Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
Auditory word identification in dyslexic and normally achieving readers
(USC Thesis Other)
Auditory word identification in dyslexic and normally achieving readers
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
AUDITORY WORD IDENTIFICATION
IN DYSLEXIC AND NORMALLY ACHIEVING READERS
by
Jennifer Lynn Bruno
A Thesis Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In partial Fulfillment of the
Requirements for the Degree
MASTER OF ARTS
(PSYCHOLOGY)
December 2005
Copyright 2005 Jennifer Lynn Bruno
UMI Number: 1435120
1435120
2007
UMI Microform
Copyright
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
ProQuest Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, MI 48106-1346
by ProQuest Information and Learning Company.
ii
ACKNOWLEDGEMENTS
Franklin R. Manis
Jonathan Nakamoto
University of Southern California
Anne J. Sperling
Georgetown University
Patricia Keating
University of California, Los Angeles
And
Mark S. Seidenberg
University of Wisconsin, Madison
iii
TABLE OF CONTENTS
Acknowledgements ii
List of Tables
iv
List of Figures v
Abstract vi
Introduction 1
Method 11
Results 18
Discussion 37
References 45
Appendix 52
iv
LIST OF TABLES
Table 1: Means and standard deviations on standardized tasks
19
Table 2: Means on the gating measures
22
Table 3: Means and standard deviations on the standardized tests for
PA groups
25
Table 4: Means and standard deviations on gating measures for PA
groups
26
Table 5: Correlations
30
Table 6: Hierarchical Regression Analyses for Three Criterion
Variables
32
v
LIST OF FIGURES
Figure 1: Average gating score based on Phonological Awareness Group 27
Figure 2: Path analysis 36
vi
ABSTRACT
The integrity of phonological processing and phonological
representation in dyslexic children was explored using a gating task in which
children listened to successively longer segments (gates) of words. At each
gate the task was to form a guess of the entire word. Responses were scored
for overall accuracy, and sensitivity to coarticulation from the final consonant.
As a group, dyslexic children were slightly less able than normally achieving
readers to detect coarticulation present in the vowel portion of the word, and
only on the most difficult items, those ending in a nasal consonant. Dyslexics
with low phonological awareness required more auditory information to make
use of coarticulation across all three consonant endings that were studied
(stops, liquids and nasals).
Hierarchical regression analyses revealed performance on the gating
task significantly predicted phonological awareness while controlling for age
and IQ, suggesting that phonological awareness mediated the relationship
between gating and word reading ability.
1
INTRODUCTION
Dyslexic children typically have difficulty with word recognition,
phonological decoding, reading comprehension, and spelling, but an absence
of major intellectual, sensory, emotional or experiential impediments to
learning. Although many theories have been postulated to explain the causes
of these deficits, there is a strong consensus
in the field that the most probable etiology involves phonological
deficits (Brady, 1997; Fowler, 1991; Snowling, 2000; Stanovich & Siegel,
1994; Wagner & Torgesen, 1987). Phonological deficits are thought to
underlie specific kinds of language processing difficulties outside the domain
of reading, including poor phonological awareness (Bruck, 1992; Liberman &
Shankweiler, 1985; Pratt & Brady, 1988; Swan & Goswami, 1997), inefficient
use of verbal working memory (Brady, Shankweiler, & Mann, 1983; Griffiths
& Snowling, 2002; McDougall, Hulme, Ellis, & Monk, 1994), and slow access
to the mental lexicon as manifested in naming tasks (Denckla & Rudel, 1976;
Wolf & Bowers, 1999), as well as critical components of the reading process,
such as the learning of spelling-sound correspondences and the development of
efficient word recognition (Bruck, 1992; Rack, Snowling, & Olson, 1992;
Share, 1995; Stanovich & Siegel, 1994).
The relationship between phonological awareness and reading
disability has received a great deal of attention within the reading research
community. Poor performance on phoneme level analysis tasks is among the
2
most persistent difficulties in dyslexic individuals (Bruck, 1992; Manis,
Custodio, & Szeszulski, 1993). Some have argued that poor phonological
awareness reflects difficulties in analyzing the sound structure of words,
particularly at the level of the phoneme. Such difficulties are directly linked
problems in learning spelling-sound correspondences in alphabetic languages
(Liberman & Shankweiler, 1985; Share, 1995). A problem with this approach
is that phonological awareness at the level of the phoneme appears not to
develop actively until the onset of reading instruction, and it may be heavily
influenced by the individual’s experience with printed words in an alphabetic
language (Morais, Cary, Alegria, & Bertelson, 1987; Perfetti, Bell, Beck, &
Hughes, 1987; Ziegler & Goswami, 2005). Others have argued that poor
performance on phonological awareness tasks may reflect incomplete or
inaccurate phonological representations, rather than analytic problems per se
(Elbro, Borstrom, & Petersen, 1998; Fowler, 1991; Snowling & Hulme, 1989;
Swan & Goswami, 1997). This broad view can be termed the phonological
representations hypothesis (Swan & Goswami, 1997).
One line of research investigating the phonological representations
hypothesis is the study of speech perception and dyslexia. If dyslexic children
do not accurately categorize phonemes in their language, it would be difficult
for them to create accurate representations for words in long-term memory.
Although several studies have reported categorical speech perception deficits
in dyslexics as a group (Chiappe, Chiappe, & Siegel, 2001; Godfrey, Syrdal-
3
Lasky, Millay, & Knox, 1981; Maassen, Groenen, Crul, Assman-Hulsmans, &
Gabreels, 2001; Reed, 1989; Serniclaes, Sprenger-Charolles, Carré, &
Demonet, 2001;Werker & Tees, 1987), many individual dyslexics show
normal speech perception (Adlard & Hazan, 1997; Joanisse, Manis, Keating &
Seidenberg, 2000; Manis, McBride-Chang, Seidenberg, Keating, Doi, Munson,
& Peterson, 1997; Manis & Keating, 2004; Pennington, van Orden, Smith,
Green, & Haith, 1990). In the Joanisse et al. (2000) and Manis & Keating
(2004) studies, a subset of dyslexics with combined oral language and
phoneme awareness impairments performed poorly on tests of speech
perception, but the remainder of the dyslexic sample did not. Several dyslexic
children with the “classic” profile of normal oral language, low phonological
awareness and low nonword reading performed normally on speech perception
tasks. This raises the possibility that (at least by the time of reading
instruction) categorical perception of speech sounds is normal in dyslexic
children, and other aspects of phonological representation and/or processing
are abnormal (Elbro, Borstrom & Peterson, 1998; Swan & Goswami, 1997).
Swan and Goswami (1997) explored the relationship between adequacy
of phonological representations and phonological awareness, using a design in
which children’s performance on phonological awareness tasks at three levels
(syllable, onset-rime and phoneme) was adjusted for their knowledge of the
word stimuli (the latter assessed via a picture naming task). After adjusting for
picture name knowledge, dyslexics performed as well as same-age normal
4
readers on the syllable and onset-rime level tasks, but remained deficient
relative to both same-age and reading-age matched normal readers on the
phoneme-level tasks. These results suggest phonological awareness might
depend both on the adequacy of the underlying phonological representation
and the ability to segment and manipulate phonemes. Swan and Goswami
(1997) cautioned that the ability to pronounce names of pictures was a
relatively crude index of underlying phonological representations.
Elbro et al. (1998) obtained a measure of the distinctness of word
pronunciations (e.g., /ælgetor/ is a less distinct pronunciation of alligator than
/NlOPgetor/) we’re just quoting them here right? because it’s very unlikely that
anyone says /tor/ at the end of alligator in two groups of kindergarten children
who either had a dyslexic parent (at-risk group) or did not. The distinctness
measure predicted phoneme awareness in second grade over and above the
contribution of expressive vocabulary, articulation accuracy and letter-name
knowledge. In addition, phonological distinctness predicted membership in the
at-risk group for children who were later categorized into a reading disabled
group in second grade, controlling for phoneme awareness and phonological
short-term memory. Elbro et al. (1998) suggest that less distinct phonological
representations interfere with the learning of spelling-to-sound mappings in
some disabled readers.
In the present study, the gating paradigm was utilized to assess
phonological representation/processing in dyslexic children. In a typical gating
5
task a listener is presented with successively longer portions of words (gates)
beginning with the onset. At each gate, the listener is asked to guess the entire
word. This type of task requires intact and highly integrated phonological
representations because subjects must use limited acoustic information to
identify a word by comparing the acoustic information to many possible stored
representations (Salasoo & Pisoni, 1985). Adults require as little as 150 ms of
acoustic input (less than half the length of a typical spoken word) to identify
high frequency words (Grosjean, 1980; Salasoo & Pisoni, 1985; Tyler &
Wessels, 1985).
