Skip over navigation

Journal Issue: Special Education for Students with Disabilities Volume 6 Number 1 Spring 1996

Learning Disabilities
G. Reid Lyon


The influence of advocacy has, in turn, contributed to a substantial proliferation in the number of children who have been identified with learning disabilities relative to other handicapping conditions (see Figure 1). Clearly, the prevalence of LD identification has increased dramatically.

The "real" prevalence of learning disabilities is subject to much dispute because of the lack of an agreed-upon definition of LD and objective diagnostic criteria.4,8,12 Some have argued that the currently recognized 5% prevalence rate is excessive and is based on vague definitions, leading to inaccurate identification. On the other hand, research efforts to identify objective early indicators of LD in basic reading skills have concluded that virtually all children scoring below the 25th percentile on standardized reading tests can meet the criteria for having a reading disorder.12 While less is known about LD in written expression, researchers estimate its true prevalence at between 8% and 15% of the school population.13 Research also indicates that approximately 6% of the school population has difficulties in mathematics which cannot be attributed to low intelligence, sensory deficits, or economic deprivation.14

Increase in Identification

The substantial increase in the identification of children with learning disabilities shown in Figure 1 has led many to question the validity and reliability of LD as a diagnostic category or its "realness" as a handicapping condition.15 In fact, it appears likely that there are both sound and unsound reasons for the increase, as is discussed later.

It should be made clear that difficulties in the identification of children with learning disabilities do not make the disabilities any less "real" to the student who cannot learn to read, write, or understand mathematics despite good intelligence, an adequate opportunity to learn, and ostensibly good teaching. However, such an anecdotal understanding of learning disability and its prevalence seems inadequate now, given the increase in diagnoses of LD, the consequences of learning failure in children, and the tremendous financial resources that are applied to the identification and teaching of children with learning disabilities. Given what is at stake, it is critical that the construct of learning disability and procedures for identifying children and adults with LD be valid and accepted by the scientific and clinical communities.

The question remains, however, of how to go about increasing the ability to identify individuals with LD accurately. Valid prevalence estimates depend upon a set of criteria for identification that are clear, observable, measurable, and agreed upon.

The Discrepancy Standard

There is currently no universally accepted test, test battery, or standard for identifying children with LD. While a discrepancy between intelligence quotient (IQ) and achievement has been a widely accepted criterion for the identification of LD and still serves as the driving clinical force in the diagnosis of LD, there is considerable variation in how the discrepancy is derived and quantified.9,16 Federal regulations and extant clinical criteria17 do not specify particular formulas or numerical values to assess discrepancy objectively. The effect of this lack of specification on both clinical and research practices is substantial. From a clinical standpoint, a child can be identified as having a learning disability in one school district but not in a neighboring district because of differences in the measure of discrepancy used. From a research perspective, different approaches to the discrepancy measurement lead to substantially different sample characteristics and different prevalence estimates, which undermine the ability to replicate and generalize findings.5,6,8,9

For the individual child, use of the discrepancy standard clearly promotes a wait-to-fail policy because a significant discrepancy between IQ and achievement generally cannot be detected until about age eight or nine. In fact, most school districts do not identify children with learning disabilities until a child is reading well below grade level, generally in third or fourth grade.18 By this time the child has already experienced at least a few years of school failure and probably has experienced the common attendant problems of low self-esteem, diminished motivation, and inadequate acquisition of the academic material covered by his classmates during the previous few years.

It is clear that the longer children with learning disabilities, at any level of severity, go without identification and intervention, the more difficult the task of remediation becomes and the harder it is for the children to respond. Specifically, the data strongly suggest that children at risk for reading failure should be identified before the age of nine if successful intervention results are to be anticipated.13 For example, a longitudinal investigation of 407 students19 found that 74% of the children whose disability in reading was first identified at nine years of age or older continued to read in the lowest quintile throughout their middle and high school years. In addition, the longer children, at all severity levels, are faced with failure in reading in the classroom setting, the greater the probability that comorbid learning and behavioral difficulties will arise, further complicating the remediation task.

Developing a Diagnostic Standard

If current definitions of learning disability are not useful and if the discrepancy standard is a poor one, why have schools not adopted other means of defining and identifying LD? There are a number of conceptual and methodological barriers to the accurate identification of learning disabilities, and these impediments lead to confusion about definitions, diagnostic issues, and rising prevalence rates.

