Handbook of psychology volume 7 educational psychology

470

Learning Disabilities

naming, and morphophonemic knowledge). The language

skills closest to normal involved the higher levels of seman-

tic knowledge. The language test profiles of the three types

were not as sharply differentiated as were their reading test

profiles.

A battery of 37 neuropsychological tests was administered

to determine the extent to which neuropsychological deficits

interacted with each type of reading disability and to estimate

the types of brain dysfunctions potentially associated with

the reading disabilities. Factor analysis of the results did not

differentiate any characteristic profiles of neuropsychologi-

cal deficit. In addition, there was very little indication of any

difference in relative incidence, localization, or severity of

cerebral dysfunction among the three types of reading

disabilities.

This study suffers from many of the same definitional

problems as most other subtype studies do. A lag of approxi-

mately 2 years on reading measures defines the reading dis-

ability, and the study did not account for the fact that 2 years

below grade level means very different things for a child in

Grade 3 as opposed to a child in Grade 8. In this study, the age

of the subjects sampled was extremely diverse (8–27 years),

which means that there were enormous variations in the

severity of the reading disability. It is hardly surprising that

qualitative and quantitative differences in the reading perfor-

mances of such a would emerge, but whether it makes sense

to speak of subtypes rather than a range of performance de-

pending on severity of deficits is questionable. The fact that

normal readers scored below average in many related mea-

sures and individuals with reading problems scored in the

normal range for some skills further confuses the interpreta-

tion. There seems to be a problem with the sensitivity of some

of the measures used or alternatively with their relevance to a

reading deficit. The volume of results at which one arrives

after all these tests are administered is formidable but at the

same time very confusing. It would probably make far more

sense to break down the reading process itself into crucial

areas and see how the participants performed in each area—

for example, oral reading in context, word recognition, word

attack, and silent reading. The findings from a more focused

and detailed analysis of fewer measures might be more mean-

ingful. The fact that 30–60% of the participants fell within

the normal range on the oral reading tests makes one really

wonder about the definition of reading disability in this study.

Furthermore, because a number of participants were clas-

sified into more than one type and had profiles characteristic

of combined types, separation into three distinct groups be-

comes even more questionable. It seems that if one tested

enough individuals (including normally functioning individ-

uals) on enough measures, one would find enormous ranges

of performances in both quality and quantity; it is question-

able, however, how worthwhile it is to define them as differ-

ent subtypes of a reading disability. The authors suggest that

there seems to be a form of continuity between types, and I

think that that is exactly what this study is all about. Most

probably the authors have taken an extremely heterogeneous

group of children in terms of reading performance and have

emerged with three groups who are at different stages of the

reading performance continuum. These styles of performing

cannot be considered subtypes; rather, they represent differ-

ent degrees of severity of a problem.

As with the language tests, the neuropsychological tests

only reflected the heterogeneity of the group or in fact perhaps

any reading-disabled group. With a more narrowed selection

of measures and a detailed error analysis of actual reading per-

formance (rather than a list of scores on so many measures),

more meaningful subtypes might emerge. This study illus-

trates how inherent heterogeneity of the population under dis-

cussion affects subtypes. Doehring et al. found a significant

overlap between reading-disabled and normal individuals on

a variety of measures that tested the child’s understanding

of basic syntax using similar but not identical measures.

There was also a great deal of variability within each of the

subtypes that they identified. Siegel and Ryan (1984) have

found virtually no overlap between reading-disabled and nor-

mally achieving children but found that the performance of

the reading-disabled children was very significantly below

that of normally achieving children of the same chronological

age. Siegel and Ryan’s subjects were homogeneous and se-

lected on the basis of low scores on a word-reading test.

Rutter, Yule, and associates proposed the existence of two

subtypes of reading disorders. They used a regression equa-

tion in which the expected reading level was predicted from

a child’s IQ score (e.g., Rutter, 1978; Rutter, Tizard, &

Whitmore, 1970; Rutter & Yule, 1973; Yule, 1967; Yule,

Rutter, Berger, & Thompson, 1974). According to these au-

thors, if the child’s reading level is significantly lower than

that predicted from his or her IQ, then the child is said to be a

retarded reader. They differentiate between backward readers

(children who are at the bottom end of the reading attainment

continuum) and retarded readers (children who are under-

achieving in relation to their chronological age and general

level of intelligence) in sex distribution, neurological corre-

lates, and association with speech and language disorders.

It should be noted that there was considerable overlap be-

tween the groups. Retarded readers who had (by definition)

higher IQs were worse in their accuracy but not in compre-

hension and were worse in spelling but had higher scores

in mathematics. We do not believe that these constitute

meaningful subgroups. It is important to note that language

Subtypes

471

measures (WISC) constituted part of the definition of IQ.

