University of Southern California Dissertations and Theses

Dyslexic's development of word recognition and phonological skill over time

(USC Thesis Other)

Dyslexic's development of word recognition and phonological skill over time

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INFORMATION TO USERS This manuscript has been reproduced from the microfilm master. UMI 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 will 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. DYSLEXIC’S DEVELOPMENT OF WORD RECOGNITION AND PHONOLOGICAL SKILL OVER TIME by Laurie Freedman A Thesis Presented to the FACULTY OF THE GRADUATE SCHOOL UNIVERSITY OF SOUTHERN CALIFORNIA In Partial Fulfillment o f the Requirements for the Degree MASTER OF ARTS (PSYCHOLOGY) December 1999 Copyright 1999 Laurie Freedman Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. UMI Number: 1409629 _® UMI UMI Microform 1409629 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 THE GRADUATE SCHOOL UNIVERSITY PARK LOS ANGELES. CALIFORNIA 10007 This thesis, written by under the direction of hi*:.—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 L A O P - I E " F fL & M P Dtmm T )„t0 December 20, 1999 THESIS COMMITTEE Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table of Contents I. Abstract (p. iv) II. Introduction (p. 1-8) III. Method (p. 8-15) a. Subjects (p. 8-10) b. Procedures (p. 10-15) 1. Standardized Measures (p. 10-12) 2. Experimental Measures (p. 12-15) IV. Results (p. 15-28) a. Correlations (p. 15-16) b. Group Differences In Word Recognition and Orthographic Skill (p. 16-22) c. Do Dyslexics Use Phonology Differently in Reading Printed Words? (p. 22-25) d. Semantic Categorization Data (p. 25-28) V. Discussion (p. 28-33) VI. References (p. 34-36) Vff. Tables (p. 37-46) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. ii. List o f Tables Table 1: Descriptives O f Each Group At Each Test Period (p. 37) Table 2: Time 1 Correlations Between Measures (p. 38-39) Table 3: Time 3 Correlations Between Measures (p. 40-42) Table 4: Orthographic ANOVA Data (p. 43) Table 5: Phonological ANOVA Data (p. 44) Table 6: Word Pronunciation ANOVA Data (p. 45) Table 7: Semantic Categorization ANOVA Data (p. 46) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Introduction: A main consensus in the reading community is that deficits in phonology underlie the cognitive deficits o f dyslexia. The development o f a variety o f phonological skills is dependent on cognitive access to phonological representations. (Snowling & Hulme. 1994) Thus, problems at this level of representation may account for the reported deficits that dyslexic children have in a variety o f skills; speech perception, speech production, naming, verbal short-term memory, and phonological awareness. (Snowling, Defty, Goulandris, 1996) Even though dyslexic children might compensate for their reading difficulties, they often show residual impairments when tested on phonological processing tasks. (Pennington, Van Orden, Smith. Green. & Haith. 1990) While the majority of cases of dyslexia appear to arise from an underlying phonological deficit, it is still important to explore other potential profiles o f reading disability. (Castles & Coltheart, 1993; Manis et al. 1996; Wolf & Bowers, 1999)The present research looks at the development o f dyslexics over time, and serves to answer the questions; “What is the nature o f this phonological deficit?” “How does it change over time?” and finally, “To what extent do dyslexics use phonology differently in reading printed words?” In their 1992 review paper, Rack. Snowling, and Olson compared dyslexic children to a younger, but reading level matched group o f children. They found that, despite being matched for reading ability, dyslexics tended across studies to perform worse on a variety o f experimental measures, including nonword reading, which is one of the main measurements of phonological skill. Rack et al. (1992) replicated this finding in their data set, and concluded that group differences were not simply a result o f differing reading levels, but were possibly due to a more profound cognitive deficit. I Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Although it has been well established that dyslexics show deficits in phonological skill, we do not know much about the nature o f this phonological deficit. One question addresses the extent to which the nature and severity o f the phonological deficit changes over time. Given that dyslexia is a developmental disorder, it is surprising that the majority o f studies comparing dyslexic and normal readers have not followed the development o f dyslexics over time. This type of longitudinal design allows one to study how dyslexics’ cognitive deficits change, rather than providing a snapshot of the problems experienced by dyslexic children at one time. At the initial testing period, most dyslexics will tend to show a phonological deficit, but a number o f possibilities present themselves as these dyslexics are followed over time. The phonological deficits can show improvement over time, maintenance over time, or even a worsening over time, due to a variety o f environmental factors. The present study utilizes a longitudinal design to adequately look at dyslexics’ performance over time. In one o f the few longitudinal studies conducted on dyslexics, LaBuda & DeFries (1989) compared dyslexic children to chronological age (CA) matched children over a period o f five years. They determined that dyslexic children did indeed fall behind the CA comparison group on reading related measures. During this time, the dyslexics’ deficits increased from 15 months to 3-4 years below the CA matched children. Dyslexics’ progress over time was related to full scale IQ and severity of initial deficit. However, La Buda & DeFries’ (1989) most interesting finding was that readers with more severe reading disabilities made better progress on reading-related tasks, including word recognition, reading compensation, and spelling. For all of their strengths, LaBuda and DeFries’ (1989) method of comparing dyslexic children to CA matched children is ~ > Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. inherently confounded because by definition, dyslexics are at a lower level on measures o f reading and related abilities. Stanovich, Nathan and Zolman (1988) conducted a study incorporating a reading level (RL) matched group of younger normal readers. Stanovich et al. (1988) reported poor progress over 2 years in a group o f less skilled readers when compared with a RL group. At follow up, the poor readers had made only 1.5 grades progress in reading, whereas the reading level matched children had improved by 2.8 grades. It is interesting to note that the cognitive profile o f the poor readers was similar to that o f the reading level matched children on tests tapping receptive vocabulary, memory, articulation speed, rhyming ability, and reading strategies. These results also suggested that the poor readers experienced a fairly broad delay in their reading development. However, since Stanovich et al (1988) used a sample of poor readers who had below average general cognitive ability, their results may not generalize to the dyslexic population. A different picture emerged from a study by Manis, Custodio, and Szezulski (1993) in which the authors used a more traditionally defined dyslexic sample. The dyslexics (age 9-15 years) were followed over two years. At time 1, the dyslexic children were reading at the second-grade reading level as measured by the Word Identification sub-scale o f the Woodcock Reading Mastery Test (Woodcock, 1987). When compared to RL normal readers, they performed more poorly on tests o f phonological and orthographic processing, including tests of phoneme deletion, pseudoword pronunciation, verification o f word and homophone spellings (yes) trials, and irregular word spelling. The dyslexic children performed at the same level as reading-age matched children only in two-syllable pseudoword pronunciation, pseudoword spelling, and homonym 3 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. verification. At time 2, the dyslexic readers had improved significantly in all o f the experimental tasks, with the exception of irregular word spelling. Their performance was compared with that of a new reading level comparison group that was reassembled at time 2, and that was composed of third and fourth graders. The dyslexic children had progressed in word recognition at a normal rate during the 2-year period, making an average gain o f 2.2 grade points on the Word Identification Sub-scale o f the Woodcock Reading Mastery Test. Although they had made these gains, they were relatively more impaired than the reading age-matched children in decoding nonwords that did not have orthographic neighbors, in phoneme deletion, and in irregular word spelling. In other words, the dyslexics failed to show significant “catch up” in any comparable skills relative to age- and reading-level-matched normal readers. (Manis et al., 1993), and actually fell behind, in particularly on phonological tasks. These results point to a persistent phonological deficit among dyslexic children. Snowling, Goulandris, & Defty (1996) compared dyslexics (age 7 to 12 years) to both “longitudinal” RL comparison group (matched at time 1 in word recognition and followed for two years) and a standard RL group which was re-selected at time 2 to match dyslexics’ word recognition skills. Snowling et al. (1996) found significant differences in phonological processing favoring the RL groups. Although the dyslexic children were reading at the same level as the standard RL group and in a similar manner, as suggested by their pronunciation o f words varying in spelling-sound regularity, they still showed nonword reading impairments. This finding was replicated across both of Snowling et al’s (1996) data sets. In addition, the dyslexic children showed impairments in verbal repetition. However, they performed as well as the standard RL group at time 2 4 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. on metaphonological tasks requiring the processing o f rhyme and at a similar level, though not for their age, on the WISC-R digit span. To summarize Snowling et al’s (1996) findings, the dyslexic children made poor progress over the two years, and in comparison with RL matched children at Time 2, showed specific difficulties in nonword reading and nonword repetition, as well as made more spelling errors. These results lend further support to the notion that the phonological deficits become more distinct with development. A shortcoming of this study and the Manis et al (1993) study, was the wide age range of the dyslexic (and hence the RL) sample, increasing the chance o f confounds in reading exposure. A series of adult studies confirm that dyslexics’ phonological deficits not only persist through childhood, but also into adulthood. Bruck (1990) compared adults with a childhood diagnosis o f dyslexia, to CA and RL matched comparison groups. She found that despite normal and often above average levels of reading comprehension, the adult dyslexics had not attained normal word recognition and spelling ability compared to peers o f equivalent IQ, and continued to experience phonological processing difficulties, including problems with nonword reading and phoneme awareness (also see Bruck, 1992). These studies suggest that dyslexic children can make progress in learning to read but that they experience persisting difficulties with nonword reading and phonological awareness, even into adulthood. Despite phonological difficulties, dyslexic adults relied strongly on phonological re-coding to moderate word recognition, a strategy that is characteristic o f younger readers. Hence, Bruck (1990) used the term “arrested development” to describe the adult dyslexics in her sample. