Rethinking Service Delivery for Students With Significant Learning Problems

Abstract

Despite recent advances in the development of instructional interventions, many students with significant learning problems do not benefit from them. This includes 25% to 50% of students with learning disabilities (LD). In this article, we identify five limitations of current instructional programs that may help to explain students’ inadequate responsiveness. The first of these is that the instructional programs often fail to address the difficulty students experience when transitioning from the primary grades to the intermediate grades. Second, many of the programs lack sufficient comprehensiveness in the strategies or the skills they address. Third, they typically do not teach for transfer. Fourth, they do not adequately address the linguistic and cognitive limitations of many struggling students. Fifth, they do not make use of implementation features that can optimize the intensity of instruction. We describe these limitations and explain how researchers might modify their instructional programs to make them more effective for children with severe learning problems.

Keywords

instruction reading literacy math policy special education classroom

Millions of America’s children and youth have severe learning problems. This large group includes students with high- and low-incidence disabilities. The academic performance of most students with disabilities (SWD) is abysmal. The National Longitudinal Transition Study–2 (NLTS-2; Wagner, Newman, Cameto, Levine, & Marder, 2003) produced a representative snapshot of the academic achievement of SWD in high school. Those with learning disabilities (LD) were on average 3.4 years behind grade level in reading; 3.2 years behind in math. The Special Education Elementary Longitudinal Study (Schiller, Sanford, & Blackorby, 2008) was meant to parallel NLTS for the earlier grades. Schiller et al. (2008) indicated that 64% of elementary-age children with LD scored below the 21st percentile in reading comprehension in three consecutive years of the study. An important cause of SWD’s unacceptably low achievement in high school and the elementary grades is that schools fail to provide sufficiently intensive instruction—not because they willfully withhold it but because they do not recognize the need for it and have lost the know-how to provide it. To understand why, we turn to research and policy making conducted in the 1980s.

Research, Policies, and Practices of the Recent Past

Low Achievers With and Without an LD Label

No research group since the early 1980s, we believe, has had stronger influence on special education policy and practice than Ysseldyke, Algozzine, and their colleagues. Throughout the 1980s, this team conducted descriptive studies that purported to show that low achievers with and without an LD label perform similarly on tests of cognition and academic achievement. Perhaps their most widely known and influential study was described in “Similarities and Differences Between Low Achievers and Students Classified Learning Disabled” (Ysseldyke, Algozzine, Shinn, & McGue, 1982). On April 2, 2014, Google Scholar showed that this article had been cited 269 times. Ysseldyke et al. (1982) recruited 50 fourth graders with a school-given LD label and another 49 students without a label from the same school districts. The 50 labeled children had previously been assessed on the Peabody Individual Achievement Test (PIAT) and obtained an average total standard score of 91.9 (SD = 8.78). The 49 unlabeled children were also deemed “low achieving” because each had performed at or below the 25th percentile on a group-administered Iowa Test of Basic Skills. The authors then tested these 99 low achievers on a variety of 49 measures so that the labeled and unlabeled groups could be compared in cognitive, academic, perceptual-motor, self-concept, and school behavior domains.

Ysseldyke and colleagues ran inferential and descriptive analyses on these test data, after which they concluded the labeled and unlabeled low achievers did not differ in any important respect. They expressed this point repeatedly. In the third paragraph on page 83 of their article, they wrote, “No psychometric differences of practical utility between the groups were observed.” In the fourth paragraph on the same page, they claimed, “No difference was observed in the performance of the two groups on psychometric measures.” In the next paragraph (same page), they reiterated, “There were no psychometric differences in the performance of the two groups of students.” Still on page 83, “In this investigation we were unable to identify psychometric measures that would differentiate the groups” (pp. 83–84).

This oft-stated conclusion was understood by numerous academics and policy makers across several decades as evidence that “LD versus non-LD” is a spurious distinction; that practitioners’ use of the LD label is unscientific, arbitrary, and capricious. We base this last point on how the writers of many subsequent articles described and made use of the Ysseldyke et al. (1982) study. In short, Ysseldyke and associates’ articles in the 1980s deepened skepticism among many stakeholders about the validity of the LD construct, if not about the meaningfulness of high-incidence disabilities more generally.

