INT. J. LANG. COMM. DIS.

, JULY –SEPTEMBER 2004,

VOL. 39, NO. 3, 401–422

Working memory and novel word learning

in children with hearing impairment and
children with specific language impairment

K. Hansson
, J. Forsberg
, A. Löfqvist
, E. Mäki-Torkko` and

B. Sahlén


Department of Logopedics, Phoniatrics and Audiology
` ENT Department, Section of Audiology, Lund University Hospital, Lund,
Sweden

(Received August 2003; accepted November 2003)

Abstract
Background: Working memory is considered to influence a range of linguistic
skills, i.e. vocabulary acquisition, sentence comprehension and reading. Several
studies have pointed to limitations of working memory in children with specific
language impairment. Few studies, however, have explored the role of working
memory for language deficits in children with hearing impairment.
Aims: The first aim was to compare children with mild-to-moderate bilateral
sensorineural hearing impairment, children with a preschool diagnosis of
specific language impairment and children with normal language development,
aged 9–12 years, for language and working memory. The special focus was on
the role of working memory in learning new words for primary school age
children.
Methods & Procedures: The assessment of working memory included tests of
phonological short-term memory and complex working memory. Novel word
learning was assessed according to the methods of Gilbertson and Kamhi
(1995). In addition, a range of language tests was used to assess language
comprehension, output phonology and reading.
Outcomes & Results: Children with hearing impairment performed significantly
better than children with a preschool diagnosis of specific language impairment
on tasks assessing novel word learning, complex working memory, sentence
comprehension and reading accuracy. No significant correlation was found
between phonological short-term memory and novel word learning in any
group. The best predictor of novel word learning in children with specific
language impairment and in children with hearing impairment was complex
working memory. Furthermore, there was a close relationship between complex
working memory and language in children with a preschool diagnosis of

Address correspondence to: Birgitta Sahlén, Department of Logopedics, Phoniatrics and Audiology,
Lund University Hospital, SE-221 85 Lund, Sweden; e-mail: Birgitta.Sahlen@logopedi.lu.se

International Journal of Language & Communication Disorders

ISSN 1368-2822 print/ISSN 1460-6984 online # 2004 Royal College of Speech & Language Therapists
http://www.tandf.co.uk/journals
DOI: 10.1080/13682820410001669887
402 K. Hansson et al.

specific language impairment but not in children with hearing impairment.

Conclusions: Complex working memory seems to play a significant role in
vocabulary acquisition in primary school age children. The interpretation is that
the results support theories suggesting a weakened influence of phonological
short-term memory on novel word learning in school age children.

Keywords: children, hearing impairment, language impairment, working memory,

vocabulary.

Introduction
Several studies have pointed to limitations of working memory in children with
specific language impairment (SLI). Few studies, however, have explored the role
of working memory for language deficits in children with hearing impairment (HI).
Working memory is considered to impinge on a range of linguistic skills, i.e.
vocabulary acquisition, sentence comprehension and reading. Specifically, successful
vocabulary acquisition has been claimed to be the single most important
determinant of a child’s eventual intellectual and educational attainments
(Gathercole 1999). One focus of interest in this study were the learning and
memory mechanisms in lexical acquisition: more specifically, the role of working
memory in novel word learning in children with a preschool diagnosis of SLI and
children with HI of primary school age.
There are two prominent models of working memory, Baddeley’s (2000)
component model and the capacity theory of comprehension (e.g. Just and
Carpenter 1992). In Baddeley’s model, working memory is a multicomponent,
capacity-limited system. One of the components, the phonological loop, includes an
articulatory control process (verbal rehearsal) that brushes up and maintains speech
material in the store for a short period. This verbal short-term memory system
supports the learning of the phonological structures of novel words. The
phonological short-term memory is considered to be reliably indexed by non-word
repetition in children before age 5 years (Gathercole 1999). Phonological short-
term memory has been shown to predict word learning, grammar and readings
skills in children with normal as well as deviant language development. It is argued
by some authors that in SLI, the underlying deficit is reduced phonological short-
term memory (Gathercole and Baddeley 1990, Adams and Gathercole 1995).
The capacity theory of comprehension (Just and Carpenter 1992) is a
computational model of a system responsible for the simultaneous processing and
storing of information. The present paper will refer to this system as complex
working memory. SLI is, by an increasing amount of authors, explained by a limited
capacity model of language processing. The limited capacity in children with SLI is
linked to a reduced working memory (Ellis Weismer and Hesketh 1996, Ellis
Weismer et al. 1999, Ellis Weismer and Evans 2002, Montgomery 2003). The
simultaneous processing and storing of verbal information was studied in 20
children with SLI and 20 age-matched children with normal language development
(NL) by Ellis Weismer et al. (1999). The authors found that the children with SLI
had poorer word recall and also a different pattern of errors in a competing
language processing task (CLPT; Gaulin and Campbell 1994), which will also be
used in the present study.
So far, few studies have sought to study whether the two above-mentioned
Working memory and novel word learning in children 403

memory systems equally contribute to language development. According to

Gathercole (1999), phonological short-term memory and complex working memory
might play different roles in supporting the acquisition and skills during childhood.
Language development is often delayed in children with HI (Yoshinaga-Itano
et al. 1998). For example, Gilbertson and Kamhi (1995), in a study on novel word
learning in children with mild-to-moderate HI, found that only 50% of the children
performed as well as normally developing children. Phonological development is
also typically hampered in children with HI. Huttunen (2001) showed that
compared with normally developing, age-matched, 4–6-year-old children, children
with HI made significantly more vowel substitutions. They also made frequent
voicing errors and deletions of final consonants. In Huttunen (2001), the frequency
of errors was related to the degree of HI. Recent studies have shown that early
diagnosis followed by an early start of habilitation is associated with better language
development (Yoshinaga-Itano et al. 1998, Moeller 2000).
Very few studies have explored the causes underlying the problems of language
and literacy in children with HI. In children with HI, not only are hearing
thresholds affected, but also frequency discrimination for suprathreshold stimuli,
resulting in restricted speech perception (Moore 1995). The early processes of
perceptual encoding of the speech signal must be successfully completed in order
to create a phonological memory trace. Although a reduced ability perceptually to
encode the speech signal might explain the linguistic deficits in children with HI,
there is no simple relationship between degree of HI, degree of speech perception
deficit, and severity of language problems in children with HI (Gilbertson and
Kamhi 1995). For example, weak vocabulary has been demonstrated in children
with very mild HI (Davis et al. 1986).

