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
/
Insights into the nature of phonological and surface dyslexia: Evidence from a novel word learning task
(USC Thesis Other)
Insights into the nature of phonological and surface dyslexia: Evidence from a novel word learning task
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UM I films
the text directly from the original or copy submitted. Thus, some thesis and
dissertation copies are in typewriter face, while others may be from any type of
computer printer.
The quality of this reproduction is dependent upon the quality of the
copy submitted. Broken or indistinct print, colored or poor quality illustrations
and photographs, print bleedthrough, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send U M I a complete manuscript
and there are missing pages, these w ill be noted. Also, if unauthorized
copyright material had to be removed, a note w ill indicate the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand comer and continuing
from left to right in equal sections with small overlaps.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6" x 9* black and white
photographic prints are available for any photographs or illustrations appearing
in this copy for an additional charge. Contact UMI directly to order.
ProQuest Information and Learning
300 North Zeeb Road. Ann Arbor. M l 48106-1346 USA
800-521-0600
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
INSIGHTS INTO THE NATURE OF PHONOLOGICAL AND SURFACE DYSLEXIA:
EVIDENCE FROM A NOVEL WORD LEARNING TASK
Copyright 1999
by
Caroline Elizabeth Bailey
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 1999
Caroline Elizabeth Bailey
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 1409616
____ _ _ < 0
UMI
UMI Microform 1409616
Copyright 2002 by ProQuest Information and Learning Company.
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, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UNIVERSITY O F SOUTHERN CALIFORNIA
THC CRAOUATC SCHOOL
UNIVERSITY PARK
LOS ANGELAS. CALIFORNIA MOOT
This thesis, written by
.Car.Qli.me-E.. bailey..
under the direction of h Thesis Committee,
and approved by all its members, has been pre
sented to and accepted by the Dean of The
Graduate School, in partial fulfillment of the
requirements for the degree of
Master o f Arts
THESIS COMMITTEE
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1 1
Table of Contents
Table 1 ............................................................................................................................page 40
Table 2 ............................................................................................................................page 41
Figure 1 page 42
Figure 2 page 43
Figure 3 page 44
Appendix A ...................................................................................................................page 47
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
iii
List o f Tables
Table 1. Means. Standard Deviations, and Group Differences o f all
eligible dyslexics, reading level matched controls, and age matched controls........... page
40
Table 2. Means, Standard Deviations, and Group Differences o f
phonological dyslexics, surface dyslexics, and reading level matched controls. ...page 41
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
iv
List o f Figures
Figure 1. The effect o f subgroup, word type, and day (includes both
learning trials) on the percentage of words correct....................................................... page 42
Figure 2. The effect o f subgroup and word type on the percentage of
words correct page 43
Figure 3. The effect o f subgroup and word type on the percentage of
words correct for the computer task page 44
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Abstract
The current study investigates cognitive correlates of surface and phonological dyslexia
using a novel nonsense word learning task. Specific word learning profiles as well as
differences in rate o f word learning are assessed using a training study design and a
delayed recall task. Results indicate that while phonological dyslexics exhibit a
distinctive developmental profile, surface dyslexics have a learning rate similar to that of
reading level matched controls. Additionally, nonverbal and verbal paired associate
learning as well as print exposure are assessed. The implications o f these findings for
current etiological theories of surface dyslexia are discussed. A core orthographic
processing deficit is implicated in the depressed performance o f surface dyslexics on the
word learning task. Connectionist interpretations for this finding are discussed.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1
Insights Into the Nature of Phonological and Surface Dyslexia:
Evidence From a Novel Word Learning Task
Developmental dyslexics are a heterogeneous group as they differ not only in
overall reading ability and amount o f reading experience but also in their specific reading
profiles (Castles & Coltheart, 1993; Manis, Seidenberg, Doi, McBride-Chang &
Petersen., 1996; Share, 1995; Stanovich, Siegel & Gottardo, 1997). Recent
investigations have identified two distinct reading profiles defined on the basis o f specific
word reading skills that subdivide dyslexics into two distinct subgroups o f phonological
and surface dyslexics. Respectively, these subgroups serve as a means o f delineating
dyslexics who experience a marked difficulty with phonological skill from those who
demonstrate a marked impairment in exception word reading skill (Castles & Coltheart,
1993; Castles & Coltheart, 1996; Castles, Datta, Gayan, & Olson, 1999; Joanisse,
Manis, Keating & Seidenberg, in press; Manis et al., 1996; Manis et al, in press;
Stanovich et al., 1997).
Case studies identifying individuals of average intellectual functioning
demonstrating pure surface and phonological dyslexic profiles have been instrumental in
defining the specific features of each subgroup classification (Coltheart et al, 1981;
Campbell & Butterworth, 1985; Snowling & Hulme, 1989; .Goulandris & Snowling,
1991; Hanley, Hastie & Kay, 1992; Castles & Coltheart, 1996). These studies assert that
an individual with a pure phonological dyslexic profile exhibits a marked impairment in
the ability to effectively implement phonological reading strategies when reading words
and nonsense words that utilize phonologically regular grapheme to phoneme
correspondences. In contrast, exception word reading falls within the normal range.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Conversely, a pure surface dyslexic case exhibits the opposite profile therefore
demonstrating normal reading o f phonologically regular words and nonsense words and
impaired exception word reading (Coltheart et al, 1981; Campbell & Butterworth, 1985;
Snowling & Hulme, 1989; .Goulandris & Snowling, 1991; Hanley, Hastie & Kay, 1992;
Castles & Coltheart, 1996).
Larger group studies o f surface and phonological dyslexia have validated these
subgroup classifications of poor readers while expanding upon the case study findings in
key ways (Castles & Coltheart, 1993; Manis et al., 1996; Stanovich et al., 1997). Using a
regression based method in which nonsense word reading was regressed on exception
word reading, Castles & Coltheart (1993) identified 10 pure surface dyslexics and 8 pure
phonological dyslexics from a sample of 53 poor readers. However, Castles and
Coltheart suggested that in comparison to the age-matched comparison group, the
majority of the dyslexics appeared to be impaired to some degree on both nonsense word
and exception word reading such that the subgroups demonstrated a greater impairment in
one type of reading task than the other in the predicted direction. Further analyses using
this criterion o f discrepant degrees of impairment on the reading tasks yielded 16 surface
dyslexics and 29 phonological dyslexics. While it appears that these subgroups of
dyslexics exhibit profiles that indicate some degree o f deficiency in both types o f reading
tasks, the authors interpreted these findings using the dual route model o f reading
(Coltheart, 1978) and thus attributed the observed deficiencies in phonological skill and
exception word reading skill among phonological dyslexics to a specific impairment in
the sub-lexical route, which is responsible for utilizing grapheme-phoneme conversion
rules. Hence, any concurrent difficulties in exception word reading are conceptualized as
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3
complications arising from a primary impairment along this sub-lexical route. Similarly,
difficulties in both whole word recognition and phonological skill observed within the
surface group were attributed to core deficiencies along the lexical route responsible for
the storage and retrieval o f whole word pronunciations.
Manis et al. (1996), provide some convergent validity for the distinct phonological
and surface dyslexic subgroup classifications. Using a similar regression based
subgrouping method the authors validated these subgrouping distinctions not only on the
basis of measures of nonsense word reading and exception word reading but also with
additional measures of phonological awareness and orthographic skill. Additionally,
Manis et al.(1996) used both same aged normal readers and younger normal readers as
comparison groups for the dyslexics. However, Manis et al.(1996) use the connectionist
model of reading (Seidenberg & McClelland, 1989) to explain these subgroup
distinctions and posit that these observed impairments in both nonsense word and
exception word reading can be attributed to disruptions occurring along a single reading
route.
Manis et al. (1996) suggest that while phonological dyslexics have a distinct
developmental profile, surface dyslexics appear to perform much like younger normal
readers across a wide array of measures including phonological awareness and
orthographic skill. Using a younger sample than that used in Manis et al., Stanovich et al
(1997) bring forward similar findings validating the subgroup distinctions and supporting
the notion that surface dyslexics behave much like younger normal readers across a wide
spectrum o f cognitive tasks. In addition, both Manis et al. (1996) and Stanovich et al.
(1997) assert that most cases of phonological and surface dyslexia do not fit the pure
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4
subgroup classification. Instead, Stanovich et al. (1997) suggest that surface and
phonological dyslexia are better described using the construct of "soft” subgroups as
opposed to the pure case classifications. This construct of soft subgroups reinforces the
notion of relative impairments in exception word and nonsense word reading skill and
suggests that most dyslexics are impaired more severely along one dimension o f reading
skill than the other.
Several investigators have posited hypotheses exploring the etiologies o f the
cognitive correlates o f surface and phonological dyslexia (Boder, 1973; Petrauskas &
Rourke, 1979; Statz & Morris, 1981; Newby, Recht & Caldwell, 1993; Castles &
Coltheart, 1993; Manis et al., 1996; Castles & Holmes, 1996; Stanovich, Siegel &
Gottardo, 1997; Castles, Datta, Gayan & Olson, 1999: Joanisse et al., in press; Manis et
al., in press). While there appears to be a general consensus indicating that phonological
deficits are responsible for the reading difficulties found in phonological dyslexics, the
core component of surface dyslexia remains debated (Castles & Coltheart, 1993; Castles
& Holmes, 1996; Castles, et al., 1999; Harm & Seidenberg, 1999; Joanisse et al., in
press; Manis et al., 1996; Manis et al., in press; Stanovich, et al., 1997).
