Introduction
Reading research over the past 20 years has greatly advanced our understanding of reading problems. We know that children who experience difficulty learning to read have phonological processing weaknesses (e.g., Liberman, Shankweiler, & Liberman; 1989) and poor word recognition skills (e.g., Ehri, 1998; Siegel, 1989; Share & Stanovich, 1995). Moreover, a number of researchers (e.g., Blachman, Ball, Black, & Tangel, 1994; Byrne & Fielding-Barnsley, 1991; D. Fuchs, Fuchs, Thompson, Al Otaiba et al., 2001; Torgesen, Wagner, & Rashotte, 1997) have demonstrated that programs emphasizing phonological awareness and decoding can significantly improve young students' reading achievement, at least in the short term.
Yet, there is a persisting problem: Not all children respond even to the most effective interventions. Researchers have reported that 20% to 30% of children at risk for reading difficulties (e.g., Blachman et al., 1994; Brown & Felton, 1990; Mathes, Howard, Allen, & Fuchs, 1998; Torgesen et al., 1999), and more than 50% of children with disabilities (e.g., D. Fuchs, Fuchs, Thompson, Al Otaiba et al., 2001; Torgesen et al., 2001), do not appear to benefit from generally effective early reading intervention. Such students have been dubbed "nonresponders." Recently, unresponsiveness-to-intervention has been suggested as an alternative criterion to the current IQ-achievement discrepancy approach for identifying students with learning disabilities (e.g., President's Commission on Excellence in Special Education, 2002). This has contributed to researchers' interest in finding the best ways to identify nonresponders and develop effective interventions to reduce unresponsiveness (e.g., Case, Speece, & Molloy, in press; O'Connor, 2000; Speece & Case, 2001; Vaughn, Linan-Thompson, & Hickman, 2003; Vellutino et al., 1996).
Identifying Nonresponders
It is generally agreed that nonresponders to reading intervention are students who do not make adequate reading progress despite their participation in evidence-based instruction. However, there is little agreement about what constitutes "adequate reading progress" (Torgesen, 2000). In an extensive review of research addressing unresponsiveness to reading instruction, Al Otaiba and Fuchs (2002) showed that researchers have used one of two basic indicators of reading progress: performance level or growth rate. In terms of performance level, researchers have identified unresponsiveness as performance below the 10th percentile to below the 50th percentile on a given measure (e.g., Foorman et al., 1998; Torgesen et al., 1999; Vellutino et al., 1996). With respect to growth, nonresponders have been identified on the basis of no growth (e.g., Berninger et al., 1999; Torgesen & Davis, 1996, Uhry & Shepherd, 1997) or limited growth (e.g., Vadasy, Jenkins, Antil, Wayne, & O'Connor, 1997; Vellutino et al., 1996).
There are serious conceptual problems related to performance-level-only and growth-rate-only approaches. For example, a child's performance level may be low but, without considering her growth rate, it is difficult to determine whether she is responsive to intervention. She may be making important growth. Likewise, using only growth to determine unresponsiveness ignores information about a child's performance relative to meaningful educational benchmarks. A child may be making steady progress, but may still be performing at such a low level that she cannot be expected to reach an adequate level of competency in a timely manner.
Developing valid methods of identifying nonresponders is a recognized goal of early reading intervention research. One alternative to the performance-level-only and growth-rate-only methods is a "dual-discrepancy" approach (L. Fuchs & Fuchs, 1998), whereby students must be discrepant from their peers in both performance level and growth rate to be considered unresponsive. Researchers are determining whether this discriminates among average readers and poor readers who do and do not respond to instruction (e.g., Speece & Case, 2001; Speece, Case, & Molloy, 2003). Other researchers are testing its utility by comparing it to alternative procedures like "median split," "normalized," and "benchmark" posttreatment scores (see L. Fuchs, 2003). The dual discrepancy approach seems to hold promise, but more work remains.
Treating Nonresponders
A second question important to the study of nonresponders is how to address such children's failure to learn to read. Many who have implemented interventions for struggling readers have reported the proportions of nonresponders (e.g., Berninger et al., 1999; Foorman et al., 1998; Torgesen et al., 1999). However, only a few have attempted multi-phased interventions in an attempt to decrease rates of unresponsiveness (e.g., Case et al., in press; O'Connor, 2000; Speece et al., 2002; Vaughn et al., 2003; Vellutino et al., 1996). In this subgroup of ambitious studies, the first phase has consisted of instruction for students identified as at risk for early reading failure, and subsequent phases have consisted of continued instruction directed at nonresponders in the previous phase. Some of these studies have defined unresponsiveness in the context of general education instruction; others in the context of more intensive, small-group instruction.
Special-education-like approaches. Some researchers have conceptualized unresponsiveness as a failure to respond to instruction resembling traditional special education service delivery (e.g., Berninger et al., 1999; Foorman et al., 1998; Torgesen et al., 1999, Vellutino et al., 1996). That is, students were temporarily removed from the classroom to receive focused supplemental reading instruction from well-trained teachers. Vellutino et al., for example, implemented an intensive, one-to-one tutoring intervention for 15 weeks with at-risk first-graders. Students who did not make substantial progress during tutoring were designated as "difficult-to-remediate" and received further intervention. Vaughn et al. (2003) implemented daily small-group instruction with at-risk second-graders. Nonresponders (i.e., students who did not meet exit criteria after 10 weeks) received a second round of tutoring. Students who did not meet exit criteria after another 10 weeks received a third phase. After 30 weeks of intervention, less than 25% of the original at-risk sample had not met the exit criteria.