Comparisons between children with HI and children with SLI for language and
literacy
Subtle acoustic processing deficits may thus explain poor performance on language
measures in children with HI, but not necessarily in children with SLI. Bishop
(1999), for example, found a weak association between fine-grained temporal
discrimination of brief tones and non-word repetition in children with SLI. Few
studies so far have explored the influence of working memory on language
development in children with HI. Comparisons between children with SLI and HI
could shed light on the relative contribution of auditory perception and cognitive
factors, such as working memory, to language development in children with HI.
Briscoe et al. (2001) compared 6–11-year-old children with SLI and children with
mild-to-moderate sensorineural HI on language and literacy skills. They found
similarities on tests for phonology and non-word repetition in children with SLI
and children with HI compared with normal controls. However, there were
qualitative differences in test results between the children with SLI and HI.
Children with SLI had more difficulties with longer and more complex non-words
than children with HI. In addition, the relationship between phonology and other
language abilities, including literacy skills, seemed to be closer for children with SLI
than for children with HI. The authors also compared the children with SLI and HI
on reading skills as measured by reading comprehension and decoding of
words and non-words. No difference between the children with HI and control
children was found on any of the measures, but marked deficits were found
404 K. Hansson et al.

for the children with SLI. According to the authors, limitations of phonological
processing in the children with HI were of the same magnitude as in the children
with SLI without leading to consequences for reading development. The authors
suggest that a speech perception deficit in children with HI can compromise
phonological short-term memory. However, this does not necessarily lead to
serious impairments in language or literacy. An explanation for this could be that
children with HI are more dependent on orthographic strategies than children with
SLI. Phonological problems in children with HI can thus represent a ‘single hit’,
whereas SLI might be the consequence of a ‘double hit’, i.e. limitations of
phonological short-term memory are necessary but not sufficient for SLI (Briscoe
et al. 2001).
The focus in Briscoe et al.’s study is phonological processing and reading
and the theoretical framework is Baddeley’s multicomponent model of working
memory. The role of phonological short-term memory in reading seems obvious
in children with SLI, but evidence is not equally convincing for children with HI.
To replicate some aspects of Briscoe et al. (2001) and to explore the role of
complex working memory in reading, which is an important aspect of language
in children beyond preschool years, the present study also included tests of literacy
skills.

Non-word repetition in children with SLI and in children with NL

Non-word repetition is claimed to predict both lexical and grammatical
development in young children. Adams and Gathercole (2000) showed that
4-year-olds with better non-word repetition skills produced speech that comprised a
wider repertoire of words, longer utterances and a greater range of syntactic
constructions than did children with relatively poor non-word repetition skills.
However, non-word repetition only predicts vocabulary development reliably up to
5 years of age (Gathercole et al. 1992, Adams and Gathercole 1995, Gathercole
1999). According to Gathercole (1999), after age 5, long-term knowledge is
integrated with temporary memory mechanisms. The integration may operate at
both the lexical and phonotactic levels. The recall of non-words becomes strongly
influenced by both the lexical knowledge and phonotactic frequency of the memory
item (Gathercole et al. 1999). This interaction makes non-word repetition a less
sensitive measure, at least if standard non-word repetition tests are used, where the
phonotactic rules of the ambient language are obeyed (Baddeley et al. 1998,
Gathercole and Pickering 1999). This was corroborated by results from Simkin and
Conti-Ramsden (2000), who found that normally developing children perform at a
ceiling on non-word repetition by the age of 10 years on the widely used non-word
repetition test developed by Gathercole et al. (1994).
Even when the memory items do not follow the phonotactic rules of the
ambient language, children become very good at repeating non-words with
increasing age, probably due to support from lexical knowledge in long-term
memory. In a study of Swedish normally developing children aged 9–12 years, Fries
and Holmberg (2001) found that the children reached a ceiling on a non-word
repetition test (Sahlén et al. 1999). Therefore, they used Russian words as memory
items and showed that 88% of the items were correctly repeated, despite violating
Swedish phonotactic rules, i.e. they contained consonant clusters that are of low
probability or impossible in Swedish.
Working memory and novel word learning in children 405

In children with SLI, several studies have shown a relationship between

phonological output and non-word repetition (e.g. Edwards and Lahey 1998),
sometimes referred to as ‘mutual output constraints’. For example, in a study on 28
5-year-old children with SLI (Sahlén et al. 1999), the phonological developmental
level was the best predictor of non-word repetition when compared with other
measures of expressive and receptive language. The mutual output constraints
hypothesis, however, has been questioned by Adams and Gathercole (2000), who
studied normally developing 4-year-olds. The authors found a significant
association between short-term memory assessed with non-spoken recall and
language development. They also found an association between language
performance and visuospatial short-term memory leading to the formulation of
a ‘mutual input constraints hypothesis’.
Considering the influence of phonological output on non-word repetition in
younger children with SLI, it is interesting to study children whose problems of
phonology are restored. According to Bishop et al. (1996), non-word repetition is ‘a
genetic marker’ of SLI, and problems in non-word repetition should ‘resist’
restoration of output phonology. In the present study, the children with a preschool
diagnosis of SLI are between 9 and 12 years of age at testing and were 5 years of
age when diagnosed. At age 5, all had phonological as well as grammatical
problems. However, present language status was not used as a selection criterion
for the present study.
To eliminate the confounding influence of speech production processes in
children with phonological impairment, Nyman (1999) constructed a non-word
discrimination task based on the non-words in the repetition test used by Sahlén
et al. (1999). In a study of 41 Swedish children with NL (mean age 5;10), a strong
correlation between non-word repetition and non-word discrimination was found,
supporting the view of a mutual dependency on a common underlying
phonological processing ability as proposed by Bowey (1996). Thus, the non-
word discrimination task was included in the present study not only to eliminate the
influence of speech production. A comparison between non-word repetition and
non-word discrimination also makes it possible to examine perceptual influence, i.e.
phoneme discrimination, on non-word repetition.

Purpose
The first aim was to compare children with HI, children with a preschool diagnosis
of SLI and children with NL in primary school age on aspects of language
considered to be influenced by working memory. Specifically, the aim was to study
the role of phonological short-term memory and complex working memory in
novel word learning. The hypothesis was that children with a preschool diagnosis
of SLI will perform worse on language measures than age-matched children with
mild-to-moderate HI. Novel word learning is probably associated with phonological
short-term memory and with complex working memory in both groups of children.
Based on the findings of Briscoe et al. (2001), a stronger association between
linguistic and working memory measures is predicted for children with SLI than for
children with HI.
The main research questions were as follows:
. What are the similarities and differences between primary school age
406 K. Hansson et al.

children with a preschool diagnosis of SLI and children with HI on language

and working memory?
. What is the role of working memory in novel word learning? Do
phonological short-term memory and complex working memory equally
contribute in primary school age?