Inclusive Practices

The Ysseldyke et al. (1982) article implicitly promoted a related and equally consequential notion, namely, that low performers with and without high-incidence disability labels would benefit—and benefit similarly—from “inclusive practices.” Others in the 1980s, like Madeleine Will (1986), Assistant Secretary of Special Education and Rehabilitation Services, were more explicit about this expectation. With Will’s help, the belief became part of a loosely defined policy that was widely known as The Regular Education Initiative. Advocates called for the elimination of resource rooms and self-contained classes. With monies to be saved by these closings, they argued for more specialists in general education and for the use of classwide interventions like the Adaptive Learning Environments Model (e.g., Wang & Birch, 1984), cooperative learning (e.g., Stevens, Madden, Slavin, & Farnish, 1987), and other “mainstreaming” programs that encouraged the participation of SWD and unlabeled low-achieving students (cf. D. Fuchs & Fuchs, 1994). Subsequent legislation and educational practices—for example, the 1997 reauthorization of Individuals with Disabilities Education Act (IDEA), accommodating the general education curriculum, coteaching—also reflected the view that better funded classroom settings with inventive instructional procedures and flexible staffing would benefit virtually all children and youth.

Underestimating the Seriousness of Many SWD’s Learning Problems and Overestimating the Capacity of General Education to Address Them

Underestimating SWDs’ Learning Problems

But if low achievers with and without an LD label are indeed very similar students requiring very similar instruction, and if inclusive practices like coteaching are equally efficacious for both groups, then why have millions of children with an LD label (and other high-incidence disability labels) performed so poorly across more than three decades? We offer an answer in two parts, the first of which requires a closer look at the Ysseldyke et al. (1982) study.

As mentioned, Ysseldyke and associates analyzed their data in multiple ways, but the two they valued most were “percentage of overlapping scores” and “number of exact pairs.” The first of these required the computation of “the percentage of scores within a common range for both [groups’ distribution of scores]” (Ysseldyke et al., 1982, p. 76). That is,

if the 49 scores for the low-achieving [unlabeled] group [on a given test] ranged from 8 to 17 and 45 of 50 of the scores for the LD group were also within that range, the percentage of overlap was calculated as 95/99 × 100 = 97% overlap. (p. 76; see Note 1)

The authors reported that “percentage of overlap . . . ranged from 82% to 100%” (p. 80). For the second analysis, the authors tallied “the number of students in the two groups who earned identical scores” (Ysseldyke et al., 1982, p. 80). Among 49 possible pairs, the number of identical scores across the various measures ranged from 19 to 44.

While Ysseldyke et al. (1982) believed these findings “raise serious concerns regarding the differential classification of poorly achieving students as either LD or non-LD” (p. 82), they gave short shrift to another set of results that tell a different story. In their Table 2, they listed the means and standard deviations of the two groups’ performances on the 19 subtests from the Woodcock–Johnson. They also displayed mean differences between the groups on the subtests, indicating the ones that were statistically significant. The table shows that the students with the LD label did significantly worse on 10 of these 19 subtests. In Table 3, the authors presented the groups’ scores on 25 additional measures representing cognitive, academic, perceptual-motor, self-concept, and school behavior domains. Among the 25 comparisons, the group with the LD label performed statistically significantly worse on 8, which included four of five subtests from the PIAT as well as the PIAT total score.

Ysseldyke et al. (1982) were dismissive of these numerous significant findings on grounds that “the magnitude of [the] mean differences is at best moderate” (p. 79). That is, the LD group’s poorer performance on many measures may have been reliable, but it lacked practical importance and, on this basis, it should be ignored. Kavale, Fuchs, and Scruggs (1994) calculated effect sizes (ESs) for each of the group comparisons displayed in Ysseldyke et al.’s Tables 2 and 3. The median ES difference across the Woodcock–Johnson subtests in Ysseldyke et al.’s Table 2 was 0.45 (range = 1.60 to −0.17; see Kavale et al., Table 3), with the unlabeled children scoring higher. The median ES among the eight tests and subtests in the achievement domain in their Table 3 was 0.80 (range: 1.10 to −0.05; see Kavale et al., Table 4), again with the unlabeled students achieving the higher scores.

Reasonable people can disagree on how large an ES must be before it may be regarded as practically important. Nevertheless, it is an incontrovertible fact that the children with the LD label in the Ysseldyke et al. (1982) study performed on average reliably worse than the unlabeled children on many measures, especially on those of academic achievement. The data suggest that teachers, school psychologists, and other building-based professionals do not arbitrarily assign the LD label to students. Rather, they tend to give it to the lowest of low achievers who arguably require more intensive instruction.