Stanovich et al. (1997) suggest surface dyslexia stems from a mild phonological
deficit in combination with environmental factors including print exposure deficits and
limited exposure to learning opportunities. While recent work suggests that surface
dyslexics may have mild phonological delays (Harm & Seidenberg, 1999; Joanisse et al.,
in press; Manis et al,. 1996; Manis et al., in press), empirical tests of the Stanovich et
al.(1997) hypothesis regarding print exposure have been attempted with inconsistent
results (Castles et al., 1999; Manis et al., in press). While Castles et al. found marginally
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
5
significant differences in print exposure (as assessed by a Title Recognition Task)
between surface and phonological dyslexic subgroups, the authors posit that surface
dyslexics should be markedly lower in print exposure compared to phonological dyslexics
and further suggest that these marginal findings do not sufficiently explain the surface
dyslexic profile. Additionally, Manis et al. (in press) did not find significant group
differences on print exposure among the dyslexic subgroups. Thus, the Stanovich et al.
account of surface dyslexia is somew hat limited as print exposure findings are nebulous.
Furthermore, the soft subtype construct does not fully account for the existence of pure
cases as do alternative hypotheses.
Castles and Coltheart (1993) explain the differences in phonological and surface
dyslexic reading profiles from the perspective o f the dual route model (Coltheart 1978)
which suggests two independent routes, the lexical and sub-lexical, are implemented in
the reading process. Hence, pure cases of phonological dyslexia are characterized by
impairments in the sub-lexical route, the route responsible for encoding and applying the
translation rules used in learning and using the symbol to sound correspondences o f the
alphabetic phonic system, while surface dyslexics are characterized by impairments in the
lexical route which is used in developing a sight word vocabulary and responsible for
both storing and retrieving orthographic representations of "whole" words as individual
units. While the dual route model provides a clear account of pure cases, it does not
account for the soft subgroup profile as it does not offer any predictions or explanations
of the process which occurs when both nonsense word reading and exception word
reading are impaired at the same time as they would be in the soft subgroup dyslexic
profile. Additionally, the dual route model suggests that difficulties in phonological
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6
processing stem from difficulty applying specific grapheme-phoneme conversion rules
and does not take into account the findings which suggest an inherent phonological deficit
manifests itself in both reading and non-reading tasks (Manis et al., 1996; Harm &
Seidenberg, 1999, Wagner & Torgesen, 1987).
The Seidenberg and McClelland (1989) connectionist model of reading proposes
and models using computer simulation a single route method o f normal word reading in
which phonological (sub-lexical) and orthographic (lexical) information is processed
along the same route. The most recent version o f this model (Harm & Seidenberg, 1999)
applies this single route model to the developmental reading acquisition process. In
essence the model "learns" words through a series of repeated exposures to a corpus o f
one syllable words in which it establishes weighted connections between word types and
pronunciations by manipulating the number hidden units associated with each connection.
These hidden units determine the extent to which the model generalizes pronunciation
rules to specific word units. For example, hidden units enable the model to simulate
word-specific learning in the sense that the more hidden units the model has available to
use and apply to different connections, the more able the model is establish unique
connections for individual words. In this sense, the quality o f the model’s output can be
adjusted by manipulating the quality of the input.
The Harm and Seidenberg model differs from the dual route model in several
important ways. The first o f these being a difference in interpretation o f the source o f the
phonological impairments characteristic of the phonological dyslexic group. While the
dual route model attributes phonological difficulties to an inability to manipulate
grapheme-phoneme conversion rules, Harm & Seidenberg (1999) adhere to the notion of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
7
a core phonological deficit (Lieberman & Shankweiler, 1985, Wagner & Torgesen, 1987)
suggesting that an inherent underlying phonological deficit evidenced in both reading and
non-reading phonological tasks is the source o f the phonological difficulties demonstrated
by phonological dyslexics. Additionally, this model poses several possible hypotheses
which may account for developmental surface dyslexia.
The first of these hypotheses proposes a general rate of learning deficit which
suggests that surface dyslexics are simply delayed in learning on a general cognitive level.
This general cognitive delay yields an individual with a relatively intact core phonological
system and a delayed sight word vocabulary. This would then explain a delay in
exception word reading. Another possible account o f surface dyslexia posited in this
model suggests surface dyslexia may result from a lack o f experience with reading. In
this account, exposure to exception word stimuli is deficient due to limited learning
opportunities including limited exposure to print and inappropriate curriculum or
instruction. Additionally, the model asserts that impaired exception word reading may
result from an input deficiency. This input deficiency can take several forms. The first
is a hidden unit deficiency which results in degraded connections between input, storage,
and retrieval o f exception words. This would manifest an arbitrary word-specific reading
deficit that would greatly effect exception word reading. Similarly, Harm and Seidenberg
(1999) suggest that surface dyslexia can be simulated by degrading the quality o f the
orthographic units. The behavioral analog of this process in the model would be an
underlying visual deficit or an underlying orthographic processing deficit which manifests
itself in a failure to develop adequate orthographic representations of words. Hence, this
model can account for pure, mixed and soft subgroup classifications o f phonological and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
8
surface dyslexia. Additionally, the Harm and Seidenberg model generates testable
hypotheses regarding the etiology of both subgroups.
It is clear that the predictions from these competing theoretical accounts of surface
dyslexia need further empirical testing. Castles and Holmes (1996) began this process in
a training study looking at differences in the ability of poor lexical readers and poor sub-
lexical readers in learning exceptional and strange pronunciations o f novel, nonsense
word stimuli. By using nonsense words, this study was able to control for aspects of print
exposure (most importantly familiarity with the stimuli) while assessing the role of
specific lexical acquisition skills in both lexical and sub-lexical readers. The stimuli were
designed such that appropriate pronunciations o f the items did not correspond with the
phonologically regular pronunciation o f the particular orthographic string. For example
* ‘macht” was pronounced "mot”. Additionally half of these items were derived from
exception words while the other half were derived from strange words. The intent behind
this was to render the application of grapheme-phoneme conversion rules ineffective in
producing the correct pronunciation of the items. The children learned these items over
the course of four 15 minute training sessions with 3 or 4 days in between sessions.
Subjects were tested using a reading aloud task as well as an orthographic choice task.
While Castles and Holmes(1996) did not find significant group differences for
accuracy on the reading aloud task, they did find a significant main effect for word type
such that the exception words were pronounced correctly more often than the strange
words for both poor lexical and poor sub-lexical readers. Additionally, while accuracy
when reading exceptional items did not differ, Castles and Holmes found that poor lexical
readers had significantly more difficulty reading the strange items correctly than poor
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
9
sub-lexical readers. However, on the orthographic choice task, Castles and Holmes did
find group differences in overall accuracy of word recognition such that the poor lexical
readers demonstrated significant difficulty producing correct responses, yet, a main effect
for word type was not found.
The current study expands upon Castles and Holmes (1996) in several important
ways. First, this study strives to differentiate surface and phonological dyslexic
subgroups from a large group of non-clinic referred poor readers using a novel nonsense
word learning task modeled after Castles and Holmes (1996). Furthermore, it is designed
to yield important information about the ways in which surface and phonological
dyslexics learn different word types. The stimuli for this study differ form those used in
Castles and Holmes (1996) in that these stimuli consist o f items that have both
phonologically regular and irregular pronunciations. As Castles and Holmes (1996) did
not use phonologically regular nonsense word items, it is important to determine the full
scope of the reading context in which surface dyslexics exhibit the pattern found in
Castles and Holmes (1996). It is equally important to investigate whether or not
phonological dyslexics exhibit a similar profile when learning phonologically regular
nonsense word items. Also o f interest is the idea o f learning rate for words which this
study examines and Castles and Holmes (1996) did not. Additionally, this study has
included measures o f nonverbal and verbal paired associate learning rate, print exposure,
and the child's reading habits to test hypothesis offered by Harm and Seidenberg (1999)
as well as Stanovich et al. (1997). Furthermore, the current study has a delayed word
recognition task and a spelling task which are used to measure the outcome o f the
training.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
10
In the current study we posit 2 distinct hypotheses. First, we expect an interaction
between word type and subgroup such that phonological dyslexics should outperform
surface dyslexics on irregular words while this pattern should be reversed for regular
words. Second, we expect that the surface dyslexic subgroup will demonstrate a profile
similar to the reading level comparison group, in terms o f learning rate, although at a
lower rate o f proficiency. In contrast, the phonological dyslexic subgroup should
manifest a distinct developmental profile.
Method
Participants
The participants were recruited from nine public elementary schools in Long
Beach, California. The participants consist o f fourth and fifth grade poor readers (n=71),
fifth grade normal readers (n=31) and first, second and third grade younger normal
readers (n=21). The demographics of this sample are representative of the population of
the Long Beach Unified School District in that the sample is composed of an ethnically
diverse, lower middle to middle class group o f children.
