<%BANNER%>

The Effects of Phonological Awareness, Rapid-Naming, and Visual Skills on Early Elementary Students' Reading Fluency


PAGE 1

THE EFFECTS OF PHONOLOGICAL AWARENESS, RAPID-NAMING, AND VISUAL SKILLS ON EARLY ELEMENTARY STUDENTS READING FLUENCY By HOYEON KIM A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENT FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2004

PAGE 2

ACKNOWLEDGEMENTS I would like to express my sincere gratitude to all those who encouraged me go forward and helped me complete this dissertation. First of all, I would like to thank my advisor, Dr. Vivian Correa. She provided a motivating, enthusiastic, and critical atmosphere and endless support and encouragement in difficult times. She always kept an eye on the progress of my work and always was available whenever I needed her advice. Her warmth, enthusiasm, and dedication have been truly inspirational. It has been my great pleasure to conduct this dissertation under her guidance. Furthermore, I owe much gratitude to the other members of my Ph. D committee, Dr. Martha League, Dr. Holly Lane, and Dr. David Miller who have shown me the way of research. They also monitored my work and took a tremendous effort in reading my dissertation and providing me with valuable comments. They could not even realize how much I have learned from them. I would like to express my special thanks to Dr. Algina. He let me audit his SEM course and provided me a possibility to change my methodology into a more challenging method. He even took a great effort in providing input for data analysis for this study. I am deeply indebted to the department of special education for providing an opportunity to work with undergraduate students as a field advisor for their internship. Ive learned so much from this experience as a teacher and researcher. Especially, ii

PAGE 3

working under a guidance of Dr. Martha League was grateful. I specially thank her wholeheartedly for the many things she has done for me. My friends have helped me not to get lost during the painful and long journey of this dissertation emotionally. Without their warm smile and kind words, my journey wouldnt have been easy. Needless to say, I am grateful to all of my colleagues, professors, and staff members at the Department of Special Education. They always made my days in UF much brighter and easier in so many ways. Thanks a lot! I feel a deep sense of gratitude for my father and mother who have formed my vision and taught me the good things that really matter in my life. Without their faith in me, never-ending love, and great support, I wouldnt have lived such a happy life with lots of dreams. I also would like to thank my sister, brother, and relatives who have shown their concerns and love during the critical time in my life. Last and foremost, my special thanks go to my husband, Jaemin Jung, whose endless love, encouragement, and concerns enabled me to complete this work. I cant imagine my life without him. Since 1977, when we both were 8 years old, weve been together as each others best friend and husband and wife. One of the best experiences that we lived through in this period is that we both got Ph. D. degrees from the University of Florida. My journey seems to be very easy because we travel together. Jaemin, you are the best and the one who completes me. iii

PAGE 4

TABLE OF CONTENTS page ACKNOWLEDGEMENTS................................................................................................ii LIST OF TABLES............................................................................................................vii LIST OF FIGURES.........................................................................................................viii ABSTRACT.......................................................................................................................ix CHAPTER 1 INTRODUCTION TO THE PROBLEM.....................................................................1 Rationale for the Study.................................................................................................5 Importance of Early Identification and Intervention.............................................5 Ongoing Debate.....................................................................................................5 A Continuing Attention on Phonological Awareness Research............................6 New Trends of Reading-Related Vision Research................................................8 A Rising Attention on Rapid-Naming Skills.......................................................10 Problems with Prior Research.............................................................................11 Why Reading Fluency?.......................................................................................13 Scope of the Study......................................................................................................13 Delimitations.......................................................................................................14 Limitations...........................................................................................................14 Brief Definition of Terms...........................................................................................14 2 LITERATURE REVIEW...........................................................................................16 Theoretical Framework...............................................................................................17 Connectionism.....................................................................................................17 Seidenberg and McClellands Connectionist Model...........................................18 Seidenberg and McClellands Model on Orthographic Deficits.........................19 Adams Reading Process Model.........................................................................21 Ehris Theory of the Development of Word Recognition...................................23 Prealphabetic phase.............................................................................................24 Partial alphabetic phase.......................................................................................24 Full alphabetic phase...........................................................................................25 Consolidated alphabetic phase............................................................................25 Deficit Theories Involving Reading Disabilities........................................................26 iv

PAGE 5

A Core Deficit of Phonological Awareness........................................................26 Double-Deficit Hypothesis.........................................................................................28 A Visual Deficit Hypothesis................................................................................29 Temporal Visual Processing Theory...................................................................33 Components of Successful Reading...........................................................................34 Reading Fluency.........................................................................................................34 Definition.............................................................................................................34 Component..........................................................................................................35 Factors of reading dysfluency.............................................................................36 Reading Comprehension.....................................................................................37 Definition.............................................................................................................38 Level of reading comprehension.........................................................................38 Factors of reading comprehension.......................................................................38 Significant Predictors of Reading Success.................................................................39 Linguistic Predictors............................................................................................40 Phonological awareness.......................................................................................40 Rapid naming.......................................................................................................43 Vision-Related Predictors....................................................................................46 Eye movement..............................................................................................46 Visual processing.........................................................................................48 Implication for Research.............................................................................................50 Conclusions.................................................................................................................52 3 METHODS.................................................................................................................54 Introduction.................................................................................................................54 Research Questions.............................................................................................54 Hypotheses..........................................................................................................55 Methods......................................................................................................................56 Settings................................................................................................................56 Participants..........................................................................................................56 Measures..............................................................................................................58 Data Collection...........................................................................................................59 Procedure.............................................................................................................59 Test Instruments..................................................................................................59 Inter-coder Reliability.........................................................................................64 Data Analysis..............................................................................................................65 Overview of Statistical Analysis.........................................................................65 Structural Equation Modeling.............................................................................65 Overview......................................................................................................65 Assumptions of SEM analysis.............................................................................67 Analytic Procedures for SEM..............................................................................68 A Proposed model...............................................................................................68 The overall model fit...........................................................................................69 4 RESULTS...................................................................................................................72 v

PAGE 6

Descriptive and Inferential Statistics..........................................................................73 Demographic Characteristics...............................................................................73 Means and Standard Deviations on All the Observable Variables......................74 Descriptive Statistics on Percentile Scores..........................................................76 Gender Differences.....................................................................................................77 Differences Between Second and Third Graders................................................78 Differences Between Students From Schools in Mid and Low SES Neighborhoods.................................................................................................79 Correlational Analysis................................................................................................79 SEM Analysis.............................................................................................................83 Measurement Model............................................................................................83 Structural Model..................................................................................................87 Regression Analysis and Correlational Analysis with SEM...............................89 Multiple regression analysis.........................................................................89 Summary.....................................................................................................................93 5 DISCUSSION.............................................................................................................95 Introduction.................................................................................................................95 Summary of the Results..............................................................................................96 Interpretations of the findings.....................................................................................97 Rapid Naming......................................................................................................97 The Relationship between Rapid Naming and Eye Movement...........................99 The Relationship between Visual Factors and Reading Fluency........................99 Discussions on Study Limitations............................................................................101 Measurement Issues...........................................................................................101 Generalizability Issues.......................................................................................102 Implication for Future Research and Practice..........................................................103 Future Research Directions...............................................................................103 Selecting Reading Assessments.................................................................103 Effective reading intervention for subtypes of RD............................................104 Advanced technology........................................................................................106 SEM...................................................................................................................107 Implications for Practices..................................................................................109 The earlier the better..................................................................................109 Early literacy with at risk children.............................................................110 Careful observations of visual problems....................................................111 Comprehensive vision screening policy............................................................112 Conclusions...............................................................................................................113 A Comprehensive Set of Indicators of RD........................................................114 A Balanced Approach........................................................................................115 APPENDIX PARENT INFORMED CONCENT FORM ..............................................119 LIST OF REFERENCES.................................................................................................121 BIOGRAPHICAL SKETCH...........................................................................................136 vi

PAGE 7

LIST OF TABLES Table page 3-1 Research Questions..................................................................................................54 3-2 Hypotheses...............................................................................................................55 3-3 Descriptive Information for Schools........................................................................56 3-4 Demographic Data....................................................................................................57 3-5 Test Variables...........................................................................................................58 4-1 Descriptive Statistics on Demographic Characteristics...........................................73 4-2 Means and Standard Deviations on All of the Observable Variables......................74 4-3 Criteria for Percentile Cutoff....................................................................................75 4-4 Descriptive Statistics on Percentile Scores..............................................................76 4-5 ANOVA Summaries on Gender Differences...........................................................77 4-6 ANOVA Summaries on Group Differences ............................................................78 4-7 ANOVA Summaries on Group Differences.............................................................80 4-8 Correlation Matrix of Phonological Awareness, Rapid-Naming, and Visual Skills and Reading Fluency................................................................................................81 4-9 Factors for Variables in the Overall Model and Goodness of Fit Results of the Measurement Model.................................................................................................85 4-10 Results of Confirmatory Factor Analysis.................................................................86 4-11 Results of Goodness of Fit.......................................................................................88 4-12 Regression Matrix of Four Independent Constructs on Reading Fluency...............91 4-13 Correlation Matrix on Independent Latent Variables..............................................92 vii

PAGE 8

LIST OF FIGURES Figure page 2-2 Adamss Reading Process Model.............................................................................22 2-3 Adamss Reading Process Model with Breakdowns................................................37 3-1 A Proposed Model....................................................................................................70 4-1 SEM Reading Fluency Model with Correlation Coefficients..................................90 4-2 Independent Latent Variables with Correlation Coefficients...................................92 viii

PAGE 9

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE EFFECTS OF PHONOLOGICAL AWARENESS, RAPID-NAMING, AND VISUAL SKILLS ON EARLY ELEMENTARY STUDENTS READING FLUENCY By Hoyeon Kim August 2004 Chairperson: Vivian L. Correa Major Department: Special Education Reading is at the core of the educational experience and its ability is an essential key to success in our society. Unfortunately, 20% of U.S. children have serious reading difficulties. As a consequence, there have been numerous efforts to search for reliable indicators for early reading problems as well as to develop early intervention programs for young children at risk for reading difficulties. In spite of these enormous efforts on early detection and intervention, some students failed to learn to read. In addition, some of those study findings regarding factors that affect reading have revealed inconsistent outcomes. In response, this current study was designed to clarify factors that have been controversial and inconsistent. Since components of reading are quite broad, this focused only on factors related to reading fluency. Two hundred and ten early elementary students participated in this study. Twelve measures were administered in order to determine the relationships of phonological awareness, rapid-naming and visual skills on reading ix

PAGE 10

fluency. Data were analyzed in various methods including ANOVA, correlational analysis, and SEM. Results from descriptive statistics indicated that there was no significant gender difference on all the measures. However, there were significant differences between second and third graders in several measures including rapid-naming, eye movement, reversal, and reading fluency rate measures. As predicted, third graders were superior to second graders on all the measures. The results indicated that mean scores of students who go to schools in higher SES neighborhoods are higher than those of students who attend to schools in lower SES neighborhoods on all the measures. Specifically, except the rapid-naming variables, there were statistically significant mean differences between these two groups on most of the measures. The results of correlational analysis revealed that there are significant relationships between reading fluency and four constructs including phonological awareness, rapid-naming/eye-movement, reversal, and visual perception skills. Finally, SEM results showed that the suggested reading fluency model has a good fit, and measures used in this study were significant indicators of their respective latent variables. No modification of the measurement model was needed since all of the fit indices met the criteria for a good fit. Furthermore, the multiple regression analysis indicated that among the four independent constructs, the rapid-naming/eye-movement construct has the highest correlation with reading fluency. These results imply that these factors should be importantly considered to indicate young childrens reading problems. x

