Professional Documents
Culture Documents
discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/221971050
CITATIONS
READS
22
453
1 author:
Marilyn A Nippold
University of Oregon
121 PUBLICATIONS 1,897 CITATIONS
SEE PROFILE
AJSLP
Viewpoint
American Journal of Speech-Language Pathology Vol. 21 183196 August 2012 A American Speech-Language-Hearing Association
183
185
Anderson, Pellowski, and Conture (2005) examined language development in preschool children (n = 90) ages 3;0
to 5;11 (Mage = 4;1). Half of the children stuttered (CWS)
and half had normal fluency (CWNS). All children spoke
Standard American English. The two groups were matched
on the basis of age, gender, race, and SES. To qualify for the
study, all children in the control group had to score at the
20th percentile or higher on a battery of norm-referenced
language tests. However, no such screening procedure was
used with children in the stuttering group, allowing their
language skills to vary. To examine language development,
the investigators administered the following tests to each
child in the study: the Peabody Picture Vocabulary Test
Third Edition (PPVTIII; Dunn & Dunn, 1997), the Expressive Vocabulary Test (EVT; Williams, 1997), and the
Test of Early Language Development, Third Edition (TELD3;
Hresko, Reid, & Hammill, 1999). Collectively, these
tests evaluated the childrens syntactic, morphologic, and
lexical development. The results showed that both groups
performed within the normal range on all tests, and that they
did not differ significantly on the PPVTIII or the EVT.
However, the CWNS outperformed the CWS on the receptive and expressive subtests of the TELD3. Although there
was no evidence of language disorders in the CWS, the
authors suggested that the overall language abilities of
the CWS were lower than those of the CWNS.
Before accepting this conclusion, caution is advised. Although the two groups had been matched for age, gender, race,
and SES, all children in the control group had to have
scored at the 20th percentile or higher on a set of normreferenced language tests, but no such requirement was made
for the children in the stuttering group, whose language
skills were free to vary. This type of differential subjectselection procedure (Schiavetti & Metz, 2006, p. 144) constitutes a serious threat to the studys internal validity. Had
the language skills of all children in the Anderson et al.
(2005) investigation been free to vary, the outcome might
have been different.
In an effort to expand the Anderson et al. (2005) study,
Coulter, Anderson, and Conture (2009) administered the same
set of norm-referenced language tests used by Anderson
et al. to additional groups of CWS and CWNS (n = 40 per
group). The children in this study ranged in age from 3;0 to
5;8 (Mage = 4;0), all spoke Standard American English, and
none had participated in the previous study. As before, to
be admitted to the control group, children were required to
have performed at least within the average range on a normreferenced language test (standard score 85), but children in
the stuttering group were not required to meet this criterion.
As with the previous study, both groups performed within
normal limits on all language tests, including the PPVTIII,
the EVT, and the receptive and expressive subtests of the
TELD3. Then, to analyze the data with greater statistical
power, the investigators combined the scores from this investigation with the scores from all of the participants in the
Anderson et al. study. The results of this analysis indicated
that the CWNS (n = 85) significantly outperformed the CWS
(n = 85) on all of the norm-referenced language tests.
Again, a note of caution is advised before concluding that
preschool CWS have weaker language skills than their peers
187
studies that examined the language abilities of CWS compared to CWNS. Studies were included in their meta-analysis
if the participants were between 2 and 8 years of age, had
been assessed using norm-referenced language tests or
language sampling tasks, and had not received a previous
diagnosis of language impairment. Studies that examined
phonological development were excluded, as the focus of the
review was on syntactic, morphologic, and lexical development. For each study included in the meta-analysis, group
differences between CWS and CWNS on the language measures were examined by calculating effect sizes. The results indicated that CWS scored significantly lower than
CWNS on measures of overall language ability, receptive
vocabulary, expressive vocabulary, and MLU. Despite these
group differences, however, in most of the studies, CWS still
performed within normal limits on the language measures,
confirming that they did not have language disorders. For
this reason, Ntourou et al. suggested that CWS, as a group,
demonstrate subtle differences in language abilities when
compared to their normally fluent peers (p. 173) and that
language may be an influential variable associated with
the difficulty that some CWS have establishing normally
fluent speech (p. 176).