Warren & Marslen-Wilson (1987) showed that adults can use the effect
of anticipatory coarticulation on the vowel to identify spoken words at a gate
that preceded the completion of the following consonant. That is, adults could
use differences in the vowel sound to anticipate the following consonant before
the consonant is articulated. This effect was independent of the frequency of
the target word. The implication is that word recognition can proceed by means
of phonetic features continuously sampled from the input, and that
identification of discrete units such as the phoneme may not strictly be
necessary for spoken word recognition. Whether this is true or not, the gating
task can be used as an overall index of the integrity of phonological
representation and/or processing that does not depend on the ability to utilize
phoneme-level segments to process spoken words. In addition, because they
require only a single, untimed response on each trial, gating tasks place a
6
minimal load on working memory and on the speed of phonological retrieval,
two processes that may be compromised in dyslexic children (e.g., Wagner &
Torgesen, 1987; Wolf & Bowers, 1999). However, general problems with
phonological retrieval (e.g., as on confrontation naming tasks, found in some
dyslexic children, e.g., Wolf & Bowers, 1999) would interfere with gating
performance.
Relatively few studies have utilized the gating paradigm with children
(Edwards, Fourakis, Beckman & Fox 1999; Elliot, Hammer, & Evan, 1987;
Elliot, Scholl, Grant, & Hammer, 1990; Griffiths & Snowling, 2001; Metsala,
1997a,b; Montgomery 1999; Murphy, Shea & Aslin 1989; Walley 1988;
Walley, Michela, & Wood, 1995; Wesseling & Reitsma, 2001; Munson, 2001).
The findings are generally that children require more gates (i.e., more of the
word) for identification than adults. For example, Metsala (1997a) assessed
gating performance in 7-, 9-, and 11-year old children, as well as adults. The
7-year-olds required significantly more gates (i.e., more acoustic input) than
11-year olds and adults to identify target words, as well as initial phonemes of
the target words. Both children and adults required fewer gates to identify
high-frequency words, and words that had a larger number of lexical neighbors
(i.e., words that share more phonemic units with the target words). The results
were interpreted in terms of the lexical restructuring hypothesis (1993; Metsala
& Walley, 1997a). According to this view, phonological representations of
young children are initially holistic. Due to the pressure of vocabulary
7
expansion representations become increasingly segmental and eventually reach
the level of the phoneme. For words with denser neighborhoods there will be
greater pressure to restructure, and restructuring should occur at a younger the
age.
Still fewer studies using the gating paradigm have included dyslexic
children in the sample. Metsala (1997b) administered a gating task to dyslexic
and age-matched normal readers. Words from dense lexical neighborhoods
were compared to words from sparse neighborhoods. This variable was
crossed with word frequency. Frequency did not interact with reader group,
although there was an interaction of reader group and neighborhood density.
In contrast to Metsala (1997a), normal readers needed fewer gates to identify
words from sparse neighborhoods than dense neighborhoods, the opposite of
the prediction made by the lexical restructuring hypothesis. Dyslexic children
needed more gates than the normal readers to identify words in sparse
neighborhoods, whereas the groups did not differ in gating performance on
words in dense neighborhoods. Metsala (1997b) claimed that the dyslexic
children resembled the younger normal readers in the Metsala (1997a) study,
in the sense of having less segmentally organized representations, but this
conclusion does not appear to follow from the data.
Griffiths & Snowling (2001) found that dyslexic and normally
achieving readers (age 8-12 years) required the same number of gates to
identify words. Subject group did not interact with neighborhood density, as
8
had been the case in the Metsala (1997b) study. The dyslexic sample showed
the commonly observed pattern of deficits in nonword pronunciation and rapid
name retrieval. Griffiths and Snowling (2001) concluded, contrary to Metsala
(1997b), that dyslexic children in fact had segmentally organized phonological
representations. They argued, based on null results for the gating task and the
presence of rapid name retrieval difficulties, that phonological deficits in
dyslexia involve problems in the generation of phonological output, rather than
the adequacy of phonological representations per se.
Metsala’s (1997b) and Griffiths and Snowling’s (2001) interpretation
of the gating task is at odds with the treatment of this task in the adult
literature, in one respect (Grosjean, 1980; Salasoo & Pisoni, 1985; Tyler &
Wessels, 1985; Warren & Marslen-Wilson, 1987). Successful performance on
a gating task requires adequate auditory processing and adequate phonological
representations of the sounds of words, but gating task performance provides
no direct information on the segmental nature of the representations. The
presence (or absence) of lexical neighborhood effects has multiple
interpretations. Furthermore, these effects have no direct bearing on the
segmental or non-segmental nature of processing in the gating task.
Neighborhood effects, if they exist, might reasonably operate at the lexical
level rather than at the level of phonemic segments. For example, it might be
the case that words in sparse neighborhoods are easier to identify, because
there are fewer competing lexical neighbors.
9
A methodological issue with both the Metsala (1997b) and Griffiths &
Snowling (2001) studies is that responses were scored correct only when
subjects guessed the exact target word for a given trial. According to this
scoring method, if the target word was /kaet/ a response of /kaep/ would be
considered incorrect. However, both words end in a stop consonant, and
because stop consonants typically have relatively little influence on the
articulation of the preceding vowel, these two words sound virtually the same
until articulation of their respective final stop consonants (/t/ or /p/). The
incorrect response is not given partial credit, so to speak, for the strong
resemblance of /kaet/ to /kaep/, as opposed to /kaen/ or /kael/.
In the present study, the gating procedure provided more direct
information about the organization of phonological representations in dyslexic
and non-dyslexic children. Warren and Marslen-Wilson’s (1987)
demonstration that listeners are sensitive to nasal coarticulation, and West’s
(1999) demonstration that listeners are sensitive to liquid coarticulation were
considered. For purposes of the present study it was thus reasoned that if
normal readers had more fully specified phonological representations, they
would be able to distinguish between words with nasals (or liquids, e.g. lateral
/l/) and other types of words (e.g., with oral stops) at earlier gates than
dyslexics. Two different scores were obtained for responses in the gating task,
first, a category score, based on whether the child named a word in the correct
category (nasal vs. lateral (let’s vs. oral stop), and second, an exact match
10
score, based on whether the child named the correct word. Hypotheses
indicated poorer performance of dyslexic children compared normal readers of
the same age on both measures.
Although the primary goal of this study was to explore group
differences between dyslexic and non-dyslexic children on the gating task,
individual differences were also investigated within the dyslexic sample. Two
critical dimensions that may be important are the degree of phonological
deficit, and the degree of language impairment (Gallagher, Frith, & Snowling,
2000; Griffiths & Snowling, 2002; Joanisse et al., 2000; Manis, Seidenberg,
Doi, McBride-Chang, & Petersen, 1996; Stanovich, Siegel, & Gottardo, 1997).
According to the phonological representations hypothesis, the degree of
phonological impairment should be the primary variable affecting gating task
performance. In order to explore this alternative a measure of phonological
awareness was obtained in the present study. According to the lexical
restructuring hypothesis, performance on the gating task may depend in part on
the child’s level of language development, with vocabulary perhaps assuming
the most important role (Metsala & Walley, 1997; Walley, 1993). Therefore,
several measures of language ability were obtained. These included measures
of receptive and expressive vocabulary, ability to follow oral directions, and
sentence memory. Covariance and multiple regression analyses were used to
parse the relationships observed among gating performance, phonological
awareness, language skill and reading.
11
METHOD
Participants
Children were recruited from elementary schools in a major United
States metropolitan area. The total number of subjects was 46 (23 dyslexics
and 23 normally achieving readers). Children ranged from 8 to 14 years of age
(dyslexics, 9-14 years and normally achieving readers, 8-14 years). There
were 14 boys and 10 girls in the dyslexic sample and 13 boys and 13 girls in
the normally achieving sample. In order to be included in the study children
were required to have a scaled score greater than 7 (corresponding to a
standard score of 85) on either the verbal (average of Vocabulary and
Similarities subtests) or performance (average of Block Design and Picture
Completion subtests) estimate of WISC-III IQ (Wechsler, 1992). This
criterion was used to avoid children who would likely have poor reading
ability due to general cognitive impairments, but to avoid overly restricting the
range of oral language ability within the dyslexic sample. Children were
excluded from the study based on the following criteria as determined from
parent report: neurological problems, uncorrected hearing or vision problems,
serious emotional or behavioral problems, and lack of English fluency.