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Studies of child dyslexics also reveal a strong reliance on phonological codes in pronouncing printed words, as indexed by significant effects of spelling-sound regularity. (Bruck, 1988; Snowling et al, 1996; Stanovich & Siegel, 1994; Szeszluski & Manis, 1987) In each case, dyslexics showed regularity effects o f the same magnitude as RL-matched normal readers. These results support the notion that dyslexics use qualitatively similar processes to pronounce printed words to normal children matched on word recognition level. It is possible that a word pronunciation task encourages children and adults to rely on phonological re-coding. (Stanovich and Siegel, 1994) Previous studies that measured the extent to which children rely on phonological re-coding, required subjects to answer orally. This is a flaw in the design, because readers do not generally read words aloud. Hence, it is important to test whether children use phonological mediation in silent reading, which is an activity that more closely resembles the normal reading environment. The present studies utilize a silent reading task called the semantic categorization task, developed by Van Orden (1987). Subjects are required to respond as to whether or not items are exemplars of a specific semantic category. In the present studies, we showed the subjects the sentence “Is this a fruit? ” and gave them a target word. Subjects had to press a button to respond either “yes” or “no." There were four types of items that could be presented: actual exemplars (e.g., pear), alternative exemplars (peach), homonym foils (pair) and visual foils (peer). Subjects had to respond “yes” to alternative exemplars and “no” to foils. The task measures the extent to which the subjects rely on phonological re-coding when reading words for meaning. Children who are unable to reject homonym foils display evidence of relying on phonological re- 6 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. coding (e.g., accept pair because it sounds exactly like pear). The visual foils were designed to share as many letters with the actual word (pear) as did the homonym foils, to control for guessing based on visual similarity. Sprenger-Charolles et al. (1994) found that young readers made more errors in homonym then visual foils, suggesting they rely on phonology in reading for meaning. The semantic categorization task has not to date been used with a dyslexic sample. This article reports the results o f a study in which a group o f 42 dyslexic children were tested at three time intervals over a two year time period. We used three comparison groups to address the issue of how phonological skills and word recognition skills develop. Not only did we compare the dyslexic children to chronological age matched children over two years, but we also compared them to a younger but normal reading level comparison group (longitudinal RL group) which was matched on word recognition level at time 1 and followed over the same two years. Because the longitudinal RL group would most likely develop at a faster rate than the dyslexics in overall reading ability would, we used a third group (the traditional or "standard” RL group) to ensure a match in word recognition skill at each age. The standard RL group was re-selected at each time period to match the present word recognition level of the dyslexic group. This method incorporates the strengths o f the longitudinal design. All of these children were tested on a variety o f tests of reading and reading-related skills to address the nature of dyslexic's phonological deficit and word recognition skill. We predict that the dyslexics will continue to demonstrate a persistent deficit in measures related to phonological skill. Another prediction concerns dyslexic's performance in word recognition. Dyslexics are predicted to demonstrate similar performance to that of 7 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the standard reading level group on both types of word recognition measures (both reading aloud and silently). Method: Subjects: A total o f 162 subjects were selected for this study from two cohorts totaling 230 children who were participating in a longitudinal study. Subjects who were not utilized in the present analyses generally had Wechsler Vocabulary scores that were too low, or Woodcock Word Identification scores that did not fit the range proscribed for a particular group. Data from 1996, 1997, and 1998, will be reported in this paper. Normally achieving readers in grades one, two, and three, and dyslexics in grade three were nominated by teachers at the outset of the study. After obtaining consent, screening criteria were applied. Children were not included in the study if they were rated by school staff as having limited English proficiency. These ratings were based on interviews with the child by a speech-language therapist. Other exclusionary criteria were severe cognitive or neurological impairments, severe hearing loss, or visual impairments, based on school records. The dyslexic group consisted of 42 children followed from grade 3 to grade 5. At the initial time o f testing, in grade 3, the dyslexics had a mean age of 8;5 (range: 7;4 to 9;1). The group was more diverse than a typical dyslexic sample, as it included some children with below average expressive language skills. The criteria for inclusion in the group was a scaled score on the WISC-III Vocabulary test (Wechsler, 1991) o f 4 or 8 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. higher, and a score below the 26th percentile on the Woodcock Reading Mastery Test- Word Identification. (Woodcock, 1987) In addition, to be included in the present analyses, each dyslexic child had a to have a standardized score above 85 on EITHER the time 3 Woodcock-Johnson Visual Closure task, OR a time 3 Peabody Picture Vocabulary Test- Revised (PPVT-R). A group o f 44 chronological age matched control children were tested in grade3 as well. All were reading at or above the 35th percentile on the Woodcock Word Identification test and had a Vocabulary scales score of 4 or higher. All o f the CA controls scored above the 85th percentile on either the Visual Closure task or on the Peabody Picture Vocabulary Test. Their mean age was 8;5 (range 7;9 to 9;3). A total o f 19 younger normal readers (called the longitudinal RL group) were tested at time 1 and then followed over the next two testing periods to chart their development over time. At time I all were reading at or above the 35th percentile on the Woodcock Word Identification test and had a Vocabulary scales score of 4 or higher. All of the Longitudinal RL controls had a time 3 standardized score on the Visual Closure task or the PPVT-R at or above 85. At time 1, their mean age was 6;7 (range 6; 1 to 7; 1). The next comparison groups to be discussed will be referred to as the three “standard” reading level comparison groups. These groups were selected and tested at only one time period. That is, at time 1 we selected a group to match the dyslexics, and then again at time 2 and time 3. These groups were re-selected at each time of testing to match the preset reading level o f the dyslexic children. The standard RL group at time 1 consisted of 37 first and second grade children. These children were only tested at time 1. All were reading at or above the 35th 9 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. percentile on the Woodcock Word Identification test, and had a Vocabulary scaled score o f 4 or higher. Their mean age was 6;8 (range 6; 1 to 8; 1). The standard RL group at time 2 consisted o f 34 first, second, and third grade children. These children were only tested at time 2. All were reading at or above the 35th percentile on the Woodcock Word Identification test and had a Vocabulary scales score o f 4 or higher. Their mean age was 6.7 (range 5;7 to 7;3). The standard RL group at time 3 consisted o f 32 first, second and third graders. These children were only tested at time 3. All were reading at or above the 35th percentile on the Woodcock Word Identification test and had a Vocabulary scales score o f 4 or higher. Their mean age was 6.7 (range 5.7 to 7.3). The initial identifying information for the subject groups in year 1 (mean and range for Woodcock Word Identification grade equivalent, percentile and WISC-III Vocabulary scores, and standardized scores on the PPVT-R and WJ-R Visual Closure Tests are shown in table 1.) Procedures: Subjects were run individually over five 30-minute test sessions in Fall 1996 (time 1), and for four 30-minute test sessions in Fall 1997 (time 2) and Fall 1998 (time 3). Standardized Measures Woodcock Reading Mastery Test- Word Identification - Form G (Woodcock, 1987) was administered to children as a measure o f sight word vocabulary in all three 10 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. years o f the study. The items on the test are a representative sample o f English words. About 25% o f the words fit the general criteria for exception words utilized in the present study (e.g. the pronunciation violates one or more common spelling-sound correspondences, or represents an orthographically “strange” combination of letters). Peabody Picture Vocabulary Test- Revised ('PPVT-R) - This task was used only at time 3 as another measure o f vocabulary .