INT. J. LANG. COMM. DIS.

, MAY–JUNE 2007,
VOL. 42, NO. 3, 307–323

Research Report
Can a ‘single hit’ cause limitations in language
development? A comparative study of Swedish
children with hearing impairment and children
with specific language impairment

Kristina Hansson{, Birgitta Sahlén{ and Elina Mäki-Torkko{

{Department of Logopedics, Phoniatrics and Audiology, Clinical Sciences,
Lund University, Lund, Sweden
{Department of Otorhinolaryngology, Section of Audiology, Linköping
University Hospital, Linköping, Sweden
(Received 15 October 2005; accepted 24 July 2006)

Abstract
Background: Studies of language in children with mild-to-moderate hearing
impairment (HI) indicate that they often have problems in phonological short-
term memory (PSTM) and that they have linguistic weaknesses both in
vocabulary and morphosyntax similar to children with specific language
impairment (SLI). However, children with HI may be more likely than children
with SLI to acquire typical language skills as they get older. It has been
suggested that the more persisting problems in children with SLI are due to a
combination of factors: perceptual, cognitive and/or linguistic.
Aims: The main aim of this study was to explore language skills in children with
HI in comparison with children with SLI, and how children with both HI and
language impairment differ from those with non-impaired spoken language
skills.
Methods & Procedures: PSTM, output phonology, lexical ability, receptive
grammar and verb morphology were assessed in a group of children with
mild-to-moderate HI (n511) and a group of children with SLI (n512) aged 5
years 6 months to 9 years 0 months.
Outcomes & Results: The HI group tended to score higher than the SLI group on
the language measures, although few of the differences were significant. The
children with HI had their most obvious weaknesses in PSTM, vocabulary,
receptive grammar and inflection of novel verbs. The subgroup of children with
HI (five out of 10) who also showed evidence of grammatical output problems

Address correspondence to: Kristina Hansson, Department of Logopedics, Phoniatrics and Audiology,
Clinical Sciences Lund, Lund University Hospital, 221 85 #, Lund, Sweden; e-mail: kristina.
hansson@med.lu.se

International Journal of Language & Communication Disorders

ISSN 1368-2822 print/ISSN 1460-6984 online # 2007 Royal College of Speech & Language Therapists
http://www.informahealthcare.com
DOI: 10.1080/13682820600933526
308 Kristina Hansson et al.

was significantly younger than the remaining children with HI. Correlation
analysis showed that the language variables were not associated with age,
whereas hearing level was associated with PSTM.
Conclusions: Children with HI are at risk for at least a delay in lexical ability,
receptive grammar and grammatical production. The problems seen in the HI
group might be explained by their low-level perceptual deficit and weak PSTM.
For the SLI group the impairment is more severe. From a clinical perspective an
important conclusion is that the language development in children with even
mild-to-moderate HI deserves attention and support.

Keywords: hearing impairment, specific language impairment, phonological

short-term memory, language.

What this paper adds

Children with mild-to-moderate hearing impairment often have weaknesses in
phonological short-term memory and in language skills. Most studies have
found that the language problems are not related to the severity of the hearing
impairment. Some studies suggest that language problems are more prevalent
in younger age groups of children with HI. The results confirm that children
with HI have weaknesses in PSTM and in language and that the language
problems tend to be more severe in younger children with HI. Therefore, it is
important to assess their language at an early age and to offer support and help
to compensate. The comparison with the SLI group confirms that the
problems seen in this group are more severe than in the HI group, probably
due to a combination of factors.

Introduction
Children with permanent sensorineural mild-to-moderate hearing impairment (HI)
have an important characteristic in common with children with specific language
impairment (SLI) i.e. children who do not develop language as expected in spite of
normal hearing and otherwise typical development (Leonard 1998), in that they
often show a significant discrepancy between non-verbal cognitive ability and
language ability already at three years of age (Yoshinaga-Itano et al. 1998). The
language delay in children with HI has been suggested to include weak vocabulary
skills (Davis et al. 1986, Moeller 2000) as well as weak verbal reasoning skills (Moeller
2000). The aim of the present study is to compare children with HI and children
with SLI on several measures commonly used to assess language skills. Such a
comparison on a range of language skills has the potential to contribute to our
understanding of the prerequisites for language development and of factors that
may obstruct its typical course.
Gilbertson and Kamhi (1995) studied different aspects of language in children
with mild-to-moderate HI aged 7;9–10;7 compared with normally hearing controls.
Based on the children’s performance on a novel word learning task Gilbertson and
Kamhi could identify two distinct subgroups within the HI group. One subgroup
(50%) performed at a much lower level on receptive and expressive vocabulary,
novel word learning and non-verbal IQ. The remaining children performed at the
Hearing impairment and language impairment 309