Children with learning disabilities have language problems

(Siegel, 1985a, 1985b; Siegel & Ryan, 1984; Vellutino, 1977,

1978, 1979); therefore, children will have lower IQ scores

when a language-based test is used as a measure of IQ. It is

impossible to measure IQ independently of language and

reading skills. A definition in which subtypes are developed

on the basis of IQ-reading level discrepancies therefore

seems to be invalid (see previous discussion). The distinction

between reading retardation and reading backwardness may

represent a difference of severity in that children who are

backward in reading have more cognitive deficits than do

those who are retarded in reading (and hence have lower IQ

scores).

Although Doehring (1984) notes that most subtype classi-

fications of reading problems involve a visual nonverbal sub-

type, I contend that the appearance of this subtype is an

artifact of the tests of reading used. When a subtype such as

this one with visual nonverbal problems is used, I contend

that they also have language problems, that they are the chil-

dren with good phonics skills but low comprehension scores,

or that their word recognition skills are adequate but because

of attention difficulties they have problems with memory,

speed, attention, strategies in relation to reading, or any com-

bination of these problems.

Therefore, conclusive and convincing evidence of sub-

types of reading disability has not emerged. In addition, there

is homogeneity within the reading-disabled population. Ap-

parent heterogeneity is a function of the definition used. As

Doehring suggests, “The one simplifying assumption that I

will continue to make for the present in my own work how-

ever is that the most profound reading disabilities involve

difficulty in acquiring lower level coding and word recogni-

tion skills rather than higher-level skills and strategies”

(Doehring, 1984, p. 211). If investigators use a word-

recognition definition, phonics (nonword) definition, or both

of reading disability, there will not be a significant amount of

variability on relevant cognitive functions within the reading-

disabled population.

Reading disability involves a problem with phonological

processing, language, and memory for verbal information.

Reading-disabled children can be differentiated from other

LD children on these variables. Visuospatial and perceptual-

motor problems may also occur in dyslexia, but they are not

the basic problem, nor are they characteristic of all dyslexic

individuals in the same way in which language problems

are.

If a logically consistent definition of dyslexia is used, all

dyslexic (reading disabled) children have problems with

language.

A Simple Model of Subtypes

What emerges from this confusing array of studies is the fact

that there are clearly two subtypes of learning disability—

namely, a reading disability (dyslexia) and an arithmetic dis-

ability. These subtypes have been validated in child and adult

populations (e.g. Rourke & Strang, 1978; Shafrir & Siegel,

1994b; Siegel & Ryan, 1988, 1989a, 1989b). Shafrir and

Siegel (1994b) compared three groups—individuals with

arithmetic disability (AD), reading disability (RD), and both

reading and arithmetic disabilities (RAD), with a comparison

group with normal achievement (NA)—on a variety of cog-

nitive and achievement measures. The main findings were as

follows: (a) Each of the groups differed significantly from the

others on tests of reading, spelling, memory, and other cogni-

tive measures; (b) both the RD and RAD groups showed a

deficit in phonological processing, vocabulary, spelling, and

short-term memory; (c) the AD group performed similarly to

the NA group on pseudoword reading and phonological pro-

cessing but performed more poorly than did the NA group on

word reading and vocabulary; (d) on many tasks, the RAD

group performed more poorly than did the other groups; and

(e) the AD and RAD groups performed more poorly than did

the NA and RD groups on a visuospatial task. Therefore, this

classification scheme for the subtyping of learning disabili-

ties in adults appears to have validity.

Spelling: A Digression

Problems with spelling do exist, but they can co-occur with

either reading or arithmetic difficulties, and it is rare to find a

child who has difficulties with spelling and no problems in

any other areas of functioning. Some children may have spe-

cific difficulties with spelling when they are required to write

the word from memory rather than when they are required to

recognize the correct spelling of a word.

In addition, English words are characterized by both regu-

lar spelling (e.g., singing, print)—that is, words in which the

letter-sound correspondences are predictable—and irregular

spelling—that is, spelling that is not predicted from the rules

of spelling-sound correspondence (e.g., island, knight). The

possibility exists that children may be able to spell regular

words but have more difficulty with the irregular words. In

any case, spelling difficulties do occur but (as noted earlier)

usually in combination with other problems. Furthermore,

studies such as Jorm (1981), and Lennox and Siegel (1993,

1996) have found significant differences in the understanding

of letter-sound correspondence rules and the orthographic

awareness between children who were poor readers and

spellers and children who were only poor spellers. On the

472

Learning Disabilities

basis of findings such as these and for the reasons discussed

previously, reading and spelling need to be treated as separate

variables.

TYPES OF LEARNING DISABILITIES

Over the past 30 years, it has become clear that there are two

major clusters of learning difficulties. The most commonly

known is a reading disability, sometimes called dyslexia.

