California in Testing Showdown With U.S. Department of Education

Includes updates and/or revisions.

California’s move to suspend most of its accountability testing for one year—and the sharp rebuke that it’s drawn from the U.S. Department of Education—illustrate the complications some states could face next spring when they juggle their own assessments with field tests of new common-core tests.

In an unusually stern statement Sept. 9, U.S. Secretary of Education Arne Duncan said that California’s plan “is not something we could approve in good conscience.” If the state proceeds with the plan anyway, “the department will be forced to take action, which could include withholding funds from the state,” he said.

While he applauded the state’s commitment to implement the Common Core State Standards, Mr. Duncan said that “letting an entire school year pass for millions of students without sharing information on their schools’ performance with them and their families is the wrong way to go about this transition,” and that it amounted to “backing away entirely from accountability and transparency.”

The missive from Washington comes amid growing discontent in some states that federal officials are meddling in local education decisionmaking and showing insufficient flexibility as they transition to the common core.

It was sparked by a California’s desire to dump most of its longtime testing program, called STAR, for Standardized Testing and Reporting, in the spring of 2014. Instead, the state would use the field tests in math and English/language arts being designed by the Smarter Balanced Assessment Consortium, one of two big groups of states designing tests for the common standards.

But since that assessment isn’t fully ready—the field test is part of its development—the results would not be used for accountability.

Late last week, both houses of the legislature had approved the measure, and Gov. Jerry Brown was reportedly ready to sign it.

Old Mandates, New Exams

The No Child Left Behind Act requires states to report student progress annually on tests of reading and math in one year of grades 3-8 and high school. The federal Education Department can withhold a portion of Title I funds from states that fail to do so.

But top California education officials feel so strongly about their plan that they are weighing accepting such a penalty as part of the price of moving forward.

“We have the right to accept their penalty,” Michael W. Kirst, the chairman of the state board of education, said. “There’s that option.

“There is no long-term defiance here,” he said, noting that California’s plan would fully comply with the NCLB law in 2015 because it would be giving the Smarter Balanced tests to all students and reporting the results as required.

“We’re trying to do the best thing in the short run,” Mr. Kirst said, “with the top priority being to implement the common core quickly and thoroughly.”

Striking back at Secretary Duncan, state schools Superintendent of Public Instruction Tom Torlakson said in a statement that California won’t reach its goals “by continuing to look in the rear-view mirror with outdated tests, no matter how it sits with officials in Washington.” He said he hoped that federal officials would “agree that withholding badly needed funds from California’s students would be a grave and serious error.”

With the urging of concerned state schools chiefs, the federal department announced this year that it would consider proposals that would allow states to avoid “double-testing”—giving their own tests as well as field tests created by Smarter Balanced and the other state consortium designing such tests, the Partnership for Assessment of Readiness for College and Careers, or PARCC.

Defining ‘Flexibility’

Federal officials said they would weigh requests to allow schools to administer only one assessment in 2013-14 to any individual student—either the current statewide assessment or the field test. For those schools, accountability labels (such as whether a school is a “priority” turnaround school) and targeted interventions would hold steady for a year, with no relaxation of accountability requirements.

Proposals for such testing flexibility are being considered now by federal officials. But that offer was made explicitly only to states that have obtained waivers from key tenets of the NCLB law. California is among the fewer than 10 states that have not yet obtained such waivers.

Explaining the federal Education Department’s priorities in an interview with Education Week, a department official said that states that wish to change their testing systems in the transition to the common core must make sure that all students in NCLB-required grades take a test in both math and English/language arts, whether those states use their own tests or a field test of one of the consortia’s assessments.

California’s plan, the official said, leaves such distinctions up to the state department of education.

But the official was less clear and emphatic about whether states must use tests that produce reportable results.

The Education Department recognizes that field tests can’t yet produce valid results, so states would not be required to use them for accountability, the official said. A state that wished to use only the field tests would have to discuss that prospect with the department.

“The clear expectation … is not that every child have results, but that every child participate in tests in both [English/language arts] and math, the state’s current tests or a field test,” the official said. “If California would like to use a mix of STAR and Smarter Balanced tests, that is something we would consider.”

“Clearly, accountability for student learning is of enormous importance,” a department spokesman added. “As states move to new assessments aligned to career and college readiness, we will work with them in a spirit of flexibility. But simply dropping assessment for a year isn’t the right answer.”

California Deputy Superintendent Deborah Sigman has said that her state would incorporate its testing plan into a waiver application. But given Secretary Duncan’s reaction, the state now appears to face a choice between revising its plan or losing of part of its Title I money.

In the past, the Education Department has penalized states by revoking a portion of the administrative part of that funding stream, but it hasn’t yet held back any of the Title I money that supports disadvantaged students. California was the target of one such action in 2009, when the federal department said that it had failed to show that its middle school math tests were properly aligned to its math standards.

The federal department has also used the threat of such a penalty to make states comply with the NCLB law. In 2006, it warned 10 states that they could lose 10 percent to 25 percent of their Title I aid if they didn’t bring their assessment systems into line with the law’s expectations. Those states avoided penalties by revising the tests.