same level as normally hearing controls. Neither age nor degree of HI could account
for the observed variation in language skills in the HI group. Gilbertson and Kamhi
(1995) argue that language impairment is more prevalent in a population of children
with mild-to-moderate HI than in children with normal hearing and that the better
performance in their higher performing subgroup is due to their higher than average
non-verbal IQ and language processing skills.
In a comparison between children with HI and children with SLI aged 5;11–
10;8, Briscoe et al. (2001) found that the children with HI as a group were as
impaired as the children with SLI in phonological processing tasks such as non-word
repetition, phoneme discrimination and non-word discrimination. They did not
differ from age controls with normal hearing and typical language development on
receptive vocabulary, receptive grammar or sentence recall. However, a subgroup of
children with HI with phonological impairment performed lower than an
unimpaired subgroup on a range of other spoken and written language measures,
in particular on receptive vocabulary. The authors conclude that both children with
HI and children with SLI apparently have impaired phonological processing skills,
but language and literacy problems are not as pervasive in the HI group as in the SLI
group. According to Briscoe et al. (2001), the two groups may share a low-level
perceptual deficit. This deficit can explain the difficulties evidenced by the children
in the HI group, but it cannot explain the whole range of language problems seen in
children with SLI. The authors suggest that SLI is the consequence of a number of
co-occurring factors, phonological processing problems being only one, perhaps
necessary, but not sufficient to cause the impairment.
Frazier Norbury et al. (2001) compared children with HI and children with SLI
aged from 5;9 to 13 on different aspects of verb morphology. As a group the
children with HI performed significantly better than the children with SLI, and at
the same level as controls. However, a substantial subgroup (22%) of the children
with HI had problems in finite verb morphology, similar to what is found in children
with SLI. Those children were among the youngest in the HI group. Poor language
ability was related to the age at testing, but not to the degree of HI, age of
identification or non-verbal ability. The authors conclude that children with HI may
have delayed, but not persistently impaired morphosyntactic development. A mild-
to-moderate HI should be treated as a risk factor for problems in language
development, but it has more pervasive consequences for phonological processing
than for morphosyntax.
In a third study Frazier Norbury et al. (2002) focused on receptive grammar.
Again, they found that their HI group, aged 5;9–10;7, performed more like controls
than like children with SLI. This strengthens the position that other factors than, or
in addition to, phonological processing problems associated with a low-level
perceptual deficit are critical in morphological and syntactic deficits.
Hansson et al. (2004) found that Swedish children with HI aged 9–13 years as a
group performed significantly better than a group of children with SLI in
approximately the same age range, and at the same level as controls on novel word
learning, receptive grammar and reading accuracy. However, 17% of the children
with HI performed below the 10th centile on receptive grammar, compared with
22% in the SLI group and with less than 10% in children with typical language.
Furthermore, the HI and SLI groups performed at the same level and somewhat
below normal age references on receptive vocabulary. The two groups performed at
a similar level in a non-word repetition task, also below what would be expected
310 Kristina Hansson et al.

given their age. Since both groups had normal phonology in their production of real
words, phonological output constraints were not considered a likely explanation of
their problems with non-word repetition. The results are interpreted to support the
suggestion made by Briscoe et al. (2001) that phonological processing problems in
children with HI are related primarily to the speech perception deficit, which is a
consequence of their HI, i.e. a ‘single hit’. The more wide-ranging and severe
problems in children with SLI, on the other hand, may be the result of limitations in
phonological processing in combination with something else, i.e. multiple risk
factors are involved (Briscoe et al. 2001).
The results from the cited studies suggest that like children with SLI, children
with HI may have problems in phonological processing (Briscoe et al. 2001),
phonological short-term memory (Briscoe et al. 2001, Hansson et al. 2004),
vocabulary skills (Gilbertson and Kamhi 1995, Briscoe et al. 2001), finite verb
morphology (Frazier Norbury et al. 2001) and receptive grammar (Hansson et al.
2004). A more or less consistent finding is that only a subgroup of children with HI
has significant problems. The problems cannot be explained by the degree of HI.
Gilbertson and Kamhi (1995) suggest that non-verbal IQ could be the explanation
of within group differences, whereas Frazier Norbury et al. (2001) found that
problems were more prevalent in younger children with HI. In a recent study
comparing phonological, lexical and grammatical skills in two age groups (6–10 and
11–15) of French-speaking children with mild-to-moderate HI Delage and Tuller
(2006) also found more severe problems in the younger group. However, the results
did not lead to the conclusion that the older group had normalized language skills.
Instead, Delage and Tuller (2006) found a stronger correlation between language
skills and degree of hearing impairment in the older group than in the younger
group. This was interpreted to confirm the hypothesis that the effects of the degree
of hearing loss may become more evident at later stages of language development as
within group variation decreases.
Even if age is a factor of relevance the results from several of the cited studies
(Briscoe et al. 2001, Hansson et al. 2004, Delage and Tuller 2006) indicate that we
need to look closer for a fuller account of language problems in children with HI.
Several accounts have been offered to explain the problems seen in children with
SLI. Some explanations focus on linguistic limitations and argue that these children
lack grammatical knowledge (e.g. Rice and Wexler 1996). Others claim that children
with SLI have perceptual limitations that cause difficulties to process grammatical
morphemes of short duration (e.g. Leonard et al. 1997). Others still argue that the
problems are due to cognitive limitations (Gathercole and Baddeley 1990). In
particular those accounts that assume limitations in the ability to process language,
due to perceptual or cognitive limitations, might be of relevance for children with
HI as well. Briscoe et al. (2001) suggest that a deficit in phonological processing
could be one (but not sufficient) factor of relevance. Another account attributes the
language problems of children with SLI to limitations in phonological short-term
memory (PSTM; e.g. Gathercole and Baddeley 1990). Non-word repetition is
considered to be a relatively pure measure of the capacity of the component of the
working memory system known as the phonological loop and highly consistent
deficits in SLI groups have been reported for this measure (e.g. Gathercole and
Baddeley 1990, Sahlén et al. 1999). PSTM supports the ability to form long-term
representations of novel phonological material, which is crucial for language
development (Baddeley et al. 1998). Given a close association between PSTM and
Hearing impairment and language impairment 311

language skills (Adams and Gathercole 2000) it is logical that phonological loop
deficits should result in language processing deficits (Baddeley 2003).
The purpose of the present study is to explore further language skills in children
with HI in relation to hearing level, age, non-verbal IQ and PSTM and in
comparison with children with SLI. We have investigated PSTM, phonology, lexical
skills, receptive grammar and morphosyntactic skills in Swedish children with HI
and children with SLI aged 5;6–9;0 years. We relate the results to an earlier study
comparing older groups (9–13 years) of children with HI and children with SLI
(Hansson et al. 2004) as well as to existing normal age references. Our specific
research questions are as follows:
N To what extent and in what areas do children with mild-to-moderate HI have
language impairments as severe as children with SLI?
N What differentiates children with HI with severe language impairments from
children with HI who have non-impaired spoken language skills?
N How can a relatively high prevalence of language impairments in children
with HI be explained?