Dyslexic children. Dyslexics were defined by a score at or below the
25
th
percentile (SS=90) on either of two subtests (Word Identification or Word
Attack) of the Woodcock-Johnson Reading Mastery Test-Revised (Form G)
(Woodcock, 1987). Both subtests are standardized measures of reading level.
12
Word Identification contains age-appropriate English words and Word Attack
contains orthographically regular nonwords whose pronunciations are scored
based on common patterns of spelling-sound correspondence in English.
Norms updated in 1993 were used to calculate percentile scores (Woodcock,
1998).
Normally achieving readers: To be classified as a normally achieving
reader a child was required to score at or above the 40
th
percentile on both
Woodcock Word Identification and Word Attack. All other criteria for
inclusion in the sample applied.
Test of Phonological Awareness
The Elision task from the Comprehensive Test of Phonological Processing
(Wagner, Torgesen, & Rashotte, 1999) was administered in the standard
format. The Elision task requires children to delete syllables or phonemes
from words spoken by the examiner. The test mostly taps phoneme awareness,
as only the first three items involve syllable deletion. The remaining 17 of the
20 test items involve the deletion of a single phoneme (from the word onset,
middle or end, including blends at the beginnings of the words). The test is
terminated if the child misses three in a row. The test manual reports
Cronbach’s Alpha reliability at age 10 of 0.91.
Oral Language Tasks
13
Language ability was assessed by means of the following standardized
tests: Concepts and Directions and Recalling Sentences from the Clinical
Evaluation of Language Fundamentals- Revised (CELF-R) (Semel, Wiig &
Secord, 1995), and the Receptive One Word Picture Vocabulary Test
(ROWPVT) (Brownell, 2000). Concepts and Directions requires subjects to
listen to a short sentence and carry out the action by pointing to black and
white geometric forms (e.g., “Point to the small, white square after you point to
the large triangles.”). Recalling Sentences requires subjects to listen to
sentences of varying lengths and repeat each one back verbatim.
Gating Task
The listener was presented (via headphones) partial word segments of
varying durations and asked to identify the words. There were a total of 25
words (3 practice, 22 test items) and each word was divided into 6 gates (see
Appendix for words and gate lengths). The gates were adjusted according to
important acoustic properties of the words and therefore were not of equal
durations. The first gate included only the initial consonant(s). The second
gate added the initial 25 msec of the vowel, that is, most of the CV formant
transition interval. From this gate the initial consonant(s) can generally be
fully perceived, along with partial information about the vowel. The third gate
gave an additional 25 ms of the vowel and included the entire CV formant
transition interval; from this gate the initial consonant(s) plus the vowel should
ordinarily be perceived. The fourth gate presented all except the last 25 msec
14
of the vowel, further ensuring correct perception of the vowel. Nasalization
and lateralization of vowels should be quite obvious in the fourth gate. The
fifth gate added the last 25 msec of the vowel, that is, most of the VC formant
transition interval. From this gate the entire word is likely to be correctly
perceived. Gate 6 added the final consonant, and hence was the full
presentation of an intact, complete word (100 percent of the auditory
information from the word.) Since the word sets contained different vowels
which inherently differ in duration, and since the final consonant affects vowel
duration, words differed in their total vowel durations within and across sets.
Thus from gate 4 through gate 6, gate durations differ as a function of original
vowel durations.
1
For the practice trials, each of three practice items was played four
times (gates 3 through 6) to familiarize subjects with the gating task. For the
experimental trials, each of 22 test items was presented in a series of five
blocks of trials, one gate per block. Presentation began with gate two. Gate 1
was not used because pilot testing indicated that children were not able to
guess the correct word or guess a word in the same category of ending sound
(nasal, lateral, stop) based on such limited acoustic information. Many
children claimed to hear only a piece of static at this first gate, and therefore,
refused to guess the word.
The order of items was randomized within each block but blocks were
always presented in ascending order from 2 through 6. Responses were
15
recorded via audio cassette for later transcription. Children were encouraged
to make a guess on each item regardless of confidence level. If a child refused
to give a response at a particular gate, the response was removed from
subsequent analyses. No feedback was given at any time except during
administration of the practice items.
All test items were monosyllabic English words. Approximately half
the words were low frequency (e.g., coal) and half were high frequency (e.g.,
cat) based on Carroll, Davies & Richman’s Word Frequency Book (1971).
Words possessed the following phonological structure: consonant, vowel,
consonant, or consonant, consonant, vowel, consonant. Groups of words that
have CVC or CCVC structures but differed in their final consonant were
constructed. The final consonant varied in terms of manner of articulation:
nasal (/kon/), lateral (/kol/), or oral stop (/kod/). Most of the stimuli were
constructed as minimal pairs (e.g. sweat/swell) but 2 sets were minimal triples
(cone/coal/code, bone/bowl/boat). The words were selected to minimize the
presence of highly familiar shorter words contained within them, therefore
reducing the tendency of subjects to identify a highly salient shorter word, and
then possibly perseverate on that response when presented with longer gates.
For example, within the cone/coal/code triple, “co” is not a word.
Train/trail/trade could not be used as a stimulus set in our experiment, since
“tray” is a word. Within feet/feel, “fee” is a word, but most likely not a highly
salient one for children.
16
The set of words was recorded by a female who’s word pronunciations
were typical of Californian speakers. The recording was made in a sound
booth to digital audio tape, which was then transferred to a computer disk and
edited using Praat software. Gates were produced by cutting at zero-crossings,
but the amplitude of the end of the gate was not ramped. Amplitudes were not
normalized during the experiment, but the recorded levels of the items were
similar.
The primary purpose for using the gating task was to determine if
children were able to identify the category of word based on differences in the
sound quality of the vowels due to the final consonant (i.e. coarticulation). For
clarity and ease of interpretation children were asked to identify the exact word
from the earliest possible gate. Their answers were assumed to reflect their
ability to use coarticulation.
Reliability (Cronbach's Alpha) was calculated for the 22 test items in
terms of gating scores for all subjects. When using the more lenient scoring
method (an item is scored as correct if a the answer is a word in the same class
as the target, the exact word need not be given) Cronbach's Alpha was .689.
Removing any one of the test items resulted in minor shift in Cronbach’s
Alpha (values ranged from .643 to .699 with removal of any one item.) When
using the first gate at which an exact match to the target word was achieved
Cronbach's Alpha was .795. Again, removing any one of the items resulted in
17
only minor shifts in Cronbach’s alpha (values ranged from .773 to .796 with
removal of any one item).
Procedure
Testing was conducted in a quiet room either at a University laboratory
or at the children’s schools. The entire battery of tests was completed in
approximately 2 hours (including breaks) and took place within a period of
time that did not exceed three weeks for any given child.
18
RESULTS
Descriptive Data
Mean scores on the standardized tests for dyslexic and normally
achieving readers are shown in Table 1. There were no group differences in
general cognitive ability as measured by the WISC-III performance estimate.
The mean verbal estimate of WISC-III was significantly higher for normally
achieving readers than dyslexics as a group, F (1, 44) = 17.38, p < .001, U
2
=
.28. The dyslexic group was also lower on all three language tasks, and on
Elision (see Table 1 for F-values).
Scoring of the Gating Task
One dependent measure for the gating task was the first gate at which
the child was able to name a word that was within the correct category of
words as determined by their final consonant (nasal, lateral or oral stop
consonant). For example, if the word was /kon/, responses of /kon/ and /kom/
would be accepted as correct because both words have a nasal consonant
following the vowel. A response of /kot/ would be incorrect in this case
because /kot/ ends in a stop (/t/) and /kon/ ends in a nasal (/n/). The gate at
which a subject first correctly identified the category of the final consonant
was used as his/her score even if the child later changed his/her answer.
Analyses by Griffiths & Snowling (2001) supported this last scoring
19
Table 1 Means and standard deviations on standardized tasks
Undifferentiated groups.
Dyslexic (N=23) Control (N=23)
Measures Mean SD Mean SD F-value,
sig.
3.13, Age in
months
142.75 16.13 134.04 17.19
p<.05
Word Ident.
SS
78.43 8.09 107.04 7.1 162.5,
p<.001
Word Attack
SS
86.83 9.61 107.43 7.73 64.21,
p<.001
WISC-III
Verbal Est.
8.67 2.55 11.52 2.06 17.39,
p<.001
WISC-III
Perf. Est.
9.98 2.18 11.09 2.23 2.91
Recalling
Sent. SS
6.74 3.39 11.13 2.44 25.47,
p<.001
ROWPVT SS 97.57 9.64 105.7 12.81 5.92,
p<.05
Con and Dir.
