You are on page 1of 19

IS DYSLEXIA A FORM OF SPECIFIC LANGUAGE IMPAIRMENT?

A COMPARISON OF DYSLEXIC
AND LANGUAGE IMPAIRED CHILDREN AS ADOLESCENTS
Author(s): Nata K. Goulandris, Margaret J. Snowling and Ian Walker
Source: Annals of Dyslexia, Vol. 50 (2000), pp. 103-120
Published by: Springer
Stable URL: http://www.jstor.org/stable/23765192
Accessed: 03-11-2015 02:13 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact support@jstor.org.

Springer is collaborating with JSTOR to digitize, preserve and extend access to Annals of Dyslexia.

http://www.jstor.org

This content downloaded from 134.29.12.206 on Tue, 03 Nov 2015 02:13:28 UTC
All use subject to JSTOR Terms and Conditions

A FORM OF SPECIFIC
IS DYSLEXIA
LANGUAGE
A
IMPAIRMENT?
COMPARISON
OF DYSLEXIC
AND LANGUAGE
IMPAIRED
CHILDREN
AS
ADOLESCENTS

Nata K. Goulandris,
University

College

London,

United Kingdom

J. Snoioling

Margaret

of York,

University

United Kingdom
Ian Walker

Max

Planck

Institute

of Cognitive

Bennewitz,

Neuroscience,

Germany

Two groups of adolescents with a childhood history of language impair


ment were compared with a group of developmentally dyslexic young
people of the same age and nonverbal ability. The study also included
two comparison groups of typically developing children, one of the same

age as those in the clinical groups, and a younger comparison group of


similar reading level to the dyslexic students. Tests of spoken and writ
ten language skills revealed that the adolescents with dyslexia were in
impairments on
from those with resolved language
distinguishable
at age
tasks, and both groups
performed
spoken
language
expected levels. However, both dyslexic readers and those with resolved
specific language impairments showed deficits in phonological aware
ness. On written language tasks, a different pattern of performance was
Annals

of Dyslexia,

Vol.

50,2000

Copyright 2000 by The International Dyslexia Association


ISSN 0736-9387
103

This content downloaded from 134.29.12.206 on Tue, 03 Nov 2015 02:13:28 UTC
All use subject to JSTOR Terms and Conditions

104

Development

Language

and Reading

Disabilities

apparent. In reading and spelling, adolescents with dyslexia performed


only as well as those with persistent oral language impairments and
younger controls. However, their reading comprehension was better. The
theoretical and educational

implications

of these findings are discussed.

(SLI) and developmental


reading
impairment
two distinct condi
are usually
considered
(dyslexia)
Association
tions (e.g., DSM-IV,
American
1994).
Psychiatric
a child who has a
The term dyslexia
is usually used to describe

Specific
disorder

language

skills despite
normal
ostensibly
language
that such
abilities, and there is general agreement
deficits (Snowling
1991). In
specific phonological
is
used
to de
contrast, the term specific language
impairment
with the acquisition
of spo
scribe children who have problems
disorder

of written

oral language
children have

ken
The

language

nonverbal
normal
despite
of SLI children
difficulties

language
wide range of linguistic
ments and grammatical
Leonard
2000).

1997).
ability (Bishop
a
tend to encompass
vocabulary
including
impair
processes
et al. 1987; Bishop and
deficits (Leonard

However,
many dyslexic children
difficulty (Rutter and Yule 1975) and
with language
to go on
impairments
1990; Tallal, Ross, and
(Scarborough
and Stollwerck
1996). Such findings

have

a history of language
it is common
for children
to have reading difficulties
Curtiss 1989; van der Lely
suggest that the two disor

ders exist on a continuum


of language
disorder, both groups of
children showing
in
awareness
and phono
deficits
phonological
and Wells
(Catts 1989, 1993, 1996; Stackhouse
logical processing
Within
this view, dyslexia
is conceptualized
either as a
1997).
mild form of language
the
impairment,
affecting only
phonolog
ical system, or as a residual problem that remains when oral lan
difficulties
guage
1984; Scarborough
An alternative

and Nation
(Aram,
Ekelman,
1990).
of these findings
is that lan
interpretation
a
risk
for
is
factor
guage
impairment
dyslexia
(Snowling,
to this view, phonologi
Bishop, and Stothard 2000). According
cal skills are critical to reading
development
(Byrne et al. 1997;
Share 1995) and, to the extent that children have phonological
failure.
difficulties,
However,
they will be at risk of reading
whether or not they show specific reading
difficulties/dyslexia
depends

on how

dyslexic

profile

have

resolved

and Dobrich

this phonological
deficit interacts with other
skills
2000). For children who
cognitive
language
(Snowling
have good semantic
a
of
is possible
skills,
degree
compensation
and Snowling
and
it
is
that
the
(Nation
common
1998),
likely
and

of better

reading

comprehension

than

word

This content downloaded from 134.29.12.206 on Tue, 03 Nov 2015 02:13:28 UTC
All use subject to JSTOR Terms and Conditions

Is Dyslexia

a Form of Specific Language

Impairment?

