Treatment validity as a unifying construct for identifying learning disabilities.

The purpose of this article is to revisit the issue of treatment validity as a framework for identifying learning disabilities. In 1995, an eligibility assessment process, rooted within a treatment validity model, was proposed that (a) examines the level of a student's performance as well as his/her responsiveness to instruction, (b) reserves judgment about the need for special education until the effects of individual student adaptations in the regular classroom have been explored, and (c) prior to placement, verifies that a special education program enhances learning. We review the components of this model and reconsider the advantages and disadvantages of verifying a special education program's effectiveness prior to placement.

**********

Over the 25-year history of the Individuals with Disabilities Education Act, the percentage of students with learning disabilities has increased dramatically so that students with learning disabilities now comprise more than 50% of all children with disabilities (U.S. Department of Education, 1999). This precipitous increase in the prevalence of learning disabilities has raised questions about the methods by which these children are identified.

Concern over the process by which students with learning disabilities are identified appears well founded. Because learning disabilities are a "soft" disability (Reschly, 1996), for which no discernable physical markers currently are known, the identification process is subjective. Moreover, because learning disabilities are associated with unexpected failure to learn, the discrepancy between intelligence and achievement is the central organizing theme of most definitions of learning disabilities. Yet, the measurement of those discrepancies has proven problematic due to poor reliability of difference scores (Reynolds, 1984); in fact, varying discrepancy formulae and test instruments identify different sets of students (Shepard, Smith, & Vojir, 1983). In addition, research documents similar underlying deficits in children with reading disabilities who do and do not demonstrate discrepancies (Fletcher et al., 1998). These problems have led some to question the viability of the learning disabilities concept.

At the same time, public awareness about the scientific controversy over learning disabilities has grown. One reason for increasing public awareness is the high cost of a special education -- more than double the cost of regular education (Parrish, 1995). Another reason is the phenomenon of overrepresentation of students of color in special education. This has prompted placement bias litigation (Larry P. v. Riles, 1979/1986; Marshall et al. v. Georgia, 1984/1985; S-1 v. Turlington, 1986), which has led to public debate over learning disabilities and mild mental retardation identification issues (e.g., Polloway, 1985; Reschly, 1984, 1988; Snow, 1984) and about the very quality and value of special education services (D. Fuchs & Fuchs, 1995a, 1995b; Reschly, 1996).

The 1982 National Research Council's (NRC) (Heller, Holtzman, & Messick, 1982) analysis of the problem of minority student overrepresentation set the stage for much of the focus, tone, and substance of this debate. That NRC report reformulated the problem of overrepresentation from one of reducing disproportionality to that of determining the conditions under which inequality constitutes inequity of treatment. According to Messick (1984), disproportion signifies inequity only when children are unduly exposed to classification because they receive poor-quality regular education, are assessed invalidly for special education, or receive an ineffective special education that hinders their educational progress.

Over the past decade, this reformulation of the overrepresentation problem, along with the increasing prevalence of learning disabilities, concern about conceptual and technical issues with the identification of learning disabilities, and questions about special education costs and effectiveness, has fostered calls for eligibility decision making that relies on treatment validity as a unifying concept. With a treatment validity approach, the value of an eligibility assessment process is judged by its capacity to simultaneously inform, foster, and document the necessity for and effectiveness of special treatment (e.g., Reschly, 1988; Reschly & Grimes, 1995; Witt & Gresham, 1985). A treatment-oriented approach to identification strives to eliminate the inequity potentially associated with overrepresentation of minority students in special education by maximizing regular education's potential effectiveness for all students and reserving judgment about the need for special education until the effects of individual adaptations in the regular classroom have been assessed and until evidence verifies that a special education program enhances learning.

In 1995, L.S. Fuchs described how curriculum-based measurement (CBM; Deno, 1985), a well-developed assessment system that permits modeling of students' responsiveness to instruction, can be used within a treatment validity approach to the identification of learning disabilities (also see L.S. Fuchs & Fuchs, 1998). One important component of that model was that special education demonstrate its value before a student is eligible for those special education services. This component has created controversy (e.g., Martin & Fuchs, 2000).

In this article, we consider the advantages and disadvantages of this component. We begin by providing an overview of the model as it originally was proposed, and we consider evidence supporting the tenability of a treatment validity approach. We then review the rationale for requiring special education to demonstrate its added value and consider the potential disadvantages of this model component. We conclude by proposing a modified treatment validity model for identification of learning disabilities.

THE TREATMENT VALIDITY MODEL AS ORIGINALLY PROPOSED

Assumptions

Most assessment frameworks for identifying students with disabilities who require intervention beyond standard classroom instruction focus on the absolute level of a student's performance, relative to similar peers, at a fixed point in time. A treatment validity approach, by contrast, requires that special education be considered when a student not only performs below the level of peers but also demonstrates a substantially lower learning rate; that is, the student manifests an unresponsiveness to the instructional environment. We refer to unsatisfactory performance along with inadequate growth as a "dual discrepancy."

An example from pediatric medicine illustrates the rationale for a focus on dual discrepancies. The endocrinologist monitoring a child's physical growth is interested not only in height at one point in time but also in growth velocity over time (Rosenfeld, 1982). Given a child whose current height places him or her below the third percentile, the endocrinologist considers the possibility of underlying pathology and the need to intervene only if, in response to an adequately nurturing environment, the child's growth trajectory is flatter than that of appropriate comparison groups. Based on long-term, large-scale normative information (Tanner & Davies, 1985), this criterion, at age 7, for example, is operationalized as an annual growth rate of less than 4 cm. So, the physician judges a 7-year-old who is short relative to peers, but who is nonetheless growing at least 4 cm annually in response to a nurturing environment, to be deriving available benefits from that environment and to be an inappropriate candidate for special intervention.

The endocrinologist's decision-making framework reflects three assumptions. First, genetic variations underlie normal development, producing a range of heights across the population. Second, in response to a nurturing environment, a short but growing child does not present a pathological profile but rather suggests an individual who may legitimately represent the lower end of the normal distribution on height -- an individual whose development is commensurate with his or her capacity to grow. Third, under these circumstances, special intervention is unlikely to increase adult height sufficiently to warrant the risks associated with that intervention. Alternatively, when a child is not manifesting normatively appropriate rate growths in a nurturing environment, additional assessment is warranted to determine whether more obtrusive forms of treatment are needed. (When questions arise about the quality of the environment, the first-level response is to remove those uncertainties by enhancing nurturance so that growth can be tested under adequate environmental conditions.)