PAGE 11

CHAPTER 1 INTRODUCTION TO THE PROBLEM The ability to read is the gateway to knowledge. Worldwide, todays economy demands a universally higher level of literacy achievement than at any other time in the past. Many service-related and information-based jobs have been developed and increased, and these jobs requires higher level of literacy. Moreover, advanced vocational or academic training is required now for a wide variety of positions that previously high school dropouts might be able to take. Therefore, ensuring advanced literacy achievement for all students is indispensable for an economic necessity. More recently, using computers and accessing the Internet to get information have become a part of our life. Today, the ability to read thus has become a basic means to live a successful life. As children progress through school, they are expected to be able to read in order to succeed academically. However, some children experience difficulties with reading and constantly experience frustration and failure. Approximately, 40% of fourth-grade students performed below basic levels on the National Assessment of Educational Progress (NAEP, 2001). This longitudinal national assessment revealed that since 1992 the proportion of students who performed below basic levels of reading chronically maintained about the 40% rate. In particular, researchers estimated that approximately 15% to 20% of all children have severe reading difficulties (Fletcher & Lyon, 1998; Lyon & Fletcher, 2001) and need special education services due to their reading-related learning disabilities. In the field of special education, 50% of all students with disabilities 1

PAGE 12

2 are identified as having learning disabilities. Further, it is estimated that approximately 85% of students with learning disabilities have severe reading problems (Lyon & Fletcher, 2001). If these children continuously experience reading difficulties in school, frustration and failure can affect their motivation and their approach to the task of learning (Smith, Simmons, & Kameenui, 1998). Eventually, reading difficulties may also affect all aspects of a persons life. Therefore, many reading researchers and the NAEP (2001) have warned that reading failure in the schools of the United States has reached alarming proportions. Early reading failure in schools may also continue to affect Americans literacy rates. The National Adult Literacy Survey (National Center for Educational Statistics, 1999) reported that nearly one-fourth of adult Americans were functionally illiterate. With this alarming caution, Congress has finally reauthorized the Elementary and Secondary Education Act (ESEA) after a 3-year period of debate and consideration. In addition, the No Child Left Behind (NCLB) Act of 2001 was signed into law by President Bush in January 2002. Since the passage of this law, reading programs for young children have been more actively implemented. Under this legislative support, we are hoping to have more positive outcomes in the near future. Preventing the problem is the most effective in many ways. However, preventing reading disabilities is dependent on identifying those factors that contribute to childrens reading disabilities. Fortunately reading disabilities (RD) have been studied from various perspectives over the past two decades, and research has provided cumulative insight into the factors that contribute to RD. There is a general consensus in the literature that deficits in phonological processing of reading are correlated with RD. For example, a large number of research studies have indicated that a childs lack of understanding of the

PAGE 13

3 sound structure of language (phonological awareness) is one of the strong contributors of RD (Adams, 1990; Torgesen, Wagner, & Rashotte, 1994; Torgesen, Wagner, & Rashotte, 1997). Most reading researchers widely believe that phonological processing abilities are crucially related to reading and that phonological awareness is the core deficit of children of RD. However, recent research has built on this consensus to extend beyond it (Stringer & Stanovich, 2000) with a multi-dimensional view of reading. As a result, other linguistic factors have been studied beyond phonological awareness. Specifically, a rapid-naming deficit received significant consideration as one of the key indicators of RD (Wolf & Bowers, 1999). Compared to linguistic factors, there is much less agreement on other factors that influence reading skills, such as vision-related factors. However, studies on factors that impact reading are continuously presented in the research literature because there is a strong belief that reading is not only a linguistic process but also dependent upon visual input (i.e., deciphering written language) (Boden & Brodeur, 1999). In particular, many connectionist theories that rely on the combined activity of numerous simple, interlinked, neuron-like processing units, suggest that knowledge is constructed by the strengthening of connections among processing units for reading (Adams, 1990; McClelland, Rumelhart, & Hinton, 1986; Seidenberg & McClelland, 1989). For example, Adams reading process model (1990) contains a system of four processing units, including the phonological processor, the orthographic processor, the meaning processor, and the context processor. The model suggests that reading is successfully processed by the connection of these four different units. Therefore, it is understandable to examine other process units rather than focusing only on the phonological process unit. In addition,

PAGE 14

4 there is another theory, the double-deficit theory, which was recently acknowledged in the field of RD research, that proposes that developmental reading disability can be characterized with respect to two specific deficits in phonological awareness and visual naming speed (Bowers & Wolf, 1993; Wolf, 1991). The second deficit concerns deficiencies in rapidly accessing and retrieving names for visual symbols (Bowers & Swanson, 1991; Bowers & Wolf, 1993). Recently, Functional magnetic resonance imaging (fMRI) added important information about reading processes. For example, studies on visual processes highlighted that the M pathways projections are to post parietal and frontal cortex, involved in the generation of saccades and attention shift (Perry & Zeki, 2000). The results of this study suggest that there is scientific evidence on the relationship between visual processing and reading. Although reading researchers and practitioners have long believed that phonological awareness is a major facet of the problem, recent neurological findings (e.g., Perry & Zeki, 2000) on dyslexia have provided additional evidence on a visual deficit as one of the important factors of RD. At the same time, some researchers constantly suggest that there may be a certain percentage of children who have RD and visual deficits affecting reading development (Eden, Stein, Wood, & Wood, 1995; Rayner, 1998). Due to empirical research findings and scientific evidence, a multidimensional viewpoint on factors related to RD has more widely been accepted recently. Therefore, it is appropriate to investigate visual aspects of reading to ascertain whether they affect RD or at least certain types of RD. In the next section, the rationale for the current study will be provided in detail with various perspectives.

PAGE 15

5 Rationale for the Study Importance of Early Identification and Intervention Unfortunately, children who perform poorly in reading in first and second grades have a tendency to continue experiencing difficulties throughout their schooling (Blachman, 2000). A large number of studies revealed that early reading failure seriously affects future reading ability. For example, Francis, Shaywitz, Stuebing, Shaywitz, and Fletcher (1996) found that approximately 75% of the children who had difficulties in reading in the third grade would still be struggling with reading at the end of high school. As these children grow, they face increasing obstacles to reading development (Torgesen & Burgess, 1998). As a result, the importance of early identification and intervention on reading has been recognized among reading researchers. A review study estimates that the number of children who are typically identified as poor readers and served through either special education or compensatory education programs could be reduced by up to 70% through early identification and prevention (Lyon et al., 2001). Effective prevention efforts, however, depend on the effective early identification of students with reading difficulties. Hence, there is a vital need to identify key indicators of RD for young children to reduce the impact of reading difficulties and prevent the problems in the earliest stage. Ongoing Debate As a consequence, much research has been conducted in order to identify the major factors associated with reading disabilities. Overall, there are two parallel perspectives on this issue. A foremost source of debate among researchers is whether reading difficulties are the result of one or more than one type of processing problem. To date, many researchers argue strongly that most reading difficulties are a result of verbal/linguistic

PAGE 16

6 processing deficits (e.g., Mann, 1991; Velluntino, 1979). Some other researchers, however, have demonstrated evidence that reading involves a combination of multiple factors related to an individuals experiences, ability, and neurological functioning. Some of the latter researchers argue that visual abilities contribute to childrens reading abilities (e.g., Eden et al.,1995; Lovegrove, Garcia, & Nicholson,1990; Rayner, 1997; Watson, 1990 ). However, such studies of reading-related visual skills yield inconsistent findings. Some researchers were able to find that there is a significant difference in visual skills between children with RD and non-RD (e.g., Morad, Lederman, Avni, Atzmon, Azoulay, & Segal, 2002) while some were not (e.g., Velluntino, 1987). In the context of discussions of visual deficits, there is an increasing consensus on the hypothesis that these processing problems may not be general but may characterize a small group of severely affected children. It means that not all children with RD have visual processing deficit but some of children may have visually related reading problems. If all children with reading disabilities receive the same type of instruction, which is solely based on phonological awareness, the programs may not be beneficial for a group of children who have reading disabilities that are associated with other factors such as a rapid-naming deficit and a visual deficit. Therefore, we should be aware of subgroups of reading disabled children for effective early identification and intervention. A Continuing Attention on Phonological Awareness Research Over the past two decades, no area of reading research has caught as much attention as phonological awareness. The term "phonological awareness" began appearing in the research literature in the late 1970s and early 1980s (Clark & Uhry, 1995). Early research on phonological awareness recognized that most poor readers have difficulty tapping out the number of phonemes in a word as well as in adding or deleting

PAGE 17

7 a phoneme or reserving the order of phonemes in a word (Liberman, 1973; Liberman, Shankweiler, Fischer, & Carter, 1974; Lundberg, Olofsson, & Wall, 1980). Since the 1980s, there has been increasing agreement among reading researchers that problems in phonological processing are a major factor in reading difficulties. Additionally, numerous prediction studies have demonstrated that understanding the sound structure of language (phonological awareness) is a strong predictor of later reading success. For instance, Catts and his colleagues observed 604 young children longitudinally and indicated over 70% of poor readers had a history of language deficits including phonological awareness and oral language in kindergarten (Catts, Fey, Zhang, & Tomblin, 1999). During the 1990s, research demonstrating the importance of phonological awareness to the reading process grew rapidly and become widely adapted in the field of education. In response, a huge number of research studies were actively conducted. On the basis of such empirical research studies, reading intervention studies, focusing on instruction of phonological awareness, have continuously been implemented (Clark & Uhry, 1995). Based on the findings, researchers (e.g., Adams, 1990; Catts et al., 1999; Torgesen, Wagner, & Rashotte, 1994; Togesen et al., 1997) strongly agree that the problem associated with phonological awareness is a confirmed factor in reading difficulties. Recently, functional magnetic resonance imaging (fMRI) revealed the brains of dyslexics show less activation in the language centers of brain (Broca's area) during phonological awareness tasks, and researchers indicated that letter identification lights up the extrastriate cortex in the occipital lobe and that the superior temporal gyrus and parts of the middle temporal and supramarginal gyri are activated during word meaning assessment (Shaywitz, 1996). Therefore, assessing skills on phonological awareness

PAGE 18

8 should be continuously considered for improving the predictability of identifying children at risk for RD as well as developing effective early intervention programs. New Trends of Reading-Related Vision Research Compared to phonological awareness research, studies of reading-related visual skills have a long history with three eras (Rayner, 1998). More than century ago, Hinshelwood (1895) concluded that poor reading ability was probably due to impaired visual memory for words and letters (as cited in Eden et al., 1995). Since his study, research has been implemented vigorously. Recently, Rayner (1998) did an extensive review of studies focusing on reading-related visionespecially eye movements. In his review he recognized the trends of prior research and divided research studies from the late 1890s up to the early 1990s into three eras. According to his review, the first era extended from the late 1890s up until the 1920s. During this first era, the basic facts of eye movements were discovered. These included saccadic suppression (i.e., do not perceive any visual information during an eye movement), saccadic latency, and the size of the perceptual span (Rayner, 1998). Findings from a variety of studies in the first era were the foundation of eye-movement research. Contrary to the first era, the second era (from 1920s until the mid-1970s) produced little research since researchers believed they had learned almost everything on eye movements, given the technology at that time (Rayner, 1998). However, the third era (since the late 1970s) is drawing a different picture. An advanced technology led to an enormous amount of research being implemented to investigate the possibility of a relationship between visual skills and reading (Rayner, 1998). The technological improvement on measurements allowed vision research to be more accurately and more easily implemented.