This apparent fact should not be interpreted as evidence or support for the LD construct. It is not. But neither should it be ignored—as it was ignored by Ysseldyke and his associates in the 1980s, by researchers in the 1990s using their own (rather than school) identification criteria to explore the validity of the distinction between IQ-achievement discrepant students from nondiscrepant students, and by policy people in the 2000s who continue to insist that parsing “garden-variety” low achievers is of little pedagogic value. By not recognizing that school-based personnel often use disability labels to distinguish their most academically vulnerable children, academics, policy makers, and others have contributed to a widespread underestimation of the severity of these children’s learning problems and the failure by educators to come to their educational rescue.

Overestimating General Education’s Instructional Capacity

The Ysseldyke et al. (1982) perspective on the identification and classification of children as LD, and its uncritical acceptance by many academics and policy makers during the last three decades, is only a partial explanation of the absence of intensive instruction and unacceptably low academic achievement of many SWDs in our schools. A related cause is that the so-called inclusionary strategies for general education classrooms—accommodating the curriculum, coteaching, and so forth—were (and continue to be) viewed by stakeholders as valid means of ensuring an appropriate education for virtually all students.

It is beyond the scope of this article to review the empirical literature on each of these well-intentioned approaches to strengthen general education instruction for SWD and their nondisabled peers. Suffice to say, most serious students of this literature have concluded that these practices and policies have not provided meaningful help to large numbers of struggling learners. This does not mean that these practices were (or are) without benefit. McMaster and Fuchs (2002), for example, reviewed the empirical literature on cooperative learning to determine the proportion of SWD who responded positively to a cooperative approach. The answer: about 50%. One likely explanation of these mixed results is that SWD differ in the severity of their learning difficulties. Those with more severe problems need an intensity of instruction that goes well beyond what cooperative learning, coteaching, and other whole-class approaches provide.

Intensive Instruction and Responsiveness to Intervention (RTI)

RTI has been adopted by thousands of teachers and administrators because of a growing belief that many students with significant learning difficulties are not benefiting from so-called inclusive practices and are requiring instruction delivered in a more intensive and timely manner. All RTI frameworks reflect the intent to provide multiple tiers of increasingly intensive instruction. They are a hallmark of the RTI approach. Tier 1 is commonly understood as general education instruction, Tier 2 as small-group instruction, and Tier 3 as something more intensive and often unspecified.

Despite broad agreement on a need for RTI—and irrespective of its exemplary execution by a minority of school-based educators—we detect an ambivalence among many toward its multiple tiers. This ambivalence, we believe, reflects lingering skepticism about a need for greater intensity of instruction, which, in turn, is rooted in the persistent beliefs that the problems of SWDs and “garden-variety” low achievers are not very different from each other, and general educators can help SWD with general education solutions. This ambivalence and skepticism help explain why “intensity of instruction” is still inadequately conceptualized and unspecified and infrequently conducted.

There are more challenges associated with RTI. Establishing a broadly effective Tier 2—let alone an effective Tier 3—has proved more elusive than many imagined. Although randomized control trials show that many at-risk students respond well to carefully constructed and expertly conducted Tier 2 instruction, the research indicates, too, that instruction at Tier 2—even when implemented with fidelity—does not help all academically needy students (e.g., Gilbert et al., 2013; O’Connor & Fuchs, 2013; Vaughn et al., 2010). Researchers estimate the equivalence of 2% to 7% of the general population failing to respond adequately to Tier 2 instruction, including 25% to 50% of students with an LD label (e.g., Al Otaiba & Fuchs, 2002; D. Fuchs, Compton, Fuchs, & Davis, 2008; D. Fuchs, Fuchs, & Compton, 2012; O’Connor & Fuchs, 2013). Moreover, these rates of inadequate response were produced when (resourced) researchers were conducting instruction. It is reasonable to expect higher rates of poor response when (lesser resourced) practitioners implement instruction.