The endocrinologist's decision-making framework translates into three related propositions for the identification of learning disabilities. First, because student capacity varies, schooling outcomes will differ; so, a low-performing child who nonetheless is learning may ultimately perform less well than peers. Second, if a low-performing child is learning at a rate similar to that of other children in the same classroom, he or she is demonstrating the capacity to profit from that educational environment. Therefore, the student probably does not require a unique form of instruction and probably is achieving commensurate with his or her capacity to learn. Third, under these circumstances, the risks and costs associated with entering the remedial or special education system are deemed inappropriate because it is unlikely, in light of the growth already occurring, that a different long-term outcome could be achieved as a function of that intervention. By contrast, when a low-performing child fails to manifest growth in a situation where others are thriving, alternative instructional methods must be tested to address the apparent mismatch between the student's learning requirements and those represented in the conventional instructional program. (As with the endocrinologist's framework, when the educational environment is insufficiently nurturing, as revealed by inadequate learning for the majority of students in a class, classroom intervention must occur to enhance the overall quality of the instructional program. That is, growth under more nurturing conditions must be indexed before any child's need for special intervention can be assessed.)

A Treatment Validity Model

Together, these assumptions led to a proposal for an alternative model of learning disability identification that relies on treatment validity as its unifying concept. With this model, Phase I assessment is used to determine whether the overall rate of responsiveness in the classroom indicates that the instructional environment is sufficiently nurturing to warrant individual student decisions. If the mean rate of growth in the classroom is low, when compared to other classrooms in the same building, in the same district, or in the nation, the appropriate decision is to intervene at the classroom level to provide a stronger instructional setting.

Once the presence of a generally nurturing regular classroom instructional environment has been established, Phase II assessment is used to identify children with dual discrepancies; that is, students whose level of performance and rate of improvement are dramatically below those of classroom peers. Phase III assessment generates the database for enhancing instruction in the general education classroom and determining whether that regular education setting can, with adaptations, produce better growth and thus be transformed into an acceptable learning situation for the individual. Only when corrective action fails to yield improved growth does consideration of special services to supplement the general education program become warranted.

If so, prior to a learning disability classification, Phase IV assessment evaluates special education effectiveness for a given student. If special education effectiveness for that child cannot be documented, no compelling rationale exists for assigning a learning disabilities label or removing the child from the classroom for instruction. This last phase is the component of the model that has generated controversy. Later in this article, we return to this topic.

CBM's Contribution to a Treatment Validity Approach to Identification

Within a treatment validity approach to identification, an assessment system must document that (a) the regular classroom is producing acceptable growth for many students and thus is nurturing, (b) important dual discrepancies on performance level and growth rate exist between a target child and classroom peers, (c) inadequate individual learning occurs even with general education adaptations, and (d) improved growth can be derived with the provision of a special education. To provide these multiple sources of documentation, an assessment method must be valid for modeling academic growth, distinguishing between ineffective general education environments and unacceptable individual student learning, informing instructional planning, and evaluating relative treatment effectiveness. CBM is one assessment method that can generate these multiple sources of documentation.

The purpose of CBM is to index academic competence and progress. In developing CBM, Deno (see Deno, 1985) sought to establish a measurement system that (a) teachers could use efficiently; (b) would produce accurate, meaningful information with which to index standing and growth; (c) could answer questions about the effectiveness of programs in producing academic growth; and (d) would provide information that helped teachers plan better instructional programs. A 15-year research program has addressed questions about features of the measurement system (what to measure, how to measure, and how to use the resulting database) against issues of technical adequacy, treatment validity, and feasibility in the domains of reading, spelling, mathematics, and written expression.

In each domain, CBM integrates key concepts from traditional measurement theory and from the conventions of classroom-based observational methodology to forge an innovative approach to assessment. As with traditional measurement, every assessment samples a relatively broad range of skills by sampling each dimension of the annual curriculum on each weekly test. Consequently, each repeated measurement is an alternate form, of equivalent difficulty, assessing the same constructs. This sampling strategy differs from typical classroom-based assessment methods, where teachers assess mastery of a single skill and, after mastery is demonstrated, move on to a different, presumably more difficult skill (see L.S. Fuchs & Deno, 1991). CBM also relies on a traditional psychometric framework by incorporating conventional notions of reliability and validity so that the standardized test administration and scoring methods have been designed to yield accurate and meaningful information.

By sampling broadly with standardized administration and scoring procedures, the CBM score can be viewed as a "performance indicator": It produces a broad dispersion of scores across individuals of the same age, with rank orderings that correspond to important external criteria, and it represents an individual's global level of competence in the domain. Practitioners can use this performance indicator to identify discrepancies in performance levels between individuals and peer groups, which helps inform decisions about the need for special services or the point at which decertification and reintegration of students with disabilities might occur.

At the same time, however, CBM departs from conventional psychometric tools by integrating the concepts of standardized measurement and traditional reliability and validity with key features from classroom-based observational methodology: repeated performance sampling, fixed-time recording, graphic displays of time-series data, and qualitative descriptions of student performance. Reliance on these methods permits slope estimates for different time periods and alternative interventions for the same individual, which in turn creates the necessary database for testing the effects of different treatments for a given student. Research also suggests that, when combined with prescriptive decision rules, these time-series analytic methods result in better instruction and learning: Teachers raise goals more often and develop higher expectations (L.S. Fuchs, Fuchs, & Hamlett, 1989a); they introduce more adaptations to their instruction (L.S. Fuchs, Fuchs, & Hamlett, 1989b); and they effect better student learning (L.S. Fuchs, Fuchs, Hamlett, & Stecker, 1991).

In addition, because each assessment simultaneously samples the multiple skills embedded in the annual curriculum, CBM can yield qualitative descriptions of performance to supplement graphed, quantitative analyses of CBM total scores. These diagnostic profiles demonstrate reliability and validity (see L.S. Fuchs, Fuchs, Hamlett, & Allinder, 1989; L.S. Fuchs, Fuchs, Hamlett, Thompson, Roberts, Kubec, & Stecker, 1994), offer the advantage of being rooted in the local curriculum, provide a framework for determining strategies for improving student programs (see L.S. Fuchs, Fuchs, & Hamlett, 1994), and help teachers plan more varied, specific, and responsive instruction to meet individual student needs (L.S. Fuchs, Fuchs, Hamlett, & Allinder, 1991).