PAGE 19

9 A variety of experimental research studies revealed differences between disabled and nondisabled readers on visual skills in the area of reading. However, research on relationship between visual problems and reading was criticized by reading researchers due to technical problems of the research designs. Specifically, methodological problems in the experimental approaches were detected. Along with such criticisms, some researchers believed that only language factors were involved in reading (e.g., Vellutino, 1987). Therefore, studies in this area were somewhat neglected among reading researchers. Furthermore, since 1980, increased attention on studies related to phonological awareness may have contributed to underestimating research in other aspects of reading. Despite heavy criticism and underestimation, many reading researchers revealed that visual skills play an important role in reading (e.g., Eden et al., 1993, 1995; Watson et al., 2003). Recently, methodological problems were reduced through using more advanced measurements with sufficient evidence of validity and reliability and through carefully selecting appropriate research designs. Moreover, the multidimensional viewpoint on reading is beginning to be more widely accepted among reading researchers. Most recently, Habib, a neuroscientist, offered scientific evidence on the roles of a visual deficit on RD (2000). Functional magnetic resonance imaging (fMRI) has provided new information concerning the neuroanatomical location of the system affected in RD (Eden & Zeffiro, 1998). By utilizing fMRI, researchers have found that both dyslexic and normal readers can see stationary dots on a screen but finding moving dots was more difficult to detect in dyslextic readers compared to normal readers. Furthermore, dyslexic participants show no activation of the MT (V5) visual cortex, which detects motion, during reading and seeing these moving dots while normal

PAGE 20

10 readers do have activation of the MT (V5) cortex (Eden & Jeffiro, 1996; Zeffiro & Eden, 2000). An impairment of the magnocellular (M) pathway might even affect the generalized timing mechanisms of the cerebellum (Fawcett, Nicolson, & Maclagan 2001). The visual cortex is divided into 6 different areas; V1, V2, V3, V4, MT/V5. In the visual cortex, there are two distinct pathways call Parvocellular Pathways (P Pathways) and Magnocellular Pathways (M Pathways) that carry information about what and where objects are. In particular, M pathway runs from area V1 to V2 to V3 and MT/V5 and represents information about dynamic properties such as motion and spatial relations among objects. With such new trends, reading-related vision research has been more actively undertaken. However, since the findings are still contradictory and inconclusive, there is a critical need to clarify whether visual factors are correlated to reading abilities of a certain group of children. A Rising Attention on Rapid-Naming Skills There is a growing consensus that rapid-naming skills are significantly related to reading ability (e.g., Swanson & Wolf, 1999; Wolf, 1997). Originally, research from the field of cognitive psychology provided evidence that individuals with RD have a deficit in naming speed (Wolf, 1997; Wolf & Obregn, 1997). Basically, a naming-speed deficit refers to the inability to retrieve rapidly the spoken referent for a visual stimuli (Badian, 1997). Traditionally, rapid-naming skills have been considered as a part of phonological skills (Wagner & Barker, 1994; Wagner & Torgeson, 1987). However, Wolf and her colleagues provided evidence that a naming-speed deficit is another core deficit, along with a phonological processing deficit, by showing a modest correlation between phonological awareness and rapid-naming variables. Researchers who argue for a double

PAGE 21

11 deficit theory also have concerns about two certain types of group (e.g., Wolf, 1991; Wolf & Bowers, 1999). Actually their research findings (Wolf & Bowers, 1999) suggest that there are three types of RD based on processing deficits: a group of children with mostly a phonological processing deficit, a group who have mostly a naming deficit, and a group who have both phonological and rapid-naming deficits. Obviously, the last two groups of children who have a rapid-naming deficit would not be well served if early intervention programs that are mainly based on phonological awareness skills were provided. In sum, although a growing body of research supports the importance of rapid-naming skills as a separate indicator from phonological awareness of RD, it is still inconclusive. Hence, there is a need to replicate such studies to clarify this issue. Problems with Prior Research While phonological awareness has captured researchers attention, reading-related vision studies have received less attention. As a result, there has been a lack of research, particularly in the field of education. Specifically, only a few correlation studies involving both visual skills and verbal skills in the regression models have been undertaken (e.g., Eden et al., 1993, 1995). Moreover, a lack of research has revealed inconsistent results. To illustrate, while McKinney, Short, and Feagans (1985) reported that only 7% of a sample with learning disabilities showed severe visual problems, Watson (1990) reported that 52% of a sample of students with reading disabilities exhibited some type of visual processing problems. In addition to inconsistency, the findings were conflicting and yielded mixed outcomes. For example, some researchers were able to indicate a significant difference in the visual abilities between poor readers and nondisabled readers (e.g., Watson, 1990). In contrast, others found no such differences between the two groups (e.g., Vellutino, 1987).

PAGE 22

12 Disagreement is also recognized in studies on a rapid-naming deficit. As mentioned before, a large body of evidence has accumulated in favor of the importance of a rapid-naming factor in reading development and a key indicator of reading disability (e.g., Wimmer, Mayringer & Landerl, 2000; Wolf & Bowers, 1999). However, the evidence is somewhat inconsistent, and there are gaps in the evidence. Moreover, some issues still remain. Although many studies have found phonological awareness and naming speed to have significant independent unique effects on current or later reading (e.g., Badian, 1997, 1998; Kirby & Parrila, 1999; Manis, Doi, & Bhadha, 2000; Scarborough, 1998; Torgesen, Wagner, Rashotte, Burgess & Hecht, 1997), the outcomes varied based on when assessment occurred, how many indicators were used in their models, and what other variables were used in their equations. Moreover, some other studies indicated that one or the other individual effect disappears when prior achievement, which is an autoregressor, is included in the regression model (e.g., Torgesen et al., 1997). This issue is discussed in further detail later in Chapter 2. Another problem is that the intervention studies are mostly limited to investigating the effectiveness of their certain programs (e.g., Torgeson et al, 1997). Almost all reading programs appeared to be very effective and helpful in improving childrens reading acquisition levels. However, the intervention researchers failed to take a close look at a group of students who have shown no gains despite an intensive reading support with specially designed approaches based on linguistic components. Among children who do not show any improvement, some children may suffer from other problems such as visual deficits that affect their reading. Hence, there is a need to clarify possible factors that are associated with people with no gains after reading interventions.

PAGE 23

13 Why Reading Fluency? Reading fluency encompasses the speed or rate of reading as well as the ability to read materials with expression. Meyer and Felton (1999) defined fluency as "the ability to read connected text rapidly, smoothly, effortlessly, and automatically with little conscious attention to the mechanics of reading, such as decoding" (p. 284). Children are successful with decoding when the process used to identify words is fast and nearly effortless or automatic. The concept of automaticity refers to a student's ability to recognize words rapidly with little attention required to the word's appearance. The ability to read words by sight automatically is the key to skilled reading (Ehri, 1998). Researchers have utilized reading rate for their studies since a student's reading rate is simply calculated by dividing the number of words read correctly by the total amount of reading time. Furthermore, Fuchs, Fuchs, and Maxwell (1988) found that students' fluency was more highly correlated with their scores on a standardized reading comprehension test than were oral and written retellings, question answering, or cloze tests. Shinn, Good, Knutson, Tilly, and Collins (1992) provided additional support for the use of fluency measures to monitor progress in reading. Therefore, reading fluency rate was selected as a dependent variable measure for this study. In response, this research was designed to clarify the relationship among reading fluency and phonological awareness, naming speed, and visual factors simultaneously with structural equation modeling. Scope of the Study This study was conducted within a limited scope. Therefore, there were delimitations and limitations in conducting the present study. They are described in this section.

PAGE 24

14 Delimitations This study was geographically limited to one school district, Alachua County, in north-central Florida. The 210 subjects were selected from five schools in this medium-sized district. The study was conducted with early elementary students, especially with secondand third-grade students since peoples visual skills are fully developed by ages 6 to 7. Therefore, kindergarten and first-grade students were excluded from this study. Subject selection did not include consideration for gender or ethnicity. Limitations This study was conducted with secondand thirdgrade students. Therefore, the findings of this study are generalized to older or younger students. Due to time limitations, some assessment subtests were excluded for this study. For example, a subtest of CTOPP (Comprehensive Test of Phonological Processing), segmentation, for assessing phonological awareness was not included as an observed variable. Brief Definition of Terms An understanding of terminology that was applied in this study is necessary to the interpretation of this examination. The following section defines relevant terms as they apply to this study. Phonological awareness refers to a conscious sensitivity to the sound structures of language. It includes the ability to detect, isolate, blend, segment, and otherwise manipulate the sounds in language (Adams, 1990; Snow, Burns, & Griffin, 1998; Torgesen & Bryant, 1994). Phoneme refers to the smallest unit of spoken language. For example, the word chin contains three phonemes ( /ch/ /i/ /n/ ).

PAGE 25

15 Rapid naming is the ability to recall quickly and to verbalize the name of a presented object. Naming speed is typically assessed by the Rapid Automatic Naming (RAN) Test (Denckla & Rudel 1974, 1976), which requires children to name familiar colors, pictured objects, digits, and letters (Wagner & Togeson, 1987, as cited in Manis, Seidenberg, & Doi, 1999) and subtests of CTOPP Comprehensive Test of Phonological Processing (CTOPP) (Wagner, Torgesen, & Rashotte, 1999). Orthographic processing refers to the use of printed symbols (i.e., orthographic information) to process written language (Wagner & Marker, 1994). Eye movement is one of the visual efficiency skills that include accommodation, binocular vision, and ocular mobility. Eye movement in reading-related research is involved in abilities of tracking and fixation. Visual processing skills refer to a group of visual cognitive skills used for extracting and organizing visual information from the environment and integrating this information with other sensory modalities and higher cognitive functions (Scheiman, 2002).

