Learning Disabilities as Operationally Defined by Schools
Donald L. MacMillan, University of California, Riverside, & Gary N. Siperstein, University of Massachusetts, Boston
Learning Disabilities Summit: Building a Foundation for the Future White Papers
This paper is available in alternative formats: | Download Word | Download pdf |
Pages: | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
CHANGES IN THE IMPORTANCE OF DIFFERENTIAL DIAGNOSIS
In the 1960s and early 1970s, the assessment/classification process was a "high stakes enterprise" in the sense that the disability classification made in the case of a given child had profound consequences for that child's educational experience (MacMillan, Gresham, Bocian, & Siperstein, 1997). The classification decision whether a child was EMR or LD carried with it consequences in terms of the administrative arrangement into which a child would be served and the curriculum/services that would be provided. Figure 1 schematically represents the shift in consequences for differential diagnosis discussed in the following paragraphs. Recall that the earlier time frame in this schematic predates passage of P.L. 94-142 (Education for All Handicapped Children Act of 1975) and the requirements for a free appropriate public education (FAPE), LRE, and IEPs.
If a diagnosis of EMR was made, the question of "where," or the administrative arrangement in which the child would receive services, was essentially automatic. Robinson and Robinson (1965) described educational services for EMR students in that era, explaining that "The consensus of special educators today definitely favors special class placement for the mildly retarded" (p. 466). So, a specific placement (self-contained special class) was linked to the diagnosis of EMR. Moreover, an alternative curriculum--that is, a functional curriculum--that differed markedly from the general education curriculum was taught to EMR students. It is important to note that special education for virtually all disabilities except mental retardation consisted of modification in how a child was taught. In the case of mental retardation, however, special education consisted of modifications in both how and what the child was taught. The curriculum for EMR students typically emphasized the promotion of prevocational skills in the elementary grades along with social and interpersonal skills and functional academics. In the secondary programs vocational training received considerable attention.
Diagnosis of LD, on the other hand, resulted most commonly in service delivery in resource settings, and yet what was taught was the standard general education curriculum. Special education for LD consisted of assistance provided by a resource teacher in order to enable the LD child to succeed in the general education curriculum. Unlike what the EMR student at that time was taught, there was seldom any consideration of an alternative curriculum for LD students.
School personnel responsible for diagnosis in the 1960s and 1970s did agonize over what "type" of child they were considering. The diagnostic decision they would make had profound consequences for the child under consideration. One diagnosis conveyed the belief that the general curriculum was appropriate (i.e., LD), and the other diagnosis (EMR) reflected the belief that an alternative curriculum was needed. To convey the significance of that decision, one of the major and most telling criticisms of the "EMR program" was that the decision was irreversible--once the child was put into the alternative curriculum for any period, the possibility of returning that child to the general curriculum was effectively blocked.
Figure 1. Importance of differential diagnosis at two points in time.
Passage of P.L. 94-142 almost imperceptibly reduced the stakes in making differential diagnoses; it took the pressure off public school personnel in their classification efforts. We believe this is why today school personnel tell us that they know the child is mentally retarded but classify him or her as LD because there is no upside to calling a child mentally retarded. Apparently they don't believe there is a downside to making an "erroneous" classification. Why? While a diagnosis of one of the disabling conditions recognized under IDEA establishes eligibility for special education and related services, other provisions of the law call for the IEP to be individually negotiated for a child. No longer does the diagnostic category under which a child's eligibility is established carry with it any consequences for the curriculum; those are negotiated. Similarly, the LRE provision of the law precludes placement of all children in a given category in a given administrative arrangement (e.g., a special class). Placement also must be individually determined on a case-by-case basis. One consequence is that a very low cognitive child being diagnosed as LD could conceivably get a functional curriculum delivered in a special day class if that is the result of the IEP process considering LRE on an individual basis. In other words, the diagnostic category is not determinative of the placement or the treatment a child will experience, making diagnosis no longer a high-stakes venture.
