State data on students who are excluded from taking the National Assessment of Educational Progress, as well as those who are given special help or accommodations during the tests, will be featured more prominently in NAEP reports, starting this fall.
Officials are making those changes to ensure a better understanding of state differences, and the limitations of such comparisons, the governing board that sets policy for the federal testing program said this month. The board will also consider ways to standardize exclusion procedures nationwide, beginning with exams scheduled for 2009.
“It’s a national assessment, and it should be given as a national assessment,” said Andrew C. Porter, a board member and the dean of the graduate school of education at the University of Pennsylvania in Philadelphia. “Now,” he said, “it’s a national assessment given by local rules.”
Revealing Information
Exclusion rates vary considerably among states. States generally don’t test students who would not be included in the states’ own assessments. That usually means children who are just beginning to learn English, or special education students whose individualized education programs, or IEPs, restrict or prohibit testing. Some states, for example, give certain students more time to take the tests, or read math and science questions to test-takers who need such help. (“States Vary on Students Excluded From NAEP Tests,” Nov. 2, 2005.)
At the National Assessment Governing Board’s regular meeting here Aug. 2-4, several of its 21 members expressed concern that information on exclusions and accommodations was not readily available.
In 2005, the state-by-state information on exclusions and accommodations was not presented in the printed reports on the assessment’s reading and mathematics exams. Instead, it was buried in the extensive NAEP database that is available on the Internet. That year, the exclusion rates ranged from just 2 percent in Alabama and Wyoming to a high of 14 percent in Louisiana.
In contrast, information from other reports has been easier to find. The trial test of NAEP given in urban districts that year included the data in the main report. It also alerted readers that the results for Houston and Austin, Texas, should be considered with caution, given the large proportion of students with special needs who were excluded from the test sample.
California officials have argued that their state’s historic poor showing on NAEP is partly the result of the state’s policy of including most children, regardless of language or academic difficulties, in the test sample. Texas, which has similarly high percentages of English- language learners, does not test as significant a proportion of those students.
Governing-board Chairman Darvin M. Winick, said such changes should be carefully considered since inevitable comparisons will be made between California and his home state of Texas.
The board agreed that the reports on the 2007 reading and math results, scheduled for release this fall, will include the information on exclusions and accommodations within the main printed version, with clear “cautionary language” about how variations in those numbers can affect state results and comparisons between states. Future test results may flag scores from states that have high exclusion rates or allow significant accommodations to test-takers.
Standardization Obstacles Ahead?
Some critics have charged that states that exclude large numbers of students from the test may be taking advantage of the loophole to improve their scores. But setting a uniform policy around exclusions and accommodations could be problematic.
Peggy Carr, the deputy commissioner of the National Center for Education Statistics, the arm of the U.S. Department of Education that administers NAEP and compiles the data, said that standardizing those practices could run up against state testing requirements and the legal rights of special education students. Before 2004, NAGB tried to impose new rules on schools to make exclusion rates more uniform, but participating schools generally followed students’ IEPs anyway, said Arnold Goldstein, a statistician with the center.
The NCES is conducting a statistical study to predict the exclusion rates among students with disabilities based on the severity of their impairments, the types of accommodations the students get when they take state tests, and historical data on state exclusion rates.