Method
Participants
The participants were 23 children aged from 5;6 to 9;0 years. All children were
monolingual speakers of Swedish. The children belonged to two different groups.
One group consisted of 11 children, five girls and six boys, aged 5;7–8;11 years with
bilateral, symmetrical, sensorineural mild-to-moderate HI, fitted with a hearing aid
in at least one ear and with spoken Swedish as their first language. The better ear
hearing level, i.e. hearing thresholds averaged over the frequencies 0.5, 1, 2 and
4 kHz (BEHL0.5–4 kHz) varied between 30.0 and 68.75 dB HL. Worse ear hearing
level (WEHL0.5–4 kHz) varied between 32.50 and 71.25 dB HL. Audiograms were less
than 6 months old for all children except two. No children showed signs of
progressive HI. Age at diagnosis was between 1;1 and 4;7 years (mean 3;0) and age
at hearing aid fitting was between 1;6 and 4;9 (mean 3;1). All the children with HI
were educated in an oral setting and they had hearing parents. They were reported to
wear their hearing aids consistently and all wore their hearing aid(s) during testing.
IQ, as assessed with Raven’s progressive matrices (RSPM; Raven et al. 1990) ranged
from 80 to 130.
Grammatical aspects of the children’s language production were assessed with
the grammatical subtest of a Swedish test for language assessment, the Lund Test of
Phonology and Grammar (LuMat; Holmberg and Stenkvist 1983) although the
results were not used as a selection criterion. The scores ranged from 31 to 47 out of
50. Data from this test were missing for one of the children. For ten of these
children the Swedish version of the Children’s Communication Checklist (CCC;
Nettelbladt et al. 2003) had been completed by a teacher who knew the child well.
They all performed within normal limits.
A second group consisted of 12 children with SLI, nine boys and three girls aged
5;6–9;0 years. The criteria for selection were that they should have normal hearing
(passed screening at 20 dB HL at 0.5, 1, 2 and 4 kHz), non-verbal cognitive skills
within the normal range and no history of social–emotional or neurological
312 Kristina Hansson et al.

impairment (secured through parental report). These children had been identified by
speech/language clinicians as having grammatical difficulties. The inclusion criterion
for the children in this group was a performance .1.25 standard deviation (SD)
below normal reference data on the grammatical subtest of the LuMat where they
scored between 23 and 45 out of 50. Their non-verbal IQ ranged between 80 and
130. For nine of these children CCC protocols had been completed by a teacher
who knew the child well. Two children showed signs of pragmatic problems.
However, one was below six years of age, i.e. below the age for which the CCC is
appropriate.
Five of the children in the SLI group attended special language units, the
remaining were in mainstream day care or school. Background data for both groups
are shown in table 1.

Tasks
A Swedish non-word repetition test (Sahlén et al. 1999) consisting of 24 non-words was
administered to the children in order to assess their PSTM. The children were
instructed by the investigator to repeat non-words presented by a female voice on an
audiotape. The words were of two, three, four and five syllables in length (six of
each). They were constructed according to Swedish phonotactic rules, but were not
considered word-like. The non-words were transcribed phonematically, using both
audio- and videotaped recordings. The children’s responses were quantified as per cent
consonants correct (PCC), computed according to Holmberg and Sahlén (2001).

Table 1. Means and standard deviations of the performance of the HI and SLI groups on the
pre-tests

HI group (n511) SLI group (n512)

Chronological age (months)
Mean 85.82 80.83
SD 14.33 15.28

BEHL0.5–4 kHz (dB HL)

Mean 47.5 , 20
SD 10.85

WEHL0.5–4 kHz (dB HL)

Mean 51.59 , 20
SD 12.15

Non-verbal IQ
Mean 103.54 106.58
SD 13.65 17.42

Grammar LuMat (maximum score 50)

Mean 40.20 36.05
SD 4.28 6.77

Pragmatic composite on CCC

Mean 153.10 140.33
SD 7.74 16.82

SD, standard deviation.

Hearing impairment and language impairment 313

Output phonology was measured as per cent consonants correct (PCC) in 58 words
elicited with a picture-naming test. All Swedish consonants and the most common
consonant clusters are represented in these words, although not in all positions.
The children’s receptive vocabulary was assessed with a Swedish translation of
the Peabody Picture Vocabulary Test (PPVT; Dunn and Dunn 1981). A maximum score
of 93 was possible. The selection of items was based on the results from an item
analysis of the performance of 20 Swedish children with normal language
development aged 5–7 years on the first 109 items of the PPVT. The items to
which less than 60% of the children responded correctly were deleted from the test
(Reuterskiöld-Wagner et al. 2000).
Lexical organization and retrieval was assessed with the auditory associations
subtest from the Swedish translation of the Illinois Test of Psycholinguistic Abilities
(ITPA; Kirk et al. 1968). In this task the child listens to a sentence presenting a
relationship and is then asked to complete a second sentence, presenting a
relationship that is analogous with the one in the first sentence (e.g. ‘Daddies are big.
Babies are …’). The task consists of 35 items with increasing degree of difficulty.
The scores for the PPVT and auditory associations were transformed into per cent
correct responses.
A third test of lexical skills was the test of Rapid Automatized Naming (RAN;
Olofsson 2000). This is a test of lexical access where the child is asked to name
pictures of ten different objects occurring repetitively on sheets, as fast as possible.
The score was the total number of seconds it took the child to name all 60 pictures.
Receptive grammar was assessed with the Swedish version of the Test for Reception
of Grammar (TROG; Bishop 1983). The tester reads a sentence to the child who has
to point to the one of four pictures that matches the sentence. The test includes 20
blocks each assessing a different grammatical structure in four test items. The child
has to respond correctly on all four items in a block and the score is the number of
blocks correct. The score was transformed into per cent blocks correct.
The children’s skills in finite verb morphology were assessed with tasks eliciting
past tense forms of known and novel verbs. Regular past tense forms in Swedish
consist of the stem (which is the imperative) of the verb to which one of the
inflections -de (if the stem ends with a vowel or a voiced consonant as in stäng-de ‘closed’)
or -te (if the stem ends with a voiceless consonant as in lek-te ‘played’) is added. Irregular
past tense forms involve a vowel change in the stem without addition of an inflection (as
in sprang ‘ran’ from the stem spring). The test for past tense forms of known verbs
consisted of 20 test items eliciting ten regular and ten irregular past tense forms. For
each verb form the child was shown a picture showing an action and the past tense
forms were elicited in the following way: ‘These children like to play. They do it every
day. So what did they do yesterday? They … (played)’.
The children were also asked to inflect eight novel verbs in a task where each
item was presented in the same way as for the past tense of known verbs (e.g. a past
form like tengde from the infinitive tenga and the present tenger). The verbs used were
the same as in Hansson and Leonard (2003).
Each type of verb, regular, irregular and novel verbs were scored separately.
Regular and irregular verb responses were scored as correct if the child provided the
target form. The children’s responses in the novel verb task were scored as correct if
the form given was a plausible past tense form of that verb. Preliminary testing of
normally speaking adults showed some variation in past tense forms suggested for
each of these verbs. Thus, forms ending with -de or -te as well as irregular forms
314 Kristina Hansson et al.

