During 1998–99, 5,541,166 students with disabilities, or 8.75% of the school-age population ages 6–21 years, received special education and related services under Part B of the federal Individuals with Disabilities Education Act (IDEA) (U.S. Department of Education, 2000). IDEA specifies 13 disability categories based on etiological groupings. The largest single category of disability served is specific learning disabilities (50.8%), with speech and language impairments the second largest category (19.4%). Children with mental retardation account for 11.0%, while children with autism, considered a low incidence disability, constitute 1% of those receiving special education and related services. Since the original passage in 1975 of IDEA's forerunner, the Education for All Handicapped Children Act, the categorical model has served as the basis for determining who qualifies for special education in accord with two premises. First, each disability category represents a separate and distinct condition that may co-occur with others but is not identical (Lyon, 1996), and second, dissimilar disability conditions, or deficits, require educational programs that differ from regular education programs.

The concept of inclusive schooling emerged during the 1980s and gained momentum in the 1990s in response to the two categorical premises of IDEA and its predecessor. Special education was viewed as a form of de facto segregation for children with disabilities. Furthermore, special education had assumed the mantle of a placement setting rather than seeing its primary function under federal regulations as the source of specially designed instruction and support services intended to meet children's unique learning needs (Giangreco, 2001). Because of these reasons, major educational reforms were necessary to transform schools and instruction (Skrtic, 1992). Educational equity and excellence for all students required “a restructured system of education, one that eliminates categorical special needs programs by eliminating the historical distinction between general and special education” (Skrtic, Sailor, and Gee, 1996, p. 146). Inclusive schooling became the democratic mechanism proposed to accomplish this restructuring. From the perspectives of curriculum and instruction, inclusive practices in a single system of education should foster engaged learning. All classrooms should be learner-centered, children should be supported to become active and self-regulated learners, and instructional practices should be grounded to theme-based units, cooperative learning, team teaching, and student-teacher dialogue that scaffolds critical inquiry in both intellectual and social realms (National Research Council, 1999, 2000).

IDEA specifies that children with disabilities must be educated in the least restrictive environment. This means that, to the greatest possible extent, a child should be educated with children who do not have disabilities and an explanation must be provided in the Individualized Educational Plan (IEP) of the degree, if any, to which a child will not participate in regular class activities (Office of Special Education and Rehabilitation Services [OSERS], 2000). In the practical implementation of this requirement in the past 10 years, the terms least restrictive environment, mainstreaming, and inclusion often have become confused (Osborne, 2002). Least restrictive environment is a legal requirement. It specifies that children can be removed from the regular education classroom for special education placement only if the nature and severity of the child's disability are such that education in regular classes with the use of supplementary aids and services cannot be satisfactorily achieved (OSERS, 2000). In other words, maintenance in the regular education classroom means that children receive special education and related services for less than 21% of the school day. Mainstreaming is an educational practice that refers to the placement of students in regular education for part of the school day, such as physical education or science classes.

Inclusion is neither specified in law nor regulations as a placement. Similar to resource rooms and self-contained classrooms as optional placement settings, inclusion is also an option along the educational continuum of the least restrictive environment. In its broadest sense, inclusion is an educational philosophy about schooling. Inclusive schools are communities of learners “where everyone belongs, is accepted, supports, and is supported by his or her peers and other members of the school community while his or her educational needs are met” (Stainback, Stainback, and Jackson, 1992, p. 5). Full inclusion is the complete integration of the regular and special education systems where all children with disabilities receive their education, including special education and related services, as an integral part of the regular education curriculum. A major criticism of full inclusion models has been that placement decisions begin to take precedence over decisions about children's individual educational needs (Bateman, 1995). In contrast, partial inclusion, which is more typical of current inclusive schooling, pertains to those situations in which one or more classrooms within a school or school district are inclusive. Most advocates of full inclusion would consider partial inclusion as inconsistent with the philosophy of inclusive schooling. Moreover, judicial tests of inclusion have primarily involved students with moderate to severe cognitive disabilities. In general, in case law situations, courts have ruled that “inclusionary placement” should be the placement of choice, with a segregated placement occurring only when the evidence is overwhelming, despite the school district's best efforts, that inclusion is not feasible (Osborne, 2002).

Depending on state education laws, speech-language pathologists may provide either special education (instructional) services or, in conjunction with special education, related (support) services. The provision of related services must meet standards of educational necessity and educational relevance (Giangreco, 2001). In full and partial inclusion, special education and related services, as specified in an IEP, may be delivered through multiple models, all of which are premised on collaboration. In this sense, collaboration means that team participants have particular beliefs and possess certain skills (Silliman et al., 1999; Giangreco, 2000). These include (1) a shared belief in the philosophy of inclusion, (2) empowerment as decision makers combined with respect for varying decision-making values, (3) flexibility in problem solving about how best to meet the language and literacy needs of individual students, (4) the shared expertise made possible through coteaching strategies, and (5) high expectations for all students, regardless of their educational and disability status.

