Fuchs’ view challenges hundreds of studies that have consistently found that inclusive educational settings have substantial benefits for the cognitive and social development of children with disabilities. That research has been instrumental in persuading lawmakers to increase funding to help schools accommodate students with disabilities, in some cases hiring extra special education teachers for every class. Roughly 15 percent of U.S. public school students have been diagnosed with a disability and receive services, according to the most recent data, so this debate over special education placement affects not only the academic prospects of students with disabilities but also the cost and structure of the whole educational system.
The paper, “Reframing the Most Important Special Education Policy Debate in Fifty Years: How versus Where to Educate Students with Disabilities in America’s Schools,” was co-authored by Allison Gilmour, a researcher of special education at the American Institutes for Research, and Jeanne Wanzek, a professor of special education at Vanderbilt. Fuchs provided me with a pre-publication draft and gave me permission to discuss it with other experts.
The core of Fuchs’ critique is that previous researchers failed to distinguish between students with disabilities who are sent to separate special education classrooms and students with disabilities who are included in general education classrooms. They are fundamentally different. Children who are placed in separate settings for a significant part or most of the day tend to have more severe disabilities and academic struggles. It should be no surprise to anyone that higher achieving students with milder disabilities end up with higher test scores than students who initially had lower test scores and more severe disabilities. That isn’t proof that a child with a disability learns more in a general education classroom. Ideally, from a research perspective, you’d want to randomly assign students with disabilities to both types of classrooms and see where they learn more. But that’s unethical, and impractical.
Researchers call this problem “selection bias” and they have tried to overcome it with statistical techniques. For example, they have compared students with disabilities who have similar demographic characteristics, such as the same race or ethnicity, similar family income and the same type of disability. Inclusion still comes out on top. However, Fuchs points out that many of these studies have still failed to account for the two most important factors: how the student was doing academically before the disability was diagnosed and the severity of the disability.
Beginning in the late 1980s, the federal government started to collect data on these two important, confounding factors – academic achievement before diagnosis and disability severity – so that policymakers could see how well students were faring under the 1975 federal law that mandates support for educating students with disabilities. Fuchs and his co-authors reviewed a 1991 analysis of this data, called the National Longitudinal Transition Study, and noted that it initially reported that high school students with disabilities learned more when they learned alongside their general education peers. But the appendix of the report disclosed that the advantage for special education inclusion disappeared when the academic gains were adjusted for prior academic achievement and measures of the students’ functional skills. Fuchs said there were no differences in outcomes between the two settings when researchers compared students who started with the same test scores and had the same disability severity.
Some recent studies with statistical sophistication still show that inclusion prevails. For example, in two studies of Indiana students with disabilities published in 2021 and 2023, researchers found that the more time that students spent in an inclusive setting, the better they did. However, Fuchs and his co-authors pointed out that more than half the students were thrown out of the 2021 study because of missing data and research design. They say the studies compared only the two extremes of students who spent 80 percent of the time or more in general education versus 80 percent of the time or more in separate classrooms, which was a very small group of students (only 75 in math and 63 in English language arts). Even with statistical adjustments for prior academic achievement, it’s hard to equate these two groups. Fuchs and his co-authors concluded that the validity of the two studies is “problematic.”
This is not the first time Fuchs has questioned the gospel that inclusion is best. In an article published 30 years ago, Fuchs criticized the wisdom of always educating children with disabilities in the general education classroom. In 2023, Fuchs published a study showing that even states with the highest rates of special education inclusion did not have consistently improving test scores for children with disabilities. Scores declined in some states.
Fuchs and his colleagues’ sharp critique of the strength of the evidence for inclusion is controversial, but they are not alone. In December 2022, the Campbell Collaboration, a widely respected international nonprofit organization that reviews research evidence for public policy purposes, also concluded that the benefits for inclusion were inconsistent and inconclusive. The Campbell reviewers threw out 99 percent of the 2,000 studies they found because of poor quality and research design, for reasons similar to those Fuchs describes. Only 15 studies survived. They found that math and reading scores, along with psychological, emotional and behavioral measures, were no higher for children with disabilities who learned in general education classrooms, on average, compared to children who learned in separate special education classrooms. Advocates for children with disabilities disputed the findings.
Lynn Newman, a researcher at SRI, a California-based research organization, has worked on multi-year studies of students with disabilities for the federal government. She said Fuchs’ paper makes some good points, but she said his argument also has some “holes” because it excludes some well-designed studies of more recent data, in which inclusion appears to be beneficial, especially among high-school students with disabilities.
Newman explained to me that there was very little support for students with disabilities in general education classrooms in the 1980s and 1990s. Inclusion has since improved, she said. She cited four studies (one, two, three, four), published between 2009 and 2021, showing that students fared better with inclusion.
I showed this research to Fuchs, who agreed that the methodology and quality were good, but he noted that these studies didn’t analyze whether students were learning more in one place than another. Instead, the studies focused on other outcomes like employment after high school. “The articles Newman identified are barking up a different tree,” he said by email.
Fuchs is concentrating on academic outcomes. He admits there may be other psychological or social benefits to learning alongside peers in general education classes. He did not study those. But those benefits could be even more important to parents, and to lifetime success. (Fuchs also did not review the evidence of how students without disabilities are affected by peers with disabilities in their classrooms. That is a different body of research.)
Measuring academic outcomes for students with disabilities is difficult. Students with disabilities are more likely to fail a general education class. Grades between the two settings – special education and general education – cannot be directly compared. Test scores are often lacking, especially before and after changes in special education placements.
Other scholars I talked to said Fuchs lumped all disabilities together. Two specialists in children with the most severe disabilities who need extensive support showed me recent studies that point to superior learning when these students are included in the general classroom, even though they rarely are. However, those students represent only 1 percent of the student population with disabilities.
In many ways, this debate shows how science responds to changing conditions. Decades ago, there weren’t many ways to help children with disabilities. Today there is a growing body of research about the best ways to teach children, especially young elementary school children, who are having difficulties with reading and math. Some of these interventions require daily instruction away from the general education classroom.
Fuchs doesn’t think his argument will lead to segregating all children with disabilities in self-contained classrooms. He envisions schools where students would be pulled out of the general education classroom on a daily basis to receive the reading and math instruction they need in a separate classroom. Some children with mild dyslexia, he said, might need only an hour a day of intensive reading instruction. Meanwhile some high-functioning children with Down syndrome might be able to remain in the regular general education class during reading time.
And just as the quality of separate, special education may be evolving, so too is the quality of inclusion in a general education classroom. Schools are getting better at supporting and accommodating students with disabilities. Clearly, a good version of inclusion will outperform a bad version of a separate classroom. And a good version of intense, specialized instruction will outperform a bad version of an inclusive classroom where the general education teacher is overwhelmed and lacks training. Too often, students aren’t getting the support they need.
School leaders are in a tough spot when they have to decide whether to invest in improving the general classroom to accommodate everyone or to create and refine interventions that happen outside of the classroom. And at the moment, research can’t really tell them what works best.
