The latest round of testing for the National Assessment of Educational Progress wrapped up last week, but the board that sets policy for the test known as “the nation’s report card” is still grappling with issues related to expanding participation of students with disabilities and students learning English.
At a March 1 meeting, board members learned that their recently adopted policy aimed at reducing the number of students excluded from taking the assessment could make it hard to compare results from year to year. Trend data are among the most important pieces of NAEP.
The National Center for Education Statistics, which administers the tests, plans to follow the method for calculating scores that it has used in the past, even though that diverges from board policy.
The center will present information to the board on what a change in calculation might mean to scores once the agency has a chance to evaluate results from the most recent test administration, said Arnold A. Goldstein, the program director for design, analysis, and reporting for the assessment division of the NCES.
The issue of test inclusion is not limited to NAEP, a voluntary test of a nationally representative sample of students. The tests, administered every other year, are given in grades 4, 8, and 12 in math and reading, with other subjects assessed periodically.
The groups tasked with creating tests to assess students on the Common Core State Standards adopted by nearly all the states are also working out their own test-accommodation policies, and in some cases are facing criticism from special education advocates who say that the policies are too restrictive.
Comparing Results
In 2010, the National Assessment Governing Board, which oversees NAEP, created a policy that would limit how many students with disabilities and English-language learners could be excluded from the testing pool. States were asked to test at least 95 percent of the students in the sample selected for assessment, including 85 percent of students with disabilities and 85 percent of students learning English.
The policy was meant to ensure that only students with severe cognitive disabilities or English-learners who had been in the country for less than a year would be excluded from taking the tests.
Before that policy change, there was no directive from the board on what percentage of students could be excluded by states from taking the test. Some states chose to exclude a higher proportion of students with disabilities and English-learners from the test than others, making NAEP results difficult to compare across states.
The policy change was in effect for the test’s 2011 administration, and the NCES at first made a “pretty informal effort” to follow it, said Cornelia Orr, the governing board’s executive director. However, states were aware of the change, and many made adjustments on their own.
The NCES noted that 45 states assessed fewer than 85 percent of the students with disabilities and English-language learners in 2009. By 2011, the number of states assessing less than that percentage dropped to 18. Even more states are expected to meet the policy goals this year.
To push states in the direction of inclusion, the 2010 policy is to count students who don’t take the tests as “refusing” to take them instead of counting them as excluded from the assessment.
New Issues
But that change distorts some of the statistical analysis the NCES uses, according to the agency’s presentation to the board.
The policy also encourages states to have students who would take a state test with accommodations to take NAEP without those accommodations.
There’s a concern that doing so can drive down test scores, said Andrés A. Alonso, the superintendent of the 85,000-student Baltimore school district and the chairman of the board’s reporting and dissemination committee. His district is one of the 21 urban districts for which NAEP scores are reported individually in the Trial Urban District Assessment, or TUDA.
“If you’re a TUDA district, it’s a different dynamic of perception and conversation,” said Mr. Alonso in a telephone interview after the meeting.
Louis M. Fabrizio, the chairman of the board’s committee on standards, design, and methodology and the director of accountability policy and communications for the North Carolina education department, was on the board when the change was made to promote greater inclusion. The board was unaware of some of the reporting issues that would result from the change, he said.
Similar Issues
Though NAEP has some unique issues connected to the way it samples students, Mr. Fabrizio said he sees that Smarter Balanced Assessment Consortium and the Partnership for Assessment of Readiness for College and Careers, the two groups creating common-core tests, will have similar issues. PARCC has released draft language that would exclude certain students from using accommodations on the test.
“All students will end up affected by these accommodation decisions,” Mr. Fabrizio said.