(Dunn & Dunn, 1981) Children were shown a page with four pictures. The experimenter read aloud a vocabulary word. The child had to name the correct picture that accurately described the target vocabulary word. The advantage o f using this task to measure vocabulary skill, is that it does not require the child to produce the target word, but requires the child to recognize its meaning. Visual Closure Task - This task was also used at time 3 as a measure of nonverbal ability. The subjects were required to look at a partially occluded object or an object with missing information in a picture and correctly name the object. Names were commonly used words familiar to children o f the target age range and alternative names were accepted (e.g. for a boat; sailboat, ship, or boat were all accepted). Items were numbered in order o f increasing difficulty. The test was stopped after 10 consecutive incorrect items. Wechsler Intelligence Scale for Children-lII Vocabulary - The Vocabulary subtest of the WISC-III (Wechsler, 1992) was administered. The task required children to define a series o f words as best they could. Clinical Evaluation o f Language Fundamentals (CELF) Recalling Sentences - The Recalling Sentences subtest o f the CELF (Semel, Wiig and Secord, 1995) was administered in year 2 and again in year 3. Children listened to tape recorded sentences 1 1 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. and repeated them back. Sentences gradually increased in length and grammatical complexity as well as word length and difficulty. Experimental Measures Pseudoword Reading - A list o f 70 pseudowords was created for the study. The items ranged from simple CVC patterns (nug) to patterns with two or more letter clusters (e.g., chome. sendee). Some two-syllable items were included as well (e.g., stining, namsion). The items were Ordered in difficulty based on pilot data. Children read the items aloud, and testing was discontinued when children made ten consecutive mistakes. The internal consistency reliability (Cronbach;s alpha) was 0.96 for year 1. The task was administered in year 2 and in year 3 with a slight re-ordering of items, and testing was discontinued when the child made six consecutive errors. Pilot data indicated that very few children pronounced any items correct above the cut-off item. Exception Word Reading - A list of 70 exception words was also created. The items were ordered in difficulty from easiest to hardest based on frequency and grade norms (Adams & Huggins 1986; Carroll, Davies, and Richman 1971) (e.g., have, people. island, yacht, silhouette), as well as our own pilot data. The task was discontinued when the children made six consecutive mistakes. Internal consistency reliability for this task was high (Cronbach’s alpha = 0.96) Again, pilot data revealed very few children pronounced items correctly above the cut-off item. Phoneme Deletion - A task o f the type devised by Bruce (1964) was administered in two parts. In part one, subjects repeated a familiar word that was spoken on a tape. 1 2 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The speaker on the tape asked the subject to repeat the word but with a specified part missing (e.g. “snow” without the /s/, “act” without the /k/). A single phoneme or a blend o f two phonemes was deleted from the beginning, middle, or end o f the word. There were 25 items and testing was discontinued if the child made five mistakes in a row. Chronbach’s alpha was 0.88. In part two, the items were all pseudowords (e.g., “kimp” without the /m/). There were 15 items and testing was discontinued if the child made five mistakes in a row. Cronbach’s alpha for the pseudowords was 0.84. Scores were combined across the two tasks for the present analyses. Three tasks were presented on Macintosh PowerBook laptop computers (Orthographic Choice. Word Pronunciation and Semantic Categorization). Responses were timed and latency was calculated for correct trials only. Orthographic Choice - This task required children to view two strings of letters displayed side by side on the screen and decide which item represented a correctly spelled word. They pressed a button to indicate which side o f the screen had the correctly spelled item. Half of the items contained at least one exception word (sponge spunge) and half contained two regular words (sheap sheep). All of the foil items were identical phonologically to the correct exemplar items if coded correctly and hence the child could not rely on phonological recoding of the items to recognize the answer. This task is thought to tap the ability to recognize the specific spellings of words, (e.g. Olson et al., 1989; Stanovich & Siegel, 1994) There were 48 items in all. Six practice trials were given. Chronbach’s alpha for the accuracy scores was equal to 0.75. somewhat lower than the other tasks. The task was adapted from a similar measure used by Olson et al. (1989). 13 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Word Pronunciation - For this task, children saw a series o f regular words, and exception words o f both high and low frequency, as well as nonwords. The words were read individually when they appeared on the computer screen, and subjects were asked to respond into a microphone with the correct word or nonword pronunciation. Responses were timed and latency was calculated for correct trials only. The experimenter pressed a button to indicate whether the response was correct or not. Semantic Categorization - For this task, children saw a printed category displayed on the screen (e.g., “a fruit”, “part o f your body”). The category was read aloud by the experimenter, disappeared from the screen, and was replaced a second later by a word. They decided if the item was a member o f the category or not and pressed a button. There are four types o f items that could be presented: actual exemplars (e.g., pear), alternative exemplars (peach), homonym foils (pair) and visual foils (peer). Only the last three types of items were shown. Subjects had to respond “yes” to alternative exemplars and “no” to foils. The task measures the extent to which the subjects rely on phonological recoding when reading words for meaning. Children who are unable to reject homonym foils display evidence o f relying on phonological recoding (e.g., accept pair because it sounds exactly like pear). The visual foils were designed to share as many letters with the actual word (pear) as did the homonym foils, to control for guessing based on visual similarity. There were sixteen categories and 40 items in all displayed (16 alternative exemplars, and 12 each o f homonym foils and visual foils). There were 6 practice trials. Cronbach's alpha for the accuracy data was 0.58. The task is adapted from a similar measure developed by Van Orden for adults (Van Orden, 1987) and by Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Sprenger-Charolles, Siegel, and Bechannec (1998) for children. The stimuli are listed in the Appendix. Results: Correlations between tasks at time 1; Word identification scores were significantly correlated with exception word reading for both the dyslexics (.84, p<.01), and the CA group (.62, p<.01). Orthographic choice was correlated with word identification only for the CA group (.36, p<.05) See table 2 for a complete chart of relevant variables. Nonword reading was significantly correlated with phoneme deletion for both the dyslexic group (.41, p <.01) and the CA group (.58, p <.01). It is interesting to note that measures o f phoneme deletion and nonword reading were not significantly correlated with other measures o f language (WISC-III vocabulary, nor to the CELF Recalling Sentences.) For both the dyslexic group and the CA group, phoneme deletion was not significantly related to Exception Word Reading, Orthographic Choice, or to any o f the semantic measures (homonym foil task, visual foil or target item accuracy.) This served to demonstrate that phonological awareness was a unique and different skill set in comparison to other measures of language or to measures o f word recognition. Nonword reading, however, was correlated with most o f the word recognition measures (see table 2 for a detailed description of time 1 correlation variables.) Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Correlations between tasks at time 3; Like time I measures, word identification scores were strongly significantly correlated with exception word reading for both the dyslexics (.74, p<.0l), and the CA group (.53, p<.01) at time 3. Orthographic choice was correlated with word identification only for the dyslexic group (.31, p<.05) See table 3 for a complete chart of relevant variables. Nonword reading was significantly correlated with phoneme deletion for both the dyslexic group (.41, p <.01) and the CA group (.58, p <.01). Similar to time 1 correlations, it is interesting to note that measures o f phoneme deletion and nonword reading were not significantly correlated with other measures o f language (PPVT-R, nor to the CELF Recalling Sentences.) Measures o f exception word reading were correlated to phoneme deletion for dyslexics only (.33, p<.05) at time 3. For both the dyslexic group and the CA group, phoneme deletion was not significantly related to Orthographic Choice , or to any o f the semantic measures (homophone foil task, visual foil or target item accuracy.) This demonstrated that phonological awareness was a unique and different skill set in comparison to other measures o f language or to measures o f word recognition. Similar to time 1, nonword reading, however, was correlated with most of the word recognition measures. Group differences in word recognition and orthographic skill Longitudinal Data. Regarding the Woodcock Word Identification data, our longitudinal data shows significant main effects of time, F(2, 204) = 420.00, p<.001, and a significant 16 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. interaction o f time by group F (2 ,204) = 31.36, p<.001, and a significant main effect of group F(2, 102) = 112.29, p<.001. Post hoes indicate that at time 1 and time 2. the dyslexics’ performance was similar to the longitudinal comparison group, but both groups performance was significantly behind the CA group on Woodcock measures (Tukey p<.01 for both groups). At time 3, the dyslexic group fell significantly behind both the RL-Longitudinal group as well as the CA group (Tukey p<.01) These results indicate that over time, the dyslexics started out performing similarly to the RL group, but then fell significantly behind the younger normals over time. These results confirm the LaBuda and DeFries, (1989) and Snowling et al. (1996) findings that dyslexics tend to fall further behind as time progresses, even when matched initially in word recognition. See table 4 for a complete chart o f relevant variables. Like the longitudinal results from the Woodcock Word Identification scores, the longitudinal data for the Orthographic choice task showed significant main effects of time, F(2, 178) = 113.3, p<.001. a significant interaction o f time by group F(4, 178) = 6.8 p<.001, and a significant main effect of group F(2, 89) = 19.8, p< 001. These results indicated that at time 1 and time 2 the dyslexics performed similarly to the longitudinal reading comparison group. By time 3, the dyslexics fell significantly behind both the reading group and the CA group (both Tukey and LSD post hoc p<.01). Also at time 3, the longitudinal RL group’s word identification skills became on par with the CA group. These results indicated that over time, dyslexics gradually fell behind on orthographic skills, whereas their younger counterparts caught up and passed easily through the initial reading stages. