I have a new piece out in Education Week that focuses on teacher-designed assessments. In it I argue that while teacher designed assessments can be more beneficial to student learning than commercially prepared assessments, teacher survey data suggests that most teachers don’t feel they have the appropriate skills to design high-quality assessments:
National teacher polling data suggest that I was not alone. A 2016 Gallup poll found that roughly 30 percent of teachers do not feel prepared to develop assessments. Less than 50 percent of teachers in low-income schools reported feeling “very prepared” to interpret assessment results, and less than 50 percent of teachers said they’d received training on how to talk with parents, fellow teachers, and students about assessment results. More alarming is that no state requires teachers to be certified in the basics of assessment development, so it’s likely that many teachers have never had any formal assessment training.
I highlight work underway in New Hampshire and Michigan to make significant investments in assessment literacy training for educators. More states should follow the lead of these exemplars and commit to equipping all educators with the tools to develop high-quality, rigorous assessments.
I have a new piece out in GOVERNING Magazine discussing innovation in state assessments, and why local and state officials should invest in improving their assessment systems instead of cutting back. I highlight work underway in New Hampshire and Louisiana, which have both received waivers from the federal government to do something different with their tests. Just as the piece came out, Georgia and North Carolina got approval from the Department of Education for their own innovative assessment plans. But there’s a lot states can do even without special federal approval.
An excerpt of my op-ed:
“Test” has become a four-letter word in schools, as many states face political pressure to cut, minimize or deemphasize their much-maligned annual standardized assessments of student achievement. The most common complaints are that these tests do little to help teachers do their jobs well and can distract from more important aspects of teaching and learning.
But if standardized state tests aren’t useful in the classroom and aren’t informing instruction, that’s a problem that can be fixed even with current federal law mandating annual tests in math and reading. Instead of indiscriminately cutting back on statewide testing, states need to think about approaching them differently and look beyond typical end-of-year tests. Reducing investment to the barest minimum could leave students and schools worse off, without good information on achievement gaps, student growth, or college and career readiness.
Just a few years ago, it seemed like most of the country was heading towards common state assessments in math and reading. Two groups of states won federal grant funds to create higher-quality tests; these became the PARCC and Smarter Balanced test consortia. Now, despite the demonstrated rigor and academic quality of those tests, the testing landscape is almost as fractured as it was before, with states pursuing a variety of assessment strategies. Some states in the consortia are still waffling. Others that have left are already scrapping the tests they made on their own with no idea of what they’ll do next.
States should think carefully before going it alone or introducing a new testing overhaul without strong justification. There are some big tradeoffs at play in the testing world, and a state might spend millions on an “innovative” new test from an eager-to-please vendor only to find that it has the same, or worse, issues as the “next generation” tests they tossed aside.
But annual testing is only half the story. That’s because Alexander’s bill doesn’t just offer two statewide testing options for policymakers to fight about. It also offers a separate testing option fordistricts on top of the state choices. And although education wonks are up-in-arms over the merits of door #1 vs. door #2 for states, most have, unfortunately, ignored the giant local testing loophole that is behind door #3.
Through it, districts could opt-out of statewide testing and use their own tests instead, regardless of whether Congress chooses door #1 or door #2. But the real kicker is that this loophole isn’t actually new at all. Alexander’s draft bill just makes it far easier for districts to take advantage of–and abuse–existing flexibility.
Districts would only need state approval that their local assessments meet the same federal requirements with which state tests comply. And given the increasing number of districts pushing back on state testing, door #3 would be an irresistible option for many, even as it undermines the comparability of data between schools for evaluation and accountability; states’ abilities to provide technical assistance, support, and professional development to districts; and state investments in new assessment systems aligned to college- and career-ready standards.
The American Federation of Teachers (AFT) and the Center for American Progress (CAP) have released a joint set of principles for ESEA reauthorization. They call for preserving statewide annual testing requirements for students, but they would base school-level accountability only on tests taken once per grade span—once in elementary school, once in middle school, and once in high school.