Based on the collaboration concept, the American Speech-Language-Hearing Association (ASHA, 1996) supported “inclusive practices” as an option for optimally meeting the educational needs of children and youth. A flexible array of service delivery models for implementing inclusive practices was also specified in accord with children's needs at different points in time. These four general models are not viewed as mutually exclusive or exhaustive. One model is direct service delivery in the form of “pull-out” services, considered appropriate only when there is a short-term objective for a child to achieve, such as acquisition of a new communication skill through direct teaching. A second model is classroom-based service delivery, in which the speech-language pathologist and teachers collaborate, for example through team teaching, to incorporate children's IEP goals across the curriculum. The composition of team teaching models varies, but the teams may include a regular education teacher, a teacher of specific learning disabilities, and a speech-language pathologist who works full-time in the classroom (e.g., Silliman et al., 1999). A third model is community-based service delivery, in which communication goals are specifically addressed in community settings, such as a vocational education program that is a focus of a transition service plan. This plan is designed to bridge between the secondary education curriculum and adulthood and must be included in an IEP for students beginning at age 14 years (OSERS, 2000). The consultative model of service delivery is a fourth model; its main characteristic is indirect assistance. This model may be most appropriate when a child's communication needs are so specific that they do not apply to other children in the classroom (ASHA, 1996).

For any inclusion model to be effective in meeting children's diverse needs, two foundations must be functional. One concerns the change process that serves to translate conceptually sound research into everyday instructional practices. The process of changing beliefs and practices must be explicitly understood, supported, and crafted to the particular educational situation (Skrtic, 1992; Gersten and Brengelman, 1996). The second fundamental support involves the building and sustaining of educational partnerships. A team's capacity to sustain innovative and educationally relevant practices requires the successful integration of collaborative principles and practices (Giangreco et al., 2000).

Few studies have evaluated the outcomes of inclusion programs for students with language impairments. At least three predicaments have contributed to this situation. The first is the co-occurring disabilities dilemma. Few inclusion programs have been reported in the literature that specifically focus on children classified as having language impairment as the primary condition. The absence of data can be attributed in part to the categorical model, which fails to account adequately for co-occurring disabilities, much less the existence of category overlap. For example, the U.S. Department of Education (2000) continues to report that, for identification and assessment purposes, “learning disabilities and language disorders may be particularly hard to distinguish … because these two disabilities present in similar ways” (p. II-32). Procedural requirements contribute one part of this dilemma. State education agencies typically report unduplicated counts of children with disabilities. Only the primary disability category is provided. For example, specific learning disability, mental retardation, or autism is often reported as the primary condition, with language impairment classified as the secondary condition. However, when duplicated counts are available, such as the count of all disabilities for each child that the Florida Department of Education provides (U.S. Department of Education, 2000), language impairment emerges as the most frequent disability associated with another disability.

Second is the broad variations in research purposes and methods. Most outcome studies have primarily addressed the social benefits of inclusion for children with severe development disabilities, such as autism or Down syndrome. The results are complicated to evaluate because of significant disparities among studies in their definitions of inclusion, sample characteristics, such as ages, grades, gender, and type of disabilities, and the instructional or intervention focus (McGregor and Vogelsberg, 1999; Murawski and Swanson, 2001).

Two major issues confront the design of future research on the efficacy of inclusive practices (ASHA, 1996). First, in studying the cognitive and social complexities of teaching and learning, multiple research methodologies pursued in a systematic manner are necessary for exploring, developing, and testing hypotheses (Friel-Patti, Loeb, and Gillam, 2001) about the efficacy of the four inclusion models. Second, in moving past the social outcome focus, research strategies should examine individual differences in the ability to benefit academically and linguistically from inclusive education as a product of variations in instructional practices, expanding the scope of investigation beyond children's performance on standardized measures of language and academic performance.

A third predicament is the broad variation in instructional practices in inclusive classrooms. The fact that disruptions in language and communication development are implicated in a wide range of disabilities but not acknowledged as central to children's literacy learning (Catts et al., 1999) also has significant ramifications for research on academic outcomes of inclusion. For example, the effects of inclusion on emerging reading skill have been reported in a limited manner, primarily for children classified with learning (severe reading) disabilities (Klingner et al., 1998), less so for children with language impairment (Silliman et al., 2000). In general, the reading instruction for these children remained undifferentiated from the reading practices used with other students in the classroom, resulting in minimal gains. Thus, shifting a child to inclusion from a special education placement, or even maintaining a child in the regular education classroom, may mean that little has changed for that child in reality.

One implication for instructional practices is that educational team members, including speech-language pathologists, should be skilled in designing multilevel instruction that takes into account the individual child's needs for the integration of oral language dimensions with evidence-based practices for learning to read, write, and spell (National Institute of Child Health and Human Development, 2000). A second implication, supported by ASHA (2001), is that speech-language pathologists are critical stakeholders in children's literacy learning. They have the professional responsibility to bring their knowledge of language to the planning and implementation of prevention, identification, assessment, and intervention programs in order to facilitate children's literacy development.