What follows is a critique of current approaches to Tier 2 instruction that have been used by us and by others. We focus on Tier 2 because, among the schools in which practitioners are working hard to make RTI work, it is Tier 2 that tends to be best understood, most popular, and most often implemented. Yet, our sense is that Tier 2 approaches collectively reflect five noteworthy limitations. First, they typically do not address the difficulties many students face when transitioning to the intermediate grades. Second, they typically lack sufficient comprehensiveness. Third, they do not explicitly teach for transfer of learning. Fourth, they do not address many of the linguistic and cognitive limitations that contribute to the low achievement of students with LD. And fifth, they infrequently make use of implementation features that can optimize intensity of intervention. We describe and illustrate these problems and explain how researchers and practitioners might expand current conceptions of intensive instruction and strengthen instruction at Tier 2 in reading and mathematics in the next decade. We focus on Grades 3 to 5 when curricular challenges increase substantially, especially now with Common Core State Standards (CCSS).

Suggestions for Strengthening Instructional Intensity

Addressing Key Problem Areas

Areas of difficulty ignored or inadequately addressed in instructional research include the transitions from stories to informational text and from whole numbers to common fractions, decimal equivalents, and algebraic thinking (FD&A). Poor comprehension of informational text undermines learning in school and after school and negatively affects overall academic achievement (Meneghetti, Carretti, & De Beni, 2006; Taraban, Rynearson, & Kerr, 2000). FD&A is foundational for algebra, for success with more advanced mathematics, and for competing successfully in the American workforce (National Mathematics Advisory Panel). These key concerns are reflected in CCSS at Grades 3 through 5, which should encourage researchers to address these topics successfully. As researchers focus on developing effective instructional programs involving informational text and FD&A for students with serious learning problems, they should not forget to emphasize word-level skills and fluency to support comprehension and arithmetic concepts and principles and whole-number skill to support FD&A. Incorporating such foundational skills in reading and mathematics instruction is consistent with the CCSS learning progressions and addresses difficulties most students with significant learning problems experience.

Broadening Instructional Content

Previous studies on informational text at Grades 3 through 5 have typically targeted one or two instructional components. For FD&A, several randomized control trials have addressed common fractions or other types of rational number word problems or prealgebraic thinking. Moreover, this research has focused nearly exclusively at middle school or high school. By contrast, we believe there should be an earlier focus on these topics, which is consistent with CCSS. Research should evaluate the overall effects of instructional packages that combine multiple validated components, as well as untested program elements. For informational text, this may mean evaluating instructional programs that integrate background knowledge, morphology, intra- and extratext inference making, and strategies to foster higher standards for text coherence. In mathematics, there can be a combined focus on FD&A as interventions simultaneously target big ideas, procedures, calculations, and word problems. Research on such interventions is needed to evaluate efficacy and to explore student characteristics associated with adequate/inadequate responsiveness.

Transferring Learning

Researchers know that transferring learning is a major challenge for children with significant learning difficulties. Prior work shows the benefits of explicit instruction. In a large randomized control trial, L. S. Fuchs et al. (2003) contrasted schema-based instruction (teaching students to recognize the underlying mathematical structure of whole-number word-problem types) with and without explicit transfer instruction. As part of transfer instruction, teachers explained that superficial problem features (e.g., response format, vocabulary) can make word problems seem unfamiliar without altering the problem type or solution strategies. Students also practiced sorting novel problems in terms of superficial problem features and searching novel problems for familiar problem types. L. S. Fuchs et al. found statistically significant and important benefits for explicit instruction on transfer word problems.

Such work has been limited to whole-number word problems and to higher performing at-risk students. It should be extended to students with severe learning difficulties. At Grades 3 through 5, this includes word-problem types involving FD&A. It also includes comprehension of informational text by explicitly teaching students to transfer to less coherent and predictable text, and to texts with different structures.

Strengthening Language Comprehension (LC) and Executive Function (EF) as Part of Academic Skills Instruction

We focus on LC and EF (e.g., working memory, attentional control) for two reasons. First, individual differences in these cognitive areas are strongly associated with reading and mathematics performance in Grades 3 through 5. Second, students with an LD label often experience deficits in LC and EF.

Most skills-based academic programs are designed to compensate for struggling students’ LC and EF weaknesses. Some researchers, aware of children’s inadequate LC, have used short simple sentences to explain why a given procedure is valuable, say, in solving math problems. Nevertheless, too many of these students respond inadequately to such instruction (e.g., L. S. Fuchs, Fuchs, & Compton, 2013). Interventions should attempt to address LC and EF weaknesses more effectively.