Because of these features, CBM meets four technical requirements of a treatment validity approach to identification. First, CBM has the capacity to model academic growth. With a common test framework administered to children within a fixed age range, CBM can be used to judge performance repeatedly over an academic year on the same raw score metric. Therefore, CBM provides equal scaling of individuals throughout the range of behavior measured over time on a construct and difficulty that remain constant. CBM taps constructs that are qualitatively constant over an academic year, for which the difficulty level remains the same.

Second, CBM has the capacity to distinguish between ineffective instruction and unacceptable individual learning. The effectiveness of regular classroom instructional settings can vary dramatically (L.S. Fuchs & Fuchs, 1998). Therefore, a student's responsiveness to instruction cannot be judged in absolute terms. For example, in Classroom A, where instruction is relatively ineffective, slopes are generally low, and a child with a low growth rate represents the typical level of responsiveness to the instruction provided. Classroom B, by contrast, is associated with much stronger rates of growth. Therefore, the individual rate of growth that mirrored the profile of Classroom A would suggest a lack of responsiveness to the effective learning environment represented in Class B and indicate a potential learning problem. Early evidence (L.S. Fuchs & Fuchs, 1998; Speece & Case, 2001) suggests that CBM can be used effectively to distinguish ineffective instruction from problematic individual benefit.

Third, CBM informs instructional planning. A well-established, long-standing research program documents how CBM can help teachers plan better instruction and effect superior achievement. Studies have examined the effects of alternative data-utilization strategies, as well as CBM's overall contribution to instructional planning and student learning (see L.S. Fuchs & Fuchs, 1998).

The final technical requirement that CBM satisfies is evaluating treatment effectiveness. To function adequately within a four-stage eligibility assessment process, the assessment method must answer the following questions: Is the nonadapted regular classroom environment producing adequate growth? Have adaptations introduced in the regular classroom setting resulted in an improved growth rate? Has the provision of specialized services enhanced the student's learning? To answer these treatment effectiveness questions, the assessment must demonstrate sensitivity to student growth and to relative treatment effects, and it must permit comparisons of the effectiveness of alternative service delivery options. And research illustrates how CBM achieves sensitivity (Allinder & Fuchs, 1994; L.S. Fuchs Fuchs, & Hamlett, 1989b; Marston, Fuchs, & Deno, 1986) as well as comparisons across alternative service delivery arrangements (D. Fuchs, Fuchs, & Fernstrom, 1993; Marston, 1987-1988).

Illustration of the Treatment Validity Model

We illustrate the treatment validity identification model in Figure 1 with CBM data for a fictitious second grader, Jose. Between September and November, the mean CBM performance of Jose's classroom peers was 75 words read correctly per minute (standard deviation = 24); their slope averaged 1.35 words increased per week (standard deviation = .14). Phase 1 assessment verified an acceptable instructional environment, as demonstrated by a classroom growth rate that conforms nicely to normative second-grade data (L.S. Fuchs, Fuchs, Hamlett, Walz, & Germann, 1993). However, Jose's CBM performance averaged only 18.1 words read correctly per minute, with a slope of -.02 correct words increased per week. This Phase II comparison placed Jose more than 2 standard deviations below the classroom mean in terms of both performance level and slope.

[FIGURE 1 OMITTED]

Due to this dual discrepancy, the classroom teacher, in consultation with the school psychologist, engaged in Phase III assessment by developing and testing two prereferral interventions: adding Accelerated Reader (with which Jose worked at the library on reading books and answering questions about those books) and incorporating drill on sight words with homework. Unfortunately, the CBM data revealed that neither pre-referral intervention substantially improved Jose's level or rate of reading improvement.

Consequently, the multidisciplinary team, which included Jose's mother, decided on a Phase IV assessment: a special education diagnostic trial for 20 minutes of daily individual reading instruction. The special education teacher first established a goal of a 2.5 word weekly increase for Jose. If met, this would result in Jose reading approximately 70 words correct at the end of the year. This desired rate of improvement appears on Jose's graph as a diagonal solid line (with "G' showing the end-of-year goal). Then, based on a phonetic analysis of Jose's current CBM reading errors, the special education teacher identified vowel teams and r-controlled words as problematic. Therefore, during an initial special education intervention, the teacher provided mnemonics instruction in decoding these two phonetic patterns, and incorporated systematic, explicit teaching for decoding these words as they appear in natural text. With this intervention, Jose's rate of growth increased to 2.46 words per week -- a level greater than that of classroom peers and commensurate with the teacher's ambitious goal.

As the special educator continued to work with Jose, however, she noticed that the last three data points indicated that performance might be leveling off. Therefore, she conducted another diagnostic analysis of Jose's CBM database, this time examining the strategies Jose brought to bear in interacting with the text. The teacher noted that Jose made a high proportion of miscues that failed to preserve the meaning of the text. Consequently, the special educator designed the next intervention to address this problem: On a daily basis, Jose read into a tape recorder; listened to himself read as he reread along silently; and identified and corrected errors that failed to preserve the meaning of the text. With this intervention, Jose maintained a strong and acceptable rate of reading growth, a 2.24-word increase per week. Moreover, his performance at the end of May (65.4 words read correctly in a minute) approximated the teacher's goal of 70 words.

At this time, a decision was made that the special education diagnostic trial had proved the effectiveness of special education for Jose. The multidisciplinary team reviewed Jose's graph, along with other sources of data, to formulate a decision that an Individualized Education Plan (IEP) would be designed for Jose, which included 20 minutes of individual reading instruction each day.

EMPIRICAL WORK SUPPORTING A TREATMENT VALIDITY APPROACH

Research Showing Differential Rates of Responsiveness for Students with Learning Disabilities

At the core of a treatment validity approach to identification is the notion that a learning disability, in its most basic sense, is a "failure to thrive." Within this conceptualization, the student is viewed as a "nonresponder" to conventional educational treatments. Some research supports the tenability of such a proposition to characterize children identified as having a learning disability.

For example, L.S. Fuchs, Fuchs, Karns, Hamlett, Katzaroff, and Dutka (1997) examined the effects of a task-focused goals treatment on poor achievers with and without learning disabilities. Regardless of their disability status, children in the task-focused goals treatment increased their effort more than contrast students. This increased effort, however, was associated with differential learning only for low achievers without learning disabilities. The effect size for amount learned was .94 for children without learning disabilities; zero for children with learning disabilities. Differential effectiveness for low achievers with and without learning disabilities also has been documented with other treatments. For example, studies reveal a pattern of smaller effects for peer-mediated learning activities for students with learning disabilities than for nonlabeled low achievers (e.g., L.S. Fuchs, Fuchs, Hamlett, Phillips, & Bentz, 1994; L.S. Fuchs et al., 1995; L.S. Fuchs et al., 1997). Similar findings have emerged for cooperative learning (Tateyama-Snizek, 1990).