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The longitudinal data for the Exception word reading task showed significant main effects of time, F (2,204) = 748.57, p<.001, a significant interaction o f time by group F (4 ,204) = 39.2 p<.001, and a significant main effect of group F(2, 102) = 125.1, p<.001. These results indicated that at time 1 and time 2 the dyslexics’ performance was matched to the longitudinal reading group. By time 3, the dyslexics fell significantly behind both the reading level group and the CA group (both Tukey and LSD post hoc pc.OOl). The longitudinal exception word reading results at time 3 differed from the results for orthographic task at time 3. Instead o f the reading group catching up with the CA group, these results indicated that over time, the reading level group also fell behind the CA group. These three measures can be taken together to demonstrate that the dyslexics were not catching up to their same age normal readers. These results also point to the fact that the dyslexics were not demonstrating a temporary developmental lag, (where they would perform similarly to the younger normal readers) but the dyslexics actually fell further behind both the younger normal readers, and the same age readers. These results indicate that dyslexics have persistent deficits in word recognition and orthographic skill, confirming LaBuda and DeFries (1989), as well as Snowling et al. (1996). Standard Reading Level Group Comparison. The following deals with the re-selected RL group at each testing period on measures o f Woodcock Regularity, Orthographic skill, and Exception word reading measurements. The following results demonstrated that once dyslexics’ performance was matched on measures o f initial reading level, they also performed similarly on our measures o f exception word reading and o f orthographic skill. 18 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Evidence indicated that the dyslexics performed similarly to the re-selected reading group at each time period on measures o f Woodcock word recognition (Time 1 F(2, 122) = 140.23, p<.001; Time 2 F (2 ,119) = 95.75, p ,.001; Time 3 F(2, 122) = 54.79, p<.001), Orthographic word reading, (Time 1 F(2, 113) = 19.75, p<.001; Time 2 F(2, 110) = 16.38, p ,.001; Time 3 F (2 ,120) = 14.60, p<.001), and Exception word reading, (Time 1 F(2, 122) = 133.40, p<.001; Time 2 F(2, 119) = 84.21, p ,.001; and Time 3 F(2, 122) = 56.65, p<.001). Post hoes for all three tasks indicated that for each time of testing, the dyslexics' performance stayed similar to the standard RL group, but both the dyslexics and the standard RL group performed significantly lower than the CA group (Tukey p<.01 for all testing times.) These results supported the notion that once the dyslexics were compared to a re-selected reading level group, they performed similarly on measures o f word recognition and orthographic skill. These results also demonstrate that our process of finding a reading level group that was matched at each testing period to the dyslexics was a valid process. In summary, we compared dyslexic subjects to three groups on a set of measures o f orthographic skill, and word identification. For all three measures used to address word identification and orthographic skill, by time 3 the dyslexics’ progress was significantly worse than both the longitudinal RL group and the CA group. When we use the standard reading level group, these effects got washed away and point to a remaining developmental lag when compared to the CA group. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Do the dyslexics* phonological deficits become more distinct over time? Another research question addresses the hallmark set o f skills that traditionally define dyslexics, and asks if they become more distinctive over time. Results from the two main measures o f phonological deficits; nonword reading, and phonemic awareness, will each be presented longitudinally, and then presented using the standard RL control group as another comparison. See table 5 for a detailed description of relevant variables. Longitudinal Data. The longitudinal data for the nonword reading task showed significant main effects of time, F(2, 204) = 190.7, p<.01, a significant interaction o f time by group F(4, 204) = 8.06 p<.01, and a significant main effect o f group F(2, 102) = 63.9, p<.01. These results indicated that at time 1 the dyslexics performed similar to the reading level groups. At time 2, and then again at time 3. the dyslexics fell significantly behind both the reading level group and the CA group (both Tukey and LSD post hoc p<.01). These results demonstrated a persistent phonological deficit. These results also indicated that phonological deficits became relatively more impaired as time passed, and replicated Snowling et al.'s (1996) findings that dyslexics' phonological deficits worsened over time. The longitudinal data for the phoneme deletion task showed significant main effects o f time, F(2, 204) = 49.1, p<.001, a significant interaction o f time by group F(4, 204) = 6.95 p<.001, and a significant main effect o f group F(2, 102) = 22.35, p<.001. Like the nonword results, these results indicate that at time 1 the dyslexics performed the same as the reading level controls. At time 2, and then again at time 3, the dyslexics fell significantly behind both the reading level group and the CA group (Time 2 LSD post Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. hoc p<.05, and for Time 3 both Tukey and LSD post hoc p<.001). These results also demonstrated a persistent phonological deficit. These results indicated that phonological deficits became relatively more impaired as time progressed. These results also replicated Snowling et al.’s (1996) findings that dyslexics’ phonological skills are improving, but not at the same rate as the longitudinal RL group and the C A group. Standard Reading Level Group Comparison. When dyslexics were matched to reading level controls at each testing period, they still rapidly fell behind on measures o f nonword reading, (Time 1 F(2, 122) = 70.30, p<.001; Time 2 F (2 ,119) = 39.26, p ,.001; and Time 3 F(2, 122) = 45.03, p<.001). At time 1, the dyslexics performed the similarly to the standard RL group according to Tukey’s post hoc test, but worse according to LSD post hoc (p<.05). For the next 2 testing periods, both types of post hoc tests indicated that the dyslexics performed significantly worse than the reading level group matched at each test period on nonword reading (Tukey and LSD p<.01 at time 2 and at time 3). To summarize the nonword reading results, both longitudinally, as well as in comparison to the re-selected RL group, dyslexic’s phonological deficits were demonstrated to persist throughout childhood. Dyslexics caught up on measures o f phoneme deletion when compared to the standard reading level group, (Time 1 F(2, 122) = 22.73, p<.001; Time 2 F(2, 119) = 84.21, p <.001; and Time 3 F(2, 122) = 56.65, p<.001). At time 1, the dyslexics and the standard reading level group performed similarly, but both performed worse than the CA group (Tukey post hoc p<.05). At time 2, the dyslexics performed significantly worse than the standard RLs only using the LSD post hoc , p<.05. By time 3, however, the 2 1 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. dyslexics and the standard RL group again performed similarly, but both performed worse than the CA group (Tukey post hoc p<.01) In summary, the longitudinal results from both the nonword reading and phoneme deletion tasks confirmed that dyslexics have persistent phonological deficits as a group. Both the nonword reading task, as well as the phoneme deletion tasks demonstrated that dyslexics’ phonological skills worsened over time. Despite similar performance in reading ability, dyslexics rapidly fell behind their longitudinal reading level counterparts. Both tasks demonstrated a persistent deficit in phonological skill. Like the results for the orthographic tasks, the phoneme deletion effects dissipated when dyslexics were compared to the standard RL group, however, the nonword reading difference did not. When we compared the dyslexics to the standard reading level group on nonword reading measures, results indicated that dyslexics’ deficits did not disappear. Do dyslexics use phonology differently than normal readers in reading printed words? We addressed dyslexics’ use o f phonology by looking how they pronounce regular and irregular words. We used both a reading aloud task, and a silent word reading task to address the issue o f whether or not dyslexic children have the same regularization errors that normal beginning readers demonstrate. To review, a '‘regularity effect” entails pronouncing regular words better than irregular words. The longitudinal RL group data for the word pronunciation task will be presented as percent differences (regular word minus irregular word) and will serve to demonstrate if similar regularity effects were characteristic o f both the dyslexics and the longitudinal RL group. The 22 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. regularity data comparing the dyslexics to the standard RL group will then be presented and analyzed in a similar manner. See table 6 for a detailed description of relevant variables. Longitudinal Data. A key question was whether the size of the regularity effect differed across groups. We compared the dyslexics to the C A group and to the longitudinal RL group on the word pronunciation task at time 2, and again at time 3. Since the word lists were different between time 2 and time 3, an overall ANOVA for the regularity difference across years could not be completed. The time 2 analyses indicated that all groups pronounced regular words better than irregular words, thereby demonstrating a “regularity effect,” F(l, 99) = 161.20, p<.001. The time 2 analyses also showed a group effect, F(2, 99) = 68.79. p<.001. Post hoes indicated that, the dyslexics mispronounced both regular and irregular words more frequently than the longitudinal RL group, and the CA group, (Tukey post hoc p<.05). There was