Initially, potential participants were recruited through teacher/principal
nomination. Teachers were asked to nominate potential participants who met the
following general criteria: fluency in English (as evaluated by the school's speech and
language therapist), a developmental history free o f neurological impairment or other
developmental disorders including visual impairments beyond that which can be
corrected with eye glasses, and normal hearing. After meeting these general inclusion
criteria, teachers were then asked to nominate children for the specific groups of interest.
Poor Readers
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
11
This group contains fourth and fifth grade students with a mean age of 124.65
months (sd = 6.45). Teachers were asked to nominate poor readers who fell within the
bottom 25% of reading ability. In addition to teacher nomination, subjects also had to
meet the additional inclusionary criteria. In order to be included in the poor reader group,
subjects were required to score at or below the 25th percentile on The Woodcock Word
Identification Test (Woodcock, 1987) and obtain a standard score of at least 85 on either
the Peabody Picture Vocabulary Test or the Visual Closure subtest o f the Woodcock Test
o f Cognitive Ability. Children were excluded from the analyses if they received a
standard score o f less than 60 on either of these measures as a means o f insuring that the
general cognitive ability of the poor reader group fell within normal range.
Reading Level Matched Comparison Group
This group contains first, second, and third grade students with a mean age of 91.1
months (sd = 9.47). All members of the reading level matched comparison group were
initially recruited during the Spring of their first grade year. Before they were chosen for
the reading level matched comparison group, potential participants were screened on the
Woodcock Word Identification test to insure that their scores fell between the 40th and
97th percentile on word recognition. This range ensured that no members of this group
performed at a level above two standard deviations above the mean. Those scoring
outside of this range were not selected for the reading level matched comparison group.
In addition to meeting the screening criteria, subjects were held to the same inclusionary
criteria as the poor readers in terms of the measures o f cognitive ability.
Age Matched Comparison Group
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
12
This group contains fifth grade students with a mean age o f 125.06 months (sd =
4.18). All members o f the same aged normal reader group scored at or above the 40th
percentile on the Woodcock Word Identification test. Similarly to the poor readers and
reading level matched comparison group, members in this group were required to obtain a
standard score of at least 85 on either the Peabody Picture Vocabulary Test or the Visual
Closure subtest o f the Woodcock Test o f Cognitive Ability. Additionally, children were
excluded from this group if they received a standard score below 60 on either of these
measures.
Procedure
Subgrouping Variables
In order to distinguish the surface and phonological dyslexic groups from the
overall reading impaired group, participants were assessed on measures o f nonsense word
reading and exception word reading. The nonsense word reading task consisted of 70
phonetic nonsense words. This list begins with items of simple CVC (consonant-vowel-
consonant) structured words C‘nug”) and also includes items which contain two or more
letter clusters and some two syllable items. The items were ordered so that the difficulty
increased sequentially based on the findings o f pilot data. Similarly, the exception word
reading task consisted of a list of 70 exception words. This list was ordered from easiest
to hardest based on frequency and grade norms (Adams & Huggins, 1986). On both of
these tasks, administration of the items was discontinued when the subject failed to read 6
consecutive items correctly at the end o f a stimulus page. The internal consistency
reliability for both the nonsense word reading task and exception word reading task was
.96 (Cronbach's alpha).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
13
Phoneme Deletion Task
A two part measure o f phoneme deletion similar to the type used by Bruce (1984)
was administered. For the first part of the task, subjects were required to delete a
phoneme from an existing word and then recombine the remaining phonemes into a new
phonological representation of an existing English word.. For example, a participant is
asked to say "split" without the "p" sound and arrive at "slit". For the second part of this
task, subjects were required to repeat a nonsense word, delete a specified phoneme and
blend the remaining sounds into a new nonsense word. For example, a correct subject
trial would involve saying the nonsense word "kimp" without the "m" and produce the
nonsense word "kip". The test items presented on audio tape. Subjects could request to
have an individual test item repeated once. Feedback intensive practice items were given,
however, no feedback was given on the test items. The task consisted of 25 word items
and 15 nonsense word items. Testing was discontinued after 5 consecutive incorrect
responses. Internal consistency reliability for this measure was .84 (Cronbach's alpha)
Orthographic Choice Task
The orthographic choice task (adapted from Olson et al., 1989) required the
participant to view two strings of letters displayed on opposite sides o f a computer screen
and choose the item that is a correctly spelled word. The subject pressed a button on a
button box to indicate the side of the screen that contained the correct answer. Half of the
items contained at least one exception word (e.g., sponge spunge) and half contained two
regular words (sheep sheap). All of the foil items were phonologically identical to the
targeted exemplar item. Hence the child could not rely on phonological reading strategies
to choose the correct answer. Both paper and computer practice items were administered
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
14
with corrective feedback on order to make sure the child understood the task. Forty eight
test items were administered. Cronbach’s alpha for accuracy scores was equal to .75.
Paired Associate Learning Tasks
In order to investigate the issue of learning rate and general paired associate
learning ability, participants were given the verbal paired associate learning and the
nonverbal paired associate learning subtests of the Weschler Memory Scale (Weschler,
1990). The accompanying recall measures were also given. The verbal paired associate
task was administered exactly as it was designed, however amendments were made to the
nonverbal paired associate learning task in order to minimize potential practice effects
originating from high familiarity with this specific type o f task. Due to the notion that
many common children's games have a high visual paired associate learning component,
we incorporated the existing practice items into the test battery and added two new
practice items to this test in order to increase the difficulty level as we believe that
children have more exposure and training on this type o f task than adults. The order in
which the items appeared was determined at random. Even though the content of the
nonverbal paired associate test was slightly altered, the standard procedures for the
administration o f both o f the subtests were followed.
Title Recognition Task
Subjects were shown a list of 45 book titles, 30 real book titles and 15 foils, on a
sheet o f paper. Pilot data was collected to ensure that these book titles were appropriate
for the sample. The experimenter instructed the child that s/he was going to be asked to
decide which o f the book titles were titles of “real” books and which ones were not. The
experimenter read the book titles aloud to the child one at a time. The subject was then
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
15
asked to decide whether the book title they just heard was a real book or not. Subjects
were encouraged not to guess and to only give affirmative answers if the knew that the
book title was the title o f a real book. The total score on the task was the proportion of
correct book titles chosen by the child minus the proportion o f incorrect titles chosen by
the child. This score corrects for guessing in the manner devised by Cunningham &
Stanovich (1990). The reliability for correct answers was .82 (Cronbach’s alpha).
Letter Matching Task
This task (based on Bigsby, 1988) required children to look at a string of five
letters on a computer screen and decide if that letter string contained two letters that had
the same name. Children recorded their answers by pressing either the “yes” or “no”
button on a button box. On some trials the matching letters appeared in the same case
(AcfAg) and in others they were in alternating case (gtaVA). This required children to
process letters to the point of recognition, rather than making comparisons strictly on the
basis of low-level physical features. Cronbach’s alpha for the accuracy scores was .86.
Teacher Questionnaire
To assess whether subgroup differences could be explained by differences in the
reading habits of participants as assessed by their teachers, a questionnaire was given to
the home-room teacher of each participant. Among the items on this questionnaire, the
teacher was asked the following questions: (1) To your knowledge, how often has this
student read at home for at least 15 minutes during the last 4 weeks? (2) Does the student
do extra-curricular reading? (3) Does the student bring in free reading books from home?
(4) During silent reading, does the student regularly have new or different books? (5)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
16
Does the student comment about book he/she read at home? and (6) Is the student
enrolled in any extra-curricular or home reading clubs through the school?
Word Learning Task
Subjects received training on 22 nonsense words over the course o f 2 separate
days o f training.. Subjects received a total of 6 training trials dived equally over the 2
training days within each session lasting 10-15 minutes. There was a three to four day
delay between training days depending on whether a weekend intervened. After
completing the training, subjects had two testing days. A spelling test was administered
on the first day and a computerized recognition test on the second. Again, the delay
between testing days was either three or four days depending on whether a weekend
intervened.
On the first day o f training, subjects were told that they were going to leam how
to read the names of 22 space aliens. At this time the children were shown a hand drawn
picture of a group of space aliens in order to maintain their interest in the task.
Subsequently, the subjects received the first training trial on the nonsense word stimuli.
The stimuli consist of 22 one syllable nonsense words ranging between five and
six letters in length (see Appendix A). The stimuli were constructed so that the vowel
patterns and consonant clusters are regular at the bi-gram level. Thus, the stimuli contain
only short vowel and double vowel patterns. Similar to Castles and Holmes (1996), the
stimuli were designed so that each individual nonsense word was assigned one of two
possible pronunciations, a regular pronunciation or an irregular pronunciation. For
example, in the regular pronunciation condition, the nonsense word "hape" would be
pronounced /hep/ as this pronunciation uses the most frequent, lawful pronunciation of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
17
the vowel pattern. However, in the irregular pronunciation condition, it would be
pronounced /heap/, making it an irregular word as this uses a pronunciation o f the vowel
pattern which is considered either exceptional or strange. The stimuli were counter
balanced for pronunciation type across two testing conditions such that a word given a
regular pronunciation in condition 1 had an irregular pronunciation in condition 2.
