Handbook of Learning Disabilities

3

Classification and Definition
of Learning Disabilities:

An Integrative Perspective

Jack M. Fletcher
Robin D. Morris

G. Reid Lyon

This chapter addresses research on the clas- al., 2001). In 1977, recommendations for
sification, definition, and identification of operationalizing the federal definition of LD
learning disabilities (LD), and implications were provide to states after passage of Pub-
for public policy. For the past 20 years, we lic Law (PL) 94-142 to help identify chil-
have been addressing issues related to the dren in this category of special education
classification and definition of LD (Fletcher, (U.S. Office of Education, 1977). In these
Lyon, et al., 2002; Fletcher & Morris, regulations, LD was defined as a heteroge-
1986; Lyon et al., 2001; Morris, 1988; neous group of disorders with a common
Morris, Satz, & Blashfield, 1981). We have marker of intraindividual variability, (i.e.,
attempted to identify classification as a cen- “unexpectedness”), representing a discrep-
tral issue in LD research, showing that the ancy between IQ and achievement. Unex-
results of any given study depend greatly on pectedness was also indicated by exclusion-
the underlying classification of LD. The ary criteria, such as sensory disorders,
classification model chosen leads to defini- socioeconomic disadvantage, inadequate in-
tions of LD and related disorders that, in struction, and emotional-behavioral disor-
turn, influence the methods used for its ders that presumably lead to “expected” un-
identification. How children are identified derachievement.
as having LD has significant influence on
the results of any study. Implementation of this model and its fo-
cus on intraindividual differences in public
Historically, LD has existed as a disorder policy have led to an industry that domi-
that was difficult to define. Implicit classifi- nates identification procedures in schools.
cations viewed LD as “unexpected” under- This industry develops IQ and achievement
achievement. The primary approach to tests, produces research on the best way to
identification involved a search for intrain- measure discrepancy, and trains a large
dividual variability as a marker for the “un- cadre of personnel who give these tests and
expectedness” of LD, along with an empha- help ensure compliance with procedural
sis on the exclusion of other causes of guidelines adopted by states. As states vary
underachievement that would be “expect- considerably in how the federal definition is
ed” to produce underachievement (Lyon et operationalized, and schools in how identi-

30

Classification and Definition of LD 31

fication methods are implemented and in- mon form of LD, how can it be described as
terpreted, there is substantial variability in “unexpected?”
which students are served in special educa-
tion as learning disabled across schools, dis- Intraindividual Differences versus
tricts, and states (MacMillan & Siperstein, Problem-Solving Models
2002; Mercer, Jordan, Allsop, & Mercer,
1996). In response to these findings, two models
have emerged, both of which are represent-
Classification research over the past 10 to ed as competing policy recommendations
15 years has provided little evidence that IQ for LD and other high-incidence disorders
discrepancy demarcates a specific type of identified in the Individuals with Disabilities
LD that differs from other forms of under- Education Act (IDEA). The first involves in-
achievement (Fletcher, Lyon, et al., 2002). dividual differences and focuses on within
This research has also questioned the classi- child ability discrepancies as the basis for
fication validity of most proposed exclu- LD (Kavale & Forness, 2000). The second,
sionary criteria, noting little evidence that commonly referred to as the problem-solv-
children with “expected” forms of achieve- ing model (Reschly, Tilly, & Grimes, 1999),
ment differ from those with “unexpected” is an outcomes-oriented approach in which
underachievement beyond the identification the child’s response to instruction is para-
criteria (Lyon et al., 2001). Although the mount. The former is a child attribute mod-
case against IQ discrepancy and exclusion is el focused toward organismic hypotheses re-
strongest for word-level reading disabilities, garding the nature of LD, whereas the latter
enough research has been completed on is more oriented toward the context in
the underlying psychometric model to cast which the child learns, focused as it is on in-
doubt on applications to other forms of LD struction. However, from a classification
in reading, math, and written expression perspective, the two models are more simi-
(Stuebing et al., 2002). lar than different, as we describe herein.
Both models are best conceptualized as di-
Other research, especially in reading, has mensional, retain the concept of “unexpect-
shown that LD appears dimensional, not edness,” are based on the notion of discrep-
categorical, and has not been able to pro- ancy, do not rely on policy-based special
duce markers that qualitatively distinguish education categories, focus on specific acad-
different forms of LD from other forms emic behaviors, and have as a goal the de-
of underachievement (Shaywitz, Escobar, velopment of effective interventions. More-
Shaywitz, Fletcher, & Makuch, 1992). In over, although the models identify children
word-level reading disabilities, for example, as LD based on different attributes, the
a major determinant of reading ability is measurement issues are virtually identical.
clearly phonological processing, which Both involve ability–expectancy discrepan-
distinguishes word-level reading disability cies and rely on individual differences for
(RD) from other forms of LD and from typ- treatment implementation. The next section
ically achieving readers, but only on a quan- describes each model in turn. We then high-
titative basis (Liberman, Shankweiler, & light the similarities and differences from a
Liberman, 1989; Share & Stanovich, 1995; classification perspective and provide three
Vellutino, Scanlon, & Fletcher, 2002). examples derived largely from the intraindi-
Greater severity of phonological processing vidual differences model (IQ discrepancy,
deficits produces more severe reading diffi- subtypes, aptitude by treatment interven-
culties, but strengths in phonological pro- tions), to illustrate the convergence of the
cessing also produce better reading. Normal two models.
variability on a continuum is also consistent
with genetic research on word-level RD, Intraindividual Differences Model
where RD is strongly heritable (Grigorenko,
2001; Olson, Forsberg, Gayan, & DeFries, The essence of this model was clearly speci-
1999). However, the same genetic suscepti- fied in a recent consensus paper from 10
bilities that lead to poor reading also ac- major groups interested in LD organized by
count for proficiency in reading (Gilger,
2002). Given the accumulation of knowl-
edge about word-level RD, the most com-

32 FOUNDATIONS AND CURRENT PERSPECTIVES

the National Center for Learning Disabili- place with new and presumably improved
ties (NCLD) and the Office of Special Edu- norm-referenced measures. In addition, this
cation Programs (2002). This paper notes trend leaves unanswered the questions of
that “while IQ tests do not measure or pre- what to do with children who do not
dict a student’s response to instruction, achieve adequately who have relatively flat
measures of neuropsychological functioning test profiles and, more important, how such
and information processing could be includ- approaches lead to better outcomes for chil-
ed in evaluation protocols in ways that doc- dren with LD. Of particular concern is the
ument the areas of strength and vulnerabili- tendency of the intraindividual differences
ty needed to make informed decisions about model to focus on behaviors that are not di-
eligibility for services, or more importantly, rectly related to intervention, such as pro-
what services are needed. An essential char- cessing skills (Torgesen, 2002). Both issues
acteristic of SLD is failure to achieve at a are often addressed by attempts to define
level of expected performance based upon subtypes of LD based on the hypothesis that
the student’s other abilities” (p. 18). more homogeneous groupings lead to im-
proved outcomes through more targeted in-
This statement clearly highlights the role terventions. But even here, interactions be-
of intraindividual differences as a marker yond the primary area of difficulty (reading,
for discrepancy and unexpected under- math) are hard to identify.
achievement. It also highlights the limita-
tions of IQ–achievement discrepancy as a Problem-Solving Model
marker for LD, largely on the basis of lack
of relationships with intervention outcomes. The second model is put forth as a marked
As opposed to a single marker such as IQ departure from historical conceptions of LD
discrepancy, unexpectedness is operational- and an alternative to the intraindividual dif-
ized as unevenness in development. The ferences model. Referred to as a problem-
child with LD has strengths in many areas solving model, it is based on the view that
but weaknesses in some core attributes that what is paramount for LD is how to treat it.
lead to underachievement. The LD is unex- Classifications, intraindividual differences,
pected as the weaknesses lead to difficulties and subtypes are all notions that have not
with achievement and adaptive functions, proved beneficial for intervention and are
but not all areas of adaptation. Building on therefore not useful (Rechsly & Tilly, 1999).
reading research, proponents of this view These notions are viewed as outgrowths of
call for better classifications that more organismic, “medical” models that require
clearly delineate the different profiles asso- knowledge of the cause in order to affect a
ciated with LD, help delineate different treatment. Thus, in its extreme version, the
types of LD, and also differentiate LD from problem-solving model is purportedly de-
other childhood disorders, such as mental void of theoretical assumptions and classifi-
retardation and behavioral disorders such cations. It reflects an empirical approach to
as attention-deficit/hyperactivity disorder the discovery of “what works” and is fo-
(ADHD). This approach leads to definitions cused largely on improvements in the be-
based on inclusionary criteria and systemat- haviors leading to identification.
ic attempts to identify children as having
LD based on characteristics that relate to in- In implementation, the model is noncate-
traindividual differences (Lyon et al., 2001). gorical, at least regarding special education
It relies heavily on norm-referenced assess- categories in IDEA. It relies on functional
ment. analyses of learning and behavior that are
ipsative, not normative. The referent popu-
A major assumption of this model is that lation is typically locally defined. For LD,
better classifications will lead to enhanced methods that involve progress monitoring,
treatment of children with LD. The weak- such as curriculum-based assessment (Fuchs
ness of the model, especially from the per- & Fuchs, 1998; Speece & Case, 2001), are
spective of the problem-solving model, is major tools for identification.
the focus on test scores in isolation of the
child’s classroom performance (Reschly & The problem-solving model implicitly re-
Tilley, 1999). The result is more testing of tains the concepts of unexpectedness and
children, reinforcing the model currently in discrepancy but bases them on assessments

Classification and Definition of LD 33

of learning and progress over time. For ex- Thus, although it is common to cast these
ample, the initial decision regarding models as opposing views, we take the posi-
whether a child is discrepant from school tion that for LD, the two perspectives are
and/or parent expectations, essential to this actually quite compatible. Whereas the in-
model, is a discrepancy classification (Ys- traindividual difference perspective may
seldyke & Marston, 1999). The decision is lead to excessive testing and a focus on clas-
wrought with the same difficult issues that sification that does not consistently opti-
the more traditional normative classifica- mize identification of children with LD, the
tion systems possess. If a child is from a problem-solving model is not independent
low-performing school, does that mean they of classification issues, or even the concept
are not poor readers if their performance is of intraindividual differences. This model
in line with school expectations? Similarly, simply uses a different type of classification
if a child is from a high-functioning school, approach that produces issues for identifica-
should parental expectations that their child tion that are not terribly different from
be an outstanding reader represent the basis those characteristic of the intraindividual-
for such decisions? Even if one uses curricu- differences model. Moreover, although the
lum-based measurements as an alternative problem-solving model assumes null results
to more traditional norm-referenced psy- for subtype by treatment (or aptitude by
chometric measures, there is always the de- treatment) interventions, there is clear evi-
cision to be made as to whether a child has dence for such interactions, and these inter-
met, or not met, the specified academic skill ventions are at the heart of the problem-
or ability level for their group. This is clear- solving model. If not, why should schools
ly a classification problem because of the attempt to provide different treatments for
need to define the comparison group, the children with reading, math, and behavioral
academic skills/abilities to be evaluated, and difficulties? In an extreme application, why
the criteria for progress. provide any differences in instruction to any
child? The key, from our perspective, is that
In the problem-solving model, the the components of the intraindividual mod-
progress of children is constantly moni- el that are especially viable focus on acade-
tored, and those who do not show adequate mic skills and rely less on dimensions in-
development of reading or math receive tar- volving processing or special education
geted interventions (Reschley et al., 1999). diagnostic categories. But much of the re-
Identification of the student as having LD is search that shows viability has focused on
based on failure to respond to intervention, models relating processes to outcomes and
another classification decision that involves classifications that form groups using ex-
explicit criteria for sorting kids into those plicit, multidimensional criteria based on
who respond and do not respond. Such a the relevant dimensions. The groupings fa-
decision is based on a model of change. Any cilitate communication but are not neces-
determination of change requires a baseline sary for the intraindividual-differences mod-
postintervention attribute comparison, el and do not necessarily require significant
which is another type of discrepancy model. normative assessment frameworks.
Thus, in the problem-solving model, deci-
sions are made about who needs interven- Differences in Models
tions and the types of interventions they
need. These are clearly decisions that reflect In these examples, we are not ignoring the
implicit classifications of students, interven- differences in these models and the assump-
tions, intervention effectiveness, and how tions they make. A model based intraindi-
they should be matched. Otherwise all chil- vidual differences focuses on ability–ability
dren would receive the same interventions. discrepancies, whereas the problem-solving
Anything in between reflects classification, model is based on changes in the same abili-
which leads to definitions, and, in turn, ty over time. The former can be either nor-
identification. The task is always to make mative based, or ipsative based, as all deci-
explicit the implicit nature of these classifi- sions are based on individual children, an
cations. When this is done, the models are inherently ipsative process. The latter is typ-
more similar than different but focus on dif- ically relative to a behavioral baseline and
ferent characteristics of the same process.

34 FOUNDATIONS AND CURRENT PERSPECTIVES

based on ipsative change, but there are al- clear markers of the interaction of past bio-
ways questions concerning the normative logical propensities and environmental ex-
basis for evaluating changes across different periences and represent a strong case for
education contexts. However, the presump- aptitude by treatment interactions, particu-
tion that ability–ability discrepancies are re- larly from an individual-differences orienta-
lated to intervention outcomes at the level tion.
of processing, a secondary level of analysis,
or even neurobiological correlates is at best The perceived incompatibility of these
weakly established and largely reflects rela- two models ultimately reflects confusion
tionships with initial status. But that does about different levels of classification, the
not mean that such outcomes are not possi- relation of classification and identification,
ble, just that past models, particularly those and a failure to recognize that no single
focused on policy-based special education classification is suitable for all purposes. In
groups, have not been found to be valid. the remainder of this chapter, we briefly re-
These types of differentiations are also view that nature of classification, highlight-
found in the problem-solving model, where ing differences in classification and identifi-
expectancy–ability discrepancy is used to cation. We discuss the past 20 years of
identify children who need intervention classification research in the context of
(Tilly, Reschly, & Grimes, 1999). As Tilly models of intraindividual differences, high-
and colleagues (1999) state when defining a lighting levels at which subtype by treat-
discrepancy, “Data collection provides ap- ment interventions has emerged. We also
propriate quantitative and qualitative de- discuss classification and identification is-
scriptions of a target behavior and of rele- sues from the perspective of problem-solv-
vant setting expectations, yielding a ing models, outlining the classification hy-
quantitative discrepancy between the two” potheses implicit in this approach as well as
(p. 311). Similarly, they state that “the mag- how such models do indeed make use of the
nitude of the discrepancy is quantified, concept of intraindividual differences. In
based on a comparison between learner per- the end, we hope to make a case for broad-
formance and local educational demands” er understanding of both perspectives as
(p. 311). The reliance on local norms is a part of a more integrated understanding of
major issue for large-scale implementation LD with significant implications for public
and begs the question of why these models policy.
eschew norm-referenced achievement tests.
Such tests have excellent reliability and va- Nature of Classifications
lidity, providing quick “snapshots” of level
of performance. Classifications are heuristics that facilitate
the partitioning of a larger set of entities
Another difference may be more histori- into smaller, more homogeneous subgroups
cal in nature than reflective of current based on similarities and dissimilarities on a
thinking, which involves Cronbach’s (1957) set of defining attributes. When entities are
disillusionment with his early statement that assigned to the subgroups making up the
“there is some best group of treatments to classification, the process is appropriately
use and some best allocation of persons to called identification, representing opera-
treatment” (p. 680). Tilley and colleagues tionalization of the definitions that emerge
(1999) suggest that this disillusionment from the classification. Diagnosis is the
came because the research could not identi- process of applying these operational defini-
fy interactions between interventions and tions to these children to decide member-
information-processing modality, neuropsy- ship in one or more partitions. Even de-
chological profiles, or learning styles and ciding that a child needs academic
orientations. They do suggest that there was interventions is a diagnostic decision and
evidence of “prior knowledge” affecting lat- does not imply the necessity of an organis-
er learning and academic outcomes but do mic or medical model. Although we use ter-
not consider this to be evidence of the type minology that describes groupings, the
of aptitude by treatment interaction pro- groupings are essentially decisions made
posed by Cronbach. From the intraindivid- about the placement of individuals on a set
ual-differences model, such findings are

Classification and Definition of LD 35

of correlated dimensions. The decisions are cation represent hypotheses that should be
somewhat arbitrary, reflecting measurement evaluated. Such hypotheses are present in
error and the fact that the dimensions are both the intraindividual-differences model
correlated. The critical issues are the validi- and the problem-solving model and can
ty and reliability of the partitions. Valid only be evaluated by using variables that are
classifications do not exist solely because different from those used to establish the
partitions can be made. Rather, the parti- classification.
tions making up a valid classification can be
differentiated according to attributes (exter- Classification, Definition, and Identification
nal variables) not used to establish the sub-
groups. In addition to these validity consid- Many of the issues involving different mod-
erations, good classifications are also els for identifying children with LD reflect
reliable (i.e., are not dependent on the confusion about the relationship of classifi-
method of classification and replicate in cation, definition, and identification. The
other samples) and have adequate coverage relationship is inherently hierarchical in that
(i.e., permit identification of the majority of the definitions derived from classifications
entities of interest). They also facilitate com- yield criteria for identifying members into
munication, prediction, and other activities, the subcomponents making up the classifi-
though different classifications may be bet- cation. Thus, definitions of LD typically de-
ter for some than other purposes (Blashfield rive from an overarching classification of
& Draguns, 1976). childhood disorders that differentiate LD
from mental retardation and various behav-
Most endeavors in the social and behav- ior disorders, such as ADHD. This classifi-
ioral sciences, as well as the natural sci- cation yields definitions and criteria based
ences, involve classification. In the behav- on attributes that distinguish LD from men-
ioral sciences, the underlying classification tal retardation and ADHD. These criteria
is often implicit and not recognized. In clas- can be used to identify children into differ-
sification research, classifications are made ent parts of the classification model.
explicit and treated as hypotheses about the
reliability, validity, coverage, and utility of a It is pretentious and inaccurate to main-
hypothetical subgrouping of interest. In tain that any form of identification is inde-
essence, classification research is concerned pendent of an overarching classification.
with the independent variables present even Moreover, when the classification is not ex-
in single-subject designs that serve to isolate plicitly articulated, identification will be-
a child, group, or other subdivision for come fuzzy and lead to unnecessarily het-
study. Any research study is an evaluation erogeneous groupings. Thus, a major step in
of a set of dependent variables as well as the the development of identification methods
independent variables that led to the specifi- for LD was the dropping of even broader
cation of the entities under investigation concepts such as minimal brain dysfunction
(Blashfield, 1993; Fletcher, Francis, Rourke, (MBD) and the recognition that MBD con-
Shaywitz, & Shaywitz, 1993; Morris & sisted of at least two groups of children:
Fletcher, 1988; Skinner, 1981). those with difficulties primarily in the acad-
emic domain (LD) and those with difficul-
In research and practice on LD, classifica- ties primarily in the behavioral domain
tion occurs in identifying children as need- (ADHD) (Satz & Fletcher, 1980). Although
ing intervention, as having LD or typically there is overlap in which children may be
achieving; as having LD versus being men- identified into these categories, one hypoth-
tally retarded or with ADHD; within LD, as esis is that this overlap reflects comorbidity,
reading versus math impaired. When exclu- or the presence of two disorders in the same
sionary criteria are applied, LD represents a child (Fletcher, Shaywitz, & Shaywitz,
subgroup of “unexpected” underachieve- 1999). Another hypothesis is that this over-
ment. It is differentiated from expected un- lap represents one disorder with dual attrib-
derachievement due to emotional distur- utes. Although some propose that differenti-
bance, economic disadvantage, cultural and ating these disorders is not essential, as in
linguistic diversity, and inadequate instruc- the recent advancement of the notion of
tion (Kavale & Forness, 2000). From a clas- atypical brain development as an overarch-
sification perspective, these levels of classifi-