Method
Participants
Three groups of children were included in the study. One group consisted of 18
children with mild-to-moderate, bilateral, sensorineural HI (Better Ear Hearing
Level [BEHL], 0.5–4 kHz, mean 41.4 dB HL, range 30.0–57.5 dB HL). They were
between 9;1 and 13;3 (median 10;10) years, and wore a hearing aid in at least one
ear. Sixteen of the 18 children had a non-verbal IQw80, two children reached IQ
73 (as assessed by Raven et al.’s 1990 progressive matrices [RSPM]) (mean IQ 91,
standard deviation [SD] 13.67]. The children were recruited from ear, nose and
throat clinics in southern Sweden. All children were monolingual speakers of
Swedish. They were all educated in an oral setting. The majority of the children
attended mainstream school; only three children attended a special school for
children with HI. They had hearing parents. Language status was not used as a
selection criterion. The time of identification of their HI varied (median age of
identification 4;3). For demographic data, see table 1.
The second group consisted of 27 children with a preschool diagnosis of SLI.
The children were recruited from two earlier research programmes and were thus
thoroughly examined as preschoolers. At the time of the present study, the age of
the children was 8;6–11;4 (median 10;0) years. Non-verbal IQ assessed by the
RSPM (Raven et al. 1990) in this group was similar to that in the HI group (mean
90, SD 13.75). Three of the 27 children did not reach a non-verbal IQ of 80 at the
time of the present study (although they all did in preschool). Their IQ scores were
77, 73 and 73. All children in the SLI group passed a hearing screening (at 0.5, 1, 2
and 4 kHz). These children were also monolingual speakers of Swedish with
hearing parents. Present language status was not used as a selection criterion. All
the children attended mainstream school, except for one child who was in a special
school. The majority of children in this group were still receiving individual support
from a special teacher in school.
The third group consisted of 38 children with NL studied by Fries and Holmberg
(2001). The children were 9;5–12;4 (mean 10;6) years. The data from this group
included the majority of the tests used for the present study. Those not included
were tests where pilot testing had showed that normally developing children of this
age have reached ceiling or where published age norms exist. For the group of
children studied by Fries and Holmberg, only group means and SDs were available.

Procedure
The children were assessed at the Department of Logopedics, Phoniatrics and
Audiology, Lund University, except for five children with HI, who were tested in a
special school for children with HI for geographical reasons. For the remaining
children with HI, all phonological memory tasks were administered in a sound-
treated room. The testings were performed by two experienced speech–language
Working memory and novel word learning in children 407

Table 1. Demographic data for the children with hearing impairment

Age at Age at start

diagnosis of hearing Progress of
(years; aid use (years; BEHL0.5–4kHz* WEHL0.5–4kHz** hearing
Sex months) months) (dB) (dB) impairment Aetiology
F 4;2 4;3 56.25 57.50 yes unknown
M 4;8 4;10 34.00 35.00 no unknown
F 3;0 3;2 57.50 58.75 no unknown
F 3;8 3;1 57.50 72.50 no unknown
F 4;7 4;9 50.00 50.00 yes unknown
F 7;5 *** 57.50 61.25 yes hereditary
M 4;1 5;0 30.00 33.75 no unknown
M 3;2 4;2 36.25 40.00 not known unknown
F 5;5 5;7 32.50 45.00 not known unknown
M 4;5 4;10 36.25 41.25 not known unknown
F 1;3 1;3 45.00 47.50 no unknown
F 2;5 3;1 33.75 42.50 no hereditary
M 4;5 5;1 31.25 36.25 no hereditary
F 4;4 4;5 40.00 42.50 no hereditary
F 4;4 4;5 31.25 36.25 no unknown
M 3;1 3;2 45.00 50.00 no hereditary
M 4;6 4;1 33.75 45.00 no unknown
M 7;0 7;3 37.50 40.00 no hereditary

*BEHL0.5–4kHz, better ear hearing level, average threshold at 0.5, 1, 2 and 4 kHz.
**WEHL0.5–4kHz, worse ear hearing level, average threshold at 0.5, 1, 2 and 4 kHz.
***No hearing aid.

pathologists. The whole procedure was audio- and video-recorded. The following
tests were performed:
. Novel word learning (Gilbertson and Kamhi 1995).
. Complex working memory (Gaulin and Campbell 1994).
. Phonological short-term memory, per cent consonants and per cent
phonemes correct in a non-word repetition task (Sahlén et al. 1999).
. Non-word discrimination (Nyman 1999).
. Output phonology, per cent consonants correct in a picture-naming test
(Hellqvist 1991).
. Lexical ability (PPVT; Dunn and Dunn 1981).
. Language comprehension (TROG; Bishop 1983).
. Literacy, reading accuracy, and reading speed for words and non-words on a
computerized reading test (TIPS; Lyxell and Holmberg 2000).
Novel word learning was assessed with a Swedish version of a procedure
described by Gilbertson and Kamhi (1995). The test consisted of acquisition and
retention phases. The acquisition phase consisted of trials to learn four novel nouns
(faj, afte, prulado, sannebille) for four novel objects. For a first exposure, two well-
known objects and one of the novel objects were placed in front of the child. The
child was asked to hide the objects, one at a time (the well-known objects first),
underneath three different hiding items (a cloth, a box and a cone). The child was
supposed to figure out by exclusion what the new item was called. The hiding
408 K. Hansson et al.

