Dodd, B., Holm, A., Crosbie, S., & McIntosh, B. Core vocabulary intervention for
inconsistent speech disorder. pp117-136
In L. Williams, S. McLeod, & R. McCauley (Eds.), Interventions for speech sound disorders
in children. Baltimore: Brookes.
Abstract
Core vocabulary intervention establishes consistent word production for a functional SSD
characterized by inconsistent pronunciation of the same lexical item, in the absence of CAS. The
intervention is suitable for children from two years of age, including those from bilingual
backgrounds and those with a cognitive disability. Children are seen twice weekly for thirty
minutes, for about eight weeks. Carers are required to carry out daily practice. Clients select a
vocabulary of around 70 functionally powerful words. Each week, children learn to produce their
best pronunciation of up to ten of these words consistently, in isolation and connected speech.
Non-treated probes measure generalization of consistency. Efficacy studies reveal gains in
consistency and accuracy of speech production.
Introduction
Children with speech sound disorders differ in severity, the types of errors made and their
response to specific intervention approaches. Given the complexity of speech processing, it is not
surprising that deficits to different abilities cause differences in speech output (Duggirala and
Dodd, 1991). There are three interactive domains: input processing (sensation, perception);
cognitive-linguistic abilities (phonological representation, derivation of language specific
phonological constraints); and, output processing (assembly of phonological plans or templates,
phonetic planning and articulation). This chapter focuses on intervention for children who have a
deficit in phonological assembly: an impaired ability to plan the sequence of phonemes that make
up a word, in the absence of any oro-motor signs of childhood apraxia of speech (CAS). The
deficit results in inconsistent pronunciation of the same word.
Inconsistent speech errors
Inconsistency characterized by multiple error types (unpredictable variation between a relatively
large number of phones) indicates pervasive speech processing difficulties (Grunwell, 1982) and
is a potential indicator of persistent SSD (Forrest, Elbert and Dinnsen, 2000). Children who make
inconsistent speech errors pronounce the same words and phonological features differently not
only from context to context, but also within the same context (McCormack and Dodd, 1996;
Holm and Dodd, 1999). For example, one 7 year old boy (Dodd, Holm, Crosbie and McIntosh,
2007) was likely to pronounce the same word differently each time he said it (e.g., [bns],
[dn], [bm] tongue; [bw::t], [bw], [bwt] witch; [deuwa], [jeia] [jedw] zebra).
The proportion of children with SSD who make inconsistent errors is around 10%. For example,
Broomfield and Dodd (2004) assessed 320 English-speaking children with SSD, finding that 30
(9.4%) pronounced at least 40% of 25 words differently when they were asked to name the same
pictures on three separate trials in one assessment session, each trial separated by a different
activity. Cross-linguistic evaluations have identified about the same proportion of children with
inconsistent speech in a range of languages indicating that the deficit in the speech processing
Dodd, Holm, Crosbie and McIntosh
2
chain is not dependent on the phonology being learned (see Holm, Crosbie and Dodd, 2005, for
review).
Differential diagnosis: inconsistent SSD and childhood apraxia of speech (CAS)
Table 1 summarizes clinical opinion (Forrest, 2003) and research (reviewed by Ozanne, 2005)
concerning criteria for the diagnosis of CAS. The characteristics of errors made by children with
inconsistent speech disorder are drawn from Holm, Crosbie and Dodd (2005). Two symptoms are
shared: inconsistent errors and increasing errors with increased word length. The two groups are
discriminated, however, by two important differences. Children with CAS have oro-motor
problems that result in inappropriate oral movements during speech (e.g. groping) and affecting
speech prosody, rate and fluency. Children with inconsistent SSD do not show these symptoms.
Children with CAS make more errors in imitation than they do in spontaneous production. The
opposite is true of children with inconsistent SSD: they make fewer errors in imitation than in
spontaneous production. Differential diagnosis is, then, straightforward.
Table 1. Differential diagnosis of CAS and inconsistent speech disorder
Childhood Apraxia of Speech
Inconsistent Speech Disorder
Inconsistent errors
Inconsistent errors
Increasing errors with increasing length
Increasing errors with increasing length
Poor sequencing of sounds (e.g., metathesis)
Wrong choice of phoneme rather than order
errors as in metathesis
Inability to imitate sounds, better spontaneously Better in imitation than in spontaneous
than in imitation
production
General oral-motor difficulties;
Oromotor skills within normal limits
Groping; silent posturing
No groping, no silent posturing
Prolongations and repetitions of speech sounds. No prolongations and repetitions of speech
sounds
Slow speech and diadochokinetic (DDK) rates
Normal speech and DDK rates
Intervention
Describing and analyzing inconsistent speech errors in terms of error patterns is not possible and
deciding what to target in intervention is difficult (Dodd and Bradford, 2000). Forrest et al.
(2000) stated that intervening with children who make inconsistent errors is problematic because
“one may not know the appropriate sound to use in contrast to the error. This may mean that
children with a variable substitution will fare worse in treatment than other children because the
available protocols for this population are not as effective as other procedures” (p. 529). An
alternative intervention, the core vocabulary approach, has been developed that targets
consistency, rather than accuracy, of whole word production (Dodd and Iacono, 1989). It has
been trialed with children as young as two years as well as preschool and school aged children
with a functional SSD who make inconsistent errors. The approach has also been successfully
used with a bilingual child (Holm and Dodd, 1999) and children who have Down syndrome
2
Dodd, Holm, Crosbie and McIntosh
3
(Dodd McCormack and Woodyatt, 1994). Broomfield and Dodd’s (2005) RCT indicated that
children with inconsistent SSD made most progress if core vocabulary intervention occurred
when they were three years-old. Early intervention is possible for children who make inconsistent
errors because their low intelligibility means that they tend to be referred early (Broomfield and
Dodd, 2004). Identification of inconsistent SSD requires assessment of multiple productions of
the same words in the same phonetic context.