36 FOUNDATIONS AND CURRENT PERSPECTIVES

ing classification of children with various dividual model involves discrepancies in dif-
developmental difficulties (Gilger & Ka- ferent abilities typically assessed at the same
plan, 2002), this concept is not very differ- time point, whereas the problem-solving
ent from the use of MBD as a syndrome en- model typically involves the assessment of
capsulating children with LD and/or the same abilities at different time points.
ADHD. Knowing that a child is identified But the measurement issues in determining
with MBD (or with atypical brain develop- significant differences between, for exam-
ment) says little about intervention or prog- ple, two abilities are identical to those in-
nosis. However, identification with LD or volved in the assessment of significant
ADHD (or both LD and ADHD) has clear changes in the same ability at two time
implications for intervention (academic re- points (Morris, Fletcher, & Francis, 1993).
mediation, medication, behavior modifica- These issues can be understood with a brief
tion) and prognosis (Fletcher et al., 1999). discussion of the concept of an ability pro-
At this level of classification, there are well- file as a representation of similarities and
established interactions of subgroup mem- dissimilarities—the essence of classifica-
bership, interventions, and outcomes. tions.
Would we proceed by putting all children
with LD on stimulant medication regardless Ability Profiles: Similarities
of the ADHD component or, conversely, us- and Dissimilarities
ing scarce resources to put a child with
ADHD and no RD into an intensive phono- Both the intraindividual differences model
logically based intervention program? and the problem-solving model involve the
assessment of unevenness (similarities and
Classification in Intraindividual Differences dissimilarities) in ability development. This
and Problem-Solving Models is clearly apparent in the intraindividual
model, where different tests are given to de-
Regardless of the model, classifications are termine achievement and cognitive process-
implicit in any attempt to identify a child as ing strengths and weaknesses. Thus, differ-
needing academic or behavioral attention, ences in IQ and achievement, reading and
as having LD, or as needing help with read- math, or language and spatial skills can be
ing and/or math. In the intraindividual- represented as a profile that displays the
difference model, the classification is often strengths and weaknesses of a child (in a
made explicit as norm-referenced assess- clinical evaluation) or group of children (in
ment batteries are completed that presum- research), reflecting dimensions on which
ably measure attributes derived from theo- the child or groups of children are similar
retical links to the achievement problem and dissimilar.
that, in turn, are derived from the classifica-
tion that lead to identification into defined Profiles vary on multiple dimensions com-
subgroups. But the problem-solving model monly represented as shape (or pattern), ele-
also uses implicit classifications. First, the vation (or level), and scatter (or variability).
act of identifying a child as needing atten- A correlation coefficient, for example, is an
tion or as not responding to intervention is index of profile similarity/dissimilarity that
an application of a classification model. is based solely on shape and scatter with no
Similarly, different outcomes are assessed in consideration of whether the profiles differ
measuring progress in reading versus math. in elevation. Differences in elevation are
Assessment of response to intervention for commonly represented along a severity di-
academic and behavioral difficulties is not mension. An index of similarity such as Cat-
equivalent. Why would one assess response tell’s rp or squared Euclidean distance takes
to a behavioral intervention with a reading into account both shape and elevation when
outcome? Although different attributes are evaluating profiles for similarity and dissim-
measured for identification purposes in the ilarity. Two profiles with the same shape
intraindividual versus problem-solving could be highly correlated but might be rep-
models, the underlying methods (and the resented at different levels of severity, so that
difficulties implementing them) are identi- incorporating both dimension of shape and
cal. The major difference is that the intrain- elevation is usually important (Morris &
Fletcher, 1988).

Classification and Definition of LD 37

In the intraindividual-differences model, These children are essentially either more
variations in attribute profiles are implicitly severe or have two disorders. The intraindi-
at the heart of most conceptual models for vidual-differences model does not account
classifying childhood disorders and the ba- well for variations in elevation.
sis for identification. If children were identi-
cal, all ability–ability profiles would be flat, In LD (and other dimensional disorders),
consistent with a normative score at the we emphasize the importance of a multi-
mean of the population, and classification variate approach (Doehring, 1978; Satz &
would be irrelevant (Morris et al., 1993). Fletcher, 1980) at the level of both the de-
Measurement error introduces variability pendent and independent variables. At this
(scatter) around the mean and will make point in the evolution of scientific research
profiles uneven in shape and different in ele- on LD, studies of single variables are not
vation. These measurement issues also limit terribly meaningful beyond the pilot phase.
the validity of hard and fast cut points for For example, research on word-level RD
diagnostic cut points as well as whether sig- that involves a dependent variable without
nificant change has occurred with an inter- considering its relationships with phonolog-
vention. For example, a child with a 14- ical awareness or word reading cannot pro-
point discrepancy between two correlated vide a strong explanation of a group differ-
abilities is not very different from one with ence or correlation of the dependent
a 7-point difference. Similarly, a child who variable with reading. Similarly, comparing
reads 14 words per minute below expecta- groups of children defined as having LD
tion (note how expectation must be intro- without specifying the area of academic im-
duced into the discussion) is not very differ- pairment (e.g., accuracy vs. fluency of sin-
ent from one who reads 7 words per minute gle-word reading, math vs. reading vs. both
below expectation. Determining the discrep- reading and math) or relationships with
ancy in outcomes and expectations depends other disorders (e.g., ADHD) is less mean-
on the reliability of the measure of reading, ingful. It will be difficult to establish
the number of time points, and the age of whether any differences are specific to the
the child. We presume that the joint effects basis for grouping or to some other corre-
of experience and biology combine to in- lated, but not measured, attribute. If we
crease the variability of ability–ability and have learned anything from research on LD
ability–time profiles and to an extent that is over the past century, it is that the results of
greater than that induced by measurement any study depend greatly on how the sam-
error. Much of what happens in research on ple is defined, which ultimately reflects the
dependent variables is the determination of nature and explicitness of the classification
the extent to which these variations can be underlying the independent variable in the
accounted for by introduction of a classifi- study. This is true even in single-subject de-
cation (e.g., comparison of LD and typically signs as a decision had to be made that the
achieving groups), with further inferences child needed to be the single subject (i.e., re-
concerning the basis by which the classifica- quired intervention). That this issue has
tion accounts for this variability (instruc- not permeated practice and the day-to-
tion, home environment, brain function). A day identification and provision of services
paradox for the intraindividual-differences to children with LD in schools—or that
model is the fact that profiles that reflect re- researchers still use school-identified
duced elevation but that are relatively flat samples—should be of concern to everyone
are produced by the same processes (envi- involved with these children.
ronment, biology) that produce unevenness.
Don’t children who are comparably im- Whereas these issues clearly apply to the
paired in reading and math have LD? They intraindividual-differences model, we stipu-
have essentially greater impairments in lan- late that they also apply to the problem-solv-
guage and working memory than children ing model, and in areas beyond the assess-
with difficulties only in reading and math. ment of change. As we stated earlier, there is
But not calling them learning disabled (un- an implicit classification of LD as children
less they fit better in another part of the with reading, math, and behavior difficulties
classification) is a true hall of mirrors. are assessed with outcomes appropriate to
the academic domain. Another type of classi-
fication occurs when children are subdivided

38 FOUNDATIONS AND CURRENT PERSPECTIVES

into those who need intervention and those el has not been useful. From the perspective
who do not. Children are often further clas- of the problem-solving model, this classifi-
sified into those who respond and those who cation is based on attributes that are not re-
do not respond to initial intervention and lated to outcomes and detract from a focus
then those who respond to, or do not re- on intervention. For the intraindividual-
spond to, increasingly intense forms of inter- differences model, this classification exem-
ventions. The bases for both decisions are es- plifies the importance of a hypotheses-test-
sentially profiles that vary in shape, ing approach. It is important to recognize
elevation, and scatter. However, these pro- that the IQ-discrepancy model is typically in
files usually represent changes in ability de- fact a two-group classification of RD. It il-
velopment over time and may be represented lustrates clearly that a classification can lack
as a learning curve. Such curves vary in level validity but with no impact on identifica-
(intercept) and shape (slope) and also are as- tion. As Lyon, Fletcher, and Barnes (in
sociated with measurement error that is ide- press) recently summarized (see also Fletch-
ally lower than that attributed to intercept er, Lyon, et al., 2002; Lyon et al., 2001),
and slope (Fuchs & Fuchs, 1998). Even in a studies of IQ–achievement discrepancy have
single-subject-design framework, the graphs taken place largely in the domain of RD.
of functional assessment results tied to The studies include two recent meta-analy-
changes in the intervention have identical ses of studies comparing cognitive and
level and shape characteristics. The psycho- achievement correlates in children in RD
metric issues underlying the evaluation of groups based on IQ-discrepancy and low
these ipsative profiles are similar to those in- achievement definitions. There are also
volved in normative profiles in the intraindi- studies that examine prognosis, response to
vidual-differences model (Morris et al., intervention, and heretibility in IQ-
1993). The profiles represent multiple time discrepant and low-achievement groups of
points and the results of such assessments children with RD. As the two meta-analyses
also depend on how the sample is defined. show, the available studies are extensive,
The learning profiles of children with both covering the age range into adults. Mea-
RD and ADHD on a reading fluency probe sures of reading outcomes predominantly
likely differ from those associated with RD involve word recognition but extend to
or ADHD as single disorders, although this reading comprehension and fluency mea-
type of question apparently has not been sures, as well as school-identified samples
asked. One of the critical questions in all where the specific identification measures
these models is the decision made regarding are loosely specified. These studies involve
which attributes should be the focus of inter- both genders and a range of socioeconomic
vention and how one decides when the status (SES) levels. Finally, other forms of
wrong attribute has been selected. LD are addressed (e.g., math disability),
along with speech and language disorders
In the next section, we consider research (Tomblin & Zhang, 1999). The psychomet-
on the intraindividual-differences model, ric issues are well understood and singularly
examining areas in which the concept of explain why an IQ-discrepancy model is not
discrepancy has been useful and not useful. likely viable (Stuebing et al., 2002).
That section addresses the issue of individ-
ual differences and their relevance to LD, as The two meta-analyses of cognitive corre-
well as the related issue of aptitude by treat- lates of RD are most instructive, represent-
ment interventions. We then address the ing about 25 years of accumulated research.
treatment of this research by the problem- Hoskyn and Swanson (2000) identified 69
solving model, largely showing how the studies conducted from 1975 to 1996, cod-
strengths of this model are compatible with ing 19 that met stringent IQ and achieve-
the results of research from the intraindivid- ment criteria. Effect sizes were computed to
ual-differences model. compare groups of students with higher IQ
and poor reading achievement (IQ dis-
IQ–Achievement Discrepancy crepant) and students with both lower IQ
and poor reading achievement (low achieve-
The issue of IQ discrepancy is an example ment, or LA). They reported negligible to
of how the intraindividual-differences mod- small differences on several measures of

Classification and Definition of LD 39

reading and phonological processing (range subtests (spatial cognition, concept forma-
= –0.02 to 0.29), but larger differences on tion) yielded small to medium effect sizes,
measures of vocabulary (0.55) and syntax also indicating higher scores by the IQ-
(0.87). They concluded, “children with RD discrepant group. Many measures outside
share a common phonological core deficit the phonological domain shared negligible
with LA achievers. However, the results in- to small effect size differences despite the
dicated that the deficits shared by the two large differences (about standard deviation)
groups are much broader than a phonologi- in IQ between the aggregated IQ-discrepant
cal core” (p. 102). and LA groups. Altogether the difference
across the 46 studies in cognitive ability was
Stuebing and colleagues (2002) identified about three-tenths of a standard deviation,
46 studies from a sample of over 300 from demonstrating substantial overlap between
1973 to 1998. These studies included mea- the groups on phonological, language, and
sures of behavior, academic achievement, nonphonological tasks.
and cognitive abilities. From these studies,
effect sizes were computed for behavior, In examining psychometric issues, Stue-
achievement, and cognitive domains. The bing and colleagues (2002) found that vari-
effect-sizes estimates were negligible for be- ation in effect sizes across studies could be
havior (–0.05; 95% confidence interval = modeled simply by the scores on the IQ and
–0.14, 0.05) and achievement (–0.12; 95% reading tasks used to define the groups (i.e.,
confidence interval = –0.16, –0.07). A small sampling variation across studies) and the
effect size was found for cognitive ability correlation of these definitional variables
(0.30; 95% confidence interval = 0.27, with the tasks used to compare the two
0.34). groups. Thus, variation in effect sizes large-
ly reflected differences in how groups are
As the effect sizes were heterogeneous in formed, clearly showing the importance of
the achievement and cognitive ability areas, classification issues, not true differences be-
specific tasks within the each domain were tween the groups as so defined.
examined. Achievement outcomes involving
word recognition, oral reading, and spelling The results of these two meta-analyses are
showed small effect sizes indicating poorer consistent despite differences in the criteria
performance by the IQ-discrepant groups. for selecting studies and do not provide
However, outcomes on reading comprehen- strong support for the validity of classifica-
sion, math, and writing yielded negligible tions based on IQ discrepancy. Other stud-
effect sizes. The small effect sizes for word ies have examined the IQ-discrepant model
recognition, oral language, and spelling may in relation to intervention outcomes and
reflect the use of word recognition tasks to prognosis. As Aaron (1997) reported in a
define poor readers in many of the studies review of earlier studies, there is little evi-
and the correlation of these definitional dence of relationships of IQ scores or
measures with similar measures not used to groupings based on discrepancy and reading
form groups. outcomes. Fletcher, Lyon, and colleagues
(2002) reviewed six recent studies that ex-
For tasks involving cognitive ability, re- amined the outcomes of remedial and pre-
sults were similar to those of Hoskyn and vention studies in relation to IQ or IQ dis-
Swanson (2000). Those cognitive abilities crepancy (Foorman et al., 1997; Foorman,
closely related to reading disability yielded Francis, Fletcher, Schatschneider, & Mehta,
negligible effect sizes: phonological aware- 1998; Hatcher & Hulme, 1999; Torgesen et
ness (–0.13; 95% confidence interval = al., 1999; Vellutino, Scanlon, & Lyon,
–0.23, –0.02), rapid naming (–0.12; 95% 2000; Wise, Ring, & Olson, 1999). Five of
confidence interval = –0.30, 0.07), memory the six studies found no relationships. The
(0.10; 95% confidence interval = –0.01, only study to identify a relationship (Wise et
0.19), and vocabulary (0.10; 95% confi- al., 1999) found that Full Scale IQ predicted
dence interval = –0.02, 0.22). Outside the 5% of the variance in word reading out-
domain of tasks closely related to reading, comes on one measure of word reading.
measures of IQ not used to define the group However, this effect was not apparent on
yielded large effect-size differences favoring, two other measures of word reading or as-
as expected, the IQ-discrepant group. Cog- sessments of phonological processing abili-
nitive skills such as those measured by IQ

40 FOUNDATIONS AND CURRENT PERSPECTIVES

ty. As Vellutino and colleagues (2000) stat- based on neuropsychological or cognitive
ed, “the IQ-achievement discrepancy does measures. The latter are more commonly
not reliably distinguish between disabled recognized as subtypes, but the former are
and non-disabled readers. . . . Neither does vital, as they not only clearly demonstrate
it distinguish between children who were subgroup by outcome interactions in several
found to be difficult to remediate and those domains but also help establish the viability
who are readily remediated, prior to initia- of the concept of LD. As such, it is easier to
tion of remediation, and it does not predict support the use of norm-referenced achieve-
response to remediation” (p. 235). ment tests in the assessment of children with
LD than the use of neuropsychological and
Similar results are apparent for studies of cognitive tests as a demonstration of in-
prognosis in naturally occurring (nonreme- traindividual differences.
diated) samples, where longitudinal out-
comes do not differentiate IQ-discrepant ACHIEVEMENT SUBTYPES
and low achieving children identified with
RD or relate strongly (Flowers, Meyer, Lo- The division of children with LD into
vato, Felton, & Woods, 2001; Francis, groups based on the level and pattern of
Shaywitz, Stuebing, Shaywitz, & Fletcher, academic underachievement has a long his-
1996; Share, McGee, & Silva, 1989; Shay- tory (Dool, Stelmack, & Rourke, 1993;
witz et al., 1999; Vellutino, Scanlon, & Fletcher, 1985; Rourke & Finlayson, 1978).
Lyon, 2000). Similarly, O’Malley, Francis, These studies, which most commonly com-
Foorman, Fletcher, and Swank (2002) re- pare children with disabilities in reading,
ported that children identified into IQ- math, and both reading and math, show
discrepant and LA groups in grade 2 were that all forms of LD are not the same in a
similar on kindergarten assessments of dif- wide range of external attributes. As such,
ferent precursor skills. they support the heterogeneity of LD and
the need to tie LD to specific domains of
Altogether, the results of these studies do academic functioning. These subdivisions
not provide evidence for the validity of extend to variations in reading disability,
models of intraindividual differences based where children can be differentiated by pat-
on the two-group classification of children terns of strengths and weaknesses in word
into IQ-discrepant and LA groups. These recognition, fluency, and comprehension.
findings involve multiple outcomes, ap- Thus, there is a significant literature com-
proaches to defining IQ discrepancy and paring children with adequate word recog-
LA, and extend to different types of LD. Es- nition and poor reading comprehension
pecially critical are the largely null results with those who have both word recognition
for relations of IQ or IQ discrepancy with and reading comprehension deficits
intervention outcomes and long- term devel- (Cornoldi & Oakhill, 1996, reviewed in
opment. Consistent with the NCLD (2002) Lyon et al., in press). There is an emerging
position paper and other consensus docu- literature on children with deficits only in
ments (Donavon & Cross, 2002), neither word recognition or fluency with those who
IQ scores nor IQ discrepancy appear rele- have deficits in both domains. Finally, there
vant for treatment planning. Thus, this is extensive literature on interactions of RD
model for intraindividual differences lacks and ADHD (Fletcher et al., 1999).
support and should be abandoned.
This literature permits some interesting
Subtypes of Learning Disability conclusions. These subgroups, typically de-
fined by patterns on achievement tests, are
A second approach to intraindividual differ- clearly differentiated on cognitive attributes
ences involves the search for subtypes of not used to define the groups, as well as
LD. In introducing this extensive area of re- from children who are typically achieving
search, it is important to recognize that any and those with ADHD and lower IQ scores
attempt to differentiate subgroups of LD is (Dool et al., 1993; Fletcher, 1985). They
a subtyping study. Thus, attempts to com- also differ in heritability and neurobiologi-
pare children who vary in academic cal correlates (Grigorenko, 2001). The cog-
strengths and weakness are just as impor- nitive differences are clearly indicated in
tant as the studies that attempt subtyping

Classification and Definition of LD 41

Figure 3.1, which compares children de- tion, procedural learning). Figure 3.1 clearly
fined on the basis of achievement, IQ, and shows profile differences in the groups on
behavioral assessments (rating scales) into the shape and level of performance on these
those with only RD (word recognition), variables. In particular, the group with RD
only math disability (computations), only shows strengths in procedural learning and
ADHD, typically achieving, and those with weaknesses in phonological awareness. The
IQ test scores below 80, representing an op- groups with ADHD and math disability
erationalization of a mental deficiency crite- show differences from one another on the
rion. In fact, this cut point is too high and measures of concept formation and proce-
most likely should be set lower to capture dural learning, which also differentiate
children with mental retardation. Children them from the typically achieving children.
were defined as having LD based on either a The low-average IQ group has a flatter pro-
low achievement (less than 25th percentile) file and is distinguished primarily on the ba-
or a 1.5 standard error discrepancy defini- sis of elevation differences. Finally, some
tion. The dependent measures were selected variables distinguish the RD and MD
from a set of cognitive and neuropsycholog- groups from those that are typically achiev-
ical tests expected to differentiate children ing, but not from one another (rapid nam-
with LD from those who are typically ing, paired associate learning, visual–motor
achieving, as well as those with RD (phono- integration). These results, which are appar-
logical awareness, rapid naming, vocabu- ent across multiple studies of subsets of
lary, paired associate learning), math dis- these variables (see Lyon et al., in press),
ability (concept formation, procedural help establish the external validity of this
learning, visual–motor integration), and classification of mental deficiency, different
ADHD (sustained attention, concept forma- types of LD, and ADHD. It should be noted

FIGURE 3.1. Comparisons of cognitive profiles for children with only reading disability (RD), only
math disability (ND), and typical achievement (NL).