made in analogy with one of the existing vowel change patterns for Swedish
irregular verbs were considered plausible past tense forms.
The variables obtained were per cent correct past tense forms of regular and
irregular known verbs and per cent of novel verbs with a plausible past tense form.
Reference data for the different tasks from children with normal hearing and
typical language development in the age ranges 5;0–5;11 and 6;4–7;4 collected for
different master’s thesis at the Department of Logopedics, Phoniatrics and
Audiology, Lund University were used for comparison.

Procedure
Screening and experimental testing sessions took place in a quiet room at the
Department of Logopedics, Phoniatrics and Audiology, Lund University. However,
the non-word repetition task was administered in a sound-treated room, used for
hearing level measurements. The testings were performed by a speech/language
pathologist and an audiologist. The sessions were all audio- and videotaped. The
examiners registered the children’s responses during the sessions. The tapes were
checked when the on-line registrations were incomplete or uncertain. All children
received the tests in the same order: phonology, PPVT, RAN, non-word repetition,
TROG, inflection of known verbs, inflection of novel verbs and auditory associations.
The whole procedure was in most cases carried out on one single occasion.
Unfortunately, not all of the children were willing to participate in all tasks. Data
are missing for auditory associations from one child in the HI group and one child
in the SLI group, for RAN also from one child in each group. Finally, one child in
the HI group did not do the novel verb inflection task. The reason not to exclude
those children was that they contributed data for the other lexical measures and verb
morphology on known verbs respectively.
The study was approved by the research ethics committee at the Medical Faculty,
Lund University. The parents received written and oral information about the project
and signed an informed consent form that they and their child agreed to participate.

Reliability
Recordings of the non-word repetition and verb inflection tasks for four randomly
selected children, two from each group, were independently transcribed and scored
by a second judge. Inter judge agreement on the non-word repetition test was
measured as the percentage of cases where both judges agreed on whether a
consonant was correctly reproduced or not. Agreement was 87.4%. Reliability on
the verb inflection tasks was measured as the percentage of cases where the two
judges agreed on whether a verb was correctly inflected for past tense or not (or, in
the case of the novel verb task, whether the child produced a plausible past tense
form or not). Agreement in this task was 93.4%.

Results
Comparison between the HI and SLI groups
The means and standard deviations for all variables for the HI and SLI groups are
shown in table 2. Normal reference data, when available, from children with normal
Hearing impairment and language impairment 315

Table 2. Means and standard deviations for the HI and SLI groups on the test variables.
Means and standard deviations from groups of children with typical language development in
the age group of 5;0–5;11 and from children in approximately the same age as the HI and SLI
groups when available are provided for comparison

HI group SLI group Controls 5;0–5;11

(n511) (n512) (n520) Age references
Non-word repetition (PCC)
Mean 72.92 48.80 75.11 83.21
SD 13.67 15.62 9.73 4.6
Output phonology (PCC)
Mean 99.08 72.0 – –
SD 1.01 23.96
PPVT (%)
Mean 77.61 67.85 72.8 84.52
SD 10.03 10.02 7.38 7.5
Auditory associations (%)
Mean 57.71 45.97 48.71 71–803
SD 16.32 12.74 4.14
RAN (seconds)
Mean 69.34 115.75 100.15 –
SD 16.67 58.50 18.47
TROG (%)
Mean 63.18 65.0 78.28 754
SD 19.27 10.87 17.09
Regular correct verb form (%)
Mean 90.91 47.50 84.5 –
SD 10.44 35.96 1.0
Irregular correct verb form (%)
Mean 60.91 35.00 38.5 –
SD 23.00 21.11 1.95
Novel verbs past form (%)
Mean 26.25 20.83 61.25 –
SD 23.16 28.37 2.27
1
Data from 20 children aged 6;4–7;6.
2
Data from 20 children aged 5:3–6;10 (Reuterskiöld-Wagner et al. 2000).
3
Stanine 5 for the age group 6;6–6;11 (Kirk et al./Holmgren, 1984).
4
Mean for the age group 6;0–6;11 (Holmberg and Lundälv, 1998).