Conceptually, the phenomenon of differential effectiveness may be viewed as an aptitude-by-treatment interaction, whereby learning disability connotes some constellation of affective and cognitive states that mediates the effects of treatments. In this way, findings support the premise that learning disabilities may constitute a distinct form of low achievement, more resistant to intervention. Because research has failed to identify a useful diagnostic battery for diagnosing the constellation of affective and cognitive states, which predicts nonresponsiveness (Speece, 1990), such distinctiveness may be assessed more defensibly through a treatment validity approach.

Research Documenting Rates of Nonresponsiveness

Some work documents rates of nonresponsiveness to otherwise effective treatments. Potentially, these rates provide insight into whether a treatment validity model for identification, whereby students are identified for special services based on their lack of responsiveness to otherwise effective teaching methods, results in the expected incidence rates for learning disabilities.

Foorman, Francis, Fletcher, Schatschneider, and Metha (1998), for example, documented the effectiveness of explicit instruction on the alphabetic code, compared to an embedded and an implicit instructional approach on the reading performance of economically disadvantaged, low-achieving first and second graders. With this effective explicit instructional method, 16% of the children remained unresponsive to treatment (compared to 44% and 46% of children in the other treatments). However, this figure of 16% is high relative to the expected incidence of learning disabilities. Nevertheless, one might assume that, with additional (Phase II, prereferral) intervention aimed at enhancing outcomes within the confines of general education support, the 16% figure could be reduced so that it more closely approximates the expected incidence of learning disabilities.

In this way, L.S. Fuchs, Fuchs, Hamlett, Phillips, and Karns (1995) randomly assigned general educators to two treatments, both of which implemented ongoing CBM as well as classwide peer tutoring in mathematics with all students in their classes beginning in September. Prior work (e.g., D. Fuchs, Fuchs, Mathes, & Simmons, 1997; L.S. Fuchs, Fuchs, Phillips, Hamlett, Karns, & Dutka, 1997) shows that students achieve better in general education settings with this combination of CBM and peer-assisted learning strategies. As with most "effective" treatments, however, some portion of students do not respond to CBM with peer-assisted learning strategies (i.e., approximately 10% of students fail to demonstrate persuasive progress; see D. Fuchs et al., 1997; L.S. Fuchs et al., 1994).

Because of the need to focus more carefully on students who manifest unacceptable performance and growth, the second treatment in this study focused on individual adaptations conducted in regular classrooms, as might be conceptualized with Phase II assessment. Beginning in November, the bimonthly CBM class reports identified up to two target students per class whose CBM progress was inadequate (i.e., low level combined with low slope relative to classmates). For these students, teachers (a) formulated an adaptation before the next two-week report; (b) implemented that adaptation at least four times in the upcoming two weeks; and (c) when CBM identified the same student multiple times over reports, modified previous adaptations to enhance progress.

Nevertheless, despite many focused, data-based attempts to enhance learning, some children proved unresponsive to regular classroom adaptations. Two cases illustrate students' differential responsiveness. Over a 12-week period, a fourth-grade teacher implemented a rich set of adaptations, relying on basic facts drill, motivational workcharts and contracts, and manipulatives. The target student, who exhibited a CBM slope of .21 digits per week when identified for adaptation, responded well to these modifications to the regular classroom and completed the school year with a slope of .63 digits per week -- the average slope for the class. This success contrasts with the experience of a third-grade teacher who also implemented a large number and rich set of adaptations, including drilling basic facts, slicing back to second-grade material, implementing a motivational workchart, and using money to work on conceptual underpinnings. Despite this teacher's similar level of effort to modify regular classroom instruction, her target student demonstrated little improvement in growth rate: He ended the year with a relatively low slope of .28 digits per week, which was similar to his slope at the time he was identified for adaptation, and which was considerably lower than his classmates' average weekly slope of .98 digits.

In all, teachers effected substantial improvement for 33% of their target students, who had been identified on the basis of dual discrepancies. This suggests that, with the assistance of rich assessment information and consultative support to formulate feasible adaptations, regular classroom teachers may address the problems of some portion of students who initially demonstrate significant learning discrepancies from classroom peers. If we multiply the estimate of initial levels of unresponsiveness to CBM and peer-assisted learning strategies (10%) with the estimate of those who remain unresponsive with additional classroom adaptation (100% of initial unresponders minus 33% who proved responsive with adaptations = 77%), we arrive at a figure of 7.7%. This better approximates the incidence of learning disabilities.

Research on a Treatment Validity Approach to Identification

Some researchers have examined more directly the tenability of a treatment validity model for identification. For example, Vellutino and colleagues (1996) identified children who demonstrated early signs of reading problems, as judged by their first-grade teachers in November. A subsample of 74 poor readers received individual daily tutoring for 15 weeks. A linear regression was run for each child using time in months between kindergarten and fall of second grade as the independent variable and reading scores as the dependent variable. From these data, Vellutino et al. estimated slopes to represent each child's reading growth. On the basis of these slopes, the tutored poor readers were classified further in terms of treatment responsiveness. Vellutino et al. then described one group of poor readers as the "difficult to remediate" children and, on the basis of results, argued that treatment provided a "firstcut diagnostic" in distinguishing between disabled and nondisabled learners. This approach is consistent with the notion of learning disability as a failure to thrive; the diagnostic framework conforms nicely to a treatment validity perspective on identification.

Research Examining the Tenability of a Dual Discrepancy

Dual discrepancy as an index of "failure to thrive" has considerable intuitive and empirical appeal. Such an index deals directly with the problem at hand (e.g., poor reading, poor math skills), reflects a dynamic rather than a static approach to learning and assessment, and is data based. However, it requires a seismic shift in beliefs, attitudes, and practice. The details of ideally implementing the treatment validity model with dual discrepancy include (a) assessing every child in every classroom every week, (b) evaluating progress on a regular basis, (c) formulating interventions within general education classrooms for children identified as dually discrepant, (d) implementing the interventions with fidelity, and (e) evaluating the effects of the intervention. Clearly, the treatment validity model requires much effort and vigilance, especially compared to diagnostic procedures that rely on identification of IQ-achievement discrepancies or simple low achievement. The question becomes, "Is it worth it?"