42 FOUNDATIONS AND CURRENT PERSPECTIVES

that although we discuss this classification show that the patterns on variables related
as a set of subgroups, a categorical classifi- to MD also suggest similarities (Klorman et
cation is not implied. In fact, this is a subdi- al., 2002). Thus, children with RD and MD
vision based on cut points on dimensional have essentially both disorders with more
assessments of reading, math, IQ, and be- pervasive impairment of working memory
havioral assessments of inattention and hy- and language. Imagine how results of differ-
peractivity. There are many different classi- ent studies of LD will vary depending on
fication models, which require the use of whether the study evaluates either, or both,
multiple dimensions, academic vectors, or math and reading.
prototypes that do not require dichotomous
diagnostic decisions. In Figure 3.3, we compare children with
RD, ADHD, and both RD and ADHD with
In Figure 3.2, we compare children with typically achieving children. Note that the
RD, MD, and both RD and MD with the group with RD and ADHD differs in eleva-
low-IQ group on the same variables. Note tion from the single-dimension groups but
that the group with both RD and MD dif- shows patterns that parallel their profiles. In
fers in elevation from the single-deficit addition, the group with ADHD shows rela-
groups and the low-IQ group. There are tively little cognitive morbidity on these
also configuration differences that essential- measures. Studies of children with ADHD
ly parallel the patterns seen in the single- that do not evaluate reading and math may
deficit group. The severe and parallel pat- exaggerate the extent of impairment on cog-
terns around the phonological awareness nitive tests. This suggestion is clearly sup-
dimension stand out, but statistical tests ported by Figure 3.4, which shows profiles

FIGURE 3.2. Comparisons of cognitive profiles for children with math disability (MD), reading dis-
ability (RD), both reading and math disability (RD-MD), and attention-deficit/hyperactivity disorder
and no learning disability (ADHD).

Classification and Definition of LD 43

Age Adjusted Standardized Score 1
NL
RD
ADHD

0.5 RD/ADH D

0

-0 .5

-1

-1 .5 P ro ced u ral Concept Ph o n olo g ical Rap id Nam in g Vo cab u lary P aire d Visual M otor
L e ar nin g F o rm atio n Aw are n ess Asso ciat e
Su s ta i n e d L e arn in g

Att en t io n

Profile Variables

FIGURE 3.3. Comparisons of cognitive profiles for children with reading disability and no ADHD
(with or without math disability) (RD), reading disability and ADHD (with or without math disability)
(RD/ADHD), only attention-deficit/hyperactivity disorder (ADHD), and typical achievement (NL).

for children with MD, ADHD, MD and dition, there are subgroup-by-treatment in-
ADHD, and typically achieving children. teractions on variables related to treatment
Again, children with both MD and ADHD and prognosis. In terms of treatment, empir-
differ largely on elevation, while patterns ical demonstrations show that children with
differentiate ADHD and MD from typically word-level RD do not respond to metacog-
achieving children. nitive strategy instruction, or to math inter-
ventions, but respond well to interventions
Other research shows distinctions among that incorporate explicit intervention in
children with RD based on patterns of word phonics and the alphabetic principle (Lovett
recognition, fluency, and comprehension. & Barron, 2002). In addition, children
For example, children with impairments on show differential responsiveness to various
fluency but not word recognition show dif- interventions based on patterns involving
ficulties with rapid naming and other mea- accuracy of word recognition versus fluency
sures that index speed of processing (Wolf of text reading (Lovett & Barron, 2002). In
& Bowers, 1999), but not phonological terms of prognosis, long-term outcomes are
awareness. Children with poor reading demonstrably poorer for children with LD
comprehension and adequate decoding that also involves ADHD, or for ADHD
show problems with language comprehen- that involves academic difficulties (Satz,
sion and metacognitive variables involving Buka, Lipsitt, & Seidman, 1998; Spreen,
inferencing ability, integration of textual in- 1989). It is likely that any LD that involves
formation, and abstraction (Cornoldi & more than one academic area, or that oc-
Oakhill, 1996; Lyon et al., in press). In ad-

44 FOUNDATIONS AND CURRENT PERSPECTIVES

1 NL
0.5 MD
ADH D
M D/ADHD

A g e A d ju ste d S ta n d a rd ize d S co re 0

-0.5

-1

-1.5 P ro ced u ral C once pt P h on o lo g ical R ap id N am in g Vo c a bu l a r y Paire d Visu al M o to r
L e arn in g F o rmat io n Awar en e ss Asso ciate
S ust ain e d Learning
At te nt io n

P r o file V ar ia b les

FIGURE 3.4. Comparisons of cognitive profiles for children with only math disability (MD), math dis-
ability and ADHD (MD/ADHD), only attention-deficit/hyperactivity disorder (ADHD), and typical
achievement (NL).

curs in conjunction with an oral language nitive models of learning and achievement
disorder, shows poorer outcomes. outcomes, none of which are necessary or
sufficient. Rather, the model is most akin to
These examples of external validity may prototype models where there are ideal
seem trivial or obvious. But they are impor- types and variation around the ideal type,
tant in the face of blanket assertions that or vector models that link academic do-
there is no evidence for subgroup by treat- mains in multidimensional space.
ment or aptitude by treatment interactions
(Tilley et al., 1999). These findings are the Where the issues emerge is with attempts
strongest support for the intraindividual- to use the variables that we have shown val-
difference model, with further support in idate classifications of children as having
neuroimaging studies and genetic studies LD as subclassifications based on cognitive
showing difference in the neurobiological or neuropsychological profiles. Thus, the
correlates of different subgroups of LD LD subtyping literature seeks more homo-
(Grigorenko, 2001; Lyon et al., in press). geneous subgroups in the belief that this ap-
This classification does not imply a categor- proach is related to intervention, prognosis,
ical model, especially as the distinctions are or neurobiological correlates. There is a vast
on correlated dimensions. The underly- literature on aptitude by treatment interac-
ing classification model is not monothetic, tions in education that also posits relations
which means that all attributes must be pre- of cognitive attributes, learning styles, and
sent for classification, or even the type of putative neurological functions. As we see
polythetic classifications used in psychiatric in the next two sections, there are justifiable
classification. Here there is a set of attribut- concerns about what this extensive body of
es, defined by theoretical links between cog- research has produced.

Classification and Definition of LD 45

COGNITIVE/NEUROPSYCHOLOGICAL who achieve age-appropriate word recogni-
tion but are deficient in the second or third
SUBTYPES OF LD phase are termed “rate disabled.”

Children with LD are heterogeneous. Even The strength of the Lovett subtype re-
within well-defined samples of readers with search program is its extensive external vali-
LD, there is large within-sample variance on dation (Lovett & Barron, 2002; Newby &
some skills. This observation may explain, Lyon, 1991). In a study of the two subtypes
in part, why readers with LD have been re- (rate-disabled vs. accuracy-disabled subtype)
ported to differ from controls on so many and a typically achieving sample matched on
variables (Doehring, 1978). The literature word recognition ability to the rate-disabled
on subtyping of LD is voluminous (Hooper group, accuracy-disabled readers were defi-
& Willis, 1989; Lyon et al., in press; Newby cient in a wide array of oral and written lan-
& Lyon, 1991; Rourke, 1985). It has largely guage areas different from the reading be-
not been linked to theoretical models of the haviors used to identify subtype members.
development of academic skills, brain func- The deficiencies of the rate-disabled group
tion, or other relevant bodies of work. were more apparent in deficient connected
Much of the research consists of dumping text reading and spelling (Lovett, 1987).
archival data sets into a statistical algo- Reading comprehension was impaired on all
rithm, so that results vary considerably measures for the accuracy-disabled group
across studies. There is little demonstration and was highly correlated with word recog-
that the patterns that emerge have implica- nition skill, but the rate-disabled group was
tions for intervention or prognosis. These impaired on only some comprehension mea-
concerns fit into a larger literature that sures. These additional subtype-treatment
questions whether there is value in assessing interaction studies (Lovett et al., 1988;
processing skills in children with LD, thus Lovett, Lacerenza, et al., 2000; Lovett, Rans-
questioning the assumptions underlying the by, Hardwick, & Johns, 1989) found some
intraindividual-differences model. Some differences between accuracy- and rate-dis-
possible exceptions are the focus of the re- abled groups on contextual reading out-
mainder of this section. One focuses on ra- comes, whereas word recognition improved
tional grouping of readers with LD into for both groups.
subtypes on the basis of clinical observa-
tions and/or theories related to reading dis- Lovett’s program is founded on explicit
ability (Lovett & Barron, 2002; Lovett, developmental reading theory, which has
Ransby, & Barron, 1988; Wolf & Bowers, been translated into a developmental model
1999; Wolf, Bowers, & Biddle, 2001). A of subtypes of reading, illustrates method-
second approach exemplifies the use of em- ological robustness, and offers detailed,
pirical multivariate statistical methods to thoughtful alternative explanations for the
identify homogeneous subtypes of readers complex external validation findings (New-
with LD (Lyon, 1985a; Morris et al., 1998). by & Lyon, 1991). Important treatment-
outcome findings are muted somewhat by
RATIONAL SUBTYPING reading gains on standardized measures that
do not move many children into the average
As an example of a rational (clinical) ap- in spite of statistically significant results.
proach to subtypes, Lovett (1984, 1987; Unfortunately, the classic question of how
Lovett, Steinbach, & Frijters, 2000) pro- one defines average (or expected) is an
posed two subtypes of reading disability unanswered classification question that may
based on the hypothesis that word recogni- be complicated due to contextual differ-
tion develops in three successive phases. The ences among school, classrooms, and so on.
three phases are related to (1) accuracy in Another problem is that if all children have
identifying printed words, (2) automatic effective instruction that moves them into
recognition, and (3) automatization fol- the average range, then the “expected” or
lowed by as components of the reading “average” changes and new children are
process become consolidated in memory. again identified as discrepant. Thus, there is
Children who fail at the first phase are little evidence for significant subtype by
termed “accuracy disabled,” whereas those treatment interactions (Lyon & Flynn,
1989), but this program is continuing.

46 FOUNDATIONS AND CURRENT PERSPECTIVES

More recent research continues to em- or rapid naming, as well as in reading, com-
phasize the importance of this basic distinc- pared to children with single deficits. An al-
tion between accuracy and rate but tends to ternative is to use an empirical approach to
rely on cognitive measures of processing. In subtyping to determine whether these sub-
the double-deficit model developed by Wolf types emerge, which we address in the next
and associates (Wolf & Bowers, 1999; Wolf, section.
Bowers, & Biddle, 2001), the authors pro-
pose that while phonological processing Prior to discussing empirical subtyping,
contributes considerably to word recogni- note that Wolf and colleagues (2001) base
tion deficits, reading involves the ability to subtyping at the level of processing, whereas
read both accurately and fluently. This view Lovett, Lacerenza and colleagues (2000)
receives especially strong support from base their schemas on patterns of reading
studies of RD in languages other than Eng- deficits. Other subtyping schemes also base
lish, where the relationship of phonology the classification on patterns of reading sub-
and orthography is more transparent. Thus, skills. For example, Castles and Coltheart
children with RD in German and Italian are (1993) found evidence for two subtypes or
still characterized by difficulties that are children with RD based on patterns of er-
more apparent in how rapidly they read rors in reading pseudowords and exception
words and text, not by accuracy (Paulescu (irregular) words. Relating these findings to
et al., 2001; Wimmer & Mayringer, 2002). a body of research on acquired disorders of
In these studies, a phonological defect is ap- reading (alexia), they distinguished children
parent in poor spelling, but a subgroup also with phonological dyslexia from those with
emerges that reads and spells adequately but surface dyslexia. However, there has been
has fluency deficits that are independent of virtually no external validation of these sub-
problems with phonological processing. types. Stanovich, Siegel, and Gottardo
When isolated deficits in fluency occur, the (1997) suggested that whereas phonological
most reliable correlate occurs on tasks that dyslexia could be validated, surface dyslexia
require rapid naming of letters and digits. appeared unstable and transitory.
Thus, Wolf and associates have postulated
the double-deficit model of subtypes. EMPIRICAL SUBTYPING METHODS

This model specifies three subtypes, one There are numerous examples of empirical
characterized by deficits in both phonologi- subtyping studies derived from achieve-
cal processing and rapid naming, another ment, neurocognitive, neurolinguistic, and
with impairments only phonological pro- combined classification models. Multiple
cessing, and a third with impairments only models of LD in reading have emerged
in rapid automatized naming. Wolf and as- through the application of multivariate sta-
sociates have summarized evidence, largely tistical approaches. An integrated analysis
rational, but with some evidence of validity, of several prominent reading-disability sub-
based on comparisons on other cognitive type systems that have been intensively in-
skills, that supports this subtyping scheme. vestigated suggests some areas of conver-
Although it has been suggested that the two gence in the literature (Hooper & Willis,
deficits are additive in children with double 1989). In particular, memory-span, phono-
deficits, and that the double-deficit group is logical, and orthographic processing in
more severe, there are inherent measure- reading appear to be central in defining
ment and methodological problems identi- subtypes. Although a dichotomy of audito-
fied by Schatschneider, Carlson, Francis, ry–linguistic versus visuospatial reading dis-
Foorman, and Fletcher (2002) and Comp- ability subtypes have been commonly pro-
ton, DeFries, and Olson (2001) with this as- posed, this division has not been effectively
sertion. When both phonological processing validated (Newby & Lyon, 1991), nor is the
and rapid naming are impaired, the child is evidence strong for any nonlinguistic vari-
more severely impaired in both dimensions, able as an explanation for the reading diffi-
which makes it difficult to match single- culties experienced by children with dyslex-
and double-deficit-impaired children. Thus, ia. Because of these findings, we would not
children with double deficits tend to have expect subgroup by treatment interactions
more severe problems on either phonology when auditory versus visual processing sub-

Classification and Definition of LD 47

types, or other older neuropsychological is clear that many of these approaches to
and information processing subtype models, subtyping are largely atheoretical and sim-
are used for classification purposes. The ply involved the application of multivariate
theoretical model behind the choice of at- statistical algorithms to cognitive and acad-
tributes for a classification model is critical emic variables. The resultant solutions were
to its ultimate validity and success. highly variable, and often unreliable. Al-
though there was some general replication
In a series of studies employing multivari- across groups in terms of the types of clus-
ate cluster-analytic methods, Lyon (1985a, ters that are identified, the subtypes them-
1985b) identified six subtypes of older read- selves were often difficult to relate to what
ers with LD (11- to 12-year-old children) is known about domain-specific reading or
and five subtypes of younger readers with other learning disabilities.
LD (6- to 9-year-old children) on measures
assessing linguistic skills, visual–perceptual One recent empirical subtyping study
skills, and memory-span abilities. The theo- provided support not only for the double-
retical viewpoint guiding this subtype re- deficit model, but also for models that sepa-
search was based on Luria’s (1966, 1973) rate “specific” forms of reading disability
observations that reading ability is a com- from garden-variety forms of reading dis-
plex behavior effected by means of a com- ability (Morris et al., 1998). This study dif-
plex functional system of cooperating zones fered from previous empirical approaches to
of the cerebral cortex and related subcorti- subtyping in that it was based on a theoreti-
cal structures. Within the context of this cal model emphasizing the role of phono-
theoretical framework, it could be hypothe- logical processing in reading disability
sized that a deficit in any one or several (Liberman et al., 1989; Stanovich, 1988). It
zones of the functional system could impede also used other theories to select potential
the acquisition of fluent reading behavior. variables. Thus, measures of rapid naming,
The identification of multiple subtypes short-term memory, vocabulary, and per-
within both age cohorts suggested the possi- ceptual skills were included. From a
bility that several different types of LD methodological perspective, the sample was
readers exist, each characterized by differ- large and was selected on an a priori basis
ent patterns of neuropsychological subskills for a subtyping study (i.e., it was not just a
relevant to reading acquisition. sample of convenience). Multiple definitions
were used to identify children. In addition
We emphasize this example as it explicat- to children defined with dyslexia, children
ed attempts to identify subtype by treat- with dyslexia and math disability as well as
ment interventions, a research priority that isolated math disabilities, permutations in-
is still infrequently addressed. Follow-up volving ADHD, and typically achieving
subtype-by-treatment interaction studies us- children were included. The application of
ing both age samples (Lyon, 1985a, 1985b) the clustering algorithms was rigorous and
only partially supported the independence followed guidelines ensuring both internal
of the subtypes with respect to response to and external validity (Morris & Fletcher,
treatment. It was found, however, that sub- 1988).
types characterized at both age levels by
significant deficits in blending sounds, Figure 3.5 portrays the nine resultant
rapid naming, and memory span did not re- subtypes. All profiles are depicted as z-
spond to intervention methods employing scores relative to the sample mean. Here it is
synthetic phonics procedures. Rather, mem- apparent that there are five subtypes with
bers of this linguistic-deficit subtype first specific reading disability, two subtypes rep-
had to learn phonetically regular words by resenting more pervasive impairments in
sight and then learn the internal phonologi- language and reading, and two representing
cal structure using the whole word as a typically achieving groups of children. Six
meaningful semantic context. Again, this of the seven reading disability subtypes
was true for both younger and older dis- share, however, an impairment in phonolog-
abled readers within the linguistic-deficit ical awareness skills. The five specific sub-
subtype. types largely vary in rapid automatized
naming and verbal short-term memory.
In the past 10 years, the frequency of em- Here we can see a large subtype in Figure
pirical subtyping studies has diminished. It

48 FOUNDATIONS AND CURRENT PERSPECTIVES

FIGURE 3.5. Cognitive profiles for nine subtypes of reading produced by cluster analysis. The two
subtypes in the upper panel are typically achieving children and are largely differentiated from other
subtypes by level of performance. The subtypes in the lower panel are lower in overall level of function,
representing pervasive impairments of language, and are differentiated by level and shape. The five sub-
types in the middle panel show specific patterns of strengths and weaknesses distinguishable largely by
shape (Morris et al., 1998). V = verbal; STM = short-term memory.