SS
7.39 3.33 11.39 2.9 18.87,
p<.001
Elision SS 7.04 3.01 10.91 2.56 22.10,
p<.001
Nonword Rep
SS
7.24 2.42 10 2.68 12.24,
p<.01
20
procedure, as they found a similar pattern of results whether counting the first
correct response or counting only consistently correct responses. A second
dependent measure was the first gate at which a subject was able to produce an
exact identification of the target word. Once again, changes in responses at
later gates were ignored in the analyses.
All words were divided into 6 gates as described in the previous
section, and since subjects were presented with gates 2-6, therefore the lowest
and best possible score for each word was 2. Success at gates 2, 3, or 4 meant
that a subject used anticipatory acoustic information; in general this was
expected to be very unlikely for gate 2 and quite likely for gate 4. If a subject
never correctly identified a word, a score of 7 was assigned for that word
(largest gate plus 1). In such a case it was assumed the subject would be able
to correctly identify the word if it was repeated or presented in context. This
was the scoring method used in the two previous gating studies with dyslexic
children (Griffiths & Snowling, 2001; Metsala, 1997b). We also report
alternative analyses in which subjects’ responses were excluded from analyses
if they made an error in identifying either the category or the exact word on the
6
th
gate.
21
Gating Task Group Results
Children were able to identify the correct category of ending consonant
somewhere between the 3
rd
and 4
th
gate, on average (Table 2). Results were
calculated for individual item type (stops, liquids, nasals) and for all items
combined. Detection at the 3
rd
/4
th
gate means many children were able to
detect the category of consonant prior to the end of the vowel. Stops were
easiest, followed by laterals and then nasals. Exact matches required more
gates on average (4 or 5). There were no group differences in the mean scores
for correct category or exact match across consonant types. Analyses of
individual item types revealed a group difference favoring the normal readers
on the nasal category.
One concern with the gating task is that some of the words might not be
recognizable, either because they were unfamiliar to subjects, or because the
audio recording was not clear enough. No single item was missed (after
presentation of gate 6) by more than 1/3 of the subjects, substantiating the
claim that the recording/playback of each item was preserved in terms of sound
quality. There were four subjects, three dyslexic children and one non-
impaired reader, who failed to identify 5 or more words by gate 6 (out of 22
items). The data was re-analyzed excluding individual subjects’ data when
failures of this type occurred.
22
Table 2 Means on the gating measures
(undifferentiated groups)
Category: first gate at which correct identification of consonant
category was achieved (maximum score: 7)
Exact: First gate at which an exact match to the target word was
achieved (maximum score: 7)
Dyslexic (N=23) Control (N=23)
Measures Mean SD Mean SD F-
value;
sign.
Mean
category
ident.
3.48 0.52 3.28 0.25 2.77
Category
– Stops
3.27 0.64 3.18 0.35 0.33
Category
– Liquids
3.43 0.6 3.34 0.46 0.38
Category
– Nasals
3.77 0.66 3.37 0.55 4.91, p
<.05
Mean
exact
ident.
4.6 0.59 4.43 0.35 1.44
Exact –
Stops
4.61 0.63 4.57 0.39 0.08
Exact –
Liquids
4.42 1.06 3.93 0.65 3.54
Exact –
Nasal
4.67 0.58 4.5 0.54 1.09
23
Excluding errors lowered all of the gating scores slightly (by about 1/10
of a gate for the composite categorization score and about 1/5 of a gate for the
composite exact match score). The group comparisons were unchanged
(significant at p < .05 only for the nasal category).
To summarize, dyslexics showed trends toward requiring more gates to
identify the category of the word or the exact word itself across all consonant
types (about 1/5 to 1/2 of a gate). However, group differences were significant
only for the most difficult item type (nasals) and then, only in the ability to
detect the correct category of the consonant. The category identification
measure is a more sensitive measure than the exact match measure used in
previous studies, but group differences were not impressively large. It is
possible that deficits in the integrity of phonological representation/processing
on the gating task would only be found for a subset of dyslexics. Therefore,
individual difference analyses were conducted.
Subgroup Analyses
Two individual difference variables from previous work might be
expected to be related to performance on phonological processing tasks,
language ability (Joanisse et al., 2000; Manis & Keating, 2004; Manis et al.,
1997) and phonological awareness (Manis et al., 1997; Griffiths & Snowling,
2001). Inspection of the individual data revealed that this sample did not
include many dyslexic children with language ability in the normal range on
two or more of the language tasks administered, but about half of the dyslexic
24
sample scored within the normal range on a measure of phonological
awareness (Elision). Accordingly, the sample was divided into a low Elision
subgroup (n = 11, subsequently refered to as low PA) and a non-impaired
Elision subgroup (n = 12, subsequently refered to as non-impaired PA). Both
groups were compared to the non-impaired reader group. Scores on language,
reading and cognitive ability tests, as well as gating measures, are shown in
Table 3 for these groups, along with the group comparisons that were
significant by post hoc test (Tukey’s procedure). The dyslexic subgroups
differed from each other, and the non-impaired PA dyslexics did not differ
from the non-impaired reader group, on the Elision task according to Tukey
test, validating the subgroup division. Nonword Repetition scores were also
in the expected direction, although only the difference between the low PA
and non-impaired reader group was significant by Tukey’s test.
Analyses comparing the subgroups and the non-impaired readers
revealed significant F-values for mean gate at which a word in the correct
category was identified, for stops, F (2, 43) = 7.66, p < .001, liquids, F (2, 43)
= 5.31, p < .01, and nasals, F (2, 43) = 3.56, p < .05 (Table 4). The low PA
group differed from both the non-impaired PA group and the non-impaired
reader group by Tukey post hoc tests, for all three consonant class measures.
Differences in mean gate for exact matches were not significant, although there
25
Table 3 Means and standard deviations on the standardized tests
For subgroups defined in terms of Elision scores
(means that do not differ by Tukey post hoc test at p < .05 share a superscript)
Low PA High PA Normal
Dyslexic Dyslexic Readers
Variable (n=11) (n=12) (n=23)
F-
values
&
signif.
Age (months)
138.0
(16.02)
1
147.08
(15.63)
1
134.04
(17.19)
1
2.46
Word Ident. SS
75.40
(7.34)
1
81.25
(7.99)
1
107.04
(7.10)
2
87.94,
p<.001
Word Attack SS
83.82
(9.49)
1
89.58
(9.24)
1
107.43
(13.52)
2
34.57,
p<.001
WISC-III Verbal
Est.
8.13
(2.20)
1
9.17
(2.83)
1
11.52
(2.06)
2
9.29,
p<.001
WISC-III Perf. Est.
9.0
(2.12)
1
10.87
(1.88)
1,2
11.09
(2.25)
2,3
3.82,
p<.05
Recalling Sent. SS
5.91
(3.62)
1
7.50
(3.12)
1
11.13
(2.44)
2
13.78,
p<.001
ROWPVT SS
96.09
(8.21)
1
98.92
(10.97)
1
105.70
(12.81)
1
3.09,
p<.06
Concept & Dir. SS
6.00
(2.61)
1
8.67
(3.50)
1
11.39
(2.90)
2
12.45,
p<.001
Elision SS
4.55
(1.21)
1
9.33
(2.19)
2
10.91
(2.56)
2
30.92,
p<.001
Nonword Rep. SS
6.11
(1.96)
1
8.09
(2.59)
1,2
9.17
(1.83)
2
3.74,
p<.05
Word Ident. and Word Attack: Woodcock Reading Mastery Word Identification
and Word Attack; WISC-III Verbal Est.: WISC-III mean scaled score on
Vocabulary and Similarities; WISC-III Perf. Est.: WISC-III mean scaled score on
Block Design and Picture Completion; ROWPVT: Receptive One-Word Picture
Vocabulary Test; Concept and Dir.: CELF-R Concepts and Directions scaled score;
Elision: CTOPP Elision; Nonword Rep: CTOPP-R Nonword Repetition.
26
Table 4 Means and standard deviations on gating measures
Subgroups are defined in terms of Elision scores (means that do not differ by
Tukey post hoc test at p < .05 share a superscript)
Variable
Low
Elision High Elision
Normal
Readers F-values
Dyslexic
(n=11)
Dyslexic
(n=12) (n = 23) & signif.