VARIABILITY WITHIN THE SI LD POPULATION
A teacher hired to teach LD students is likely to encounter a very heterogeneous group of students identified as LD by the three-stage process described. However, the degree of heterogeneity and the nature of the LD students they encounter will vary as a function of the state in which they live and the school building in which they are employed. Simply looking at the prevalence rates for LD in the public schools across states reveals considerable variability in the percentage of children identified as LD. The Twentieth Annual Report to Congress on the Implementation of the Individuals With Disabilities Education Act (U.S. Department of Education, 1998) reported that 3.28% of the children in Georgia and 3.40% of those in Kentucky are classified as LD, while 9.62% in Rhode Island and 9.90% in Massachusetts are so classified. While the variability in LD rates across states is less pronounced than is found for MR or ED (Hallahan, Keller, & Ball, 1986), it is far greater than one would expect for a clinical entity reliably assessed. Table 2 shows the percentage for all states. In all probability, the children served as LD in Georgia and Kentucky do differ in important ways from those served in Rhode Island and Massachusetts. The key here is that there is variability in LD prevalence across states attributable to a number of factors such as different criteria for eligibility and different perspectives on classifying children as MMR. It is also important to keep in mind that data aggregated at the state level tend to obscure variations within states at the district level.
In addition to differences in the rate of LD identification, it is important to factor in the point raised earlier that the nature of the SI LD child varies for urban and suburban school districts. Evidence has been presented herein clearly documenting that LD students in urban settings represent very different learning problems that do those in suburban districts. They score lower on measures of intelligence and require that we consider sociocultural factors as contributors, if not causes, of their learning difficulties. They frequently come from families living in poverty. Thus, factors such as low birthweight, exposure to lead, exposure to prenatal substance abuse, living in crowded residences, being raised by parents with less formal education, and various other conditions associated in the past with the condition of MMR are clearly salient to the current urban LD population.
Variability is also evident in the breakdown of SI LD by gender. For example, national projection data from the 1997 Office of Civil Rights survey reveals that 68.35 percent of LD are male, with 31.65 percent being female. While this gender disproportion has been widely accepted in most of the mild disability categories, it is not as apparent when surveys employ RI LD samples. For example, in the Connecticut Longitudinal Study (Shaywitz, Fletcher, Holahan, & Shaywitz, 1992; Shaywitz, Shaywitz, Fletcher, & Escobar, 1990), a two-stage probability sample used RI LD criteria of two types: (a) an IQ-based regressed discrepancy of 1.5 SD and (b) low reading achievement (using an age-adjusted standard score #25% on the Reading Cluster of the Woodcock-Johnson). Neither the discrepancy criteria nor the low reading achievement criterion yielded differential rates by gender. It seems reasonable to speculate that the gender difference so apparent in SI LD populations arises from differential expectations and/or concomitant externalizing behaviors accompanying reading difficulties for males and factors associated with teacher referral behavior. MacMillan, Gresham, Lopez, & Bocian (1996) found gender differences in referral rates of teachers for White and Black students, but not for Hispanic students.
When rates of SI LD by racial or ethnic group are aggregated at the national level and examined, the issue of overrepresentation of Black students found in the mental retardation category is not found for LD (MacMillan & Reschly, 1998). However, it is noteworthy that the percentages of White, Black, and Hispanic children served in the LD category have also increased dramatically. Yet, because the increase has been "proportionate" (at approximately the same rate for all three ethnic groups), it is not viewed as a problem. In 1978 the projected national figures showed that 2.23% of Black students were being served in LD; in the 1997 survey 6.15% of Black students were being served in LD. For White students the change in the same period was from 2.23% to 5.53%. Using the 1997 survey data, there are considerable variations across ethnic groups. For example, the percentage of Asian/Pacific Island students served as LD is only 1.90% in contrast to the following percentages for other ethnic groups: American Indian/Alaskan, 6.41%; Hispanic, 5.99%; Black, 6.15%, and White, 5.53%. Hence, comparing the rates to that of White students suggests no overrepresentation for Black and Hispanic students; however, if one uses the rate for Asian/Pacific Island students, there are considerably higher rates for all other ethnic groups.
ISSUES RAISED ABOUT THE CURRENT PROCESS
Inherent in the current process is the apparent belief that differentiation of the broad band of children presenting with low achievement is essential in order to match treatment to individual need. Federal and state regulations promote such distinctions, providing criteria to be used in the identification of, and distinction between, children with LD, mental retardation, speech and language impairments, and emotional disturbance. On the other hand, evidence has been presented herein suggesting that the public schools give lip service to this process but have increasingly used the LD category in a manner quite different from that suggested in these regulations--as a cross-categorical designation. As we have argued elsewhere (MacMillan, Gresham, Siperstein, & Bocian, 1996; MacMillan et al., 1997, 1998), the time has come to consider the limitations of current policy and to examine the consequences of current practices. Let us turn to the issues we believe need to be addressed in the identification process and then turn to the consequences that we anticipate will result from the failure to resolve these issues.