Classification and Definition of LD 49

3.5 that has impairments in phonological ences in cognitive correlates, heritability,
awareness, rapid naming, and verbal short- and prognosis, but it pales in relationship to
term memory. There are two subtypes with classifications based on academic and be-
impairments in phonological awareness and havior subtypes. For academic subtypes, the
verbal short-term memory, varying in lexi- interactions with outcomes are obvious.
cal and spatial skills, a subtype with phono-
logical awareness and rapid naming difficul- Where these studies have been helpful is
ties, and a subtype that is not impaired in the identification of components of inter-
phonological awareness but has deficits on vention that are essential for promoting en-
any measure that required rapid processing, hanced achievement in children with LD.
including rapid naming. This latter subtype At least for RD, progress in the develop-
is not impaired in word recognition but has ment of interventions proceeded directly
difficulties on measures of reading fluency from research on the cognitive development
and comprehension, consistent with double- of language and reading and on intraindi-
deficit model (Wolf & Bower, 1999). The vidual differences among poor readers.
five specific subtypes can be differentiated The notion in the problem-solving model
from the garden-variety subtypes on the ba- that one simply intervenes until effective
sis of their vocabulary development. Those treatments are identified has not been actu-
children with specific subtypes of reading alized and would not be expected to help
disability have vocabulary levels that are in develop interventions for LD. Indeed, re-
the average range; children with more per- search that focus largely on “what works”
vasive disturbances of reading and language in the absence of a cumulative, integrated
have vocabulary levels that are in the low- body of knowledge does not yield effective
average range. interventions. Thus, as Cummins (1999)
suggested, we lack methods for the effective
Altogether, these results are consistent instruction of English-language learners
with the central role of phonological pro- precisely because the field has been focused
cessing in word-level RD, as well as Wolf on program evaluation outcomes and not
and Bower’s double-deficit model. The re- on a cumulative body of knowledge that
sults are also consistent with Stanovich’s includes mechanism. At the same time, fo-
(1988) phonological core-variable-differ- cusing on intraindividual differences does
ences model. This model postulates that not necessarily lead to classifications that
phonological processing is at the core of all facilitate intervention, as in a problem-solv-
word-level RD. But RD is often more than ing model.
just phonological processing problems.
Children may have problems outside the Aptitude by Treatment Interactions
phonological domain that do not contribute
to the word recognition difficulties. These Whereas the intraindividual-differences
problems could be represented by impair- model focuses on different subtype hypothe-
ments in vocabulary that would interfere ses and derives from cognitive psychology
with comprehension, more pervasive distur- and neuropsychology, the problem-solving
bances of language that would lead to a gar- model focuses on the failure of aptitude by
den-variety form of RD, or even fine motor treatment interactions research as outlined
and visual perceptual problems that are in the special education literature. In the ap-
demonstrably unrelated to the reading com- titude by treatment interaction literature,
ponent of RD. The value of these distinc- the focus is not so much on subgroups as it
tions in relation to treatment, however, is is on within-child traits. These traits repre-
largely unexplored. sent natural characteristics of the child—
either strengths or weaknesses—that should
Summary: Subtyping Studies be matched to characteristics of different in-
terventions. Thus, children might be classi-
Studies of subtypes based on processing fied as “auditory” or “visual” learners, or
skills do not suggest much evidence for sub- “left-brained” or “right-brained.” The in-
type by treatment interactions. Such sub- tervention would attempt to either strength-
types can be evaluated on other external en a deficit or bypass the deficit by focusing
variables. There is some evidence for differ- on a strength. The left-brained child, for ex-

50 FOUNDATIONS AND CURRENT PERSPECTIVES

ample, would be taught to verbally mediate Foorman and colleagues (1998) demon-
math problems to compensate for “right strated that children who were weakest in
brain” weaknesses. phonological awareness showed the best re-
sponse to basal curriculums that taught the
Years of investigation of these hypotheses alphabetic principal explicitly. Lovett and
have led to largely null results. Interactions Barron (2002) predict treatment outcomes
involving modality, learning styles, or puta- based on assessments of phonological
tive neurological factors have not interacted awareness and rapid naming. This informa-
with treatment characteristics. Similarly, ev- tion was captured by interactions of initial
idence that children in different categories status and curriculum. Initial status as relat-
of special education (e.g., LD vs. mental re- ed to outcomes is vital for both the intrain-
tardation) need or respond differently to dividual-differences and problem-solving
various interventions is not apparent. As models.
Reschly and colleagues (1999) point out,
the effect of the search for aptitude by treat- An Integrated Model
ment interactions leads in directions that do
not result in better outcomes for children in Throughout this chapter, we have argued
special education and may be iatrogenic, as that the intraindividual-differences model
the model does not result in a focus on di- and the problem-solving model have
rect treatment of the target behaviors, such evolved to a point where they implicitly re-
as reading and math. flect common themes and assumptions
about LD. In this respect, research on the
We do not dispute the null results for ap- intraindividual-differences model shows
titude by treatment interactions. However, that the underlying classifications are di-
it is important to recognize that this is an mensional, representing a set of correlated
older literature where cognitive models of vectors regarding child attributes. The mod-
the development of reading and math skills el is strongest when focused on the primary
were seriously underdeveloped. Moreover, manifestations of LD, which involve read-
rejection of interactions of special educa- ing, math, and writing. It is weakest in look-
tion categories in policy does not negate the ing for ability–ability discrepancies as a
relevance of the underlying dimensions marker for LD, especially in relation to in-
themselves, just the classification in federal tervention issues. Processing strengths and
regulations. Here the criteria are demon- weaknesses do not have strong relationships
strably short on validity and reflect purpos- with treatment outcome, although they may
es other than a scientifically based classifi- have relationships with etiology.
cation. Nonetheless, it is not apparent that
children with LD benefit from interventions At the same time, there are relationships
that do not focus on the actual area of defi- with discrepancies in academic behaviors,
ciency—reading, math, and so on. In addi- treatment outcomes, and prognosis that
tion, identification for special education support the viability of a version of the in-
services must hinge on more than a score traindividual-differences model. The nor-
on a norm-referenced test, given the mea- mative basis is especially useful for making
surement issues that are involved. Here the decisions about level of performance and
value of the problem-solving model in fo- comparability of treatments. But normative
cusing identification on the results of an in- test scores or variations in them are not
tervention with convergence from other strong markers of intervention outcomes,
sources of data is apparent. Concern, how- especially if the goal is to monitor progress
ever, that the rejection of aptitude by treat- as an indicator of outcome. These tests are
ment models invalidates classifications and usually designed to be highly stable and are
automatically involves a noncategorical not sensitive to small units of change (Fran-
model is misleading. To reiterate, the prob- cis, Shaywitz, Stuebing, Shaywitz, & Fletch-
lem-solving model does involve the notion er, 1994). These models lead to paradoxes
of discrepancy. It is dimensional and does where more severely impaired children are
use information about differences among not viewed as having LD because their pro-
children. When the effects of initial status files are flatter. In fact, flatness occurs be-
or severity are considered, there is evidence
of “aptitude” by treatment interactions.

Classification and Definition of LD 51

cause of variations in severity and the corre- ferent attributes of the child. The intraindi-
lation among the attributes. It is very diffi- vidual-differences model is misleading in the
cult to argue that children with garden-vari- belief that ability–ability discrepancies in
ety LD versus specific LD are qualitatively themselves are markers of biological vari-
different as opposed to representing more ability as test scores are a product on biolo-
severe impairment in key component skills gy and environment (Fletcher & Taylor,
that in turn suppress other correlated abili- 1984). In addition, the rejection of the
ties. problem-solving model of norm-referenced
testing is also inadequate, leading to exces-
The problem-solving model is less fo- sive reliance on local norms that could be
cused on within-child variation but retains wasteful in terms of resources. Indeed,
the concept of discrepancy with environ- methods based on curriculum-based assess-
mental (i.e., class and school) or social ex- ment and progress monitoring—critical for
pectations. The discrepancy is usually rela- problem-solving models—typically attempt
tive to expectations that apply to the local to establish a broader normative base than
context. This model also relies on dimen- just a local educational context.
sional classifications, even though these
classifications are rarely articulated. It is Integrating these models requires that we
strongly focused on outcomes and purports reorganize the inherently multilevel nature
to be atheoretical and disinterested in with- of children in schools. These multiple levels
in-child characteristics, though exactly how are depicted in Figure 3.6, which shows that
such a model would decide for whom a the child is nested within the classroom,
reading (vs. math) intervention was war- which is nested within the school, which in
ranted or establish a need for more or less turn is nested in the community. The child
intervention absent classification considera- attributes are measured over time. Any
tions is difficult to conceptualize. The mea- modeling of differences in child attributes,
surement issues underlying how children are or changes over time, must take into ac-
identified under the problem-solving model count these different levels of analysis. In
are identical to those in the intraindividual- particular, response to intervention will in-
differences model despite the focus on dif- volve interactions of the children with char-

Nested Analysis of Growth in Reading

School

Classroom

Child

T1, T2, T3, T4

FIGURE 3.6. Multilevel model of growth in reading skills. Time is nested within the child, which is
nested within the classroom, which is nested within the school. Not depicted is the nesting of the school
within the community.

52 FOUNDATIONS AND CURRENT PERSPECTIVES

acteristics of the curriculum, classroom, or dominated special education eligibility de-
teacher level of the model. These two levels terminations in IDEA. Ability discrepancies
must always be considered. that only involve child attributes are clearly
not adequate. But children should not be el-
The type of classification will always de- igible for services solely based on deviations
pend on its purpose. If the goal is to identify from local expectations or an absence of
children for special education services, test normative low achievement. The notion of
scores and ability discrepancies are not suf- discrepancy and unexpectedness is main-
ficient. Children should not be placed in tained in both models. All children are ca-
special education without evidence of fail- pable of learning, and there are no true non-
ure to respond to quality instruction. At the responders—just slow responders (Denton
same time, there are different models of re- & Mathes, in press). It is absurd not to call
sponse to instruction, and it is not correct children with profiles that are relatively flat
that norm-referenced tests are completely learning disabled, or to avoid the weakness-
insensitive to change (Reschly & Tilley, es of distinctions with mild mental retarda-
1999). It depends on the model of change. tion in the absence of some intervention
As Torgesen (2002) has demonstrated, the need that emerges in the child’s response to
child who responds should change his or instruction. Thus, for identification and eli-
her rank in the population, moving into the gibility purposes, LD should be conceptual-
average range. This begs the question of ized as “unexpected” largely in the absence
“average” in that its definition will change of response to adequate instruction, and the
as interventions affect many children. But “discrepancy” a matter of not learning to
the key is to measure children in multiple expectations. Regardless, any policy-based
ways over time and look at the strengths decisions should be evaluated on the basis
and weaknesses of these different assess- of whether children benefit from the revi-
ments. Thus, while progress monitoring us- sions in the underlying classification model,
ing curriculum-based measures is vital for which are implicitly represented in any such
identification (Fuchs & Fuchs, 1998), the policy.
snapshots of behavior provided by norm-
referenced achievement tests can comple- References
ment this approach. Both types of measures
complement identification for special educa- Aaron, P. G. (1997). The impending demise of the
tion services and help remove the inherently discrepancy formula. Review of Educational Re-
relative orientation of curriculum-based search, 67, 461–502.
measures. Such approaches also provide
broader pictures of the child, particularly if Blashfield, R. K. (1993). Models of classification as
the assessment includes domains of achieve- related to a taxonomy of learning disabilities. In
ment and behavior. But even here we are not G. R. Lyon, D. B. Gray, J. F. Kavanagh, & N. A.
suggesting that, in practice, children need Krasnegor (Eds.), Better understanding learning
significant batteries of tests. Rather, the goal disabilities: New views from research and their
should be to identify intervention needs and implications for education and public policies
monitor a child’s progress, which can be (pp. 17–26). Baltimore: Brookes.
complemented by methods derived from
both models. Children do learn in social Blashfield, R. K., & Draguns, J. G. (1976). Towards
contexts, so that specifications of the learn- a taxonomy of psychopathology. The purposes of
ing environment are essential to establishing psychiatric classification. British Journal of Psy-
response to intervention. This extends to the chiatry, 129, 574–583.
school and community level of analysis in
Figure 3.6. Schools that are dysfunctional, Castles, A., & Coltheart, M. (1993). Varieties of de-
or communities that do not interact with velopmental dyslexia. Cognition, 47, 149–180.
the school, can be expected to produce large
numbers of children with learning difficul- Chronbach, L. J. (1957). The two disciplines of sci-
ties. entific psychology. American Psychologist, 12,
671–684.
In policy, the strengths of these two mod-
els must be put in place in a revision of the Compton, D. L., DeFries, J. C., & Olson, R. K.
intraindividual-differences model that has (2001). Are RAN- and phonological awareness
deficits additive in children with reading disabili-
ties? Dyslexia, 7, 125–149.

Cornoldi, C., & Oakhill, J. (Eds.). (1996). Reading
comprehension difficulties: Processes and inter-
vention. Mahwah, NJ: Erlbaum.

Cummins, J. (1999). Alternative paradigms in bilin-

Classification and Definition of LD 53

gual education research: Does theory have a Schatschneider, C., & Mehta, P. (1998). The role
place? Educational Researcher, 29, 26–32. of instruction in learning to read: Preventing
Doehring, D. G. (1978). The tangled web of behav- reading failure in at-risk-children. Journal of Ed-
ioral research on developmental dyslexia. In A. L. ucational Psychology, 90, 37–55.
Benton & D. Pearl (Eds.), Dyslexia (pp. Francis, D. J., Shaywitz, S. E., Stuebing, K. K.,
123–137). New York: Oxford University Press. Shaywitz, B. A., & Fletcher, J. M. (1994). The
Donovan, M. S., & Cross, C. T. (2002). Minority measurement of change: Assessing behavior over
students in special and gifted education. Wash- time and within a developmental context. In G.
ington, DC: National Academy Press. R. Lyon (Ed.), Frames of reference for assessment
Dool, C. B., Stelmack, R. M., & Rourke, B. P. of learning disabilities (pp. 29–58). Baltimore:
(1993). Event-related potentials in children with Brookes.
learning disabilities. Journal of Clinical Child Francis, D. J., Shaywitz, S. E., Stuebing, K. K.,
Psychology, 22, 387–398. Shaywitz, B. A., & Fletcher, J. M. (1996). Devel-
Fletcher, J. M. (1985). Memory for verbal and non- opmental lag versus deficit models of reading dis-
verbal stimuli in learning disability subgroups: ability: A longitudinal, individual growth curves
Analysis by selective reminding. Journal of Ex- analysis. Journal of Educational Psychology, 88,
perimental Child Psychology, 40, 244–259. 3–17.
Fletcher, J. M., Foorman, B. R., Boudousquie, A. B., Fuchs, L. S., & Fuchs, D. (1998). Treatment validi-
Barnes, M. A., Schatschneider, C., & Francis, D. ty: A unifying concept for reconceptualizing iden-
J. (2002). Assessment of reading and learning dis- tification of learning disabilities. Learning Dis-
abilities: A research-based, intervention-oriented abilities Research and Practice, 13, 204–219.
approach. Journal of School Psychology, 40, Gilger, J. W. (2002). Current issues in the neurology
27–63. and genetics of learning-related traits and disor-
Fletcher, J. M., Francis, D. J., Rourke, B. P., Shay- ders: Introduction to the special issue. Journal of
witz, S. E., & Shaywitz, B. A. (1993). Classifica- Learning Disabilities, 34, 490–491.
tion of learning disabilities: Relationships with Gilger, J. W., & Kaplan, B. J. (2002). Atypical brain
other childhood disorders. In G. R. Lyon, D. B. development: A conceptual framework for un-
Gray, J. F. Kavanagh, & N. A. Krasnegor (Eds.), derstanding developmental learning disabilities.
Better understanding learning disabilities: New Developmental Neuropsychology, 20, 465–482.
views from research and their implications for Grigorenko, E. L. (2001). Developmental dyslexia:
education and public policies (pp. 27–56). Balti- An update on genes, brains, and environments.
more: Brookes. Journal of Child Psychology and Psychiatry, 42,
Fletcher, J. M., Lyon, G. R., Barnes, M., Stuebing, 91–125.
K. K., Francis, D. J., Olson, R. K., Shaywitz, S. Hatcher, P., & Hulme, C. (1999). Phonemes,
E., & Shaywitz, B. A. (2002). Classification of rhymes, and intelligence as predictors of chil-
learning disabilities: An evidenced-based evalua- dren’s responsiveness to remedial reading instruc-
tion. In R. Bradley, L. Danielson, & D. P. Halla- tion. Journal of Experimental Child Psychology,
han (Eds.), Identification of learning disabilities: 72, 130–153.
Research to policy (pp. 185–250). Mahwah, NJ: Hooper, S. R., & Willis, W. G. (1989). Learning dis-
Erlbaum. ability subtyping: Neuropsychological founda-
Fletcher, J. M., & Morris, R. (1986). Classification tions, conceptual models, and issues in clinical
of disabled learners: Beyond exclusionary defini- differentiation. New York: Springer-Verlag.
tions. In S. J. Cici (Ed.), Handbook of cognitive, Hoskyn, M., & Swanson, H. L (2000). Cognitive
social, and neuropsychological aspects of learn- processing of low achievers and children with
ing disabilities (pp. 55–80). Hillsdale, NJ: Erl- reading disabilities: A selective meta-analytic re-
baum. view of the published literature. School Psycholo-
Fletcher, J. M., Shaywitz, S. E., & Shaywitz, B. A. gy Review, 29, 102–119.
(1999). Comorbidity of learning and attention Kavale, K. A., & Forness, S. R. (2000). What defin-
disorders: Separate but equal. Pediatric Clinics of itions of learning disability say and don’t say: A
North America, 46, 885–897. critical analysis. Journal of Learning Disabilities,
Fletcher, J. M., & Taylor, H. G. (1984). Neuropsy- 33, 239–256.
chological approaches to children: Towards a de- Klorman, R., Thatcher, J. E., Shaywitz, S. E., Fletch-
velopmental neuropsychology. Journal of Clinical er, J. M., Marchione, K. E., Holahan, J. M., Stue-
Neuropsychology, 6, 39–56. bing, K. K., & Shaywitz, B. A. (2002). Effects of
Flowers, L., Meyer, M., Lovato, J., Wood, F., & Fel- event probability and sequence on children with
ton, R. (2001). Does third grade discrepancy sta- attention deficit/hyperactivity, reading, and math
tus predict the course of reading development? disorder. Biological Psychiatry, 52, 773–846.
Annals of Dyslexia, 51, 49–71. Liberman, I. Y., Shankweiler, D., & Liberman, A.
Foorman, B. R., Francis, D. J., Winikates, D., (1989). The alphabetic principle and learning to
Mehta, P., Schatschneider, C., & Fletcher, J. read. In D. Shankweiler & I. Y. Liberman (Eds.),
(1997). Early interventions for children with Phonology and reading disability: Solving the
reading disabilities. Scientific Studies of Reading, reading puzzle (pp. 1–34). Ann Arbor: University
1, 255–276. of Michigan Press.
Foorman, B. R., Francis, D. J., Fletcher, J. M., Lovett, M. W. (1984). A developmental perspective