Category –
all
3.78
(0.57)
1
3.20 (0.27)
2
3.28 (0.25)
2
9.24,
p < .001
Category –
Stops
3.65
(0.64)
1
2.92 (0.41)
2
3.18 (0.35)
2
7.66,
p < .001
Category –
Liquids
3.77
(0.49)
1
3.12 (0.53)
2
3.34 (0.46)
2
5.31,
p < .01
Category –
Nasals
3.95
(0.75)
1
3.59 (0.54)
2
3.37 (0.55)
2
3.56,
p < .05
Exact –
all
4.81
(0.56) 4.41 (0.56) 4.43 (0.35) 2.65
Exact –
Stops
4.83
(0.49)
1
4.41 (0.70)
1
4.57 (0.39)
1
2.01
Exact –
Liquids 4.73 (1.0)
1
4.15 (1.09)
1,2
3.93 (0.65)
2
3.13
Exact –
Nasals
4.83
(0.66)
1
4.53 (0.49)
1
4.50 (0.54)
1
1.37
27
Figure 1 Gating Scores Based on Phonological Awareness Group
Average gating score based on Phonological Awareness group.
Scores represent the first gate at which a category identification or
exact match was made. Bars represent standard deviations.
28
was a trend toward longer gates for the liquids. No differences were
found between the non-impaired Elision dyslexic and the non-
impaired reader groups on any of the gating measures.
It is apparent in Table 3 that the dyslexic subgroups were lower than
the non-impaired reader group in performance IQ estimate, verbal IQ estimate
and language ability. Because subgroup differences in phonological awareness
were thus confounded with differences in performance IQ estimate and
measures of linguistic ability, we repeated the analyses of variance on the
gating measures, covarying out performance IQ, verbal IQ and language ability
(composite of the three language tasks). Group differences on two of the
gating measures remained significant: stops, F (5, 40) = 4.80, p < .025, and
liquids, F (5, 40) = 4.29, p < .025.
To summarize, the subgroup analyses revealed that dyslexic children
with low phonological awareness were particularly impaired on the gating task
even compared to dyslexics with non-impaired phonological awareness.
Differences between these groups remained robust on the stop and liquid
items, even when group differences in estimated performance IQ and verbal
ability were statistically controlled. The primary difference between the
subgroup analyses and the whole group analyses is that differences emerged on
the less difficult items in the subgroup analyses (stops and liquids).
29
Correlation and Regression Analyses
Subgroup analyses were useful because they revealed more severe
problems on the gating task for a subgroup of low performers on the
phonological awareness task. However, this approach makes use of only a
portion of the data. Whole group correlation and regression analyses were
used to relate gating performance to phonological awareness and determine the
strength of the relationship for the sample as a whole. In addition, this allowed
us to determine how much these variables overlapped with the other variables
in the study, such as language ability and estimated IQ.
Two composite gating measures were correlated, category
identification and exact identification (collapsing in both cases across the three
item types), with the other tasks (see Table 5). Category identification scores
were correlated with standard score measures of Word Identification, Word
Attack, Performance IQ estimate, Concepts and Directions, ROWPVT,
Recalling Sentences, and Elision. Exact identification correlated only with
category identification and was excluded from further analyses.
30
31
Several hierarchical regressions (Table 6) were used to clarify the
relationship between gating and reading, taking into account variations in age,
IQ, language ability and phonological awareness. Gating performance is
thought to index the overall quality of phonological representations, but
obviously there are other measures that would be expected to depend on
phonological representations, such as the measures of phonological awareness
and phonological decoding. One possibility is that gating performance
contributes directly to individual differences in reading skill, based on the
importance of intact phonological representations to word decoding and
recognition. Alternatively, gating performance might contribute indirectly to
reading, via phonological awareness or language ability.
Three sets of hierarchical regressions predicted standard scores for
Word Identification, Word Attack and Elision. As category identification
score was the gating variable with the strongest correlations (suggesting it was
the most sensitive), it was used in all regressions. We collapsed the verbal and
performance IQ estimates into a single estimated IQ measure. Age was
entered first in all regressions, followed by estimated IQ, average language
score, category identification score and Elision on the remaining four steps, in
various orders to determine their effects on the final regression equation. Two
orderings of the variables, designed to show the unique contribution (if any) of
category identification and Elision, when entered on the last step of the
32
Table 6. Hierarchical Regression Analyses for Three Criterion Variables
Word Identification:
Variable R2 Change in R2
Final Beta
Weight
1. Age (in months) .09 .05* -.141
2. IQ Estimate .381 .291*** .143
3. Language Ave. .487 .106** .289
4. Category Ident. .493 .006 .061
5. Phoneme Elision .598 .105** .446**
----------------------------------------------------------------------------------------
4. Phoneme Elision .596 .109** .446**
5. Category Ident. .598 .002 .061
*
p < .05
**
p <.01
***
p < .001
Word Attack:
Variable R
2
Change in
R
2
Final Beta
Weight
1. Age (in
months)
.085 .085* -.124
2. IQ Estimate .32 .235*** .071
3. Language
Ave.
.433 .113** .272
4. Category
Ident.
.450 .017 .022
5. Phoneme
Elision
.574 .124** .486***
-----------------------------------------------------------
4. Phoneme
Elision
.574 0.141*** .486***
5. Category
Ident.
.574 .000 .022
*
p < .05
**
p <.01
***
p < .001
33
Table 6. Continued
Phoneme Elision:
Variable R
2
Change in
R
2
Final Beta
Weights
1. Age (in
months)
.056 .056 .161
2. IQ Estimate .273 .217*** .068
3. Language
Ave.
.381 .108** .369*
4. Category
Ident.
.472 .091* -.340*
*
p < .05
**
p <.01
***
p < .001
34
analysis, are shown in Table 6 for Word Identification and Word Attack. A
third analysis predicting Elision scores is shown in the lower part of the table.
In the regressions for Word Identification and Word Attack, the gating
measure accounted for significant variance on the second step, but when either
IQ or average language ability was entered on the third step, the contribution
of the gating measure became non-significant. Elision accounted for 10.5%
and 12.4% unique variance (i.e., when it was entered last in the equation) but
category identification accounted for less than 1% unique variance in Word
Identification or Word Attack scores when entered on the last step. In an
additional analysis (not shown in Table 4), we entered just age and the
performance estimate of IQ along with the gating measure. The contribution
of the gating variable remained significant. This demonstrates that it is the
verbal ability measures (estimated verbal IQ or average language score) and
the phonological awareness measure (Elision) that reduced the contribution of
the gating measure to non-significance. In other words, the common variance
between gating and the reading measures is shared with the language and
phonological measures, but gating performance did not account for unique
variance in reading beyond these measures.
The effect of one variable, Elision, in reducing the contribution of
gating performance to variability in reading scores was particularly strong.
Accordingly, we ran a hierarchical regression controlling for age, average IQ
and average language ability, exploring the contribution of gating to Elision
35
performance. Category identification accounted for 9.1% unique variance in
Elision, when entered on the fourth step (see the third panel of Table 6).
Another way to represent the relationships among the variables is by means of
path analysis. Figure 1 shows the beta weights for each variable’s unique
contribution to Elision, and to Word Identification. Age and IQ did not make
significant unique contributions to Elision, Word Identification or Word
Attack, so they are not shown in the diagram. Average language ability and
gating make indirect contributions to Word Identification and Word Attack,
mediated by their relationship with Elision. The overall model accounted for
59.8% of the variance in Word Identification, and 57.4% of the variance in
Word Attack.
Data represented in the path analysis were collected concurrently,
hence a directional relationship can not be conclusively determined. It is
possible that the development of reading skills has an impact on phoneme
awareness or gating performance, and the proper description of the
relationships among the variables would be bi-directional and reciprocal.
36
37
DISCUSSION
The purpose of the study was to compare the amount of auditory input
necessary to identify spoken words in dyslexic children and normally
achieving readers using a gating task. Dyslexic children required more gates to
generate a word response in the correct category for words ending in nasals,
indicating they were less sensitive to anticipatory co-articulation. The
differences were small, however, and there were hints in the correlational data
that phonological awareness might be an important variable. Accordingly, the
dyslexic sample was divided about the median into groups based on
phonological awareness (Elision task performance). The low phonological
awareness subgroup required more gates to identify a word in the correct
consonant category (for all three consonant types) than dyslexics with non-
impaired phonological awareness (and non-impaired readers). The differences
on the stops and liquids were robust across statistical controls for IQ and
language ability.
In previous work on gating with dyslexic samples (Griffiths &
Snowling, 2001; Metsala, 1997b), group differences were not observed. One
possibility is that measures of anticipatory coarticulation (used here) are more
sensitive than word identification measures used in previous studies. In
addition, present results reveal that primarily dyslexic children with low
phonological awareness show deficits on the gating task. Because anticipatory
coarticulation is perceived as a phonetic feature on the vowel (e.g., nasalized
38
vowel), it blends information from the vowel and consonant, and hence yields
information about the adequacy of the phonological representation (Warren &
Marslen-Wilson, 1987), and does not shed light on whether the representations
are organized into phonemic segments (as has been previously argued –
Griffiths & Snowling, 2001; Metsala, 1997). The fact that dyslexics performed
more poorly on this task in turn implies that their phonological representations
for common words are less well integrated than those of non-impaired readers.