no significant interaction of condition by group at time 2. In other words, the regularity effect for the dyslexics was similar in magnitude to both the longitudinal RL group, and to the CA group. The regularity effect for the dyslexics was 11% (74%- 63%), 17% (87%- 70%) for the longitudinal RL group, and 13% (96%- 83%) for the CA group. At time 3 we still found a significant effect o f condition, indicating that all groups pronounced regular words better than irregular words. F(l, 81) = 164.97 p<.001.). There was also a significant effect of group F(2, 81) = 40.766, p<.001. Post hoes indicated that the dyslexics mispronounced both regular and irregular words more frequently than the longitudinal RL group, and the CA group. Post hoes also indicated that now the 23 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. longitudinal RL group’s word pronunciation skills were on par with the CA group (Tukey post hoc p<.05). Here we see the longitudinal RL group catching up to the CA group, whereas the dyslexics still lag behind. Time 3 results also yielded a condition by group effect, F(2, 81) 33.880, p<.00. This interaction was due to a larger regularity effect for the dyslexics (20%) as compared to longitudinal RL group (10%) and to the CA group (5%). The time 3 results indicate that the dyslexic’s performance was different than both the longitudinal RL group, and the CA group. Not only were the dyslexics mispronouncing more words than either group, but they were also still demonstrating a larger regularity effect. Standard RL Comparison Group. ANOVA data for the word pronunciation task at time 2 showed significant main effects of condition, with all groups pronouncing regular words better than irregular words, F(l. 113)= 159.016, p<.001, and a significant effect of group F(2, 113) = 50.27 p<.001. Post hoes indicated that both the standard reading level group and the dyslexics performed similarly, but both mispronounced more regular and irregular words than the CA group (both Tukey and LSD were p<.001). At time 2, there was no significant effect o f condition by group, meaning that all three groups demonstrated similar regularity effects. The dyslexics demonstrated a regularity effect of 11%, the standard RLs showed a regularity effect of 14%, and the CA group showed a regularity effect of 13% These results demonstrated that although dyslexics’ could not pronounce as many words as their CA counterparts, they demonstrated similar regularity effects. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. The ANOVA data for the word pronunciation task at time 3 again showed similar main effects o f condition, with all groups pronouncing regular words better than irregular words, F( 1,95) = 243.396, p<.001. There was also a significant effect o f group F(2,95) = 29.725, p<.001. Post hoes again indicated that the dyslexics and the standard RLs still performed similarly, and both mispronounced more words than the CA group. Unlike the time 2 word pronunciation data, at time 3 there was a significant effect o f condition by group, F(2, 95) 28.473, p<.001. Now, the CA group demonstrated a smaller regularity effect (5%) than both the standard RLs (19%) and the dyslexics (20%). Once again, when the dyslexics were compared to the standard RL group, they performed similarly, mispronouncing more words and demonstrating a larger regularity effect than the CA group. In summary, for both time 2 and time 3, results dyslexics resembled their younger age matched counterparts in their use of spelling sound correspondences and showed a developmentally less mature pattern than the longitudinal RL and CA comparison group. These results suggest that dyslexics’ use of phonological recoding fits a developmental delay, rather than a deficit pattern. Semantic Categorization Data A key question we asked was whether dyslexics rely differently on phonological recoding when reading words for meaning. We assessed this using the semantic categorization task, and looked at the degree o f difference between homonym foils and visual foils for each group. Children who relied on phonology in this task would be unable to reject homonym foils, (e.g., the homonym foil pair for pear) but would have 25 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. less difficulty rejecting the visual foils (e.g. the visual foil peer for pear.) The longitudinal data for each condition o f the semantic categorization task (the accuracy score, the homonym foil score, and the visual foil score). Each will be presented as percent differences to determine if similar effects were characteristic of both the dyslexics and the longitudinal RL group. The data comparing the dyslexics to the standard RL group will then be presented and analyzed in a similar manner. Longitudinal Data. We looked at the three conditions over three test periods, using 3 groups. We conducted a 3x3x3 ANOVA to look at performance over time between the groups for the accuracy score, the homonym foil score and the visual foil score. Results indicated a significant effect of time, F(2, 178) = 69.12, p<.001. o f time by group, F (4, 178) = 4.07, p<.01, o f condition, F(2, 178) = 270.48, p<.001, o f condition by group, F(4, 178) = 4.53. p<.01, o f time by condition, F(4, 356) = 13.12, p<.001, and of time by condition by group F(8, 356) 4.04. p<.001. Post hoes regarding the accuracy score indicated that over all three testing periods, the dyslexics were less accurate than the CA group.(Tukey post hoes for time 1, time 2, and time 3 all had p<.01), but were as accurate as the longitudinal RL group. These results demonstrated that the dyslexics were able to keep pace over time with the younger readers followed over the same time period. These results also indicate that all groups understood the semantic categorization task itself. Post hoes regarding the homonym foil score minus the visual foil score over time indicated that the dyslexics always made more homonym foil errors than the CA group. The dyslexics’ homonym foil effects grew over time, (25% at time 1, 32% at time 2, and 34% at time 3), whereas 26 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. the CA group’s homonym foil effects lessened over time (41% at time 1,34% at time 2, and 22% at time 3). Post hoes also indicated that dyslexics’ performance on these tasks fell behind the longitudinal RL group as well. The overall pattern o f the longitudinal RL group’s homonym foil effects was a lessening over time (32% at time 1, 37% at time 2, and 27% at time 3). These results demonstrate that dyslexics were chronically unable to correctly reject homonym foils, indicating that they were over-relying on phonological codes to semantically categorize words. Both the longitudinal RL group, as well as the CA group demonstrated growth in their ability to correctly reject the homonym foils. The notion o f dyslexics over-relying on a faulty phonology when reading words for meaning will be explored further in the discussion section. Standard RL Comparison Group. We conducted 3 separate 3x3 ANOVAS to demonstrate condition effects, (between the accuracy score, the homonym foil score, and the visual foil score) group effects, (dyslexic group, standard RL group, and CA group) and group by condition effects at each time o f testing. To review, the standard RL comparison group was re-selected at each test period, and thus we could not factor the time variable into the present ANOVAS. At each testing period, there was a significant effect o f condition (time 1 F(2,216) = 176.30, p<.01; time 2 F(2, 224) = 239.36, p<.01; and time 3 F(2, 226) = 242.73, p<.01), and a significant effect o f group, (time 1 F(2, 108) = 28.68, p<.01; time 2 F(2, 112) = 23.03, p<.01; and time 3 F(2, 113) = 22.75, p<.01.) For only time 1, and time 3 (not time 2) there was a significant interaction o f condition and group, (time 1 F(4,216) = 5.22, p<.01, and time 3 F(4, 226) = 6.87, p<.01). Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Post hoes regarding the accuracy score indicated that over all three testing periods, the dyslexics were still less accurate than the CA group.(Tukey post hoes for time 1, and time 3 had p<.05, time 2 Tukey post hoc p<.01), but were as accurate as the standard RL group. Like the longitudinal results, these results demonstrated that the dyslexics were able to keep pace over time with the younger readers reassembled at each time period. These results also indicated that all groups understood the semantic categorization task itself. Post hoes regarding the homonym foil score minus the visual foil score for the standard RL group indicated that they performed similarly to the dyslexic group, but differently than the CA group. The homonym foil effects for the standard RL group steadily increased from time 1 to time 2 to time 3 (30% at time 1, 35% at time 2, and 39% at time 3.) This pattern was similar to the performance o f the dyslexics’ increasing homonym foil effects. As mentioned earlier, the CA group demonstrated a lessening homonym foil effect. These results once again demonstrate that when dyslexics were compared to a group of readers were resembled to match the reading profile of the dyslexic group at each time period, they demonstrated similar patterns of performance. Discussion: The results from the word recognition data (Woodcock Word Identification, orthographic choice, and exception word reading tasks) indicate that over time, the dyslexics started out performing similarly to the longitudinal RL group, but then fell significantly behind the younger normals over time. These results confirm the LaBuda and DeFries, (1989) and Snowling et al. (1996) findings that dyslexics tend to fall further 28 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. behind as time progresses, even when matched initially in word recognition skill. When the dyslexics were compared to the standard RL group however, they performed similarly. These three measures can be taken together to demonstrate that the dyslexics were not catching up to either the same age normal readers as well as the younger normal readers. These results also point to the fact that the dyslexics were not demonstrating a temporary developmental lag, but they actually fell further behind both the younger normal readers, and the same age readers. These results indicate that dyslexics have persistent deficits in word recognition and orthographic skill. To answer the question as to whether or not dyslexics’ phonological deficits become more distinct over time, the longitudinal data for both the nonword task and the phoneme deletion task indicated that phonological deficits became relatively more impaired as time passed. The results from the nonword tasks showed the dyslexics lagging further and further behind both the longitudinal