105

level decoding
will ensue.
children
who have poor
However,
semantic
or syntactic resources
show a global delay in reading.
semantic
difficulties
verbal
Moreover,
IQ, and
compromise
such children
receive
a
definition
of
rarely
discrepancy-based
a
be
outcome
cannot
assumed
for
Therefore,
dyslexia.
dyslexic
children
with preschool
it is
Indeed,
language
impairments.
of children who attract such a "di
likely that the characteristics
will change with the age of the child, along with both
agnosis"
the demands
of reading
at that stage in development
and the
received.
with this idea,
Consistent
teaching
they have
et al. (1992)
that different children
from
Shaywitz
reported
their epidemiological
fulfilled criteria for specific read
sample
at
different
retardation
et al. (2000) re
ing
ages, and Snowling
that
the
of
difficulties
ported
prevalence
specific
reading
with
a
children
of
among
history
preschool
impair
language
from 6 percent to 24 percent between
ment increased
the ages
of 8 and

15 years.
In light of the theoretical
and clinical interest in this issue, it
is surprising
that there have been few direct comparisons
of the
and
skills
of
children
as
reading
cognitive
diagnosed
having
difficulties and those with developmental
lan
specific reading
disorders.
Our
aim
in
the
was
to
examine
guage
present study
the notion

of a continuum
of language
disorder
directly by in
the
and
vestigating
cognitive, linguistic,
literacy profiles of ado
lescents
of school-leaving
who
had
a childhood
age
history of
either SLI or developmental
To
a
enable
dyslexia.
rigorous com

between
these groups, they were compared
with age
parison
and reading-level
matched
controls of similar nonverbal
ability.
adolescents
with either resolved
or persistent oral
By including
it was possible
to address
the following
language
impairments,
hypotheses:
1.

Dyslexia

is a resolved

pothesis,
children

with

children

ment whose

form of SLI.

with

a preschool
oral language

dyslexia

According
perform

history of language
impair
difficulties have resolved but

have been shown


to have residual
deficits (Bishop,
North, and
processing
Stothard et al. 1998).
who

2.

to this hy
to
similarly

phonological
Donlan
1996;

to
is a risk factor for reading difficulties. According
children with dyslexia
have a qualita
this hypothesis,
tively different profile from children with SLI because
deficits
and
more specific
phonological
they have
and
skills.
vocabulary
comprehension
stronger

SLI

This content downloaded from 134.29.12.206 on Tue, 03 Nov 2015 02:13:28 UTC
All use subject to JSTOR Terms and Conditions

106

Language

Development

and Reading

Disabilities

METHOD
PARTICIPANTS
Three

of 15- to 16-year-old
clinical
groups
pated in this study, one with a childhood
mental
and two with a preschool
dyslexia
The
impairment.
IQ and compared

language
nonverbal

adolescents

partici
of develop
history of specific
matched
for age and
history

were
groups
with two groups
of normal read
selected
to be of the same level of

ers, an age-matched
group
nonverbal
ability, and a younger
group
matched
controls for the dyslexies.

selected

as reading

age

The dyslexic
of 20 adolescents
with child
group consisted
hood
of
14
to
18
10
diagnoses
dyslexia,
aged
years
years
= 15
with a mean
of
15
9
months
(SD
months,
years
age
of the adolescents
were drawn from a cohort
months).
Eighteen
that had

in a longitudinal
study of developmental
five
after
their first assessment.
dyslexia,
approximately
years
this
20
children
in reading
Initially,
group comprised
ranging
from
8
at
6
to
all
of
least
whose
IQ,
age
years,
reading
average
participated

performance
lagged at least 18 months behind their chronologi
cal age (Snowling,
and Defty 1996). The data we
Goulandris,
had about
the early language
of these children
development
was from parental
About
one-third
of
reports.
parents reported
preschool
language
delays in their children.
to participate
Ninety percent of the original sample
agreed
in this follow-up;
one refused on the grounds
that he did not
wish to undertake
to
any more tests and the other was unable
attend because
of illness. 12/18 of the childhood
still
dyslexies

met

the criterion

of 18 months
below
grade level. Two more
with dyslexia, both reporting long-standing
reading
at this stage according
to the same se
difficulties, were added
lection criteria.
The participants
with a history of speech and language
diffi
culties were drawn from a sample
recruited
between
originally
3:9 and 4:2 years of age by Bishop
and Edmundson
(1987).
adolescents

children
in the study
(68 with SLI) participated
Eighty-seven
and were assessed
at 4, 4:6 and 5:6 years of age. At 5:6, 30 of
these
children
had
resolved
their
difficulties
language
38 had persistent
(Resolved
SLI), while the remaining
language
difficulties
in this paper
(Persistent
SLI). The data presented

were

collected

as part of a follow-up
of these groups of children
Stothard
et
al.
SLI
(1998). The resolved
reported by
19
children whose mean age at the time of the
comprised

at 15 years,
group

This content downloaded from 134.29.12.206 on Tue, 03 Nov 2015 02:13:28 UTC
All use subject to JSTOR Terms and Conditions

Is Dyslexia

a Form of Specific Language

Impairment?

107

assessment
was 15 years 5 months (SD = 3 months), being those
that matched the dyslexies
on the basis of chronological
age and
Block Design
of nonverbal
score, a measure
IQ. The persistent
SLI group consisted
of 20 children aged 15 years 7 months (SD
= 4 months),
selected
from the 30 available
to match on the
basis

of chronological
score. The reading
age and Block Design
status of these children was not known prior to the study.
Nineteen
readers whose mean age was
normally developing
=
15 years 9 months
3
formed the age-matched
(SD
months)
These
children
were drawn
comparison
group (CA-controls).
from the normative
tested
et
al.
(1998) by
sample
by Stothard
a range match by nonverbal
IQ and age to the
performing

children
with a mean age of 10
younger
dyslexies.
Eighteen
= 4 months)
formed a second
years 4 months (SD
comparison
readers. These children were se
group of normally
developing
and to
lected to be of similar nonverbal
ability to the dyslexies
at a similar

perform
readers

level

to the dyslexic and the persistent SLI


score on the Wechsler
Objective

to raw

according
Dimensions
test of basic reading
Wechsler
(WORD;
Reading
1993), a test of single-word
reading.
are summarized
in table 1. There
Details of the participants
the groups
were no significant
differences
between
for Block
Table

I.

of adolescents

Characteristics
language

with

Resolved
Dyslexic

(n = 20)
Age

and

RA,

impairment,

a history
of dyslexia,
CA- controls.

