States Eyeing Expense of Hand-Scored Tests in Light of NCLB Rules
When students put down their pencils at the end of Connecticut’s testing each year, another intensive process begins. Hundreds of trained evaluators work day and night for about a month to score the written responses.
Although expensive, the use of open-ended questions drives the kind of instruction that state leaders say they want in their schools. So they balked when federal officials recently suggested using multiple-choice tests to meet the requirements of the No Child Left Behind Act.
“From our point of view, it would have us dumbing down our tests,” said Betty J. Sternberg, Connecticut’s commissioner of education.
Many states have weighed the cost of hand-scored tests in light of the federal legislation, but the issue is especially pertinent for Connecticut, where Attorney General Richard Blumenthal has pledged to sue over the federal education law. Calling it an unfunded mandate, he cites estimates that the state must spend a total of $8 million of its own money by 2008 to fulfill the law’s testing provisions.
The state tests students in grades 4, 6, and 8, but the federal law requires that students be assessed annually in reading and math in grades 3-8. Those tests are to be used in judging whether schools meet state-set performance targets known as adequate yearly progress.
Federal officials dispute Connecticut’s estimate of what it would cost to comply with No Child Left Behind. In a May 3 letter to Ms. Sternberg, U.S. Secretary of Education Margaret Spellings wrote that the kinds of assessments included in the state’s calculations are more extensive than spelled out in the NCLB law.
“[S]ome of the costs of the system are attributable to state decisions in these areas,” she wrote. “While these decisions are instructionally sound, they do go beyond what was contemplated by NCLB.”
The debate shows how politically charged questions about adequate funding for the 3-year-old No Child Left Behind law revolve around ideas of what makes for appropriate testing, said Robert M. Palaich, a partner with Augenblick & Myers, a Denver-based consulting firm that helped Connecticut come up with its cost estimates.
“The secretary is correct that it is possible to do this in a way that costs less money,” he said. “But if you believe that states have the right to decide their standards and how to assess those standards—and now you are going to incorporate their assessment systems in your AYP calculations—then it does seem that the state has some good ground to stand on.”
Connecticut includes open-ended and multiple-choice items on all of its student assessments. In math, students are asked to explain how to solve problems. In reading, they’re asked to write about passages of text.
Scoring such answers means training evaluators on what makes for an adequate response. The challenge of doing so was evident last year, when Connecticut changed testing companies. The test results came back so far off from previous years that the state had the vendor rescore them.
“It’s a very intensive, very complicated process to assure reliable scoring,” Commissioner Sternberg said in a recent interview.
To find out how much it would cost to scale up its testing system under the No Child Left Behind Act, Connecticut used a process devised by Augenblick & Myers for the Council of Chief State School Officers. Eleven other states are using the same process to do their own cost studies, but Connecticut was the first to release its findings.
The analysis concludes that to expand the state’s testing system to the full set of grades required by the federal law would cost $41.6 million by 2008, when all of those tests must be in place. At current funding levels, the state by then will have received a total of $33.6 million in federal money for test implementation, according to the study.
Connecticut’s cost study wasn’t the first to suggest such a gap. Two years ago, the congressional investigative agency called the Government Accountability Office reported that, based on projected spending levels, many states would not get enough federal funding for test implementation to expand the type of tests they were then using to include all of the assessments that the law calls for.
The GAO found that the key to assessment costs was the type of test items that states used. Many states that used open-ended items would not have enough money if they kept the same number of such items in any new tests that they added, the study found.
Some states are indeed cutting back on the number of open-ended items on each assessment. Illinois, which until now has tested in math and reading only in grades 3, 5, and 8, used to include two items on each test that required students to write. Next year, when the state begins testing in grade 3-8, each assessment will have one question involving writing.
Open-ended items have traditionally made up half of Maine’s assessments. To meet federal requirements, the state plans to use tests in additional grades next year in which such questions make up about 20 percent of the exam. Maine is completing a cost study akin to Connecticut’s, and the state legislature’s joint education committee recently approved a bill authorizing the state to sue over the No Child Left Behind Act.
“Our preliminary findings indicate that we have a multiple-million-dollar gap between what the feds are providing and what our costs are going to be,” said Patrick Phillips, Maine’s deputy commissioner of education.
The U.S. Department of Education counters that states are getting ample federal money to meet the letter of the law. No Child Left Behind, they say, doesn’t call for expanding the same testing programs that states have been using.
In her letter this month to Ms. Sternberg, Secretary Spellings noted that Connecticut’s cost estimate includes scaling up the state’s largely hand-scored writing assessment. But the federal law’s requirement to test in grades 3-8 only specifies reading and mathematics.
It also requires testing in reading and math in one grade in high school—as Connecticut already does—plus science testing in one grade each in elementary, middle, and high school.
Raymond J. Simon, the acting deputy secretary of education, made a similar point last month when he said, at a meeting with Ms. Sternberg in Washington, that Connecticut could fulfill the law’s requirements by using multiple-choice assessments in the grade levels that it must add.
“What’s most important is that for every year a kid is in school, that a parent can know at the end of the year how well their students have learned what they’ve been taught,” Kerri L. Briggs, a senior policy adviser to Mr. Simon, said in an interview.
Theodor Rebarber, a testing expert who is the chief executive officer of the Washington-based Education Leaders Council, agreed that Connecticut isn’t obligated under the law to use tests that require students to write. The council has commissioned its own studies showing that federal allocations are enough for states to implement the law’s testing provisions.
“If that’s what Connecticut thinks it needs to spend, then it should,” Mr. Rebarber said. “And it should view it as a relatively small investment in terms of their overall public expenditure to ensure quality public schools.”
In fact, Ms. Sternberg sees little value in conducting additional statewide annual testing. She has sought a waiver from the federal Education Department to allow her instead to have districts give periodic assessments throughout the year in the additional grades. Such a strategy would do more to improve achievement, and cost less, she argues.
But the federal agency nixed the idea, leading Mr. Blumenthal, the state attorney general, to announce plans to sue the department, claiming that the law represents an illegal unfunded mandate. ("Connecticut Pledges First State Legal Challenge to NCLB Law," April 13, 2005.)
“This might be decided in court,” said Commissioner Sternberg, whose state, as of press time last week, had not yet filed suit. “It also might be decided in the court of public opinion, because if this pushes us to lower our rigorous standards, then what has this done?”
Vol. 24, Issue 38, Pages 1,21