54 FOUNDATIONS AND CURRENT PERSPECTIVES

on reading dysfunction: Accuracy and rate crite- Learning disabilities as operationally defined by
ria in the subtyping of dyslexic children. Brain schools. In R. Bradley, L. Danielson, & D. P. Hal-
and Language, 22, 67–91. lahan (Eds.), Identification of learning disabili-
Lovett, M. W. (1987). A developmental approach to ties: Research to policy (pp. 287–333). Mahwah,
reading disability: Accuracy and speed criteria of NJ: Erlbaum.
normal and deficient reading skill. Child Devel- Mercer, C. D., Jordan, L., Allsop, D. H., & Mercer,
opment, 58, 234–260. A. R. (1996). Learning disabilities definitions and
Lovett, M. W., & Barron, R. W. (2002). The search criteria used by state education departments.
for individual and subtype differences in reading Learning Disability Quarterly, 19, 217–232.
disabled children’s response to remediation. In D. Morris, R. (1988). Classification of learning disabil-
L. Molfese & V. J. Molfese (Eds.), Developmen- ities: Old problems and new approaches. Journal
tal variations in learning (pp. 309–338). Mah- of Consulting and Clinical Psychology, 56,
wah, NJ: Erlbaum. 789–794.
Lovett, M. W., Lacerenza, L., Borden, S. L., Frijters, Morris, R., & Fletcher, J. M. (1988). Classification
J. C., Steinbach, K. A., & DePalma, M. (2000). in neuropsychology: A theoretical framework
Components of effective remediation for devel- and research paradigm. Journal of Clinical and
opmental reading disabilities: Combining phono- Experimental Neuropsychology, 10, 640–658.
logical and strategy-based instruction to improve Morris, R. D., Fletcher, J. M., & Francis, D. J.
outcomes. Journal of Educational Psychology, (1993). Conceptual and psychometric issues in
92, 263–283. the neuropsychological assessment of children:
Lovett, M. W., Ransby, M. J., & Barron, R. W. Measurement of ability discrepancy and change.
(1988). Treatment, subtype, and word type ef- In I. Rapin & S. Segalovitz (Eds.), Handbook of
fects in dyslexic children’s response to remedia- neuropsychology (Vol. 7, pp. 341–352). Amster-
tion. Brain and Language, 34, 328–349. dam: Elsevier.
Lovett, M. W., Ransby, M. J., Hardwick, N., & Morris, R., Satz, P., & Blashfield, R. (1981). Neu-
Johns, M. S. (1989). Can dyslexia be treated? ropsychology and cluster analysis: Potential and
Treatment-specific and generalized treatment ef- problems. Journal of Clinical Neuropsychology,
fects in dyslexic children’s response to remedia- 3, 79–99.
tion. Brain and Language, 37, 90–121. Morris, R. D., Stuebing, K. K., Fletcher, J. M.,
Lovett, M. W., Steinbach, K. A., & Frijters, J. C. Shaywitz, S. E., Lyon, G. R., Shankweiler, D. P.,
(2000). Remediating the core deficits of reading Katz, L., Francis, D. J., & Shaywitz, B. A. (1998).
disability: A double-deficit perspective. Journal Subtypes of reading disability: Variability around
of Learning Disabilities, 33, 334–358. a phonological core. Journal of Educational Psy-
Luria, A. R. (1966). Higher cortical functions in chology, 90, 347–373.
man. New York: Basic Books. National Center for Learning Disabilities & the Of-
Luria, A. R. (1973). The working brain. New York: fice of Special Education Programs. (2002). Spe-
Basic Books. cific learning disabilities: Finding common
Lyon, G. R. (1985a). Educational validation studies ground. New York: Author.
of learning disability subtypes. In B. P. Rourke Newby, R. F., & Lyon, G. R. (1991). Neuropsycho-
(Ed.), Neuropsychology of learning disabilities: logical subtypes of learning disabilities. In J. E.
Essentials of subtype analysis (pp. 228–256). Obrzut & G. W. Hynd (Eds.), Neuropsychologi-
New York: Guilford Press. cal foundations of learning disabilities: A hand-
Lyon, G. R. (1985b). Identification and remediation book of issues, methods, and practice (pp.
of learning disability subtypes: Preliminary find- 355–385). New York: Academic Press.
ings. Learning Disability Focus, 1, 21–35. Olson, R. K., Forsberg, H., Gayan, J., & DeFries, J.
Lyon, G. R., Fletcher, J. M., & Barnes, M. C. (in C. (1999). A behavioral-genetic analysis of read-
press). Learning disabilities. In E. J. Mash & R. ing disabilities and component processes. In R.
A. Barkley (Eds.), Child psychopathology (2nd M. Klein & P. A. McMullen (Eds.), Converging
ed.). New York: Guilford Press. methods for understanding reading and dyslexia
Lyon, G. R., Fletcher, J. M., Shaywitz, S. E., Shay- (pp. 133–153). Cambridge, MA: MIT Press.
witz, B. A., Torgesen, J. K., Wood, F. B., Schulte, O’Malley, K. J., Francis, D. J., Foorman, B. R.,
A., & Olson, R. (2001). Rethinking learning dis- Fletcher, J. M., & Swank, P. R. (2002). Growth
abilities. In C. E. Finn, Jr., A. J. Rotherham, & C. in precursor reading skills: Do low-achieving
R. Hokanson, Jr. (Eds.) Rethinking special educa- and IQ-discrepant readers develop differently?
tion for a new century (pp. 259–287). Washing- Learning Disability Research and Practice, 17,
ton, DC: Thomas B. Fordham Foundation and 19–34.
Progressive Policy Institute. Paulescu, E., Demonet, J. F., Fazio, F, McCrory, E.,
Lyon, G. R., & Flynn, J. M. (1989). Educational Chanoine, V., Brunswick, N., Cappa, S. F., Cos-
validation studies with subtypes of learning dis- su, G., Habib, M., Frith, C. D., & Frith, U.
abled readers. In B. P. Rourke (Ed.), Neuropsy- (2001). Dyslexia: Cultural diversity and biologi-
chological validation of learning disability sub- cal unity. Science, 291, 2165–2064.
types (pp. 243–242). New York: Guilford Press. Reschly, D. J., & Tilly, W. D. (1999). Reform trends
MacMillan, D. L., & Siperstein, G. N. (2002). and system design alternatives. In D. J. Reschly,

Classification and Definition of LD 55

W. D. Tilly, & J. P. Grimes (Eds.), Special educa- ical & Experimental Neuropsychology, 11, 389–
tion in transition: Functional assessment and 408.
noncategorical programming (pp. 19–48). Long- Stanovich, K. E. (1988). Explaining the differences
mont, CO: Sopris West. between the dyslexic and the garden-variety poor
Reschly, D. J., Tilly, W. D., & Grimes, J. P. (Eds.). reader: The phonological-core variable difference
(1999). Special education in transition Functional model. Journal of Learning Disabilities, 21,
Assessment and Noncategorical Programming. 590–604.
Longmont, CO: Sopris West. Stanovich, K. E., Siegel, L. S., & Gottardo, A.
Rourke, B. P. (Ed.). (1985). Neuropsychology of (1997). Converging evidence for phonological
learning disabilities: Essentials of subtype analy- and surface subtypes of reading disability. Jour-
sis. New York: Guilford Press. nal of Educational Psychology, 89, 114–128.
Rourke, B. P., & Finlayson, M. A. J. (1978). Neu- Stuebing, K. K., Fletcher, J. M., LeDoux, J. M.,
ropsychological significance of variations in pat- Lyon, G. R., Shaywitz, S. E., & Shaywitz, B. A.
terns of academic performance: Verbal and visu- (2002). Validity of IQ-discrepancy classifications
al-spatial abilities. Journal of Pediatric of reading disabilities: A meta-analysis. American
Psychology, 3, 62–66. Educational Research Journal, 39, 469–518.
Satz, P., Buka, S., Lipsitt, L., & Seidman, L. (1998). Tilly, W. D., Reschly, D. J., & Grimes, J. (1999).
The long-term prognosis of learning disabled Disability determination in problem-solving sys-
children. In B. K. Shapiro, P. J. Accardo, & A. J. tems: Conceptual foundations and critical com-
Capute (Eds.), Specific reading disability: A view ponents. In D. J. Reschly, W. D. Tilly, & J. P.
of the spectrum (pp. 223–249). Timonium, MD: Grimes (Eds.), Special education in transition:
York Press. Functional assessment and noncategorical pro-
Satz, P., & Fletcher, J. M. (1980). Minimal brain gramming (pp. 285–321). Longmont, CO: Sopris
dysfunctions: An appraisal of research concepts West.
and methods. In H. E. Rie & E. D. Rie (Eds.), Tomblin, J. B., & Zhang, X. (1999). Language pat-
Handbook of minimal brain dysfunctions: A crit- terns and etiology in children with specific lan-
ical view (pp. 669–715). New York: Wiley. guage impairment. In H. Tager-Flusberg (Ed.),
Schatschneider, C., Carlson, C. D., Francis, D. J., Neurodevelopmental disorders (pp. 361–382).
Foorman, B. R., & Fletcher, J. M. (2002). Rela- Cambridge, MA: MIT Press.
tionships of rapid automatized naming and Torgesen, J. K. (2002). Empirical and theoretical
phonological awareness in early reading develop- support for direct diagnosis of learning disabili-
ment: Implications for the double deficit hypoth- ties by assessment of intrinsic processing weak-
esis. Journal of Learning Disabilities, 35, nesses. In R. Bradley, L. Danielson, & D. P.
245–256. Hallahan (Eds.), Identification of learning dis-
Share, D. L., McGee, R., & Silva, P. A. (1989). IQ abilities: Research to policy (pp. 563–613) Mah-
and reading progress: A test of the capacity no- wah, NJ: Erlbaum.
tion of IQ. Journal of American Academy of Torgesen, J. K., Wagner, R. K., Rashotte, C. A.,
Child and Adolescent Psychiatry, 28, 97–100. Rose, E., Lindamood, P., Conway, J., & Garvan,
Share, D. L., & Stanovich, K. E. (1995). Cognitive C. (1999). Preventing reading failure in young
processes in early reading development: Accom- children with phonological processing disabili-
modating individual differences into a model of ties: Group and individual responses to instruc-
acquisition. Issues in Education: Contributions tion. Journal of Educational Psychology, 91,
From Educational Psychology, 1, 1–57. 579–594.
Shaywitz, S. E., Escobar, M. D., Shaywitz, B. A., U.S. Office of Education. (1977). Assistance to
Fletcher, J. M., & Makuch, R. (1992). Evidence states for education for handicapped children:
that dyslexia may represent the lower tail of a Procedures for evaluating specific learning dis-
normal distribution of reading ability. New Eng- abilities. Federal Register, 42, G1082–G1085.
land Journal of Medicine, 326, 145– 150. Vellutino, F. R., Scanlon, D. M., & Fletcher, J. M.
Shaywitz, S. E., Fletcher, J. M., Holahan, J. M., (2002). Research in the study of reading disabili-
Shneider, A. E., Marchione, K. E., Stuebing, K. K., ty (dyslexia): What have we learned in the past
Francis, D. J., & Shaywitz, B. A. (1999). Persis- four decades? Unpublished manuscript.
tence of dyslexia: The Connecticut longitudinal Vellutino, F. R., Scanlon, D. M., & Lyon, G. R.
study at adolescence. Pediatrics, 104, 1351–1359. (2000). Differentiating between difficult to reme-
Skinner, H. (1981). Toward the integration of clas- diate and readily remediated poor readers: More
sification theory and methods. Journal of Abnor- evidence against the IQ Achievement discrepancy
mal Psychology, 90, 68–87. definition of reading disability. Journal of Learn-
Speece, D. L., & Case, L. P. (2001). Classification in ing Disabilities, 33, 223–238.
context: An alternative approach to identifying Wimmer, H., & Mayringer, H. (2002). Dysfluent
early reading disability. Journal of Educational reading in the absence of spelling difficulties:
Psychology, 93, 735–749. A specific disability in regular orthographies.
Spreen, O. (1989). Learning disability, neurology, Journal of Educational Psychology, 94, 272–
and long-term outcome: Some implications for 277.
the individual and for society. Journal of Clin- Wise, B. W., Ring, J., & Olson, R. K. (1999). Train-

56 FOUNDATIONS AND CURRENT PERSPECTIVES

ing phonological awareness with and without ex- ceptual review. Journal of Learning Disabilities,
plicit attention to articulation. Journal of Experi- 33, 387–407.
mental Child Psychology, 72, 271–304. Ysseldyke, J., & Marston, D. (1999). Origins of
Wolf, M., & Bowers, P. G. (1999). The double categorical special education services in schools
deficit hypothesis for the developmental dyslexi- and a rationale for changing them. In D. J.
as. Journal of Educational Psychology, 91, Reschly, W. D. Tilly, & J. P. Grimes (Eds.), Spe-
415–438. cial education in transition: Functional assess-
Wolf, M., Bowers, P. G., & Biddle, K. (2001). Nam- ment and noncategorical programming (pp.
ing-speed processes, timing, and reading: A con- 1–18). Longmont, CO: Sopris West.

4

Learning Disabilities and the Law

Cynthia M. Herr
Barbara D. Bateman

If a special educator of the 1950s or 1960s and, rarely, such a practice. Clearly, law has
had been asked how likely it was that by the an impact on special education practice.
late 1970s every aspect of the daily practice
of special education would be governed by We also discuss the possibility that LD re-
detailed federal law, the answer probably search and practice can or should affect the
would have been, “Not likely.” However, it law. We examine, for example, what influ-
has come to pass, and now few special edu- ence, if any, our ever-increasing knowledge
cators remember what special education of the effectiveness of specific methodolo-
looked like before it was subject to legal gies for students with LD has had on the
mandate at every turn. law. We focus primarily on court decisions
in major cases involving students with LD.
The focus of this chapter is how the law, The cases presented are representative, not
both legislation and litigation, affects the exhaustive. State hearing decisions have not
field of learning disabilities (LD). We exam- been presented because they are too numer-
ine the impact of law on (1) evaluation and ous, often inconsistent, and do not consti-
eligibility for services determinations, (2) the tute legal precedent.
development of individualized education
programs (IEPs) and the provision of a free Before looking specifically at LD cases, it
appropriate public education, and (3) the is important to look briefly at the law be-
provision of education in the least restrictive hind the cases.
environment. For example, the Individuals
with Disabilities Education Act requires Litigation and Legislation
IEPs for all eligible students. These IEPs (at
least in law) control the special education Litigation (i.e., case law) can prompt legisla-
and related services to be delivered to stu- tion, which, in turn, produces more litiga-
dents. If an eligible student’s behavior inter- tion, which then interprets and applies the
feres with learning, IDEA requires that the legislation. That litigation can then result in
IEP contain a behavior intervention plan new or amended legislation. And the cycle
which must, of course, be implemented. Pri- goes on. An early example of the effect of
or to IDEA, there was no such requirement special education litigation on legislation

57

58 FOUNDATIONS AND CURRENT PERSPECTIVES

was the influence that two cases in the early now in a cycle of litigation which will inter-
1970s had on legislation passed in 1975. pret and clarify those changes, as well as
The Pennsylvania Association for Retarded continue to resolve other disputes.
Children (PARC) v. Pennsylvania (1972)
and Mills v. D. C. Board of Education Three major pieces of federal legislation
(1972) cases arose out of grassroots move- affect the education of individuals with dis-
ments by parents of children with disabili- abilities, including those with LD: (1) IDEA,
ties to force public schools to provide edu- (2) Section 504 of the Vocational Rehabili-
cation for their children, many of whom tation Act of 1973, and (3) the Americans
had been excluded from public schools or with Disabilities Act (ADA) of 1990. There
segregated in separate schools for children is some overlap across these three acts. Be-
with disabilities. From these cases came the cause the great majority of cases in the field
impetus for passage of the Education of All of LD have been filed under IDEA, we deal
Handicapped Children Act (EAHCA) of only with this legislation in this chapter.
1975. Originally called Public Law (PL) 94- IDEA, first enacted in 1975, governs the
142, it is now called the Individuals with provision of special education and related
Disabilities Education Act, or IDEA. The services to all children with disabilities ages
major provisions of IDEA, which relate to 3–21 who qualify under IDEA.
zero-reject, free appropriate public educa-
tion (FAPE), discipline, child find, due Because IDEA coverage extends only
process hearings, and other procedural pro- through secondary school graduation or age
tections, were patterned closely after the 21, cases of alleged discrimination against
court orders of PARC and Mills. college students and employees with LD are
brought under Section 504 and/or the ADA.
The IDEA legislation then resulted in In most of these cases, individuals are seek-
new litigation. The landmark special educa- ing accommodations in admissions, assess-
tion case of Hendrick Hudson Central ment, or the workplace or working condi-
School District Board of Education v. Row- tions. A review of these cases is beyond this
ley (1982) illustrates how case law (litiga- discussion. With a few notable exceptions
tion) clarifies and interprets legislation. (e.g., Guckenberger v. Boston University
Even though the purpose of IDEA is to pro- [1997]), most plaintiffs with LD have not
vide FAPE to every child with a disability, prevailed. The common issues are whether
the statute itself fails to fully define the crit- the plaintiff with LD is actually a “dis-
ical word “appropriate.” The U.S. Supreme abled” person as defined by Section 504
Court, therefore, defined “appropriate edu- and ADA (i.e., whether she or he has a
cation” in Rowley: “[T]he education to “substantial limitation” in a major life ac-
which access is provided [must] be suffi- tivity) and/or whether the disputed accom-
cient to confer some educational benefit modation would involve a fundamental al-
upon the handicapped child” (pp. 200– teration in the program or job. The
201). The Court further concluded that remainder of this chapter presents cases
“the ‘basic floor of opportunity’ provided brought under IDEA.
by the Act consists of access to specialized
instruction and related services which are Evaluation and Eligibility
individually designed to provide education-
al benefit to the handicapped child” (p. The processes of evaluating individuals for
202). LD and determining their eligibility for ser-
vices and protection from discrimination
Later cases such as Honig v. Doe (1988), raise two main issues which various courts
which dealt with the discipline of students have addressed: (1) whether an individual
with disabilities, and Delaware County In- has a learning disability and (2) whether a
termediate Unit No. 25 v. Martin K. (1993), student’s academic problems are the result
dealing with methodology for a student of a learning disability or some other factor.
with autism, led to IDEA amendments in Embedded in the first issue is the highly
1990 and 1997. The cycle of litigation– controversial issue of what standard should
legislation–litigation–legislation will contin- be used to determine whether an individual
ue in the field of special education. Follow- has a learning disability.
ing the 1997 amendments to IDEA, we are

Learning Disabilities and the Law 59

Does an Individual Have a cy between the students’ achievement and
Learning Disability? their intellectual abilities. In Welton v. Lib-
erty 53 School District (2001), the court
IDEA defines a learning disability both by also agreed with the district that Eric Wel-
what it is and what it is not: ton did not have a learning disability in
reading and math, as the parents insisted,
The term means a disorder in one or more ba- because Eric’s academic performance in
sic psychological processes involved in under- these areas met state criteria. The district
standing or in using language, spoken or writ- and the court agreed, however, that Eric did
ten, that may manifest itself in an imperfect have a learning disability in the area of writ-
ability to listen, think, speak, read, write, ten language because his performance fell
spell, or to do mathematical calculations, in- below the state’s criterion in this area.
cluding conditions such as perceptual disabili-
ties, brain injury, minimal brain dysfunction, Generally, courts have tended to look at
dyslexia, and developmental aphasia. . . . three factors when deciding whether a stu-
dent has a learning disability: (1) the pres-
The term does not include learning prob- ence of a severe discrepancy between
lems that are primarily the result of visual, achievement and ability, (2) the student’s
hearing, or motor disabilities, of mental retar- need for special education, and (3) state cri-
dation, of emotional disturbance, or of envi- teria for a learning disability. For example,
ronmental, cultural, or economic disadvan- in Ridgewood Board of Education v. N.E.
tage. (34 C.F.R. § 300.7(b)(10)) (1999), the court found that a student had a
learning disability because both an indepen-
Of all the categories of disability defined dent consultant and the district’s child study
in IDEA, the category of specific learning team found that the student demonstrated a
disability is the only category for which spe- severe discrepancy between his ability and
cific evaluation procedures and criteria are his achievement even though the district
provided: claimed that the student did not meet New
Jersey’s requirement that he have a “percep-
A team may determine that a child has a spe- tual impairment.”
cific learning disability if—
Rarely does a court look beyond the
(1) The child does not achieve commensurate above factors to determine whether a stu-
with his or her age and ability levels in one dent has a learning disability. The court
or more of the areas listed . . . if provided in Corchado v. Board of Education of
with learning experiences appropriate for Rochester City School District (2000) is a
the child’s age and ability levels; and notable and refreshing exception to this
practice. Despite the fact that the evidence
(2) The team finds that a child has a severe indicated that the student had average acad-
discrepancy between achievement and in- emic performance, the Corchado court said:
tellectual ability in one or more of the fol-
lowing areas: The IHO’s [independent hearing officer] rea-
soning, in effect, precludes a child whose aca-
(i) Oral expression. demic achievement can be described as “satis-
(ii) Listening comprehension. factory” from being able to demonstrate that
(iii) Written expression. documented disabilities adversely affected the
(iv) Basic reading skill. student’s academic performance. This should
(v) Reading comprehension. not and cannot be the litmus test for eligibility
(vi) Mathematics calculation. under IDEA.
(vii) Mathematics reasoning.
Each child is different, each impairment is
(34 C.F.R. § 300.541(a)) different, and the effect of the particular im-
pairment on the particular child’s educational
The majority of litigation which relates to achievement is different. (p. 375)
LD evaluation and eligibility involves dis-
agreements between parents and districts The Corchado court decision illustrates
about whether a student has a learning dis- perfectly one of the major controversies that
ability. In both Hiller v. Board of Education exists today concerning the determination
of the Brunswick Central School District of whether an individual has a learning dis-
(1990) and Norton v. Orinda Union School
District (1999), the courts agreed with the
districts that the students did not have LD
because there was no significant discrepan-