Before accepting these conclusions, several issues must
be considered. First, it is essential that studies that are designed to compare the language abilities of CWS to those
of CWNS ensure that their groups are matched on factors
that are known to influence childrens performance on language tasks (see Nippold, 1990; Watkins & Johnson, 2004).
Most importantly, these include age, gender, and SES or
factors related to SES such as maternal education (Dollaghan
et al., 1999; Hoff, 2003; Hoff & Tian, 2005; Turnbull &
Justice, 2012). Of the 22 studies included in the metaanalysis by Ntourou et al. (2011), only four (18%) matched
the stuttering and nonstuttering groups on age, gender, and
SES; 13 (59%) matched them on age and gender; three
(14%) matched them only on age; and two (9%) matched
them only on gender (see Ntourou et al., 2011, Table 1,
pp. 167171). Because children from higher SES levels frequently score higher on language measures than children
from lower SES levels (Hart & Risley, 1995; Hoff, 2003;
Hoff & Tian, 2005), SES must be controlled in studies that
compare CWS and CWNS on language. Without properly
matching the groups, it is premature to draw conclusions
regarding language differences between them.
It should also be noted that of the four studies included in
the meta-analysis (Ntourou et al., 2011) that matched the
groups on age, gender, and SES (Anderson, 2008; Anderson
et al., 2005, 2006; Silverman & Bernstein Ratner, 2002),
only two of them (Anderson et al., 2005; Silverman &
Bernstein Ratner, 2002) found statistically significant
differences between the groups on the norm-referenced
language tests, with both favoring the CWNS. However,
these results should be interpreted cautiously. In the Anderson
et al. (2005) study, a prescreening process had been employed such that the CWNS were required to have scored
at the 20th percentile or higher on the language tests,
but the CWS were not required to meet this criterion; thus,
the language skills of the CWS were free to vary, making it
difficult to interpret the studys findings. As mentioned
189
it seems that scores on the CNRep and scores on the language tests would bear at least a weak relationship to one
another, with statistically significant coefficients possibly in
the low to low-moderate range. The absence of such findings
casts further doubt on the claim that CWS have subtle
language deficits.
It is also unclear to what extent a childs overt or covert
stuttering behavior may interfere with his or her ability to
repeat nonsense words such as brasterer, glistering, and
empliforvent (Gathercole et al., 1994). If stuttering is a deficit
in speech motor programming that hampers a speakers
ability to advance through a word, as some have argued
(Packman, Code, & Onslow, 2007), then poor performance
on nonword repetition tasks may have little to do with language ability in CWS. Rather, it may reflect motoric difficulties with speech production. Support for this position
comes from a growing body of research on neurogenic
factors associated with developmental stuttering. For example, in discussing this body of research, Watkins, Smith,
Davis, and Howell (2008) stated that stuttering is a disorder
related primarily to disruption in the cortical and subcortical neural systems supporting the selection, initiation and
execution of motor sequences necessary for fluent speech
production (p. 50). If future research confirms the view
that childhood stuttering is primarily a deficit in speech
motor programming, then continued efforts to link the
disorder to language deficits in CWS are likely to be
unproductive.
191
Conclusions
Studies published in peer-reviewed journals since 1990
that addressed a connection between stuttering and language
ability in preschool and school-age children do not provide
strong evidence to support the view that stuttering and
language ability are linked. Despite multiple efforts using
varied research designs, no study found convincing evidence
of language deficits in CWS compared to CWNS in terms of
receptive or expressive syntactic, morphologic, or lexical
development. Rather, the studies reported that CWS, like
CWNS, manifest the full range of language abilities. Other
critical findings that fail to support such a connection include
the fact that scores on language measures were not significantly correlated with the frequency of stuttering behaviors,
that stuttering onset or persistence was not associated with
language deficits, and that CWS use age-appropriate language even during challenging discourse tasks.
Future research. Nevertheless, given the enduring appeal of this topic and the persistent belief that CWS, as a
TABLE 1. Correlation coefficients (with p values) calculated
by this author (Nippold) between measures of language ability
(Test of Early Language Development, Second Edition [TELD2;
Hresko, Reid, & Hammill, 1991], Peabody Picture Vocabulary
Test, Third Edition [PPVTIII; Dunn & Dunn, 1997], mean length of
utterance [MLU]) and measures of stuttering severity (Stuttering
Severity Instrument [SSI; Riley, 1994], within-word disfluency
[WWD]) for children who stutter, using data provided by Anderson
and Conture (2000) (see their Table 1, p. 287).