objects were taken away and for a second exposure including a comprehension
task, two more novel objects were placed in front of the child. The child now had
five objects in front of her/him. The child was asked to hand three objects,
including the novel object representing the target word, to the test leader. In the
next step, the production part, the child was asked to name the object representing
the target novel word. If the child did not name the object correctly, he/she was
given three alternative words for the novel object (recognition). One of the novel
words in the recognition part was the correct target word, one was a phonologically
similar word, one was the child’s own production, if incorrect. The retention phase
was carried out after 30 min. The test leader showed each of the four novel objects
and asked the child to name them. If the child’s production was erroneous, he/she
was again given the three alternatives for recognition. (During the 30 min, other
tests in the procedure were administered.) To avoid order effects, two versions (A,
B) of the novel word-learning test were administered where the same novel words
were used but in a different order. Fifty per cent of the children were given version
A and 50% were given version B.
Analysis and scoring: the child was given a score of 1 if he/she recognized the
correct word, 2 if he/she produced a word that altered the target word with more
than one phoneme, and 3 if the word was phonologically similar (only one
phoneme altered) or correctly produced.
Complex working memory was assessed with the Competing Language
Processing Task (CLPT; Gaulin and Campbell 1994, Swedish version by Pohjanen
and Sandberg 1999). Thirty sentences constructed as semantically acceptable or
semantically unacceptable propositions, divided into 265 sets, were administered.
The child was asked to judge semantic acceptability (yes/no) for each sentence and
recall the last word of the sentences in each set.
Analysis and scoring: the children’s responses were scored as correct or
incorrect for both the true/false and word recall components of the CLPT. The
sequence of word recall did not have to match that of sentence presentation.
Responses were scored as correct if the child produced the last word of sentences
within the target group. For every word correctly recalled, a score of 1 p was given;
thus, a maximum score of 30 was possible. In addition to the quantitative analysis,
a qualitative analysis was made. The responses were classified according to Ellis
Weismer et al. (1999) into the following categories: 1, no response; 2, primacy error,
if the child produced the first word of a sentence within the group of sentences
presented at that specific level; 3, medial error, if the child produced the middle
word (verb) in a sentence within the group of sentences presented at that level; 4,
within set repetition, if the child produced the same target word twice in the same
group of sentences; 5, target intrusions, if the child produced a target word from
any prior group or level; 6, semantic substitutions, if the child produced a word that
was semantically related to the target, including antonyms; or 7, other, i.e. errors
that did not fit within any of the categories: words that had not been presented in
any of the stimulus sentences and did not seem to be related semantically to the
target words.
Phonological short-term memory was assessed with a Swedish non-word
repetition test (Barthelom and Åkesson 1995, Sahlén et al. 1999), consisting of six
non-words of two, three, four and five syllables each (totally 24 non-words, e.g.
glyvo, lebosuf, luttosaluk and dallabellhime). The words were administered to the
Working memory and novel word learning in children 409

children from a tape recording. The non-words were constructed according to

Swedish phonotactic rules.
Analysis and scoring: the non-words were transcribed, using both audio- and
videotaped recordings. The per cent consonants correct (PCC) was computed,
according to Holmberg and Sahlén (2000). An analysis of stress patterns and
syllable deletions was also made. A score of 1 was given if the number of syllables
was correct and an additional point if the stress pattern was correct. The reliability
of transcription between two independent transcribers on 10% of the data was
96%. Interjudge reliability for scoring for phonemes correctly produced (PCC) was
100%.
Non-word discrimination (Nyman 1999) was assessed with a task where 16 of
the non-words from the previous task, four from each level of syllable length, were
selected to create 32 non-word pairs for auditory discrimination (same–different).
Each non-word was paired once with itself (same, e.g. sallotan–sallotan) and once
with another non-word differing from it by one phoneme (different, e.g. sallotan–
sallovan).
Analysis and scoring: for each of the 16 non-words, both pairs had to be
correct for a score of 1. The maximum score was thus 16.
Output phonology was assessed with a picture-naming test, a shortened version
of the Phonemetest (Hellqvist 1991), where per cent PCC was counted in the same
way as for the non-word repetition test. In Hansson and Nettelbladt (2002), 5-year-
old children with NL had on average 98% of consonants correct at the same test.
Thus, the level expected for 10-year-olds would be 100%.
Lexical ability was assessed with a Swedish translation of the first 109 words in
the PPVT (Dunn and Dunn 1981). No normative data exist for this age group of
children. For the present paper, the authors collected data from a reference group
of 20 normally developing children, mean age 10;5 (range 9;0–11;8) years.
Grammatical comprehension was assessed by the TROG (Bishop 1983),
Swedish version by Holmberg and Lundälv (1998), including Swedish norms.
Analysis and scoring: the number of blocks correct was computed. A maximum
of 20 was possible. All four items in a block had to be correct in order for the
block to be considered as correct.
To assess reading, parts of a computerized reading test (Text Information
Processing System [TIPS]; Lyxell and Holmberg 2000) was administered to the
children. In the present study, tasks assessing the decoding of 30 words and 30
non-words (accuracy and speed) were used.
Analysis and scoring: the children’s reading of the words and non-words was
scored as correct or incorrect. Small phonological deviations were accepted. A
maximum score of 30 on each part of the test was possible. Speed was registered
by computer and measured in seconds.

Results
The children with HI were treated as one group, since a subdivision of the children
with HI into two groups—(1), 13 children with mild HI (BEHL0.5–4kHzv50 dB HL);
and (2), 5 children with moderate HI (BEHL0.5–4kHzw50 dB HL)—revealed that
the two groups did not significantly differ on any of the measures. Descriptive data
(absolute scores) for the children with SLI and HI are shown in table 2.
410 K. Hansson et al.

Table 2. Descriptive data for children with hearing impairment (HI; n~18) and specific
language impairment (SLI; n~27)

HI SLI NL

Mean SD Mean SD Mean SD

Novel word learning (acquisition) maximum: 12 9.33 2.79 9.44 2.72 10.95 1.33
Novel word learning (retention) maximum: 12 4.17 1.43 3.26 1.35 3.58 1.83
Complex working memory, maximum: 30 17.83 3.73 13.30 6.11 17.97 2.22
Non-word discrimination, maximum: 16 14.06 1.35 14.15 2.20 – –
Non-word repetition, phonemes correctly produced 82.75 9.18 79.06 11.84 – –
Non-word repetition, PCC 88.48 6.26 83.52 9.73 – –
Phonology phonemes correctly produced 99.17 2.32 98.11 7.68 – –
TROG maximum: 20 16.94 2.53 15.37 2.40 17.00 1.30
PPVT maximum: 109 96.17 6.50 93.07 6.80 98.88 5.05
Reading speed, words 56.29 20.18 59.63* 28.70 32.88 6.59
Reading speed, non-words 95.09 34.45 80.33* 29.13 65.15 17.18
Reading accuracy, words maximum: 30 28.50 2.04 27.82* 4.70 29.00 1.38
Reading accuracy, non-words maximum: 30 25.89 3.97 25.71* 5.23 25.82 4.18

*Task was not administered to all children with SLI (n~21–25) because they could not participate, i.e.
did not have sufficient reading skills.

