Small States Find Benefits in Jointly Developed Tests
As students in New Hampshire sit down this month to take new state mathematics and reading tests, they won’t be alone. Their fellow students in Rhode Island and Vermont will be taking the same exams along with them.
In the first collaboration of its kind under the federal No Child Left Behind Act, those three New England states have joined forces to create the reading and math tests in grades 3-8 that states must give starting this school year. They’ve also jointly developed writing tests in grades 5 and 8.
By pooling expertise and resources, state officials say, they’ve devised better tests at less cost: just shy of $7 million a year for all three states combined over the life of their five-year contract with a commercial test-maker.
“For Vermont, our total assessment budget really hasn’t changed that much,” said Michael L. Hock, the director of assessment for the state education department. “What it’s done is given us a kind of test that we never could have afforded on our own. It’s given us an extremely good, custom-designed test at off-the-shelf test prices.”
Working with the Dover, N.H.-based Measured Progress, the test developer, and the Center for Assessment, a nonprofit group also based there, the states jointly crafted grade-level expectations for the tested subjects that served as the blueprint for the tests.
Content-area teams of teachers from the three states helped set the expectations and reviewed test items. The states also agreed to give the tests within the same dates, Oct. 3-26, under the same conditions, and with the same list of accommodations available for students who need them.
Come January, the states will set common performance standards for the exams and common ways of reporting results. Teachers also will be involved in setting the standards.
“All of us are very proud of these tests,” said Mary Ann Snider, the director of assessment and accountability for the Rhode Island Department of Education. She added that the exams “have lots of constructed-response items” that require students to write short answers, “and I don’t know that we would have been able to have as many constructed-response items if we had gone into this individually.”
The states and Measured Progress also brought in Norman L. Webb, a professor of education at the University of Wisconsin-Madison, to provide training on how to design items that truly reflect the depth of knowledge described in the grade-level expectations.
In addition, they worked with the National Center on Educational Outcomes at the University of Minnesota-Twin Cities and the Center for Applied Special Technology, in Wakefield, Mass., to produce tests based on the principles of “universal design,” an approach that aims to make tests accessible to the widest possible range of students.
All of the math items, for example, were reviewed to remove unnecessary wording that might trip up students with language deficits, Mr. Hock of Vermont said. And while students need to describe how they solved problems, the written responses will be scored only on math content, not on writing skills.
Ms. Snider estimates that over the period of the testing contract, Rhode Island will save more than $5 million from the cost of giving an off-the-shelf test developed by a commercial test publisher. Measured Progress estimates each state will save about $1 million a year compared with the costs of having done the tests alone.
“Particularly for smaller states, states that maybe aren’t quite as well off financially, it seems to me that there are some real efficiencies here,” said Stuart D. Kahl, the company’s president.
Shift From Spring
In an unusual move, the states shifted their grades 3-8 testing from the spring to the fall and will measure what students learned the previous school year.
The shift will enable the states to get results back to schools in the winter, and to let schools know whether they have met their performance targets under the No Child Left Behind law in the spring, instead of in late August. That schedule will give schools more time to plan if they have to offer students school transfers or supplemental services under the federal law. And it will provide them with information for designing any teacher professional development that occurs over the summer.
In addition, said Ms. Snider, “a lot of teachers felt that March testing really put a lot of pressure on them to shut down their classrooms during one of their peak instructional times.”
“We thought it kind of made sense,” she said, “for kids to come back into school in late August and early September, engage in some review work, and we thought that review work could also serve as test preparation.”
Fall testing has proved unpopular, though, with some school administrators, who worry that students will have forgotten over the summer what they learned the preceding year. Still, by delaying testing from spring until fall, Mr. Hock said, “we know we’re testing enduring skills.”
To make the transition, the three states did not give state reading and math tests last spring. Instead, they received permission from the U.S. Department of Education to use other academic indicators to measure whether elementary and middle schools made adequate yearly progress—the central requirement for ensuring accountability under the nearly 4-year-old federal law—in the 2004-05 school year. Rhode Island, for example, used student-attendance rates, while Vermont also used a developmental-reading assessment given to 2nd graders.
“We felt it would be silly to test the same cohort of kids twice in a six-month period, which is what would have happened,” Mr. Hock said.
The three states hope to have their standards-and-assessment systems peer-reviewed by the federal government in May to determine whether they have met the NCLB requirements.
“We tried to talk the [federal department] into doing one big review among all three states, but they weren’t crazy about that idea,” said Mr. Hock, “so we’ll all go through the review process in May, but we will each have individual submissions.”
Probably the greatest challenge, according to those involved in the new tests’ development, has been working out the details of test design and administration in states with different assessment cultures and histories.
“I spend at least a day a week driving up to New Hampshire for meetings,” Ms. Snider of Rhode Island said.
The experience has succeeded so well, officials say, that the states are about to release a joint request for proposals to develop science tests in elementary, middle, and high school, which also are required under the federal law starting in 2007-08.
And they may collaborate on high school reading and math exams. Under the NCLB law, beginning this school year, students must be tested once in those subjects during high school for federal accountability purposes. All three states currently are using their existing high school tests for that purpose.
Vol. 25, Issue 08, Pages 1, 14Published in Print: October 19, 2005, as Small States Find Benefits In Jointly Developed Tests