Dyslexic subjects were assigned to one of two training conditions based upon
their subgroup assignment from the previous year. Reading level matched comparison
group members were assigned to testing conditions based upon their Woodcock Word
Identification Grade Equivalency score to assure that the two training conditions were
equated on word recognition. Each individual subject was taught 11 regular
pronunciations and 11 irregular pronunciations over the course of six training trials. The
stimuli appeared in bolded, 64 point type and were centered on laminated sheets o f 8.5
by 11 inch white paper organized inside a binder. The items appeared in a new order for
all training and testing trials. The order of the stimuli was quasi randomly determined to
control for word type (i.e., regular or irregular) such that no more than three words o f the
same type could appear in succession.
The first trial of the first training day was an experimenter trial in which the
experimenter taught the child the correct pronunciation o f each individual item and asked
the subject to repeat it. Praise was given for correct repetitions and corrective feedback
given for errors. If the subject made an error, the experimenter simply repeated the
training trial until the subject could pronounce the item correctly.
The remaining training trials on the first training day were subject generated.
Subjects were shown each item and asked to generate the proper pronunciation. This
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
18
process was feedback intensive. Experimenters were instructed to give animated praise to
the children for correct responses and were instructed to give corrective feedback for
incorrect responses while praising the child for effort made in attempting the trial.
Specifically, if the subject responded correctly, the experimenter gave the child positive
feedback which included an additional exposure to the stimulus. For example, the
experimenter might say "good job with hape" to reward a subject for reading the stimulus
correctly. This extra exposure to the word was to insure that every subject had an equal
amount of exposure to the correct pronunciation of the item. Similarly, if the subject
made an incorrect response the experimenter provided corrective feedback and asked the
child to attempt that particular item again.
The first training trial for the second day was a subject generated trial. However,
it was considered a reminder trial used to reacquaint the subjects with the stimuli. Thus,
while accuracy data were recorded for this trial these data were not used in the analyses.
The two remaining trials were identical in procedure to the last two training trials on the
first day. Accuracy data were recorded for all trials but the first trial of the first day.
On the first testing day (the third day of the overall program), subjects were given
a spelling test. Subjects did not receive any additional training on the stimuli. Subjects
instructed to try and remember how the alien names were spelled and to do their best to
spell the alien names on their answer sheet. The experimenter pronounced each item
once for the child and asked the child to spell the alien's name on their answer sheet. The
experimenter repeated each item once for the subject. No feedback was given on this
task. Accuracy and error data were recorded.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
19
On the final day o f the experiment, subjects were given a computerized, delayed
word recognition task. Subjects sat in front o f a Macintosh powerbook computer and
looked at the screen. A fixation cross appeared on the screen to signal the onset of each
trial. Subjects were told that they were going to see the alien names on the computer
screen. They were instructed to look at the screen and pronounce the item as soon as they
could. Accuracy data were recorded by the experimenter using a button box.
Subgroup Assignment
Subgroup assignment for the dyslexics was based on a Exception Word Reading
minus Nonsense Word Reading discrepancy score. If a subject demonstrated a
discrepancy of an absolute value o f .4 or more standard deviations from the mean score
s/he was assigned to the appropriate, corresponding subgroup. A cut off of .4 was chosen
such that the subjects who fall within this range represent the bottom quarter of the
sample in terms of their nonsense word or exception word reading ability. This method
of subgrouping was chosen as it provides a clear description of differences in
phonological and orthographic skill while allowing for wide variability in reading ability
due individual differences in degree of overall reading impairment. This subgrouping
method is similar to the Castles, et al., (1999) method as they also used discrepancy
scores to create their subgrouping assignments. Individual scores o f the Nonsense word
Reading and Exception Word Reading measures were converted in to Z scores based on
the scores of a norm-group (n= 71). The norm-group is composed o f a select group of
fifth graders representative o f each quartile of reading ability. Children that yielded a
discrepancy with a absolute value o f .4 or greater between their Exception and Nonsense
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
20
Word Reading scores were considered eligible for either the surface or phonological
dyslexic groups.
Before subjects were assigned to either the phonological or surface group, they
were matched on Woodcock Word Identification grade equivalency scores such that both
dyslexic groups had equal word recognition ability and equal sample size. Specifically,
each potential surface dyslexic was matched with a potential phonological dyslexic whose
word recognition level was within five tenths o f a grade point according to Woodcock
Word Identification score. This process yielded 13 phonological dyslexics with a mean
age o f 121.38 (sd = 5.78) and 13 surface dyslexics with a mean age of 124.77 (sd =6.99).
Results
Initial Comparison of all Dyslexics to both Comparison Groups
Before analyzing any potential subgroup differences in performance on the novel
tasks manufactured for the current study, an initial comparison was made between all the
dyslexics and both reading level and age-matched comparison groups. As specified
above, to be considered a dyslexic, a participant must have met the following criteria: (1)
scored at or below the 25th percentile on the Woodcock, (2) had a standard score scored
of 85 or higher on either the Peabody Picture Vocabulary Test or the Woodcock Visual
Closure Test, and (3) had a standard score no lower than 60 on both the Peabody Picture
Vocabulary Test and the Woodcock Visual Closure Test. Using these criteria, 71
children in our study qualified as dyslexics. In addition, there are 21 reading level and 31
age-matched comparison subjects in the current sample. Table 1 contains the means and
groups differences of these three groups on a variety o f measures.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
21
Several clear observations are possible from the data presented in Table 1. First,
the dyslexic group shows a typical profile in that they manifest significantly lower word
recognition (i.e., Woodcock Word Identification grade equivalency) and orthographic
knowledge (i.e., exception word reading and orthographic choice) relative to an age-
matched comparison group. In addition, the dyslexic sample exhibited significantly
lower performance on phonological awareness tasks (i.e., phoneme deletion and non
word reading) compared to both reading level and age-matched groups. Second, although
dyslexics showed lower phonological awareness relative to the reading level comparison
group, these two groups did not differ on tasks of orthographic knowledge (i.e., exception
word reading and orthographic choice). Finally, the dyslexic sample appears somewhat
low in verbal skill evidenced by a Peabody Picture Vocabulary standard score mean (i.e.,
91.7) that is approximately 9 points lower than their Woodcock Visual Closure standard
score mean of 100.4.
Initial Comparison of Phonological and Surface Dvslexics
Using the criteria discussed earlier, 13 phonological dyslexics, 13 surface
dyslexics, and 21 reading level controls were selected for comparison. Initial analyses of
group differences on a variety o f tests among these three groups is details in Table 2. As
seen in Table 2, phonological dyslexics, surface dyslexics, and the reading level
comparison group do not significantly differ in Woodcock grade equivalency scores (i.e.,
3.09, 3.02, and 3.09, respectively). In addition, consistent with expectations, we obtained
no subgroup differences in such tasks as Peabody Picture Vocabulary or Woodcock
Visual Closure.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Given that the two dyslexic subgroups were selected based on discrepancies in
exception word and non-word reading scores, overall mean differences between the
groups should emerge. As seen in Table 2, this is in fact the case as phonological
dyslexics significantly outperform surface dyslexics on exception word reading whereas
this result is reversed on the non-word reading task. As a measure to assess the internal
validity of the subgroup selection process, subgroup comparisons were also performed on
measures (i.e.. orthographic choice and phoneme deletion) which were not used as part of
the selection criteria but tap into the same underlying constructs o f orthographic and
phonological skills. Consistent with expectations, the two subgroups differ significantly
on these two measures with phonological dyslexics achieving higher scores on the
orthographic choice measure and surface dyslexics scoring higher on phoneme deletion.
In summary, the phonological and surface subgroups significantly differ in the predicted
direction on all measures where one would anticipate differences based on theoretical
reasons while the two groups do not differ on any measure where a significant difference
was not expected.
Although there are no theoretical reasons to believe that phonological and surface
dyslexics would differ on Peabody Picture Vocabulary or Woodcock Visual Closure,
other tasks included in our battery are relevant to hypothesized differences between
phonological and surface dyslexics and can be used to address questions concerning the
nature o f surface dyslexia. First, do surface dyslexics demonstrate significantly lower
print exposure relative to phonological dyslexics? The current study did not show
statistically significant differences in print exposure levels between the two subgroups.
Second, do surface dyslexics show significantly lower performance on a task concerned
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
23
with the visual analysis of letters (i.e., letter matching)? Again, we obtain no statistical
support for this hypothesis. In fact, phonological dyslexics were less accurate (i.e.,
79.2%) when compared to surface dyslexics (i.e., 89.9%), but this trend could be due the
greater speed with which phonological dyslexics completed the task (i.e., 2048 ms)
relative to surface dyslexics (i.e.. 2302 ms). As such, there is a possible speed/accuracy
trade-off with phonological dyslexics apparently sacrificing accuracy for increased speed.