60 FOUNDATIONS AND CURRENT PERSPECTIVES

ability. The controversy centers on the use as promised by OSEP in the final regula-
of discrepancy formulas to determine tions for IDEA 1997 (64 Fed. Reg. 12637),
whether a “severe discrepancy” between to determine whether changes should be
achievement and ability exists. Nowhere in proposed to the procedures for evaluating
the regulations for IDEA is the term “severe children suspected of having a learning dis-
discrepancy” defined. A discrepancy formu- ability.
la was proposed for IDEA in 1976 but was
soundly rejected by the field. Nevertheless, Are Academic Difficulties Caused by a
almost all schools rely on a state- or district- Learning Disability or Other Factor?
level discrepancy formula, even though
these are usually intended only as guide- Although the issue of whether an individual
lines. New York, for example, demands a has a learning disability is the most fre-
“discrepancy of 50% or more between sug- quently litigated issue in the area of evalua-
gested achievement and actual achieve- tion and eligibility, it is not the only issue
ment” (8 N.Y.C.R.R. § 200.1(mm)(6)). Re- that reaches the courts. Several cases have
liance on discrepancy formulas has addressed the issue of whether a student’s
flourished and is now nearly universal. The academic difficulties were really the result
thrust of IDEA in identifying children who of a learning disability or whether they were
have LD is reliance on professional judg- the result of other factors such as speech or
ment, based on a full and thorough individ- behavior problems. Kelby v. Morgan Hill
ual and individualized evaluation. However, Unified School District (1992) illustrates
practice has moved increasingly toward re- the difficulties districts, hearing officers,
liance on a formula and on a predeter- and courts can have in trying to sort out
mined, one-size-fits-all, limited selection of whether a student’s academic problems are
standardized tests. Factors other than law the result of a learning disability or behav-
have weighed heavily in shaping these prac- ior problems. Parents often argue that the
tices. One factor, ironically, is a fear of legal behavior problems are the result of a stu-
actions. Many psychologists and educators dent’s frustration because of academic fail-
mistakenly believe that reliance on a quanti- ure, while districts may argue that the be-
tative formula and the one-size-fits-all test- havior problems cause the student to have
ing procedure can be more easily defended academic problems. In Kelby, the district,
in a hearing or court than can professional the hearing officer, the district court, and
judgment in selecting and interpreting as- the circuit court all agreed that Richard Kel-
sessment procedures. Other factors limiting by’s academic problems were due to behav-
the law’s influence include the system being ioral issues even though two independent
overwhelmed by increasing numbers of stu- evaluators stated that Richard had a learn-
dents needing special education and inade- ing disability due to processing deficits. In
quate resources, including insufficient num- Capistrano Unified School District v. War-
bers of personnel trained and experienced in tenberg (1995), the Court of Appeals for the
evaluating and recognizing LD. Ninth Circuit disagreed with the district’s
decision that Jeremy Wartenberg’s academic
This is clearly an area where neither the failure was due to his misbehavior. In reach-
law nor LD practice has positively influ- ing its decision, the court held that “Jere-
enced the other. Instead both the law and my’s social maladjustment could not be sep-
LD practice seem stuck 30 years in the past arated out from his organic disorder, and
when the use of discrepancy models was that his misconduct was primarily caused by
considered appropriate (Stanovich, 1999). his organic disorder rather than a non-
covered problem” (p. 810).
In an acknowledgment of the controver-
sies that surround the field of LD, including Though some may argue that a student’s
issues of evaluation and eligibility, the Of- disability label does not matter as long as
fice of Special Education Programs (OSEP) the student receives appropriate services,
of the federal Department of Education held the court in Friedman v. Vance, Mont-
a Learning Disabilities Summit in Washing- gomery County Board of Education (1996)
ton, DC, in August 2001. One of the pur- disagreed. Here the issue was whether
poses of that summit was to begin discus- Alexander should receive services under a
sions and examination of research findings,

Learning Disabilities and the Law 61

speech/language label or under an LD label. to access records, participate in decision
The district court concluded “by a prepon- making, obtain an independent educational
derance of evidence, that Alexander ha[d] a evaluation, receive prior written notice of
learning disability” (p. 657). The court went district proposals and refusals, consent to
on to point out that the disability designa- initial evaluation and special education
tion was not irrelevant: “By missing the placements, present complaints, initiate a
learning disability problems, over and hearing and appeal, have the student’s
above the speech and language problems, placement maintained while disputes are
the IEP lacks a full set of goals and objec- pending, and more. The Rowley court par-
tives, and those that are present consist of ticularly emphasized the importance of the
mere sketches of the full range of services procedural aspects of IEP development:
needed” (p. 657). “Congress placed every bit as much empha-
sis upon compliance with procedures . . . as
Although the issues of evaluation and eli- it did upon the measurement of the resulting
gibility are not insignificant, legislation and IEP against a substantive standard” (pp.
litigation have had a far greater effect on 205–206).
learning disability practice in the area of
IEPs and the provision of FAPE. Lake (2000) has summarized about 150
hearings and cases involving IEP procedural
IEPs and FAPE errors related to timeliness, notice, team
members and participation, IEP form and
The law has clearly had a significant impact development, IEP content, and implementa-
on LD practice in the areas of the develop- tion. Some procedural errors are deemed
ment of IEPs and the provision of FAPE to “harmless” and do not result in a denial of
students with LD. Before 1975 and the earli- FAPE. Those serious violations which do
est versions of IDEA, there was no legal re- constitute a denial of FAPE are those that
quirement for IEPs and school districts were result in an actual loss to the student of edu-
not required to provide children with dis- cational opportunity or benefit or that deny
abilities with FAPE. Whether or not one be- parents the opportunity for full and mean-
lieves that IEPs serve a useful purpose in the ingful participation in the IEP process.
provision of special education to students
with learning or other disabilities, they are Courts have ruled that LD students were
mandated by law as a means to ensure that denied FAPE by (1) districts’ failure to give
students served under IDEA receive FAPE. parents adequate notice of their procedural
Many cases have been filed by parents claim- protections (Hall v. Vance County Board of
ing that their children with LD have been de- Education, 1985; Briere v. Fair Haven
nied FAPE because of either procedural or Grade School District, 1996), (2) districts’
substantive errors in the development of the failure to follow IEP procedures such as
IEP or in the document itself. The U.S. timelines and essential content of IEPs (Hall
Supreme Court in Rowley (1982) made it v. Vance, 1985; Briere v. Fair Haven Grade
clear that states must comply with both the School District, 1996; Evans v. Board of
procedural and substantive requirements of Education of the Rhinebeck Central School
IDEA: “Therefore, a court’s inquiry in suits District, 1996; Susquenita School District v.
brought under §1415(e)(2) is twofold. First, Raelee, 1996; Flowers v. Martinez Unified
has the State complied with the procedures School District, 1993; Egg Harbor Town-
set forth in the Act? And second, is the indi- ship Board of Education v. S.O., 1992; Las-
vidualized educational program developed cari v. Ramapo Indian Hills Regional High
through the Act’s procedures reasonably cal- School District, 1989), (3) a district’s unilat-
culated to enable the child to receive educa- eral decision to graduate a student (Kevin T.
tional benefit” (p. 206)? v. Elmhurst Community School District,
2001), and (4) a district’s failure to follow
Procedural Issues IDEA’s guidelines for evaluating students
with specific learning disability (Evans v.
The extensive and detailed procedural re- Board of Education of the Rhinebeck Cen-
quirements of IDEA relate to parents’ rights tral School District, 1996).

Other courts have determined that similar
procedural violations did not reach the level

62 FOUNDATIONS AND CURRENT PERSPECTIVES

of denying FAPE to students (Salley v. St. ized inquiry. The Rowley court, therefore,
Tammany Parish School Board, 1996; Wel- declined to establish any one test for the ad-
ton v. Liberty 53 School District, 2001; Ju- equacy of the benefit, beyond making it
dith S. v. Board of Education of Community clear that IDEA does not require the school
Unit School District No. 200, 1998; Liv- to provide the best program possible or to
ingston v. Desoto County School District, maximize each child’s potential. Within
1992; Doe v. Defendant I., 1990; Hiller v. those limits, several federal circuit courts of
Board of Education of the Brunswick Cen- appeal have weighed in.
tral School District, 1990).
The Fourth Circuit examined the “how
The major way that these cases have much benefit” issue in Hall v. Vance County
affected LD practice is in the heightened Board of Education (1985). The district
awareness among professionals of the many contended that the student with LD’s social
procedural requirements regarding the de- promotions plus minimal improvement on
velopment of IEPs and the provision of ser- standardized reading tests over a 4-year pe-
vices to students with LD. However, as the riod showed sufficient benefit to constitute
following section demonstrates, a height- FAPE. The court considered the student’s
ened awareness among professionals does progress in light of the student’s own intel-
not always mean that students receive ap- lectual capabilities and observed:
propriate services.
Rowley recognized that a FAPE must be tai-
Substantive Issues lored to the individual child’s capabilities and
that while one might demand only minimal re-
In this section we discuss the most signifi- sults in the case of the most severely handi-
cant issue courts have dealt with since IDEA capped children, such results would be insuffi-
was enacted, which is: To how much benefit cient in the case of other children. Clearly,
is a student with a disability entitled in the Congress did not intend that a school system
provision of FAPE? We also discuss a sec- could discharge its duty under the EAHCA
ond issue, that of methodology, which is [predecessor to the IDEA] by providing a pro-
gaining increasing importance in special ed- gram that produces some minimal academic
ucation, especially in the provision of an ap- advancement, no matter how trivial. (p. 636)
propriate education to students with LD.
The Third Circuit, in the case of a dyslex-
THE BENEFIT STANDARD ic student, said that IDEA “calls for more
than a trivial educational benefit” (Ridge-
When Rowley was decided in 1982, it was wood, 1999, p. 247). In Capistrano (1995),
necessary for the Supreme Court to deter- the Ninth Circuit was explicit in detailing
mine the type of standard (maximize poten- an appropriate program to meet Jeremy’s
tial, provide equal opportunity, etc.) to be educational needs and then concluding that
used to judge program “appropriateness.” Jeremy’s IEP did not meet those needs but
The Supreme Court chose the “reasonably that the private LD school did. Among Jere-
calculated to enable the child to receive edu- my’s named needs were frequent feedback,
cational benefit [italics added] standard,” clear commands, structured school environ-
now widely called the “ed benefit” or the ment, small class size, consistent behavior
Rowley standard. In each individual case, a management, and individualized attention.
second inquiry is required, namely, to how The court also found that Jeremy’s public
much benefit is this child entitled? Benefit is school IEP moved him too much between
often measured by progress. In this context, classes; assigned him too many teachers;
an appropriate program for an intellectually and failed to provide the structure, consis-
gifted 15-year-old would potentially allow tency, and attention that Jeremy needed.
her to progress or benefit more than an ap-
propriate program would benefit a 4-year- In contrast to the Ninth Circuit’s Capis-
old who had severe mental retardation. trano decision, the Eighth Circuit in Fort
Zumwalt v. Clynes (1997) concluded that
Thus, the question of whether FAPE (in- Nicholas Clynes had received sufficient edu-
cluding sufficient benefit) has been provided cational benefit even though, after 5 years
must always be a fact-specific, individual- of school, Nicholas’s word attack skills
were at a first-grade level and he did not

Learning Disabilities and the Law 63

know the alphabet. The court was im- ‘floor of benefit’, [citation deleted] they are
pressed that Nicholas earned passing marks certainly one indicator” (p. 7). Evans v.
in the third grade and that he had been pro- Board of Education of the Rhinebeck Cen-
moted to fourth grade just before his par- tral School District (1996) held that the dis-
ents removed him from the public school. trict’s program could not provide educa-
The dissenting judge had a different view: tional benefit to the student because the
proposed teacher “was not qualified to
[T]he IEP for 1991–1992 was not designed to teach adolescents or to instruct, train or
provide “personalized instruction with suffi- otherwise consult with teachers as to how
cient support services to permit [Nicholas] to to work with Frank [the student] using the
benefit educationally from that instruction.” approach he requires” (p. 348). Since the
By submitting an IEP substantially similar to 1997 IDEA amendments, several courts
others that had previously produced so few have increased the bar of what constitutes
positive results, and by exhibiting an unwill- enough educational benefit to provide a
ingness to explore any different approaches, FAPE.
Fort Zumwalt did not extend to Nicholas the
free and appropriate education mandated by A federal district court in Indiana recently
IDEA. . . . Nicholas’s achievements, particu- dealt in detail with LD and FAPE issues
larly in the area of reading skills, can at best (Nein v. Greater Clark County School Corp.,
be described as trivial. This cannot be the sort 2000). From kindergarten through fourth
of education Congress had in mind when it en- grade, Lucas attended an elementary school
acted IDEA. (p. 617) in the Greater Clark District (Indiana). He
was identified in first grade as having a se-
When appellate judges disagree so vere learning disability. In first grade, Lucas’s
sharply on whether a student received Wechsler Intelligence Scale for Children
FAPE, it should not be surprising that we (WISC; Wechsler, 1991) Full Scale IQ was
are sometimes confused. Many district 95. Nevertheless, in January of fourth grade,
courts have also wrestled with the issue of Lucas was reading below second-grade level
how much benefit a child with a learning and was spelling at first-grade level. In No-
disability should expect under IDEA. In vember of Lucas’s fifth-grade year, a hearing
Pascoe v. Washingtonville Central School officer (as quoted by the district court) found
District (1998), the court ruled that a 17- the district had denied Lucas FAPE: “The
year-old student with severe dyslexia had re- ability to read is a fundamental ingredient in
ceived enough benefit because the student a free appropriate education that can be di-
earned credits toward a high school diplo- minished only by a finding that the disabled
ma and had passed the New York Regent’s child is clearly incapable of achieving read-
Competency Tests even though the student’s ing skills transferable to life settings. The
reading and writing skills were at about a failure to use an approach that will provide
second-grade level. The court also noted Student with the tools to become an inde-
that although several experts expressed the pendent reader is alone an important reason
opinion that the student could have made why the LEA did not provide an appropriate
substantial progress in reading and writing education” (p. 970).
in the right environment, “such potential
does not establish that the IEP in issue was The district court went on to explain, “If
inappropriate” (p. 35). an IEP must be designed to take into ac-
count a child’s individual educational needs,
Even before IDEA’s 1997 amendments it logically follows that the child’s capacity
significantly increased the focus on level of to learn should also be considered in evalu-
benefit and on objective evidence of ating the IEP” (p. 974).
progress, some district courts held that lim-
ited educational progress was not sufficient After noting that district personnel did
to prove educational benefit. In Egg Harbor not demonstrate any expertise or significant
Township Board of Education (1992), the training in teaching students with dyslexia,
court stated that “while the benefits or lack the court reviewed and refuted the school
of them actually realized by a child are not district’s claims that they had provided suf-
dispositive of the question of whether the ficient benefit to Lucas (i.e., a Chevrolet ed-
program was sufficient to satisfy Rowley’s ucation) while the parents were demanding
a Cadillac:

64 FOUNDATIONS AND CURRENT PERSPECTIVES

At the risk of carrying these metaphors too far, specially designed instruction as “adapting,
for a student like Lucas, the ability to read is as appropriate to the needs of an eligible
truly the key that opens the door to all other child under this part, the content, methodol-
aspects of an education. In terms of the auto- ogy (emphasis added) or delivery of instruc-
motive metaphor, Greater Clark was provid- tion . . .” (34 C.F.R § 300.26(b)(3)).
ing the Neins with a Chevrolet without a
transmission—even if the engine might run, no As significantly, Attachment 1 (titled
power ever reached the wheels. Because the Analysis of Comments and Changes) to the
Milestones program produced no transferable 1999 IDEA regulations (64 Fed. Reg.
progress in three years, as both the initial 12552) discusses the foregoing change to
hearing officer and the Board of Appeals the definition of specially designed instruc-
found, the program was plainly failing to pro- tion:
vide even a minimally adequate educational
benefit. (p. 977) Case law recognizes that instructional method-
ology can be an important consideration in the
The court also saw through the district’s context of what constitutes an appropriate ed-
claim that adequate benefit was shown by ucation for a child with a disability. At the same
Lucas’s grades (in fact, they were modified) time, these courts have indicated that they will
and his promotions (retention was against not substitute a parentally-preferred methodol-
district policy) as did the court in R.R. v. ogy for sound educational programs developed
Wallingford Board of Education (2001), by school personnel in accordance with the
which held: procedural requirements of the IDEA to meet
the educational needs of an individual child
Despite the student’s attainment of passing with a disability.
grades and his regular advancement from
grade to grade, we are not persuaded by the In light of the legislative history and case
Board’s argument that the student was making law, it is clear that in developing an individu-
satisfactory progress. The record is replete alized education there are circumstances in
with test results indicating that, despite having which the particular methodology that will be
been placed in a mainstream ninth grade class, used is an integral part of what is “individual-
the student had not progressed in reading abil- ized” about a student’s education and, in
ity beyond a third or fourth grade level. (p. those circumstances will need to be discussed
123) at the IEP meeting and incorporated into the
student’s IEP (emphasis added). For example,
Indisputably, the benefit standard has for a child with a learning disability who has
been newly highlighted and emphasized, not learned to read using traditional instruc-
and most of us agree the standard has been tional methods, an appropriate education may
raised. Congress’s insistence on improving require some other instructional strategy.
special education outcomes requires, among
other changes, greater utilization of effec- In all cases, whether methodology would be
tive teaching practices (i.e., effective addressed in an IEP would be an IEP team de-
methodology). cision. (64 Fed. Reg. 12552)

METHODOLOGY AS A SUBSTANTIVE ISSUE If we recall that every IEP must contain
“a statement of the special education . . . to
OF FAPE be provided” (20 U.S.C. § 1414(d)(1)
(A)(iii)), that special education is specially
This section examines recent changes in the designed instruction and that methodology
law and the effects of those changes. How- is included therein, we can only conclude, as
ever, we must begin with an important non- has the U.S. Office of Education, that there
change. are times at which methodology is an essen-
tial ingredient in FAPE and must, therefore,
From 1975 to the present time, IDEA has be included in the IEP.
defined special education as “specially de-
signed instruction, at no cost to parents, to Though some degree of deference may
meet the unique needs of a chid with a dis- still be due the school’s choice of methodol-
ability . . .” (34 C.F.R. § 300.26(a)). There ogy under the 1997 IDEA, it seems clear
is no change here. But, in the 1999 regula- that any presumption of appropriateness of
tions, the U.S. Office of Education defined that school choice now is properly seen as
rebuttable. Under IDEA, the parents have
the right to dispute any matter related to