Multidisciplinary Nature of the Field

Opinions about what constitutes a learning disability vary6,10 in part because LD is the concern of many disciplines and professions, including education, psychology, neurology, neuropsychology, optometry, psychiatry, and speech and language pathology, to name a few. Each of these disciplines has traditionally focused on different aspects of the child or adult with learning disability, so divergent ideas and contentious disagreements exist about the importance of etiology, diagnostic methods, intervention methods, and professional roles and responsibilities.10 It is not surprising that so many children are identified because each professional may view the child through his or her own idiosyncratic clinical lens. For example, optometrists may identify a child as having a learning disability if the youngster displays difficulties in visual tracking. Speech and language pathologists, on the other hand, become concerned if the child's vocabulary and syntactic development are not commensurate with expectations. Educators become concerned primarily when development in reading, writing, and mathematics is deficient.

Lack of Specific Identification Criteria

Probably the most significant and persistent problem in the field is the lack of a precise definition and a theoretically based classification system that would allow (1) the identification of different types of learning disabilities and (2) a means of recognizing distinctions and interrelationships between types of learning disabilities and other learning disorders such as mental retardation, attention deficit disorder, speech and language difficulties, and general academic underachievement.20 At present, the field continues to construct and use vague and ambiguous definitions that rely heavily on the exclusion of alternative diagnoses, such as the IDEA definition shown in Box 1.

Overly Broad Label

Some observers argue that the term "learning disability" is too broad to be of any diagnostic value. Stanovich,16 a leading proponent of this view, proposes that the general term learning disabilities be abandoned and that definitional and research efforts focus on the specific types of disabilities that are now identified in ambiguous terms.

As noted earlier, the generic term learning disabilities encompasses disabilities in seven categories: (1) listening, (2) speaking, (3) basic reading skills, (4) reading comprehension, (5) written expression, (6) mathematical calculation, and (7) mathematical reasoning. Given the complexity and heterogeneity of each of these disabilities, it seems unrealistic to expect that any definitional clarity can be achieved by grouping them together under one label. To do so only obscures the critical features of each disability and makes research findings difficult to interpret.

Definitions of specific learning disabilities can be more easily and successfully operationalized than generic definitions, as the research on disability in basic reading skills shows.8 To establish valid prevalence estimates for the number of individuals with learning disabilities, the first step should be to establish explicit diagnostic criteria for each of the seven specific disability domains. At present, the greatest progress toward this goal has been in the area of disability in basic reading skills.8

LD as a Sociological Phenomenon

The simplest explanation for the increasing numbers of children identified with learning disabilities and for the difficulty in understanding and defining LD is that "LD" is not a distinct disability, but an invented category created for social purposes. Some argue that the majority of students identified as having learning disabilities are not intrinsically disabled but have learning problems because of poor teaching, lack of educational opportunity, or limited educational resources.15 In addition, because the label of LD is not a stigmatizing one, parents and teachers may be more comfortable with a diagnosis of LD than with labels such as slow learner, minimal brain dysfunction, or perceptual handicap. A diagnosis of LD does not imply low intelligence, emotional or behavioral difficulties, sensory handicaps, or cultural disadvantage. Thus, more positive outcomes are expected for children with learning disabilities than for those with mental retardation or emotional disturbance.

Reasons for Increase in Identification of LD

As pointed out, the substantial increase in the identification of LD, as shown in Figure 1, has caused many researchers to question the validity of the data. No doubt, the failure to develop an agreed-upon, objective, operational definition of learning disability gives credence to the concern about the validity of the identification process. Thus, it seems reasonable to assume that at least some of the increase in prevalence can be linked to conceptual, methodological, social, and political factors that spuriously inflate the identification of children with learning disabilities.5 However, despite the conceptual and methodological shortcomings that have plagued the field with respect to definition and identification practices, there exist a number of possibly sound reasons that could account for an increase in the number of children identified with LD.

Some Sound Reasons

As knowledge about learning disabilities grows, some academic difficulties not previously recognized as LD can be identified as such. Greater knowledge also affects the behavior and practices of teachers and parents. Sound reasons for the increase in identification rates are described and discussed in the sections that follow.


  • Better Research. Research in the past decade measures underachievement in reading as it occurs naturally in large population-based samples12,13 rather than as identified by schools, which use widely varying criteria. In addition, much of this new research is longitudinal and has been replicated, providing the necessary foundation for epidemiological studies.2,12,13,19,21-24 Finally, many of these studies have been specific to LD in reading, rather than LD in general, allowing greater precision.