It should also be noted that surface dyslexics did not differ significantly from the
reading level comparison group on the phoneme deletion task but they did scored
significantly lower than the reading level comparison group on the test of non-word
reading. Given these findings, it would seem that the surface dyslexics have a mild
phonological deficit relative to the reading level comparison group and visual inspection
of Table 1 indicates that the surface dyslexic group certainly shows even lower
performance on these two phonological tasks compared to the age-matched comparison
group. However, this mild phonological impairment is only one aspect o f surface
dyslexia.
Subgroup Comparison on the Learning Trials of Regular and Irregular Words
To assess whether phonological and surface dyslexics showed different learning
rates for regularly and irregularly pronounced novel words, a 3 (subgroup: phonological
dyslexics / surface dyslexics / reading level comparison group) x 2 (word type: regular /
irregular) x 2 (day: day 1 / day 2) repeated measures ANOVA was performed on the
percentage of words answered correctly with the last two factors as within-subjects
variables. Results indicate a main effect for subgroup, F(2,44) = 10.83, p<.001, a main
effect for day, F(l,44) = 310.42, p<.001. a main effect for word type, F( 1,44) = 94.54,
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
24
p<.001, a significant subgroup x word type interaction, F (2,44) = 6.35, p<.01, and a
marginal day x subgroup interaction, F(2,44) = 2.98, p=.061. However, all of these
effects are qualified by a significant subgroup x word type x day interaction, F(2,44) =
6.99, p<.01. This triple interaction is still significant when the reading level control
comparison group is removed, F(l, 24) = 10.38, p < .01. Figure 1 displays the percentage
of words correct for each subgroup across both regular and irregular words for both
subject trials on day 1 and day 2.
Since the interpretation of a triple interaction is often difficult given the many
post-hoc comparisons which are possible, two potential ways o f understanding this
interaction are presented. First, a series o f separate 2 (word type: regular / irregular) x 2
(day: day 1 / day 2) repeated measures ANOVAs were performed for each of the three
subgroups (i.e., phonological dyslexics, surface dyslexics, and the reading level
comparison group). Results indicated that only the phonological group displayed a
statistically significant word type x day interaction, F(l,12) = 5.93, p<.05. Specifically,
phonological dyslexics showed greater progress over time learning irregular words
relative to regular words whereas surface dyslexics and the reading level comparison
group showed approximately equal progress on both types o f words. Visual inspection of
Figure 1 shows that the gap between regular and irregular word performance narrowed
significantly for the phonological group, but that this same gap remained fairly constant
for both the surface dyslexics and reading level comparison group.
A second way to understand the triple interaction o f subgroup, word type, and day
is to examine the potential interaction of subgroup and word type separately for day 1 and
day 2. Performance o f the appropriate 3 (subgroup: phonological dyslexics / surface
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
25
dyslexics / reading level comparison group) x 2 (word type: regular / irregular) repeated
measures ANOVAs indicated no interaction o f subgroup and word type for day 1, F(2,44)
= 1.78. p>. 10, but a significant subgroup x word type interaction for day 2, F(2,44) =
9.48, p<.001. Removal o f the reading level comparison group from both analyses did not
alter the findings.
Further analysis o f the day 2 data is warranted for two reasons. First, the previous
analysis indicated that a significant subgroup x word type interaction was present only for
day 2. Second, since learning novel nonsense words is an extremely difficult task for
young readers, it is reasonable to assume that differences in the patterns of learning
regular and irregular words will become more pronounced with time. Given this fact, it is
important to look at the day 2 data since this is more likely to show differential subgroup
recall accuracy for regular and irregular words. Theory would predict an interaction
between subgroup and word type such that phonological dyslexics should be better than
surface dyslexics on irregular words but the opposite trend should hold for regular words.
To test this hypothesis, a 3 (subgroup: phonological dyslexics / surface dyslexics / reading
level control) x 2 (word type: regular / irregular) repeated measures ANOVA was
performed with the second factor as a within-subjects variable. The dependent measure
used in this analysis was a composite of the two learning trials held on the second day of
the study. It was reasoned that a composite variable would provide more stability than
either one of the measures individually. In addition, reliability coefficients for the two
regular and irregular scores were .91 and .83, receptively. It should be noted that this
analysis is conceptually identical to the repeated measures ANOVA performed in the
preceding paragraph.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
26
Results o f this analyses showed a significant main effect for word type, F(l,44) =
60.66, p<.001, with a higher recall percentage of regular compared to irregular words.
More important, and consistent with expectation, there was a significant interaction
between subgroup and word type, F(2,44) = 9.48, p<.001. This same interaction held
when the reading level comparison group was removed from the analysis, F(l,24) =
16.06, p=.001. Figure 2 displays the means for phonological and surface dyslexic groups
for both regular and irregular words. Inspection o f this figure reveals that phonological
dyslexics showed the theoretically predicted pattern o f better performance than surface
dyslexics on regular words but this trend was reversed on irregular words with surface
dyslexics outperforming phonological dyslexics. Post-hoc analyses comparing both
subgroups on each o f the word types showed that there was a significant difference
between phonological and surface dyslexics on regular words, t(24) = 2.13, p<.05, but
only a nonsignificant trend for irregular words, t(24) = 1.53, p=.139.
Subgroup Comparison on Regular and Irregular Words for the Spelling Task
To further investigate the interactive effect of subgroup and word type, a 2
(subgroup: phonological dyslexics / surface dyslexics) x 2 (word type: regular / irregular)
repeated measures ANOVA was performed on the spelling data collected on day 3 or the
training program. Although the trend of the means was identical to the day 2 data and in
the theoretically predicted direction (i.e., surface dyslexics showing higher performance
on regular words and lower performance on irregular words relative to phonological
dyslexics), the interaction between subgroup and word type did not reach standard levels
of statistical significance, F(l,24) = 2.44, p=.13. This result is potentially due to the
increased difficulty o f this particular task relative to all other tasks in the training
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
27
program. Specifically in the day 1 recall, day 2 recall, and the delayed recall task the
participant only needed to pronounce the word correctly when it was shown to them. In
contrast, the current task necessitated that the participant spell the word correctly.
Consistent the complicated nature of this task, neither subgroup exceeded a mean
accuracy rate of over 30% on either word type.
Subgroup Comparison on Regular and Irregular Words for the Delaved Recall Task
In order to assess whether differential subgroup performance on regular and
irregular words extended beyond the learning trials conducted on the first two days o f the
program, a computer-based outcome measure was added on the fourth day of the
program. This outcome measure had two distinct qualities. First, and most important,
the measure was given on the fourth day of the training program which occurred one
week after the data from the two separate day #2 learning trials was collected. It was this
latter data that formed the composite measure that is shown in Figure 2. Given this gap in
time between day#2 and day#4 o f the training program, we can attempt to assess the
robustness over time o f the finding shown in Figure 2. Second, the method by which the
stimuli for the current outcome measure differs slightly from the data collected during the
training trials. Specifically, during the training trials on the first and second day of the
program, the experimenter would show the words to the participant on 8 1 / : x II inch
pieces o f paper. In contrast, the format of the outcome measure had participants sit in
front o f a computer screen and the stimuli words were presented sequentially. As such,
the computer-based outcome measure can speak to the robustness o f the research finding
across slightly different stimuli presentation formats.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
28
In order to assess the effect of subgroup and word type on the computer-based
outcome measure, a 3 (subgroup: phonological dyslexics / surface dyslexics / reading
level control) x 2 (word type: regular / irregular) repeated measures ANOVA was
performed with the second factor as a within-subjects variable. Identical to previous
findings, results indicated a main effect for word type, F(l,44) 16.57, p<.001, with higher
recall of regular words relative to irregular words. However, this main effect is qualified
by a significant subgroup x word type interaction, F(2,44) = 5.60, p<.01. As with
previous analyses, this interaction remains significant even after the reading level
comparison group is removed from the analysis, F(l,24) = 9.22, p<.01. Figure 3 shows
the means for phonological and surface dyslexic groups for both regular and irregular
words. Consistent with earlier findings, phonological dyslexics displayed better levels of
performance than surface dyslexics on regular words but this trend was reversed on
irregular words. Post-hoc analyses comparing both subgroups on each of the word types
showed that there was a marginally significant difference between phonological and
surface dyslexics on regular words, t(24) = 1.73, p=.096, but only a nonsignificant trend
for irregular words, t(24) = 1.02, p>. 10.
Title Recognition and Questionnaire Data
To assess potential subgroup differences in reading experience, data was collected
using both a title recognition task completed by each participant and a questionnaire that
was filled out by the primary teacher of each participant asking about the participant's
observable reading habits. As discussed earlier, and as seen in Table 2, there is no
significant difference between surface and phonological dyslexics the title recognition
task (i.e., print exposure) which is a measure o f reading experience. In addition, although
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
29
not statistically significant, the reading level comparison group does display a lower level
o f reading experience which is expected due to their younger age.