PAGE 26

CHAPTER 2 LITERATURE REVIEW In order to have a better understanding of reading development based on roles of cognitive processing, this section reviews three areas of the literature regarding theoretical background based on connectionist theories: key deficit theories underlying childrens reading disabilities, the components of successful reading, and significant predictors of reading success in the previous and current literature. As a theoretical framework for this study, connectionism and connectionist reading models were employed and are discussed in Chapter 2. The latter section of Chapter 2 includes four general deficit theories including core deficit of phonological processing, a double-deficit hypothesis, a visual deficit theory, and a visual temporal processing theory, identifying causes of RD in the literature to date. For early identification of reading disabilities, two significant linguistic indicatorsphonological awareness and rapid namingand two visual related indicatorseye movement and visual processingare discussed in the final section of Chapter 2 based on prior literature that includes findings of empirical studies. Currently, researchers in diverse fields as neuroscience, cognitive psychology, educational psychology, special education, and reading have influenced this topic. The present study reviewed journal articles and books from many of these disciplines. Therefore, within this literature review, the terms reading disability (RD) and dyslexia were used interchangeably to describe a person who has a specific and significant impairment in reading abilities, unexplainable by any kind of deficit in general 16

PAGE 27

17 intelligence, learning opportunity, general motivation or sensory acuity (Habib, 2000, p. 2374). The current research provides a reliable set of predictors of reading achievements focusing on both verbal and visual factors. Theoretical Framework The theoretical framework of understanding childrens reading process in this literature review section is primarily grounded in connectionist theories that were originally based on interaction models of the information processing theory. In this section, general characteristics of connectionism, Seidenberg and McClellands connectionist model for visual word recognition, and Adams reading acquisition model are briefly discussed in order to explain the connections between reading behaviors and the brains cognition and perception. In addition, Ehris theory of reading development is also explained to understand the procedures of reading developments of young children. Connectionism Simply stated, connectionism is a theory of information processing. Connectionists originally adopted the view that the basic building block of the brain is the neuron. Connectionism offers a new approach to explaining human cognitive activities. While classical cognitive theories use explicit, often logical, and rule-based theories in a serial manner, connectionist systems rely on the combined activity of numerous simple, interlinked, neuron-like processing units. In order to begin the reading process, Seidenberg and McClelland (1989) developed a connectionist model based on an interactive theory in an information-processing theory. Commonly, all connectionist models, including a Seidenberg and McClelland word recognition model, have several similar properties. They all have a set of interconnected simple processing units. The types of units include input units, output units, and hidden units. Input units accept the

PAGE 28

18 information either from the sensory channels or from other parts of the network. Output units put out the information. They either control the information directly or send those to other parts in the network. Hidden units communicate with the input and output units internally. Most of the important work in a connectionist model occurs in the units and their connections. Furthermore, there is no central control within the connections, and the processing is spread across the entire network. In fact, all units have their own activation value. However, each unit itself is meaningless and not active since the cognitive processing produces output when the connections between units occur. In particular, to understand the act of reading as connectionism, we should carefully consider Seidenberg and McClellands parallel distributed processing model. Seidenberg and McClellands Connectionist Model Seidenberg and McClelland (1989) developed a connectionist word recognition model by describing a parallel distributed processing (PDP) system consisting of four different processors: context, meaning, orthographic, and phonological. In their PDP model, they developed a general connectionist framework for thinking about how lexical knowledge is obtained, represented in the brain, and used in processing (Manis et al., 1999). Simply stated, their reasoning is that reading involves a series of associations or connections resulting in accumulated lexical knowledge (Clark & Uhry, 1995). Actually, the concept of PDP is based on the belief that intelligence emerges from the interaction of a large number of simple processing units, each sending excitatory and inhibitory signals to other units (McClelland et al., 1986, p. 10). In addition, their PDP model involves connections in both directions between context and meaning and then additional two-way connections between meaning, orthography, and phonology (see Figure 2-1).

PAGE 29

19 That is, connections between any two processors can trigger other associations in any direction and contribute to the overall reading process. In order to increase the processing capabilities inherent in networks, they include a set of hidden units in their model. Hidden units are considered important since they mediate between the pools of representational units (Seidenberg & McClelland, 1989). More interestingly, Seidenberg and McClelland used computational simulations of word learning to provide evidence of the interactive nature of the relationship between phonology and orthography. By reducing the number of hidden units, Seidenberg and McClelland were able to simulate poorer performance in fluency, naming, and visual processing, which are characteristics found in people with dyslexia. The following section of their model focuses on the role visual processing plays within their model. It provides the reason for poor reading when visual processing is not working properly in this model. Seidenberg and McClellands Model on Orthographic Deficits Seidenberg (1992) schematically illustrated the main component of the existing model. One component considers the orthographic representations. The model starts by encoding a letter string as a pattern of activation across the orthographic units (Seidenberg, 1992, p. 262). Letter recognition is a computational process. However, the process may fail to activate appropriated orthographic representation when this component fails to operate. This could occur due to abnormalities in visual processes or damage to the orthographic representations themselves (Seidenberg, 1992).

PAGE 30

20 Hidden units Hidden units Hidden units Output /mAk / Input MAKE Phonological Processor Orthographic Processor Context Processor Meaning Processor Figure 2-1. Seidenberg and McClellands Word Recognition Model The model performed as in the simulations with 100 hidden units. Actually, the normal model was configured with 200 hidden units, but for this simulation they reduced a number of hidden units for the simulation. They reported that with fewer computational resources, the model was still able to learn; however, it performed poorly on irregular words and non-words even after extensive training (Seidenberg, 1992, p.

PAGE 31

21 263). For another simulation, they damaged the input orthographic representation and found that the outcomes of two damage simulations produced similar results. Through these findings, they argued that two very different types of impairments yielded a dyslexic pattern of performance; only one of these corresponds to a visual processing deficit (Seidenberg, 1992, p. 265). Hence, Seidenberg (1992) noted that the simulations confirm that there must be more than one source causing RD. He also described that this might go some way toward explaining why only some children whose behavior fits the dyslexic profile exhibit visual processing deficits (Seidenberg, 1992, p. 265). Adams Reading Process Model With a thorough review of reading research, Adams (1990) framed her view of the reading process based on connectionist theory. Since 1990, Adamss connectionist model has been used as a foundation for understanding reading processes. Basically, Adams reading model includes four interrelated processors: orthographic, phonological, meaning, and context (see Figure 2-2). She accentuated the connectivity among four processors that are crucial for a skilled reader. Although each processor functions independently in reading, the four units should work simultaneously for successful reading.

PAGE 32

22 Context Processor Orthographic Processor Phonological Processor Meaning Processor Print Speech Figure 2-2. Adamss Reading Process Model As seen in Figure 2-1, each processor is systematically connected and works cooperatively to obtain meaning from reading. According to Adams (1990), the system begins its work when the orthographic processor obtains visual information from printed materials and delivers visual input to the meaning processor. If the printed word was recognized and connected to a readers lexical memory, the meaning processor automatically interprets the word. However, if the word is unfamiliar to a reader, the phonological processor begins to compensate for the orthographic processor. At that time, a reader simply tries to sound out the word and determines whether the word is unfamiliar or familiar (i.e., in the readers listening vocabulary). Based on the cooperation of these two processors, the activities of these two processors can connect to the meaning processor. In addition, the context processor helps a reader to understand

PAGE 33

23 context and predict future developments. Thus, the context processor reinforces the connectivity among these three processors for a reader. As a result, its work increases the processing speed of the other three units. Furthermore, the context processor helps a reader to verify and choose the real meaning of a word based on its context, which also may improve comprehension. For a skilled reader, each of the four reading processors should play a unique role but also function simultaneously, collaboratively, and automatically with little conscious effort (Adams, 1990). Ehris Theory of the Development of Word Recognition Understanding how normal reading develops is crucial to the understanding of what goes wrong during the development of reading in children with RD (Clark & Uhry, 1995). Many researchers in the reading community have proposed theories to explain how readers develop reading abilities. In particular, Ehris theory proposes the four developmental stages of word recognition by sight word. Adams (1990) mentioned that the ability to read words by sight is dependent upon the strength of the connections that have been developed between the orthographic and phonological processors. Ehris theory highlights such connections between the orthographic and phonological processors (Adams, 1990). According to Ehri and Saltmarsh (1995), readers read sight words by accessing connections that they have formed between letter in the spellings specific words and phonemes detected in pronunciation (p. 298). Eventually, her viewpoint was developed under the connectionist view (Ehri, 1992). She described that a readers ability to read progresses through a series of phases, and each phase is defined by the kind of connections that are formed between visual cues seen in print and information about a specific word stored in memory (Ehri, 1992, p. 123). The four phases are described below.

PAGE 34

24 Prealphabetic phase During this first phase, children begin to recognize words in their environment (Ehri, 1991, 1992, 1995). Since children under this phase have little knowledge about letters, they create connections between salient features of a word and the meaning of the word (Ehri, 1991, 1992, 1995). The visual cues selected by children in this stage might be part of the spelling, shapes, or colors. Simply, a large yellow arch shape logo of McDonald can be a typical example of a visual cue (Ehri, 1992). Therefore, Ehri (1991) indicated that the identification of the word occurs at the semantic rather than the phonological level, and the connections are visual and nonphonetic in this first stage of word recognition. She also described that because most connections formed between spellings and meaning are arbitrary, they are easily forgotten unless practiced frequently (Ehri, 1992, p. 125). However, to become a skilled reader, children should be able to form the connections systematically between the printed word and the meaning. In the second phase of Ehris theory the reader finally begins to use letter-sound relations. Partial alphabetic phase During this second phase, learners begin using their letter knowledge to form visual-phonetic connections between letters seen in spellings and sounds detected in pronunciation of words they are learning to read (Ehri, 1992, p. 126). Compared to the first phase where connections are arbitrary, those in the second phase are systematic. Moreover, if only one connection occurs in the first stage, more connections are formed in the second phases since a reader uses letters. Another main difference between two phases is that the primary connection in Phase 1 is formed between a spelling and its meaning, whereas the main connection in Phase 2 is between a spelling and its pronunciation (Ehri, 1992). They use more reliable and systematic information, but they

PAGE 35

25 are not yet able to use the full cipher in this stage. Therefore, this stage is referred to as the partial alphabetic phase (Ehri, 1991). Novice readers in this stage use incomplete letter-sound correspondences to access words in memory. They may use the first or last letter sounds, ignoring the medial phonemes in the work (Ehri, 1992) such as recognizing the word book by the /b/ and the /k/. However, the readers should be able to utilize the full cipher for their skillful reading. Full alphabetic phase Ehri (1992) described that cipher sight word reading refers to the process of reading sight word by setting up connections in memory between the entire sequence of letters in spelling and phonemic constituents in the words pronunciation (p. 132). In contrast to readers in Phase 2 who may use some letters of spellings, readers in the third phase can form connections out of the complete array of letters connected to the phonemic structure of the word (Ehri, 1992, p. 134). For example, in a word with four phonemes, desk, a reader at the cipher reading phase would be able to create connections out of all four letters and correspond phonemes to determine the correct pronunciation (Ehri, 1999). Children who gain the ability to utilize the cipher can become proficient readers (Gough, Juel, & Griffith, 1992). Consolidated alphabetic phase Finally, mature readers can analyze the common spelling patterns such as -ing, -cal, and com-. It means that they are able to recognize these patterns as a chunk. If children encounter such combinations of letters more often, they may have stronger concepts of such combinations and the speed of orthographic processing will be faster. Therefore, connections to new words become more efficient and occur more rapidly. As a result, readers in this final stage can improve their ability of word recognition