RL group, as well as the CA group over time. The phoneme deletion results indicated that dyslexics initially preformed similar to the longitudinal RL group, but over time fell significantly behind both the RL and the CA group. Longitudinal results for both the nonword and the phoneme deletion tasks demonstrated a persistent phonological deficit for the dyslexic sample. Using the standard RL group, we showed that even when dyslexics were matched to similar reading level groups at each time of testing, the nonword results still indicated that dyslexics’ performance became relatively more impaired as time progressed, but the phoneme deletion results did not. The standard RL group showed a similar performance pattern when compared to the dyslexics on measures of phoneme deletion. The nonword reading results demonstrate that the dyslexics performed 29 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. chronically worse than both the longitudinal RL, the standard RL, and the CA group, and thus replicated Snowling et al.’s (1996) findings that dyslexics’ phonological skills are not improving at the same rate as other relatively stable skills. Looking at the three main longitudinal studies assessing the development o f word recognition skills (Rack et al., (1992); Snowling et al. (1996); and the present study) it is apparent that dyslexics’ phonological deficit becomes more distinct over time mainly for nonword reading, but not for phoneme deletion. Perhaps nonword reading is a more demanding cognitive task, and therefore might be a more sensitive measure o f phonological deficits. The final question we addressed in the present studies asked whether or not dyslexic children use phonology differently than normal readers in reading printed words. To address the nature of dyslexics’ use o f phonology, we used both a reading aloud task, and a silent word reading task to see if dyslexic children relied on phonological recoding to the same extent as normal readers do at various levels of reading ability. The longitudinal results for the word pronunciation task showed that dyslexics’ regularity effects increased over time, whereas the younger normal group’s and the CA group's regularity effects decreased over time. The dyslexics were persistent in relying on an immature word reading strategy, whereas young normal readers were moving away from this strategy. Using the standard RL group, results once again showed that their regularity effects increased in a similar manner to the dyslexics’ increase. Overall, the semantic categorization data yielded very similar results to the word pronunciation data. The accuracy results demonstrate that longitudinally, the dyslexic child has a similar ability to semantically categorize words than both the younger normal child as well as CA child. The longitudinal results for the homophone effects 30 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. (homophone foil - visual foil) again demonstrated that dyslexics’ homophone effects increased over time, whereas the longitudinal RL group’s, as well as the CA group’s homophone effects decreased over time. Both comparison groups were having less and less difficulty rejecting the homophone foils, whereas the dyslexics demonstrated persistent difficulties. Once again, when the dyslexics were compared to the standard RL group, they showed an increasing homophone effect that was similar to the dyslexics’ increasing homophone effects. To review, children who are unable to reject homonym foils display evidence o f relying on phonological recoding (e.g., accept pair because it sounds exactly like pear). The dyslexics in our sample demonstrated this pattern. At this point we can pose the question, as Bruck (1988) posed, why do dyslexics rely on phonology to read words, when they have phonological deficits? We attempt to answer this question in three ways. Firstly, we can explain these findings by mentioning that phonological recoding is a mandatory phase in reading. One cannot escape phonology when trying to read words. When first learning how to read, children over-rely on phonology because they have not encountered many words, and thus need to use phonology to recognize each word. As a reader develops, this over-reliance on phonology dissipates as the reader is able to recognize more and more words. In other words, the process of reading for normal readers becomes automated, whereas the process o f reading for a dyslexic reader who has a faulty phonological system does not. Another explanation for dyslexics’ over-reliance in phonology could be due to their phonics remediation instruction. In encountering any word, a dyslexic is taught to ‘‘sound it out.” A final explanation is grounded in the connectionist approach to reading. Using the connectionist approach to reading, the 31 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. dyslexic’s over-reliance on phonology is mainly due to faulty orthographic to semantic connections. In encountering a new word, if the dyslexic’s reading system is impaired, the normal route o f phonology to orthography to semantics is not automatic, and thus the dyslexic reader perseveres on the lower stages o f reading; phonology. In the present studies, the longitudinal younger normal sample passed through these early reading stages, whereas the dyslexic sample did not. This pattern has been shown throughout the present analyses. These results replicate Bruck et al.’s (1990) findings that dyslexics showed “arrested development,” in their use o f phonology. This study is an improvement o f other studies that have looked at dyslexics' performance mainly because of its large sample size. We also used a single cohort for the dyslexics as well as for the comparison groups, to factor out effects due to a wide age range. We used a variety o f comparison groups to address dyslexics’ development over two years in an adequate manner. Using this improved methodology we still demonstrated that dyslexics have persisting phonological difficulties as evidenced by their deficient performance on nonword reading. These results confirmed Rack et al (1992) and Snowling et al. (1996) who found that, despite being matched for reading ability, dyslexics tended to perform worse on a variety of experimental measures of phonological skill, including nonword reading. The present study showed a different pattern than the Manis et al. (1993) study. In the present study, the dyslexics did not have the same rate o f growth as the CA group, whereas in the Manis et al. (1993) sample the dyslexics demonstrated 2 grades o f growth in 2 years. These dyslexics were keeping pace with their CA counterparts, and developing at the same rate. Manis et al. (1993) used a highly remediated sample from a private school with a large scale reading 32 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. intervention program. With all the added academic help, it is not very surprising that the Manis et al. (1993) dyslexics were on par with the development of the chronological age matched group. The subjects in the present study, were children with reading difficulties selected from middle class public schools throughout Long Beach. The present sample was more representative, and therefore more generalizable to other public school samples. Even in the Manis et al. (1993) study, the highly remediated sample still showed nonword deficits, again demonstrating that even remediated dyslexics still show phonological deficits. Bruck (1990, 1992) confirmed these results, and also argued that dyslexics have “arrested development.” The present studies are consistent with the hypothesis that dyslexics word recognition strategies may be arrested or at least severely delayed. We demonstrated dyslexics’ arrested development in terms o f their increasing homonym effects, confirming Sprenger-Charolles et al. (1994). As evidenced, dyslexic readers have persistent phonological deficits. These results indicate that public schools need to institute programs specializing in intensive phonics remediation to best serve the needs of dyslexic students. For dyslexics with severe and persistent phonological deficits, it might also be beneficial to develop reading remediation programs that can adequately support their phonological deficits so that they can read successfully. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 33 References Adams, M .J., & Huggins, A.W.F. (1985). The growth of children’s sight word vocabulary. A quick test with educational and theoretical implications. Reading Research Quarterly, 20. 262-281. Baron, J. (1979). Orthographic and word-specific mechanisms in children’s reading of words. Child Development. 5 0 .60-72. Bruck, M. (1990). Word-recognition skills o f adults with childhood diagnoses of dyslexia. Developmental Psychology. 26, 439-454. Bruck, M. (1992). Persistence o f dyslexics’ phonological awareness deficits. Developmental Psychology. 28, 874-886. Castles, A. & Coltheart M. (1993). Varieties o f developmental dyslexia. Cognition. 47 149-180. Coltheart, M., Curtis, B., Atkins, P., & Haller, M. (1983). Models o f reading aloud: Dual route and parallel-distributed processing approaches. Psychological Review. 100. 589-608. Dunn & Dunn (1981). Peabody Picture Vocabulary Test- Revised. American Guidance Service, Minnesota. Felton, R.H., Naylor, C .E., & Wood, B E. (1990). Neuropsychological profile of adult dyslexic children. Brain and Language. 39,485-497. Fowler, A. (1991). How early phonological development might set the stage for phoneme awareness. In S. A. Brady & D. P. Shankweiler (Eds.), Phonological processes in literacy: A tribute to Isabelle Liberman (p.97-117). Hillsdale, NJ: Earlbaum. LaBuda, M.C., & DeFries, J.C. (1989). Differential prognosis o f reading-disabled children as a function o f gender, socio-economic status, IQ and severity: A longitudinal study. Reading and Writing. I , 25-36. Manis, F. R., Custodio, R., & Szeszulski, P.A.(1993). Development o f phonological and orthographic skill: A 2-year longitudinal study o f dyslexic children. Journal of Experimental Child Psychology. 5 6 .64-86. Manis, F.R., Seidenberg, M.S., Doi, L.M., McBride-Chang, C., and Peterson, A. (19%). Cognition. 5 8 .157-195. Pennington, B. F., Van Orden, G. C., Smith, S. D., Green, P. A., & Haith, M. M. (1990). Phonological processing skills and deficits in adult dyslexic children. Child Development. 61. 1753-1778. 34 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Rack, J. P., Snowling, M. J., & Olson, R. K. (1992). The nonword reading deficit in dyslexia: A review. Reading Research Quarterly. 2 7 ,29-53. Seidenberg, M., (1992). Dyslexia in a computational model of word recognition in reading. In B. P. Gough, L. C., Ehri, & R. Treiman (Eds.), Reading acquisition (pp. 243-274). Hillsdale, N. J.: Erlbaum. Seidenberg, M., & McClelland, J. (1989). A distributed, developmental model o f word recognition and naming. Psychological Review. 