Comparison of means for children with HI, SLI and NL

Figures 1a and b show a comparison between the children with HI and SLI. To
enhance comparability, all results were computed in mean per cent, except reading
speed, which was measured in seconds. The dotted lines in figures 1a and b indicate
means (%) for children with NL in the third group on novel word learning,
working memory, reading accuracy and speed according to Fries and Holmberg
(2001). The 38 children with NL performed at ceiling on non-word discrimination,
non-word repetition and output phonology at age 10 years. For the PPVT, a
reference group of 17 Swedish children aged 9;0–11;8 years reached a mean of
90.7% (mean 99.28, SD 5.18) on the PPVT. A maximum score of 109 was possible
(data collected by the present authors). According to Holmberg and Lundälv
(1998), Swedish children with NL at age 10 reach a mean of 83% on the TROG.
Since individual data from the children with NL were lacking, t-tests were used
to assess the differences between the results of the children with NL and the two
subject groups: the children with HI and SLI. To adjust for an elevated type I error
rate due to multiple comparisons, alpha levels were selected by dividing 0.01 and
0.05 by the number of comparisons made. The t-tests revealed that reading speed
for words was the only measure where the children with NL and the children with
HI differed significantly (d.f.~54, t~26.653, pv0.05). There were significant
differences between the children with NL and the children with SLI on four
measures: complex working memory (d.f.~63, t~4.492, pv0.05), TROG
(t~3.536, pv0.05), PPVT (t~3.903, pv0.05) and reading speed for words
(t~25.785, pv0.05).
No control data exist for non-word discrimination or output phonology, but
10-year-olds are reasonably expected to reach a ceiling on both. Non-word
repetition in children with SLI and HI could not be directly compared with
10-year-old controls in Fries and Holmberg (2000), since the same non-word
Working memory and novel word learning in children 411

Figure 1a and b. Group means (%) for children with hearing impairment and specific language
impairment. The dotted lines indicate the means for age-matched reference groups.
412 K. Hansson et al.

repetition test was not included in the procedure for this group. However, it is
known that children of this age perform at a ceiling at non-word repetition tests
with words obeying the phonotactic rules of their language (Simkin and Conti-
Ramsden 2000). Given a performance on non-word repetition in the SLI group
ranging from 47.7 to 96.4 (mean 79.1) and from 63.1 to 97.3 (mean 82.75) for the
HI group, it is reasonable to assume that a difference would have been found
between the children with HI and SLI and age controls.

Comparison between children with SLI and HI

Comparing children with SLI and HI, t-tests revealed that the children with HI
performed better on the following measures: complex working memory
(t(43)~2.82, pv0.01), novel word learning (retention) (t(43)~2.16, pv0.05),
language comprehension, i.e. TROG (t(43)~2.10, pv0.05), and reading accuracy for
both words (t(43)~264.47, pv0.001) and non-words (t(43)~229.73, pv0.001).
Overall, there was a great variation of performance in the subject groups.
Regarding the language comprehension test, the TROG, 28% of the children
with HI and 56% of the children with SLI performed below the 25th percentile.
Three children with HI and six children with SLI performed below the 10th
percentile.
On the RSPM, eight children with HI (44%) performed below the 25th
percentile. In the SLI group, 13 children (48%) performed below the 25th
percentile on the RSPM. In the HI group, all the children below the 25th percentile
on TROG were also below the 25th percentile on RSPM. Three were below on
RSPM only. In the SLI group, 10 children were below the 25th percentile on both
TROG and RSPM, three were below on RSPM only, and five were below on
TROG only.
In the complex working memory task, 49% of the words were recalled after
30 min of the children in the SLI group and 61% in the HI group. There were also
some qualitative differences between the groups. The distribution of overt errors
can be seen in figure 2. In both groups, primacy errors and target intrusions were
most common. In the children with SLI, they constituted 48%, and in the children
with HI, they constituted 74% of the overt errors. For both groups, few errors
were medial errors, semantic substitutions or within set repetitions. In the bottom
panel, the corresponding results from Ellis-Weismer et al. (1999) is shown for
comparison.
No significant differences were found for non-word repetition (per cent
consonants correctly repeated). The children with SLI and HI performed equally
well on two- and three-syllable non-words. However, there was a tendency for the
children with HI to perform slightly better than the children with SLI on four- and
five-syllable non-words (figure 3).

Within-group correlations
Correlations were computed using Pearson’s product moment correlation. The
correlations were based on raw scores transformed into z-scores.
As shown in tables 2a and b, there is no correlation between acquisition and
retention in the novel word-learning task for the children with HI or SLI. Since the
aim in the present paper was specifically to explore the relationship between novel
Working memory and novel word learning in children 413

word learning, complex working memory and phonological memory, the focus will
be on the correlations relevant for this question. Novel word learning correlates
with complex working memory in both groups of children. For the children with

Figure 2. Distribution of overt errors in the children with hearing impairment (median age 10;10
years) and specific language impairment (median age 10;0 years) in the present study and the
children in the study by Ellis Weismer et al. (1999).
414 K. Hansson et al.

SLI, the complex working memory task correlates with almost all measures (6 out
of 8). In the children with HI, complex working memory correlates significantly
with 3 out of 8 measures.
Phonological memory as assessed by non-word repetition did not correlate

Figure 3. Non-word repetition in two-, three-, four- and five-syllable non-words in children
with specific language impairment and hearing impairment.
 Working memory and novel word learning in children
Table 3a. Within-group correlations for children with specific language impairment

1 2 3 4 5 6 7 8 9 10
1. Word learning (acquisition)
2. Word learning (retention) n.s.
3. Working memory 0.515** 0.678**
4. Non-word repetition n.s. n.s. 0.640**
5. Non-word discrimination n.s. 0.390* 0.427* n.s.
6. PPVT n.s. n.s. 0.477* 0.581** n.s.
7. TROG n.s. 0.409* 0.613** n.s. 0.455* 0.492**
8. Reading speed, words n.s. n.s. 20.669** 20.736** n.s. n.s. n.s.
9. Reading speed, non-words n.s. n.s. 20.501* 20.535** n.s. n.s. n.s. 0.901**
10. Reading accuracy, words n.s. n.s. n.s. 0.552** n.s. n.s. n.s. 20.475* n.s.
11. Reading accuracy, non-words n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. 0.449*

*pv0.05; **pv0.01; n.s., non significant.

