Secretary of Education Arne Duncan has just drawn a very confusing line in the sand over standardized testing. A new announcement this week allowing states flexibility to avoid double-testing students on both existing state assessments and the coming field tests for new PARCC and Smarter Balanced assessments is likely to have far-reaching and not entirely positive implications.
Let us explain. The new flexibility allows any state (and any district in those states) to count field tests toward NCLB’s annual testing requirement in the spring of 2014 as long as all students are tested on some assessment. However, those field tests will not result in student-, school-, district-, or state-level scores, so theoretically a state could administer the field test to all of its students and have no transparent or actionable data. At the same time, Duncan has threatened to withhold money from California for its legislation limiting testing in the spring of 2014 only to districts with the technical capacity to administer the field tests. California’s plan would also not result in actionable testing data next spring.
The only difference between the two is that California doesn’t care if some students are given a one-year testing reprieve. Both Duncan’s waiver and California’s plan would result in large numbers of students, schools, districts, and (potentially) entire states without testing data for this school year. In fact, Duncan’s plan could result in testing kids solely for the sake of testing them—little to no actionable information for educators and insufficient data to maintain state-level accountability measures.
This policy has important implications. Do you want to measure student growth for any purpose? Well, that requires measuring student achievement at two points in time, and this policy would mean that the spring of 2014 could not be one of them. We may not get student-growth data until 2016, when we’re scheduled to have two years of results on the new assessments.
Do you want to set student-achievement goals? Without a state test, schools and districts will have to find something else on which to measure student achievement.
Postponing test results for a year is a dream come true for accountability opponents, because it halts a decade of regular testing and sets in motion multiple years of confusion and the political conditions to give low-performing schools a reprieve and anti-accountability forces the opportunity to roll back decades of progress. The waivers are being offered as only a one-time opportunity, but it’s a dangerous genie to let out of the bottle—lots of places are going to like it and want to preserve this new no-testing reality.
What’s more, the Obama administration has just made all of its policy initiatives much more complicated. The definition of “student growth” it has used in Race to the Top, NCLB waivers, the Teacher Incentive Fund, and a host of other grants and initiatives includes student scores on state tests. For one example of how this is playing out, the Department is threatening to revoke state waivers over states’ inability to include state test results as one component of teacher and principal evaluations. If there are no state tests, this definition and this argument are moot.
The transition to new assessments was always going to be rough, and as two former US Department of Education officials, we appreciate the challenges the Secretary and his team faced. But this was not the right way to handle it. Duncan should have set a clear line that all students need to be tested on assessments that produce reportable, actionable data. When states adopted new assessments in the past, they conducted field testing on top of existing statewide tests or made the necessary data decisions to translate field-test results into usable scores in the same year. Duncan decided he didn’t want states to have to choose between these less-than-ideal options. It’s not clear how many states or districts will participate in the PARCC and Smarter Balanced field testing, but he also could have limited his new flexibility to a set percentage of students and schools to at least minimize the impact. (Duncan hinted at setting a 20 percent cap in regard to California, but the federal guidance does not mention any limitations.)
Instead, we could have the worst possible scenario play out next spring. All students will be required to take tests, but individual students won’t know how they did, the public won’t know how schools performed, and nothing will happen as a result of poor performance. Testing for testing’s sake will make no one happy.
-Andy Smarick and Chad Aldeman
This first appeared on the Fordham Institute’s Flypaper blog.