Assessing consistency
Research indicates that children who score 40% or more on the Inconsistency Subtest of the
Diagnostic Evaluation of Articulation and Phonology (Dodd, Crosbie, Zhu, Holm and Ozanne,
2002) benefit from a core vocabulary approach to intervention. In this assessment children are
asked to name 25 pictures (e.g., girl, dinosaur, elephant) on three trials within one session. Each
trial is separated by another activity. If a child says a word identically on all three trials, s/he
receives a score of 0 for that word. If the child produces a word differently on at least one of the
trials, then he receives a score of 1 for that word. The child’s score, for all words produced three
times, is converted to a percentage, with a score of 40% being the criterion for diagnosis. The
40% criterion is justified by consistency of production data from typically developing and speech
disordered children (Holm, Crosbie and Dodd, 2007; McCormack and Dodd, 2000). Ways of
describing inconsistency in more detail are outlined in Holm, Crosbie and Dodd (2005).
Theoretical Basis
When children begin to produce recognizable words, at around 12 months, their phonology is
idiosyncratic. It is characterized by a great deal of individual variation in the phones used,
phonetic variability of production and inconsistency of error types (Grunwell, 1982). Ferguson
and Farwell (1975) concluded that children’s initial phonology is whole-word based. They argued
that ‘a phonic core of remembered lexical items and articulations which produce them is the
foundation of an individual’s phonology’ (p. 437). Ingram (1976) claimed that once children’s
vocabularies begin to expand past 50 words, phonological error patterns (e.g. stopping, fronting)
begin to occur across lexical items, suggesting reorganization from a whole word to a segmental
phonological system. Similarly Velleman and Vihman (2002) argued that while initially there are
no consistent error patterns, this is followed by a stage characterized by dominant production
patterns attributed to the development of word templates. The templates are abstract phonetic
production patterns that integrate the adult target with the child’s most common vocal patterns,
resulting in explicit word learning.
Reorganization of their phonological system from whole words to phonemic segments seems to
occur by most children’s second birthday. By three years mean inconsistency of word production
is 13% and by four years, 6%. About 10% of those referred for assessment of suspected SSD,
however, continue to make inconsistent errors while showing no oro-motor signs of childhood
apraxia of speech. Their speech is often unintelligible, even to their mothers. Carers cannot learn
how children say particular words, because their productions are inconsistent. What type of
deficit in speech processing underlies inconsistent speech errors? A series of experiments
compared groups of children who make inconsistent errors but have no symptoms of CAS, with
typically developing children and those who make consistent speech errors (for a summary of
published results see Dodd, Holm, Crosbie and McCormack, 2005).
Case history factors
3
Dodd, Holm, Crosbie and McIntosh
4
Fox, Howard and Dodd (2002) examined parental reports of risk factors for 66 children with SSD
according to types of surface speech errors and 48 controls. Positive family history was
statistically higher for the SSD group than controls, but did not discriminate between subgroups.
A reported family history of communication difficulty, however, was the only risk factor for
children who consistently made atypical errors (e.g., backing, bilabial fricatives for clusters).
According to parental report, children with delayed phonological development were more likely
than other groups to have a history of multiple middle ear infections. Children who made
inconsistent errors were more likely to have a reported history of prenatal or perinatal
complications (e.g., resuscitation, maternal infections, prematurity). This finding is intriguing
given the literature on phonological disorder in aphasia. A deficit in phonological planning
(termed ‘phonological assembly’ in Berndt & Mitchum, 1994) is reported to underlie inconsistent
speech errors in adults with acquired speech difficulties that cannot be attributed to dyspraxia.
Input processing
Studies of auditory processing abilities on children with SSD have yielded contradictory results
(Bird and Bishop, 1992) probably due to the assessment of heterogeneous populations. One study
(Thyer and Dodd, 1996), investigated the auditory processing skills of 30 children with SSD (10
phonologically delayed, 10 consistently using atypical error patterns and 10 making inconsistent
errors) and 30 typically developing (TD) controls. None of the SSD groups performed differently
from the matched controls on standard auditory-processing assessments of speech discrimination,
auditory figure-ground separation or binaural integration. The findings indicate that auditory
processing difficulties are unlikely to be a general explanation for SSD.
Cognitive Linguistic processing
Lexical impairment
The relationship between phonology and the size of the lexicon is important given that consistent
production of words occurs once the lexicon reaches a critical size, and that children with larger
lexicons produce a wider range of speech sounds and speech-sound sequences (Storkel and
Morrisette, 2002). One experiment (Dodd et al 2005) compared the expressive vocabulary
performance of subgroups of SSD on the Hundred-Pictures Naming Test (Fisher and Glenister,
1992). Children with an inconsistent disorder performed more poorly than all the other subgroups. No other differences between the subgroups were significant. This finding might reflect a
word-finding difficulty, since qualitative analysis of errors indicated that 46% of the inconsistent
children’s errors were related to the target word (e.g., net, “a thing for catching fish in”; nest,
“eggs”; owl, “bird that comes out at night”), compared with 28% of the controls’ errors, 33% of
the delayed children’s errors, and 33% of the consistent children’s errors. That is, the children
who made inconsistent errors often seemed to know what the picture was, but couldn’t access the
phonological shape of the word. An impaired ability to access full phonological specifications of
words might contribute to inconsistent word production.
Linguistic knowledge
An experiment tested the preference for phonological legality of subgroups of children with SSD
and controls (Dodd, Leahy and Hambly, 1989). Children had to choose between two nonsense
words to give names to pictures of animals. The 12 nonsense word pairs differed by one phoneme
that made one of the words phonologically legal, and the other phonologically illegal. For
4
Dodd, Holm, Crosbie and McIntosh
5
example, // is legal, whereas // is illegal because /sv/ does not occur in
Australian-English. Only the children who consistently made atypical errors showed no
preference for legal nonsense words. There were no significant differences between the other
three groups.
A more recent experiment (Holm, Farrier and Dodd, 2007) measured the syllable segmentation,
rhyme awareness and alliteration awareness of 61 preschool children: 46 with SSD and 15 TD
controls. Children who made inconsistent errors performed poorly on the syllable segmentation
task but no differently from controls on the other two tasks. In contrast, the children who made
consistent atypical errors showed poor performance on rhyme and alliteration but appropriate
performance on syllable segmentation.