One way of addressing this question is through comparison with other identification procedures. Speece and colleagues (Speece & Case, 2001; Speece, Molloy, & Case, 2000) compared the dual discrepancy method to IQ-reading achievement discrepancy and low reading achievement definitions of reading disabilities in an epidemiological sample of first- and second-grade children to assess the validity of the dual discrepancy approach. Only the validity of using dual discrepancy as a marker of reading disability was assessed. This corresponds to Phase II of the treatment validity model in which children are identified as being below their classmates on level and rate of progress. However, the determination of dual discrepancy status was based on CBM oral reading fluency measures collected across approximately six months of a school year. This approach was taken to reflect reading problems that were likely more severe than those based on several weeks, as proposed in the initial problem certification phase of the four-phase treatment validity model.

The population (N = 694) was screened on CBM letter-sound fluency (first grade) and oral reading fluency (second grade) probes to identify at-risk (N = 144) and purposive (N = 129) samples. The at-risk children comprised the lowest 25% of children in each classroom. To form the purposive sample, five children representing a range of skill above the 30th percentile on the screening measures were selected from each classroom. The at-risk and purposive sample children were followed across the school year and were administered a minimum of 10 CBM oral reading fluency probes to determine dual discrepancy status (CBM-DD). Other measures of intelligence and reading achievement were used to form the IQ-reading achievement discrepancy (IQ-DS) and low reading achievement (LA) groups. Validation measures included several phonological processing variables, teachers' ratings of academic competence, problem behaviors, and social skills, age, gender, and race.

The poor reader groups were formed hierarchically. All children who exhibited a dual discrepancy were assigned to the CBM-DD group. The remaining children who met the classification criteria were assigned to the IQ-DS group for the first set of comparisons and to the LA group for the second, and parallel, set of comparisons. Using these procedures, 47 children were identified as CBM-DD, 17 as IQ-DS, and 28 as LA. The prevalence of CBM-DD was 8.1% compared to 5.9% for IQ-DS (based on all children who met the classification criteria regardless of whether they qualified for either of the other groups). By definition, the LA group would reflect approximately 25% of the population because we used a standard score below 90 on a normreferenced measure as the criterion. Thus, it appears that the CBM-DD procedures are likely to identify a reasonable number of children with interventions implemented as proposed in the full treatment validity model. It should be noted that the CBM-DD group obtained significantly lower reading scores compared to the at-risk and purposive sample children not classified as poor readers.

Comparisons between the poor reader groups on the continuous measures are reported as effect sizes in Table 1. The CBM-DD group, compared to IQ-DS, was more impaired on every measure with the exception of reading and word reading efficiency. There were fewer and more modest differences between CBM-DD and LA. Notably, the CBM-DD group had poorer phonological awareness skills at second grade and were rated lower on academic competence, problem behaviors, and social skills. Interestingly, the four cross-sectional comparisons that used grade (including IQ-DS) indicated that CBM-DD children may become more impaired over time (age could not be used as a covariate in the inferential analyses because the assumptions of the test were not met). It is notable that the CBM-DD group was younger than either of the other poor reader groups. This finding has implications for early identification and intervention, which is a major educational conundrum (Good, Simmons, & Kameenui, 2001; Lyon, 1995).

Not reflected in Table 1 are the results for race and gender. All three methods included equal numbers of females and males (approximately 50-50), likely because research criteria were used to form the groups. For both the CBM-DD and LA groups, racial distributions reflected the proportions of majority and minority children in the schools. The IQ-DS group included a disproportionately high number of majority children

Messick (1989, 1995) proposed that validity be judged relative to the meaning (construct validity, relevance and utility) and use (social consequential validity and values) of test scores, broadly defined. In the present case, scores refer to classification as dually discrepant. The data presented by Speece and Case (2001) are most relevant to construct and social consequential validity.

The group comparisons support construct validity in that the CBM-DD group demonstrated more problems in skills that underlie beginning reading. Further, teachers viewed CBM-DD children as less academically competent and as exhibiting more social and classroom problem behaviors. These effects were stronger in comparison with IQ-DS than with LA. However, small effect sizes (i.e., .20) are not trivial (Cohen, 1992). Further, the dual discrepancy method demonstrated social consequential validity in that decisions to identify children in this manner would reflect gender and racial equity. Also, the children were younger, a consequence valued by educators, and considerably fewer were identified compared to a low achievement definition. The CBM-DD classification was the only one of the three methods to reflect favorably the three social consequences that were investigated (gender and racial equity, early identification).

In response to the question, "Is it worth it?," we suggest the answer is "yes," keeping in mind that demonstration of validity is an ongoing process (Messick, 1995). Another important question is whether the reading disability community, and by extension the learning disability community, can embrace the implications of the model (assuming that additional evidence can be marshaled). As noted earlier, the model requires a great deal of change, which is not a defining feature of American education (Tharp & Gallimore, 1988). As research is completed with respect to validity, we need to ask if and how changes may be implemented in the service of more valid procedures.

THE CONTROVERSIAL FOURTH PHASE: REQUIRING SPECIAL EDUCATION TO DEMONSTRATE ADDED VALUE PRIOR TO PLACEMENT

The Rationale for a Fourth Phase

As originally proposed, the treatment validity model for identification includes a fourth phase, which requires that the added value of special education be demonstrated before a child qualifies for special services. The decision to include this fourth phase was motivated by two concerns.

The first is minority student overrepresentation within special education. As mentioned, the NRC's analysis (Heller et al., 1982) of overrepresentation reformulated the problem from one of reducing disproportionality to that of determining the conditions under which inequality constitutes treatment inequity. According to Messick (1984), disproportion signifies inequity only when children are unduly exposed to classification because they receive poor-quality regular education, are assessed invalidly for special education, or receive an ineffective special education that hinders their educational progress. The fourth phase of a treatment validity approach to identification addresses Messick's last point. This fourth phase strives to eliminate the inequity potentially associated with overrepresentation of minority students in special education by reserving judgment about the need for special education until evidence verifies that a special education program enhances learning. If, on the other hand, special education effectiveness at the individual level cannot be documented, no compelling rationale exists for assigning a learning disability label or removing the child from the classroom for instruction. So, equity is achieved when, prior to placement, evidence verifies that special education is valuable and desirable because it enhances the learning of the individual. Consequently, this fourth phase of the assessment process may be important to the achievement of equity.

