The Myth of the Special Education Burden

The debate already swirling around the upcoming renewal of the major federal law addressing special education, the Individuals with Disabilities Education Act, draws much of its energy from a widely shared but probably false premise: that schools are increasingly swamped with disabled children who are diverting scarce resources away from other students. Both the education establishment and IDEA reformers tend to accept this premise, although they draw different conclusions from it. The establishment wants the federal government to cover the costs of special education more fully so that, as these groups see it, public schools will finally have the wherewithal to deliver on the promise of improving the achievement of nondisabled students. Cost-conscious reformers, on the other hand, want the costs of special education contained, figuring that money devoted to general education is likely to yield better results than money devoted to special education.

But what makes everyone think the schools are being inundated by more and more children with learning problems? It’s true that the proportion of children in special education has increased significantly since the mid-1970s, when the IDEA began. The percentage of K-12 students identified as needing special education rose from 8.3 percent in 1976-77 to 11.8 percent in 1998-99. But an increase in the percentage of students identified as needing special education does not necessarily mean that there has been an increase in the percentage of students with disabilities, any more than an increase in reports of domestic violence necessarily means there is more domestic violence. Trends in these statistics are sensitive to reporting biases and shifting definitions.

A close examination of different categories of special education enrollments suggests that, while the identification of special education students has increased, the actual number of disabled youngsters has remained approximately steady. Almost the entire increase in special education enrollments since 1976 can be attributed to a rise in one category, called “specific learning disability,” which has more than tripled, from 1.8 percent of the student population in 1976-77 to 6 percent in 1998-99. All other categories of special education combined, including mental retardation, serious emotional disturbance, deafness, blindness, autism, and head injury, have actually declined from 6.5 percent to 5.8 percent of the student population during the same period.

If a general increase in the proportion of students with learning problems were truly under way, it should be evident in more than just one category of special education. It is highly implausible that something has caused there to be more children in one disability category without also causing more mental retardation, serious emotional disturbance, and so on. It is more likely that the large increase in the category of specific learning disabilities can be attributed to a greater likelihood of diagnosing children with those problems than to a true increase in the incidence of those learning problems (and only those learning problems) in the student population.

This seems especially likely when we recognize that this high-growth category, specific learning disability, consists of learning problems that are more subjective in their diagnosis and less expensive in their treatment than other categories of special education. The relative subjectivity of identifying specific learning disabilities makes it possible for changes in the number of children with that diagnosis to be caused by an increase in the propensity to assign that label. The relatively low cost of treating specific learning disabilities may further incline schools and educators to assign that label, especially if the additional funds produced by identifying a child with a specific learning disability exceed the marginal cost of providing that student with relatively minimal services.

Whether or not schools make a “profit” off these children, they may face other incentives to identify students in this way. They may be more willing to diagnose students with specific learning disabilities so as to exempt them from accountability testing, to reduce expectations about their academic performance, to solve behavior problems in “regular” classrooms, to get additional help to a child who is struggling academically (whether disabled or not), or in response to parental requests for special treatment. And, of course, one cannot rule out the possibility that schools have increased their rate of diagnosing specific learning disabilities because they have become better informed about such disorders and are more adept at recognizing them.

The malleability of the “specific learning disability” diagnosis is underscored by the odd way in which those disorders are identified. To be diagnosed as having a specific learning disability, students must perform significantly worse in a subject area, like math or reading, than is indicated by their cognitive potential, as typically measured by an IQ score. But this mismatch between potential and achievement is not always caused by a “disorder in one or more of the basic psychological processes involved in understanding or in using spoken or written language,” as the official definition of the disorder requires. It could also result from exposure to ineffective teaching methods. Schools may not be receiving more students with learning problems; in an era of results-based accountability, they may simply be more likely to identify students who have been the victims of educational malpractice.

Whatever the causes of the increase in children diagnosed with specific learning disabilities, it is extremely unlikely that nonschool factors are producing more children afflicted with such disabilities while producing no increase in all other categories of disability combined. But this is not how defenders of the education status quo see it. They persist in arguing that schools are being drowned by a tidal wave of children with learning problems caused by forces outside the school system.

Consider the arguments advanced by Sheldon Berman, Perry Davis, Ann Koufman-Frederick, and David Urion in their chapter in Rethinking Special Education for a New Century, a volume released a year ago by the Thomas B. Fordham Foundation and the Progressive Policy Institute. The authors are convinced that “the increases [in special education that] schools have been experiencing have not been caused by school district policy and practice.” Instead, they insist, “these cost increases have been primarily due to the increased numbers of children with more significant special needs who require more costly services.” In particular, they identify three nonschool phenomena that, they say, have increased the numbers of children with learning problems: improvements in medical technology, deinstitutionalization of children with serious difficulties, and increases in childhood poverty.

All these putative causes have an air of plausibility to them, but all also turn out to be inconsistent with the broader facts. It’s true that improvements in medical technology have saved the lives of more low-birth-weight babies and others with health problems, leading to an increase in children who manifest learning problems later in life. But improvements in medical technology and other improvements in public health, such as reductions in lead paint and safer car seats, have also helped prevent some children from developing disabilities at all, or have reduced the severity of those problems. Improvements in medical care and public health would thus lead to a net reduction in the number of children with learning difficulties.