The findings of these two experiments indicate that children who make inconsistent errors had
relatively strong phonological awareness skills. In contrast, children who consistently made
atypical errors had phonological awareness deficits that are associated with reading difficulties.
The children who make inconsistent errors, however, had specific difficulty with syllable
segmentation. If their underlying deficit was one of phonological planning, they would have
difficulty assembling words for subvocal rehearsal. Subvocal rehearsal is an essential step in a
number of phonological awareness tasks. For example, it allows the counting of how many
syllables there are in statistical.
Output processing
To investigate the nature of the deficit underlying the inconsistent sub-group’s variable
production of words, Bradford and Dodd (1996) compared the sub-groups’ ability to establish
motor plans for words. A nonsense-word learning task assessed children’s ability to generate and
execute motor plans for the articulation of nonsense words. An attractively illustrated book,
depicting the story of ‘The Three Little Pigs,’ was used. Each pig was given a legal disyllabic
nonsense name (i.e., [pazi], [] and []), that was modified to allow accurate
production for each child. To differentiate visually between the pigs, who were portrayed ten
times throughout the book, one wore glasses, another, a tie, and the third, a hat. The children
learned the names of the ‘Three Little Pigs.’ The story was told and children were encouraged to
imitate the nonsense names. The children then had to tell the story. Spontaneous production of
each test word was elicited five times. Children’s comprehension of the names was assessed by
asking them to point to named characters.
The inconsistent group performed more poorly than the other three groups on the expressive
naming task. The control, delayed and consistent groups performed equally well. Comparison of
the four sub-groups’ performance on the receptive task showed that the groups performed equally
well, although there was a trend for the inconsistent group to perform more poorly. The finding
that children who make inconsistent errors have a more general motor-planning problem raises
the question about the nature of their deficit.
Interpretation
Velleman and Vihman (2002) argue for a word ‘template’ that contains the phonological
specifications for word production – a phonological plan. The phonological plan assembled when
children have derived the wrong phonological constraints results in a pronunciation that differs
from both the target adult form and the developmental error form of a child, of the same age,
5
Dodd, Holm, Crosbie and McIntosh
6
following the normal course of development. Nevertheless, the errors made, and therefore the
phonological plan assembled, is the same when children pronounce the same word on different
occasions, and when the same phonological feature is produced in different words (e.g., using a
bilabial fricative to mark consonant clusters). In contrast, the evidence reviewed in this section
leads to the conclusion that children whose speech is characterized by inconsistent errors may
have difficulty selecting and sequencing phonemes (i.e., in assembling a phonological template or
plan for production of an utterance). Since their phonological knowledge seems intact, it is
unlikely that they have an input or cognitive linguistic difficulty. Since they have no oro-motor
difficulties, it seems unlikely that they have phonetic planning or implementation difficulties.
Their deficit appears to be at the level of phonological planning.
Level of consequences
The World Health Organization (2007) classification of learning difficulties relevant for speechlanguage pathology specifies the need for clinicians to address impairment, activity and
participation. Core vocabulary intervention focuses on the impairment underlying the functional
limitation of unintelligible speech. Targeting the underlying deficit leads to improved
intelligibility that enhances children’s communication (activity) and their academic and social
skills (participation).
Treatment Target
Core vocabulary intervention targets an underlying deficit in the speech processing chain: the
ability to generate consistent phonological plans for words. The long-term goal of intervention for
children who make inconsistent errors is to establish consistent (as opposed to correct) production
of words in spontaneous speech. Unlike phonological contrast therapy, it does not target surface
error patterns. Learning to consistently say a set of high frequency, functionally powerful words
targets the phonological planning deficit. The ability to create a phonological plan on-line is
improved by providing detailed specific information about a limited number of words and drilling
the use of that information with systematic practice.
Empirical Basis
The primary reason for targeting inconsistency is the impact it has on intelligibility. Children who
make inconsistent speech errors usually have a high degree of unintelligibility even to family
members. Consequently they are more likely to be referred for assessment of a speech disorder by
their parents rather than another referral agent, and they are more likely to be referred at three,
rather then four, years (Broomfield and Dodd, 2005). Intervention target selection is very
difficult. A child with inconsistent speech disorder may use a range of sound substitutions that
differ in manner of production, place of production or voicing. For example, Amy (the case study
described later in this chapter) marked /s/ with a [b, f, v, t, d, s] or deleted the sound. It was
impossible to select the appropriate error to contrast given the range of substitutions. Further, it
would not be effective to take an articulatory approach that targets a single sound when a child
has adequate oro-motor control and sometimes produces the target accurately or, if not, is
stimulable for the sound.
Children with inconsistent speech disorder are resistant to phonological contrast or articulation
therapy. Forrest, Dinnsen and Elbert (1997) conducted a retrospective post-hoc analysis of 14
children with speech disorder. The children were divided into three groups: those who made
6
Dodd, Holm, Crosbie and McIntosh
7
consistent sound substitutions for sounds not present in their inventories (e.g., /k/ always
produced as [t]), those who had inconsistent sound substitutions across word positions (e.g., /v/
substituted by [b] word initially, but [f] word finally), and those that used a different sound
substitution within (word initial /s/ being substituted by /v, f, d, b/) and across word positions.
The three groups were matched for severity of phonological impairment and all received
phonological contrast therapy targeting a single error in a single word position. The children with
consistent sound substitutions learned the sound and generalized it to other word positions. The
children with inconsistent sound substitutions across word positions learned the sound but only in
the treated position. The children with variable sound substitutions within and across word
positions did not learn the sound in the treated or untreated word position.
Another intervention study (Forrest, Elbert and Dinnsen, 2000) compared two groups of children:
those with consistent sound substitutions and those with variable sound substitutions. The
children with inconsistent substitutions did not respond to traditional articulation therapy. The
children with consistent substitutions learned and generalized the taught sound. Forrest et al
(2000) acknowledged the limitation of these approaches in treating children with inconsistent
SSD, concluding that “the challenge is to develop treatment protocols that instill learning and
generalization, despite the complex pattern of errors that these children demonstrate” (p. 530).