Second, as detailed above in the method section, there were 6 relevant items from
the teacher questionnaire that assessed participant’s reading habits. Analyses of each
individual question using t-tests did not reveal significant differences between subgroups
even though statistical power was increased by not controlling for inflation of Type I error
rate across the six tests. It should also be noted that the reading level comparison group
had higher mean values (which corresponds to more active reading habits) for all 6
questions relative to both dyslexic groups. Post-hoc analyses show that these differences
were statistically significant relative to at least one of the dyslexic groups for 5 of the
questions with a marginally significant result for the sixth question.
In summary, no differences were detected between phonological and surface
dyslexics for any measure o f reading experience or exposure in the current study. In
contrast, relative to the dyslexic groups, the reading level comparison group appeared to
have a slightly lower knowledge o f books (as measured by the print exposure task), which
is most likely due to their younger age. However, according to the data collected from
their teachers, the reading level comparison group appears to be composed of more active
readers.
Paired Associates Learning Data
As seen in Table 2, there are no differences among the phonological dyslexics,
surface dyslexics, and the reading level comparison group in either the verbal or non
verbal paired associates tasks. As such, it is unlikely that any o f the main findings
presented in this study can be attributed to subgroup differences in the ability to learn
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
30
arbitrary paired associates. More specific to various hypotheses concerning the possible
nature of surface dyslexia, it should be noted that surface dyslexics did not show a
significantly slower general learning rate relative to either phonological dyslexics or the
reading level comparison group. In addition, phonological dyslexics did not show a faster
learning rate compared to the reading level comparison group. This last fact is important
to point out given that phonological dyslexics appear to make faster progress on the
learning of irregular words relative to the other two groups (see Figure 1).
Discussion
The current study posited two major hypotheses. The first was an interaction
between subgroup and word type such that surface dyslexics would out perform
phonological dyslexics on regular words but this trend would be reversed for irregular
words. The data from day 2, the spelling task, and the delayed recall measure all clearly
support this hypothesis. This pattern of means is very consistent as it occurs not only in
different points in time in the study but also across different presentation formats. The
second hypothesis was that surface dyslexics would perform in a similar manner to the
reading level comparison group in terms of their rate of learning the novel stimuli.
However, although the pattern will be similar, surface dyslexics should demonstrate a
lower level of accuracy. Statistical analyses of the data presented in Figure 1 demonstrate
that indeed the surface dyslexics do perform in this fashion while phonological dyslexics
exhibit a distinct profile. Specifically, phonological dyslexics showed an increased rate
of learning of irregular words compared to regular words. This trend was not the same
for the other two groups.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
31
One potential way to understand the theoretical implications o f the triple
interaction presented in Figure 1 is to examine the relative performance of each subgroup
separately for both regular and irregular words. As seen in figure 1, there are initial
discrepancies between the performance o f the reading level comparison group, the surface
dyslexics, and the phonological dyslexics that remain fairly constant through out all four
training trails on regular words. Specifically, the reading level comparison group
performs at the highest level, phonological dyslexics perform at the lowest level and the
surface dyslexics are in between. This observation is consistent with the idea o f a core
phonological impairment as outlined in Harm and Seidenberg (1999) in that the
differential performance in the subgroups may be attributed to different levels o f inherent
phonological skill. The phonological dyslexics clearly demonstrate the most extreme
phonological impairment o f the three groups and consequently exhibit the poorest
performance on the regular words. As discussed earlier, the surface dyslexics have a mild
phonological impairment and as such perform below the level o f the reading level
comparison group and above the phonological group on regular words. As the reading
level comparison group would appear to be free of phonological impairments, they
perform the best. As any type o f phonological impairment will take time to remediate, it
is reasonable to believe that the consistency of subgroup differences on regular words will
remain constant across short time periods as they do in the current study. These findings
are highly consistent with the idea of core phonological impairments in that in the
absence of remediation, an inherent phonological difficulty will remain consistent. Thus
performance on phonological tasks should also remain consistent.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
32
Conversely, these findings are inconsistent with the predictions o f the dual route
model which asserts that surface dyslexics have relatively good sub-lexical processing
skills but impaired lexical processing skills in comparison to same aged normal readers.
While the dual route model does not expressly rule out the idea o f surface dyslexics with
mild phonological impairments of the type our surface dyslexic group demonstrates (as is
evidenced in the nonsense word reading measure), it does specify that these impairments
are due primarily to corruptions along the lexical route. If this is the case, the dual route
model would predict that while surface dyslexics may not perform as well as same aged
normal readers, they should be able to perform as well as a younger reading aged matched
comparison group on the Word Learning Task for regular words as surface dyslexics have
relatively well intact sub-lexical skills. While dual route predictions may hold true for
surface dyslexics with normal nonsense word reading, Figure 1 demonstrates that this is
not the case in this sample o f surface dyslexics with mild phonological impairments.
The pattern o f subgroup performance on the irregular words as seen in Figure 1
tells a different story. Unlike the regular word items where subgroup differences remain
constant across time due to the robust nature of phonological impairment, relative
subgroup performance on the irregular words did change across training trials.
Specifically, although the phonological and surface subgroups demonstrated similar
performance on irregular words during the training trials on day 1, with surface dyslexics
slightly out performing phonological dyslexics, the pattern o f results changes dramatically
for day 2. First, the performance o f the phonological dyslexics surpasses that o f the
surface dyslexics. Second, by the final training trail, phonological dyslexics have not
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
33
only surpassed the surface dyslexics, but have essentially matched the performance o f the
reading level comparison group on irregular words.
One potential explanation of the discrepant performance o f the phonological
group centers around the idea of word-specific learning strategies. As the irregular words
are designed such that phonological strategies are ineffective in producing correct
pronunciations, these pronunciations must essentially be produced using word-specific
learning strategies. Unlike phonological skills, word-specific learning strategies can be
implemented quickly as they are highly sensitive to practice effects. Hence,
implementing word-specific learning strategies can generate improvements in
performance over a short period of time. Hence, it appears that the phonological
dyslexic's rapid increase in learning rate can be attributed to highly developed word-
specific learning skills. One possible reason for the marked discrepancy in this ability
between the phonological dyslexics and the other subgroups is that phonological
dyslexics use these skills more often than the other subgroups as a means of
compensating for the specific reading difficulties they encounter due to an underlying
phonological impairment. In essence, most phonological cues are ineffective for
phonological dyslexics. Thus they must use word-specific learning strategies in order to
read most words. Visual inspection of Figure 1 indicates that the phonological dyslexics
performed at equal levels on the regular words and the irregular words. One possible
interpretation o f this is that they used word-specific learning strategies to learn both types
of words. Chance alone would indicate that if this was the case, the phonological
dyslexics would learn an equal number of regular and irregular words as is evidenced in
Figure 1. Given this hypothesis, it is quite likely that phonological dyslexics are utilizing
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
34
these word-specific learning skills rather than phonological strategies across both word
types as they did not demonstrate the regularity effect seen in the other groups.
Another interesting component of the pattern of performance on the irregular
word stimuli has to do with learning rate. While the surface dyslexics exhibit poorer
performance than the other subgroups on the irregular word items, the surface dyslexic’ s
rate o f learning irregular words is similar to their rate of learning regular words.
Additionally, the surface dyslexic group has a general rate o f word learning that is
relatively similar to that o f the reading level matched comparison group for both word
types. As surface dyslexics learn at the same rate as reading level matched controls who
are presumed to be free o f learning impairments, it can be argued that surface dyslexics
do not have a rate o f learning deficit. This finding suggests that while surface dyslexics
do not seem to experience difficulty employing word-specific learning strategies as they
are able to learn just as "fast" as reading level comparison group members, they do
experience some initial difficulty, quite possibly due to an orthographic processing
deficiency, that depresses performance on irregular word specific learning tasks. These
findings have several implications for theories of the nature o f surface dyslexia.
Proponents o f the dual route model would characterize depressed performance on
exception word reading tasks as a specific difficulty in storing and retrieving whole word
representations stemming from an impairment in the lexical route (Castles & Coltheart,
1993). However, dual route theorists also produce findings that seem inconsistent with
the dual route interpretation of exception word reading. Interestingly, Castles and
Holmes (1996) point out that phonological and surface dyslexics experience more
difficulty implementing word-specific learning strategies when learning strange words
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
35
(e.g.. “ache”) in comparison to exception words (e.g., “have”) and that this difficulty is
especially pronounced for surface dyslexics. As Castles and Holmes (1996) constructed
the exception word stimuli such that the items "were derived from exception words
(whose orthographic structures are not irregular but whose pronunciations are)" (p. 132)
and constructed the strange word stimuli such that the "orthographic structure and
pronunciations are irregular" (p. 132) it appears that all dyslexics, especially surface
dyslexics, are in some way sensitive to regularity components inherent in the orthographic
structure of the stimuli. This is evidenced by the finding that surface dyslexics
experience pronounced difficulty with strange word stimuli which use irregular
orthographic structure and pronunciations and do not experience the same degree of
difficulty with exception words which only use irregular pronunciations. Dual route
interpretations would not predict a difference between exception word and strange word
learning performance as the lexical route does not treat these word types differently. This
provides evidence that there is more to irregular word specific reading skill than the
simple storage and retrieval of whole word entities characteristic of lexical route
functions. Hence, Harm and Seidenberg's (1999) connectionist conceptualization of
irregular word specific learning may better explain this concept.