hearing and typical language development in the age range 5;0–5;11 and in
approximately the same age range as the HI and SLI groups are also shown. The HI
group scored higher than the SLI group on all variables except the TROG.
Independent samples t-tests showed that the difference was significant for non-word
repetition (t(21)53.74, p50.001, eta250.42), output phonology (t(11.04)53.91,
p50.002, eta250.42), PPVT (t(21)52.33, p50.03, eta250.21), RAN
(t(11.76)522.52, p50.03, eta250.25) and correct past tense forms of regular
(t(12.99)54.00, p50.002, eta250.43) and irregular (t(21)52.82, p50.01, eta250.27)
known verbs. Following Cohen (1988) all effect sizes can be considered large
(.0.14).
An independent samples t-test of the results from the non-word repetition task
showed that the children in the HI group (aged 5;7–8;11) did not significantly differ
from 5–6-year-old references (t(27)520.51, p50.62). However a one-sample t-test
indicated that they performed significantly below the level of references from
children aged 6;4–7;4 (t(10)522.50, p50.03). According to an independent samples
316 Kristina Hansson et al.

t-test the SLI group performed significantly below the level of 5–6-year-old
references (t(28)525.70, p,0.001, eta250.54). The HI group performed close to
ceiling on output phonology whereas the SLI group had a significantly lower mean
and very large variation. The severe output phonology problems in the SLI group
clearly influences their performance in the non-word repetition task, which puts into
question whether the difference between the HI and SLI groups can be assigned to
a difference in PSTM or whether it is a consequence of the difference in output
phonology. In the SLI group non-word repetition correlated strongly and
significantly with output phonology (r50.76, p50.004). Due to a ceiling effect for
output phonology the relation between these two variables was much weaker and
non-significant (r50.19, p50.58) in the HI group. Thus, both the assumption of
linearity and the assumption of homogeneity of regression slopes were violated,
which prevented us from conducting an analysis of covariance with output
phonology as covariate.
The HI group scored significantly higher than the SLI group on the PPVT. Both
groups scored below the mean in reference data from a group of children aged 5–7
years who had 91% correct (Reuterskiöld-Wagner et al. 2000). A one-sample t-test
showed that this was significant for the HI group (t(10)522.28, p50.046) as well as
for the SLI group t(11)525.76, p,0.001).
The HI group was significantly faster than the SLI group on RAN. An
independent samples t-test revealed that they were also significantly faster than 5–6-
year-old references (t(26)524.37, p,0.001, eta250.42) whereas the SLI group,
showing an extreme variation, did not significantly differ from the 5–6 year olds
(t(11.23)50.86,1 p50.41).
The HI group also scored significantly higher than the SLI group on both
regular and irregular inflection of known verbs. Compared with typically developing
5–6 year olds the HI group scored at the same level on regular verbs (t(29)51.68,
p50.10) but significantly higher on irregular verbs t(29)52.87, p,0.001, eta250.22)
according to independent samples t-tests. The SLI group, on the other hand, scored
significantly lower than the 5–6 year olds on regular verbs (t(12.03)523.481,
p50.04, eta250.29), but did not differ from the references on irregular verbs
(t(30)520.48, p50.64).
The measures where the two groups did not differ were on auditory associations,
receptive grammar and past tense inflection of novel verbs. Reference data only exist
from 5–6-year-olds for auditory associations and inflection of novel verbs. However,
both auditory associations and TROG have been standardized for Swedish. When
the scores from auditory associations were transformed into stanine scores (Kirk
et al. 1968), the mean for the HI group was 3 and for the SLI group 2. Five children
in the HI group (45%) and eight children in the SLI group (67%) performed at
stanine 2 or lower. When the TROG scores were transformed into centiles (Bishop
1983) the mean centile for the HI group was 24.09 and for the SLI group 24.58. Five
of the children in the HI group (45.5%) and five of the children in the SLI group
(41.7%) performed below the 25th centile. Thus, both groups performed lower than
expected given their age on both of these tests. On past tense of novel verbs, finally,
both the HI and the SLI group scored far below 5–6 year olds. Independent samples
t-tests confirmed this (t(28)523.37, p50.002, eta250.29 for the HI group and
t(30)523.90, p,0.001, eta250.34 for the SLI group).
To summarize the results from the comparison between the two groups the HI
group performed significantly better than the SLI group on non-word repetition,
Hearing impairment and language impairment 317

output phonology, PPVT, RAN and correct form of known verbs. The low
performance in non-word repetition in the SLI group is to some extent related to
their severe output phonology problems. Comparison with existing reference data and
published norms for Swedish indicated that the HI group scored significantly lower
than expected for their age on non-word repetition, PPVT, auditory associations,
TROG and inflection of novel verbs. The HI group has more or less age-adequate
performance in output phonology, RAN and inflection of known verbs.

Lower and higher performing subgroups of HI

Five out of ten children in the HI group (50%) were found to meet the inclusion
criterion used for the SLI group on the LuMat2 (i.e. .1.25 SD below normal age
references; for group mean see table 1). The two subgroups of children with HI
were compared using the Mann–Whitney U-test. Medians and ranges for the two HI
subgroups are shown for all variables in table 3. The lower performing group was
significantly younger than the higher performing subgroup. The two subgroups did
not differ with respect to BEHL, age at identification or non-verbal IQ. The higher
performing group scored higher than the lower performing group on non-word
repetition and on all three lexical measures, but the difference was significant only
for the PPVT. There was also a significant difference between the two groups in per
cent correct blocks at the TROG. However, when the scores were transformed into
centiles, the difference did not reach significance (z521.62, p50.11). Four of the
five lower performing children and one of the five higher performing children
scored below the 25th centile.
The two HI subgroups did not significantly differ on inflection of known regular
verbs or inflection of novel verbs whereas the higher performing group scored
significantly higher on correct forms of irregular verbs.

Correlations in the HI group

Correlation analyses were made using Pearson r in order to investigate the
associations between age, hearing level, non-verbal IQ and PSTM and the language
variables in the HI group. Age correlated significantly only with past tense inflection
of novel verbs and this correlation was negative (r520.77, p50.009), i.e. younger
children tended to score higher on novel verb inflection (see also above). While
BEHL and age at diagnosis were not significantly associated with any of the
language variables, WEHL correlated negatively with non-word repetition
(r520.60, p50.05) and past tense forms of novel verbs (r520.74, p50.014).
Non-verbal IQ correlated positively and significantly with the PPVT (r50.68,
p50.02), auditory associations (r50.64, p50.047) and the TROG (r50.80,
p50.003). Non-word repetition, finally, correlated positively with grammar as
assessed with the LuMat (r50.66, p50.036) and with the PPVT (r50.73, p50.012).