There is no difference in meaning between the terms dyslexia

and reading disability. Another equally prevalent but less

commonly known disability is an arithmetic (mathematics)

disability, sometimes called nonverbal learning disability,

developmental output failure, writing-arithmetic disability, or

visual-spatial disability. Although there is, admittedly, some

heterogeneity within the two major clusters, they do share

enough common characteristics to be considered as specific

entities.

Reading Disabilities

Depending on the theoretical bias of the particular investiga-

tor, the country, the circumstances, and so on, the word

dyslexia may be used instead of reading disability. However,

there is no difference between dyslexia and a reading disabil-

ity; they are exactly the same.

Dyslexia involves difficulties with phonological process-

ing, including such abilities as knowing the relationship be-

tween letters and sounds and phonological awareness—that

is, the ability to segment the speech stream into separate ele-

ments. Over the years, a consensus has emerged that one core

deficit in dyslexia is a severe difficulty with phonological

processing (e.g., Rack, Snowling, & Olson, 1992; Siegel,

1993b; Snowling, 1980; Stanovich, 1988a, 1988b). Children

with a reading disability have a core deficit in phonological

processing. The evidence that is available clearly demon-

strates that adults with dyslexia have deficits in phonological

processing (e.g., Bruck, 1990, 1992; Elbro, Neilsen, &

Petersen, 1994; Felton, Naylor, & Wood, 1990; Gottardo,

Siegel, & Stanovich, 1997; Greenberg, Ehri, & Perin, 1997;

Pennington, Van Orden, Smith, Green, & Haith, 1990; Pratt &

Brady, 1988; Read & Ruyter, 1985; Russell, 1982;

Scarborough, 1984; Shafrir & Siegel, 1994a, 1994b; Siegel,

1998). Most individuals with dyslexia show problems in the

area of memory and language (Siegel & Ryan, 1984, 1988;

Snowling, 1980; Stanovich, 1988a, 1988b; Vellutino, 1978).

Usually individuals with dyslexia have spelling problems,

but the presence of spelling difficulties without reading diffi-

culties does not indicate dyslexia. A definition of dyslexia

that captures the other problems that often co-occur with it is

illustrated in Padget et al. (1996):

Dyslexia is a language-based learning disorder that is biological

in origin and primarily interferes with the acquisition of print lit-

eracy (reading, writing, and spelling). Dyslexia is characterized

by poor decoding and spelling abilities as well as deficit in

phonological awareness and/or phonological manipulation.

These primary characteristics may co-occur with spoken lan-

guage difficulties and deficits in short-term memory. Secondary

characteristics may include poor reading comprehension (due to

the decoding and memory difficulties) and poor written expres-

sion, as well as difficulty organizing information for study and

retrieval. (p. 55)

Current theories of the development of reading skills in

English stress that phonological processing is the most

significant underlying cognitive process (Stanovich, 1988a,

1988b, 1988c). Children and adults with a reading disability

have difficulty with phonological processing. Phonological

processing involves a variety of functions, but in the context of

the development of reading skills, the most significant in the

association of sounds with letters or combinations of letters.

This function is referred to as the understanding of grapheme-

phoneme conversion rules, and because of the irregular nature

of the correspondences in English, the learning of these rules

is a very complex process. The child who is learning to read

must map oral language onto written language by decompos-

ing the word into phonemes and associating each letter (or

combination of letters) with these phonemes.

The task of the beginning reader is to extract these

grapheme-phoneme conversion rules. The alternative is sim-

ply to memorize each word as a visual configuration and to

associate a meaning with it. This kind of learning may occur,

but it is inefficient and makes tremendous demands on visual

memory. In English, no one-to-one correspondence exists

between a letter (or letters) and a sound. The same letter rep-

resents different sounds, and the same sound may be repre-

sented by different letters.

In an alphabetic language such as English, the best mea-

sure of phonological processing skills is the reading of

pseudowords—that is, pronounceable combinations of letters

that can be read by the application of grapheme-phoneme con-

version rules but that are, of course, not real words in English.

Examples, include such pseudowords as shum, laip, and

cigbet. Pseudowords can be read by anyone who is familiar

with the grapheme-phoneme conversion rules of English even

though they are not real words and have not been encountered

in print or in spoken language before.