Persistent

SLI

SLI

(k = 19)

( = 20)
(n

specific

Young

Older

Controls

Controls

( = 18)

(n = 19)

(years)
15.79a
15.793

15.44

1.30

SD

a
3

15.62a
15.623

10.39b

15.68a
15.683

0.41

0.39

0.25

11.00a
11.003

9.65a
9.653

9.33

3.30

2.74

2.00

0.31

Block Design
(Scaled Score)
M

11.15

a
3

2.89

SD
Word

a
3

10.74a
10.743

2.45

reading

(WORD)
Mean

SS

SD
Note

all F ratios

Means
p<.05

level

84.80a
84.803

97.26b

85.20

11.85

11.90

15.33

are significant

at the .001

the same superscript


having
or Games-Howell
on Tukey

a
3

102.22b

10.42

104.89b

6.95

level.

are not significantly


post hoc tests.

different

at the

This content downloaded from 134.29.12.206 on Tue, 03 Nov 2015 02:13:28 UTC
All use subject to JSTOR Terms and Conditions

Language

108

Development

and

Reading

Disabilities

were
standard
that the groups
scores,
showing
Design
=
=
MSe
matched
for nonverbal
7.39).
1.76,
ability (F[4,91]
TESTS

AND

well

PROCEDURES

to pro
on a battery of tests designed
of spoken
and written language
vide a broad characterisation
skills of the participants.
processing
of the Wechsler
Nonverbal
The Block
Design
ability.
- ZZJLJK
was adminis
Scale
Children
1992)
(Wechsler
Intelligence
for
In
this
the child was
tered as a measure
of nonverbal
task,
ability.
The

children

were

tested

to reconstruct
two dimensional
using red and
patterns
for use with children age 6 to
blocks. This test is designed
16 years and has a split-half reliability for children age 15 of .92.
The Long Form of the
skills.
Language
Receptive
vocabulary.
British Picture Vocabulary Scale (BPVS)
(Dunn,
Dunn, Whetton,
In this test, the
and Pintilie 1982) assesses
vocabulary.
receptive
asked

white

a picture from a selection


of four that most closely
The test is standard
the word spoken by the examiner.
in
3
17 years and has a
ized on children
from
to
ranging
age
15-16
of.91.
for
children
split-half reliability
aged
child

chose

matched

This
Expressive
vocabulary.
taken from the Snowling
picture
and Stafford 1988),
Wagtendonk,
Graded Naming Test (McKenna
and
a range
posite measure
provided

six easy items


test comprised
test (Snowling,
van
naming
and the first 24 items of the
1983). This com
Warrington

of pictures
suitable
thought
for a wide range of abilities.
The Clinical
Evaluation
Sentence
of Language
processing.
- Revised (CELF-R)
Fundamentals
1986)
(Semel, Wiig, and Secord
subtest was used to measure
Sentences
Recalling
grammatical
sensitivity.
ing length
for children
to 16 of .80.

In this test, the child repeated


a sentence
of increas
and syntactic
This
test
is
standardized
complexity.
age

PHONOLOGICAL

5 to 18 and

PROCESSING

has a reliability

for children

age

15

SKILLS

Nonword
To assess
children's
repetition.
ability to repeat
unfamiliar
we administered
the Children's
Nonword
words,
Test
et
al.
This
test
com
(CN
(Gathercole
1994).
Repetition
Rep)
40
non
words
with
10
items
of
and
five
two,
three,
four,
prised

and "blonterstaping").
The test-retest re
(e.g., "ballop"
for
children
in
the
5
to
7
was
.77. The non
liability
age range
words were presented
via audiotape
to all the groups except the
syllables

who
dyslexies,
was important

the nonwords
after the experimenter.
It
repeated
to include
nonword
because
this test
repetition

This content downloaded from 134.29.12.206 on Tue, 03 Nov 2015 02:13:28 UTC
All use subject to JSTOR Terms and Conditions

Is Dyslexia

a Form of Specific Language

Impairment?

109

was thought to be too sensitive to underlying


difficul
language
ties (Bishop, North, and Donlan
1996).
awareness
skills were exam
Spoonerisms.
Phonological
ined using a version
In this
of a test devised
Perin
(1983).
by
the initial sounds
from the be
test, the child had to transpose
bear > badding
ginning of two spoken words (e.g., paddington
ton pear; hot dogs > dot hogs). Eleven different two-word
items
were

in this test. One

for each
point was allocated
on
that
word
order
condition
phoneme
transposed
correctly
was respected.
The maximum
score
on
this
test
is 22.
possible
The test has an internal reliability for children age 15 to 16 of .79
presented

(Stothard
LITERACY

et al. 1998).
SKILLS

The Wechsler

(WORD)
(Wechsler
Objective Reading Dimensions
to provide
of current literacy
measures
1993) was administered
attainment.
This test contains
three subtests:
Word
Single
and
All
the
Recognition,
Spelling,
Reading
Comprehension.

subtests

were

entered

into the analyses


to investi
individually
The WORD
is standardized
on

differences.

gate possible
group
the ages of 6 and 16 years and has a split-half
children between
.91 for
for children
reliability
age 15 to 16 of .88 for reading,
spelling, and .82 for reading comprehension.
In this test, the participant
Nonword
read a set of
reading.

and 10 two-syllable
nonwords
10 one-syllable
(e.g. "blem",
Test (Snowling,
from the Graded Nonword
"tegwop")
Reading
(1
1996). Five more difficult nonwords
Stothard, and MacLean
were
two- and 4 three-syllable
nonwords,
e.g., "pragendent")
added
to increase the level of difficulty. This test has a split-half
reliability for children age 15 to 16 of .77.