Table 3b. Within-group correlations for children with hearing impairment

1 2 3 4 5 6 7 8 9 10
1. Word learning (acquisition)
2. Word learning (retention) n.s.
3. Working memory 0.478* 0.548*
4. Non-word repetition n.s. n.s. n.s.
5. Non-word discrimination n.s. n.s. n.s. n.s.
6. PPVT 0.603** n.s. 0.525* n.s. n.s.
7. TROG n.s. n.s. n.s. n.s. n.s. n.s.
8. Reading speed, words 20.537* n.s. 20.495* n.s. n.s. n.s. n.s.
9. Reading speed, non-words n.s. n.s. n.s. n.s. n.s. n.s. n.s. 0.751**
10. Reading accuracy, words 0.485* n.s. 0.561* 0.511* n.s. n.s. n.s. 20.738** 0.538*
11. Reading accuracy, non-words 0.578* n.s. n.s. n.s. n.s. 0.486* n.s. n.s. 0.617** 0.626**

415
*pv0.05; **pv0.01; n.s., non significant.
416 K. Hansson et al.

significantly with novel word learning in the children with SLI or HI. Non-word
discrimination, however, correlates with novel word learning (retention) in the
children with SLI. Non-word repetition correlates with vocabulary knowledge
(PPVT) in the children with SLI but not in the children with HI.
As for the reading tests, the test of complex working memory, the CLPT,
correlates significantly with reading speed for words and non-words in the children
with SLI, and with reading accuracy and speed for words in the children with HI.
Non-word repetition correlates with reading accuracy for words in the children
with HI, and with reading speed and accuracy for words and reading speed for
non-words in the children with SLI.
A step-wise multiple regression analysis with novel word learning (retention) as
the dependent variable and non-word discrimination, non-word repetition and
complex working memory as independent variables revealed that the best predictor
for the retention of novel words in the children with SLI is complex working
memory (r~0.678**, r2~0.459). For the children with HI, complex working
memory is also the best predictor for novel word learning (r~0.548*, r2~0.300).
The other factors did not contribute in either group.

Discussion
The results should be interpreted with caution. There is a range of limitations in the
present study, some of them inevitable in studies carried out in a small country like
Sweden. Thus, the sample sizes are fairly small and there is also an overall lack of
standardized language measures. The discussion will first focus on the present
research question about similarities and differences between children with a
preschool diagnosis of SLI and children with HI on language and working memory.
Second, it will discuss the role of working memory in novel word learning.

Language skills in children with HI and in children with a preschool diagnosis

of SLI
Present language status was not used as a selection criterion for the children with
HI or for the children with a preschool diagnosis of SLI. However, in both groups,
there were children demonstrating linguistic weaknesses. The TROG is the only
existing standardized language measure for Swedish. Therefore, it was not possible
to confirm or disconfirm the diagnosis of SLI. According to many authors (e.g.
Baker and Cantwell 1987, Ekholm et al. 1999), SLI has long-lasting effects. In the
present study, almost 60% of the children with a preschool diagnosis of SLI
performed below the 25th percentile on the TROG and 22% performed below the
10th percentile at age 10 years. For three children with SLI, there seems to have
been a drop in IQ. Although their non-verbal IQ scores were w80 in preschool (at
age 5), their scores had now decreased to 73. Note that all the children with a
preschool diagnosis of SLI, except one girl, were placed in mainstream schools.
One girl was placed in a special school and she performed at the 10th percentile on
the TROG and still had obvious expressive language problems for grammar and
phonology. The majority of the children in this group still had individual support
from a special teacher.
Regarding the children with HI in the present study, the only statistically
significant difference found between them and the children with NL was for
Working memory and novel word learning in children 417

reading speed for words. They also performed slightly below the children with NL
on the novel word-learning task (acquisition phase), on the PPVT and on the
reading test (reading accuracy for words). On the TROG, 28% in the HI group
performed below the 25th percentile and 11% performed below the 10th
percentile. There is an increasing amount of evidence that late identification
hampers the linguistic and cognitive development in children with HI (Yoshinaga-
Itano et al. 1998). The children with HI in the present study differed considerably
regarding their age at diagnosis. According to Borg et al. (2000), the mean age at
identification for Swedish children with HI is 3;6. It is noteworthy that the children
in the present study were identified later (median age 4;3). Perhaps the linguistic
weaknesses found in one group of children with HI would not have been present if
their HI had been identified at an earlier age.

Working memory in children with HI and in children with a preschool diagnosis

of SLI
Group data revealed statistically significant differences between the children with
SLI and HI on complex working memory, where the children with HI had a
significantly better word recall. The fact that the children with HI were slightly
older (10 months) should be taken into consideration. However, the authors do not
believe that this fully explains the difference.
There were some qualitative differences between the groups for word recall in
the complex working memory task. Responses classified as ‘no response’ were more
frequent in children with HI (95% of all incorrect responses) than in children with
SLI (68%). The two most frequent errors categories among the overt errors were
primacy errors and target intrusions for both groups.
As shown in figure 2, the two most frequent categories of errors for children
with SLI and NL in Ellis Weismer et al. (1999) were also primacy errors and target
intrusions. The distributions of these errors were different for the two groups.
Children with SLI had 59% of primacy errors and 29% of target intrusions,
whereas children with NL demonstrated the opposite pattern, i.e. 36% of primacy
errors and 50% of target intrusions. Ellis Weismer et al.’s (1999) interpretation of
these findings is that the relatively higher frequency of target intrusions in children
with NL than in children with SLI reflect a more obvious interference from
activation of items stored in long-term memory. The relatively higher percentage of
primacy errors in children with SLI might reflect an inability to inhibit lexical
information within the task.
In the present study, the children with a preschool diagnosis of SLI and HI
demonstrated no difference for the relative distribution of the two types of errors.
The children with SLI responded less often with primacy errors and target
intrusions than the children with SLI in Ellis Weismer et al. (1999), probably
because the latter group of children was on average more than 1 year younger.
Regarding phonological short-term memory, there were in both groups children
scoring less than 80% of consonants correct on the non-word repetition test, on
which children at this age usually score 100% correct (Fries and Holmberg 2001).
Regarding non-word repetition, Briscoe et al. (2001) found that children with SLI
had relatively greater difficulties repeating longer non-words than children with HI.
A similar tendency was found in the present study. It is probable that if non-words
longer than five-syllables had been used, a similar difference between the groups
418 K. Hansson et al.

would have emerged. The present authors are currently exploring suprasegmental
aspects of non-word repetition (number of syllables and stress patterns) in the
children with SLI and HI. Preliminary data do not indicate any difference between
the groups for imitation of suprasegmental features. Both groups of children in the
present study were better at imitating ‘the whole than the parts’. The present data
so far seem to support the findings of Carter et al. (2002), who found that deaf
children with cochlear implants were significantly better at imitating suprasegmental
than segmental aspects of non-words.
Complex working memory correlated significantly with most measures of
language and reading in the children with SLI. For the children with HI, the
relationship was not so obvious. The present results might be an indication of a
stronger relationship between working memory and language in children with SLI
than in children with HI as proposed by Briscoe et al. (2001).