The Harm and Seidenberg (1999) model asserts that specific irregular word
reading deficits can occur from many possible sources and involve more than just storage
and retrieval of words as whole word entities. The first o f these potential sources is an
environmental learning/experience deficit. While this hypothesis may be true o f some
surface dyslexics it does not explain the findings of the current study. The notion that the
novel nonsense word aspect of the Word Learning Task controls for exposure to the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
36
stimuli and amount o f learning experience, in combination with the data indicating that
there are no subgroup differences on the Title Recognition Task and the Teacher
Questionnaire, suggest that print exposure or reading experience deficits are not present
in this sample. Hence, the “lack o f experience hypothesis” does not explain the
performance o f the surface dyslexics in our sample.
Also o f interest is the general cognitive delay hypothesis posited by Harm and
Seidenberg (1999). The data from the current study support this notion in the sense that
the paired associate learning data in combination with the letter matching data indicate
that surface dyslexics do behave in a similar fashion to the reading level comparison
group across a variety of cognitive tasks. However, this general delay hypothesis does
not sufficiently explain the reading specific behaviors of surface dyslexics as they
perform below the reading level matched comparison group on the Word Learning Task.
If the root o f this specific irregular word reading deficit was a general cognitive delay,
then the surface dyslexics should behave like the reading level matched comparison group
across both word types. As the data indicate that this is not the case in this sample, the
general cognitive delay hypothesis does not seem to fit these data.
However, hypotheses offered by Harm and Seidenberg (1999) pertaining to input
deficiencies are more consistent with the findings o f the current study. The first o f these
hypotheses, the visual processing deficit, is in need of further exploration. Initial results
o f the letter matching task address the issue of an underlying visual processing deficit in a
preliminary fashion. While this task is not designed to pick up any visual processing
difficulties that occur on the word level, the results do suggest that visual processing skill
is not a problem at the letter processing level for either of the dyslexic groups. These
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
37
results warrant further investigation into the visual processing deficit hypothesis, thus this
hypothesis should not be ruled out. However other aspects o f the input deficiency
hypothesis are more consistent with the findings of the current study.
The notions of a hidden unit deficiency or a problem establishing adequate
orthographic representations o f words provide a rich theoretical context in which to
interpret these findings.
The results of analyses on the paired associate learning tasks indicate that the dyslexic
groups do not differ in comparison to each other or in comparison with the reading level
matched comparison group in their ability to learn general arbitrary paired associate
relationships. These findings support the notion that the difficulty surface dyslexics
experience in learning irregular words is not due to a global paired associate learning
impairment, but rather a irregular word specific mechanism. Additionally, these data help
support the interpretation that surface dyslexics do not have a general rate o f learning
deficit, but rather a core deficit that suppresses irregular word reading performance.
These data lend credence to the notion o f the hidden unit deficit may be responsible for
this core deficit as Harm and Seidenberg (1999) posit that having a deficient number of
hidden units can result in impairments in irregular word specific reading ability. In
addition, the hidden unit deficiency theory does not exclude the idea that orthographic
skill and phonological skill play some role in irregular word specific reading. In fact, a
hidden unit deficiency could play a role in disabling these phonological and orthographic
reading processes. While further investigation is needed to determine the viability of the
hidden unit deficiency hypothesis, these initial findings are encouraging.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
38
Alternatively, these data also speak to the idea that this core deficit experienced by
surface dyslexics may be an orthographic processing deficit that manifests itself in a
marked difficulty developing adequate orthographic representations o f words. As the
findings in the current study suggest that environmental deficits, visual impairments,
general cognitive delay, rate o f learning, and phonological deficit do not fully account for
the word learning profile of surface dyslexics in this sample, it seems logical that
orthographic representation deficits are a worthwhile construct to investigate and a likely
component of this core irregular word reading deficit. While the work o f other theorists
should by no means be discounted, further investigations of these connectionist ideas may
yield fruitful information regarding the etiology of surface dyslexia.
Lastly, it is important to mention that these data also have relevant implications
for the Stanovich et al. (1997) hypothesis concerning the nature o f surface dyslexia.
While these data support the notion that surface dyslexics as a group do have a mild
phonological impairment relative to a reading age matched comparison group as well as
an age matched comparison group, we find no evidence of print exposure differences
between the subgroups. Statistical analysis o f the title recognition task does not yield
significant differences between the dyslexic subgroups even though dyslexics as a group
do have significantly lower levels o f print exposure when compared to same aged normal
readers. However, as the same aged normal readers are more skilled readers, it is not
surprising that they have higher levels of print exposure as assessed by our title
recognition task. Moreover, this finding is supported by the teacher questionnaire data.
Results o f these analyses indicate that members o f the reading level comparison group
tend to be more active readers than the dyslexics, yet, there are no significant group
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
39
differences between the dyslexic subgroups. Again, this finding is not surprising that
normal readers read more than disabled readers. However, it is important to note that the
reading habits of the dyslexic subgroups appear very similar. This suggests that deficits
in print exposure or reading opportunities in combination with a mild phonological deficit
do not account for the findings in the current study.
While these data presented in the current study do provide interesting insights into
differing theoretical accounts of the etiologies and specific reading behaviors associated
with developmental surface and phonological dyslexia, more empirical work is needed to
fully understand the true nature of these constructs. For instance, the current study should
be expanded to include a same aged normal reader group. Additionally, the exception
word stimuli should be modified to include nonsense words with irregular pronunciations
that are less difficult for children to master. Furthermore, replicating this study with a
more pure surface dyslexic group may reveal more specific information about the etiology
o f surface dyslexia.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
40
Table 1. Means, Standard Deviations, and Group Differences o f all eligible dyslexics,
reading level matched controls, and age matched controls.
(VARIABLES)
Dyslexics
(n = 71)
Reading Level
Comparison
Group
(n = 21)
Age-Matched
Comparison
Group
(n = 31)
Woodcock Word Identification
Grade Equivalency
3.29a (0.66) 3.09a (0.63) 5.99b (1.96)
Woodcock Word Identification
Standard Score (SS)
11. t (9.0) 116.2b (9.2) 102.7C (8.8)
Peabody Picture Vocab. (SS)
91.7a (14.5) 104.9b (14.6) 106.4b (13.6)
Woodcock Visual Closure (SS)
100.4a (13.7) 108.1b (12.3) 101.43 (9.2)
Orthographic Choice (%)
77.6a (9.1) 76.5a (8.8) 87.0b (5.8)
Phoneme Deletion (%)
49. l a (19.6) 60.5b (24.0) 74.9C (15.1)
Verbal Paired Associates (%)
7 I.7a (17.4) 71.7a (13.8) 80.7b (12.9)
Non-Verbal Paired Associates (%)
58.3a (19.6) 56.4a (18.5) 61.23 (21.8)
Exception Word Reading (%)
58.5a (14.7) 52.0a (17.1) 84.2b (5.5)
Non-Word Reading (%)
36.2a (17.2) 51.8b (15.9) 73.3C (17.8)
Print Exposure (%)
29. l a (20.1) 17.5b (17.8) 38.6C (16.0)
Letter Matching Accuracy (%)
86.8a (11.9) 83.9a (17.8) 90.93 (7.4)
Note: Mean scores within rows that do not share a common superscript differ
significantly in post-hoc pairwise comparison analyses (p<.05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
41
Table 2. Means, Standard Deviations, and Group Differences o f phonological dyslexics,
surface dyslexics, and reading level matched controls.
(VARIABLES)
Phonological
Dyslexics
(n = 13)
Surface
Dyslexics
(n = 13)
Reading Level
Comparison
Group
(n = 21)
Woodcock Word Identification
Grade Equivalency
3.09a (0.58) 3.02a (0.64) 3.09a (0.63)
Woodcock Word Identification
Standard Score (SS)
76.4a (7.5) 73.9a (9.0) 116.2b (9.2)
Peabody Picture Vocab. (SS)
86.7a (14.6) 91.3a (17.4) 104.9b (14.6)
Woodcock Visual Closure (SS)
96.1a (12.9) 101.8ab(13.1) 108.l b (12.3)
Orthographic Choice (%)
80.8a (8.3) 72.4b (8.4) 76.5ab (8.8)
Phoneme Deletion (%)
31.9a (21.0) 56.7b (18.5) 60.5b (24.0)
Verbal Paired Associates (%)
70.5a (19.6) 63.8a (19.2) 71.7a (13.8)
Non-Verbal Paired Associates (%)
50.0a (16.7) 55.l a (21.9) 56.4a (18.5)
Exception Word Reading (%)
58.6a (11.9) 47.2b (11.9) 52.0ab (17.1)
Non-Word Reading (%)
21.9a (12.3) 37.l b (15.3) 51.8C(15.9)
Print Exposure (%)
25.3a (25.2) 29.5a (19.6) 17.5a (17.8)
Letter Matching Accuracy (%)
79.2a (19.1) 89.9a (9.7) 83.9a (17.8)
Letter Matching RT (ms)
2048a (507) 2302a (592) 2523a (1090)
Note: Mean scores within rows that do not share a common superscript differ
significantly in post-hoc pairwise comparison analyses (p<.05).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
42
Figure 1. The effect of subgroup, word type, and day (includes both learning trials) on
the percentage of words correct.