96, 523-568. Snowling, M. (1980). The development o f grapheme-phoneme correspondences in normal and dyslexic readers. Journal o f Child Psychology and Child Psychiatry. 32, 49-77. Snowling, M. (1991). Developmental reading disorders. Journal o f Experimental Child Psychology. 58, 160-180. Snowling, M., Defty, N., Goulandris, N. (1996). A longitudinal study of reading development in dyslexic children. Journal o f Educational Psychology. 88, 653-669. Snowling, M., & Hulme, C. (1989). A longitudinal case study of developmental phonological dyslexia. Cognitive Neuropsychology. 6* 379-401. Snowling, M., & Hulme, C. (1994). The development of phonological skills. Transactions of the Roval Society B. 346. 21-28. Snowling, M., Hulme, C.. & Goulandris, N. (1994). Word recognition in developmental dyslexia: A connectionist interpretation. Quarterly Journal of Experimental Psychology. 47A. 895-916. Snowling, M., Hulme, C., Smith, A., & Thomas. J. (1994). The effects o f phonetic similarity and word length on children’s sound categorization performance. Journal of Experimental Child Psychology. 58. 160-180. Sprenger-Charolles, L., Siegel, L.S., Bechennec, D. (1998). Phonological mediation and semantic and orthographic factors in silent reading in French. Scientific Studies of Reading. 2, 3-29. Stanovich, K. E. (1992). The development o f rhyme production and reading skills. Unpublished manuscript, University o f London, England. Stanovich, K. E. (1986a). Cognitive processes and the reading problems o f learning disabled children: Evaluating the assumptions of specificity. In J. K. Torgeson & B. Y. L. Wong (Eds.), Psvhchological and educational perspectives on learning disabilities, (pp.87-131). San Diego, CA: Academic Press. 35 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Stanovich, K. E. (1986b). Matthew effects in reading: Some consequences o f individual differences in the acquisition o f literacy. Reading Research Quarterly. 21, 360-407. Stanovich (1988). Explaining differences between the dyslexic and the garden variety poor reader: the phonological -core variable difference model. Journal of Learning Disabilities. 21. 590-612. Stanovich, K. E. (1994). Does dyslexia exist? Journal o f Child Psychology and Psychiatry. 35. 579-596. Stanovich, K. E., Nathan, R.G., & Zolman, J. E. (1988). The developmental lag hypothesis in reading: Longitudinal and matched reading level comparisons. Child Development. 59. 71-86. Stanovich, K. E., & Siegel, L. S. (1994). Phenotypic performance profile o f children with reading disabilities: A regression-based test o f the phonological-core variable- difference model. Journal o f Educational Psychology. 86, 1-30. Van Orden, G. C., Pennington. B. F., & Stone, G. C. O. (1990). Word identification and the promise o f subsymbolic psycholinguistics. Psychological Review. 97, 488-522. Wagner, R.K., & Torgesen, J.K. (1987). The nature o f phonological processing and its causal role in the acquisition o f reading skills. Psychological Bulletin. 101. 192-212. Wechsler, D. (1974). Wechsler Intelligence Scale for Children-Revised. New York: Psychological Corporation. Woodcock, R. W. (1987). Woodcock Reading Mastery Tests-Revised. Circle Pines, MN: American Guidance Service. Woodcock, R.W. & Johnson, M.B. (1989). Woodcock- Johnson Psycho-Educational Battery- R., Allen TX: DLM Teaching Resources. Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. 36 Table 1 Descriptives of Each Group at Each Test Period TIM E 1 U U t2 TIM E 3 WWPR WWPR2 WWPR3 N Mean SD Dyslexic 42 9.84 7.63 CA group 44 68.30 16.54 Longitudinal RL 19 89.47 8.14 Standard RL 38 81.47 1.41 WWGE N Mean SD Dyslexic 42 2.15 0.38 CA group 44 4.03 0.59 Longitudinal R L 19 2.10 0.33 Standard RL 38 2.22 0.39 VOCSS N Mean SD Dvslexic 42 8.45 2.29 CA group 44 10.45 2.82 Longitudinal RL 19 13.11 3.21 Standard RL 38 11.89 3.74 N Dyslexic 42 CA grouD 44 Longitudinal RL 19 Standard-BL 38 Mean SD Mean SD 12.45 9.84 10.81 9.24 61.36 16.68 49.30 22.56 86.05 13.02 79.05 12.98 89.67 10.80 80.49 14.15 WWGE2 WWGE3 Mean SD Mean SD 2.87 0.51 3.53 0.63 4.78 1.16 5.58 1.87 3.13 0.41 4.52 1.01 2.83 0.55 3.46 0.70 VOCSS2 PPVT-R3 Mean SD Mean SD 9.31 2.90 93.90 13.66 10.68 2.85 104.16 16.14 12.26 3.71 104.05 14.12 11.71 3.62 102.76 13.20 Visual Closure3 Mean SD 100.21 12.47 101.77 10.25 104.16 14.28 102.73 12.10 Page 37 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. X A X * X T 2. C M •Q x » O O o o o co in ^ o CM C M O O CM d CO co o co C M CO in c o b co x- o CM XT d o w « 2 0 3 2 O < a 1 1 1 u > o c S I CM o CO C M o o o « C O CD C M CO o co (J < « B I L < > a c K 0 3 (fl ri in P o v v ' A Page 38 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. T im e 1 Correlations Between Measures. CM M - 0 0 O o co IO CO CM CO CM 8 § s E 1 - o »- o o *- x a * CM O IO r~- o M - O CM O cn co m t t o *- o d co co o o o o o U| < LL (fl 2 CM M " * S < a X c 0 9 2 X 0 5 £ 2 CM M - C M M - M - s .o 2 . CM ■ Q M - 2 05 CM ay o o> £ CO o « in oo o o O o o o o o C O o ▼ — o CO CM CM « « r*. C M C M CO T — CN CN o o G O o o o o d o o d « CO 0 . 2 9 0 . 2 2 . 5 8 * * « « o i n 2 0 0 0 . 2 5 o o C O o - 0 11 CM o • 0 . 0 4 - 0 . 1 3 CO CN o C O o o i - 0 . 2 5 C O C M o 1 o o 0 2 7 - 0 . 3 3 « « o> CN v* CO CN « « CN e « 00 o o i n CN e m CO o <p o d o o If) o (*) o « « o * 9 ; 0 . 1 2 . 3 9 * 0 0 0 0 . 2 0 « e e " * 00 1 - 0 . 2 6 CN d 0 . 1 4 « e o o CN 05 o CM o O O « e O O o £ 8 o o o o r - ' m o o i o • 2 2 . CM ■ O M - CM o co o co CM o M C fl s 03 a U - i c U J > (J 8 o o o CM o o : o *- a C O in o o o O < < B a < > C a H 0 9 0 3 £ •« P o £ V P C L Page 39 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. T im e 3 Correlations Between TasKs. in (O 8 o c o co o o o o C O C N o CO CN m CN o cn co m x ») 2 s C N T3 N T C N 0 0 O o co o o C O C O co m C ' - CN d m CO » — o o U < o o s x a > C N N - 2 . C N T O N - fs. O o co co o d o d o o o o C O C N o N - C N o ■ o - o o a O) in C M C M © CO o d d o CN o o o c n d C A B C IL _J UJ o O N - . o X ffi c o ■ o ' ^ o o d C N C N CO o d o o rs. o co C O c o * — d c n C N d cn co C N o o co CN o CN o d cn d CN CN o CO o co CN o m o o o co c - co * 5 1 til (A .J a (A CN O o CN o o o o o o *- CO o o CO o o o o C N o d C N c o C N o o o o CD d CN O CO d s d co CN d c o o d 5 o o o d o o IS. o o e • e « a a in 0 0 in < 0 m 8 O r * . C M © in m O o CN o d C N C N o 4 e a a a e a C M in CO 0 0 <0 2 0 9 f l 09 09 2 ii C O I u O — I < 09 5 I U u u o CN « < C O co C M d in o o n C O I < * > 1 C O C 09 U U < 09 < < s H IL IL < U z > H a a a a a (A 09 09 C L O . V « CL Page 40 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Reproduced w ith permission o f th e copyright owner. Further reproduction prohibited without permission. Table 3 Time 3 Correlations Between Tasks. eoa NW3 3DHEAC3 SDVFAC3 SDTARAC3 dyslexic CA dyslexic CA dyslexic CA dyslexic CA dyslexic CA N 42 44 42 44 42 42 42 42 42 42 WWRS3 .46** .48** .65** .55** 0.12 0.14 .51** -0.06 0.23 0.27 VISCLSS3 -0.02 -0.14 0.02 .32*- -0.16 0.10 0.16 0.02 -0.23 -0.04 CELFRS3 .39* 0.29 0.28 0.02 -0.19 -0.02 0.13 0.22 -0.19 0.23 OCAC3 -0.24 -0.21 -0.10 -0.04 .40* .339* 0.24 0.29 .40* 0.25 EXC3 .33* 023 .53** .41** 0.27 0.28 .51** 0.25 0.27 .49** PD3 1.00 1.00 .74** .53** -0.21 -0.06 0.22 -0.11 -0.21 0.19 u r n .74** .53** 1.00 1.00 -0.07 0.08 .34* -0.08 -0.07 0.00 SDHEAC3 -0 21 -0.06 -0.07 0.08 1.00 1.00 0.06 0.08 -0.01 0.28 SDVEAC3 0.22 -0.11 .34* -0.08 0.06 0.08 1.00 1.00 0.18 0.18 SDTARAC3 -0.10 0.19 0.10 0.00 -0.01 0.28 0.18 0.18 1.00 1.00 PBPCTSS3 .31* 0.19 0.14 0.15 -0.30 0.29 0.04 0.03 -0.30 .36* **p<01 *p<05 Reproduced with permission o f th e copyright owner. Further reproduction prohibited without permission. PBPCTSS3 dyslexic CA N 42 42 WWRS3 0.16 0.55** VISCLSS3 0.08 0.17 CELFRS3 .37* .30* OCAC3 -0.19 0.01 EXC3 0 11 .53** PD3 .306* 0.19 NW3 0.14 0.15 SDHFAC3 -0.30 0.29 SDVFAC3 0.04 0.03 SDTARAC3 0 08 .36* PBPCTSS3 1.00 1.00 **p<01 *p<05 Table 3 Time 3 Correlations Between Tasks. Table 4 Orthographic ANOVA Data Woodcock Word Identification Time 1 Time 2 Time 3 Mean SD Mean SD Mean SD dyslexic 38.88 8.66 51.05 7.17 57.93 6.89 Longitudinal.-RL 38.11 7.40 53.95 4.56 66.68 6.16 Standard-RL 39.84 8.21 49.97 7.31 57.11 7.69 CA group 62.89 5.32 68.64 6.30 72.80 8.54 ExceDtion Word Readina Time 1 Tima2 Time 3 Mean SD Mean SD Mean SD dyslexic 20.24 7.78 32.24 8.27 42.26 8.33 RL-Long 20.21 7.13 36.16 6.55 53.00 4.16 RL-Xsect 22.35 8.39 30.74 10.08 41.97 11.22 CA group 44.64 6.80 51.34 5.79 58.48 3.82 OrthograDhic Choice Time 1 Time 2 Time 3 Mean SD Mean SD Mean SD dyslexic 0.65 0.11 0.73 0.01 0.79 0.01 RL-Long 0.63 0.01 0.73 0.01 0.86 0.01 RL-Xsect 0.63 0.13 0.71 0.01 0.78 0.01 CA group 0.77 0.01 0.82 0.01 0.86 0.01 Page 43 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Table 5 Phonologial ANOVA Data Nonword Reading Time 1 Iime.2 Time 3 Mean SD Mean SD Mean SD Dyslexic 10.50 8.81 19.50 10.17 26.76 12.55 Longitudinal RL 16.11 9.47 30.37 10.84 45.37 11.00 Standard RL 15.68 9.94 26.44 12.49 36.57 11.35 CA group 36.16 12.51 41.45 12.50 51.45 12.42 Phoneme Deletion Tim el Time 2 Time 3 Mean SD Mean SD Mean SD Dyslexic 15.50 7.22 18.07 8.00 21.67 8.42 Longitudinal RL 18.90 6.94 22.79 8.30 30.26 6.92 Standard RL 19.78 8.32 22.09 7.59 24.92 8.78 CA group 26.95 8.33 27.66 7.19 29.27 6.58 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Page 44 Dyslexics Longitudinal RL Standard RL CA group Dyslexics Longitudinal RL Standard RL CA group Table 6 Word Pronunciation Data Regular Word Irregular Word Time 2 Time 2 Mean SD 0.74 0.12 0.87 0.01 0.80 0.17 0.96 0.00 Time 3 Mean SD 0.80 0.16 0.92 0.01 0.83 0.15 0.97 0.00 Mean SD 0.63 0.11 0.70 0.01 0.66 0.13 0.83 0.01 Time 3 Mean SD 0.60 0.18 0.82 0.11 0.64 0.22 0.92 0.01 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission. Page 45 Table 7 Semantic Categorization ANOVA data Time 1 SDTARAC SDHFAC SDVFAC Mean SD Mean SD Mean SD Dyslexic 0.86 0.13 0.40 0.20 0.65 0.20 Longitudinal RL 0.87 0.12 0.44 0.23 0.76 0.16 Standard RL 0.85 0.15 0.45 0.23 0.75 0.17 CA group 0.92 0.10 0.53 0.20 0.94 0.10 Time 2 SDTARAC SDHFAC SD Y EA C Mean SD Mean SD Mean SD Dyslexic 0.85 0.10 0.44 0.18 0.76 0.18 Longitudinal RL 0.89 0.01 0.55 0.20 0.92 0.01 Standard RL 0.86 0.01 0.52 0.20 0.87 0.15 CA group 0.93 0.01 0.59 0.18 0.93 0.01 Time 3 SDTABAC SDHEAC SDVFAC Mean SD Mean SD Mean SD Dyslexic 0.88 0.01 0.55 0.17 0.89 0.12 Longitudinal RL 0.89 0.01 0.71 0.19 0.98 0.00 Standard RL 0.89 0.01 0.60 0.19 0.99 0.00 CA group 0.93 0.01 0.75 0.17 0.97 0.00 Page 46 Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Insights into the nature of phonological and surface dyslexia: Evidence from a novel word learning task