One plausible treatment approach would be to target whole words, rather than segments. While
Ingram and Ingram (2001) advocated a whole-word approach to phonological analysis and
described how that approach would translate into a set of treatment goals, no treatment study was
done. The core vocabulary approach for inconsistent SSD has included treatment case studies that
explored the appropriateness of the approach; a group efficacy study comparing different
treatments using a cross-over design; and, an effectiveness study that was part of a randomized
control trial carried out in a Speech and Language Therapy service in the north-east of England.
Case Studies
Developing a novel therapy
MW a boy, aged 4;2, who was part of a clinical research study evaluating phonological contrast
treatment for phonological disorders (Dodd, and Iacono,1989) made no progress despite weekly
therapy for eight months. The other six children in the study had improved and been discharged.
He differed from them in that he made inconsistent errors (e.g., he pronounced tomato as
[] and [] and TV as [] and []). He showed no oro-motor signs of CAS.
To establish consistency of production of specific lexical items, a novel therapy approach was
developed. The approach focused on whole words to establish a vocabulary of highly functional,
powerful words that were produced consistently, although not necessarily correctly. After two
months of weekly therapy establishing a consistent vocabulary of 50 words, consistency
generalized and a phonological contrast approach was successfully reintroduced.
Comparing different intervention approaches for different types of SSD
Bradford-Heit (1996) reported six treatment case studies that compared three therapy methods for
children with different types of speech disorder (consistent atypical errors, inconsistent errors and
CAS). Two children in the study (see Dodd and Bradford 2000) presented with speech disorder
characterized by inconsistent productions. A multiple baseline with alternating treatments
evaluated the effect of phonological contrast therapy (targeting error patterns), core vocabulary
therapy (targeting consistency) and PROMPT system therapy (targeting articulatory gestures).
7
Dodd, Holm, Crosbie and McIntosh
8
Children were seen individually, twice weekly, for 12 (30 minute) sessions in each treatment
block with a 3-week withdrawal period between therapy blocks. The order of treatments was
randomized across the six children in the study.
MC, aged 4;3, received core vocabulary therapy first. Consistency of word production was
established and generalized to untreated words (31% pre-treatment to 69% post treatment of
untreated words). He also benefited from the second block of therapy that targeted a phonological
process (liquid and glide contrast) although generalization to untreated words was limited. MC
did not benefit from the PROMPT therapy approach. TN, aged 3;7, received core vocabulary
followed by PROMPT and then phonological therapy. The results indicate that TN benefited
from core vocabulary (consistency 36% pre-treatment, 75% post-therapy on untreated words).
The other two blocks of therapy did not result in improvement. These findings suggest that once
consistency is established, it may be possible to immediately target phonological contrasts
effectively, as was done with MW, if accuracy is still at unacceptable levels.
Core vocabulary intervention with ‘special’ children
Three preschool boys, who were referred to an efficacy study focusing on phonological contrast
intervention, did not meet the inclusion criteria because they made inconsistent speech errors. All
three had additional complicating factors. Treatment case studies were undertaken to examine the
effects on core vocabulary intervention outcome of previous intervention, the use of default
preferred word plans and behavior disorder (McIntosh and Dodd, in press). Andrew had a history
of previous intervention that had focused on the articulation of /s/ which had over-generalized.
Ben had a behavior disorder and limited attention. Cameron appeared to have a number of
pervasive default plans for his word production. He most often began words with /w/, used a
velar plosive syllable finally, employed default syllables (with variable vowels) // or //
word finally or as whole word substitutes (e.g., [] apple, [] orange, []
lighthouse, [] watch, [] swing, [] giraffe). All three boys showed gains in
intelligibility, accuracy and consistency of word production, although their individual differences
required clinical adaptation of the approach (see Table 2). While Andrew received only six hours
of intervention, Ben received 13.5 hours and Cameron 19 hours. It proved difficult to suppress
preferred word shapes, but progress was achieved by an intermediary step of accepting a
consistent production that included an aspect of his default word plan (e.g. accept [] swing)
before that was modified to [] and then []. The case studies indicated that previous
treatment focusing on articulation did not affect the outcome of core vocabulary intervention,
although it did make it difficult to diagnose because the inconsistent use of an intrusive /s/ might
have reflected previous therapy rather than inconsistency. The cases with behavior disorder and
default templates for words required more intervention hours than had previously been found
necessary for positive outcome.
Table 2. Case studies: percent increase in consistency and accuracy after core vocabulary
intervention
Children
Consistency
Accuracy (PCC, PVC)
Pre
Post
Pre
Post
Andrew 3;8
44
90
46, 88
94, 99
Cameron 4;2
36
90
34, 78
69, 95
Ben 3;9
36
60
22, 53
52, 86
8
Dodd, Holm, Crosbie and McIntosh
9
Group design
Efficacy study: alternating treatments design
The study (Crosbie, Holm and Dodd, 2005) evaluated the effect of two different types of
intervention on the speech accuracy and consistency of word production of children with
consistent (N = 8) and inconsistent SSD (N = 10) SSD. The 18 children (aged 4;08 to 6;05 years)
with severe SSD participated in an intervention study comparing phonological contrast and core
vocabulary intervention. All children received two 8-week blocks of each intervention, receiving
twice weekly, 30 minute sessions. A multiple baseline design with alternating treatments was
used. Two pre-therapy baseline measures were taken three weeks apart and showed no
spontaneous change in number or type of speech errors. Once eligibility was confirmed children
were allocated to one of the two therapies by order of referral. Treatment 1 was implemented
after the baseline period followed by a 4-week withdrawal period, followed by treatment 2. The
method of allocation to treatment ensured children in both subgroups of speech disorder received
the blocks of therapy in both possible orders (core vocabulary followed by phonological contrast;
phonological contrast followed by core vocabulary).