ELIGIBILITY USING A ONE-TIME-ONLY ASSESSMENT
The current process establishes a child as eligible based on assessments conducted at one point in time (i.e., after referral and failure to respond to prereferral interventions). We see this as problematic in two ways. First, there is a problem with assessing a child, let's say in third grade, and finding a psychometric profile permitting eligibility as LD. One of the benchmarks currently employed is low academic achievement (usually in reading) discrepant from expected level of achievement (based on IQ). Assessing current level of functioning, however, is unable to inform us as to "why" the child's academic performance is low. Among the low scorers in third grade are some children who, in fact, do have problems processing information despite good instruction in first and second grade. Others who score low have no serious problems processing information yet score low because they have not been instructed well in first and/or second grade. The current process, which uses a "one-time-only" assessment, fails to differentiate between these two possibilities. Hence, it is difficult to refute the contention of those who argue that a child identified as LD in third grade is not simply a failure of general education. This issue is independent of the debate over discrepant versus nondiscrepant low achievement, and rather derives from the provision regarding exclusion in the reauthorization of IDEA 1997 to include children who have not had the opportunity to learn--they are not to be identified as having a learning disability.
Given the previous data provided on children in urban districts identified as LD, this distinction becomes even more important. We know that inner-city schools are staffed more often by new teachers (currently in some districts, the majority of which lack credentials) and teachers who are unable to secure transfers to more affluent schools. Urban districts often have older school buildings with poorer equipment, fewer amenities such as adequate computer facilities, and student bodies often coming from more disadvantaged backgrounds. The likelihood of confusing "disability" with "disadvantage" in such settings is great, yet the current process provides little direction for distinguishing between the two. How is one to establish that the low reading achievement exhibited by a child in third grade is not due to "a lack of opportunity to learn"? Currently, this requirement is met in a cavalier fashion--at best by attesting to the prereferral intervention efforts. However, prereferral interventions are a nonspecific, often very weak intervention seldom targeted to the problem prompting referral and seldom, if ever, implemented with fidelity.
Elsewhere in this volume (see work of Torgesen) is presented converging evidence that in the area of reading, there are validated procedures which if implemented are successful in promoting significant improvement in reading skills in children presenting with reading disabilities (see, for example, Foorman, Francis, Winikates, Mehta, Schatschneider, & Fletcher, 1997; Torgesen, Alexander, Wagner, Rashotte, Voeller, & Conway, 2001). Less systematic work has been done in the area of mathematics. However, the benefit of structured instruction in mathematics has been demonstrated to reduce the gap evident upon entry to school for low socioeconomic status (SES) students in comparison with high SES students (Case, Griffin, & Kelly, 1999; Griffin & Case, 2000). The Number Knowledge Test (see Griffin & Case, 2000, Appendix A) provides a quick reliable assessment of the child's developmental stage in elementary math. Moreover, 6-year-old children who attended school in low-income areas and were ready to enter first grade performed at the 3- to 4-year-old level on the Number Knowledge Test. A program called Rightstart was developed (the name changed to Number Worlds later) and appears successful for many children in promoting number facility in initially low-scoring, low SES students, enabling them to perform at a level comparable to that of high SES children. Empirically validated procedures are in place that could be used to ensure "the opportunity to learn" reading and math. Before using a one-time-only assessment to establish the presence of a disability, it seems reasonable to ask that validated interventions implemented with integrity are provided and that the child's responsiveness to these interventions is examined prior to labeling.
A second concern with the current process derives from the mere fact that whether a child will exhibit the requisite "severe discrepancy" is, in part, a function of the age/grade level at which the assessment occurs. Requiring a discrepancy between achievement and intelligence has been characterized as a "wait and fail" method of classification because several years of schooling are usually required in order to obtain a sufficiently large discrepancy to qualify as LD. Failure to target reading problems early, as one waits for the discrepancy to be achieved, has been a concern of many (e.g., Fletcher & Foorman, 1994; Speece & Case, submitted). Delays in targeting treatment permit the child to flounder, experience additional failure, and reduce the probability that treatments will be effective once they are finally introduced.