The second rationale for a fourth phase of decision making within a treatment validity approach is to encourage special education accountability. In 1995, when the treatment validity model was originally proposed, the general education reform movement was (and continues to be) driven by accountability for children's achievement of high standards. And the focus on statewide assessment programs, with strong consequences to schools for student learning, was gaining momentum. In special education, however, no parallel push for accounting for student outcomes was occurring. Even today, with the amendments of the Individuals with Disabilities Education Act's requirement that all students with disabilities participate in the general education accountability system, serious questions remain about the nature and force of accountability for the achievement of students with learning disabilities. For example, it is unclear whether the content of the state assessments adequately addresses the acquisition of basic reading and mathematics competence, which constitutes the heart of the instructional programs of most students with learning disabilities (see Fuchs & Fuchs, 2000, for discussion). In addition, the many unanswered questions about accommodations on high-stakes, group-administered tests raise concerns about whether performance on these assessments validly represents outcomes for students with learning disabilities (see, for example, Koretz & Hamilton, 2000).

Alternatively, with a diagnostic special education trial, which serves to assess the added value of special education services, accountability for special education is addressed using the very learning outcomes that created concern in the first place. That is, if a student is identified for special attention because he or she manifests a dual discrepancy in reading that proves resistant to adaptations within the regular classroom, then special education accountability is demonstrated when the student's rate of reading growth improves, as revealed by the same assessment device through which treatment unresponsiveness initially was documented. In this way, the fourth phase within a treatment validity approach to identification creates important consequences associated with special education accountability: If special education is to exist (i.e., is to have students to serve), special education must demonstrate its added value on the front end, prior to placement.

Arguments Against a Fourth Phase

Despite the salience of overrepresentation and accountability issues in support of a fourth phase, two counterarguments are compelling. First, among the students who enter a fourth phase (those who manifest dual discrepancies that prove resistant to general education adaptations), those who respond favorably within the limited timeframe of a diagnostic special education run may represent false positives. At the same time, the children who fail to respond within a limited timeframe to the more intensive, individualized forms of instruction available within special education may represent the population of learners for whom special education is most necessary -- those who, in fact, have the serious underlying deficits that a disability denotes. In this way, the children who are identified via a fourth phase (those who do respond) may be the very children who should be excluded from consideration. And, in a complementary way, those who are excluded from a special education (those who do not respond) may be the appropriate candidates for special education.

The second argument against the fourth phase of the treatment validity model also concerns students who reveal serious learning problems in Phases II and III, but fail to demonstrate enhanced growth with special education and, therefore, do not meet eligibility requirements. Given a straightforward implementation of this decision-making process, these students are returned to the regular classroom without additional support when, clearly, some action is required. And both arguments inevitably lead to the conclusion that special education is most appropriate for the very children it cannot help. Alternatively, however, these difficulties with the fourth stage of the treatment validity identification model lead to a model that attempts to address these conceptual and operational problems.

A Revised Treatment Validity Model

Addressing the difficulties associated with the original model's fourth phase, while capitalizing on its potential, requires revision. In implementing the treatment validity model recently within a Nashville public school, we revised the last phase in the following way. We referred to the revised last phase of the identification model as an "extended assessment plan." Consistent with the original model, this extended assessment plan includes a diagnostic special education trial (approved by the multidisciplinary team on which the parent participates), and is designed to assess the potential contribution of special education to the student's learning. During the diagnostic trial, the use of CBM continues to determine whether special education reduces the student's dual discrepancy. No later than eight weeks into this phase, the team reconvenes with the parent to review the diagnostic CBM data. When successful student progress has been demonstrated within the diagnostic trial period, the successful intervention continues to be implemented, and an IEP is developed to continue and to enhance the program.

When differential progress is not demonstrated during the diagnostic trial period, however, the child is not simply returned to regular education without the benefit of additional intervention. Instead, assessment continues. The assessment team, in collaboration with the parent(s), reviews options and collects additional assessment information with which to describe and address the dual discrepancy. These options include, but are not limited to: (a) placing the student in general education with accommodations that teach the student how to access the general education curriculum in effective ways despite basic skill limitations; (b) continuing the diagnostic trial period for a designated period of time; (c) continuing the diagnostic trial period in a more restrictive placement that offers additional resources for effecting progress; and (d) continuing the special education diagnostic trial in another school where other special educators or alternative resources are available to address the student's needs.

In this way, an unsuccessful Phase IV diagnostic special education trial might lead to the identification of students for whom alternative curricula are appropriate, where literacy or numeracy goals are deemphasized and where the use of prosthetic devices to overcome handicaps might assume priority. Alternatively, an unsuccessful diagnostic special education trial could lead school districts to proactively consider and empirically test results associated with more intensive models of special education. Regardless of which option is selected, the student's progress continues to be monitored via CBM: When the student is returned to general education, CBM is used to problem solve on the student's behalf; when the diagnostic trial is extended, CBM is used to assess whether a data-based rationale can be mounted on behalf of special education; and when an IEP is opened, CBM is used to develop an optimal plan in response to the student's actual progress and to determine the earliest opportunity for productively exiting the student from special education.

With these modifications, the last stage of the treatment validity model provides a more flexible system of alternatives for problem solving. It avoids the problematic situation where students who fail to respond quickly within a diagnostic special education trial are matter-of-factly returned to regular education without additional support. At the same time, the last stage of the treatment validity model also sets the stage for districts to develop an array of alternative responses to students who fail to demonstrate progress -- rather than to permit the most difficult-to-teach students to languish in ineffective instructional settings. Moreover, progress is monitored continuously, and decisions are formulated about entry to special education based on these accountability data, which have the capacity to reveal special education's value. And finally, once placement decisions are made, progress is monitored continuously so that decisions about exiting special education can be formulated responsively.

CONCLUDING COMMENT

The revised treatment validity model provides four guarantees. First, special education decisions are formulated on the basis of students' lack of responsiveness to instruction, in the face of dual discrepancies that prove resistant to prereferral adaptations; second, the diagnostic process provides evidence that special education benefits the students it serves; third, a flexible problem-solving system permits consideration of alternative responses when the diagnostic special education trial reveals continued unresponsiveness; and fourth, students exit special education as soon as objective data indicate that they can make adequate progress within the confines of regular education. In these ways, a treatment validity model, which includes the proposed revision to its final stage, should simultaneously provide a stronger conceptual framework and technical basis for identifying students as having learning disabilities, while avoiding the controversy associated with disproportionate representation of students of color within special education.