Yet, many interventions designed to improve LC and EF deficiencies apart from academic instruction have not strengthened children’s academic performance (e.g., Kearns & Fuchs, 2013; Melby-Lervåg & Hulme, 2013). A different approach, one perhaps with greater promise, is to embed LC or EF instruction into explicit academic skills instruction. Research is needed, for example, on embedding LC instruction in FD&A word-problem narratives to focus on the many LC challenges that are specific to FD&A word problems. Another example is to build students’ attentional control for executing the multiple cognitive processes involved in building text representation and a situation model. EF games might be developed that require students to store, process, and update an increasing number of key elements from passages, while gradually increasing the number of off-goal elements they must inhibit. The intention here is to strengthen LC and EF in the context of explicit skills intervention.

Optimizing Instructional Intensity

Two variables associated with increased implementation intensity are group size and duration of instruction. Across reading and mathematics, few studies have manipulated group size as an independent variable to examine its effects on learning. But syntheses of studies not designed to compare group size suggest 3 to 4 students per group may be optimal for students with LD (Lou, Abrami, Spence, & Poulsen, 1996; Wilkinson & Fung, 2002). However, task-relevant verbal interactions between tutor and tutee and among students appear superior in pairs (Thurlow, Ysseldyke, Wotruba, & Algozzine, 1993; Trowbridge & Durnin, 1984; Webb, 1989). Moreover, the influence of group size may differ as a function of intervention procedures. Research is necessary to contrast one-to-one interventions in reading and mathematics with groups of varying sizes (e.g., 2:1 vs. 4:1).

Duration of instruction refers to the length of instructional sessions, the frequency with which sessions are delivered, and the number of weeks the intervention lasts. So, strategies for increasing intervention duration are to increase (a) the amount of time for each session and/or (b) the number of times per week the sessions are scheduled and/or (c) the number of weeks the intervention lasts. Here, too, additional research is needed to identify maximum duration, and whether it is affected by methods of intervention, the personnel who deliver intervention, and student characteristics such as attentiveness of their behavior.

Significant Learning Difficulties Require a Significant Instructional Response

Millions of students with high-incidence disabilities have failed and are failing in schools across America because the severity of their learning problems has been and continues to be grossly underestimated. This “misread” of their instructional needs is caused in part by academics and policy makers who for decades have uncritically accepted groundless assertions that students with and without high-incidence disability labels are, for all intents and purposes, the same students who will respond equally well in “inclusive” arrangements characterized by accommodating the curriculum, coteaching, and universal design. Many of these “inclusive practices,” however, are not evidence based. They have not been shown to be efficacious for the students for whom they were supposedly designed. Other inclusive practices have proved helpful to some SWD, but unhelpful to many more. An important reason for this is that they lack instructional intensity. Many children and youth with disabilities will continue to perform poorly until the significance of their learning problems is recognized and they are provided an appropriate education, which is their legal and civil right.

Footnotes

Acknowledgements

We thank two anonymous reviewers for their helpful comments on an earlier draft.

Authors’ Note

We are solely responsible for the content, which does not necessarily represent the views of the U.S. Department of Education or the National Institutes of Health.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Research described in this article is supported in part by Grant R324D130003 from the Institute of Education Sciences in the U.S. Department of Education and by Core Grant HD15052 from the Eunice Kennedy Shriver National Institute of Child Health and Human Development.

Notes

References

Al Otaiba

Fuchs

(2002). Characteristics of children who are unresponsive to early literacy intervention: A review of the literature. Remedial and Special Education, 23, 300–316. doi:10.1177/07419325020230050501

Fuchs

Compton

D. L.

Fuchs

L. S.

Davis

G. C.

(2008). Making “secondary intervention” work in a three-tier responsiveness-to-intervention model: Findings from the first-grade longitudinal study at the National Research Center on Learning Disabilities. Reading and Writing: An Interdisciplinary Journal, 21, 413–436. doi:10.1007/s11145-007-9083-9

Fuchs

L. S.

(1994). Inclusive schools movement and the radicalization of special education reform. Exceptional Children, 60, 294–309.

Fuchs

L. S.

Compton

D. L.

(2012). Smart RTI: A next-generation approach to multi-level prevention. Exceptional Children, 78, 263–279. doi:10.1177/0022219412468889

Fuchs

L. S.

Fuchs

Compton

D. L.

(2013). Intervention effects for students with comorbid forms of learning disability: Understanding the needs of nonresponders. Journal of Learning Disabilities, 46, 534–548.