An alternative to using one-time assessments to document a severe discrepancy has been described by Lynn and Doug Fuchs (Fuchs, 1995; Fuchs & Fuchs, 1998). Their approach, based on tenets of treatment validity, uses curriculum-based measurement (CBM) that provides for repeated measurement and is sensitive to change or growth. The approach contrasts the entry-level skills to those of classmates (level) and rate of progress over time (slope) in comparison to classmates receiving the same quality of instruction delivered by the same teacher. Fuchs and Fuchs (1998) used the term "dual discrepancy" to capture the criteria employed by their approach to LD eligibility: both level and rate of progress have to differ from that of peers to be eligible. In a recently reported study, Speece and Case (submitted) compared children identified as dually discrepant (level and slope using CBM measures) with others exhibiting an IQ-reading achievement discrepancy. The dually discrepant group was lower in IQ and younger than the IQ-reading achievement discrepant group; however, the two groups did not differ on reading, phonological awareness, social skills, or SES measures. Interestingly, neither approach identified a gender disproportion, while the dual discrepancy approach did identify a larger proportion of younger children. Obviously, we do not know which approach identified the "real LD" cases; however, the use of repeated measurements tracking growth strikes us as an attractive feature. The downside of this approach is its labor-intensiveness, thereby reducing the likelihood of it being adopted widely in the public schools.
Going beyond the "opportunity to learn," there is the possibility that really poor instructional programs may not only prevent optimal development, but may actually have a deleterious effect on the child. Consider the work of Kellam, Ling, Merisca, Brown, and Ialongo (1998), albeit dealing primarily with aggressive behavior, as it bears on the influences of classroom context on the course of behavior. In this study, involving 19 public elementary schools, schools and teachers were randomly assigned to intervention or control conditions and children within each school were assigned sequentially to classrooms from alphabetized lists. Despite these efforts to randomize, "...classroom differences in levels of aggressive behavior emerged as early as the first quarter of first grade, suggesting that the very origins of variation in classroom aggression levels came from the classroom teacher and/or the mix of students and the teachers" (p. 181). That is, classrooms (as opposed to poverty, school building) were associated with levels of aggression, and the levels of aggression evident in classrooms were bimodally distributed--either high or low levels of aggression being evident. When these students were followed into middle school, aggressive males from these "chaotic classrooms" were at increased risk for serious conduct disorders. The authors summarized the importance of first-grade classroom contextual factors as follows:
We suggest the following hypothetical model to better understand the classroom effect on the more aggressive males. The skills of the teachers in highly aggressive, disruptive classrooms were not sufficient to promote an integrative prosocial classroom social system. Effective classroom behavior management appears to be essential in the socialization of young children, and for boys whose initial behavior response is aggressive, disruptive, the lack of providing teachers [with] sufficient background training is critical. Teacher training does not provide effective methods and experience in classroom management. This, coupled with the lack of staff support for the classroom, places such children at greater risk of later conduct disorder and related academic problems. (p. 182)
In their conclusions, Kellam et al. stated, "The implications of these findings are that the first grade classroom has a critical impact on the developmental course of aggressive behavior for the boys whose initial responses to classrooms are aggressive, disruptive" (p. 184).
What is the likelihood that a first-grade teacher who has difficulty controlling behavior also has difficulty teaching reading? If the students do not attend to instruction, that instruction is unlikely to have an effect. Moreover, if the inability to control behavior is an indicator of a poor or inexperienced teacher, then we would likely see similar weaknesses in the instructional program provided to students in the class. Where do we tend to find more inexperienced or ineffective teachers? We submit it is in the same schools where we find "at risk" students because of the presence of poverty, second-language issues, deteriorating classrooms, higher mobility rates for students, and so forth. When one conducts a one-time assessment in third grade and finds a discrepancy between aptitude and achievement, how risky is a conclusion that it reflects a within-child problem? When an IEP team excludes inadequate prior instruction as a contributor to the child's learning deficiency, on what basis is that determination made? The need for consideration of such contextual factors is paramount, but heretofore too often ignored in the LD field (Keogh & Speece, 1996).
In contrast to the field's willingness to consider environmental factors to explain or understand mild mental retardation, the field assumes that achievement deficits exhibited by students labeled LD are due to within-child, neurological factors. The denial evident in the exclusive focus on intrinsic factors will not serve the field well. The population of students served as LD includes substantial numbers of children whose academic performance deficits are clearly linked to environmental influences and contextual factors, traditionally excluded from consideration by the LD field.