All of the children increased their consonant accuracy during intervention (see Table 3). Core
vocabulary therapy resulted in greater change in children with inconsistent speech disorder and
phonological contrast therapy resulted in greater change in children with consistent speech
disorder. The results of this study provide strong evidence that treatment targeting the speechprocessing deficit underlying a child’s speech disorder will result in efficient system-wide
change. Differential response to intervention across subgroups provides evidence supporting
theoretical perspectives regarding the nature of SSD: different underlying deficits result in
different types of speech errors requiring different types of therapy.
Table 3.
Group summary of change in inconsistency score and PCC following each
intervention
Inconsistency (% mean, SD)
PCC (mean, SD)
Group
Core
vocabulary
Phonological
contrast
Core
vocabulary
Phonological
contrast
Consistent (n=8)
5.00 (7.63)
9.50 (12.99)
6.75 (3.01)
24.62 (12.88)
Inconsistent (n=10)
24.6 (9.14)
4.20 (7.57)
15.80 (9.05)
9.70 (6.57)
15.89 (12.99)
6.56 (10.35)
11.78 (8.28)
16.33 (12.22)
Overall
Randomized Control Trial
Effectiveness Study
Broomfield and Dodd (2005) reported a randomized control trial that included 320 children with
SSD who were referred to the pediatric Speech and Language Therapy Service in Middleborough
UK, between January 1999 and April 2000. The study differentially diagnosed subtypes of speech
impairment, specified the content of the therapeutic approaches implemented with each subgroup
9
Dodd, Holm, Crosbie and McIntosh
10
and described the clinical pathway (including service delivery). Statistical analyses investigated
treatment versus no treatment outcome. In addition, the effect of a range of factors on outcome
was reported: dosage, age for cost-effective intervention for specific impairments, gender, case
history factors and co-morbidity with other language disorders. Only data relevant to children
with inconsistent SSD are discussed here.
The type and duration of the intervention offered was determined by the nature and severity of
each child’s SSD. Thirty children had inconsistent SSD, twenty of them classed as having a
severe or profound level of unintelligibility. Core vocabulary therapy was the specified treatment
for a diagnosis of inconsistency and all clinicians (N = 12) employed by the service used the
treatment approach when it was indicated by diagnosis. After assessment, children typically
attended up to six ‘diagnostic’ group sessions focusing on attention, listening, speech sound
awareness, language concepts and categories and oro-motor and speech skills to ensure all
difficulties that a child was having had been identified by the standardized assessments routinely
administered. These initial sessions also allowed children to become familiar with the clinicians
and confident in the clinic environment. Children receiving core vocabulary intervention then
received six individual, weekly, 30-45 minute treatment sessions. After a break from intervention
(a consolidation period) children were reassessed and, if their consistency remained atypical,
received a second episode of core vocabulary intervention.
In order that a standard measure was available for comparison, assessment findings were
converted to z scores. Each change in (accuracy) performance was measured in comparison with
an age change, with six months passing between each RCT assessment. Consequently, when a
difference in z scores showed improvement, the child had progressed more than would have been
expected for their change in age. Two of the children moved away and did not complete
intervention.
The 18 children with inconsistent disorder who received intervention performed better (z = 0.37)
than the 10 children who received no intervention (z = 0.10). There is evidence, then, that a core
vocabulary approach is effective for inconsistent SSD at a clinical service level. The result
suggests, however, that children who made inconsistent errors made less progress in treatment
than other subgroups of speech disorder (e.g., treated consistent atypical disorders: z =0.59;
untreated z = 0.03). A number of factors may contribute to this finding. Children who make
inconsistent errors often have a severe, pervasive disorder that may need to be addressed by more
intervention sessions An alternative explanation is that many of the children in this group were
younger (around 3 years old) than children in the other subgroups. Consequently, some children
may have had poorer attention or less motivation to comply with intervention tasks. Additionally,
unlike other evaluations, treatment occurred once per week rather then twice. Further, clinicians
had only recently been trained in the use of core vocabulary and may have been less confident in
its implementation than for other treatment approaches. Finally, for comparison with other
children with SSD, the outcome measure focused on accuracy rather than consistency. While not
assessing the primary target of therapy, the findings demonstrate that improved accuracy is a ‘byproduct’ of intervention targeting consistency.
Summary of Clinical Findings
Robey and Schultz (1998) identify clinical reports and case studies of a treatment as the first steps
in research on clinical outcomes. These studies can explore a treatment’s potential and allow
10
Dodd, Holm, Crosbie and McIntosh
11
precise definition of the treatment and the population the treatment will benefit. The outcome of
the seven treatment case studies briefly described all had a positive outcome. The group efficacy
study, which used a multiple baseline alternating treatments research design, also provided
positive findings. This design has been argued to have greater power than RCTs (e.g., Dodd,
2007) because all participants receive the same amount of intervention and children act as their
own controls. The highly controlled research design allows stronger conclusions to be drawn
about the relative benefit of a particular therapeutic approach using a constant service delivery
model. The benefit of core vocabulary therapy was confirmed by an effectiveness study: an RCT
carried out in a clinical service.
Practical Requirements
The service delivery model recommended for core vocabulary intervention is individual, twice
weekly sessions lasting around 30 minutes, with the parent present for at least one of those
sessions. While the number of sessions varies according to severity and teacher and carer input, a
clinician should not expect to exceed 16 half hour sessions to attain consistency of production, as
well as enhanced PCC, that generalizes to untreated words. Some studies of older children have
held one session per week at school and the other at the SLP health-based clinic to ensure liaison
with both the child’s teacher and parents.
Roles of Parent and Teachers
The intervention approach depends on the child’s family and teacher to reinforce use of the core
vocabulary and carryout daily practice. Both must be involved from the outset of intervention.
For example, after assessment and diagnosis, but before the first intervention session, the child,
family and teacher are asked to contribute to a list of 70 words that are frequently part of the
child’s functional vocabulary. These words are used as the basis for the treatment sessions.
Family and school must be aware of, and monitor the consistency of, their child’s best production
of the words being targeted that week, in practice and spontaneously. It must be emphasized to
carers that the primary target of the intervention is to make sure their child says a word in exactly
the same way each time they attempt to say it, not necessarily an error-free production.