Learning Disabilities and the Law 65

the provision of FAPE (34 C.F.R. § objectives set forth in the IEP were reasonable.
300.507(a)(1)). Methodology is, in many The goals must be realistic and attainable, yet
cases, the key element in FAPE and must be more than trivial and de minimis. . . .
subject to dispute resolution.
Third, the school district must show that
If any further evidence is required that the methodology that it employed was tailored
methodology may be and sometimes must to meet the annual goals, benchmarks, and
be on the IEP, one has only to look at the short-term objectives set forth in the IEP. Stat-
new mandates for the IEP team to address ed differently, the special education and relat-
Braille for students with visual impairments, ed services must be tailored to reasonably ac-
to consider positive behavior interventions complish the goals in the IEP. (p. 610)
for students with behavior issues, and to
conduct functional behavior assessments as The court in Nein (2000) also dealt with
necessary. These are all squarely methodolo- methodology issues in determining whether
gies. When any techniques and/or method- Lucas had received an appropriate educa-
ologies are on an IEP, they, of course, may tion. In the following excerpt from the case,
be disputed by the parent. Ms. Hoeppner is the school’s special ed-
ucation teacher who had used a whole-
Hearing officers and courts are under- language method with Lucas unsuccessfully.
standably reluctant to be involved in these Ms. Dakin is one of the experts who testi-
disputes over methodology, and in spite of fied about dyslexia.
the emphasis on methodology found in the
1997 IDEA, some courts still refuse to en- [This testimony] does not show that Ms.
tertain the issues of methodology. The court Hoeppner either had actually implemented
in Kugler v. Vance (1999), for example, de- Ms. Dakin’s recommendations or was plan-
clared, “The question of appropriate ning to do so. Ms. Dakin made numerous rec-
methodology for providing related services ommendations, but the recommendation most
to Matthew is precisely the type of dispute at issue here was that Greater Clark imple-
that is inappropriate for judicial resolution” ment a direct teaching reading program using
(p. 751). Even though many courts are re- multisensory, structured, sequential tech-
luctant to address issues of methodology, niques. Ms. Hoeppner did not testify that Lu-
there are times at which courts must, as the cas’ fifth grade IEP included the use of such a
provision of FAPE requires no less. teaching program, or that she was planning to
use such a teaching program. In fact, Lucas’s
A federal district court has recently pro- IEP does not provide for the use of a direct
vided cogent guidance, well grounded in teaching method or any other particular teach-
both law and special education, for deci- ing technique to be used to improve his read-
sion makers faced with evaluating the ap- ing skills. As Ms. Hoeppner explained, be-
propriateness of a school district’s IEP. In cause there are never any specific instructions
Board of Education of the County of in a student’s IEP regarding teaching method-
Kanawha v. Michael M. (2000), “the entire ology or technique, Ms. Hoeppner would de-
dispute rests on the issue of whether the termine what teaching techniques to use with
methodology in the IEPs [italics added] was a particular student by looking at the broad
reasonably tailored to accomplish the goals goals contained in the student’s IEP. There is
set forth in the IEPs” (p. 611). In approach- simply no evidence in the record indicating
ing this dispute, the court laid out the fol- that, if Ms. Hoeppner had had the opportuni-
lowing steps: ty to implement the fifth grade IEP, she was
planning to use a teaching method or tech-
For a school district to sustain its burden of nique different from those she had used unsuc-
proving that its IEP was reasonably calculated cessfully with Lucas for the prior three years.
at the time of creation to provide some educa- (p. 979)
tional benefit, the school district cannot sim-
ply provide conclusory statements that the IEP In fairness to the judicial system, it should
was adequate. The school district must show be said that although many courts have flat-
the following concrete information. First, the ly refused to examine methodology, a few,
school district must show that it set forth the in addition to Kanawha and Nein, have
proper elements of the IEP. . . . braved the waters. Among these is the
Evans (1996) court. Frank Evans was a 15-
Second, the school district must show that year-old with severe dyslexia. The Rhine-
the annual goals, benchmarks, and short-term

66 FOUNDATIONS AND CURRENT PERSPECTIVES

beck School District proposed a program oral test in American History. The court fur-
for Frank that consisted of one-on-one mul- ther awarded $15,000 damages, which the
tisensory instruction in reading and writing history teacher was required to pay to Dou-
for 60 minutes 4 days a week along with en- glas and his parents.
rollment in a special education class for
English; support services in math and sci- Recently, parents of students with LD set-
ence; and modifications in testing, class- tled a class action suit against the state of
work, and homework. Dyslexia experts Oregon. The suit alleged that Oregon’s
agreed that due to the severity of Frank’s statewide system of assessment (OSAS) dis-
LD, he needed an intensive program of indi- criminated against children with LD. In the
vidualized, integrated, multisensory, sequen- settlement, A.S.K. v. Oregon State Board of
tial training in order to receive academic Education (2001), Oregon agreed to adopt
benefit. The court found that “despite her extensive recommendations of a panel of
[Frank’s special education teacher] intensive experts appointed to study Oregon’s assess-
individual instruction eight times per week, ment system. In particular, the panel recom-
and homework and classwork modifica- mended:
tions, Frank’s performance declined” (p.
348). As a result of such evidence and the Accommodations should be considered allow-
testimony of dyslexia experts, the court able, valid, and scorable if they are used dur-
found that the Rhinebeck School District ing instructional or on classroom assessments
had not used a methodology designed to en- and are listed on a student’s IEP until research
sure an appropriate education for Frank. evidence invalidates the score interpretation.
Rather than consider all accommodations first
We can expect to see more cases come be- invalid until proven to be valid, ODE should
fore the courts which require decisions consider all accommodations valid unless and
about methodology for students with LD, until research provides evidence that an ac-
especially those who have severe reading commodation alters the construct or level of
disabilities. It may be that in a field in which the OSAS measure. (Disability Rights Advo-
the controversy about how best to teach cates, 2001, p. 30)
reading has long raged, law and litigation
will actually have as much effect on chang- As more and more states adopt statewide
ing LD practice as has the abundance of re- assessment systems that require students to
search on reading from the last 50 years! pass tests in order to receive a high school
diploma or qualify for college entrance or
ACCOMMODATIONS FOR STUDENTS WITH LD scholarships, this issue of appropriate ac-
commodations is likely to be raised more
The issue of providing accommodations for frequently. We can expect to see more of
students with LD is one that has only infre- these cases in the future.
quently been raised in the courts. In Doe v.
Withers (1993), the parents of a student with Placement and Least
LD claimed that their son’s teachers and Restrictive Environment
school officials had refused to provide the
accommodations required by his IEP. Dou- The IDEA legal requirements related to
glas’s IEP allowed him to take tests orally for placement are remarkably simple. Every eli-
his mainstream classes. This accommodation gible student is entitled to an individualized
had been provided when Douglas was in ele- placement decision based on his or her IEP
mentary and middle school. When Douglas and selected from a full continuum of alter-
entered high school, all his teachers but one native placements. When the student’s edu-
agreed to comply with the oral-testing ac- cation cannot be “achieved satisfactorily”
commodation. Because his history teacher in a regular classroom, another setting is al-
refused to allow him to take tests orally, lowed. According to the federal Office of
Douglas failed his history class. As a result, Special Education Programs (Letter to Tra-
he was barred from participating in ex- han, 1998):
tracurricular activities. The court ordered
the district to provide all necessary tutoring The overriding rule in any placement under
and reteaching to prepare Douglas to take an Part B is that the child’s placement must be in-

Learning Disabilities and the Law 67

dividually determined based on his or her ner that satisfies the government’s purpose.
unique abilities and needs. Recognizing that In the 1960s, institutionalized patients with
regular class placement may not be appropri- mental illness began to sue successfully for
ate for every disabled child, the Part B regula- the right to move into the most “ordinary”
tions require that school districts make avail- programs and facilities (unlocked wards,
able a range of placement options, known as weekend passes into the community, etc.) in
the continuum of alternative placements, to which they were capable of responsibly par-
meet the unique educational needs of students ticipating, rather than being kept in unduly
with disabilities. 34 C.F.R. § 300.551(a). This restrictive and often horribly inhumane
requirement for the continuum reinforces the conditions. This LRE principle would also
importance of the individualized inquiry, not a apply, for example, if a prison or govern-
“one size fits all” approach, in determining ment hospital has to choose between a
what placement is least restrictive for each stu- lobotomy or medication to control a pris-
dent with a disability. (p. 403) oner’s otherwise uncontrollable violent be-
havior.
What Is a “Placement”?
The applicability of this least drastic, in-
In 1977, when IDEA first went into effect, trusive, or harmful doctrine to public school
most special educators thought “place- special and general education classrooms
ment” meant where a program (curriculum, seems forced, at best. Similarly, the proposi-
instruction, and related activities) was deliv- tion that chronological age (rather than
ered. The program was the “what” of the mental age, social age, needs, interests, abil-
child’s education; the placement was the ities or performance level) is the only appro-
“where.” However, it was soon evident that priate basis for grouping students in school
the courts did not necessarily see it the same is a stretch for many.
way.
The term “LRE” is used in at least two
In Concerned Parents v. New York City fundamentally different ways in special edu-
Board of Education (1980), the Second Cir- cation—either as one particular place on the
cuit court held that “the term ‘educational continuum of placements (i.e., a regular
placement’ refers only to the general type of classroom) or as whatever placement maxi-
educational program in which the child is mizes options, functioning levels, or possi-
placed” (p. 753). This, of course, makes it bilities for a given student and allows his or
difficult to interpret the many references in her education to be “achieved satisfactori-
IDEA to identification, evaluation, pro- ly.” Some courts have used the first notion,
gram, and placement. Ordinarily in law, dif- reflexively and without analysis, to mean
ferent words have different referents. The placement in the mainstream of public edu-
end result is some confusion over when and cation. More thoughtful courts have used
where, if at all, a line can be drawn between the second and have sometimes acknowl-
program and placement. edged that the mainstream is not the LRE
for a particular student.
Over time, the prevailing issue in place-
ment has become one of balancing IDEA’s LRE AS MANDATED MAINSTREAMING
requirement for an appropriate education
with the philosophy of including all stu- In Amann v. Stow School System (1992),
dents in regular classes. There have been the court denied the parents’ choice of a pri-
many cases filed by parents of students with vate school placement for a 14-year-old
LD which dispute a district’s choice of with LD because the private school provid-
placement for a student. ed no mainstreaming. The court quoted
Roland M. in saying, “Mainstreaming may
Least Restrictive Environment not be ignored, even to fulfill substantive
educational criteria” (p. 621). The Amann
The concept of least restrictive environment court held this position in spite of the fact
(LRE) in IDEA apparently has its historical that the Massachusetts’s educational benefit
roots in the legal principle that when the standard called for “maximum possible de-
government abridges or restricts life, liber- velopment.” Likewise, the court in Robert
ty, or property, it must do so in the least in- M. v. Hickok (2000) rejected the parents’
trusive, least drastic, or least harmful man-

68 FOUNDATIONS AND CURRENT PERSPECTIVES

choice of a private school placement for streaming and appropriateness of program.
their son Robert based on the erroneous “Mainstreaming which results in total fail-
premise that “federal law requires that chil- ure, where separate teaching would produce
dren like Robert, whose intellectual abilities superior results, is not appropriate and sat-
are only slightly less than those of his peers, isfactory. Congress expressly limited its pre-
be incorporated into regular classroom set- sumption in favor of mainstreaming to cases
tings” (p. 531). where mainstreaming is ‘appropriate’ and
mainstream education can be provided ‘sat-
In a case in which parents and the school isfactorily’” (p. 812).
district had been unable to come to agree-
ment on an IEP for the student in spite of Even when private school placement is
numerous meetings, the Fourth Circuit not at issue, courts must sometimes balance
found in favor of the district because the IDEA’s preference for neighborhood school
parents failed to demonstrate that their pro- placement (34 C.F.R. § 300.552(c)) against
posed placement would be the LRE for the other factors. The court in Greenbush
student (DeLullo v. Jefferson County Board School Committee v. Mr. and Mrs. K.
of Education, 1999). In a similar case, (1996) said, “The default placement for a
Board of Education of the City School Dis- student under the Act is his or her local
trict of the City of New York (1998), the school, however, an IEP can override this
court ruled that parents could not show that default in situations where the student
a private school placement was the LRE for would not receive an educational benefit at
the student. However, the court also ruled the local school” (p. 203). In this case, the
that the district’s proposed program was not court found that the extreme animosity be-
appropriate! The court left unanswered tween the parents of a student with LD and
what placement was appropriate for the stu- the local school staff along with the stu-
dent. The issue of whether LRE properly dent’s “gripping” fear of attending the local
applies to parents’ placements is discussed school were sufficient to prevent the student
later. from receiving educational benefit at the
neighborhood school. The court, therefore,
BALANCING LRE AND EDUCATIONAL BENEFIT ordered the district to place the student at a
different, nearby school in the same district.
A number of courts have looked at the con-
cept of LRE more broadly and have The First Circuit, in Milford School Dis-
weighed educational benefit in balancing trict v. William F. (1997) reminds us that
LRE with appropriate program. Courts placement decisions must be made by a
have considered numerous factors when de- team of people who consider a number of
ciding what constitutes an appropriate factors. “The guidelines for a placement de-
placement. In Egg Harbor (1992), the court cision in New Hampshire law as in federal
said, “Mainstreaming, however, is not the law provide for involving many interested
primary issue in this case. The question be- persons and a wide variety of factors in the
fore us is the appropriateness of the educa- choice among alternative potential place-
tional program designed for S. by Egg Har- ments, and the law does not specify that any
bor” (p. 18). The court went on to quote one factor or any one person’s opinion must
the Burlington case: “The least restrictive be given decisive weight” (p. 26).
environment guarantee . . . cannot be ap-
plied to cure an otherwise inappropriate Even the fact that a private school is reli-
placement” (Burlington School Committee giously affiliated and not a special educa-
v. Massachusetts Department of Education, tion school does not automatically preclude
471 U.S. 359 [1985]) (p. 789). Following that school from being the LRE for a stu-
this reasoning, the Egg Harbor court ruled dent. In Matthew J. v. Massachusetts De-
that Landmark School, a private school for partment of Education (1998), the court
students with LD, was an appropriate place- ruled that the Master’s School, a private,
ment for S. and required the Egg Harbor non-special education, college-preparatory
district to pay for the placement. Christian school, appropriately addressed
Matthew’s need for no aggressive peers and
In Capistrano (1995), the Ninth Circuit a structured and supportive environment
also ruled for a balance between main- and was, therefore, the LRE.

Two recent cases delineate the critical im-

Learning Disabilities and the Law 69

portance of weighing educational benefit does and without any analysis have ruled
against mainstreaming in determining the against the parents. However, in Florence
LRE. In both cases, the district involved had County School District Four v. Carter
failed to provide an appropriate education (1993), the U.S. Supreme Court ruled:
for a student with LD, and, as a result, the
student had made negligible academic There is no doubt that Congress has imposed
progress. In Nein (2000), the court stated: a significant financial burden on States and
“Mainstreaming is not required in every school districts that participate in IDEA. Yet
case. . . . [I]t must be determined whether public educational authorities who want to
the child is benefitting educationally from avoid reimbursing parents for the private edu-
mainstreaming. The evidence in the record cation of a disabled child can do one of two
here demonstrates that, for three years, Lu- things: give the child a free appropriate public
cas did not benefit educationally from education in a public setting, or place the child
Greater Clark’s educational plans. . . . Thus, in an appropriate private setting of the State’s
there is a good reason in this case to dis- choice. This is IDEA’s mandate, and school of-
count Greater Clark’s reliance on the IDEA’s ficials who conform to it need not worry
‘strong preference’ for mainstreaming” (p. about reimbursement claims. (p. 15)
981).
The few courts that have looked at the
The court in R.R. (2001) ordered an out- question carefully since then have conclud-
of-district placement at a private LD school ed that when the public sector fails to offer
for an eighth-grade student who read at a an appropriate program, LRE does not ap-
third- or fourth-grade level in spite of the ply to bar reimbursement for a private
district’s argument that the student would placement, even if it is more restrictive. In
be deprived of elective courses and social- Cleveland Heights-University Heights City
ization with nondisabled peers. The court School District v. Boss (1998), the Sixth
ruled, “On balance, we find the opportuni- Circuit pointed out the fallacy of expecting
ties the student will miss in the Board’s pro- parents to pay for a private school educa-
gram pale beside his need for ‘an intensive tion for their child when the district failed
and unique program’ in order to remedy his to provide an appropriate program.
learning disability” (p. 124).
The District would have us read the IDEA to
Issues of LRE continue to be raised in cas- say, in effect: “If we fail to provide a disabled
es in which parents seek reimbursement for child with an appropriate education, the par-
private school placements, even if those ents must pay for a private education, or let
placements are residential, for their children their child languish in our institution if the
with LD because the parents believe that only placement more suitable to her needs and
district programs are inappropriate. more closely approximating the ideal envi-
sioned by the IDEA than what we offer is a
REIMBURSEMENT AS AN LRE ISSUE specialized private school that admits only
learning disabled students.” Congress did not
Many LD placement cases arise when the intend to place beneficiaries of the IDEA in the
parents believe the public school has failed position of having to choose only among these
to offer FAPE by not providing an appropri- unpalatable alternatives. (p. 400)
ate instructional methodology and/or by not
employing trained and experienced LD The parents of Raelee, a ninth-grader
teachers. Typically, the parents remove the with LD, withdrew her from public school,
student to a private LD school and then placed her in a private school, and request-
seek reimbursement. Analytically, these cas- ed reimbursement from the school district.
es pose two questions: Did the district pro- The court, in Susquenita School District v.
vide FAPE, and if not, is the private place- Raelee (1996), ruled: “Although the Janus
ment proper under IDEA? However, a School did not provide Raelee the least re-
critical question is embedded in whether the strictive setting possible, it was an appropri-
private placement is proper, that is, whether ate placement in light of her educational
IDEA’s LRE preference applies to the par- needs and in view of the fact that Susqueni-
ents’ chosen placement. To date, the majori- ta failed to offer an appropriate placement
ty of courts have simply assumed that it in a less restrictive setting” (p. 127).

70 FOUNDATIONS AND CURRENT PERSPECTIVES

Courts have used similar reasoning to or- a fourth-grade student requested a residen-
der reimbursement to parents for residential tial placement for their daughter after she
placements when the residential placement spent some time in a psychiatric hospital.
provided the only appropriate program. In The court found that “Many of the stu-
Lascari (1989), the court said: dent’s difficulties were related to her family
relationships and were more adequately
We are sensitive to the possibility that parents treated through counseling rather than re-
may select a private school that affords their moval to a residential facility” (p. 879). The
child an education that is more elaborate than court ruled that the district had offered an
is required. Conceivably, parents might select appropriate program that was less restric-
a boarding school even though a day program tive than the residential placement the par-
would furnish their child with an appropriate ents requested.
education. It would be anomalous, however,
to recognize the parents’ right to reimburse- In Walczak v. Florida Union Free School
ment, but to deny completely that right merely District (1998), the district proposed an IEP
because they selected a school that furnished for a student with a learning disability
an education beyond that which the district is which called for her placement in a self-
obliged to offer. It would also be anomalous to contained class for developmentally dis-
deny parents the right to reimbursement when abled students. The Walczaks objected to
the district failed to provide their child with an this placement and argued that their daugh-
appropriate education and the only school ter’s needs could not be met in a day pro-
that the parents could find was a boarding gram and that she required a residential
school. (p. 572) placement. The Walczaks also objected to
the size (12 students) and composition (stu-
In a case in which the parents of a student dents with developmental disabilities) of the
with a serious learning disability and a se- district’s proposed class. The court found
vere language disorder placed her in a pri- that “a clear preponderence of that evidence
vate, residential school, the court found that demonstrates that B.W. could make satisfac-
“no non-residential alternatives to Maple- tory academic and social progress in a
brook School were proposed by the District, twelve-student class in the BOCES day pro-
nor were any such programs known to ex- gram” (p. 1143). The court, therefore, de-
ist” (Briere v. Fair Haven Grade School Dis- nied the parents’ request for a residential
trict, 1996, p. 63). placement.