In order to examine
the participants'
Nonword
spelling.
unfamiliar
20 one- and
words,
they spelled
ability to encode
Written spellings
were considered
cor
two- syllable nonwords.
re
in the target were correctly represented
rect if all the sounds
was selected.
gardless of which grapheme
in a single
All the children were tested individually
either at the university lab, their school, or their home.

session,

RESULTS
performance
profile of the children
controls
and
their
age-matched
groups
The

from the three clinical


the test battery

across

This content downloaded from 134.29.12.206 on Tue, 03 Nov 2015 02:13:28 UTC
All use subject to JSTOR Terms and Conditions

110

Language

Development

and Reading

Disabilities

is shown
in figure 1, expressed
in terms of z scores relative
to
the mean and SD of the younger
controls. A number
of impor
tant differences
in profile can be seen. First, the Persistent
SLI

were impaired
across
all tasks and showed
group
especially
deficits
in
sentence
and
nonword
whereas
strong
repetition,
the Resolved
SLI group
showed
mild
for
only
impairments
their age on spoken
tasks
but
language
poorer
reading,
and phonological
than their oral skills pre
awareness
spelling,
dicted.
the dyslexic
children
at the same
Second,
performed
level as those with Resolved
SLI on tests of spoken
language
and in reading
but they were impaired
in
comprehension
and
awareness.
reading,
spelling,
phonological
As performance
approached
ceiling in the CA-control
group
on many of the tests administered,
transformations
were con
ducted
on all variables
that departed
from normality
or homo

Figure

1.

Performance of dyslexie, resolved SU, persistent SU, and


CA-controls across tests of spoken and written language;
scores are standardized against the mean and SD
of R
controls.

This content downloaded from 134.29.12.206 on Tue, 03 Nov 2015 02:13:28 UTC
All use subject to JSTOR Terms and Conditions

Is Dyslexia

a Form of Specific Language

Impairment?

Ill

and Fidell 1996). Reflect and


(see Tabachnik
geneity of variance
were applied
to the following
tests:
square root transformations
Sentence
Nonword
WORD
Repetition,
Repetition,
Spelling,
WORD
Nonword
and Nonword
Comprehension,
Reading,

Reflect and logarithm transformations


were applied
to
Spelling.
WORD
and
the
of
as
results
However,
Reading
Spoonerisms.
the analyses
data only differed marginally
using transformed
from those on raw data, analyses
on raw data are reported here.
variance
Multivariate
of
were conducted
for
analyses
separately
the spoken
and written
ANOVAs
ses, univariate

tests. Following
these analy
language
and multiple comparisons
were under
taken using
HSD
for variables
with homogeneous
Tukey's
variances.
When
of
variance
was present,
the
heterogeneity
Games
and Howell
the
tests
were
Since
post-hoc
applied.
pre

low statistical
sent study had relatively
power, effect sizes are
to
the
mean
and
the
relative
SD
of
reported
younger controls, to
differences.
clarify marginal
group
and
and

Table II shows the performance


of the three clinical groups
the two comparison
across
the spoken
groups
language

tasks. The MANOVA


on the lan
phonological
processing
tests indicated
that there were significant
guage
group differ
tasks A(20, 289) = 7.14, pc.001.
The adolescents
ences across
with dyslexia
at the same level as the
generally
performed
SLI group and the CA-controls,
and were superior
Resolved
to
the Persistent
SLI group
and younger
controls.
In turn, the
Persis-tent SLI group scored below the level of the Resolved
SLI
of
the
oral
tests.
on
most
language
group
that there were significant
Univariate
tests showed
group
= 21.56, MSe - 187.05,
difference in receptive
(F[4,91]
p c.001)
= 14.46,
and expressive
MSe = 13.06,
(F[4,91]
vocabulary
in
the
<.001
Post-hoc
demonstrated
case
of both re
).
analyses
p
with
and
that
the
adolescents
ceptive
expressive
language
dyslexia performed
and the CA-controls

at the same
and

level

as the Resolved

better than the Persistent

SLI group
and the

SLI

younger controls. There were also significant group differences


This
in Sentence
(F (4,91)=7.81,
MSe=57.86,
p<.001).
Repetition
revealed
that
the
and Howell
time, Games
post-hoc
analyses
than
less
well
the
Persistent SLI group performed
significantly
other four groups, while there was no significant difference be

of the dyslexic,
Resolved
the performance
SLI, younger,
controls.
on both phono
differences
There were significant
group
=
= 19.07,
tasks:
7.25, MSe
F[4,91]
logical
(Spoonerisms:
= 21.51,
=
MSe
Nonword
12.26,
F[4, 91]
p <.001;
Repetition
tween

and older

This content downloaded from 134.29.12.206 on Tue, 03 Nov 2015 02:13:28 UTC
All use subject to JSTOR Terms and Conditions

112

Language

Table

II.
tests

language

Mean

Development

scores

and Reading

on standardised

for adolescents

RA-

impairment,

and

Resolved

experimental

of dyslexia,

CA-

spoken

specific

language

controls.

Persistent

(m= 20)

SLI
(n = 19)

SLI
(n = 20)

122.90a

120.89"

104.60b

Dyslexic

Receptive

and

a history

with

Disabilities

Young

Older

Controls

Controls

(n = 18)

( = 19)

Vocabulary

(BPVS)1
92.61b

127.84"

SD

12.03

14.87

15.06

12.58

13.53

Range

100-141

88-140

76-131

72-112

96-147

22.00"

Expressive

Vocabulary

(Naming)2
M

19.95"

20.11"

15.25b

15.00b

SD

3.87

3.36

3.84

3.53

3.40

Range

13-26

13-26

9-24

7-21

15-27

66.60

66.68

57.15

66.67

67.74

Sentence

Repetition3

SD

6.54

Range
Nonword

6.51

11.67

4.00

6.79

54-76

54-77

26-77

60-74

57-78

33.50"

32.21"

25.90b

34.06"

35.16"

Repetition4

SD

3.85

4.28

6.88

4.20

2.85

Range

25-39

23-39

12-36

27-39

30-39

15.35"c

16.79"b

12.35e
12.35c

17.38"b

19.53b

SD

4.22

3.63

6.63

3.79

2.12

Range

7-21

11-22

0-20

8-22

15-22

Spoonerisms5

Note

all F-ratios

Means

are significant

level.

are not significantly


superscript
or Games
and Howell
post hoes.

having
level on Tukey
p<.05
1 Raw
2
score
Maximum
4

at .001

the same

Maximum

= 40

= 30

5- Maximum

3 Maximum

different

at the

= 25

= 24

revealed
that on the Spoonerisms
p < .001). Post-hoc
analyses
at
a
task, CA-controls
performed
higher level than the adoles
cents with dyslexia
and the Persistent
SLI group. The Resolved
SLI

made

more

missed

just
did not

errors

than

significance
differ from one

the CA-controls
but the difference
= .058). The three clinical
(p
groups
another
but performed
at the same

This content downloaded from 134.29.12.206 on Tue, 03 Nov 2015 02:13:28 UTC
All use subject to JSTOR Terms and Conditions

Is Dyslexia

a Form of Specific Language

Impairment?