SSI
WWD
Note.
TELD2
PPVTIII
MLU
.22 (p = .35)
.42 (p = .06)
.18 (p = .45)
.20 (p = .40)
.37 (p = .13)
.52 (p = .03)*
*statistically significant.
syntax is not being questioned, nor is the association between frequency of stutters and utterance length and complexity. Rather, the point is that these factors alone do not
provide logical evidence of a link between stuttering and
language ability, and studies that have addressed the proposed
link do not provide empirical evidence of it.
An alternative perspective is to regard stuttering as a
speech disorder involving a motor control deficit, and not a
language disorder as Bloodstein (2006) had argued. In describing developmental stuttering, Olander, Smith, and
Zelaznik (2010) explained that during the disfluencies that
characterize stuttering, the speech motor system fails to
generate and /or send the motor commands to muscles that
are necessary for fluent speech to continue (p. 876). Similarly, as argued by Packman et al. (2007), developmental
stuttering is a problem in syllable initiation in which the
child is unable to move forward in speech because the speech
planning system is compromised. Further, they explained
that this difficulty is first noticed when the child attempts
to produce multisyllabic utterances requiring complex
sequential movements and varied linguistic stress patterns
across syllables to communicate the intended meaning.
According to Packman et al., children do not stutter when
babbling or producing first words (Bloodstein, 2006) because
these additional speech motor demands are not yet present.
In closing, this article has explained why it is reasonable
to question the view that stuttering and language ability are
linked. As with any long-running debate, this effort will
likely generate many passionate discussions among readers.
It is therefore hoped that this response will prompt new ideas
for studies that will continue to expand the knowledge
base in stuttering.
References
Adams, M. R. (1990). The demands and capacities model I: Theoretical elaborations. Journal of Fluency Disorders, 15, 135141.
Anderson, J. D. (2008). Age of acquisition and repetition priming
effects on picture naming of children who do and do not stutter.
Journal of Fluency Disorders, 33, 135155.
Anderson, J. D., & Conture, E. G. (2000). Language abilities of
children who stutter: A preliminary study. Journal of Fluency
Disorders, 25, 283304.
Anderson, J. D., Pellowski, M. W., & Conture, E. G. (2005).
Childhood stuttering and dissociations across linguistic
domains. Journal of Fluency Disorders, 30, 219253.
Anderson, J. D., Wagovich, S. A., & Hall, N. E. (2006).
Nonword repetition skills in young children who do and do not
stutter. Journal of Fluency Disorders, 31(3), 177199.
Arndt, J., & Healey, E. C. (2001). Concomitant disorders in
school-age children who stutter. Language, Speech, and
Hearing Services in Schools, 32, 6878.
Bajaj, A. (2007). Analysis of oral narratives of children who stutter
and their fluent peers: Kindergarten through second grade.
Clinical Linguistics and Phonetics, 21(3), 227245.
Bernstein Ratner, N. (1997). Stuttering: A psycholinguistic
perspective. In R. F. Curlee & G. M. Siegel (Eds.), Nature and
treatment of stuttering: New directions (2nd ed., pp. 99127).
Boston, MA: Allyn & Bacon.
Bernstein Ratner, N., & Sih, C. C. (1987). Effects of gradual increases in sentence length and complexity on childrens dysfluency.
Journal of Speech and Hearing Disorders, 52, 278287.
Nippold: Stuttering and Language
193
195
Yaruss, J. S., LaSalle, L. R., & Conture, E. G. (1998). Evaluating stuttering in young children: Diagnostic data. American
Journal of Speech-Language Pathology, 7(4), 6276.
Zimmerman, I., Steiner, V., & Pond, R. (1979). Preschool
Language Scale. San Antonio, TX: The Psychological
Corporation.
Zimmerman, I., Steiner, V., & Pond, R. (1992). Preschool
Language ScaleThird Edition. San Antonio, TX: The
Psychological Corporation.