This finding extends previous results obtained with gating tasks in dyslexic
samples (Griffiths & Snowling, 2001; Metsala, 1997b), and supports the
growing literature suggesting that dyslexics have less complete phonological
representations of printed words (Elbro, et al., 1998; Swan & Goswami, 1997;
Wesseling & Reitsma, 2001).
Regression analyses were conducted to provide a finer-grained picture
of the relationships between gating performance, language, IQ and
phonological awareness. Hierarchical regression analyses indicated that gating
task performance contributed unique variance to the phonological awareness
task, which in turn contributed unique variability to word and nonsense word
reading (see Figure 1); gating made indirect contributions to the reading
variables, mediated by phonological awareness. Gating did not make direct
contributions to the reading tasks, once age, IQ, language ability and
particularly, phonological awareness, were taken into account.
39
There are at least two theoretical interpretations of the pattern of group
differences and the regression findings. First, as previously argued, the gating
task may index fundamental qualities of the phonological representation and
processing of spoken words. Even subtle deficiencies at this level may
interfere with the development of phonological awareness, which in turn
interferes with the learning of spelling-sound correspondences, and hence
reading progress in general. This would mean that dyslexic children have
separable deficits in phonological representations and in the development of
phonological awareness, particularly at the level of the phoneme. It is possible
that dyslexic children approach the reading task with a deficit in phonological
representation or access to these representations (Ziegler & Goswami, 2004)
and acquire the phonological awareness deficit as a consequence of attempting
to apply inadequate phonological representations to the demanding task of
learning spelling-to-sound correspondences. This argument is consistent with
Swan and Goswami’s (1997) finding that correcting for knowledge of the
names of spoken words erased dyslexic-non-impaired reader differences on
rime-onset and syllable level phonological awareness tasks, but not phoneme-
level awareness tasks. In addition, Ziegler and Goswami (2004) report that
cross-nationally, difficulties in phoneme segmentation appear to arise
developmentally during the period of letter-sound correspondence learning,
regardless of the child’s age at the outset of literacy, and to be exacerbated in
languages where the spelling-sound correspondences are irregular and/or
40
complex, such as English, Danish and French. Whether this view is correct or
not in its entirety, it seems almost certainly to be the case that learning to read
has reciprocal influences on phonological awareness (Morais et al., 1979;
Perfetti et al., 1987). A conclusion has not been drawn regarding whether the
experience of reading has reciprocal influences on gating task performance.
A second interpretation of our group differences and regression
findings is that dyslexic children have intact holistic representations of spoken
words, but fail to organize their phonological representations at the segmental
level, particularly phonemic segments (Metsala, 1997b). Metsala (1997b)
argued that her data fit the lexical restructuring hypothesis, in which holistic
phonological representations of young children gradually give way to
phonemically segmented representations owing (as a result of growth in oral
vocabulary) to the need to make more fine-grained distinctions between
similar-sounding words (e.g., Walley, 1993). Metsala (1997b) found that
dyslexic children required more gates to identify words with few lexical
neighbors, but did not differ from non-impaired readers in the number of gates
needed to identify words with many lexical neighbors. Metsala (1997b)
argued that dyslexics were showing a temporary lag in the development of
segmental representations. It looks like 3 arguments against this interpretation.
First, result is not previously robust (Griffiths & Snowling, but also Metsala
according to the intro); second (mentioned in intro but not here), Metsala’s 2
studies are contradictory so she can’t really have a story about lexical
41
neighborhoods; third, gating doesn’t bear on segmentation in the first place. Is
it possible that we’ve already made these points in the intro and don’t need to
repeat them here as if there’s really a viable second interpretation? However,
Griffiths and Snowling (2002) failed to replicate this result. In addition, the
argument may be flawed on logical grounds, as gating performance provides
an index of the overall quality of the phonological representation/processing of
words, and does not shed light on their segmental nature. This is a more
categorical statement than earlier in paper. In particular, the anticipatory co-
articulation effects observed in our study, and the fact that dyslexics were less
sensitive to them, suggest there is an impairment in phonological
representations, or in the ability to utilize them to process partial auditory
information in spoken words.
One of the present findings, the difficulty of identifying words ending
in nasalizations regardless of group membership, deserves further scrutiny.
The finding suggests greater difficulty in anticipating nasalization of a vowel
relative to lateralization or the control stimulus, a vowel proceeded by an oral
stop. It was on this most difficult of tasks that dyslexics as a group showed the
most impaired performance compared to normally achieving readers. This
could indicate that phonological representations for words ending in
nasalization (or more specifically for nasal phonemes) are particularly
impaired among dyslexic children for some reason. Alternatively, nasals were
more difficult to identify for both subject groups, suggesting that we were
42
simply seeing greater group differences on the more demanding task (a
measurement characteristic). To distinguish between these two possibilities
dyslexic and non-dyslexic readers would need to be tested on a wider range of
stimuli, incorporating anticipatory features of pronunciation that are equal to or
greater in difficulty than nasalization and lateralization, such as coarticulation
between vowels and velars. If dyslexics have an overall impairment in
phonological representations, it should be present on any demanding holistic
perception task.
The study is limited in several respects. First, reading-level matched
younger normal readers were not included. The typical argument for this
group is that the task in question (in this case, gating performance) might vary
as a function of reading experience, as indexed by reading level. A strong test
of the hypothesis that the gating deficit is a core deficit in dyslexia would be
the observation of differences favoring reading level-matched younger normal
readers. However, given the lack of group differences even between
chronological-age matched normal readers and dyslexics in some conditions or
age levels in past studies (Griffiths & Snowling, 2002; Metsala, 1997b), it was
important first to establish that such group differences existed. Our sample had
5 younger normal readers who could be equated to a subset of the dyslexic
sample (15 dyslexic children) based on the mean and range of raw scores on
the Word Identification test. When these two groups were compared on the
gating task, group differences failed to emerge. However, there are obvious
43
power limitations for such an analysis. Future studies will need to include a
reading level comparison group.
A second limitation is that the dyslexic sample did not have the same
overall IQ as the normal reader sample, as is often the case in traditional
studies of dyslexia. This was primarily due to the fact that dyslexic children
with low oral language skills (this tends to deflate the verbal portion of the IQ-
estimate) were admitted to the sample. Language ability accounted for some
variance in both Word Identification and Elision scores for the sample as a
whole, but it did not detract from the relationship of gating to Elision.
However, caution must be observed in generalizing our findings to other
dyslexic samples. It is possible that differences would be difficult to observe if
only dyslexics with average oral language ability were tested.
To conclude, the use of a gating task enabled exploration of the
integrity of phonological representation and processing in dyslexic children by
means of a novel method for the dyslexia literature (sensitivity to anticipatory
coarticulation). Past studies using the gating paradigm did not consistently
find group differences. There were group differences on the measures of
coarticulation sensitivity, particularly for dyslexic children who scored below
average on a test of phonological awareness. In addition, gating performance
was not directly related to word reading and decoding skills, but appeared to be
related indirectly through its relationship with phonological awareness. The
results join the findings of other investigations utilizing different techniques
44
(e.g., Elbro et al., 1998; Swan & Goswami, 1997) in pointing to a basic deficit
in phonological representation and processing in dyslexic children.
45
REFERENCES
Adlard, A., & Hazan, V. (1998). Speech perception in children with specific
reading disabilities (dyslexia). Quarterly Journal of Experimental
Psychology, 51, 153-177.
Bishop, D.V.M. (1997). Uncommon understanding: Development and
disorders of language comprehension in children. Hove. UK:
Psychology Press.
Brady, S. A. (1997). Ability to encode phonological representations: An
underlying difficulty for poor readers. In B. Blachman (Ed.),
Foundations of reading acquisition and dyslexia: Implications for early
intervention (pp. 21-48). Mahwah, NJ: Erlbaum.
Brady, S. A., Shankweiler, D., & Mann, V. A. (1983). Speech perception and
memory coding in relation to reading ability. Journal of Experimental
Child Psychology, 35, 345-367.
Brownell, R. (Ed.) (2000). Receptive One-Word Picture Vocabulary Test, 2
nd
ed. Academic Therapy Publications, Novato, CA.
Bruck, M. (1992). Persistence of dyslexics’ phonological awareness deficits.
Developmental Psychology, 28, 874-886.
Carroll, J.B., Davies, P., & Richman, B. (1971). Word Frequency Book.
Boston, MA: American Heritage.
