Are Students Not Counted?

K-12 Testing

A series of Associated Press news articles has spawned a new "urban myth" by claiming that the scores of nearly two million racial minority students are not being reported by their subgroup due to a "loophole" in the No Child Left Behind law, and that states are actively helping districts do use the loophole. Some derivative articles have reduced the story to a claim the students are not counted at all. In response, 14 Democratic legislators, led by Rep. George Miller, one of the authors of NCLB, have asked the Department of Education to close the alleged "loophole." As Republicans quickly joined the fray, Secretary of Education Margaret Spellings pledged to do so. But this is a fake controversy, useful only to those who seek to ensure that almost all schools in the nation become labeled as "failing."


In fact, The No Child Left Behind (NCLB) law mandates that almost all children are tested and their scores included in school and district level reports. Scores must also be reported for subgroups such as racial minorities, provided there are enough students in the subgroups so that individuals cannot be identified. Subgroups also must contain enough students to ensure group scores are reliable. Because the number of particular subgroup students is not large in many schools, that group's scores are not reported, But the students are included in the score reports of the whole school and district.


The provision in the law was written primarily to protect children. Parents do not want their children to be able to be publicly identified by their test scores. If the number of any group, e.g., African Americans, is small enough, then individuals could be identified.


Secondly, the minimum "cell size" requirement for subgroup reporting guards against statistical errors. Before NCLB was even signed, researchers Douglas Staiger and Thomas Kane discovered that year-to-year scores in schools varied greatly, particularly in any one grade. One year's students often scored higher or lower than the previous year's class for reasons that had little to do with anything the school did (see Examiner, Summer 2001). They concluded that measuring year-to-year changes among groups smaller than about 67 would lead to mis-identifying many schools as subject to NCLB sanctions. Because NCLB penalizes schools that do not increase their scores sufficiently for every subgroup, using small groups makes it almost certain that schools will be labeled failing for what are no more than normal fluctuations in scores. Professor Walt Haney separately concluded that to ensure accuracy, group sizes should not be below 100.


The news stories are accurate in stating that many states have increased their minimum subgroup size over the past few years. And arguably, they have done so to help reduce the numbers of schools being labeled failing (see Examiner, Fall 2005). This is not unreasonable given the overall irrationality and punitive nature of NCLB (See FairTest's report, Failing Our Children, for more on this issue.)


But every one of these states has had its plans approved by the U.S. Department of Education headed by Secretary Spellings, using the law promoted by Rep. Miller. In every case, the cell sizes are smaller than the low limit of 67 recommended by Kane and Staiger.