Discussion
In summary, the present results indicate that children with HI in the age range 5;6–
9;0 as a group have relative weaknesses in PSTM, vocabulary and grammar.
However, their language problems are less severe than the problems seen in children
318 Kristina Hansson et al.

Table 3. Means and standard deviations for the variables and z- and p-values from the group
comparison (Mann–Whitney U-test) between the lower and higher performing HI groups.
The scores for non-word repetition and output phonology are in PCC and for RAN in seconds.
The remaining scores represent the per cent correct responses

Lower performing Higher performing

HI group (n55) HI group (n55) Statistics
Age (months)
Mean 76.20 93.20 z522.0
SD 11.08 13.26 p50.046
BEHL0.5–4 kHz
Mean 47.0 47.75 z520.53
SD 3.14 16.85 p50.60
WEHL0.5–4 kHz
Mean 52.5 51.25 z520.73
SD 10.83 15.76 p50.47
Age at identification (months)
Mean 35.4 32.80 z520.21
SD 15.14 14.45 p50.83
IQ
Mean 96.8 105.0 z521.05
SD 12.72 8.34 p50.30
Non-word repetition (PCC)
Mean 67.70 75.12 z520.94
SD 15.61 11.24 p50.35
Output phonology (PCC)
Mean 98.60 99.38 z521.28
SD 1.16 0.76 p50.20
PPVT (%)
Mean 70.53 81.71 z521.98
SD 9.79 4.10 p50.047
Auditory associations (%)
Mean 50.0 56.0 z521.24
SD 9.48 8.94 p50.22
RAN (seconds)
Mean 77.51 62.13 z521.23
SD 16.97 16.57 p50.22
TROG (% blocks correct)
Mean 48.0 75.0 z522.22
SD 18.57 7.07 p50.026
Regular correct verb form (%)
Mean 92.0 88.0 z520.44
SD 8.37 13.04 p50.66
Irregular correct verb form (%)
Mean 44.0 76.0 z521.95
SD 25.10 5.48 p50.051
Novel verbs past form (%)
Mean 32.5 20.0 z520.85
SD 20.92 25.92 p50.40

with SLI and problems with grammatical production are more common in a
younger subgroup.
The first question was to find out to what extent children with HI have the same
problems with language as children with SLI. The comparison between the two
groups in relation to normal references indicate that the areas where the children
with HI have their strengths are output phonology, lexical access (the RAN) and
Hearing impairment and language impairment 319

inflection of known verbs. In the HI group even the youngest children performed
almost at ceiling in output phonology. They were also close to ceiling at past tense
inflection of regular known verbs, whereas the children with SLI had significant
difficulties with regular past tense inflection, which is in accordance with earlier
studies (Hansson and Leonard 2003).
The HI group was significantly faster than the SLI group on the RAN.
Gilbertson and Kamhi (1995) found that their children with HI were faster than
typically developing peers on object naming. Data from more closely age-matched
controls for the HI group on this task would have been interesting in order to see if
this pattern had been replicated. Although the HI group scored significantly higher
than the SLI group on the PPVT they still scored significantly lower than
approximate age controls. Thus, they do have difficulties with receptive vocabulary,
although not as severe as the children in the SLI group.
Non-word repetition was another area where the HI group scored significantly
higher than the SLI group. We suspect that this might be partly a consequence of
phonological output problems in the SLI group. Whereas the HI group had close to
normal output phonology, many of the children with SLI had severe phonological
problems. In addition, their PCC in non-word repetition strongly correlated with
their PCC in output phonology, which was not the case in the HI group. The HI
group as well as the SLI group performed significantly lower than approximate age
controls. This is in accordance with Briscoe et al. (2001) as well as with Hansson et al.
(2004) who found no difference in non-word repetition between older children with
HI and children with SLI with no output phonology problems.
The HI group did not differ from the SLI group on auditory associations,
TROG and past tense inflection of novel verbs. On both auditory associations and
TROG the two groups scored in the low range, given their age. Thus, in contrast to
Briscoe et al. (2001) and Frazier Norbury et al. (2002) our results indicate that
receptive grammar as well as vocabulary can be an area of weakness in children with
HI. A possible explanation for the different results may be that the children in
these two studies are on average one year older than the children in our study (see
below).
Although the HI group apparently master past tense inflection of known verbs,
they were as impaired as the SLI group on past tense inflection of novel verbs. Boyle
and Gerken (1997) found that lexical familiarity is an important factor influencing
the use of inflection in children with typical language development. It can be
assumed that inflection of novel verbs is a task relying on both PSTM and
morphological skills. Since both groups have weak PSTM this may be part of the
explanation for the extremely large difference between the inflection of known vs.
novel words. What we see might be a delay in automatizing the application of
productive morphological rules in a task where vocabulary knowledge cannot be
relied upon. Like children with SLI, children with HI may need more input and
more practice, hence more time, in order to be fully proficient in morphosyntax.
We also wanted to find out what differentiates subgroups of children with HI
with and without impaired spoken language skills. We explored this issue by dividing
the HI group into a low performing and a high performing group based on their
results on an assessment of grammar. The results showed that those children in the
HI group who performed below the grammatical inclusion criterion for children
with SLI (. 1.5 SD below the mean from normal reference data on the LuMat)
were significantly younger than the rest of the group. This is largely in accordance
320 Kristina Hansson et al.