Like the Education Trust, we think this is a bad idea. Grade-span accountability solves none of the problems of our current system while making other problems worse. Namely:
It doesn’t address concerns about over-testing. Students could be taking the same number of tests as they have in the past, particularly if districts don’t reduce the number of duplicative and unnecessary local tests. CAP has rightly cited these local tests as the root of the problem, but this proposal would not reduce the number of federally mandated tests.
Rather than decreasing the stakes on standardized tests, the AFT/CAP proposal would amplify them. Under their plan, a 5th grader would no longer be taking tests that reflect just on the 5th grade. His or her results would be the basis on which their entire school was judged. How, exactly, does this help “de-link [academic standards] from high-stakes tests”, as AFT President Randi Weingarten suggested a year ago?
It makes it even harder to focus on specific subgroups. NCLB held schools accountable for every subgroup that had a sufficient number of students (called the minimum “n-size”). But under the CAP/ AFT proposal, a school’s 5th grade African-American, ELL, or SWD groups could be too small to meet the minimum n-size and the whole school’s disadvantaged students could go uncounted. This may sound wonky and technical, but it becomes a pretty huge issue even at relatively small n-sizes (such as 10 or 20 students). Arne Duncan has estimated that hundreds of thousands of students were invisible to state accountability systems because of n-size issues. CAP has praised states in the past for lowering their n-sizes, but their plan to have fewer students “count” toward a school’s accountability rating would mean less attention on important subgroups of students.
We already have anecdotes about teachers who prefer to avoid tested grades and subjects. They may prefer teaching in 2nd grade, where there are no required standardized tests, than 3rd grade, where there are. But it’s tough to avoid the current tests altogether because they’re given in 3rd through 8th grade. Grade-span testing would make it even tougher to attract teachers into those few areas with much higher stakes. Who wants to be a 5th grade teacher when they might responsible for their entire school? In most places, they won’t even earn any extra money for all the added pressure! Moreover, this is exactly the kind of policy the AFT previously opposed for teacher evaluations. A year ago, Weingarten wrote: “In Florida, the system went completely haywire, giving teachers value-added scores for students they had never taught.” If it’s not okay for educator accountability, why is it okay for school accountability?
Standardized tests are often criticized for merely reflecting student demographics. While states and districts have been slow to implement accountability systems that incorporate student growth, with annual statewide testing, we at least had a hope of shifting attention to how much progress students make over time. CAP and the AFT once shared this hope. Yes, not all that long ago AFT advocated for an ESEA that “judges school effectiveness—the only valid and fair basis for accountability—by measuring the progress that schools achieve with the same students over time.” With longer gaps between tests that count for accountability purposes, we’re more likely to lean even heavier on raw test scores, measures that are highly correlated with student demographics.
Under this plan, students and families would still get a sense of how much progress they’re making. That’s important, but it’s odd to then turn around and suggest that states and school districts should ignore this same information for determining school progress. As CAP’s 2011 NCLB recommendations suggest, “Measuring and reporting student data is not sufficient to improve our nation’s schools. Congress should take several steps to ensure schools act on that data to boost student outcomes.” We assume they did not mean several steps backwards.
AFT and CAP pitch their proposal as targeting interventions to schools with large achievement gaps. That’s true, it would identify schools with gaps. But, ironically, it would give no credit to schools that are actually closing those achievement gaps. CAP used to support annual gap-closing goals. But now, schools with large concentrations of economically disadvantaged and minority students, English Language Learners, or students with disabilities would all be penalized unfairly, worse than they are under NCLB.
Ultimately, this plan would move us closer to how other countries do testing: fewer tests with much higher stakes. Rather than having regular check-ups on student progress, with relatively low stakes on those results, we’d have much higher stakes attached to a smaller number of test scores. Fortunately, AFT and CAP have already told us why this is a bad idea.