Novel word learning and working memory

The retention task in the novel word-learning test in the present study seemed to
be difficult for children with NL, SLI and HI (table 1). Few words were recalled
after 30 min in the three groups. Somewhat surprisingly, there was no correlation
between acquisition and retention of novel words in the group of children with SLI
or in the group of children with HI. A so-called missing link between acquisition
and retention was discussed by Towse et al. (1998), who challenged the concept of
a trade-off between processing and storing in working memory. In a series of
studies on concurrent storage loads, Towse et al. showed that processing speed in
children was not necessarily related to recall of memory items over time. Retention
duration is a more important general property of working memory in children
according to these authors.
Within-group correlations revealed that phonological memory as measured by
the per cent of consonants correct on non-word repetition did not predict novel
word learning in children with HI or SLI in the present study. A lack of correlation
should be interpreted with caution since the number of children was small. This
negative finding, however, is in line with findings by Gilbertson and Kamhi (1995).
The lack of a relationship between phonological short-term memory and recall of
novel words has been explained by Metsala (1999). Vocabulary growth in children
around 10 years of age might have improved the segmented lexical skills and
thereby reduced the dependency on phonological memory in vocabulary
acquisition. Metsala suggested that stored phonological representations in long-
term memory fill in incomplete traces in phonological short-term memory before
output. For the children with HI and SLI in the present study, complex working
memory seemed to contribute more to the learning of new words than
phonological short-term memory. For the children with NL, multiple regressions
were not made, but the correlation between the non-word repetition task (Russian
words) and novel word learning hardly reached significance, whereas the correlation
between complex working memory and novel word learning turned out to be solid
(pv0.01). The present authors are currently comparing 5–6-year-olds with HI and
SLI using a similar procedure as in the present study. There is reason to expect a
stronger influence of phonological memory on novel word learning in younger
children. For children with NL at age 5–6 years, a strong correlation between non-
word repetition and novel word learning was found by Hörman and Kring (2001).
Working memory and novel word learning in children 419

The complex working memory task assessing simultaneous processing and

storing of verbal information had the strongest predictive power for novel word
learning not only in the children with SLI, but also in the children with HI in the
present study. This was also shown for the children with NL by Fries and
Holmberg (2001). The present results seem to support theories emphasizing the
role of complex working memory in vocabulary acquisition particularly for older
children.
In speech production, there were few phonological errors and only some
articulation errors in both groups (table 2). Only one child with a preschool
diagnosis of SLI still had a phonological impairment. Phonological output
constraints can thus hardly explain the weak performance on non-word repetition
in the children with a preschool diagnosis of SLI and in the children with HI.
Adams and Gathercole (2000) found support for a ‘mutual input constraints’
hypothesis for the underlying deficit in SLI. The present results, however, do not
seem to support this hypothesis. At least Adams and Gathercole (2000) do not
indicate a common underlying process for non-word repetition and non-word
discrimination for children this age. Neither in children with a preschool diagnosis
of SLI nor in children with HI did the present authors find a correlation between
non-word repetition and non-word discrimination, which is in line with the results
of Bishop (1999), who also found a lack of a relationship between measures of
phonological processing and non-word repetition.

Conclusions
The first research question regarded the similarities and differences of children with
HI and SLI. Group data revealed significant differences between the children with a
preschool diagnosis of SLI and the children with HI. The children with HI
performed better on tests of complex working memory, novel word learning and
language comprehension, and they were more accurate at reading words and non-
words. No significant difference was found for phonological short-term memory
(non-word repetition). In longer non-words, the children with a preschool
diagnosis of SLI seemed to have more difficulties than the children with HI, but
the finding did not reach significance. Complex working memory was linked to all
language measures in the children with SLI. In the children with HI, the link was
less evident. Thus, the speculation is that there is a closer connection between
working memory and language in children with SLI than in children with HI.
The second research question regarded the role of working memory in novel
word learning and the contribution of the two different memory systems. The
results showed that novel word learning was best predicted by complex working
memory in both groups of children. This finding must be considered as being fairly
robust, whereas the lack of correlation between phonological memory and novel
word learning must be interpreted with caution. However, the interpretation is that
the contribution of complex working memory to novel word learning is stronger
than the contribution of phonological short-term memory for children in this age.
The clinical implications are evident. Language and working memory skills
should be thoroughly examined even in children with mild HI. Working memory
tasks should be included to a higher extent in the assessment procedure for
children with SLI. However, in order to be used for younger children, the complex
working memory task must be modified. The authors are currently using a simpler
420 K. Hansson et al.

version of the CLPT and visual working memory task in a comparative study of
5–6-year-old children with SLI and HI.

Acknowledgements
The authors thank audiologists Eva Holst, Christina Öfverberg-Nestor and Ingrid
Lennart, who performed the hearing screening of the children. The authors are very
grateful to all the children and parents for their participation, to Master’s students
for collecting normative data, and to Tina Ibertsson for help with the manuscript.
The study was financed by a grant (No. 2000-0171:01) from the Bank of Sweden
Tercentenary Foundation.

References
ADAMS, A.-M. and GATHERCOLE, S. E., 1995, Phonological working memory and speech production in
preschool children. Journal of Speech and Hearing Research, 38, 403–414.
ADAMS, A.-M. and GATHERCOLE, S. E., 2000, Limitations in working memory: implications for language
development. International Journal of Language and Communication Disorders, 35, 95–116.
BADDELEY, A., 2000, The episodic buffer: a new component of working memory? Trends in Cognitive
Sciences, 4, 417–423.
BADDELEY, A., GATHERCOLE, S. and PAPAGNO, C., 1998, The phonological loop as a language learning
device. Psychological Review, 105, 158–173.
BAKER, L. and CANTWELL, D. P., 1987, A prospective psychiatric follow-up of children with speech/
language disorders. Journal of the American Academy of Child and Adolescent Psychiatry, 26, 546–553.
BARTHELOM, E. and ÅKESSON, M., 1995, Konstruktion, testning och utvärdering av nonord.
Examensarbete i logopedi (Lund: Institute for Logopedics and Phoniatrics, Lund University).
BISHOP, D. V. M., 1983, The Test for Reception of Grammar (TROG) (available from the author at the MRC
Applied Psychology Unit, 15 Chaucer Road, Cambridge CB2 2EF, UK).
BISHOP, D. V. M., 1999, Different origin of auditory and phonological processing problems in children
with language impairment: evidence from a twin study. Journal of Speech, Language and Hearing
Research, 42, 155–168.
BISHOP, D. V. M., NORTH, T. and DONLAN, C., 1996, Nonword repetition as a behavioural marker for
inherited language impairment: evidence from a twin study. Journal of Child Psychology and
Psychiatry, 37, 391–403.
BORG, A., MCALLISTER, B., RISBERG, A., UNDEMAR, B.-M., EDQUIST, G., REINHOLDSON, A.-C., WIKING-
JOHNSSON, A. and WILLSTEDT-SVENSSON, U., 2000, Hörselskadade barns språkutveckling (HSS). Rapport
från Ahlséns Forskningsinstitut, 2B.
BOWEY, J. A., 1996, On the association between phonological memory and receptive vocabulary in five-
year-olds. Journal of Experimental Child Psychology, 63, 44–78.
BRISCOE, J., BISHOP, D. V. M. 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.
CARTER, A., DILLON, C. and PISONI, D., 2002, Imitation of non-words by hearing impaired children with
cochlear implants: suprasegmental analysis. Clinical Linguistics and Phonetics, 16, 619–638.
DAVIS, J. M., ELFENBEIN, J., SCHUM, R. and BENTLER, R. A., 1986, Effects of mild and moderate hearing
impairments on language, educational, and psychological behaviour of children. Journal of Speech
and Hearing Disorders, 51, 53–62.
DUNN, L. M. and DUNN, L. M., 1981, Peabody Picture Vocabulary Test—Revised (Circle Pines: American
Guidance Services).
EDWARDS, J. and LAHEY, M., 1998, Nonword repetitions of children with specific language impairment:
exploration of some explanations for their inaccuracies. Applied Psycholinguistics, 19, 279–309.
ELLIS WEISMER, S. and EVANS, J. L., 2002, The role of processing limitations in early identification of
specific language impairment. Topics in Language Disorders, 3, 15–29.
ELLIS WEISMER, S., EVANS, J. and HESKETH, L. J., 1999, An examination of verbal working memory
Working memory and novel word learning in children 421

