statewide, Common Core-aligned assessments, the outlook was decidedly less rosy. Half of students were at least “proficient” in reading, but only 37% of low-income students met that benchmark. In writing? 36% and 27%, respectively. The results in math were most distressing: only one-third of students were proficient, and just one in five low-income students could say the same. Minority students struggle, too. And all of the results were below state averages.

Proficiency rates are imperfect measures, though, often correlated with student characteristics. For that reason, they aren’t especially useful for diagnosing how Bate is contributing to its students’ learning. Further, this is a school in improvement–it’s reasonable to expect that the proficiency data wouldn’t be great. That’s why it’s so important that Kentucky has annual statewide assessments that can measure student growth. Is Bate, with its project-based learning experiment, accelerating learning more than other schools in the state?

The unfortunate answer is “not really.” Growth is only measured in reading and math, because they are the only statewide tests given annually (it’s not feasible to measure growth using grade-span tests, or the local measures–a big reason to keep annual state tests). In reading, 56% of Bate students were found to be making typical annual growth toward college and career readiness. Put another way, nearly half of students, even with this new project-based approach, were not making typical growth, let alone the high levels Bate needs for its most disadvantaged kids to get on track academically. The math results were, again, worse: only 43% of students were making typical growth, while the majority were not.

I’m not saying Bate should give up its project-based experiment. These data are by no means a thorough program evaluation, and there were other benefits to the reforms, based on Kamenetz’s reporting–like increased student engagement, community involvement, and educator morale. That should not be discounted or overlooked, and other results could improve with time.

What I’m saying is that Bate should not give up statewide annual assessments. Not yet. The only way to know whether “districts of innovation,” like Danville, are working is by examining all the evidence, not just the evidence local officials like. And in this case, it’s pretty clear the local approach is not ready to be “validated.”

Maybe the education field will be ready for a system of district “validated” assessments one day that are both high-quality and comparable to state tests. But that day is not today. And the only way that day will come for these reforms, whether it’s project-based learning or competency-based education, is to show that they work. And that’s going to require statewide annual testing for the foreseeable future. Instead of talking out of both sides of their mouths on testing–for annual tests and for district flexibility–CCSSO and CAP should clarify: we need statewide annual assessments now… and to demonstrate that district flexibility is a worthy alternative in the future.

Anne Hyslop is a senior policy analyst at Bellwether Education.