with Frazier Norbury et al. (2001) who found that children with mild-to-moderate
HI with weak PSTM, vocabulary and grammar were younger than children with
stronger language skills. In contrast, Gilbertson and Kamhi (1995) found that a
subgroup of children with HI with language problems differed from a subgroup
without language problems on non-verbal IQ.
The crucial question is how to account for the consistent finding of a fairly large
subgroup of children with HI with language problems. The approach adopted by
Gilbertson and Kamhi (1995) is to explain the performance of the higher
performing group as a consequence of stronger than average non-verbal cognitive
and language processing skills. This account was discarded by Frazier Norbury et al.
(2001) who found no difference between higher and lower performing groups on
non-verbal ability. This was also the case in the present study (table 3). However,
since several of the language variables correlated significantly with non-verbal IQ in
the HI group in the present study we cannot conclude that it is without importance.
Limited PSTM is one of the accounts offered for SLI, but since the higher and lower
performing HI groups did not differ in PSTM we cannot argue that weak PSTM is
responsible for language problems in children with HI. The lack of difference
between the higher and lower performing groups on hearing levels and the lack of
significant correlations between hearing level and language variables suggest that the
degree of hearing impairment cannot explain the language problems. However, it
might be possible to argue that there is an indirect association between hearing level
and lexical skills given that hearing level correlated with non-word repetition, which
in turn correlated with receptive vocabulary.
The age difference between the higher and lower performing HI groups, also
found by Frazier Norbury et al. (2001), indicates that language development, or at
least some aspects of it, is delayed rather than severely impaired in children with HI.
The results from TROG and LuMat indicate that the delay involves both
comprehension and production of grammar. The mean centile on TROG for the HI
group was 24, the same as for the SLI group. In Hansson et al. (2004) the mean
centile for an older HI group was 40, which was higher (although not significantly
so) than in an older SLI group whose mean centile was 23. In Frazier Norbury et al.
(2002) four out of 19 children with HI, somewhat older than the ones in the present
study scored markedly low on TROG, but as a group, they outperformed children
with SLI. However, the lack of associations between age and language variables in
the HI group shows that the relationship between age and language in children with
HI is not straightforward. In particular, there is no evidence that children with HI
acquire typical language skills as they get older, only that their problems are more
severe when they are younger. This is in accordance with the findings by Delage and
Tuller (2006) and Hansson et al. (2004).
In contrast to Frazier Norbury et al. (2001), our lower performing HI group did
not show significant impairment in the verb inflection task (except for problems
with novel verb inflection). Cross-linguistic differences may account for this.
Although Swedish children with SLI do have significant problems with verb
morphology, their problems do not seem to be as severe as the problems seen in
English-speaking children (Leonard et al. 2004).
Although the lack of age-matched controls does not allow us to draw any far-
reaching conclusions the results indicate that, as a group, it is apparent that the
children with HI have significant problems with PSTM as well as with language.
However, the fact that the HI group outperformed the SLI group on several
Hearing impairment and language impairment 321

measures suggests that children with SLI do have more severe and wide-ranging
difficulties than children with HI. A plausible account for this difference is that SLI
is the consequence of a combination of several factors (perceptual, cognitive and/or
linguistic), a ‘multiple hit’, whereas the somewhat milder and less persisting
problems in children with HI stem from a ‘single hit’, a factor which might be
related to their perceptual deficit as suggested by Briscoe et al. (2001). In the present
study the significant correlation between WEHL on the one hand and non-word
repetition and inflection of novel verbs on the other could be taken to point in this
direction.
When interpreting the results it is important to have in mind the weaknesses of
the present study, which make definite conclusions hard to draw. First, the sample
size is small. In particular the comparison between the lower and higher performing
subgroups of children with HI suffers from the low number of children. Another
problem is the lack of age-matched controls. Only TROG and auditory associations
have norms for Swedish. For the remaining tasks we had to rely on reference data
from children at the middle and lower end of the age distribution of the participants.
Finally, individual variation was large in both groups.

Conclusions
To conclude, what most clearly distinguish the HI group from the SLI group are
phonological skills, lexical access and inflection of known verbs. The two groups
are more similar in PSTM, vocabulary and receptive grammar. Language problems
are more severe in younger children with HI but are not as visible in verb
morphology as they are in children with SLI. Non-word repetition shows some
relation to hearing level in the HI group. This suggests that their perceptual deficit
may influence their PSTM.
The results have clinical implications for language assessment and intervention
in children with mild-to-moderate HI. Although part of the language problems seen
in younger children with HI is often a delay this does not mean that they do not
need support. In view of the PSTM problems and lexical weakness seen in older
children with HI it is important to assess their language and to offer support and
help to compensate for their deficit. In addition, language impairment ought to be as
prevalent in hearing impaired as in normally hearing children. A cognitive deficit in
combination with a hearing impairment is likely to have severe consequences.
Therefore, early identification and intervention to reduce the impact of the
impairment is of great importance. In both clinical practice and in research special
attention should be devoted to their lexical skills and more aspects of lexical and
grammatical abilities should be further explored.

Acknowledgements
This investigation was supported by Research Grant No. 2000-0171:01 from the
Bank of Sweden Tercentenary Foundation. The authors would like to thank Tina
Ibertsson who collected the data from the children with HI and performed the
hearing screening of the children with SLI. They are also very grateful to Jennifer
Forssén, Karin Göransson, Jonas Lindsjö, and Cajsa van der Pals who collected data
from typically developing children for their Master’s theses. They also warmly
322 Kristina Hansson et al.

acknowledge all the children and their parents for participation. Finally, they are very
grateful for very useful comments from two anonymous reviewers.

Notes
1. For these two comparisons, Levene’s test for equality of variances showed that equal variances
could not be assumed. The t-values are compensated for this.
2. For one child the record from the LuMat was missing. Therefore, this child was removed from the
analysis.