  • Broader Definitions. Prevalence is directly linked to definition. LD in reading has been defined in recent research as significant difficulties in reading single words accurately and fluently, in combination with deficits in phonological awareness.8 Using this definition and stronger longitudinal research methods outlined above, the prevalence for reading disability alone has increased from estimates of less than 5% in l976 to approximately 17% in l994.12

Phonological awareness is a critical attribute in learning to read, and children who lack this awareness can be identified in late kindergarten and early first grade. Typical diagnostic questions for kindergartners or first graders involve rhyming skills (for example, "Tell me three words that rhyme with 'cat'") and phoneme deletion skills (for example, "Say 'cat' without the /t/ sound"). The majority of children pick up phonological awareness skills easily by six to seven years of age, but a large minority of children (about 17%) have significant difficulty with these skills and will have great difficulty learning to read, regardless of their intelligence, unless these skills are acquired.


  • Identification of LD in Girls. A substantial portion of this increase can be attributed to the fact that females have been found to manifest reading disabilities at rates equal to males, in contrast to previous reports that males with reading disabilities outnumbered females with reading disabilities at a ratio of four to one.25 This finding necessarily increases the prevalence rate.


  • Increased Awareness. Information disseminated in the past decade, particularly concerning the characteristics of reading disability, has increased the number of children referred for assessment of a learning disability.6


  • Understanding of the Impact. There has been an increase in the recognition that even "mild" deficits in reading skills are likely to portend significant difficulties in academic learning and are, therefore, worthy of early identification, diagnosis, and intervention.26,27

Some Unsound Reasons

There is no shortage of horror stories about the misidentification of LD and reports that the category serves as a "catch all" for any youngster who is not meeting the expectations of parents and teachers. Are there legitimate reasons for these criticisms? The answer appears to be yes. Examples are described and discussed in the sections that follow.


  • Ambiguous Definitions. The ambiguity inherent in the general definitions of LD (see Box 1) leaves the identification process open for wide interpretation and misinterpretation. Flexible identification decision making allows some children to be identified as having learning disabilities when they do not, while others with learning disabilities may be overlooked.5 This latitude can be manipulated to increase prevalence rates in response to financial incentives (for example, to qualify for increased state funding), to decrease prevalence rates in response to political movements (for example, inclusion), or to abandon programs that appear too costly.28


  • Social and Political Factors. Social and political factors also contribute to the inflation of prevalence rates for learning disabilities. In 1976–77, the first year of full implementation of Public Law 94-142, 2.16% of all school children were served in programs for children with mental retardation (MR) and 1.80% in programs for children with learning disabilities (Figure 1). By the 1992–93 school year, placements for children with MR had decreased to 1.1% while placements for children with LD had increased to 5.4% of the total school population (Figure 1). While these reversed trends mask substantial variations among states, the dramatic changes in identification rates of the two types of disability suggest that attempts to apply less stigmatizing labels may be influencing the identification process.


  • Number of Professional Specialties Involved. The large number of professional specialties involved in the identification process provides fertile ground for the overidentification of LD because each specialty brings its own set of diagnostic assumptions, theories, and measures to the assessment task. Inconsistent identification practices allow prevalence rates to escalate. This is a significant problem when there are financial incentives to encourage identification (see the article by Parrish and Chambers in this journal issue). Both market and legal forces can stimulate the development of new professional specialties (such as language/learning disorder specialist) whose members have financial incentives to diagnose students with learning disabilities, which the specialists will often be employed to treat. Although it may be uncomfortable to mention these factors, they exist and play some role in the increase of prevalence of LD. At the same time, the majority of professionals serving children with learning disabilities appear well intentioned and well informed.


  • Inadequate Preparation of Teachers. Unfortunately, a major factor contributing to invalid prevalence estimates may be the inadequate preparation of teachers by colleges of education. Recent studies have found that a majority of regular classroom teachers feel that they are not prepared to address individual differences in learning abilities within classroom settings.29 Even more alarming, research suggests that special educators themselves do not possess sufficient content knowledge to address the language and reading needs of children with learning disabilities.30 Without adequate preparation, teachers have a tendency to overrefer children for specialized assistance because they feel ill-equipped to provide the necessary services.31
Interpreting Prevalence Rates

The prevalence of learning disabilities is completely dependent upon the definition used. In most areas, the identification of LD is based largely upon the discrepancy standard and, thus, provides a count of the number of older elementary students (third grade and above) who are achieving significantly below expectations based on IQ. This is, at best, an incomplete definition of LD and one that, for the majority of students with learning disabilities, is based upon an invalid criterion, namely, the discrepancy standard.

Clearly, current definitions allow both Overidentification and underidentification of LD. Depending upon the magnitude of financial incentives and upon unrelated factors (for example, class size, goals for increasing test scores) that often shape the decisions of classroom teachers to refer students with special needs, an individual school district may drastically overidentify or underidentify students with learning disabilities. Therefore, local or national statistics on identification rates for students with LD must be interpreted with caution.