The authors contend that increasing numbers of surviving low-birth-weight babies resulted in an estimated increase in children expected to have mental retardation, from 4,550 between 1980 and 1985 to 12,375 between 1995 and the present. However, the overall number of mentally retarded children in schools actually declined from 961,000 in 1976-77 to 597,000 in 1998-99.

Obviously, low-birth-weight babies contribute only a small amount to the total number of children with mental retardation. But it appears that the increase in mental retardation attributable to medical technology saving low-birth-weight babies has been more than offset by a significant reduction in the number of mentally retarded children, a reduction attributable to other improvements in medical care and public health.

The claims about deinstitutionalization and childhood poverty are no less misleading. While the authors provide no numbers, they contend that the deinstitutionalization of mentally retarded children in particular has placed a growing burden on school systems. Yet, as we have already seen, the total number of mentally retarded children served by schools under the IDEA has steeply declined. Children with specific problems in math or reading were never institutionalized, so deinstitutionalization cannot explain the rise in the one category of special education that has grown.

Nor is childhood poverty a plausible explanation for the increase in specific learning disabilities. First, the federal government’s definition of specific learning disabilities explicitly excludes learning problems that result from economic disadvantage: “The term does not include children who have learning problems which are primarily the result of visual, hearing, or environmental, cultural, or economic disadvantage.” Second, poverty among children under age 6 is actually about the same now (16.9 percent in 2000) as it was when the IDEA began (17.7 percent in 1976). It’s true that, during the intervening years, including recessions in the early 1980s and early 1990s, childhood-poverty percentages were sometimes higher, but this particular measure of poverty can be misleading because it excludes noncash benefits, such as public housing, Medicaid, and food stamps. Those governmental programs act as a cushion (albeit an imperfect one) to reduce basic material deprivation for children during economic slowdowns.

Poverty comparisons over time are also complicated by the shifting goal post of the poverty line. The larger reality is that, as our gross domestic product grows over time, America is becoming a wealthier society, and this is true for those at the lower end of the income scale as well as (if not as dramatically as) those at the upper end. In 1976, the average family in the lowest quintile of income earned $12,696 (in 2000 dollars), compared with $14,232 for the average bottom-quintile family in 2000. Economic deprivation is simply not a plausible explanation for a tripling in the percentage of children with specific learning disabilities (and only children in that category) especially when the definition of the disorder explicitly excludes economic causes.

If I am correct that there has been no large secular increase in the true incidence of learning problems since the mid-1970s, but rather an increase in the identification of students with those problems, several important policy conclusions follow.

First, money devoted to special education should not be viewed as money taken away from the general education of students. Schools have simply shifted more students who previously would have been in general education into special education. Whether schools were previously underdiagnosing or are currently overdiagnosing, the basic fact is that schools have been given additional money over time to educate a population of students with the same spectrum of learning problems that they used to have.

If nonschool factors have not stuck schools with more disabled students, then we would expect schools to have produced better outcomes with the additional resources. They have not. Over the last four decades, per-pupil spending in real dollars has increased from $2,360 to $7,086.Yet student outcomes, as measured by the U.S. Department of Education’s National Assessment of Educational Progress and high school graduation rates, have been relatively flat.

Whatever the causes for this productivity crisis in education (spending more without improving outcomes), it is not reasonable to blame special education for consuming extra dollars or burdening schools with more difficult-to-educate students. Schools have more money to educate the same distribution of students than they used to have. Shifting students and money into the special education category does not alter this basic fact.

Second, if the federal government were more fully to compensate schools for the costs of the IDEA mandates, we should expect a further increase in the shifting of students into the special education category. The additional outside funds that learning-disabled students bring to a school may already play a role in the tripling of students with that diagnosis.

As with Medicare and Medicaid, policymakers need to balance the desire to cover necessary costs with the need to prevent abuse of the special education program. Improvements in the diagnosis system could take several forms. Washington might institute a system of spot-check auditing of diagnoses. Perhaps diagnostic patterns in schools can be compared against demographic profiles to identify schools that appear to be overdiagnosing or underdiagnosing specific disorders. These might be subject to closer review.

But the most powerful reform would “voucherize” all students who receive a special education diagnosis. This is actually happening in Florida under the McKay Scholarship program, whereby disabled youngsters may take the money allocated for their education to private schools if they wish.

Making all special education students eligible for vouchers not only expands the options available to them and their families, it also provides a disincentive to public schools to overdiagnose students, since public schools will not want to lose these students to private schools. (Any resulting problem of underdiagnosis could be handled through a state-level appeal system or, as today, through the legal system.)

Much is wrong with American special education. But one thing is not wrong: Schools are not suffering under the burden of a growing population of children with special needs. The evidence suggests that schools have the same distribution of students with learning difficulties that they used to have. Any reforms considered as part of the Individuals with Disabilities Act renewal should take this fact into account.