The Department of 91制片厂视频 plans to re-evaluate how many students鈥 test scores districts and schools will be permitted to exclude when determining whether they鈥檝e met annual educational goals under the No Child Left Behind Act.
At a June 13 hearing before the House 91制片厂视频 and the Workforce Committee, Deputy Secretary of 91制片厂视频 Raymond J. Simon assured lawmakers that large numbers of students鈥 test scores were not going uncounted when it came to evaluating whether districts and schools were raising student achievement. But he also said the 91制片厂视频 Department would hold a conference in the fall to determine whether states were excluding too many test scores to avoid sanctions under the federal law.
Under the No Child Left Behind Act, districts and schools must break down test scores by subgroups to determine whether all students are making adequate yearly progress.
Subgroup categories:
鈥 Students from major racial and ethnic groups, typically, African-American, Hispanic, white non-Hispanic, American Indian/Native Alaskan, and Asian/Pacific Islander
鈥 Economically disadvantaged students
鈥 Students with disabilities
鈥 Students with limited English proficiency
Districts and schools may choose not to count subgroups that are so small they are statistically unreliable for accountability purposes. Those numbers, known as 鈥淣-size,鈥 differ by state, based on approval by the U.S. Department of 91制片厂视频, and range from 5 to 75.
SOURCE: 91制片厂视频 Week
Under the NCLB law, which President Bush signed in January 2002, districts and schools must reach annual achievement goals not only overall, but also for specified subgroups of students, such as members of racial and ethnic minorities and students with disabilities.
Districts and schools are permitted to omit test scores from subgroups of students that are small enough that they would provide statistically unreliable information on a school鈥檚 performance or raise concerns about students鈥 privacy.
鈥淲e don鈥檛 want false positives and we don鈥檛 want false negatives,鈥 Mr. Simon said after the hearing. 鈥淭he better we can get at making the structure work correctly, then the more likely the children who need the services are going to get them.鈥
The minimum number below which districts and schools may exclude a given subgroup, known as the 鈥淣-size,鈥 differs from state to state and must be approved by the federal 91制片厂视频 Department. While districts and schools in Maryland may discount subgroups of five students or fewer, for example, West Virginia districts and schools may exclude subgroups of 50 or fewer. A larger N-size generally makes it easier for districts and schools to comply with the No Child Left Behind law.
鈥楳inimums鈥 Too Large?
An analysis released in April by the Associated Press found that the test scores of nearly 2 million U.S. students, mostly minorities, were not being counted for subgroup performance. The news service concluded that many schools were escaping accountability for the progress of racial and ethnic subgroups as a result of exclusion practices. (鈥淎nalysis Finds Minority NCLB Scores Widely Excluded,鈥 April 26, 2006.)
If schools receiving federal Title I aid for disadvantaged students fail to make adequate yearly progress, or AYP, for two years in a row, they must provide students with an option to transfer to a better-performing public school, and after three years, they must also provide free tutoring to students.
Both Republican and Democratic members of the education committee expressed worries at last week鈥檚 hearing.
鈥淚 am concerned that states are being allowed to establish minimum subgroup sizes that are too large and are thereby failing to disaggregate data for far too many students,鈥 said Rep. Howard P. 鈥淏uck鈥 McKeon, R-Calif., the chairman of the House education committee. 鈥淭his certainly was not the intent of Congress when we passed No Child Left Behind in 2001.鈥
Rep. George Miller of California, the panel鈥檚 ranking Democrat, said it was 鈥渢roubling to learn that some states appear to be circumventing the primary goal of the law and exploiting a legal loophole in order to exclude from accountability measures the scores of children from racial and ethnic minority groups and those with disabilities.鈥
John C. Brittain, the chief counsel of the Lawyers鈥 Committee for Civil Rights Under Law, a watchdog group based in Washington, told the House panel that the debate over N-sizes, at its core, was a civil rights issue.
Schools that are not identified as needing improvement under the law because they do not have to take into account some small subgroups, he pointed out, don鈥檛 have to provide extra help, such as tutoring, to students.
鈥淲hen students鈥 scores don鈥檛 count and their schools continue to make annual progress, students are denied statutory rights and benefits under NCLB,鈥 he said.
Committee members peppered Deputy Secretary Simon with questions about what the 91制片厂视频 Department was doing to address the issue. He stressed that although a subgroup鈥檚 size may have been too small to count toward a school鈥檚 performance for AYP purposes, those same students are often measured as members of other subgroups or, at least, in the district鈥檚 overall numbers.
鈥淭hus, the law, with built-in redundancies partnering with creative teachers,鈥 he said, 鈥渆nables us to get as close to 100 percent accountability as we possibly can.鈥
A Closer Look
In a June 13 letter to Rep. McKeon and other members of the House education panel, Secretary of 91制片厂视频 Margaret Spellings said she intended to take a close look at N-sizes and would schedule the fall conference to 鈥渉elp states improve their systems for ensuring the validity and reliability of their accountability decisions.鈥
鈥淲e will work with states to acquire new impact data on school and student inclusion rates and discuss with them a process for justifying how their specific N-size is necessary for valid and reliable results,鈥 she wrote.
Reginald M. Felton, the director of federal relations for the Alexandria, Va.-based National School Boards Association, said in an interview he welcomed the scrutiny and believes that federal officials will find states鈥 N-size limits are, for the most part, appropriate.
鈥淲e still hold firm that their purpose is to deal with statistical relevance and validity, and they do that,鈥 he said of the minimum sizes. 鈥淭he debate needs to shift to what is being done to ensure that students identified as in need of improvement get the help they need.鈥