Whereas researchers working within this "special-education-like" framework have demonstrated that many poor readers improve when intensive intervention is in place, they have not examined whether or how general educators might modify their instruction, curricula, or materials to accommodate the learning needs of at-risk students. This, of course, was not the purpose of their research, and we mean no criticism of it. Yet, the role of the general education teacher and the nature of mainstream instruction seem pivotal in identifying the most difficult-to-teach students and best ways to meet their needs.
We offer two reasons for this view. First, as Vellutino et al. (1996) have suggested, some students' reading difficulties are no doubt due to inadequate instruction rather than a true reading disability. Improving general education instruction may be sufficient to help many struggling readers and to identify those in need of more intensive instruction at a lower cost than providing intensive instruction to all at-risk students. Second, current education reforms emphasize evidence-based, general classroom interventions and modifications as a first step in addressing students' academic difficulties (e.g., President's Commission on Excellence in Special Education, 2002). Thus, for both pragmatic and policy-related reasons, the quality and effectiveness of classroom instruction seem important. More comprehensive examinations of unresponsiveness should begin by determining whether modifying instruction in the general education classroom is effective.
General education approaches. O'Connor (2000) and Case et al. (in press) have done precisely this. O'Connor (2000) implemented four increasingly intensive levels of beginning reading interventions to kindergartners. Intervention at the first level was an evidence-based, whole-class, phonological awareness program conducted by general education teachers. Unresponsive students then received one-to-one tutoring from teaching assistants. Children who remained unresponsive received small-group instruction from their teachers at the beginning of first grade. Finally, still unresponsive first-graders received one-to-one tutoring from a researcher. The proportion of nonresponders decreased with each level of intervention. O'Connor's findings suggest that some poor readers benefit from evidence-based classroom instruction, whereas others require more intensive, individualized instruction.
Unlike O'Connor (2000), Case et al. (in press) did not implement an evidence-based classroom intervention before identifying nonresponders; rather, they defined unresponsiveness as a dual discrepancy between poor readers and their peers participating in regular classroom instruction. Case et al. then worked with the classroom teachers of the unresponsive students to design interventions supported by research, such as phonological awareness and phonics instruction, partner reading, or computer programs to address behavior and motivational needs. Most interventions were delivered only to the nonresponders. Speece et al. (2002) found that students who received the classroom interventions in the Case et al. study made greater academic gains than those who did not. Like O'Connor's findings, these results support the idea that evidence-based intervention within the general education classroom may be beneficial for at least some nonresponders.
Comparing levels of service delivery. The "special-education-like" approach described above is important to research examining unresponsiveness, in part because it identifies the most-difficult-to-teach students in relatively few steps: Students unresponsive to intensive, individualized instruction are likely to be among those most in need of ongoing, specialized services (e.g., Vellutino et al., 1996). But given the current emphasis on providing all students access to the general education curriculum, it is likely that most schools will encourage teachers to implement evidence-based classroom instruction and modifications before removing lagging students for individualized or small-group instruction (Vaughn, Gersten, & Chard, 2000). Moreover, schools typically do not have the resources for intensive, individualized interventions for all at-risk children. Thus, we believe that students' responsiveness to general education instruction (modified or unmodified) should be included in research focusing on how to identify and help nonresponders. As we have indicated, researchers working within a general education framework have begun to do this. What is still needed is a better understanding of approaches that are both effective for many nonresponders and practical to implement. To examine these issues, we directly compared the effectiveness of
- evidence-based classwide intervention delivered by the general education teacher,
- individualized modifications to generally effective classroom instruction, and
- more intensive one-to-one intervention in strengthening struggling students' reading achievement.
Purpose of the Study
The purpose of this study was twofold. First, we wished to further explore the validity of the dual-discrepancy approach (L. Fuchs & Fuchs, 1998); specifically, to refine a process that
- distinguishes a "risk pool" of lowest-performing readers,
- monitors their progress using valid measures, and
- identifies nonresponders to intervention in a way that minimizes false positives and false negatives.
In this study, we identified nonresponders to a generally effective classwide intervention, Peer-Assisted Learning Strategies (PALS; e.g., D. Fuchs, Fuchs, Thompson, Svenson et al., 2001). Both performance level and growth rate were measured using Curriculum Based Measurement (CBM; see Deno, 1985). Students dually discrepant from average readers with respect to performance level and growth rate were identified as nonresponders to PALS.
Our second purpose was to compare the achievement of nonresponders who continued receiving PALS to nonresponders who received one of two treatments designed to represent parts of the continuum of services available in schools for most students who are referred to or identified for special education: modified classroom intervention (Modified PALS), or individualized pull-out instruction (Tutoring). In this study, PALS served as a control, enabling us to compare the reading achievement of students who received increasingly individualized interventions to students who did not. PALS and Modified PALS may be considered best practices to meet the needs of most students in the general education classroom through evidence-based classwide interventions or through modifications (e.g., D. Fuchs & Fuchs, 1994). Tutoring reflects a more intensive, costly, and, according to some, "restrictive" level of intervention because the student is temporarily removed from his or her peers. Whereas this more intensive level, with its "pull-out" dimension, may not currently be desired by all, support for it is emerging from reading research (e.g., Ehri & Robbins, 1992; Torgesen et al., 1999; Uhry & Shepherd, 1997; Vadasy et al., 1997; Vellutino et al., 1996).
Previous Page | Next Page
(Abstract) | (Method: Participants)
