The latest reading and mathematics results from the National Assessment of Educational Progress suggest that, despite federal efforts to curb the practice, states still vary widely in the numbers of students with disabilities and English-language learners whom they excuse from taking the tests.
States鈥 鈥渆xclusion rates鈥 for the 4th grade reading test, for example, range from a high of 14 percent of students in Louisiana to as low as 2 percent in Wyoming and Alabama. On the 4th grade math tests, the rates were lower and a little less variable, ranging from 1 percent to 6 percent.
Keeping a check on how many students in each state are excluded from the congressionally mandated tests is important, experts say, because the practice can lead to an inaccurate picture of states鈥 educational progress.
View the accompanying item,
Read the related story,
鈥淭here鈥檚 just no question about it: You exclude more kids and you mangle the state-to-state comparisons,鈥 said Richard G. Innes, an education analyst for the Bluegrass Institute for Public Policy Solutions in Bowling Green, Ky. 鈥69传媒 are excluded for a reason.鈥
Mr. Innes鈥 criticism is not new to federal officials who oversee the testing program, which has come to be known as 鈥渢he nation鈥檚 report card.鈥 In an effort to increase the percentages of students who participate in the tests, federal testing officials in 1996 allowed states to provide certain accommodations so that more students with disabilities and English-language learners can take the tests.
On the math tests, students who meet the criteria for one of those categories are allowed, for example, to have questions read to them or to take the tests in their native languages. 69传媒 with disabilities can be given more time on the reading tests.
Steps Taken
Arnold A. Goldstein, the program director for reporting and dissemination in the assessment division of the U.S. Department of Education鈥檚 National Center for Education Statistics, said the changes have helped increase participation rates. At the 4th grade level, for example, exclusion rates for the nation as a whole decreased by half, dropping from 6 percent in 1992 to 3 percent in 2005. On the reading tests, though, the nationwide average has hovered around 4 percent. Declines were dramatic, however, for black and Hispanic test-takers.
鈥淚 think it鈥檚 because we鈥檝e been able to provide more accommodations in math,鈥 said Mr. Goldstein. 鈥淵ou still cannot read reading items aloud.鈥
For the 2003 and 2005 tests, the department also hired independent contractors to analyze whether exclusions were skewing the data. The most recent evaluation, conducted by the Washington-based American Institutes for Research, found little correlation between the overall state scores and exclusion rates. But the study did find that states that raised their exclusion rates tend to have scores that go up on the next assessment and vice versa.
鈥淏ut far from the entire increase could be attributed to that factor alone,鈥 Mr. Goldstein said.
Notes of Caution
Such concerns led Louisiana state school officials to sound a cautionary note last month in their press releases on the state鈥檚 own test results. Between 2003 and 2005, the percentage of Louisiana students who sat out the 4th grade reading test rose from 6 percent to 14 percent.
鈥淎t grade 4, frankly, our overall scores led the nation in growth, but, knowing we had an exclusion problem, we didn鈥檛 want to say that,鈥 said Scott M. Norton, the state鈥檚 testing director. Mr. Norton said state officials could not explain the spike in exclusion rates for reading since educators have to follow similar procedures on the federal and state tests.
In other states, though, high exclusion rates on the NAEP tests seem to crop up when the state testing rules are more permissive than NAEP鈥檚 on which students can be included in the tests and the kinds of testing accommodations they can have.
In Delaware, for example, some students with disabilities can have the state reading test read to them. Some of those students can also use calculators on the entire state math test, rather than just a portion of it, as NAEP allows, according to Patsy Kersteter, that state鈥檚 testing director.
By comparison, educators follow identical rules for state and federal tests in Alabama, where only 2 percent of 4th graders did not take the 2005 NAEP reading test.
Federal officials, in this year鈥檚 NAEP reports, did not flag the states with higher-than-average exclusion rates. Tables in the reports鈥 appendices, though, provide details about the exclusion rates for every state.
鈥淲e really didn鈥檛 have a basis for determining at what point the flag should be,鈥 Mr. Goldstein said.
Indeed, federal officials admit the exclusion issue is puzzling, particularly when it comes to testing students with disabilities. Under federal law, many such students have individualized education programs that require schools to provide them with certain academic accommodations, some of which NAEP does not allow.
鈥淭he problem is, very candidly, we鈥檙e not sure what to do,鈥 said Darvin M. Winick, the chairman of the National Assessment Governing Board, which sets policy for NAEP. 鈥淭hat鈥檚 one of the loose ends of the whole assessment system in this country.鈥