PAGE 36

26 enormously. Consequently, Ehris theory of word recognition implies why some children are not very skillful at reading words (Ehri, 1992). They may not have opportunities to put through the sequential processes. More specifically, the following section discusses predominant deficit theories that explain a causal role of RD. Deficit Theories Involving Reading Disabilities Different theories are often used to explain childrens reading disabilities. These theories serve as highlights of the different processing systems involved in reading. Through the huge amount of prior reading research, everyone probably agrees with the fact that all children with reading disabilities are not the same. Therefore, it is essential to know the basis of the RD and to match the program of instruction to the inefficient information processing system. Here are four general deficit theories identifying causes of RD in the literature to date. Recently, studies on process factors involving reading have been extensively conducted underlying the progress of neuroscience A Core Deficit of Phonological Awareness A consensus among reading researchers is that phonological processing is the core deficit in RD. During the last 25 years, researchers and educators have been looking for valid and reliable predictors of reading difficulties in order to prevent problems and provide effective early intervention. Moreover, based on findings from a large amount of reading research, many intervention programs for children with reading difficulties have been developed over the years. One intervention that has attracted much recent attention involves teaching phonological awareness skills to children. One important reason why educators are so interested in phonological awareness is that research indicates that it is the best predictor of early reading acquisition, better even than IQ, vocabulary, or listening comprehension (Stanovich, 1988). In general, phonological awareness often

PAGE 37

27 refers to the conscious awareness of or sensitivity to the sounds of speech apart from their meaning (Snow et al., 1998; Torgeson, 2000). Many researchers have specifically indicated that symptoms such as slow rate of reading, spelling errors, and errors on function words in reading and writing could be accounted for poor phonological skills (e.g., Aaron & Joshi, 1992). In addition, many reading researchers indicated that phonological awareness interventions have been successful. For example, Torgeson et al. (1997) reported that two types of research support the phonological deficit perspectives as the core-causing factor of reading difficulties. First, predictive research studies have determined that a measure of phonological awareness is a major predictor of later success in reading. Torgeson et al. (1997) used a set of predictors for their logistic regression model and indicated that letter-name knowledge and phonological awareness predicted, with 91% accuracy, which children would be in the bottom 10% when they become second graders. Second, intervention studies have shown that instruction in phonological awareness improve reading acquisition. Researchers also have highlighted that a group of children who were assigned to have instructions in phonological awareness have a significantly low retention rate in special classes (9%) compared to two other groups with different instructions and a group with non-treatment (25%, 30%, and 41%, respectively) (Torgeson et al., 1997). Through vigorous research during the past two decades, it has become obvious that a lack of ability in phonological awareness is the most widespread barrier to reading acquisition (Adams, 1990; Goswami, 1993; Torgeson et al., 1997). Until recently, however, researchers were unable to map cognitive functioning in the brain during

PAGE 38

28 cognitive processes. Now, researchers are able to identify regions of the brain that are activated during cognitive processes such as reading through functional magnetic resonance imaging (fMRI). Anatomical work and functional brain imaging studies support the notion of a left perisylvian dysfunction as a basis for the phonological deficit (e.g., Pugh et al., 2000; Temple et al., 2003; Shaywitz et al. 2000). The phonological theory, however, does not explain the other deficits of dyslexia. For example, there is a limitation of explaining dyslexia with the phonological theory because the visual system on reading process cannot be accounted for by phonological theory. Double-Deficit Hypothesis The double-deficit hypothesis of reading disabilities is currently popular among those who investigate the neurological components of reading disabilities. The double-deficit hypothesis was developed by Wolf and Bowers (2000) based on their research findings showing that there are only modest correlations between phonological awareness measures and Rapid Automatized Naming (RAN) performance in groups of dyslexic children. They also found that dyslexic children as a group exhibit both phonological awareness deficits and naming-speed deficits. Originally, naming speed deficits were considered by some to be part of a phonological deficit (Wagner & Torgesen, 1987). However, through such findings, Wolf and her colleagues argued that a naming speed deficit is a separate deficit from a phonological deficit and affects reading acquisition independently from phonological processing (Olson, Forsberg, & Wise, 1994). Further, they suggested that there are different degrees of reading disability and different types of processing impairment. More specifically, there are three subtypes of dyslexic children: one having a dominant phonological deficit but little naming-speed deficit, one exhibiting the naming

PAGE 39

29 speed deficit but little phonological awareness deficit, and one with both deficits (Wolf & Bowers, 1999). Obviously, the second and the third groups of children are important in this theory because the two subgroups with only a naming-speed deficit and both deficits may not be well supported if reading intervention, which is only based on phonological awareness approaches, is provided (Wolf & Bowers, 1999). Wolf and Bowers (1999) mentioned that researchers have found that about a fifth of dyslexic children mainly suffered from naming-speed deficit and not from a phonological awareness deficit. More interestingly, Lovett, Steinbach, and Frijters (2000) found that more than half of the participants suffered from both deficits. Furthermore, extensive data replicated the existence of these three subtypes of reading disabilities in several language systems (e.g., German, Dutch, and Finnish), and they provided various evidence on the existence on the subtypes based on double deficit hypothesis (e.g., Deeney, Gidney, Wolf, & Morris, 1999). It is interesting that speed of processing and semantic fluency are included in some of the recently produced tests for dyslexic children. Beyond a double deficit hypothesis, some researchers even argue for a triple deficit hypothesis. For example, Badian (1997) purported evidence for a triple deficit hypotheses implying that orthographic factors involving visual skills should also be considered. The author found that there was a significant difference between the RD group and the non-RD group on an orthography test, the Jordan Left-Right Reversal test, which is a visual perception test. A Visual Deficit Hypothesis Another hypothesis is that problems in reading acquisition are linked to deficient visual perception. More than a century ago, James Hinshelwood, an ophthalmologist practicing in Scotland, called his patients with RD congenital word blindness and

PAGE 40

30 believed that these problems were related to the left hemisphere of the brain, which controls the storage of visual memory because these children seemed to have difficulty remembering names of letters and words (Clark & Uhry, 1995). In 1928, Samuel Orton, an American neurologist, theorized that people perceive visual inputs from both right and left visual fields and relay this input as mirror images to the visual cortices (Clark & Uhry, 1995). For most people, one side of brain becomes dominant over the other and suppresses the image from the non-dominant hemisphere; however, he believed some people who poorly perform dominance, may perceive backward images. He believed this poor dominance might produce reading problems (Clark & Uhry, 1995). Therefore, from the early 20th century, vision research related to RD has been continuously undertaken. However, researchers revealed that reversal and sequencing errors did not appear in a large proportion of people with RD, and most of people with RD showed more oral reading errors. Moreover, Vellutino (1978, 1983) indicated that poorer readers are as good as nondisabled readers on copying confusable letters and words from memory. Rather, they showed problems at naming and pronouncing the items. With these findings, he concluded that visual deficits could not explain RD. Moreover, several major reviews of the evidence concluded that visual deficits are not a major cause of reading disability (Stanovich, 1982; Velluntino, 1979). Since then, numerous studies have demonstrated that understanding the sound structure of language (phonological awareness) is a strong predictor of reading success. In contrast, studies on visual processing factors have been under heavy attack by reading researchers during the last three decades. As discussed earlier, inconsistent and conflicting findings have fueled the debate and have hindered research on visual

PAGE 41

31 processing within reading research. In addition, researchers pointed out that many of the processing measures used in education lack reliability and validity (Moats & Lyon, 1993). Some even strongly stated that the case is completely closed (Stanovich, 1992). Nevertheless, many researchers recently tried to revisit this area of reading and provide reliable evidence on differences between disabled and nondisabled children in visual recognition experiments. More recently, one visual deficit theory, saccadic-suppression deficit, emerged surrounding eye movement and visual pathways. When peoples eyes view visual stimuli such as letters, words, and any stationary things, they fixate briefly (for roughly 200 to 300 ms). After viewing, it takes about 10 to 40 ms to construct an image in the brain before the next fixation. It means visual information from the saccade is suppressed. If this saccadic suppression, which insures a series of discrete images, is not achieved, the images may be overlapped (Pollatsek, 1993). Since erratic eye movement has been observed in children with RD, researchers have continuously indicated that poor readers do have characteristic difficulties on visual ability measures (e.g., Eden et al., 1995). The measurements used for recent studies have been more carefully selected, and most of them are nationally standardized instruments (e.g., Developmental Eye-Movement test or VisualMotor Integration test). Since many reading researchers have begun to have a multi-dimensional view of the reading process, the case against visual processing factors seems to have become weaker, and in recent years more studies in this area have been introduced in educational journals. Even though many reading researchers agree that a phonological factor is a major predictor of reading difficulties, some of them also believe visual processing may play an important role (Eden et al., 1993, 1995; Rayner 1998;

PAGE 42

32 Temple et al., 2000). In particular, many studies in the medical field still focus on determining the relationship between reading and visual skills. To illustrate, one study found that children with reading difficulties had a specific disability in the shifting of attention and were only able to maintain attention for a short period of time (Facoetti, Paganoni, Turatto, Marzola, & Mascetti, 2000). As a result, researchers assumed that deficits of visual attention may hinder efficient visual processing during reading (Facoetti et. al., 2000). Casco, Tressoldi, and Dellantonio (1998) also found significant differences between disabled and nondisabled readers on visual search tasks. These findings suggested that children with the lowest performance on tasks that included searching for a target letter in a set of background letters showed both a significantly slower reading rate and a higher number of visual search errors in comparison to children with higher performances on such visual search tasks. Other studies have compared visual and verbal abilities as indicators for prediction of reading disabilities. One study compared the phonological and visuo-spatial abilities of a selected group of 17 children with reading disabilities and 21 without reading disabilities (Eden et. al., 1993). Interestingly, this study recognized that there are some visual tasks such as eye movement and visual search that are equally reliable at distinguishing between good and poor readers as phonological tests (Eden et al., 1993). In sum, in recent years vision research has begun to be more widely accepted by reading researchers. Due to advanced instruments and a multi-dimensional viewpoint on the reading process, vision research has been vigorously implemented and profoundly approached. Further, such research findings suggest there may be a certain group of children who have RD and visual problems affecting reading development.

PAGE 43

33 Temporal Visual Processing Theory The ability to process and retain pieces of information presented in a rapid sequence is a vital part of learning to read. Specifically, this is related to a role of visual processing in the reading process, the speed of process. Simply, there are two systems that transmit information from the eye to brain. The sustained or parvocellular system operates during the fixation, and the transient or magnocellular system works during saccades or movements to the next fixation (Clark & Uhry, 1995) In order to explore this theory, a number of researchers conducted experiments and demonstrated that the transient system operates slowly in some poor readers in comparison with nondisabled readers (Boden & Brodeur, 1999; Livingstone, 1993; Lovegrove, 1992). For example, Boden and Brodeur (1999) hypothesized that adolescents with RD are slower at processing visual information than normal readers and thus experience greater visual persistence. Through the findings from two types of computer-based activities, the authors concluded that the adolescents with RD demonstrated deficits in visual temporal processing, which is the processing of rapidly presented visual information, due to visual persistence. Previously, Di Lollo Hanson, and McIntyre (1983) stated that children with RD show slower visual information processing and longer duration of visual persistence compared to the control group. Similarly, Richards, Samuels, Turnure, and Ysseldyke (1990) found that children with RD require longer response times on variations of visual tasks; thus, the authors concluded that children with RD demonstrate slower information processing. Based on the findings, we can assume there may be a relationship between rapid visual naming and visual processing skills. Hence, further investigation should be implemented to clarify the relationship.