PDF

Auditory word identification in dyslexic and normally achieving readers

PDF

Determining competent/resilient outcomes across multiple domains in sexually abused and nonabused females

PDF

Emergent literacy differences in Latino and African American children: Culture or poverty?

PDF

Cognitive functioning and dementia following cancer: A Swedish twin study

PDF

Head injury and dementia: A co-twin control study of Swedish twins

PDF

Alcohol expectancies and consumption: Age and sex differences

PDF

Ethnic identity in Parsee teenagers

PDF

Adolescents' social attitudes: Genes and culture?

PDF

Communications plan for a HIV/AIDS education project

PDF

Depression in offspring following severe prenatal stress

PDF

Articulated thoughts about intentions to commit anti-gay hate crimes

PDF

Depression and suicidality in Latino adolescents: A study of acculturation and gender role beliefs

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

Ethnic identity, acculturation, self-esteem and perceived discrimination: A comparison study of Asian American adolescents

PDF

Hedonic aspects of conditioned taste aversion in rats and humans

PDF

Affection and conflict in family relationships

PDF

Expectancies for alternative behaviors predict drinking problems: A test of a cognitive reformulation of the matching law

PDF

History of depression, antidepressant treatment, and other psychiatric illness as risk factors for Alzheimer's disease in a twin sample

Asset Metadata

Creator Freedman, Laurie Beth (author)

Core Title Dyslexic's development of word recognition and phonological skill over time

School Graduate School

Degree Master of Arts

Degree Program Psychology

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag education, special,OAI-PMH Harvest,Psychology, clinical,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-c16-288013

Unique identifier UC11341295

Identifier 1409629.pdf (filename),usctheses-c16-288013 (legacy record id)

Legacy Identifier 1409629-0.pdf

Dmrecord 288013

Document Type Thesis

Rights Freedman, Laurie Beth

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