Non-Word Repetition (NWR) - Is it a linguistic marker of SLI?
Reading: Graf Estes, K. Evans, J. & Else-Quest, N. 2007. Differences in the Nonword
Repetition: Performance of Children With and Without Specific Language Impairment: A
Meta-Analysis. Journal of Speech, Language and Hearing Research, 5,177-195
Children with Specific Language Impairment show difficulties in repeating multisyllabic nonwords (Bishop, North, & Donlan, 1996; Dollaghan & Campbell’ 1998; Edwards & Lahey,
1998; Ellis Weismer et al., 2000; Gathercole & Baddeley, 1990; Montgomery, 1995), with a
significant chromosome linkage (The SLI Consortium 2002, 2004)
What is a Non-Word Repetition task?
In a non-word repetition task the child in asked to repeat non-words on 1-5 syllables, with or
without a cluster at the onset, middle or coda.
Sample words for English (from GAPS - GRAMMAR AND PHONOLOGY SCREENING
TEST, van der Lely): dep, pif, dremp, klest, tobilf, difimp, bademper, fakester, padrepper,
difrimple
What contributes to the complexity of these words? Are they English like? Could they be
used as is in a Hebrew test? Which considerations should apply?
What is NWR testing?
Phonological working memory (phonological loop)
“Working memory (WM) according to Baddeley (1986) is a multicomponent, capacitylimited system that comprises a controlling ‘‘central executive’’ and that includes an
articulatory loop system. The central executive ... is thought to regulate information flow
within WM, retrieval of information from other memory systems, and the processing and
storage of information. The articulatory loop … includes a capacity-limited phonological
short-term store and an articulatory control process … that acts to refresh and maintain speech
material in the store for a brief period. The articulatory loop’s function is to store verbal input
temporarily, especially novel phonological input (Baddeley, Gathercole, & Papagno, 1998),
while other cognitive tasks such as auditory comprehension take place. The ability to
temporarily store novel material also allows the listener the opportunity to create long-term
phonological representations of that material (Baddeley et al8991).” (Montgomery 2003, p.
222)
Digit Span
Gray, S. 2003. Diagnostic accuracy and test-retest reliability of nonword repetition and digit
span tasks administered to preschool children with specific language impairment. Journal of
Communication disorders 36, 129-151
SLI (4-5;11)
NL (age)
Digit Span
Mean
S.D.
1.86
0.99
3.73
1.78
Min
0.00
1.00
Max
4.00
9.00
NWR [N=20]
Mean
S.D.
8.68
3.39
16.41
1.79
Min
3.00
12.00
Max
18.00
19.00
Linguistic Knowledge
Knowledge of phonotactic components
Knowledge of lexical composition



Size of vocabulary
Abstract representation / Poor representational system
Less efficient mechanism for using lexical knowledge
Speech output
Questions:
Does morpho-lexical information help?
Is there a difference between TD children and children with SLI
Some Standardized NWR tests:
Children's Test of Nonword Repetition (CNRep) - Gathercole and Baddeley (1996)
Nonword Repetition Test (NRT) - Dollaghan & Campbell (1998)
How are they different?
# of words
# of Syllable
Clusters
Weak syllable with reduced
vowels (hampent)
Lexical components
(morphemes)
Prosodic pattern
Score
Test
CNRep
40
2-5
+
+
NRT
16
1-4
- (CV structure, no late acquired
phonemes)
- (acoustically salient only)
+
-
Natural for English
Online
(correct/incorrect)
Morpho/Lexical abilities
+ ….
Equal stress on each syllable
From transcription (percentage of
correct phoneme)
Basic phonological processing or
memory skills
Hypothesis:
If ____________________________________
then ______________________________________
0-Hypothesis:
If ____________________________________
then ______________________________________
Archibald, L. M. D. & Gathercole, S. E. 2006. Nonword repetition: A comparison of
tests. Journal of Speech, Language and Hearing Research, 49:5, 970-983.
Subjects: 36 native speakers: 12 SLI (9;8), 12 TD age (9;9), 12 TD language (6;1). Same SES
Scores on standardized test:
Findings
Within group: CNRep is easier than NRT
Across group: The gap is more significant on CNRep
What influences these results?
What do the findings tell us about the predictions?
Graf Estes, K. Evans, J. & Else-Quest, N. 2007. Differences in the Nonword Repetition:
Performance of Children With and Without Specific Language Impairment:A MetaAnalysis. Journal of Speech, Language and Hearing Research, 5,177-195
1. Are all versions of the nonword repetition task interchangeable?
2. Are there age differences in the magnitude of the nonword repetition deficit in children
with SLI?
3. Are there circumstances under which the nonword repetition skills of children with SLI are
comparable to those of children with NL?


Dissimilar effect sizes for the nonword repetition deficit in SLI >> “design
characteristics of the measures affect performance and indicate that the measures are
not interchangeable gauges of the same deficit.” (p.182)
Similar magnitude effect sizes >> “despite differences in the characteristics of the
measures such as wordlikeness, complexity, and nonword length, the magnitude of the
difference between children with and without SLI is robust.” (p 182)
Effect size (d), is the difference between the scores of control group (children with NL) and
children with SLI, divided by the pooled/mean standard deviation for both groups.
d = (Mean Control – Mean SLI) / Mean SD (within group)
Positive effect sizes indicate higher performance by control group children (d = 1.0 indicates
that children with NL performed one standard deviation better than children with SLI).
The magnitude of this deficit varied significantly across studies. Children with SLI performed
between 0.62 (Washington&Craig, 2004) and 4.34 standard deviations (Gathercole &
Baddeley, 1990) below children with NL.
1) Measures of nonword repetition are not interchangeable. The CNRep and the NRT
yielded the most extreme effect sizes.
2) No significant relation between effect size and the age of children in the SLI samples.
3) Children with SLI were impaired at all nonword lengths, with significantly greater
deficits for longer nonwords (three and four syllables) than for shorter nonwords (one
and two syllables).
“The CNRep may differentiate some groups of children with and without SLI to a greater
degree than the NRT. The CNRep could tap subtle articulation difficulties and a lack of
support from lexical or sublexical mediation, in addition to problems with phonological
memory resources ... It suggests that measures that compound influences on performance
promote the largest differences in performance between children with and without language
impairments.” (p. 193)