113

level as younger
controls.
The analysis
of between-group
dif
ferences in nonword
revealed
that
the
the
repetition
dyslexies,
Resolved
SLI group, and the two sets of controls performed
at
a comparable
level which was higher than that achieved
by
the Persistent
SLI group.
a range effect here may
However,
have obscured
true group
differences.
It is notable
that, in
terms of absolute
level of performance,
both adolescents
with
and
those
with
Resolved
SLI
scored
less
well
than
dyslexia
controls

some

five years

younger

(effect

sizes

were

-.13

and

-.43, respectively).
The performance

of the groups
across the literacy tasks is
shown in table III. MANOVA
on these tests indicated
that there
was a significant main effect of group, A(20, 290) = 5.34, p <.001,
confirmed
on all the written language
by univariate
analyses
tasks. Importantly,
tests revealed
a different picture
follow-up
of group differences to that seen on the spoken
mea
language
sures. There were significant
main effects of group in single

word reading (F[4, 91] = 8.59, MSe = 35.64, p <.001 and spelling

= 11.60, MSe = 25.14,


p <.001). In the case of single word
the
for their age and
reading,
dyslexic
group were impaired
at
the
same
level
as
the
Persistent
SLI
performed
group and the
controls
without
a
of
The
younger
history
language
delay.
Resolved
SLI group
read
more words
than
the
correctly
SLI and the dyslexic
Persistent
this
latter
groups,
although
difference
failed to reach conventional
levels
of significance
In
the
adolescents
with
once again
<.08).
(p
spelling,
dyslexia
at
the
same
level
as
the
Persistent
SLI
performed
group and the
(F[4,91]

controls.
Their performance
was significantly
below
younger
that of the Resolved
SLI and the CA-control
groups.
There were also significant
main effects of group for non
word reading (F [4, 91] = 7.06, MSe = 23.12, p <.001) and for non
word spelling (F[4, 91] = 10.30, MSe = 9.95, p <.001). In nonword
controls per
reading, the three clinical groups and the younger
formed

similar

with

at a statistically
SLI had
Resolved

a less

level, but it is clear that those


severe impairment
(effect sizes
were -.09 for Resolved
and -.92 for
SLI, -74 for dyslexies,
Persistent SLI). Indeed, only the dyslexies and Persistent SLI per
formed significantly
less well than CA-controls.
The dyslexies
at a comparable
also performed
level to the Persistent
SLI and

younger controls on nonword


spelling. Once again, the resolved
those of the adolescents
with
SLI group attained scores between
and the CA-controls
and between-group
differences
dyslexia
Effect sizes were -.47 for
were marginally
(p <.063).
significant
dyslexies,

-.45 for Resolved

SLI and -1.04

for Persistent

SLI.

This content downloaded from 134.29.12.206 on Tue, 03 Nov 2015 02:13:28 UTC
All use subject to JSTOR Terms and Conditions

114

Language

Table

III.

history

Mean

Development

scores

of dyslexia,

on written

specific

language

impairment,
Persistent

SLI
( = 19)
(

(n = 20)
word

tests

language

Resolved
Dyslexic

Single

and Reading

SLI
(n = 20)

Disabilities

for adolescents
RA-

and

with

CA-controls.

Young

Older

Controls

Controls

(n == 18)
(tt
18)

(n = 19)

reading

(WORD: Raw)
M

41.853

Range

48.84a

41.94a

41.00a

50.21b

6.54

5.13

8.23

5.53

2.90

26-49

31-52

22-52

31-49

40-53

30.70a

40.05e
40.05c

SD

Spelling
(WORD raw)
M

35.68bc

33.25ab

32.44ac

SD

5.15

5.23

6.09

4.98

3.01

Range
Range

20-40

27-46

22-43

22-38

33-47

28.00a

29.58a

Reading
Reading

Comprehension

(WORD raw)
M

28.80a

24.25b

23.67b

3.85

3.64

4.60

4.67

4.09

22-35

18-34

13-30

14-30

18-35

16.85ab

20.00ab

r<
vdo O

20.44ab

23.16b

SD

4.25

5.08

6.55

4.85

2.12

Range
Range

6-22

6-25

4-25

5-25

17-25

12.75abc

15.32bcd

SD

2.90

2.89

5-17

9-19

SD
Range
Range
Nonword
Von word

Von word
Nonword

Reading1

Spelling1

Range
Range
Note

all F-ratios

are significant

at .001

11.15a
11.15s

14.06e
14.06c

4.56

2.80

1.94

1-18

9-18

11-20

level.

Means

the same superscript


are not significantly
having
level on Tukey
or Games
and Howell
p<.05
post hoes.
1 Maximum
= 25

Finally, a significant main effect


the reading comprehension
subtest
<.001
In
this
).
case, the dyslexic
p
ment and gained
scores
comparable
and CA-controls.
The Persistent
groups

did significantly

less well

17.1 ld

different

at the

of group was also found on


(F[4, 91] = 8.01, MSe = 17.49,
showed
no impair
readers
to those

SLI

and

with Resolved

younger
but in line with each

SLI

control
other.

This content downloaded from 134.29.12.206 on Tue, 03 Nov 2015 02:13:28 UTC
All use subject to JSTOR Terms and Conditions

Is Dyslexia

a Form of Specific Language

Impairment?

