Nearly half of states are expanding tests into more grades in 2005-06 school year
You could describe it as the year of the tests. Twenty-three states are expanding their testing programs to additional grades this school year to comply with the federal No Child Left Behind Act.
“Every group I’ve been talking to, I’ve just said, ‘Be patient with us this year,’ ” said Alexa E. Posny, the deputy commissioner of education in Kansas, which is adding reading and math tests in four grades in each subject.
“First, it’s the sheer volume,” she said. “In the past, we would develop 4,000 test items; we’re developing 18,000 items. Second is the number of teachers who have never participated in state assessments, so it’s a whole new ballgame for them. And then there’s the overwhelming amount of data that will be available because there are so many more grades.”
Forty-eight states and the District of Columbia will give standards-based tests in reading and mathematics in grades 3-8 and at least once in high school this school year, as required by the nearly 4-year-old federal law, according to a survey by the Editorial Projects in Education Research Center.
The holdouts are Iowa and Nebraska. Districts in Iowa give the Iowa Tests of Basic Skills, a national test not designed to measure state or local content standards, while districts in Nebraska craft their own tests, except for a state writing exam.
In devising the new tests, most states have defied predictions and chosen to go beyond multiple-choice items, by including questions that ask students to construct their own responses.
But many state officials are worried that, over the long run, such decisions could push their assessment programs into the red. In addition, despite efforts aimed at getting test results back faster, they fear that the sheer volume of assessments could generate delays and errors in releasing scores.
In general, states have filled in the gaps in their testing programs with assessments that mirror those in other grades. Many analysts had predicted that, because of costs, states would rely solely on multiple-choice tests for grades and subjects in which they had not tested previously.
View a complete collection of stories in this Education Week special report, Testing Takes Off.
A 2003 study by the investigative arm of Congress estimated that it would cost states $1.9 billion to meet the testing requirements over the six years of the federal law’s authorization if they relied solely on multiple-choice questions that could be machine-scored. But it would cost states $3.9 billion if they used a mix of multiple-choice and open-ended items, and up to $5.3 billion if the tests required hand-scored, written responses, according to the agency, now called the Government Accountability Office.
“We’re developing tests that have the same format, blueprint, rigor, as the tests that we already have in place,” said Jeffrey Nellhaus, the deputy commissioner in the Massachusetts Department of Education. The reading and math tests require students to construct their own responses to some items, in addition to answering multiple-choice questions.
The Massachusetts program is expected to cost “somewhere in the order of $10 million to $12 million a year,” Mr. Nellhaus estimated, compared with the just over $7.6 million the state is receiving from the federal government this year to cover such costs.
Beyond Multiple Choice
In Nevada, which added reading and math tests in grades 4, 6, and 7 this year, the money has so far been adequate to create tests with a mix of multiple-choice and constructed-response questions, said Paul M. La Marca, the state’s assistant deputy superintendent.
“We know that tests are driving curriculum, to some extent, so we think it’s important to have items that stretch the cognitive demand of the students,” he said. “You can do that with multiple-choice items, but you have a better chance of doing it with other types of items.”
1. The District of Columbia previously used the SAT-9 in grades 1-11. In 2005-06, it will use a standards-based exam in grades 3-8 and 10.
2. Louisiana previously administered the Iowa Test of Basic Skills. In 2005-06, it will introduce an assessment program that combines the ITBS and criterion-referenced items, in grades 3, 5, 6, 7, and 9.
3. New Hampshire, Rhode Island, and Vermont suspended testing for 2004-05. These states have implemented jointly developed assessments in grades 3-8 in 2005-06.
SOURCE: EPE Research Center
Still, Mr. La Marca said, the state doesn’t have as many constructed-response items as it might want. “The balance of the tests is skewed toward multiple-choice,” he said. “It’s more than just a cost issue. In our state, we have significant pressure for quick turnaround time, so that almost hamstrings us a little bit.”
New Jersey officials announced Nov. 16 that, in addition to the tests the state already gives in grades 3, 4, 8, and high school, it would add a commercial test aligned to its reading and math standards in grades 5, 6, and 7 for the 2005-06 school year as an interim measure, while it works to revise its entire testing program to provide better diagnostic information for educators.
“Ultimately, we want to build a more robust, more rigorous state system, hopefully to incorporate a performance assessment which we have been piloting for the last three years,” said Acting Commissioner of Education Lucille E. Davy. She said the state worked closely with the U.S. Department of Education to ensure the plan could meet federal requirements, although the system still has to go through a peer-review process before it can be approved, as is true for all states.
In Mississippi, meanwhile, officials decided to drop all short-answer questions in the 2005-06 school year to facilitate speedier scoring of test results. It now has only multiple-choice items, except for a state writing test.
Kansas has suspended the use of all items that require an extended response from students and can’t be machine-scored.
“It doesn’t mean in the future we may not add those items,” Ms. Posny said. “We want to figure out how we can do that and still … score [the test] online.”
Fifteen states in total will rely solely on multiple-choice items to measure student knowledge, with the exception of their writing tests, according to the EPE Research Center survey: Arizona, California, Georgia, Idaho, Iowa, Kansas, Mississippi, North Carolina, Oklahoma, Oregon, South Dakota, Tennessee, Texas, Virginia, and Utah.
The District of Columbia this year switched from a multiple-choice test that was not aligned with its academic-content standards to a new set of standards and tests based on those given in Massachusetts.