References
ADAMS, A.-M. and GATHERCOLE, S., 2000, Limitations in working memory: implications for language
development. International Journal of Language and Communication Disorders, 35, 95–116.
BADDELEY, A., 2003, Working memory and language: an overview. Journal of Communication Disorders, 36,
189–208.
BADDELEY, A., GATHERCOLE, S. and PAPAGNO, C., 1998, The phonological loop as a language learning
device. Psychological Review, 105, 158–173.
BISHOP, D., 1983, The Test for Reception of Grammar (TROG). Swedish trans. Holmberg, E. and Lundälv, E.,
1996 (Gothenburg: SIH Läromedel).
BOYLE, M. and GERKEN, L., 1997, The influence of lexical familiarity on children’s function morpheme
omissions: a nonmetrical effect? Journal of Memory and Language, 36, 117–128.
BRISCOE, J., BISHOP, D. and FRAZIER NORBURY, C., 2001, Phonological processing, language, and literacy: a
comparison of children with mild-to-moderate sensorineural hearing loss and those with
specific language impairment. Journal of Child Psychology and Psychiatry, 42, 329–340.
COHEN, J., 1988, Statistical Power Analysis for the Behavioral Sciences (Hillsdale, NJ: Erlbaum).
DAVIS, J., ELFENBEIN, J., SCHUM, R. and BENTLER, R., 1986, Effects of mild and moderate hearing
impairments on language, educational, and psychosocial behavior of children. Journal of Speech and
Hearing Disorders, 51, 53–62.
DELAGE, H. and TULLER, L., 2006, Mild-to-moderate hearing loss: Language progression from childhood to
adolescence. Poster presented at the Latsis Colloquium, Geneva, Switzerland, 26–28 January 2006.
DUNN, L. and DUNN, L., 1981, Peabody Picture Vocabulary Test — Revised (Circle Pines, MN: American
Guidance Services).
FRAZIER NORBURY, C., BISHOP, D. and BRISCOE, J., 2001, Production of English finite verb morphology: a
comparison of SLI and mild–moderate hearing impairment. Journal of Speech, Language, and
Hearing Research, 44, 165–178.
FRAZIER NORBURY, C., BISHOP, D. and BRISCOE, J., 2002, Does impaired grammatical comprehension
provide evidence for an innate grammar module? Applied Psycholinguistics, 23, 247–268.
GATHERCOLE, S. and BADDELEY, A., 1990, Phonological memory deficits in language disordered children:
Is there a causal connection? Journal of Memory and Language, 29, 336–360.
GILBERTSON, M. and KAMHI, A., 1995, Novel word learning in children with hearing impairment. Journal of
Speech and Hearing Research, 38, 630–642.
HANSSON, K., FORSBERG, J., LÖFQVIST, A. J., MÄKI-TORKKO, E. and SAHLÉN, B., 2004, Working memory and
novel word learning in children with hearing impairment and children with specific language
impairment. International Journal of Language and Communication Disorders, 39, 401–422.
HANSSON, K. and LEONARD, L., 2003, The use and productivity of verb morphology in specific language
impairment: an examination of Swedish. Linguistics, 41, 351–379.
HOLMBERG, E. and LUNDÄLV, E., 1998, T.R.O.G. svensk manual (Gothenburg: SIH Läromedel).
HOLMBERG, E. and SAHLÉN, B., 2001, Nya Nelli — neurolingvistisk undersökning för barn med språkstörningar
(Malmö: Pedagogisk Design).
HOLMBERG, E. and STENKVIST, H., 1983, Nya Lundamaterialet: Kartläggning och Bedömning av Barns Språkliga
Förmåga (Malmö: Utbildningsproduktion, AB).
KIRK, S., MCCARTHY, J. and KIRK, W., 1968, Illinois Test of Psycholinguistic Abilities. Swedish trans. Holmgren,
B., 1984 (Gothenburg: SIH Läromedel).
LEONARD, L., 1998, Children with Specific Language Impairment (Cambridge, MA: MIT Press).
Hearing impairment and language impairment 323

LEONARD, L., EYER, J., BEDORE, L. and GRELA, B., 1997, Three accounts of the grammatical morpheme
difficulties of English-speaking children with specific language impairment. Journal of Speech,
Language, and Hearing Research, 40, 741–753.
LEONARD, L., HANSSON, K., NETTELBLADT, U. and DEEVY, P., 2004, Specific language impairment in children:
a comparison of English and Swedish. Language Acquisition, 12, 219–246.
MOELLER, M., 2000, Early identification and language development in children who are deaf and hard of
hearing. Pediatrics, 106, 1–9.
NETTELBLADT, U., RADEBORG, K. and SAHLÉN, B., 2003, Svenska CCC — Children’s Communication Checklist
(Lund: Department of Logopedics, Phoniatrics and Audiology, Lund University).
OLOFSSON, Å., 2000, Naming speed, phonological awareness and the initial stage of learning to read.
Logopedics Phoniatrics Vocology, 25, 35–40.
RAVEN, J., COURT, J. and RAVEN, J., 1990, Raven’s Standardized Progressive Matrices (Oxford: Psychologists
Press).
REUTERSKIÖLD-WAGNER, C., SAHLÉN, B., RADEBORG, K. and TIDEMAN, E., 2000, Speed and context. The effect
of a sentence prime on naming speed in children with language impairment. Clinical Linguistics
and Phonetics, 14, 369–385.
RICE, M. and WEXLER, K., 1996, Toward tense as a clinical marker of specific language impairment in
English-speaking children. Journal of Speech, Language, and Hearing Research, 39, 1239–1257.
SAHLÉN, B., REUTERSKIÖLD WAGNER, C., NETTELBLADT, U. and RADEBORG, K., 1999, Non-word repetition in
children with language impairment — pitfalls and possibilities. International Journal of Language and
Communication Disorders, 34, 337–352.
YOSHINAGA-ITANO, C., SEDEY, A., COULTER, D. and MEHL, A., 1998, Language of early and later identified
children with hearing loss. Pediatrics, 102, 1161–1171.