How School Administrators Cheat the Accountability Rules

06/01/2003
Summary of working paper 9307
Featured in print Digest

Some school administrators 'game the system' when faced with serious consequences if it seems likely that their students will not perform adequately in special testing programs. Schools reclassified students as disabled, putting them into 'special education' programs exempt from the state tests, and therefore ineligible to contribute to the school's aggregate test scores.

Schools in dozens of states face accountability measures based on aggregate student performance on standardized tests. The consequences of poor performance in these accountability systems include giving parents increased choice of schools, either within the public sector or through vouchers for private schools; reconstitution of the school; or closure, in the event of persistent identified failure of a school to improve. Just as some students respond to high-stakes exams by cheating, some school administrators "game the system" when faced with serious consequences if it seems likely that their students will not perform adequately in special testing programs.

On January 8, 2002, President George W. Bush signed into law the reauthorization of the Elementary and Secondary Education Act, also known as the No Child Left Behind Act of 2001 (NCLB). The centerpiece of this legislation involves implementing a system of school accountability. States must design systems of school report cards based on the fraction of students demonstrating proficiency in reading and mathematics.

If students do not make adequate yearly progress, then schools and districts face consequences, including mandatory public school choice and the possibility of school restructuring. In addition, states risk the loss of federal administrative dollars as federal funds are redirected elsewhere. Another risk is that, if the schools are graded lower by the government, the district will become less attractive to potential and current residents and that will hurt prices in the housing market.

In Accountability, Ability and Disability: Gaming the System (NBER Working Paper No. 9307), Research Associate David Figlio and co-author Lawrence Getzler study the effect of the introduction of the Florida Comprehensive Assessment Test (FCAT) in 1996 in six large counties in Florida that are not identified in the paper. Examining highly detailed data from a panel of as many as 4.1 million students, the authors find that the schools did "game the system" by reshaping the test pool. Schools reclassified students as disabled, putting them into "special education" programs exempt from the state tests, and therefore ineligible to contribute to the school's aggregate test scores. Following introduction of the testing regime, schools reclassified low income and previously low performing students as disabled at significantly higher rates. Moreover, these behaviors were concentrated among the low-income schools most likely to be on the margin of failing the state's accountability system.

Figlio and Getzler conclude that the introduction of the FCAT test is associated with an increase in the likelihood that a student will be classified as disabled by 5.6 percentage points. Altogether, 8.9 percent of the sample of students is identified as having a test-excludable disability. So the tests resulted in more than a 50 percent higher rate of disability classification in these six counties. Moreover, schools are more likely to switch low-performing students from a test-included category to a test-excluded disability following the introduction of the testing regime. Those schools with a higher rate of poverty, indicated by how many students are eligible for free lunches, tend to be more aggressive in reclassifying previously low-performing students as disabled, apparently hoping to avoid being classified as a failing school.

The reclassification of students as disabled "profoundly affects the student's individual educational experience," the authors note. It also reduces the accuracy in the grades or classifications given to schools based on the accountability exams, and thereby reduces the potential effectiveness of a public policy aimed at improving the educational system. Some students may end up in special education but would be better off in traditional classes. The trend also has an impact on total school costs, since special education on average costs 1.9 times as much per student as regular education.

Under the new federal NCLB Act, all students, including those with disabilities, will be included in the accountability testing system. Students must meet or exceed the state's proficient level of academic achievement by the end of the school year 2013-14, with intermediate goals along the way. Nonetheless, there will remain incentives to game the system, the authors state. For one thing, NCLB does permit giving those students with disabilities additional time to take the tests. Also, since schools will be required to have the same minimum percentage of students meeting proficiency, or at least to decrease the percentage of non-proficient students by 10 percent annually, they will continue to have an incentive to place relatively high-achieving students, say those with mild dyslexia, into the disability category to improve the probability of attaining adequate yearly progress.

-- David R. Francis