The Clinicians’ Role
The first session each week is devoted to eliciting the child’s best production of up to 12 selected
words and establishing their production. The number of words targeted depends on the child’s
ability to achieve a word’s best possible word production. The second weekly session focuses on
drilling the newly learned words in order to monitor production, providing appropriate feedback
when the best production is not produced. Parents and teachers must be encouraged to perceive
the child’s productions accurately and learn how to provide appropriate feedback. Clinicians need
to be sensitive to parent and teacher feedback that could influence the content of the therapy
session (e.g., words or phrases needing additional teaching, new words to target). Finally, an
essential clinical role is to regularly monitor generalization of consistency to untreated probes.
Materials
Equipment is minimal: pictures of the target words (with the written name beneath the picture); a
box or bag for the vocabulary pictures; a wall-chart where words produced consistently can be
glued; games that elicit multiple productions of target words rapidly. A DVD of text resources
(e.g., teacher and parent information sheets, probe word lists) is available from the authors.
11
Dodd, Holm, Crosbie and McIntosh
12
Key Components
The core vocabulary intervention approach should only be used for children who make
inconsistent errors but have no oro-motor signs of CAS. The ultimate goal of intervention is
intelligible speech. The long-term goal for a block of core vocabulary intervention is for the child
to produce at least 50 target words consistently, that is, to produce a word exactly the same way
each time it is produced. Generalization of consistency is expected once the child has mastered
the consistent production of 50 words. The short-term goals are target specific, however two
general goals can be applied to each set of target words. The first goal is for the child to achieve
an appropriate productive realization of each target based on the child’s phonological system and
phonetic inventory. This ‘best production’ may be correct or contain a developmental error. The
second goal is for the child to consistently use the established ‘best production’.
Selection of core vocabulary
A core vocabulary of up to 70 functionally powerful is selected by the child, parents, and teacher.
The types of words commonly included on children’s lists are names (e.g., teacher, friends, pet),
places (e.g., address, toilet), function words (e.g., please, sorry), foods (e.g., cornflakes, juice,
chips) and the child’s favorite things (e.g., Polly Pocket, Spiderman, games). The words are not
selected according to word shape or segments. They are chosen because the child frequently uses
these words in their functional communication. The child’s intelligible use of the functionally
powerful words selected motivates the use of consistent productions.
Establishing best production
In the first session each week, the clinician teaches the child between up to ten target words,
selected at random from a bag containing the vocabulary chosen. Words are taught sound-bysound, using cues such as syllable segmentation, imitation and cued articulation as outlined in
Passy (1990). For example: to teach Joseph, the clinician would explain that Joseph has two
syllables – [do] and [sf]. The first syllable [do] has two sounds, /d / and /o/, and the
second syllable [sf] has three sounds /s/, // and /f/. The child attempts the first syllable - [do]
- receives feedback and makes further attempts after being given models and receiving feedback
about each attempt. When the child’s best production of the first syllable has been established, the
second - [sf]’ is targeted, and then the two syllables are combined - [do-sf]’. For some
children it is effective to link sounds to letters. Children with inconsistent SSD are usually able to
imitate all sounds. If it is not possible to elicit a correct production then the best production may
include developmental errors (e.g., [dosf] for Joseph, [tmra] for camera).
Drill
It is important that the child practises the target words daily as well as receiving feedback on
those words in everyday communication situations. The second session each week with the
clinician involves practice of the target words. Games are used to elicit a high number of
repetitions. Any game that the child is highly motivated to participate in can be used to elicit
productions. Initial picture naming games (e.g., stepping stones – with more than one picture on
each stepping stone) can be followed by those requiring the target in a carrier phrase (e.g., picture
lotto) and finally by story generation (asking for one, two or three of the target words). Elbert,
Powell and Swartzlaner (1991) suggest a child should produce about 100 responses in 30
12
Dodd, Holm, Crosbie and McIntosh
13
minutes. Although this number may sound like a high rate of response it is not difficult to elicit
150 –170 responses in a 30-minute session of core vocabulary intervention.
Treatment on error
It is important to provide children with feedback when their best production is not uttered. Leahy
(2004) reported that children do not always understand why they attend treatment or what they
are required to do. Consequently it is important to be explicit about the purpose of intervention,
the nature of the error made, and how it can be corrected. If a child produces a target that deviates
from the best production the clinician can imitate the production and explicitly explain that the
word differed and how it differed. For example if the child’s target word was ‘sun’ and he
produced [] the clinician would say “[], that’s different to how we say it. That had a []
sound at the start but we need to make it a [s], []. Have another try at telling me what this
picture is.” Clinicians should avoid simply asking for an imitation of the target word that provides
a phonological plan. Instead, clinicians should provide information about the plan requiring
children to generate their own plan for the word..
Monitoring consistent production
Towards the end of the second session each week, the child is asked to produce, three times, the
set of target words that have been the focus of therapy for the past week. Any word that the child
can produce consistently is removed from the list of words to be learned. It may be placed on a
chart showing what the child has achieved. Words produced inconsistently remain on the list (go
back in the bag of words yet to be learned). Even though there are 50 target words that form a
core vocabulary for the child’s 8 weeks of intervention, such monitoring allows for words that
have not been mastered to be readdressed in another week.
Assessment and Progress Monitoring to Support Decision Making
Generalization
Core vocabulary intervention aims to stabilize a child’s system leading to consistent word
production. The therapy would not be beneficial if the effect of therapy was limited to the treated
target items. To monitor generalization, a set of untreated items (10 words) should be elicited
three times, once a fortnight in one therapy session, to monitor system change (i.e., determine
when consistency has generalized). Once untreated probes become consistent the 25 Word Test
(DEAP, Dodd et al 2002) should be re-administered to ensure generalization has occurred.