90 n
80
70
g 60
o
U 50
s >
00
| 40
CJ
« 3 30 4
cu
20
10
0
— Phonological - Reg
— Phonological - Irreg
• • Surface - Reg
— Surface - Irreg
— RL-Reg
*>—RL- Irreg
Day 1
Trial# 1
Dayl
Trial#2
Day 2
Trial# 1
Day 2
Trial#2
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Percentage Correct
43
Figure 2. The effect of subgroup and word type on the percentage of words correct.
90 i
80 -
- Phonological
Surface
— RL Control
60 -
50 -
40 -
Regular Irregular
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
44
Figure 3. The effect of subgroup and word type on the percentage of words correct for
the computer task.
80 -
- Phonological
Surface
- RL Control
H
Urn
a 70 -
U
y
c
y
y
60 -
50 -
Regular Irregular
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
45
References
Adams, M.J., & Huggins, A. W.F. (1985). The growth o f children’ sight word
vocabulary: A quick test with educational and theoretical implications. Reading
Research Quarterly. 20. 262-281.
Bigsby, Pamela (1990). Abstract letter identities and developmental dyslexia.
British Journal o f Psychology. 8 1 .227-263.
B oder,E . (1973). Developmental dyslexia; a diagnostic approach based on
three typical reading-spelling patterns. Developmental Medicine and Child Neurology.
15, 663-687.
Campbell, R., & Butterworth, B. (1985). Phonological dyslexia and dysgraphia
in a highly literate subject: A developmental case with associated deficits of phonemic
processing and awareness. Quarterly Journal o f Experimental Psychology. 35. 435-475.
Castles, A., & Coltheart, M. (1993). Varieties o f developmental dyslexia.
Cognition. 47. 149-180.
Castles A.,& Coltheart, M. (1996). Cognitive correlates o f developmental
surface dyslexia: A single case study. Cognitive Neuropsychology. 13. 25-50.
Castles, A., Datta, H., Gayan, J. & Olson, R. K. (1999). Varieties of
developmental reading disorder. Genetic and environmental influences. Journal of
Experimental Child Psychology. 72 (2). 73-94.
Castles, A., & Holmes, V.M. (1996). Subtypes o f developmental dyslexia and
lexical acquisition. Australian Journal of Psychology. 48. 130-135.
Coltheart, M., (1978). Lexical access in simple reading tasks. In G. Underwood
(Ed.). Strategies of information processing (pp. 151-216). London: Academic Press.
Cunningham, A.E.,& Stanovich, K.E. (1990). Assessing print exposure and
orthographic processing skill in children; a quick measure o f reading experience. Journal
of Educational Psychology. 28. 733-740.
Goulansdris, N.K., & Snowling, M. (1991). Visual memory deficits: a plausible
cause of developmental dyslexia? Evidence from a single case study. Cognitive
Neuropsychology. 8. 127-154.
Hanley, R., Hastie, K., & Kay, J. (1992). Developmental surface dyslexia and
dysgraphia: An orthographic processing impairment. Quarterly Journal of Experimental
Psychology. 44. 285-319.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
46
Harm, M.W., & Seidenberg, M.S. (1999). Phonology, reading and dyslexia.
Insights from connectionist models. Psychological Review. 106 (31. 491-528.
Joanisse, M.F., Manis F.R., Keating, P., & Seidenberg M.S. (in press). Language
deficits in dyslexic children: Speech perception, phonology, and morphology. Journal of
Experimental Child Psychology.
Lieberman, I.V., & Shankweiler, D. (1985). Phonology and the problems of
learning to read and write. Remedial and Special Education. 6. 8-17.
Manis, F.R. (1985). Acquisition of word identification skills in normal and
disabled readers. Journal of Experimental Child Psychology. 77. 78-90.
Manis, F.R., Seidenberg, M.S., Doi, L.M., McBride-Chang, C., & Petersen, A.
(1996). On the basis o f 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. (in press). Development o f dyslexic subgroups: A one
year follow up. Annals o f Dyslexia.
Olson, R.K., Wise, B., Connors, F., Rack, J., & Fulker, D. (1989). Specific
deficits in component reading and language skills: Genetic and environmental influences.
Journal of Learning Disabilities. 22. 339-348.
Seidenbeg M.S., & McClelland, J.L. (1989). A distributed, developmental model
of word recognition and naming. Psychological Review. 96. 523-568.
Snowling, M.J., & Hulme, C. (1989). A longitudinal case study of
developmental phonological dyslexia.. Cognitive Neuropsychology. 6. 379-603.
Share, D.L., (1995). Phonological recoding and self-teaching: The sine qua non
of reading acquisition. Cognition. 55. 151-218.
Stanovich, K.E., Siegel, L.S. & Gottardo, A. (1997). Converging evidence for
phonological and surface subtypes of reading disability. Journal o f Educational
Psychology. 89. 114-128.
Wagner, R.K., & Torgesen, J.K. (1987). The nature o f phonological processing
and its causal role in the acquisition of reading skills. Psychological Bulletin. 101. 192-
212 .
Woodcock, R.W. (1989). Woodcock Reading Mastery Test-Revised. Circle
Pines, MN: American Guidance Service.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
47
Appendix A
Item Key Word Word Type
1. trom “prom” regular
“home” irregular
2. zide “ride” regular
“lid” irregular
3. veep "jeep” regular
“zip” irregular
4. duite “suit” regular
“suite” irregular
5. tauge “dodge” regular
“gauge” irregular
6. lource “sour+s” regular
“source” irregular
7. stoom “broom” regular
“some”
irregular
8. brote “wrote” regular
“but” irregular
9. grast “grass” regular
“graced” irregular
10. mouge “gouge” regular
“Scrooge” irregular
11. torps “tore+ps” regular
“corps” irregular
12. kuise “bruise” regular
“guise” irregular
13. dieve “believe” regular
“give” irregular
14. sloam “home” regular
“mom” irregular
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
48
15. choip
16. plish
17. mearse
18. smune
19. froupe
20. cheam
21. hape
22. glait
“choy+p”
“kwipe”
“dish”
“ply+sh”
“pierce”
“hearse”
“moon”
“sun”
“frau+p”
“troop”
“beam”
“jam”
“ape”
'lap ”
“late”
“wet”
regular
irregular
regular
irregular
regular
irregular
regular
irregular
regular
irregular
regular
irregular
regular
irregular
regular
irregular
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Dyslexic's development of word recognition and phonological skill over time
PDF
Determining competent/resilient outcomes across multiple domains in sexually abused and nonabused females
PDF
Applying the theory of reasoned action to condom use: The effect of immediate consequences on intention-behavior consistency
PDF
Auditory word identification in dyslexic and normally achieving readers
PDF
Effects of personal resource sufficiency on perceived difficulty and desirability of earthquake preparedness
PDF
Head injury and dementia: A co-twin control study of Swedish twins
PDF
Adaptation to sine-wave gratings selectively reduces the sensory gain of the adapted stimuli
PDF
Adolescents' social attitudes: Genes and culture?
PDF
Ethnic identity, acculturation, self-esteem and perceived discrimination: A comparison study of Asian American adolescents
PDF
Depressed children and the social and behavioral attributes of their best friend
PDF
Aging and the use of context and frequency information in ambiguity resolution
PDF
Articulated thoughts about intentions to commit anti-gay hate crimes
PDF
Cognitive functioning and dementia following cancer: A Swedish twin study
PDF
Factors influencing the interpretation of novel words as adjectives in 4-year-old Spanish -speaking children
PDF
Depression in offspring following severe prenatal stress
PDF
History of depression, antidepressant treatment, and other psychiatric illness as risk factors for Alzheimer's disease in a twin sample
PDF
Effects of bilateral stimulation and stimulus redundancy on performance in processing nonword letter trigrams
PDF
An examination of affective modulation, psychopathy, and negative schizotypy in college and community samples
PDF
Asymmetries in the bidirectional associative strengths between events in cue competition for causes and effects
PDF
Affection and conflict in family relationships
Asset Metadata
Creator
Bailey, Caroline Elizabeth
(author)
Core Title
Insights into the nature of phonological and surface dyslexia: Evidence from a novel word learning task
School
Graduate School
Degree
Master of Arts
Degree Program
Psychology
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
OAI-PMH Harvest,Psychology, clinical,psychology, cognitive,psychology, developmental
Language
English
Contributor
Digitized by ProQuest
(provenance)
Advisor
Manis, Franklin (
committee chair
), [illegible] (
committee member
)
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-c16-286886
Unique identifier
UC11341507
Identifier
1409616.pdf (filename),usctheses-c16-286886 (legacy record id)
Legacy Identifier
1409616-0.pdf
Dmrecord
286886
Document Type
Thesis
Rights
Bailey, Caroline Elizabeth
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
psychology, cognitive
psychology, developmental