Castles, A., & Coltheart, M. (1993). Varieties of developmental dyslexia.
Cognition, 47, 149-180.
Chiappe, P., Chiappe, D. L., & Siegel, L. S. (2001). Speech perception,
lexicality and reading skill. Journal of Experimental Child Psychology,
80, 58-74.
Denckla, M., & Rudel, R. G. (1976). Rapid ”automatized” naming (RAN):
Dyslexia differentiated from other learning disabilities.
Neuropsychologia, 14, 471-479.
Edwards, J., Fourkakism M., Beckman, M.E., & Fox, R.A. (1999).
Characterizing knowledge deficits in phonological disorders. Journal
of speech and Language and Hearing Research, 42, 169-186.
46
Elbro, C., Borstrom, I., & Klint Peterson, D. (1998). Predicting dyslexia from
kindergarten: The importance of distinctness of phonological
representations of lexical items. Reading Research Quarterly, 33, 36-
57.
Elliot, L. L., Hammer, M. A., & Evan, K. E. (1987). Perception of gated,
highly familiar spoken monosyllabic nouns by children, teenagers, and
older adults. Perception & Psychophysics, 42, 150-157.
Elliot, L. L., Scholl, M. E., Grant, W. K., & Hammer, M. A. (1990).
Perception of gated, highly familiar spoken monosyllabic nouns by
children with and without learning disabilities. Journal of Learning
Disabilities, 23, 248-259.
Fowler, A. (1991). How early phonological development might set the stage
for phonemic awareness? In S. Brady & D. Shankweiler (Eds.),
Phonological Processes in Literacy: A tribute to Isabelle Y. Leiberman
( pp. 97-117). Hillsdale New Jersey: Lawrence Erlbaum Associates.
Gallagher, A., Frith, U., & Snowling, M. J. (2000). Precursors of literacy-
delay among children at genetic risk of dyslexia. Journal of Child
Psychology & Psychiatry, 41, 203-213.
Godfrey, J. J., Syrdal-Lasky, A. K., Millay, K. K., & Knox, C. M. (1981).
Performance of dyslexic children on speech perception tests. Journal
of Experimental Child Psychology, 32, 401-424.
Goswami, U, Schneider, W, Scheurich, B. (1999). Picture naming deficits in
developmental dyslexia in German. Developmental Science, 2, 53-58.
Griffiths, Y.M., Snowling, M.J. (2001). Auditory word identification and
phonological skills in dyslexic and average readers. Applied
Psycholinguistics, 22, 419-439.
Griffiths, Y.M. & Snowling, M.J. (2002). Predictors of exception word and
nonword reading in dyslexic children: the severity hypothesis. Journal
of Educational Psychology, 94, 34-43.
Grosjean, F. (1980). Spoken word recognition processes and the gating
paradigm. Perception and Psychophysics, 28, 267-283.
Joanisse, M.F; Manis, F.R; Keating, P; Seidenberg, M.S. (2000). Language
deficits in dyslexic children: Speech perception, phonology, and
morphology. Journal of Experimental Child Psychology. 77, 30-60.
47
Liberman, I. Y. & Shankweiler, D. (1985). Phonology and the problems of
learning to read and write. Remedial and Special Education, 6, 8-17.
Luce, P. A. (1986). Neighborhoods of words in the mental lexicon. Research
on speech perception (Tech. Rep. No. 6). Bloomington, IN: University
of Indiana, Department of Psychology, Speech Research Laboratory.
Maassen, B., Groenen, P., Crul, T., Assman-Hulsmans, C., & Gabreëls, F.
(2001). Identification and discrimination of voicing and place-of-
articulation in developmental dyslexia. Clinical Linguistics and
Phonetics, 15, 319-339.
Manis, F. R., Custodio, R., & Szeszulski, P. A. (1993). Development of
phonological and orthographic skill: A 2-year longitudinal study of
dyslexic children. Journal of Experimental Child Psychology, 56, 64-
86.
Manis, F.R. & Keating, P. (2005). In Catts, H.W & A.G. Kamhi (Eds). The
connections between language and reading disabilities. Mahwah, NJ:
Erlbaum.
Manis, F. R., McBride-Chang, C., Seidenberg, M. S., Keating, P., Doi, L. M.,
Munson, B., & Petersen, A. (1997). Are speech perception deficits
associated with developmental dyslexia? Journal of Experimental
Child Psychology, 66, 211-235.
Manis, F.R., Seidenberg, M.S., Doi, L.M., McBride-Chang, C., & Petersen, A.
(1996). On the basis of two subtypes of developmental dyslexia.
Cognition, 58, 157-195.
Manis, F.R., Seidenberg, M.S., Stallings, L., Joanisse, M.F., Bailey, C.E.,
Freedman, L.B., & Curtin, S. (1999). Development of dyslexic
subgroups: A one year follow up. Annals of Dyslexia, 49, 105-136.
McBride-Chang, C. (1996). Models of speech perception and phonological
processing in reading. Child Development, 67, 1836-1856.
McDougall, S., Hulme, C., Ellis, A., & Monk, A. (1994). Learning to read:
The role of short-term memory and phonological skills. Journal of
Experimental Child Psychology, 58, 112-133.
48
Metsala, J.L. (1997a). An examination of word frequency and neighborhood
density in the development of spoken word recognition. Memory and
Cognition, 25, 47-56.
Metsala, J.L. (1997b). Spoken word recognition in reading disabled children.
Journal of Educational Psychology, 89, 159-169.
Metsala, J. L., & Walley, A. C. (1997). Spoken vocabulary growth and the
segmental restructuring of lexical representations: Precursors to
phonemic awareness and early reading ability. In J. L. Metsala & L. C.
Ehri (Eds.), Word recognition in beginning literacy (pp. 89-120).
Mahwah, NJ: Erlbaum.
Montgomery, J.W. (1999). Recognition of gated words by children with
specific language impairment: an examination of lexical mapping.
Journal of Language and Hearing Research, 43, 735-743.
Morais, J. (2003) Levels of phonological representation in skilled reading and
in learning to read. Reading and Writing: An Interdisciplinary Journal,
16, 123-151.
Morais, J., Cary, L., Alegria, J., & Bertelson, P. (1979). Does awareness of
speech as a sequence of phones arise spontaneously? Cognition, 24,
323-331.
Pennington, B. F., Van Orden, G. C., Smith, S. D., Green, P. A., & Haith, M.
M. (1990). Phonological processing skills and deficits in adult
dyslexics. Child Development, 61, 175301758.
Perfetti, C. A., Beck, I., Bell, L. C., & Hughes, C. (1987). Phonemic
knowledge and learning to read are reciprocal: A longitudinal study of
first grade children. Merrill-Palmer Quarterly, 33, 283-319.
Pratt, A.C. & Brady, S. (1988). Relation of phonological awareness to reading
disability in children and adults. Journal of Educational Psychology,
80, 319-323.
Rack, J. P., Snowling, M. J., & Olson, R. K. (1992). The nonword reading
deficit in developmental dyslexia: A review. Reading Research
Quarterly, 27, 29-53.
Reed, N. (1989). Speech perception and discrimination of brief auditory cues
in reading disabled children. Journal of Experimental Child
Psychology, 48, 270-292.
49
Salasoo, A., & Pisoni, D. B. (1985). Interaction of knowledge sources in
spoken word identification. Journal of Memory & Language, 24, 210-
231.
Semel, E., Wiig, E.H., & Secord, W.A. (1995). Clinical Evaluation of
Language Fundamentals-Revised. San Antonio, TX: The Psychological
Corporation.
Serniclaes, W., Sprenger-Charolles, L., Carré, R., & Demonet, J.-F. (2001).
Perceptual discrimination of speech sounds in developmental dyslexia.
Journal of Speech, Language and Hearing Research, 44, 384-399.
Share, D. L. (1995). Phonological recoding and self-teaching: Sine qua non of
reading acquisition. Cognition, 55, 151-218.
Shaywitz S.E., Shaywitz B.A. (2003). Dyslexia (specific reading disability).
Pediatrics In Review, 24, 147-153.
Snowling, M. J. (2000). Dyslexia: A cognitive developmental perspective, 2
nd
ed. Oxford: Blackwell.
Snowling, M. & Hulme, C. (1989). A longitudinal study of developmental
phonological dyslexia. Cognitive Neuropsychology, 6, 653-659.
Stanovich, K. E., & Siegel, L. S. (1994). The phenotypic performance profile
of reading-disabled children: A regression-based test of the
phonological-core-variable-difference model. Journal of Educational
Psychology, 86, 24-53.