However, a parent will not be entitled to As with most IDEA issues addressed by
reimbursement for a residential placement the courts, the question of what is the LRE
when a court finds that the district offered for a particular student with a learning dis-
an appropriate program. In some courts, ability is complex and requires the balanc-
parents may also be denied reimbursement ing of many factors.
if the court finds that the child did not need
a residential placement to benefit education- Conclusions and Further Thoughts
ally. In Lenn v. Portland School Committee
(1992), the parents of Daniel, a 17-year-old The statutory and case law that addresses
student with a learning disability, argued services for students with LD is vast and
that their son needed a residential place- sometimes conflicting. How has the field of
ment after being hospitalized for depres- LD been affected? Our response to this
sion. The court ruled that the district’s pro- question is not based solely on the litigation
posed program was reasonably calculated and legislation we have reviewed in this
to be of significant educational benefit to chapter but also on our experiences as
Daniel. The court noted that the district scholars and practitioners in the field of
program “would address Daniel’s needs LD.
for specialized education . . . while enabling
him to remain in his home community and Arguably, legislation and litigation have
interact daily with non-disabled peers” (p. had a detrimental affect on the practices of
617). evaluating and determining the eligibility of
students suspected of having a learning dis-
In a similar case, Salley v. St. Tammany ability. Many students who are probably
Parish School Board (1995), the parents of

Learning Disabilities and the Law 71

not learning disabled are receiving services fought promise of IDEA will be empty, indeed.
under that label, and some students who (p. 389)
should receive services are not. The IDEA
definition of learning disability has lead to In making decisions about placement for
widespread misuse of standardized tests and students with LD, districts have relied on
discrepancy formulas. Even now that defini- the concept of “least restrictive environ-
tion and the methods used to identify stu- ment” to eliminate many specialized classes
dents with LD are being debated by nation- for students with LD. In what sense, one
ally recognized researchers in the field of might ask, is it “less restrictive” for a stu-
LD (Council for Exceptional Children, dent to be with only nondisabled peers who
2001). Regardless of whether the definition do not and cannot share his or her perspec-
of LD is changed in the statute, a need will tives and experiences?
remain for district professionals who are
truly knowledgeable about LD. School per- Similarly, what is the underlying message
sonnel must begin to used broader-based as- when a student with LD is told that the
sessment procedures which go beyond sim- only peers who are suitable are those who
ply comparing two scores from the do not have LD? Is it really healthier for
WISC-III (Wechsler, 1991) and the Wood- students with LD to be in regular class-
cock-Johnson III (Woodcock, McGrew, & rooms where their disability puts them visi-
Mather, 2001). There must be a move to- bly and publicly on the bottom rung of
ward (or perhaps back to) greater reliance dozens of daily ladders rather than being
on teacher input, students’ work products, with other students with LD where they oc-
observations, and other strategies to deter- cupy all the rungs of the daily ladders, from
mine whether or not students are achieving top to bottom and in between? Little, if
at appropriate levels. any, consideration is given by some districts
to the self-esteem issues raised for students
Special educators and administrators with LD placed all day in mainstream class-
must be trained to recognize when method- es.
ology constitutes an integral part of a stu-
dent’s IEP. The new legal recognition that Some courts, at least, have recognized
methodology may sometimes be required to that program effectiveness as measured by
be included on IEPs, especially for students student progress is at least as important as
with LD, is potentially important and posi- mainstreaming for students with LD. We
tive. The inclusion of methodology on IEPs firmly believe that it does students with LD
could be a major impetus for improved in- little good to be mainstreamed and social-
structional practice for many students with ized in regular education classrooms for 12
LD. For this to happen, special educators years if the result is that those students
and administrators must attend to the re- leave high school reading at a second- or
search on effective teaching methodology third-grade level and with serious self-
for students with LD, and they must be con- esteem issues. Although the original intent
vinced that it is critical to implement such of IDEA may have been to ensure access
effective teaching procedures in special edu- to public schools for students with disabili-
cation programs and regular education ties, the current IDEA regulations make
classrooms which serve these students. Spe- it clear that the appropriateness of a stu-
cial educators must once again become dent’s special education program must be
teachers of “specialized” instruction. As judged, in large part, by the progress that
Zigmond (1997) has said: student makes toward his or her education-
al goals.
Special education was once worth receiving; it
could be again. In many schools, it is not now. We in the field of learning disabilities
Here is where practitioners, policymakers, ad- have learned much in the 40 years since
vocates, and researchers in special education Samuel Kirk (1962) coined the term “learn-
need to focus—on defining the nature of spe- ing disability.” We now know how best to
cial education and the competencies of the teach students with LD so that they learn
teachers who will deliver it. Here is where the the skills and content that their nondisabled
research-to-practice gulf must be bridged. peers learn. The law provides an avenue for
Here is the issue we must resolve, or the hard- ensuring that students with LD benefit from
this knowledge if we as practitioners follow

72 FOUNDATIONS AND CURRENT PERSPECTIVES

the spirit of the law and provide truly indi- Guckenberger v. Boston University, 974 F. Supp.
vidualized, special education to students 106 (D. Mass. 1997).
with LD.
Hall v. Vance County Board of Education, 774 F.2d
References 629 (4th Cir. 1985).

Amann v. Stow School System, 982 F.2d 644 (1st Hendrick Hudson Central School District Board of
Cir. 1992). Education v. Rowley, 458 U.S. 176 (1982).

Americans with Disability Act, 42 U.S.C. §§ 12101 Hiller v. Board of Education of the Brunswick Cen-
et seq. (1994). tral School District, 743 F. Supp. 958 (N.D.N.Y.
1990).
A.S.K. v. Oregon State Board of Education (Feb.,
2001). Honig v. Doe, 484 U.S. 305 (1988).
Individuals with Disabilities Education Act, Pub. L.
Board of Education of the City School District of
the City of New York, 30 IDELR 64 (Review Of- 105–17, 111 Stat. 37 (1997) (codified at 20
ficer Decision, 1998). U.S.C. §§ 1499–1487).
Judith S. v. Board of Education of Community. Unit
Board of Education of the County of Kanawha v. School District No. 200, 28 IDELR 728 (N.D. Ill.
Michael M., 95 F. Supp. 2d 600 (S.D. W. Va. 1998).
2000). Kelby v. Morgan Hill Unified School District, 18
IDELR 831 (9th Cir. 1992).
Briere v. Fair Haven Grade School District, 948 F. Kevin T. v. Elmhurst Community School District
Supp. 1242 (D. Vt. 1996). No. 205, 34 IDELR 202 (N.D. Ill. 2001).
Kirk, S. A. (1962). Educating exceptional children.
Burlington School Committee v. Massachusetts De- Boston: Houghton Mifflin.
partment of Education, 471 U.S. 359 (1985). Kugler v. Vance, 30 IDELR 749 (D. Md. 1999).
Lake, S. (2000). IEP procedural errors: Lessons
Capistrano Unified School District v. Wartenberg, learned, mistakes to avoid. Horsham, PA: LRP
59 F.3d 884 (9th Cir. 1995). Publications.
Lascari v. Ramapo Indian Hills Regional High
Cleveland Heights-University Heights City School School District, 560 A.2d 1180 (N.J. 1989).
District v. Boss, 144 F.3d 391 (6th Cir. 1998). Lenn v. Portland School Committee, 19 IDELR 615
(D. Me. 1992).
Concerned Parents of New York City Board of Edu- Letter to Trahan, 30 IDELR 403 (Sept. 3, 1998).
cation, 629 F.2d 751 (2d Cir. 1980). Livingston v. Desoto County School District, 782 F.
Supp. 1173 (N.D. Miss. 1992).
Corchado v. Board of Education of Rochester City Matthew J. v. Massachusetts Department of Educa-
School District, 86 F. Supp. 2d 168 (W.D.N.Y. tion, 989 F. Supp. 380 (D. Mass. 1998).
2000). Milford School District v. William F., 129 F.3d 1252
(1st Cir. 1997).
Council for Exceptional Children. (2001). The con- Mills v. D.C. Board of Education, 348 F. Supp. 866
troversy over learning disabilities continues. To- (D.D.C. 1972).
day, 8(4), 1, 5, 15. Nein v. Greater Clark County School Corp., 95 F.
Supp. 2d 961 (S.D. Ind. 2000).
Delaware County Intermediate Unit No. 25 v. Mar- Norton v. Orinda Union School District, 168 F.3d
tin K., 831 F. Supp. 1206 (E.D. Pa. 1993). 500 (9th Cir. 1999).
Pascoe v. Washingtonville Central School District,
DeLullo v. Jefferson County Board of Education, 29 IDELR 31 (S.D.N.Y. 1998).
194 F.3d 1304 (4th Cir. 1999). Pennsylvania Association for Retarded Children
(PARC) v. Pennsylvania, 343 F. Supp. 279 (E.D.
Disability Rights Advocates. (2001). Do no harm— Pa. 1972).
High stakes testing and students with learning Ridgewood Board of Education v. N.E., 172 F.3d
disabilities. Oakland, CA: Author. 238 (3rd Cir. 1999).
Robert M. v. Hickok, 32 IDELR 169 (E.D. Pa.
Doe v. Defendant I., 898 F.2d 1186 (6th Cir. 1990). 2000).
Doe v. Withers, 20 IDELR 422 (W. Va. Cir. Ct. R.R. v. Wallingford Board of Education, 35 IDELR
32 (D. Conn. 2001).
1993). Salley v. St. Tammany Parish School Board, 57 F.3d
Education for All Handicapped Children Act, 20 458 (5th Cir. 1995).
Stanovich, K. E. (1999). The sociopsychometrics of
U.S.C. § 1400 et seq. (1975). learning disabilities. Journal of Learning Disabil-
Egg Harbor Township Board of Education v. S.O., ities, 32, 350–361.
Susquenita School District v. Raelee, 96 F.3d 78 (3d
19 IDELR 15 (D.N.J. 1992). Cir. 1996).
Evans v. Board of Education of the Rhinebeck Cen- Vocational Rehabilitation Act of 1973, Pub. L. 93-
112, 87 Stat. 394.
tral School District, 930 F. Supp. 83 (S.D.N.Y.
1996).
Florence County School District Four v. Carter, 510
U.S. 7 (1993).
Flowers v. Martinez Unified School District, 19
IDELR 898 (N.D. Cal. 1993).
Fort Zumwalt School District v. Clynes, 119 F.3d
607 (8th Cir. 1997).
Friedman v. Vance, Montgomery County Board of
Education, 24 IDELR 654 (D. Md. 1996).
Greenbush School Committee v. Mr. and Mrs. K.,
949 F. Supp. 934 (D. Me. 1996).

Learning Disabilities and the Law 73

Walczak v. Florida Union Free School District, 142 N. (2001). Woodcock-Johnson III—Tests of
F.3d 119 (2d Cir. 1998). Achievement. Itasca, IL: Riverside.
Zigmond, N. (1997). Educating students with dis-
Wechsler, D. (1991). Wechsler Intelligence Scale for abilities: The future of special education. In J. W.
Children—Third Edition. San Antonio, TX: The Lloyd, E. J. Kameenui, & D. Chard (Eds.), Issues
Psychological Corporation. in education students with disabilities (pp.
377–389). Mahwah, NJ: Erlbaum.
Welton v. Liberty 53 School District, 35 IDELR 63
(W.D. Mo. 2001).

Woodcock, R. W., McGrew, K. S., & Mather,

APPENDIX 4.1. Matrix of Refe

Case namea Evaluation
and

eligibility Procedur

Amann v. Stow School System (1992)

A.S.K. v. Or. State Board of Education (February 2001)

Board of Education of the City School District of the
City of New York (1998)

Board of Education of the County of Kanawha v.
Michael M. (2000)

Briere v. Fair Haven Grade School District (1996) ț
Capistrano Unified School District v. Wartenberg (1995)

Cleveland Heights-University Heights City School District
v. Boss (1998)

Concerned Parents of New York City Board of
Education (1980)

Corchado v. Board of Education Rochester City School ț
District (2000)

DeLullo v. Jefferson County Board of Education (1999)

Doe v. Defendant I. (1990) țț

Doe v. Withers (1993)

Egg Harbor Township Board of Education v. S.O. (1992) ț
ț
Evans v. Board of Education of the Rhinebeck Central
School District (1996)

Flowers v. Martinez Unified School District (1993) ț
Fort Zumwalt School District v. Clynes (1997)

Friedman v. Vance, Montgomery County Board of ț
Education (1996)

Greenbush School Committee v. Mr. and Mrs. K. (1996)

Hall v. Vance County Board of Education (1985) ț

erenced IDEA Cases and Issues LRE
IEPs/FAPE

How much Methodology Accommodations Placement Reimbursement
ral benefit
țț
ț

țț

ț

țț

ț țț
ț

ț țț
țț țț
ț
ț țț
ț țț

ț țț
țț

țț
țț

Hendrick Hudson Central School District Board of ț ț
Education v. Rowley (1982) ț
Hiller v. Board of Education of the Brunswick Central ț
School District (1990) ț ț
Judith S. v. Board of Education of Community Unit ț
School District No. 200 (1998) ț ț
Kathleen H. v. Massachusetts Board of Education (1998)
Kelby v. Morgan Hill Unified School District (1992)
Kevin T. v. Elmhurst Community School District
No. 205 (2001)
Kugler v. Vance (1999)
Lascari v. Ramapo Indian Hills Regional High School
District (1989)
Lenn v. Portland School Committee (1992)
Livingston v. Desoto County School District (1992)
Matthew J. v. Massachusetts Department of Education
(1998)
Milford School District v. William F. (1997)
Nein v. Greater Clark County School Corp. (2000)
Norton v. Orinda Union School District (1999)
Pascoe v. Washingtonville Central School District (1998)
Ridgewood Board of Education v. N.E. (1999)
R.R. v. Wallingford Board of Education (2001)
Robert M. v. Hickok (2000)
Salley v. St. Tammany Parish School Board (1995)
Susquenita School District v. Raelee (1996)
Walczak v. Florida Union Free School District (1998)
Welton v. Liberty 53 School District (2001)

aFor full citation, see reference list at the end of the chapter.

ț țț
ț
țț
țț țț
ț
țț
țț țț
ț țț
ț
țț
țț

țț
țț
țț
țț
țț
țț
ț

5

Learning Disability as a Discipline

Kenneth A. Kavale
Steven R. Forness

At a fundamental level, a discipline is de- question remains: How far has the LD disci-
fined as a branch of learning. All disciplines pline progressed?
possess the primary goal of providing a
comprehensive understanding of a particu- Historical Foundations
lar phenomenon. Given the formal defini-
tion, learning disability (LD) would appear Origins
to qualify as a legitimate discipline but one
far from achieving its primary goal. The rea- Wiederholt (1974) provided a useful history
son is found in the fact that, at present, we of LD that demonstrated its phased devel-
appear to “know” far more than we “un- opment from several types of disorders,
derstand” about LD. The consequences of most notably language, reading, and cogni-
this limited understanding are found in the tive process problems. From the seminal
inability to answer a seemingly facile ques- study of exogenous mental retardation by
tion: What is a learning disability? A disci- Strauss and Werner (see Strauss & Lehtinen,
pline should be able to define itself without 1947) to the rousing endorsement of Kirk’s
ambiguity. (1963) term “learning disabilities,” LD was
viewed as a neurologically based disorder
A discipline should also demonstrate the manifested by unique processing distur-
quality of continually evolving into a more bances that selectively interfered with ac-
inclusive and structured domain. The LD quiring and assimilating basic academic in-
discipline did not spring full grown from the formation.
brow of Kirk, Cruickshank, Kephart, and
others. The LD discipline has evolved but The LD discipline also developed along
not in any straight-line form making for social dimensions. Kavale and Forness
easy progression. Because LD developed un- (1995) showed how, within the structure of
der the stress of practical exigencies, it special education during the 1960s, there
shows gaps and bypaths as well as unfound- was a compelling need for a category such
ed leaps of faith. Therefore, an essential

76

Learning Disability as a Discipline 77

as LD. From misclassification of students in to the development of essentially divergent
categorical special classes to lack of school- disciplinary perspectives. On one side, there
based services or educationally focused in- is LD as a scientific discipline whose goal
terventions for particular problems, a new was to predict and to explain LD. On the
category such as LD could resolve diagnos- other side, there is LD as a political disci-
tic predicaments and provide needed ser- pline whose goal is advocacy, policy direct-
vices. ed at creating programs and services to
meet the needs and interests of students
The Problem of Definition with LD. Because these two disciplinary
perspectives require different interpreta-
By the 1970s, the LD discipline was experi- tions of the LD definition, there was little
encing problems with the most enduring be- association between the goals and objec-
ing the lack of consensus about definition tives of the scientific and the political LD
(Doris, 1993). The origins of the LD defini- disciplines.
tion are found in the National Advisory
Committee on Handicapped Children The Scientific Discipline
(NACHC) (1968) report, but the present In-
dividuals with Disabilities Education Act The Strauss and Werner Paradigm
(IDEA) definition has not in fact changed
substantially from the original NACHC def- The scientific discipline of LD can be traced
inition. This definition, however, lacked to investigations of brain function and dys-
precision then and now with its inherent function where Goldstein (1939) demon-
ambiguity resulting in widely varying inter- strated that brain injury rarely caused spe-
pretation. cific disorders but rather usually included a
variety of perceptual, cognitive, and behav-
In a later survey, Tucker, Stevens, and Ys- ioral disturbances that formed a syndrome.
seldyke (1983) found consensus among “ex- Werner and Strauss (1940) continued Gold-
perts” about the viability of the LD catego- stein’s research program and their findings
ry but considerable variability of opinion established the rudiments of the LD concept
about almost all other issues. Critiques of (see Strauss & Lehtinen, 1947). These ideas
the LD definition became so pervasive (e.g., were reinforced by Clements’s (1966) report
Reger, 1979; Siegel, 1968) that the founda- on minimal brain dysfunction (MBD) where
tion was laid for discussions about whether “specific learning disabilities” were one of
LD really existed as a discrete entity. Kavale the 10 most frequently agreed upon charac-
and Forness (2000) analyzed available LD teristics.
definitions and concluded that, “LD has not
been defined with much exactitude . . . [the The MBD label did not receive general
definition] provides only a generalized pic- acceptance until Kirk’s (1963) suggestion
ture of a portion of the school population that the term “learning disability” might
experiencing academic difficulties . . . [but] better focus on educational problems; might
accord about definition does not imply uni- avoid medical implications; and might be
form interpretation, and any variation is better accepted by parents, teachers, and
likely to prevent precision in describing the students. The LD concept was thus based
nature of LD” (p. 245). primarily on the Strauss and Werner para-
digm: (1) LD is associated with or caused by
The fundamental “problem of definition” neurological dysfunction; (2) LD academic
adversely affected the LD discipline. The problems are related to process disturbance,
LD definition belongs to the class of defini- most notably in perceptual–motor function-
tion termed “stipulative,” which possesses ing; and (3) LD is associated with academic
the quality of not needing to be true, only failure defined by discrepancy notions.
useful, and “as long as there is consensus
and a perceived heuristic value, the defini- However, the evidence did not support
tion is accepted and used” (Kavale & For- the foundation of the Strauss and Werner
ness, 2000, p. 247). A stipulative definition paradigm because there were few useful
also need not be used for a common pur- group distinctions of a magnitude “that
pose, and varying interpretations may lead