115

DISCUSSION
The

of adolescents
with childhood
of
comparison
diagnoses
and
revealed
similarities
dyslexia
specific language
impairment
as well as differences
between
the groups
on tests of spoken
and written language
differences
be
processing.
Interestingly
tween

adolescents
with dyslexia
and those with Resolved
and
Persistent SLI turned on the kind of processing
domain
tapped
tasks with which they were assessed.
by the various
Dyslexic
readers were indistinguishable
from age-matched
controls and
those

with

resolved

on oral language
tasks
language
problems
and grammatical
mea
tapping vocabulary
sensitivity
arguably
sures of general
both adoles
However,
language
competence.
cents with dyslexia
and those with Resolved
SLI did less well
than expected
for their age on Spoonerisms,
a test requiring ex

awareness.
A similar pattern of results was
repetition, although
group differences did not
reach conventional
level of significance.
On all of these oral lan
the group with Persistent
SLI were impaired
guage measures,
for their age and, at best, performed
as
well as younger
only
children some five years their junior.
On tests of written language,
the profile of the groups
re
plicit phonological
seen for nonword

vealed

a somewhat

altered

the fact that the


perspective.
Despite
from
the
Resolved
SLI
dyslexic
group
indistinguishable
on
were
more
in
tasks, they
group
spoken
language
impaired
and
nonword
Indeed
on
all
these
reading,
spelling,
reading.
as poorly as those with Persistent SLI and
tasks, they performed
the severity of their written lan
controls, confirming
younger
disorder.
On
basic
tests
of
guage
reading and spelling, the perfor
mance of the group with Resolved
SLI was slightly impaired
in
relation to age-matched
controls, and they were almost as poor as
with dyslexia
in nonword
adolescents
In the introduc
spelling.
were

two different hypotheses


about the relationship
tion, we posed
between dyslexia and SLI. These results force us to reject a strong
version of the hypothesis
that dyslexia is a form of Resolved
SLI.
the
readers
like
those
whose
Although
dyslexic
performed
early
difficulties had resolved
on tests of oral language,
language
they

did demonstrably
less well on tests of reading and spelling. On
these tests, they performed more like those with persisting prob
their reading
was better.
lems,
although
comprehension
of the group similarity is that
Arguably, a better characterization
the group with Resolved
SLI has residual
phonological
process
but they
ing difficulties that affect their ability to spell nonwords
have

developed

literacy skills in the normal

range.

This content downloaded from 134.29.12.206 on Tue, 03 Nov 2015 02:13:28 UTC
All use subject to JSTOR Terms and Conditions

Language

116

Development

and Reading

Disabilities

Our second hypothesis


was that SLI is a risk factor for read
by both the
ing difficulties. The shared difficulties encountered
task and
SLI and dyslexic
Resolved
group on the Spoonerisms
the
with
this
idea.
in nonword
are
consistent
However,
spelling
not
failure
is
fact that the Resolved
SLI group avoided
reading
would
within this view. The implication
easily accommodated

be that children

with

Resolved

SLI

have

weaknesses
children
with
that
dyslexic
argued

for phonological
compensate
than
reading
development

a greater

to
capacity
the
of
course
during
Nation
and
dyslexia.
children can learn to

(1998) have
Snowling
read by relying on global strategies underpinned
by vocabulary
nonword
for word
reading
despite
persisting
recognition,
and Olson 1992). It seems unlikely that
deficits (Rack, Snowling,
than
could
do this more effectively
the Resolved
SLI group
were
those in the dyslexic group because
indistinguishable
they

tasks administered.
from them on the spoken language
that
An alternative
is
dyslexic children carry the
hypothesis
children who have
risk
of
as
same
dyslexia
language-impaired
also have some,
readers
deficits. However,
dyslexic
phonological
A
candidate
area of
as yet unidentified,
additional
impairment.

be in creating links between


print and pronun
difficulty would
ciation, a difficulty that goes beyond
phonology
per se. A num
in
ber of recent theories suggest just such a deficit in dyslexia
(Wolf
representations
developing
orthographic
word-specific
and Bowers 1999; Wimmer, Mayringer,
and Landerl 1998).
It is important to note that the present study departs from oth
and dyslexic children
ers that have compared
language-impaired
in the age of the sample
tested. Kamhi and Catts (1986) investi
gated groups of much younger children, age 6 to 8 years, and it is

SLI group would


not have
at
this
for language
stage. Figure 2
impairment
shows the profile of the dyslexic group against that of those with
data
at four years (i.e., pooling
preschool
language
impairment
in
from the Resolved
and Persistent SLI groups).
When viewed
relevant

to note

that our Resolved

filled the criteria

this way, the profiles of the two groups


are strikingly similar;
while the adolescents
with dyslexia had more significant spelling
and better developed
both
problems
comprehension,
reading
in
showed
deficits
in
awareness
and
the
use
groups
phonological
of phonological
reading and spelling strategies.
These findings provide
some confirmation

for the hypothe


sis posed
above.
and
children
SLI
Therefore, dyslexic
carry the
same risk of dyslexia,
deficits.
It
seems,
namely
phonological
that dyslexic children use higher-order
skills
however,
language
as a compensatory
resource
read
(as revealed
by their superior

This content downloaded from 134.29.12.206 on Tue, 03 Nov 2015 02:13:28 UTC
All use subject to JSTOR Terms and Conditions

Is Dyslexia

0
CD
cc
o
o
O
o
QQ.

oc

U)
w
Figure 2.

a Form of Specific Language

CO
"D
"O
1
i_
o
O

u
T3
0
vu
CO
CD
0
0
cc

(0
0
TD
u
X
U_
O
o

CD
0
w
Q.
Q
Q.
(O
(J)
CO

Impairment?