Costs a Concern
Although many states said the federal government has provided enough money to cover the development of new tests, they worry about the costs in future years.
Washington state’s tests include a 50-50 mix of multiple-choice and open-ended questions. While federal funding has so far been sufficient to expand the tests to grades 3, 5, 6, and 8, said Greg B. Hall, the state’s assistant superintendent for assessment and research, starting next fiscal year the state will run a deficit in its testing budget that is expected to increase over time.
The reason, he said, is the cost involved in scoring so many additional open-ended items, “which we haven’t had to do yet.”
|Number of States|
|States testing in grades 3-8, high school||48*|
|States adding tests in reading or math||23|
|States with all multiple choice except for writing||15|
*Does not include the District of Columbia
SOURCE: EPE Research Center
Rhode Island, New Hampshire, and Vermont have jointly developed grade 3-8 reading and math tests to meet the federal law’s requirements. While federal aid has covered those costs so far, “we’re all worried about what will happen when this money goes away,” said Mary Ann Snider, the director of assessment and accountability for the Rhode Island Department of Education.
In August, Connecticut became the first and, so far, only state to sue the federal government over the No Child Left Behind law, charging that federal funding falls short of what is needed to meet the law’s requirements.
Connecticut officials have sought unsuccessfully to get out of expanding their testing in core subjects beyond grades 4, 6, and 8. An estimate by the state education department pegs the cost of providing tests in the additional grades required under the NCLB law at $41.6 million by 2008, compared with $33.6 million that the state is slated to receive from the federal government by then for testing. (“Connecticut Files Court Challenge to NCLB,” Aug. 31, 2005.)
The annual testing requirement is a linchpin of the federal law. Schools and districts are required to meet annual performance targets for their student populations as a whole and for certain subgroups of students. Those that receive federal Title I money and that fail to meet their targets for two or more years face penalties.
Earlier Testing Dates
At least some states are shifting their testing dates to try to get results back sooner. New Hampshire, Rhode Island, and Vermont started giving their new jointly developed tests in the fall, rather than in the spring.
That timing was intended, in part, to ensure that schools would know by the winter whether they have met their performance targets under the federal law and would be subject to any of the law’s sanctions, Ms. Snider said.
Michigan shifted from January to October testing, but not without some mishaps. “The idea is to give data back to teachers while they still have the students, and they still have the energy to do something about the results,” said Edward D. Roeber, who directs the state testing program.
Still, the change “cut out four months in the test-preparation schedule, which made the summer very challenging,” he said. And when the state’s test contractor, Pearson Education Inc., failed to get enough tests delivered to school districts on time, the state was forced to extend the testing window by two weeks.
Other states have made more minor adjustments. Kansas, for instance, moved its testing dates up by one week, to March 1. It also has put its writing, science, history, and government tests on hold for a year, while it gets the new reading and math exams in place.
Massachusetts consolidated all of its reading tests in April, so that they would not compete with tests in other subjects during May. New York state adopted a flexible schedule for the statewide administration of its grades 3-8 exams for this school year, to permit schools to give one session per day at each grade level. Maine has opted to replace an existing state test with the SAT college-admissions exam at the high school level, starting next spring.
Even so, state officials generally anticipate that it will take longer to report results this school year, as they try to craft new performance standards in the additional grades and cope with the extensive amounts of data.
“It will be delayed, there is no doubt in my mind,” said Ms. Posny of Kansas.
Potential for Error
One concern is that the sheer volume of tests will, inevitably, create logistical problems for schools and an increase in administration and scoring errors.
“The more you add to the test contractors’ plate and the quicker they have to report the results back, the probability goes up that errors are going to be made,” said George Madaus, a professor emeritus of education at Boston College who co-wrote a May 2003 report on the widespread errors in standardized tests. “You’re stretching the capacity of a limited number of companies that do this work.”
Among the states that have suffered from scoring glitches is Nevada. In 2002, the state board of education required the San Antonio-based Harcourt Educational Measurement to pay penalties totaling $425,000 because of a mistake that threw off the scores of nearly 31,000 students who had taken the state’s high school exit exam in math.
“We’re very worried about accuracy from the vendor because we have, unfortunately, been snake-bitten,” Mr. La Marca said.
In Michigan, which is giving 216 different test forms in grades 3-8 this year, state officials are doing “a lot of extra checking,” said Mr. Roeber.
“There are just more chances to screw up,” he said. “My staff and I have been working 18 hours a day and some weekends.”
And state officials aren’t alone. “School districts are kind of reeling,” Mr. Roeber said.
One issue for schools is simply having enough staff members to administer the tests and to provide accommodations, such as more time, for all the students who need them because of disabilities or limited English skills.
“As with most states, we have an extensive list of accommodations, which is great for students but puts the school staff in a difficult position, because many of the accommodations require alternate settings and additional staff,” said Tim Kurtz, the director of assessments for the New Hampshire Department of Education. “Every single adult human being is involved in testing somehow.”
At least one thing is clear: With more tests, in more grades, soon many more teachers will be focused on test results.
“The existing evidence suggests that when a grade level is tested, teachers pay a lot more attention to what’s on the test,” said Joan L. Herman, the co-director of the Center for Research on Evaluation, Standards, and Student Testing, or CRESST, at the University of California, Los Angeles.
“So as more grade levels are being tested,” she said, “you can now expect that every teacher will be paying attention to what’s on the test and, in the best case, aligning their instruction with standards and, in the worst case, engaging kids in a curriculum of test preparation.”