PAGE 44

34 Components of Successful Reading What is reading? Reading is a comprehensive and active process in which readers shift between sources of information (what they know and what the text says), elaborate meaning and strategies, check their interpretation (revising when appropriate), and use the social context to focus their response (Walker, 1996, p. 4). Other authors defined reading as the act of simultaneously reading the lines, reading between the lines, and reading beyond the lines (Manzo & Manzo, 1993, p. 5). The first part of their definition, reading the lines, refers to the act of decoding the words in order to construct the author's basic message. The next part, reading between the lines, refers to the act of making inferences and understanding the author's implied message. Finally, reading beyond the lines involves the judging of the significance of the author's message and applying it to other areas of background and knowledge. In other words, reading is more than translating print into the spoken word (decoding). Reading is getting meaning from print. People who have not developed automatic word recognition skills may have comprehension problems because their energy is focused on identifying words rather than thinking about what they mean. Many of these children may read very slowly and often have to read things more than once to understand them. Others have automatic word recognition skills but cannot comprehend what they read. In that sense, the reading process can be divided into two areas: reading fluency and comprehension. Reading Fluency Definition Unfortunately, there are still no consensual definitions of reading fluency (Wolf & Katzir-Cohen, 1999). However, it is closely related to the subset of time-related terms and frequently related with other terms such as automaticity, speed of processing, reading

PAGE 45

35 rate/speed, and word recognition rate/proficiency. During the last two decades, research on reading fluency was actively undertaken, and many reading researchers suggested definitions of reading fluency. Through a thorough review of previous fluency literature, Wolf and Katzir-Cohen (1999) proposed their own working definition: In its beginning, reading fluency is the product of the initial development of accuracy and the subsequent development of automaticity in underlying sublecxical processes, lexical processes, and their integration in single-word reading and connected text. These include perceptual, phonological, orthographic, and morphological processes at the letter, letter-pattern, and word levels, as well as semantic and syntactic processes at the word level and connected-text level. After it is fully developed, reading fluency refers to a level of accuracy and rate where decoding is relatively effortless; where oral reading is smooth and accurate with correct prosody; and where attention can be allocated to comprehension. (Wolf & Katzir-Cohen, 2001, p. 219) Wolf and Katzir-Cohen (2001) looked at fluency not so much as an outcome but as a developmental process that is shaped and influenced by all the linguistic systems that give us knowledge about words. Component Three major types of processes have been recognized as prominent in the development of fluency: orthographic, phonological, and semantic. Interestingly, Adams (1990) added emphasis to morphological and syntactic knowledge systems. For example, Adams proposed that both orthographic and semantic information are necessary for learning morphological knowledge and that both patterns and roots (e.g., Latin and Greek roots) make up many words. Researchers emphasize the connections that link orthographic, semantic, phonological, and morphological systems. In other words, reading fluency involves every process and subskill involved in reading. Consequently, language comprehension processes and higher-level processes affecting language comprehension (the application of world knowledge, reasoning, etc.) do not become fully

PAGE 46

36 operative in comprehending text until the child has acquired reasonable fluency (Adams, 1990; Hoover & Gough, 1990; Perfetti, 1985; Stanovich, 1991; Sticht & James, 1984; Vellutino et al., 1991; Vellutino, Scanlon, & Tanzman, 1994). Factors of reading dysfluency Why do some people fail to read fluently? There must be specific factors that affect reading fluency. During the last two decades, many reading researchers have provided evidence on factors related to reading fluency (Wolf, 1991; Wolf, Bowers, & Biddel, 2000). They suggested that there is evidence that impaired readers break down across the continuum of processes. Through a thorough review of previous fluency literature, Meyer and Felton (1999) provided a summary of the existing research on explanations of dysfluency. Three major areas were identified within their review: 1. A breakdown takes place in the lower level processes. In Figure 2-3,(1) represents a breakdown in this lower level At this level, dysfluent reading results from deficits in phonological processing and /or orthographic processing systems that affect the timing and coordination of these systems. Researchers indicated that the largest contribution to slowed word-reading rate among dyslexic readers in the slowed speed of processing in perceptual stages (Farmer & Klein, 1995; Tallal, Miller, & Fitch, 1993). 2. In this level, a breakdown takes places when people make higher order semantic and phonological connections between words, meaning, and ideas.(2) represents a breakdown in connections to meaning processor. In this level, dysfluent reading is resulted from deficits that occur after perceptual identification has been completed. 3. This breakdown may involve the slowed retrieval of names, meaning, or both (German, 1992; Wolf et al., 2000). 4. A breakdown in syntactic processing can be the last factor of dysfluency.(3) represents the last level of breakdown. In this level, dysfluent readers may have deficits exhibited in a lack of prosody and rhythm in oral reading and a lack of sensitivity to prosodic cues.

PAGE 47

37 Meaning Processor Orthographic Processor Phonological Processor 2 2 3 Context Processor 1 1 Print Speech Figure 2-3. Adamss Reading Process Model with Breakdowns Reading Comprehension As mentioned previously, reading is an interactive, holistic process involving the reader, the text, and instruction (Maria, 1990). The skillful reader is able to orchestrate a complex system of skills and knowledge. Since reading is a holistic process it is difficult to divide the concepts into fluency and comprehension. To be an effective reader, a well-developed system of decoding and comprehension skills is required (Rumelhart, 1977). Readers who decode quickly and accurately coordinate this process with the processes involved in the comprehension of the text (Anderson, Heibert, Scott, & Wilkinson, 1985).

PAGE 48

38 Definition Reading comprehension is the ability to understand or get meaning from text (any type of written material). Fountas and Pinnell (1996) emphasized that reading comprehension is not something that only happens after reading. He defined reading comprehension as the thinking done before, during, and after reading. (Fountas & Pinnell, 1996, p. 323). Others researchers defined it as the interaction between the text being read and the reader's existing prior knowledge, and expectations will generate meaning and comprehension (Leu & Kinzer, 1995). Comprehension is the reason for reading and a critical component of all content learning. Level of reading comprehension Specifically, Vacca and Vacca (1996) divided reading comprehension into three levels: 1. Literal (reading the lines): People get information explicitly from the text. 2. Interpretive (reading between the lines): People put together information and use information to express opinions and form new ideas. 3. Applied (reading beyond the lines): People perceive relationships and make inferences. Factors of reading comprehension Broadly, many researchers identified three factors affecting reading comprehension: reader, text, and activity. In addition, variations in the context in which they occur play an important role. However, these factors are intertwined and do not operate independently of the others in any authentic act of reading comprehension. First, effective readers bring to the task of reading an array of capabilities and dispositions. Reader differences in such capabilities as fluency in word recognition, oral language

PAGE 49

39 ability (Pearson & Fielding, 1991; Pressley, 2000), and domain knowledge ((Dole, Duffy, Roehler, & Pearson, 1991), along with differences in such dispositions as the readers motivation, goals, and purposes, are important sources of variability in reading comprehension. Second, text is the second element affecting reading comprehension. Duke (2000) documented that children who attend schools in poor districts have many fewer texts available than do children who attend schools in richer areas; the availability of texts in homes and libraries varies similarly. Texts that treat certain social issues or that require an interpretation and appreciation of alternative perspectives may be considered inappropriate by parents from some cultural or religious groups. Text factors thus interact with reader, activity, and context in determining the difficulty of comprehension. Third, activity refers to the acts a reader engages in with a text, and it encompasses purpose, operations, and consequences. Reading comprehension can be improved with strategies when readers themselves apply across a number of different texts (National Reading Panel, 2000). For example, teachers may teach students to activate their own background knowledge, to draw inferences as they read, or to restate information in the text (Chan, Cole, & Barfett, 1987). The successful activation of two areas of readingfluency and comprehensioncompletes reading. Significant Predictors of Reading Success Many studies including longitudinal studies have been conducted to identify the predictors of current and later reading ability. Such studies yielded many important predictors of RD. Currently, phonological awareness and rapid-naming variables are very frequently involved in prediction studies. Further, a critical role of phonological and naming speed factors in classifying reading disabilities have been identified by many reading studies (e.g., Morris et al., 1998). Within the field of reading-related vision

PAGE 50

40 research, eye movement and visual processing factors are considered as key indicators of reading difficulties. Therefore, these four individual predictors are discussed further in this section. Linguistic Predictors Among various linguistic indicators, the two most frequently identified factors that are significantly correlated to later reading ability are phonological awareness and rapid-naming deficit. During the last 25 years, research has provided a greater understanding that phonological awareness is a foundation of reading acquisition. As a consequence, an assessment of phonological awareness skills always has been included in early identification of RD. Along with phonological awareness, a wider variety of recent studies has highlighted rapid-naming skills as being highly correlated with their later reading abilities (e.g., Wolf, 1999; Wolf & Bowers, 1999). Therefore, the following section provides a neutral overview of the two skills and evidence of empirical studies as key linguistic predictors of RD. Phonological awareness In the 1970s, Isabelle Liberman and Donald Shankweiler introduced the significant concept of phonological awareness to the field of reading (Liberman, 1973; Liberman et al., 1974). They argued that an alphabetic system requires that readers have awareness that words are made up of individual speech sounds (i.e., phonemes). Hence, if a child is not sufficiently aware of the speech segments in spoken words, the child may have difficulties in reading. These researchers and their colleagues studied young children and poor readers and confirmed that they were lacking in awareness of the sound structure of words (Liberman et al., 1974); that is, the children had difficulties performing activities that require one to reflect on the speech sounds in words (e.g., counting the number of

PAGE 51

41 phonemes in a word) (Liberman et al., 1974). Currently phonological awareness is the widely accepted term and generally is used as an umbrella label for all levels of conscious awareness of the sound elements in words. Simply, phonological awareness is an ability to understand a words sound structure, and it is critical for the efficient decoding of printed words and the ability to form connections between sounds and letters when spelling (Torgeson, 1997). In general, researchers indicate phonological awareness includes five levels of language skills (e.g., Chard & Dickson, 1999). According to Chard and Dickson (1999), these five levels of phonological awareness skills (rhyming, phoneme identification, blending, segmentation, and manipulation) emerge following a developmental hierarchy. The earliest developing phonological awareness skill was rhyming and is the most challenging phonological awareness skill. It involves the ability to manipulate phonemes (Chard & Dickson, 1999). More specifically, Adams (1990) further described five levels of phonological awareness in terms of abilities to (a) hear rhymes and alliteration as measured by knowledge of nursery rhymes, (b) do oddity tasks (comparing and contrasting the sounds of words for rhyme and alliteration), (c) blend and split syllables, (d) perform phonemic segmentation (such as counting out the number of phonemes in a word), and (e) perform phoneme manipulation tasks (such as adding, deleting a particular phoneme, and regenerating a word from the remainder). As explained above, there are various dimensions of phonological awareness. A tapping task, which was first used by Liberman (1973), is involved in developing childrens phonological awareness. Liberman had children use a wooden dowel to tap the number of sounds they could hear in spoken sentences and then in words (Liberman, 1973). Over the last 25 years, a huge amount of research has suggested that students who enter school