Table 1

Effect Sizes Calculated Between the CBM Dual Discrepancy Group and Two
Other Poor Reader Groups

Variable                           IQ-DS             LA

Intelligence                    -1.55           -.19
Age                             -1.16           -.69
Reading                          -.10            .17
Fall Word Reading Efficiency     -.01            .24 Grade 1
                                                 .08 Grade 2
Fall Rapid Object Naming (a)      .50           -.08
Spring Phonological Awareness     .50 Grade 1   -.15 Grade 1
                                -1.39 Grade 2   -.84 Grade 2
Academic Competence              -.71           -.29
Problem Behavior (a)              .54            .21
Social Skills                    -.57           -.29

Note. Adapted from Speece and Case (2001). Positive effect sizes favor
the CBM Dual Discrepancy Group. IQ-DS = group defined by the
regression-based discrepancy between intelligence and reading
achievement; LA = group defined by low (< 90) reading achievement.

(a) Low scores are desirable; thus positive effect sizes favor the
IQ-DS or LA groups.

NOTES

This article is based on a paper prepared for the Workshop on IQ Testing and Educational Decision Making, National Research Council, National Academy of Sciences, May 11, 1995 and a previously published paper in Learning Disabilities Research and Practice.

Requests for reprints should be addressed to: Lynn S. Fuchs, Box 328 Peabody, Vanderbilt University, Nashville, TN 37203.

REFERENCES

Allinder, R. M., & Fuchs, L. S. (1994). Alternative ways of analyzing effects of a short school break on students with and without disabilities. School Psychology Quarterly, 9, 145-160.

Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155-159.

Deno, S. L. (1985). Curriculum-based measurement: The emerging alternative. Exceptional Children, 52, 219-232.

Fletcher, J. M., Francis, D. J., Shaywitze, S. E., Lyon, G. R., Foorman, B. R., Stuebing, K. K., & Shaywitz, B. A. (1998). Intelligent testing and the dual discrepancy model for children with learning disabilities. Learning Disabilities Research and Practice, 13, 186-203.

Foorman, B. R., Francis, D. J., Fletcher, J. M., Schatschneider, C., & Mehta, P. (1998). The role of instruction in learning to read: Preventing reading failure in at-risk children. Journal of Educational Psychology, 90, 37-55.

Fuchs, D., & Fuchs, L. S. (1995a). Special education can work. In J. Kaufman, J. Lloyd, D. Hallahan, & T. Astuto (Eds.), Issues in educational placements of students with behavior disorders (pp. 363378). Hillsdale, NJ: Eribaum.

Fuchs, D., & Fuchs, L. S. (1995b). What's "special" about special education? Phi Delta Kappan, 76, 522-530.

Fuchs, D., Fuchs, L. S., & Fernstrom, P. J. (1993). A conservative approach to special education reform: Mainstreaming through transenvironmental programming and curriculum-based measurement. American Educational Research Journal, 30, 149-178.

Fuchs, D., Fuchs, L. S., Mathes, P., & Simmons. D. (1997). Peer-assisted learning strategies: Making classrooms more responsible to student diversity. American Educational Research Journal, 34, 174-206.

Fuchs, L. S. (1995). Curriculum-based measurement and eligibility decision making: An emphasis on treatment validity and growth. Paper presented at the Workshop on IQ Testing and Educational Decision Making, National Research Council, Washington, DC.

Fuchs, L. S., & Deno, S. L. (1991). Paradigmatic distinctions between instructionally relevant measurement models. Exceptional Children, 57, 488-501.

Fuchs, L. S., & Fuchs, D. (1998). Treatment validity: A unifying concept for reconceptualizing the identification of learning disabilities. Learning Disabilities Research and Practice, 13, 204-219.

Fuchs, L. S., & Fuchs, D. (2000). Accountability and assessment in the 21st centry for students with learning disabilities. Philadephia: Pew Charitable Trusts.

Fuchs, L. S., Fuchs, D., & Hamlett, C. L. (1989a). Effects of alternative goal structures within curriculum-based measurement. Exceptional Children, 55, 429-438.

Fuchs, L. S., Fuchs, D., & Hamlett, C. L. (1989b). Effects of instrumental use of curriculum-based measurement to enhance instructional programs. Remedial and Special Education, 10(2), 43-52.

Fuchs, L. S., Fuchs, D., & Hamlett, C. L. (1994). Strengthening the connection between assessment and instructional planning with expert systems. Exceptional Children, 61, 138-146.

Fuchs, L. S., Fuchs, D., Hamlett, C. L., & Allinder, R. M. (1989). The reliability and validity of skills analysis within curriculumbased measurement. Diagnostique, 14, 203-221.

Fuchs, L. S., Fuchs, D., Hamlett, C. L., & Allinder, R. M. (1991a). Effects of expert system advice within curriculum-based measurement on teacher planning and student achievement in spelling. School Psychology Review, 20, 49-66.

Fuchs, L. S., Fuchs, D., Hamlett, C. L., Phillips, N. B., & Bentz, J. (1994). Classwide curriculum-based measurement: Helping general educators meet the challenge of student diversity. Exceptional Children, 60, 518-537.

Fuchs, L. S., Fuchs, D., Hamlett, C. L., Phillips, N. B., & Karns, K. (1995). General educators' specialized adaptation for students with learning disabilities. Exceptional Children, 61, 440-459.

Fuchs, L. S., Fuchs, D., Phillips, N. B., Hamlett, C. L., Karns, K., & Dutka, S. (1997). Enhancing students' helping behavior during peer-mediated instruction with conceptual mathematical explanations. Elementary School Journal, 97, 223-250.

Fuchs, L. S., Fuchs, D., Hamlett, C. L., & Stecker, P. M. (1990). The role of skills analysis in curriculum-based measurement in math. School Psychology Review, 19, 6-22.

Fuchs, L. S., Fuchs, D., Hamlett, C. L., & Stecker, P. M. (1991). Effects of curriculum-based measurement and consultation on teacher planning and student achievement in mathematics operations. American Educational Research Journal, 28, 617-641.

Fuchs, L. S., Fuchs, D., Hamlett, C. L., Thompson, A., Roberts, P. H., Kubec, P., & Stecker, P. M. (1994). Technical features of a mathematics concepts and applications curriculum-based measurement system. Diagnostique, 19(4), 23-49.

Fuchs, L. S., Fuchs, D., Hamlett, C. L., Walz, L., & Germann, G. (1993). Formative evaluation of academic progress: How much growth can we expect? School Psychology Review, 22, 27-48.