CO
TD
i
O
5
z
"O
TJ
CO
CD
So.
CC
DC

117

CO
"O
1
o
5
z
"03
Q.
CO
C/D

Performance of dyslexie
dyslexic children relative to the whole
sample of adolescents who had a language impairment at
4 years (SLI combined).

A consequence
of their greater reliance on
ing comprehension).
context is that they develop
less well-specified
orthographic
than those who rely less on top-down
representations
process
children.
ing to learn to read, such as language-impaired
When sampling
differences are taken into consideration,
it is
clear that the present findings align well with those of Kamhi
and Catts (1986)
who found that dyslexic
children
and those
with persisting
both
language
impairments
performed
poorly
on phonological
tasks and in nonword
processing
reading.
the superior
However,
SLI confirms the view

of the group with Resolved


at which oral language
im
of
predictor
prognosis.

performance
that the age

resolve is an important
pairments
A limitation
of the present study was
children
classified
into the three clinical
Indeed,

the

results

children

that the numbers


of
were
small.
groups

differ somewhat
from those
language
from whom we selected
the present groups

impaired
(Stothard et al. 1998). The lower statistical power in the present
differences
fre
study means that between-group
performance
failed
to reach
conventional
levels
of significance
quently

This content downloaded from 134.29.12.206 on Tue, 03 Nov 2015 02:13:28 UTC
All use subject to JSTOR Terms and Conditions

118

Language

Development

and Reading

Disabilities

there were distinct trends pointing to nonword


although
and spelling deficits in those with Resolved
SLI. These
the
to
issues
definition
of
point
diagnostic
surrounding
and lead us to be cautious
about our interpretation.
We

reading
findings
dyslexia

suggest
that the concept of a dyslexic spectrum may be one with consid
erable clinical utility that circumvents
the issue of fuzzy category
We propose
that a dimension
of phonological
boundaries.
pro
is at the core of this spectrum
(cf. Stanovich
cessing impairments

and Siegel 1994). The impact of these impairments


will be modi
fied both by language
resources
outside
of the phono
processing
domain
and
more
(Nation
1998),
logical
general
Snowling
such as processing
cognitive resources
speed, and aspects of at
tention control.
The

of the present
findings
the
factors
that place
considering

study make clear that, when


a child at risk of dyslexia,
it is
to
take
view
of
the
demands
of
a
important
developmental
to
In
it
read.
is
to
consider
how
learning
particular,
necessary
difficulties impact on children's
language
reading development
at different stages, placing
them at differential risk of dyslexia.

to read is an interactive
that draws
not only
Learning
process
on phonology,
but also on other language
as well
resources
et
al.
As
have
others
(Plaut
1996).
argued, language
impairment
outcome
places a child at risk of reading failure, but a dyslexic
is not the only scenario for children from this population.
An in
teraction of their language
and
will
weaknesses
deter
strengths
mine their outcome,
which will be more favorable
in the context
of appropriate
intervention
1996).
(Snowling

ACKNOWLEDGMENTS
This

from Wellcome
study was carried out with support
grant
to the second
author.
Ian Walker
was sup
040195/Z/93/A
Vacation
We thank Sue
ported
by a Wellcome
scholarship.
Stothard
and Dorothy
for help at various
Bishop
stages in the
research and Janice Brown, who assisted
in data collection
sup
ported

by a vacation

award

from the Nuffield

Foundation.

Address correspondence
to: Margaret J. Snowling, Department of
Psychology, University of York, York YO10 5DD (mjsl9@york.ac.uk);
Dr. Nata Goulandris, Department of Human Communication
Science,
Dr. Ian Walker,
University College London, (a.goulandris@ucl.ac.uk);
Max Planck Institute
of Cognitive
Neuroscience
(MEG
group),
D-04828 Bennewitz, Germany (walker@cns.mpg.de).

This content downloaded from 134.29.12.206 on Tue, 03 Nov 2015 02:13:28 UTC
All use subject to JSTOR Terms and Conditions

Is Dyslexia

a Form of Specific Language

Impairment?

119

References

American

Association.
Psychiatric
Disorders TV. Washington,
DC:
Aram, D., Ekelman,
B., and Nation,

and Statistical
Manual
Diagnostic
of Mental
American Psychiatric Association.
with language
disorders:
10
J. 1984. Preschoolers
and
later.
Research
27:232-44.
Journal
years
of Speech
Hearing
Press.
Bishop, D. V. M. 1997. Uncommon Understanding. Hove: Psychology
D. V. M., and Edmundson,
A. 1987. Language
Bishop,
impaired
4-year-olds:
transient
Distinguishing
Disorders 52:156-73.
D. V. M.,

1994.

from persistent

Journal of Speech and Hearing

impairment.

L. C. 2000. Speech and Language


From
Impairments:
Press.
Psychology
Bishop, D. V. M., North, T., and Donlan, C. 1996. Nonword
repetition as a behavioural
Evidence
marker for inherited language
from a twin study. Journal
impairment:

Bishop,

and

Leonard,

Research to Practice. Hove:

of Child Psychology and Psychiatry 37:391-403.


the child's and
R., Ashley, L., and Larsen, K. 1997. Assessing
Byrne, B., Fielding-Barnsley,
the environment's
contribution to reading acquisition:
What we know and what
we don't know. In Foundations of Reading Acquisition and Dyslexia: Implications for
Catts,

Early Intervention, B. Blachman,


H. W. 1989. Defining dyslexia

ed. Mahwah,
NJ: Lawrence Erlbaum Associates.
as a developmental
disorder. Annals
language

Dyslexia 39:50-64.
between
Catts, H. W. 1993. The relationship
speech
Journal of Speech and Hearing Research 36:948-58.
Catts,

H. W. 1996.

and

reading

disabilities.

as a developmental
disorder:
An ex
dyslexia
language
in
Topics
Language Disorders 16(2):1429.
L. M., Whetton, C., and Pintilie, D. 1982. British Picture Vocabulary

Defining

view.