PAGE 52

42 with little knowledge of phonological awareness experience less success in reading than students who have sufficient knowledge of phonological awareness (Adams, 1990; Mann & Brady, 1988; Stanovich, 1988). Overall, the studies of phonological awareness for young children are twofold: First, the power of phonological awareness to predict childrens reading ability enables early identification of students at risk for RD (e.g., Mann, 1993; Torgeson et al., 1997; Watson et al., 2003), and, second, the effectiveness on instruction, based on phonological awareness, has been recognized when compared to other types of instruction or no instruction (OConnor, Jenkins, Leicester, & Slocum, 1993; Torgeson et al., 1997; Vellutino & Scanlon, 1987). Research firmly established the strong relationship between phonological awareness and reading acquisition. Along with phonological awareness, some intervention studies suggest that naming speed may be a critical component in successful reading, and it may be a reason why not all children respond to phonological awareness (e.g., OConnor et al., 1993; Torgeson Wagner, Simmons, & Laughton, 1990). Although phonological awareness has been considered as the strongest indicator of RD, until recently we were not able to form an image of the brain of a living person engaged in reading (Clark & Uhry, 1995). The relationship between phonological awareness and RD has been confirmed through recent studies using fMRI (e.g., Pugh et al., 2000; Temple et al., 2000; Temple et al., 2003; Shaywitz et al., 2000; Wood, Relton, Flowers, & Naylor, 1991). For example, Temple et al. (2000) examined dyslexic and normal readers (age ranged from 8-12) by using fMRI during reading. They used fMRI in order to examine neurological differences between dyslexic and normal readers during phonological and orthographic tasks of rhyming and matching visually present letters. During letter rhyming for a phonological task, both dyslexic and

PAGE 53

43 normal readers had activity in left frontal brain regions, whereas only normal readers had activity in left temporo-parietal cortex. During letter matching performance for an orthographic task, while normal readers showed activity throughout extrastriate cortex, dyslexic readers had little activity in that region. Their findings suggest that children with RD may be characterized by disruptions in the neural bases of both phonological and orthographic processes (Temple et al., 2000). Rapid naming Rapid naming is the ability to recall quickly and to verbalize the name of a presented object. Naming speed is typically assessed by Rapid Automatic Naming Test (RAN; Denckla & Rudel 1974, 1976), which requires children to name familiar colors, pictured objects, digits, and letters (Wagner & Togeson, 1987, as cited in Manis et al., 1999). Originally, Denckla and Rudel (see Clark & Uhry, 1995) found that the length of time it took to name these stimuli varied with age and reading ability. Older children were able to name quicker than younger children, and children with higher reading ability were quicker that poor readers. This hypothesis was confirmed through continuing studies. Recently, there is a growing consensus that RAN is highly correlated with reading ability (e.g., Bowers & Swanson, 1991; Wolf, 1991) and serves as an accurate, early, time-efficient predictor of reading achievement (Bowers & Wolf, 1993; Wolf, 1991). Over the last two decades, an extensive body of research indicated that rapid-naming skill is strongly related to reading development (Badian, 1993; Bowers, 1996; Wolf, Bally, & Morris, 1986; Wolf & Bowers, 1999). Unfortunately, researchers have provided mixed findings, and such findings vary greatly across studies. For example, the relationship between rapid-naming and phonological deficits in RD becomes a subject of

PAGE 54

44 controversy. More specifically, Manis and his colleagues (1999) indicated two remaining issues in rapid naming: 1. Is a rapid-naming skill is simply a part of phonological skills? 2. What processing skills are involved in rapid naming? As described previously, several researchers argued that rapid naming is correlated with reading because these two are highly related to phonological coding skills (Share, 1995; Wagner & Togeson, 1987). Notably, Torgeson et al, (1997) pointed out that one limitation of predictive studies is the autoregressive effects of prior reading skills. When they did not control the effect of prior reading skills, they were able to indicate separate contributions of phonological and rapid-naming factors. On the other hand, when the autoregressive effect was factored out, rapid naming was no longer accounted for reading outcomes, but phonological awareness was still accounted for variance in reading. In contrast, Meyer, Wood, Hart, and Felton (1998) reported somewhat opposite findings. They conducted a study with older children, Grades 5 and 8, and yielded different results. They found that rapid naming was the only variable that significantly predicted later word identification for children with RD (not for average readers) when the autoregressive variables (Grade 3 word identification, IQ, and socioeconomic status) were controlled. Further, when autoregressor was factored out, rapid naming still accounted for variance in later reading development, but phoneme segmentation did not. Researchers (Bowers & Wolf, 1993; Wolf, 1997; Wolf & Bowers, 1999) argued that even though rapid naming may rely on and share variance with some phonological tests, some aspects of rapid naming such as speed of processing and sensitivity to temporally ordered information cannot be explained by phonological processing (Manis et al., 1999). Some researchers have tried to provide evidence by examining correlations between

PAGE 55

45 phonological skills and rapid-naming skills of study participants (Felton & Brown, 1990) and identifying subgroups of children with RD (e.g., Korhonen, 1991; Lovett, 1995; Wolf & Bowers, 1999). For example, Korhonen (1991) conducted cluster analyses to identify subgroups among third graders with learning disabilities (LD). They were able to identify one subgroup of children who were characterized by their slow rapid naming. They further noted that this particular subgroup showed the least progress in reading achievement. Similarly, Lovett (1995) was also able to identify three subtypes among children with severe RD by assessing pretreatment diagnostic testing. Three subtypes were identified according to whether they had a phonological deficit only, a naming speed deficit only, or both deficits. Compared to children in a control group, all subjects, in general, made significant progress after 35 hours of intervention, which involved phonological awareness and word identification strategy trainings. However, the subgroup with a phonological deficit only made much further gains compared to the other two groups. This finding suggests that the remedial program, which is mainly based on phonological awareness, is not very effective for children who mainly have a naming speed deficit. In that sense, a rapid-naming predictor is a useful avenue of identifying children who may need additional assistance above and beyond the kinds of remedial approaches that are effective for poor readers with no naming deficit (Scarborough, 1995). Additionally, a reading-related vision research added a naming variable for the multiple regression model (Eden et al., 1995) and indicated that naming to confrontation on the Boston Naming Test proved to be significantly more difficult for the RD group than for the non-RD group.

PAGE 56

46 Vision-Related Predictors Since many studies estimated that 70% to 85% (e.g., Getman, 1981; McGurk & McDonald, 1976) of what we learn comes through vision. For example, McGurk and McDonald (1976) addressed the degree of development of how peoples visual skills significantly determine how effectively one is able to learn. According to Getman (1981), there is a significant relationship between vision and school achievement. Specifically, empirical studies indicate that 80% of all learning during a childs first 12 years is obtained through vision (Getman, 1981). One interesting study examined the link among learning disabilities, delinquency, and visual ability and found that significant numbers of juvenile offenders had one or more visual deficiencies (Johnson, Nottinghamn, Stratton, & Jaba, 1996). Further, Johnson and his colleagues (1996) reported that 97% of students with behavioral problems and learning difficulties failed at least one of the vision subtests (tracking, fusion, acuity-near, acuity-distant, convergence, hyperopia, color vision, and visual motor integration). Byrne (1992) described several concerns in the field of reading research. One important viewpoint he highlighted was the difficulties in visual perception as underlying RD may not be a dead issue (p. 2). Although the visual aspects of RD were largely ignored until recently, a wider variety of literature indicated visual and ocular motor abnormalities, that is, eye movement and a visual processing factor in subjects with dyslexia (Fisher & Weber, 1990) as risk factors of reading difficulties. Therefore, these two skills are discussed further in this section. Eye movement Specifically, visual efficiency skills include accommodation, binocular vision, and ocular mobility. Some visual efficiency skills such as eye movement, tracking, and fixation are frequently mentioned in reading-related research. Reading begins with the

PAGE 57

47 input of information through the visual-motor system. Basically, a variety of eye-related muscles are involved in eye movement. Measuring eye movement involves an examination of the number of fixations, duration of fixation, and the sequence of picture elements fixated. These eye fixations are called saccades and become more articulated as a child grows. When a preschooler looks between different objects, he or she typically uses head and neck movements rather than eye movements. By age 6 or 7, the child should be able to read a book or a board with proper eye movements and without moving his or her head or body (Schieman, 2002). Due to body or neck movement, a student who has low functioning eye-movement skills may have difficulties in keeping his or her place while reading or adding numbers in columns (Scheiman, 2002). Further, reading speed may be reduced due to slow eye movements, recurrent loss of place, or due to a small span of recognition (the width of text that is processed per fixation). Therefore, we may conclude that underachieved or incorrectly developed eye movement may be associated with reading difficulties. A study conducted with a selected group of 17 children with reading disabilities and 21 without reading disabilities found that visual tasks such as eye movement and visual search are able to differentiate good and poor readers as well as phonological tests (Eden et al., 1993). Specifically, Eden and his colleagues (1995) found some visual deficits including vertical tracking and vergence control in eye movement skills were stronger predictors than some other verbal skills. Their findings strongly support the hypothesis that some children with reading difficulties have visual problems. In summary, Rayner (1995) recapitulated the process of reading and eye movements by introducing five principal issues: (a) the span of effective vision, (b) integration of information across eye movements, (c) eye movement control (where to

PAGE 58

48 fixate), (d) eye movement control (where to move), and (e) models of eye movement control. His research confirmed that the average span of perception can be determined by computing the number of words per fixation. As a result, slow reading speed may be highly correlated with eye movement. Visual processing Visual processing skills refer to a group of visual cognitive skills used for extracting and organizing visual information from the environment and integrating this information with other sensory modalities and higher cognitive functions (Scheiman, 2002). Visual processing skills are classified as visual spatial skills, visual analysis skills, and visual motor integration skills. As mentioned earlier, Adams (1990) stated that reading is accomplished through complex and interrelated processes, one of which is the visual process. Therefore, when determining the relationships between vision and reading, tests of visual processing should be involved rather than limiting the evaluation to eye health and visual acuity. For a skilled reader, the identification of a printed word begins with a visual process that operates on the visual forms of letters that make up a word and then instantaneously generates other processes including the phonological decoding process, to which printed letters and the sounds of the language correspond (Adams, 1990). Many researchers have considered visual processing one of several factors that may affect childrens reading abilities. There are some concrete examples. Kavale (1982) used meta-analysis to establish the relationship between visual-perceptual skills and reading. In his review, the 161 studies produced a total of 1,571 correlation coefficients, of which 1.062 indicated the relationship. Across all visual perceptual skills (visual discrimination, visual memory, visual closure, visual spatial relationship, visual-motor integration, visual association, figure ground discrimination, and visual-auditory