Fuchs

L. S.

Fuchs

Prentice

Burch

Hamlett

C. L.

Owen

. . . Janeck

(2003). Explicitly teaching for transfer: Effects on third-grade students’ mathematical problem solving. Journal of Educational Psychology, 95, 293–304. doi:10.1037/0022-0663.95.2.293

Gilbert

J. K.

Compton

D. L.

Fuchs

L. S.

Bouton

Barquero

L. A.

Cho

(2013). Efficacy of a first-grade responsiveness-to-intervention prevention model for struggling readers. Reading Research Quarterly, 48, 135–154. doi:10.1002/rrq.45

Kavale

K. A.

Fuchs

Scruggs

T. E.

(1994). Setting the record straight on learning disabilities and low achievement: Implications for the policymakers. Learning Disabilities Research & Practice, 9, 70–77.

Kearns

D. M.

Fuchs

(2013). Does cognitively focused instruction improve the academic performance of low-achieving students? Exceptional Children, 79, 263–290.

10.

Lou

Abrami

P. C.

Spence

J. C.

Poulsen

(1996). Within-class grouping: A meta-analysis. Review of Educational Research, 66, 423–458.

11.

McMaster

K. L.

Fuchs

(2002). Effects of cooperative learning on the academic achievement of students with learning disabilities: An update of Tateyama-Snizek’s review. Learning Disabilities Research & Practice, 17, 107–117.

12.

Melby-Lervåg

Hulme

(2013). Is working memory training effective? A meta-analytic review. Developmental Psychology, 49, 270–291. doi:10.1037/a0028228

13.

Meneghetti

Carretti

De Beni

(2006). Components of reading comprehension and scholastic achievement. Learning and Individual Differences, 16, 291–301. doi:10.1016/j.lindif.2006.11.001

14.

O’Connor

Fuchs

L. S.

(2013). Responsiveness to intervention in the elementary grades: Implications for early childhood education. In Buysse

Peisner-Feinberg

Cantler

(Eds.), Handbook of response to intervention (RTI) in early childhood education (pp. 41–56). Baltimore, MD: Brookes.

15.

Schiller

Sanford

Blackorby

(2008, August). A national profile of the classroom experiences and academic performances of students with learning disabilities: A special topic report for the Special Education Elementary Longitudinal Study. Menlo Park, CA: SRI International.

16.

Stevens

R. J.

Madden

N. A.

Slavin

R. E.

Farnish

A. M.

(1987). Cooperative integrated reading and composition: Two field experiments. Reading Research Quarterly, 22, 433–454.

17.

Taraban

Rynearson

Kerr

(2000). College students’ academic performance and self-reports of comprehension strategy use. Reading Psychology, 21, 283–308. doi:10.1080/027027100750061930

18.

Thurlow

M. L.

Ysseldyke

J. E.

Wotruba

J. W.

Algozzine

(1993). Instruction in special education classrooms under varying student-teacher ratios. The Elementary School Journal, 93, 305–320.

19.

Trowbridge

Durnin

(1984). Results from an investigation of groups working at the computer. Washington, DC: National Science Foundation.

20.

Vaughn

Cirino

P. T.

Wanzek

Wexler

Fletcher

J. M.

Denton

C. A.

. . . Francis

D. J.

(2010). Response to intervention for middle school students with reading difficulties: Effects of a primary and secondary intervention. School Psychology Review, 39, 3–21.

21.

Wagner

Newman

Cameto

Levine

Marder

(2003). Going to school: Instructional contexts, programs, and participation of secondary school students with disabilities. A report from the National Longitudinal Transition Study-2 (NLTS2). Menlo Park, CA: SRI International. Retrieved from www.nlts2.org/reports/2003_12/nlts2_report_2003_12_complete.pdf

22.

Wang

M. C.

Birch

J. W.

(1984). Effective special education in regular classes. Exceptional Children, 50, 391–398.

23.

Webb

N. M.

(1989). Peer interaction, problem-solving, and cognition: Multidisciplinary perspectives. International Journal of Educational Research, 13, 1–119.

24.

Wilkinson

I. A. G.

Fung

I. Y. Y.

(2002). Small-group composition and peer effects. International Journal of Educational Research, 37, 425–447.

25.

Will

M. C.

(1986). Educating children with learning problems: A shared responsibility. Exceptional Children, 52, 411–415.

26.

Ysseldyke

J. E.

Algozzine

Shinn

M. R.

McGue

(1982). Similarities and differences between low achievers and students classified learning disabled. The Journal of Special Education, 16, 73–85. doi:10.1177/002246698201600108