Mr. Duncan has made no secret of his willingness to use that lever. Last week, the Education Department denied Texas’ request to cut back on testing elementary students in higher-performing schools.

Threading the Needle

Some observers said the standoff between California and the federal Education Department might give states pause as they consider seeking double-testing flexibility.

“The secretary has said he wants to give states more flexibility on this, and that’s what California asked for,” said Christopher T. Cross, the chairman of Cross & Joftus, a California firm that advises districts on education issues. “They’re trying to thread the needle from getting from where they are today to the new assessments. But given the secretary’s statement, I can’t see how a state could do that.”

In navigating that transition, California has advanced key arguments that resonate in many states as they move toward the consortium tests in 2015. It sought to avoid an undue testing burden on students and schools, an issue the federal department acknowledged by offering the flexibility to waiver states.

California also argued that sticking with its current system risks a harmful mismatch between what’s being taught and what’s being tested.

“The question here is accountability for what?” said Mr. Kirst, the state board chairman. “For an old set of standards on an old test? That confuses teachers and pushes them in two different directions.”

Ms. Sigman, the deputy state chief, said that an earlier move to the Smarter Balanced system—even with just the field-test version—would facilitate a smoother shift to the common standards.

“We think this is a great opportunity for teachers and students and parents and administrators to get a look at what these tests are,” she said. “Part of this is giving schools and districts the time and space to implement the common core.

“We don’t want to give an assessment that’s not aligned to our current set of standards,” she said. “We want to have people look forward, not backward.”

The state proposes to give the math field test to 10 percent of its students and the English/language arts field test to another 10 percent. That would meet the Smarter Balanced consortium’s need for a scientific sample to advance the development of the test.

The rest of the state’s students would be expected to take either the math or the English/language arts field test as well, Ms. Sigman said, except where districts aren’t technologically capable of giving the computer-based test. The savings from cutting out most STAR testing could be applied to the broader administration of the field tests, she said.

All 44 states participating in one of the assessment consortia face a potential double-testing problem next spring. But the computer-adaptive format of the Smarter Balanced assessment makes the double-testing issue more pronounced in that consortium’s states. That’s because an adaptive test demands a much larger bank of test items, which in turn requires a bigger sample of students to take the field test.

The PARCC exam is computer-based, but its questions will not adapt to the test-taker’s ability level.

Smarter Balanced consortium anticipates its field testing will require the participation of 2 million students across its 25 member states. PARCC field-testing will involve 1 million students in its 20 member states.

Equity Concerns

Some activists fear that California’s plan to suspend most accountability testing would create a host of problems. Chief among those is the absence of information that helps the public understand how its schools and students are doing. But the absence of test data for 2014 has sparked additional concerns.

Test scores are a pivotal part of the decisionmaking about how best to serve students with disabilities and whether to reclassify English-language learners as proficient in English, said Arun Ramanathan, the executive director of the Education Trust-West, which came out strongly against the Golden State’s plan to suspend most STAR testing. Doing without such data would complicate judgments on education for millions of students, he said.

Additionally, he’s concerned about the state’s plan to insist on Smarter Balanced tests only in districts that are technologically able to give them.

That approach, Mr. Ramanathan said, essentially creates a divide between wealthier and poorer districts. If a key goal of giving the Smarter Balanced test more broadly is to help schools transition to the common core, then districts that have less technological capacity are put at a disadvantage, he said. Students in those districts would also lose out on time to get used to taking computer-adaptive tests.

“There’s a learning curve there, so it raises some real equal-protection questions,” Mr. Ramanathan said.

He is urging state officials to consider sharing the results of the field tests—if not for accountability purposes, then for their value in providing feedback to parents, teachers, and students.

“It’s to the detriment to test several million students and not provide results to parents,” Mr. Ramanathan said. “It’s testing for testing’s sake. What’s it going to prove?”

‘How Things Work’

Assessment experts, however, are generally skeptical of using field-test results for instructional or evaluative purposes, since they are by definition still in development. Consortium officials have long said that they could not set passing scores until after feedback from the 2014 field tests informs adjustments to the assessment designs.

“Field testing is an opportunity to find out how things work when you go to scale,” said Derek C. Briggs, who serves on both consortia’s technical-advisory committees and is the chairman of the research- and evaluation-methodology program at the University of Colorado at Boulder.

“There are all kinds of variables that need to be examined,” he said, from whether each item measures what test designers want it to measure, to the logistics of getting every student in front of a computer during a state’s defined testing window.

It’s “theoretically possible” to generate some very limited results from a field test, he said, if a core set of questions is common to all test-takers. But any such results would be limited to a narrow subsection of standards in the test.

Mr. Briggs said that no valid results could likely be generated from one of the greatest areas of promise in the new tests—deeper, more nuanced information yielded by technology-enhanced items and performance tasks—because they are such new and unfamiliar additions to the testing regimen.

“You need to be very careful about the results of field tests,” he said, “because everything is still in flux.”