The development of the ability to read pseudowords

has been studied extensively (e.g., Calfee, Lindamood, &

Types of Learning Disabilities

473

Lindamood, 1973; Hogaboam & Perfetti, 1978; Siegel &

Ryan, 1988; Venezky & Johnson, 1973). Ample evidence in-

dicates that children with dyslexia have a great deal of diffi-

culty reading pseudowords. Studies such as those of Bruck

(1988), Ehri and Wilce (1983), Snowling (1980), Siegel and

Ryan (1988), and Waters, Bruck, and Seidenberg (1985) have

shown that disabled readers have more difficulty reading

pseudowords than do normal readers matched on either

chronological age or reading level. For example, Siegel and

Ryan (1988) studied the development of the ability to read

pseudowords in normal and disabled readers aged 7–14

years. By the age of 9, the normal readers were quite profi-

cient and performed at almost a perfect level for even the

most difficult pseudowords, with—in some cases—as many

as three syllables. Similarly, Backman, Bruck, Hebert, and

Seidenberg (1984) showed that 10-year-olds perform as well

as do adults on tasks involving the reading of pseudowords;

however, Siegel and Ryan (1988) found that the performance

of the reading-disabled children was quite different. These

children appear to acquire these reading skills very late in de-

velopment, and even reading-disabled children at the age of

14 were performing no better than were normal readers at the

age of 7.

To control (at least partially) for experience with print,

Siegel and Ryan (1988) used a comparison of disabled and

normal readers matched on reading grade level. Even when

the disabled readers and the normal readers were matched on

reading level (hence, the disabled readers were considerably

older than the normal readers), the performance of the

reading-disabled individuals on a task involving the reading

of pseudowords was significantly poorer than that of the

normal readers.

Thus, difficulties with phonological processing seem to be

the fundamental problem of children with reading disability,

and this problem continues to adulthood. Many adults with a

reading disability become reasonably fluent readers but still

have difficulty reading pseudowords or read them very slowly

(e.g., Barwick & Siegel, 1996; Bruck, 1990).

For children learning to read English, the learning of

grapheme-phoneme conversion rules is a result of systematic

instruction, and the extraction of the rules is a result of re-

peated encounters with print. No evidence is available as to

how much of the development of decoding skills is a result of

specific instruction in the grapheme-phoneme conversion

rules and how much is a result of experience with print. In

any case, the understanding of the grapheme-phoneme con-

version rules develops rapidly in the first years of experience

with print under normal conditions.

Some individuals have difficulties only with writing,

spelling, or both. Because these written language problems

usually occur in the context of problems with reading prob-

lems, arithmetic and mathematics problems, or both, the exis-

tence of a separate written language disability is not clearly

established, nor is there a clear definition of it, especially in the

adult population. Spelling difficulties can occur in the absence

of severe reading disabilities (e.g., Bruck & Waters, 1988;

Lennox & Siegel, 1993). There also may be problems with un-

derstanding or producing language. These problems have not

been documented as distinct learning disabilities and are often

components of dyslexia. If learning disabilities are to be

treated as measurable entities and if individuals are to receive

educational services based on the presence of a single or mul-

tiple learning disabilities, it is then obviously important to

determine what these learning disabilities are.

Arithmetic Disabilities

Individuals with developmental output failure or writing-

arithmetic disability have difficulty with computational

arithmetic and written language, typically in the absence of

reading difficulties, although this disability can co-occur with

dyslexia. They often have difficulties with spelling and have

problems with fine-motor coordination, visuospatial process-

ing, and short-term and long-term memory (e.g., multiplica-

tion tables), but they usually have good oral language skills

(Fletcher, 1985; Johnson & Mykelbust, 1967; Kinsbourne &

Warrington, 1963; Kosc, 1974; Levine, Oberklaid, &

Meltzer, 1981; Morrison & Siegel, 1991a, 1991b; Rourke,

1991; Rourke & Finlayson, 1978; Shafrir & Siegel, 1994b;

Siegel & Feldman, 1983; Siegel & Linder, 1984; Spellacy &

Peter, 1978). Rourke and his associates (e.g., Rourke, Del

Dotto, Rourke, & Casey, 1990; Rourke & Tsatsanis, 1996)

have described a syndrome called nonverbal learning dis-

abilities that is similar to writing-arithmetic disability. How-

ever, the operational definition of this learning disability is

problematic; it is not clear how a diagnosis can be made.

Often, these individuals have verbal IQ scores significantly

higher than performance IQ, but this discrepancy is neither

necessary nor sufficient to make the diagnosis. Often, they

have lower arithmetic scores than reading scores, but the dif-

ferences between these scores are not always significant (e.g.,

Rourke et al., 1990). (For an extended discussion of the defi-

nitional issue and conceptualization of this disability, see

Morrison & Siegel, 1991a.) 

Investigators (e.g., Fletcher, 1985b; Rourke & Finlayson,

1978; Rourke & Strang, 1978; Siegel & Feldman, 1983;

Siegel & Ryan, 1984, 1988, 1989a, 1989b) have found evi-

dence that children with specific arithmetic deficits and aver-

age or above-average word recognition scores on the WRAT

appear to have a variety of cognitive and neuropsychological