What to do when the child’s speech is consistent
Core vocabulary intervention will increase the consistency of a child’s speech production. The
effect this has on the child’ speech system can vary. Case studies suggest that most children’s
speech became both more consistent and more accurate, and that the children’s speech was
characterized by developmental, not atypical, error patterns. For some children, however, more
than one intervention approach may be necessary to achieve age-appropriate speech. For
example, Dodd and Bradford (2000) report a case study of a boy with inconsistent speech
production. Once consistency was established he benefited from phonological contrast therapy
that targeted his remaining developmental error patterns. Assessment on a standardized
phonological assessment, preferably after a consolidation break from treatment, would allow
decisions about the need for further, different intervention.
13
Dodd, Holm, Crosbie and McIntosh
14
Children from culturally diverse backgrounds
A bilingual Punjabi-English speaking child had speech was characterized by inconsistent errors
(Holm and Dodd, 1999). Hafis was successfully treated using a core vocabulary approach, in
English, which targeted consistency of production. Hafis received 16 (30 minute) sessions over
an 8-week period. His consistency of production increased. On untreated items, his inconsistency
in English fell from 56% to 21% and in Punjabi from 45 to 30%. His accuracy, measured by
percent of consonants correct, increased significantly in Punjabi by 16% and in English by 26%.
The significance of the generalization of intervention across languages is that it highlights the
effect of targeting the underlying deficit rather than the surface speech error patterns.
Case study (from Crosbie et al, 2005)
Amy was assessed mid-way through her pre-school year when she was aged 4;8. She is
monolingual, an only child, living with both parents. Her speech had always been difficult to
understand. Amy’s birth, medical and developmental histories were normal. She had no hearing
problems. Amy was reported to be conscious of her SSD, having difficulty establishing
friendships at pre-school, and upset and frustrated when she was not understood.
Standardized assessment indicated age-appropriate receptive language and non-verbal
intelligence. The Articulation, Inconsistency and Phonology Assessments of the Diagnostic
Evaluation of Articulation and Phonology (DEAP - Dodd et al., 2002) indicated that Amy could
produce all speech sounds in imitated simple syllables or in isolation. Her oro-motor skills were
age-appropriate. The Inconsistency Assessment indicated that her speech was 56% inconsistent (a
standard score of 3). The Phonology Assessment indicated that Amy’s pre-intervention PCC was
50% and her PVC 92%, performance at the 1st percentile. Reassessment three weeks after initial
assessment showed no spontaneous change.
Intervention
Treatment with the clinician occurred individually, twice weekly. Therapy sessions were
alternately conducted in Amy’s home and pre-school to allow liaison with both her teacher and
parents. There were 16 (30 minute) sessions over an eight-week period. Prior to intervention,
Amy, her parents and her teacher collaborated to produce a list of 50 words that were functionally
‘powerful’ for her. The clinician explained the principles of core vocabulary therapy to Amy’s
parents and teacher. Each week 10 words were drawn randomly from the set of 50 target words.
Amy was taught the 10 words by the clinician, and then those words were targeted consistently
by her parents and teacher throughout the week and revised in the second session with the
clinician. Some of the taught words were correct. For others, developmental errors were accepted.
Production was drilled sound by sound. After the initial session where Amy learned the target
words, her parents and teacher consistently required her to produce those 10 words in the same
way throughout the week. Practice of the ten words occurred three times each day, as well as
being reinforced when they occurred in everyday communication situations. Her parents and
teacher used the same teaching strategies as the clinician. During the second weekly session with
the clinician the words were drilled, and then Amy had a “test” where she had to produce the 10
words three times, three trials separated by another activity. Untreated probes (a set of 10
untreated words) were also elicited three times to monitor generalization fortnightly. Amy’s
progress was drawn on to her chart and her parents implemented a reward scheme linked to her
progress on the weekly words.
14
Dodd, Holm, Crosbie and McIntosh
15
Amy found the intervention program difficult at times, particularly when learning the new words
each week. However she was well supported at home and school and regularly did her practice.
The intervention period was for a pre-determined eight-week period so that everyone involved
was able to commit to it knowing when it was going to end.
Progress
Amy learned to produce 57 words consistently in eight weeks. When she was reassessed on the
consistency tasks used in the initial and baseline assessments her inconsistency had decreased
from 56% to 32%. Her consonant accuracy increased by 10%. In addition she was more
consistent in her substitution patterns following intervention. Instead of the almost free variation
between up to six phonemes evident in her pre-intervention assessments Amy generally only used
either the correct phoneme or one other phoneme. Her speech was still affected by developmental
phonological error patterns but there was no evidence of atypical error patterns in her speech at
the final assessment.
Future Directions
For core vocabulary intervention to be accepted as a useful treatment approach, clinical efficacy
and effectiveness studies need to be carried out by other groups of clinical researchers to validate
the findings already made. Although the use of the approach has been anecdotally reported to be
successful by clinicians, it still needs to be rigorously tested. Future research could address some
issues needing clarification, listed below in the form of research questions.
- Does client choice of the vocabulary lead to better outcome than clinician choice that
would manipulate word shape or phoneme contrast?
- How many words have to be taught before consistency generalizes?
- Does accepting ‘best production’ rather than ‘correct production’ result in differences in
outcome of therapy for accuracy of production?
- For bilingual children, only one case study of a child with inconsistent SSD has shown
generalization of core vocabulary intervention gains from the treated to the untreated
language. This finding has important theoretical and clinical implications and needs
repeating with other language pairs.
References
Berndt, R. & Mitchum, C. (1994). Approaches to the rehabilitation of ‘phonological assembly’. In G.
Humphreys & M Riddoch (eds.), Cognitive Neuropsychology and Cognitive Rehabilitation, London:
Lawrence Erlbaum Associates.
Bird, J. & Bishop, D. (1992). Perception and awareness of phonemes in phonologically impaired
children. European Journal of Disorders of Communication 27, 289-311.
Bradford, A & Dodd, B. (1996). Do all speech disordered children have motor deficits? Clinical
Linguistics and Phonetics 10, 77-101.