capacity in children with specific language impairment. Journal of Speech, Language and Hearing
Research, 42, 1249–1260.
ELLIS WEISMER, S. and HESKETH, L., 1996, Lexical learning by children with specific language
impairment: effects of linguistic input presented at varying speaking rates. Journal of Speech and
Hearing Research, 39, 177–190.
FRIES, P. and HOLMBERG, A., 2001, Verbal working memory, novel word learning and reading in 9–12
year old Swedish children with normal language development. Masters thesis, Department of
Logopedics and Phoniatrics, Lund University.
GATHERCOLE, S. E., 1999, Cognitive approaches to the development of short-term memory. Trends in
Cognitive Sciences, 3, 410–419.
GATHERCOLE, S. E. and BADDELEY, A. D., 1990, Phonological memory deficits in language disordered
children: is there a causal connection? Journal of Memory and Language, 29, 336–360.
GATHERCOLE, S. E. and PICKERING, S. J., 1999, Estimating the capacity of short-term memory. International
Journal of Psychology, 34, 378–382.
GATHERCOLE, S. E., FRANKISH, C. R., PICKERING, S. J. and PEAKER, S., 1999, Phonotactic influences
on short-term memory. Journal of Experimental Psychology: Learning, Memory and Cognition, 25,
84–95.
GATHERCOLE, S. E., WILLIS, C. S., BADDELEY, A. D. and EMSLIE, H., 1994, The children’s test of non-word
repetition: a test of phonological working memory. Memory, 2, 103–127.
GATHERCOLE, S. E., WILLIS, C. S., EMSLIE, H. and BADDELEY, A. D., 1992, Phonological memory and
vocabulary development during the early school years: a longitudinal study. Developmental
Psychology, 28, 887–898.
GAULIN, C. and CAMPBELL, T., 1994, Procedure for assessing verbal working memory in normal school-
age children: some preliminary data. Perceptual and Motor Skills, 79, 55–64.
GILBERTSON, M. and KAMHI, A. G., 1995, Novel word learning in children with hearing impairment.
Journal of Speech and Hearing Research, 38, 630–642.
HELLQVIST, B., 1991, Fonemtestet (Malmo: Pedagogisk Design).
HOLMBERG, B. and SAHLÉN, B., 2000, Nya Nelli—neurolingvistisk undersökning för barn med språkstörningar
(Malmo: Pedagogisk Design).
HOLMBERG, E. and LUNDÄLV, E., 1998, T. R. O. G. (Gothenburg: SIH Läromedel).
HÖRMAN, C. and KRING, M., 2001, Novel word learning, phonological short term memory and
phonological processing in Swedish children with normal language development aged 4 to 6.
Masters thesis, Department of Logopedics and Phoniatrics, Lund University.
HUTTUNEN, K. H., 2001, Phonological development in 4–6-year-old moderately hearing impaired
children. Scandinavian Audiology, 30 (Suppl. 53), 79–82.
JUST, M. and CARPENTER, P., 1992, A capacity theory of comprehension: individual differences in
working memory. Psychological Review, 99, 122–149.
LYXELL, B. and HOLMBERG, I., 2000, Visual speech-reading and cognitive performance in hearing-
impaired and normal hearing children (11–14 years). British Journal of Educational Psychology, 70,
505–518.
METSALA, J. L., 1999, Young children’s phonological awareness and nonword repetition as a function of
vocabulary development. Journal of Educational Psychology, 91, 3–19.
MOELLER, M. P., 2000, Early identification and language development in children who are deaf and
hard of hearing. Pediatrics, 106, 1–9.
MONTGOMERY, J. W., 2003, Working memory and comprehension in children with specific language
impairment: what we know so far. Journal of Communication Disorders, 36, 221–231.
MOORE, B. C. J., 1995, Perceptual Consequences of Cochlear Damage (Oxford: Oxford Medical).
NYMAN, A., 1999, Nonordsrepetition, nonordsdiskrimination och metafonologisk förmåga. Masters
thesis, Department of Logopedics and Phoniatrics, Lund University.
POHJANEN, A. and SANDBERG, M., 1999, Arbetsminnet hos svenska fem-, sju- och nioåriga barn med
normal språkutveckling. Masters thesis, Department of Logopedics and Phoniatrics, Lund
University.
RAVEN, J. C., COURT, J. H. and RAVEN, J., 1990, Raven’s Standardized Progressive Matrices (Oxford: Oxford
Psychologists Press).
SAHLÉN, B., REUTERSKIÖLD WAGNER, C., NETTELBLADH, 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.
SIMKIN, Z. and CONTI-RAMSDEN, G., 2000, Non-word repetition and grammatical morphology:
422 Working memory and novel word learning in children

Normative data for children in their final year of primary school. International Journal of Language
and Communications Disorders, 36, 395–404.
TOWSE, J., HITCH, G. J. and HUTTON, U., 1998, A re-evaluation of working memory capacity in children.
Journal of Memory and Language, 39, 195–217.
YOSHINAGA-ITANO, C., SEDEY, A. L., COULTER, D. K. and MEHL, A. L., 1998, Language of early and later
identified children with hearing loss. Pediatrics, 102, 1161–1171.