Fuchs, L. S., Fuchs, D., Karns, K., Hamlett, C. L., Katzaroff, M., & Dutka, S. (1997). Effects of task-focused goals on low-achieving students with and without learning disabilities. American Educational Research Journal, 34, 513-544.

Good, R. H. III, Simmons, D. C., & Kameenui, E.J. (2001). The importance and decision-making utility of a continuum of fluency-based indicators of foundational reading skills for third grade high-stakes outcomes. Scientific Studies of Reading, 5, 257-288.

Heller, K. A., Holtzman, W. H., & Messick, S. (Eds.). (1982). Placing children in special education: A strategy for equity. Washington, DC: National Academy Press.

Koretz, D., & Hamilton, L. (2000). Assessment of students with disabilities in Kentucky: Inclusion, student performance, and validity. Educational Evaluation and Policy Analysis, 22, 255-272.

Larry P. v. Riles. 343 F. Supp. 1306 (N.D. Cal. 1972) (preliminary injuction). Aff'd 502 F. 2d 963 (9th cir. 1974); 495 F. Supp. 926 (N.D. Cal. 1979) (decision on merits). Order modifying judgment, C-71-2270 RFP, Sept. 25, 1986.

Lyon, G. R. (1995). Research initiatives in learning disabilities: Contributions from scientists supported by the National Institute of Child Health and Human Development. Journal of Child Neurology, 10, Supplement Number 1, S120-S126.

Marshall et al. v. Georgia. U.S. District Court for the Southern District of Georgia, CV482-233, June 28, 1984; Aff'd (llth cir. no 84-8771, Oct. 29, 1985).

Marston, D. (1987-1988). The effectiveness of special education: A time-series analysis of reading performance in regular and special education settings. The Journal of Special Education, 21(4), 13-26.

Marston, D., Fuchs, L. S., & Deno, 5. L. (1986). Measuring pupil progress: A comparison of standardized achievement tests and curriculum-related measures. Diagnostique, 11, 71-90.

Martin, E., & Fuchs, D. (2000). Reconceptualizing the definition of learning disabilities. Presentation at the annual meeting of the Council for Exceptional Children, Vancouver.

Messick, S. (1984). Assessment in context: Appraising student performance in relation to instructional quality. Educational Researcher, 13(3), 3-8.

Messick, S. (1995). Validity of psychological assessment: Validation of inferences from persons' responses and performances as scientific inquiry into score meaning. American Psychologist, 50, 741-749.

Messick, S. (1989). Validity. In R. L. Linn (Ed.), Educational measurement (3rd ed., pp. 13-103). New York: MacMillan.

Parrish, T. B. (1995). What is fair? Special education and finance equity. Palo Alto, CA: American Institutes for Research.

Polloway, M. (1985). Identification and placement in mild mental retardation programs: Recommendations for professional practice. Education and Training of the Mentally Retarded, 20, 218-221.

Reschly, D. J. (1984). Beyond IQ test bias: The national academy panel's analysis of minority EMR overrepresentation. Educational Researcher, 13(3), 15-19.

Reschly, D. J. (1988). Minority MMR overrepresentation and special education reform. Exceptional Children, 54, 316-323.

Reschly, D. J. (1996). Disproportionate minority representation in general and special education programs: Patterns, issues, and alternatives. Des Moines: Iowa Department of Education.

Reschly, D.J., & Grimes, J. P. (1995). Best practices in intellectual assessment. In A. Thomas & J. Grimes (Eds.), Best practices in school psychology (Vol. III, pp. 763-773). Washington, DC: National Association of School Psychologists.

Reynolds, C. R. (1984). Critical measurement issues in learning disabilities. Journal of Special Education, 18, 451-476.

Rosenfeld, R. G. (1982). Evaluation of growth and maturation in adolescence. Pediatrics in Review, 4, 175.

Ross, R. P. (1995). Best practices in implementing intervention assistance teams. In A. Thomas & J. Grimes (Eds.), Best practices in school psychology (Vol. III, pp. 227-237). Washington, DC: National Association of School Psychologists.

S-1 v. Turlington. Preliminary injunction, U.S. District Court, Southern District of Florida, Case No. 79-8120-Civ-CA WPB, June 15, 1979. Aff'd United States Court of Appeals, 5th Circuit, Jan. 26, 1981, 635 F 2d 342 (1981). Trial on Merits, May 19-June 4, 1986.

Shepard, L. A., Smith, M. L., & Vojir, C. P. (1983). Characteristics of pupils identified as learning disabled. American Educational Research Journal, 20, 309-332.

Snow, R. E. (1984). Placing children in special education: Some comments. Educational Researcher, 13(3), 12-14.

Speece, D. L. (1990). Aptitude-treatment interactions: Bad rap or bad idea. Journal of Special Education, 24, 139-149.

Speece, D. L., & Case, L. P. (2001). Classification in context: An alternative approach to identifying early reading disability. Journal of Educational Psychology, 93, 735-749.

Speece, D. L., Molloy, D. E., & Case, L. P. (2000, Feb.). Toward validating a model of reading disability identification based on response to treatment. Presented at the annual Pacific Coast Research Conference, La Jolla, CA.

Tanner, J. M., & Davies, P.S.W. (1985). Clinical longitudinal standards for height and height velocity for North American children. Journal of Pediatrics, 107, 317.

Tateyama-Snizek, K. M. (1990). Cooperative learning: Does it improve the academic achievement of students with handicaps? Exceptional Children, 56, 426-437.

Tharp, R. G., & Gallimore, R. (1988). Rousing minds to life. New York: Cambridge University Press.

U.S. Department of Education. (1999). Twenty-first annual report to Congress on the implementation of the Individuals with Disabilities Act. Washington, DC: Government Printing Office.

Vellutino, F. R., Scanlon, D. M., Sipay, E. R., Small, S. G., Pratt, A., Chen, R., & Denckla, M. B. (1996). Cognitive profiles of difficult-to-remediate and readily remediated poor readers: Early intervention as a vehicle for distinguishing between cognitive and experiential deficits as basic causes of special reading disability. Journal of Educational Psychology, 88, 601-638.

Witt, J. C., & Gresham, F. M. (1985). Review of the Wechsler Intelligence Scale for Children-Revised. In J. Mitchell (Ed.), Ninth mental measurements yearbook (pp. 1716-1719). Lincoln, NE: Buros.

LYNN S. FUCHS Ph.D., is a professor, Vanderbilt University

DOUGLAS FUCHS, Ph.D., is professor, Vanderbilt University.

DEBORAH L. SPEECE, Ph.D., is professor, University of Maryland.