Stanovich, K.E., Siegel, L.S., & Gottardo, A. (1997). Converging evidence for
phonological and surface subtypes of reading disability. Journal of
Educational Psychology, 89, 114-128.
Swan, D & Goswami, U. (1997) Phonological awareness deficits in
developmental dyslexia and the phonological representations
hypothesis. Journal of Experimental Child Psychology 66, 18-41.
Tallal, P., & Stark, R. E. (1982). Perceptual/motor profiles of reading impaired
children with or without concomitant oral language deficits. Annals of
Dyslexia, 32, 163-176.
Torgesen, J.K., Wagner, R.K., & Rashotte, C.A. (1999). Test of Word Reading
Efficiency. Austin, TX: Pro-Ed.
50
Tyler, L. K., & Wessels, J. (1985). Is gating an on-line task? Evidence from
naming latency data. Perception & Psychophysics, 34, 409-420.
Wagner, R.K., Torgesen, J.K., & Rashotte, C.A. (1999). Comprehensive Test
of Phonological Processing. Austin, TX: Pro-Ed.
Wagner, R.K. & Torgesen, J.K. (1987). The nature of phonological processing
and its causal role in the acquisition of reading skills. Psychological
Bulletin, 101, 192-212.
Walley, A. C. (1988). Spoken word recognition by children and adults.
Cognitive Development, 3, 137-165.
Walley, A. C. (1993). The role of vocabulary development in children’s
spoken word recognition and segmentation ability. Developmental
Review, 13, 286-350.
Walley, A.C., Michela, V.L., & Wood, D.R. (1995). The gating paradigm:
Effects of presentation format on spoken word recognition by children
and adults. Perception and Psychophysics, 57, 343-351.
Warren, P., & Marslen-Wilson, W. (1987). Continuous uptake of acoustic cues
in spoken word recognition. Perception & Psychophysics, 41, 262-275.
Wechsler, D. A. (1992). Wechsler Intelligence Scale for Children III. San
Antonio, TX: Psychological Corporation.
Welch, B.L. (1938). The significance of the difference between two means
when the population variances are unequal. Biometrika, 29, 350-362.
Werker, J., & Tees, R. (1987). Speech perception in severely disabled and
average reading children. Canadian Journal of Psychology, 41, 48-61.
Wesseling. R., & Reitsma. P. (2001). Preschool phonological representations
and development of reading skills. Annals of Dyslexia, 51, 203-229.
West, P. (1999). Perception of distributed coarticulatory properties of English
/l/ and /r/. Journal of Phonetics, 27, 405-426.
Wilcox, R.R. (2003). Applying Contemporary Statistical Techniques. Boston,
MA: Elsevier Science.
51
Wolf, M. & Bowers, G. (1999). The double-deficit hypothesis for the
developmental dyslexias. Journal of Educational Psychology, 91, 415-
438.
Woodcock, R.W. (1987). Woodcock Reading Mastery Test-Revised. Circle
Pines, MN: American Guidance Service.
Woodcock, R.W. (1998). Woodcock Reading Mastery Tests-Revised-
Normative Update. Circle Pines, MN: American Guidance Service.
Yuen, K.K. (1974). The two-sample trimmed t for unequal population
variances. Biometrika, 61, 165-170.
Ziegler, J. C., Goswami, U. C. (2005). Reading acquisition, developmental
dyslexia and skilled reading across languages: A psycholinguistic grain
size theory. Psychological Bulletin, 131, 3-29.
52
APPENDIX
Words used in Gating task
Practice Gate Duration (ms)
items
gate
1
gate
2
gate
3
gate
4
gate
5
gate
6
cheat 228.6 251.3 278.5 373.7 395.4 527
fail 255.4 285 309.9 505.6 527.2 668
pain 106.9 131.8 156.8 390.8 415.8 535
Gate
Duration Mean Scores
Items (ms) Dyslexics Controls
gate
1
gate
2
gate
3
gate
4
gate
5
gate
6 FCID FM FCID FM
boat 31.76 56.62 81.46 253.5 278.4 466 3.04 3.96 3.17 4.35
bone 34.84 59.08 84.08 331.4 356.7 535 4.13 4.52 3.96 4.48
bowl 42.4 67.5 92.24 310 334.9 522 3.74 4.91 3.43 4.78
cat 107.7 132.3 157.4 321.2 344.7 451 2.30 2.91 2.26 2.52
can 121.5 146.2 171.1 409 434.3 624 3.83 4.22 3.52 4.17
code 100.6 125.4 150.3 406.9 431.3 532 2.91 4.70 2.78 4.39
cone 100.8 125.9 150.3 382.5 407.1 569 4.22 4.65 3.91 4.26
coal 100.6 125.6 150.4 295 319.9 491 3.43 5.22 2.57 5.43
cloud 146.1 170.9 196.2 467.6 491.5 576 2.74 4.26 2.78 4.09
clown 120.2 145.1 159.5 404.4 428.7 566 5.04 5.13 3.96 4.83
dot 34.95 59.89 84.7 231.8 256.9 390 3.00 5.09 2.39 5.09
dawn 34 59.12 84.2 366.3 391.8 573 4.96 5.78 4.13 5.22
feet 331.4 356.9 382.1 494.3 519.1 567 3.65 4.52 3.52 4.26
feel 257.6 283.2 308.7 484.7 510.3 649 3.48 4.04 3.30 3.70
pad 71.98 96.78 121.6 424 449.1 547 2.57 5.39 2.26 5.52
pan 41.24 65.37 91.44 379.8 403.9 558 2.96 4.48 2.87 3.57
rag* 274.8 274.8 299.6 627.6 652.6 612 2.57 4.91 3.00 4.87
rang 104.6 129 154.1 443.1 466.6 609 4.26 5.87 3.83 5.57
seat 269.4 294.2 318.5 400.8 426.1 499 3.57 4.74 3.61 4.52
scene 230 254.8 279.7 499.9 523.6 689 4.26 4.78 3.70 4.43
sweat 369.7 394.7 419.9 472.1 497.5 625 2.39 3.30 2.30 3.26
swell 330.8 356.3 380.6 507 509.4 692 3.78 4.09 3.83 4.13
* NOTE: for “rag” gate 1 and 2 have same duration due to immeasuably
short duration of initial /r/ sound
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Dyslexic's development of word recognition and phonological skill over time
PDF
Insights into the nature of phonological and surface dyslexia: Evidence from a novel word learning task
PDF
A longitudinal analysis of word decoding and reading comprehension in a bilingual population
PDF
Ethnic identity in Parsee teenagers
PDF
Depression in offspring following severe prenatal stress
PDF
Adolescents' social attitudes: Genes and culture?
PDF
Distributional cues and subject identification in the production of subject -verb agreement
PDF
Helmets: The safe choice. A study on the benefits of helmet safety while skiing and/or snowboarding: A public relations program to encourage helmet use
PDF
Ethnic identity, acculturation, self-esteem and perceived discrimination: A comparison study of Asian American adolescents
PDF
Emergent literacy differences in Latino and African American children: Culture or poverty?
PDF
Does first-year treatment intensity predict outcome in young autistic children receiving Lovaas' ABA intervention?
PDF
Determining competent/resilient outcomes across multiple domains in sexually abused and nonabused females
PDF
Intragroup evaluations, attitude source, and in-group member derogation
PDF
Effects of threat and self-focus on consensual bias in majorities
PDF
Affection and conflict in family relationships
PDF
Effects of personal resource sufficiency on perceived difficulty and desirability of earthquake preparedness
PDF
Bilingual education in California: A policy debate
PDF
History of depression, antidepressant treatment, and other psychiatric illness as risk factors for Alzheimer's disease in a twin sample
PDF
Effects of parenting style and ethnic identity on European American and Asian Indian adolescents' academic competence and self esteem
PDF
Aging and the use of context and frequency information in ambiguity resolution
Asset Metadata
Creator
Bruno, Jennifer Lynn
(author)
Core Title
Auditory word identification in dyslexic and normally achieving readers
School
Graduate School
Degree
Master of Arts
Degree Program
Psychology
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
language, linguistics,OAI-PMH Harvest,psychology, developmental
Language
English
Contributor
Digitized by ProQuest
(provenance)
Advisor
Manis, Franklin (
committee member
), Nakamoto, Jonathan (
committee member
)
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c16-50637
Unique identifier
UC11338248
Identifier
1435120.pdf (filename),usctheses-c16-50637 (legacy record id)
Legacy Identifier
1435120.pdf
Dmrecord
50637
Document Type
Thesis
Rights
Bruno, Jennifer Lynn
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Access Conditions
The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...
Repository Name
University of Southern California Digital Library
Repository Location
USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA
Tags
language, linguistics
psychology, developmental