Dunn,

language

of

panded
L. M., Dunn,

Scale. Windsor,

England: NFER-Nelson.
A. D., and Emslie, H. 1994. The children's test of
S. E., Willis, C., Baddeley,
nonword repetition: A test of phonological
working memory. Memory 2:103-27.
an understanding
lan
of developmental
Kamhi, A. G., and Catts, H. W. 1986. Toward
Gathercole,

guage and reading disorders.


Leonard,
L., Sabbatini,
L., Leonard,
ment in children:
R., and

McKenna,

Journal of Speech and Hearing Disorders 5:337-47.


J., and Volterra, V. 1987. Specific language
impair
A cross-linguistic
study. Brain and Language 32:233-52.

Warrington,

E. K. 1983.

Graded

Naming

Test. Windsor,

England:

NFER-Nelson.
Nation,

M. J. 1998. Individual
differences in contextual
facilitation:
Snowling,
and poor reading
from dyslexia
Child Development
comprehension.
69:996-1011.
K., and

Evidence
Perin,

D.

1983.

Phonemic

segmentation

and

spelling.

British

Journal

of Psychology

74:245-57.
M. S., and Patterson, K. 1996. Understanding
J. L., Seidenberg,
in quasi-regular
word reading: Computational
principles
Psychological Review 103:56-115.

Plaut, D. C., McClelland,


normal

and

domains.

impaired

Rack, J., Snowling,


M., and Olson, R. 1992. The non-word
review. Reading Research Quarterly 27:28-53.

reading

deficit in dyslexia:

Yule, W. 1975. The concept of specific reading retardation.


Journal of
Child Psychology and Psychiatry 16:181-97.
in dyslexic
Child
deficits
children.
H. 1990. Very early language
Scarborough,
Development 61:1728-43.
Rutter, M., and

Scarborough,
delay.

H., and

Dobrich,

W. 1990.

Journal of Speech and Hearing

of children
Development
Research 33:70-83.

with early language

This content downloaded from 134.29.12.206 on Tue, 03 Nov 2015 02:13:28 UTC
All use subject to JSTOR Terms and Conditions

120

Language

Semel,
Share,

Development

and

Reading

Disabilities

W. 1986. The Clinical


Evaluation
E., Wiig, E. K., and Secord,
of Language
Fundamentals-Revised.
New York: The Psychological
Corporation.
D. 1995. Phonological
Sine qua non of reading acqui
recoding and self-teaching:
sition. Cognition 55:151-218.

B. A., Fletcher, J. M., and Makugh,


R. 1992.
S. E., Escobar, M. D., Shaywitz,
Shaywitz,
Evidence
that dyslexia may represent the lower tail of a normal distribution
of
reading ability. New England Journal of Medicine 326:145-50.
M. J. 1991. Developmental
Journal of Child Psychology and
reading disorders.
Psychiatry 32:49-78.
M. J. 1996. Annotation:
to the teaching of reading.
Snowling,
Contemporary
approaches
Journal of Child Psychology and Psychiatry 37:139-48.
Snowling,

Snowling,

M. J. 2000. Language

Language
Leonard,

and literacy skills. Who is at risk and why? In Speech and


From Research
to Practice,
D. V. M. Bishop
and L. C.
Impairments:
eds. Hove: Psychology
Press.

M. J., Bishop, D. V. M., and Stothard, S. E. 2000. Is pre-school


language
a risk factor for dyslexia in adolescence?
Journal of Child Psychology
Psychiatry 41,587-600.

Snowling,

pairment

im
and

M. J., Goulandris,
N., and Defty, N. 1996. A longitudinal
study of reading de
in dyslexic children. Journal of Educational Psychology 88(40):653-69.
velopment
M. J., Stothard, S. E., and MacLean,
J. 1996. Graded Nonword Reading Test.
Snowling,
Snowling,

Thames Valley Test Company.


Bury St. Edmunds,
England:
M., Van Wagtendonk,
B., and Stafford, C. 1988. Object

Snowling,

velopmental
J., and

Stackhouse,

Whurr

dyslexia.
Wells,

naming deficits in de
Journal of Research in Reading ll(2):67-85.
B. 1997. Children's Speech and Literacy Difficulties. London:

Publishers.

K. E., and Siegel, L. S. 1994. The phenotypic


performance
profile of reading
A regression-based
children:
test of the phonological-core
variable
difference model. Journal of Educational Psychology 86:24-53.
M. J., Bishop, D. V. M., Chipchase,
C. 1998.
Stothard, S. E., Snowling,
B., and Kaplan,
Stanovich,

disabled

A follow-up
in adolescence.
Journal of Speech,
pre-schoolers:
impaired
and Language Research 41:407-18.
Statistics (3rd ed.). New York:
Tabachnick,
G., and Fidell, L. 1996. Using Multivariate
Harper Collins College Publishers.
in specific language
Tallal, P., Ross, R., and Curtiss, S. 1989. Family aggregation
impair
ment. Journal of Speech, Hearing and Language Research 27:987-98.
Language

Hearing

Van

der Lely, H., and Stollwerk,


in children. An autosomal

Wechsler,

D. 1992.

Psychological
D. 1993.
Wechsler

L. 1996. A grammatical
specific language
impairment
Brain and Language 52:484-504.
inheritance?
Wechsler Intelligence Scale for Children - Third Edition. UK. New York:
dominant

Corporation.
The Wechsler

Objective

Reading

Dimensions

(WORD).

New

Wimmer,

Corporation.
H., Mayringer,
H., and

York:

Psychological

K. 1998. Poor reading:


A deficit in skill
Landerl,
or a phonological
deficit? Scientific Studies of Reading 2:321-40.
P. G. 1999. The double
deficit hypothesis
for developmental
Bowers,
Journal of Educational Psychology 91:154-68.

automatization
Wolf,

M., and
dyslexia.

This content downloaded from 134.29.12.206 on Tue, 03 Nov 2015 02:13:28 UTC
All use subject to JSTOR Terms and Conditions