Bradford-Heit, A (1996) Subgroups of children with developmental speech disorder : identification
and remediation. Unpublished PhD Thesis: University of Queensland
Broomfield, J. & Dodd, B. (2004) The nature of referred subtypes of primary speech disability. Child
Language Teaching and Therapy 20, 135-151
15
Dodd, Holm, Crosbie and McIntosh
16
Broomfield, J. & Dodd, B. (2005) Clinical effectiveness. In: Dodd, B. (ed). Differential Diagnosis
and Treatment of Children with Speech Disorder. London: Whurr, pp 211-230
Crosbie, S Holm, A & Dodd, B (2005) Intervention for children with severe speech disorder: A
comparison of two approaches. International Journal of Language and Communication Disorders,
40, 467–491.
Dodd, B. (2007) Evidence-based practice and speech-language pathology. Folia Phoniatrica
Logopedia, 59,118-129.
Dodd, B. & Bradford, A. (2000) A comparison of three therapy methods for children with different
types of developmental phonological disorders. International Journal of Language and
Communication Disorders 35, 189-209
Dodd, B. & Iacono, T. (1989) Phonological disorders in children: changes in phonological process
use during treatment. British Journal of Communication Disorders 24, 333-351
Dodd, B., Crosbie, S., Zhu, H., Holm, A. & Ozanne, A. (2002) The Diagnostic Evaluation of
Articulation and Phonology. London: Psych-Corp
Dodd, B. Leahy, J. & Hambly, G. (1989) Phonological disorders in children: underlying cognitive
deficits. British Journal of Developmental Psychology 7, 55-71
Dodd, B., McCormack, P. & Woodyatt, G. (1994) An evaluation of an intervention program: The
relationship between children’s phonology and parents communicative behavior. American Journal
of Mental Retardation, 98, 632-645
Dodd, B., Holm, A., Crosbie, S. & McCormack, P. (2005) Differential diagnosis of phonological
disorders. In: Dodd, B. (ed). Differential Diagnosis and Treatment of Children with Speech Disorder.
London: Whurr, pp 44-70
Dodd B, Holm A, Crosbie S, McIntosh B: (2006). A core vocabulary approach for management of
inconsistent speech disorder. Advances in Speech-Language Pathology, 8, 220-230.
Duggirala, V. & Dodd, B. (1991) A psycholinguistic assessment model for disordered phonology .
In: Congress for Phonetic Sciences, Aix-en-Provence, Universite de Provence, pp 342-345.
Elbert, M., Powell, T. & Swartzlander, P. (1991) Toward a technology of generalization: How many
exemplars are sufficient? Journal of Speech and Hearing Research, 34, 81-87.
Fisher, J. & Glenister, J. (1992) The Hundred Word Pictures Naming Test. Melbourne: ACER.
Ferguson, C. & Farwell, C. (1975) Words and sounds in early language acquisition. Language 51,
39-49.
Forrest, K. (2003) Diagnostic criteria of developmental apraxia of speech used by clinical speechlanguage pathologists. American Journal of Speech-Language Pathology 12, 376-380.
Forrest, K., Dinnsen, A. & Elbert, M. (1997). Impact of substitution patterns on phonological
learning by misarticulating children. Clinical Linguistics and Phonetics 11, 63-76.
Forrest, K., Elbert, M. & Dinnsen, D. (2000). The effect of substitution patterns on phonological
treatment outcomes. Clinical Linguistics and Phonetics 14, 519-531.
Fox A., Howard, D. & Dodd, B. (2002) Risk factors in speech disorder. International Journal
Language and Communication Disorder, 37, 117-132.
Grunwell, P. (1982). Clinical Phonology. London: Croom-Helm.
Holm, A. & Dodd, B. (1999). An intervention case study of a bilingual child with phonological
disorder. Child Language Teaching and Therapy, 15, 139-158.
Holm, A., Crosbie, S. & Dodd, B. (2005). Treating inconsistent speech disorders. In: Dodd, B. (ed.),
Differential Diagnosis and Treatment of Children with Speech Disorder. London: Whurr, 182-201
16
Dodd, Holm, Crosbie and McIntosh
17
Holm, A., Farrier, F., & Dodd, B. (2007). The phonological awareness, reading accuracy and spelling
ability of children with inconsistent phonological disorder. International Journal of Language and
Communication Disorders, 42, 1-23.
Ingram, D. (1976). Phonological disability in children. New York: Elsevier.
Ingram, D & Ingram, K. (2001). A whole-word approach to phonological analysis and intervention.
Language, Speech and Hearing Services in School, 32, 271-283.
Leahy, M. M. (2004) Therapy talk: analyzing therapeutic discourse. Language, Speech, Hearing
Services in Schools 35, 70-81
McCormack, P. & Dodd, B. (1996). A feature analysis of speech errors in subgroups of speech
disordered children. In McCormack, P. & Russell, A. (eds). Proceedings of the Sixth Australian
International Conference on Speech Science and Technology. Adelaide: AASTA, pp 217-222.
McCormack, P.F. & Dodd, B. (1998). Is inconsistency in word production an artefact of severity in
developmental speech disorders? Child Language Seminar, Sheffield.
McIntosh, B. & Dodd, B. (in press) Evaluation of Core Vocabulary intervention for treatment of
inconsistent phonological disorder: Three case studies. Child Language Teaching and Therapy.
Ozanne,A. (2005) Childhood apraxia of speech. In: Dodd, B. (ed.), Differential Diagnosis and
Treatment of Children with Speech Disorder. London: Whurr, pp 71-82.
Passy, J. (1990). Cued Articulation. Hawthorn, Victoria: ACER.
Robey, R & Schultz, M (1998). ‘A model for conducting clinical outcome research: An adaptation
of the standard protocol for use in aphasiology’ Aphasiology, 12: 787-810.
Storkel, H. & Morrisette, M. (2002) The lexicon and phonology: Interaction in language
acquisition. Language Speech and Hearing Services in Schools 27, 215-230.
Thyer, N. & Dodd, B. (1996). Auditory processing and phonologic disorder. Audiology 35, 37-44
Velleman, S. & Vihman, M. (2002). Whole-word phonology and templates: Trap, bootstrap, or
some of each. Language, Speech and Hearing Services in Schools 33, 9-23
Word Health Organization (2007) International Classification of functioning and health.
http://www.who.int/classifications/icf/en/
17