PAGE 59

49 integration) and reading achievement skills, the 1,062 coefficients yielded an average correlation, 376 with a standard deviation of .139 and a standard error of .004. The correlation ranged from .01 to .89 with a median of .408. Among the visual perceptual skills, he identified that visual memory and visual discrimination skills were most significantly related to reading achievement. More interestingly, even after IQ was factored out, visual discrimination ability as measured in these studies was still highly related to reading (Kavale, 1982). The findings of his extensive meta-analysis imply that visual perceptual skills appear sufficiently associated with reading achievement to be considered among the complex of factors to the prediction of reading ability (Kavale, 1982). In order to examine the proportion of children with RD, Watson and Willow (1993) reviewed 14 clinical studies and indicated that approximately 20% of low-achieving readers have some type of visual problem such as a visual-spatial problem, a visual-perceptual deficit, or a visual memory problem. Furthermore, Aaron and Joshi (1992) reported that children are more likely to process printed words as visual information rather than as phonological information. They also found that some participants tend to disregard the orientation of letters and words and then make reversal errors. Vellutino and Scanlon (1988) also concluded that poor readers make more reversal errors due to their weakness in decoding skills and that they are therefore less coherent in their visual processing of printed words. After those poor readers were taught to use strategies regarding visual processing, they made very few reversal errors. With more wide perspectives, one research study revealed the relationship between general academic performances and visual perceptional skills. Watson et al. (2003) administered standardized sensory, perceptual linguistic intellectual, and cognitive tests to 470 children

PAGE 60

50 who entered the first grade over a 3-year period. The results of 36 tests and subtests were divided into four factors, including reading-related skills (phonological awareness, letter and word recognition); visual cognition (visual perception, visual spatial perception, visual memory); verbal cognition (language development, vocabulary verbal concepts); and speech processing (the ability to understand speech under difficult listening conditions). Not surprisingly, the profiles of factor scores showed that reading-related skills have the biggest impact on reading and other academic performance. Notably, the second strongest predictor of both reading and mathematics grades was the visual cognition factor. Overall, vision researchers addressed vision as not entirely innate but as a skill that continues to develop after birth and is influenced by the visual environment and the individuals experience. If a childs visual system has not developed to its optimal level, visually demanding activities such as reading may appear to be inefficient or uncomfortable for the child. Although a child has a good visual acuity (clarity of sight), it does not necessarily assure that the child does not have any visual problems. An extended view of the definition of vision may enhance the understanding of children with reading difficulties in terms of visual skills. Implication for Research The review of the literature provided a theoretical basis and empirical evidence on early identification of reading disabilities. Empirical studies to date have presented knowledge of factors necessary for early identification of reading disabilities. Overall, such evidence indicated that phonological awareness and other factors including visual-naming skills and visual-processing abilities are associated with the development of reading skills. However, those evidence are still not clear. Some researchers have argued

PAGE 61

51 that questions and disagreements are still remained about the visual factors as one of the contributors to reading disabilities. However, the findings of many other studies from various perspectives allow us to assume that there may be a considerable relationship between a visual deficit and RD, at least for a certain percentage of children. There are the remaining issues that need to be further studied in the future. First, this present review revealed that there are still some questions surrounding the visual deficit factor, and it is worthy of further investigation. Although a limited number of children with RD may have visual characters, we need to identify risk factors at the earliest stage to prevent or reduce influences of RD. Although previous literature indicated mixed findings on the relationship, there has been an increasing interest in the area of visual skills since more recent empirical research provided concrete evidence. Furthermore, there is a growing consensus that at least a group of children with RD have vision-related RD. However, little research has been done to indicate the specific proportion of children with RD who have visual deficits. Therefore, it is important to understand better what proportion of children with RD have visual deficits and whether these deficits are truly related to reading achievement (Feagans & Merriwether, 1990). Second, the autoregressor effects are an unsolved issue in rapid-naming research. As noted previously, Torgeson et al. (1997) discussed a limitation of predictive studies in terms of the autoregressive effects of prior reading skills. Torgeson et al. (1997) argued that rapid naming is not a separate predictor of RD and the rapid-naming skills are included in phonological skills by providing evidence. When they did not control the effect of prior reading skill, they were able to indicate a separate contribution of phonological and rapid-naming factors. However, when the autoregressive effect was

PAGE 62

52 factored out, rapid naming no longer accounted for variances in reading outcomes but phonological awareness still accounted for variances in reading (Torgeson et al., 1997). On the other hand, Meyer et al. (1998) found contradictive results. When autoregressive variables were controlled, rapid naming still accounted for variance in later reading development, but phoneme segmentation did not. Therefore, it is important that autoregressor effects should be clearly considered when researchers conduct prediction studies. Finally, a growing body of research established the necessary measurements for effective identification of early elementary students with reading disabilities. However, questions still remain about the most effective and reliable set of measurements for early primary graders of reading disabilities. Scarborough (1998) analyzed predictive studies and suggested that researchers use a multivariate approach to early, accurate screening of children. In her review, she indicated that such studies use a wide variety of variables that ranged from cognitive skills to family and school-based risk factors. In addition, combinations of variables used in each study vary greatly across studies. Although various types of measures are effective to detect children with RD in the early ages, we cannot give every possible assessment to each student. This approach is not cost and time effective. Therefore, it is essential to identify the most promising combination of predictors of early reading achievement that directly relates to the reading process (Goswami & Bryant, 1990). Conclusions The review of literature adds support to two conclusions: 1. Phonological processing is the most powerful cognitive variable determining later reading ability.

PAGE 63

53 2. Phonological awareness is not enough to explain all types of reading disabilities. Along with phonological awareness, the current review searched for other significant variables that may be associated with RD. Many connectionist theories and empirical researchers provided an insight that there may be other aspects of reading related to RD. From this review process, rapid-naming and visual deficits were seriously considered as significant variables determining later reading development. Overall, studies of phonological awareness, naming speed, and visual deficits as predictors of reading development have value in helping to understand the nature of the cognitive processes underlying reading. Therefore, the purpose of this study was to determine the effects of three variables including phonological awareness, rapid-naming, and visual skills on the reading levels of early elementary students. Further, this study investigated how the three variables are interrelated to their reading levels.

PAGE 64

CHAPTER 3 METHODS Introduction The purpose of this study was to determine the effects of various variables including phonological awareness, rapid-naming, and visual skills on the reading levels of early elementary students. In addition, this study investigated how these variables are interrelated to their reading levels. In Chapter 3, the research methods of the study are described including the participants, research instruments, data collection procedures, and methods of data analysis employed by this study. Research Questions Three research questions were investigated (Table 3-1). Table 3-1. Research Questions Research Questions Question 1 Question 2 Question 3 Is the combination of phonological awareness, rapid-naming, eye-movement/orthographic skills and visual perception skills significantly associated with the level of reading fluency? Among phonological awareness, rapid-naming, eye-movement/orthographic skills and eye-movement/orthographic skills, what is the most strongly correlated with the level of reading fluency? Among phonological awareness, rapid-naming, eye-movement/orthographic skills and visual perception skills what latent variables should be included in this reading fluency model? 54

PAGE 65

55 Hypotheses Based on the research questions and prior discussion, the following null hypotheses were generated for this study (Table 3-2). Table 3-2. Hypotheses Null Hypotheses Hypothesis 1 Hypothesis 2 Hypothesis 3 Hypothesis 4 Hypothesis 5 Hypothesis 6 Hypothesis 7 Hypothesis 8 Hypothesis 9 Hypothesis 10 Hypothesis 11 Phonological awareness, rapid-naming, and visual skills are not significantly associated with reading fluency. Phonological awareness measures are not significantly associated with reading fluency. Rapid-naming measures are not significantly associated with reading fluency. Eye-movement/Orthographic measures are not significantly associated with reading fluency. Visual perception measures are not significantly associated with reading fluency. Phonological measures are not significantly correlated with rapid-naming measures. Phonological measures are not significantly correlated with Eye-movement/Orthographic measures. Phonological measures are not significantly correlated with visual perception measures. Rapid-naming measures are not significantly correlated with eye-movement/orthographic measures. Rapid-naming measures are not significantly correlated with visual perception measures. Eye-movement/orthographic measures are not significantly correlated with visual perception measures.

PAGE 66

56 Methods Settings Five school sites in a north central Florida district were selected for this study. The schools were selected to represent a distribution of both lowand middle-income populations. The socioeconomic level of the schools is based on the percentage of students in the school receiving lunch support. The specific percentage of students who participate in the free and reduced lunch program and percentage of nonwhite students enrolled in each school are presented in Table 3-3. Table 3-3. Descriptive Information for Schools School Total Enrollment % of Nonwhite Students % of Students that receive lunch support 1 2 3 4 5 611 441 577 620 985 28 43 59 76 43 28 53 80 79 31 Participants For this investigation, 210 secondand third-grade students from five public schools in an urban area of north-central Florida participated in this study. Reading researchers have come to believe that success or failure in reading is largely determined by the students experiences in Grades Kindergarten through Grade 4. As a consequence, reading research on the determinants of processes of reading achievement has been focused on this age range. Therefore, this study took secondand third-grade students as study participants who were falling into the most commonly researched developmental

PAGE 67

57 stage. Students with diverse academic achievement levels were recruited; however, inclusion was limited to children who were primarily English speaking and currently had no special education due to mental retardation, visual impairments, and hearing impairments. The sample size was appropriate for an SEM model with three factors for this study. According to Stevens (1996), a good rule of thumb is 15 cases per predictor in a standard ordinary least squared multiple regression analysis. Since SEM is closely related to multiple regression in some respects, 15 cases per measured variable in SEM can be reasonable. Bentler and Chou (1987) suggested that researchers who use SEM may go as low as five cases per parameter. More generally, Loehlin (1992) reported the result after reviewing the literature; in the class of model with two to four factors, the investigator should plan on collecting at least 100 cases, with 200 being better (if possible). Demographic data in this study are shown in Table 3-4. Table 3-4. Demographic Data Variables Total N=210 Frequency Percentage(%) Gender Male Female Ethnicity Caucasian African-American Hispanic Asian-Pacific Grade 2nd Grade 3rd Grade 97 113 112 65 21 12 111 99 46.2 53.8 53.3 31.0 10.0 5.7 46.2 53.8 Forty six percent (n=97) were male students and 54% (n=113) were female students. The sample of five public schools included a diverse population; 53% were Caucasian; 31% were African-American; 10% were Hispanic; and 6% were Asian

PAGE 68

58 Pacific Islanders. Forty six percent were second graders, and 54% were third graders. Their ages ranged between 7 and 10 with the majority being 8 or 9. Measures The early elementary students who participated in this study completed twelve assessments with the instruments. Table 3-5 presents all the variables that were examined in the study. In addition, descriptions of the test instruments are presented in detail in the next section. Table 3-5. Test Variables Variable Latent Variable: Factor Observed Variable (Test Instrument) Phonological Awareness 1. Phoneme elision (CTOPP) 2. Blending phonemes into words (CTOPP) RapidNaming 1. A serial naming of digits (CTOPP) 2. A serial naming of letters (CTOPP) Orthographic and Eye Movement Skills 1. Orthographic (J LRRT) 2. Eye movement (DEM)-Vertical 3. Eye movement (DEM)-Horizontal Independent Variable Visual Processing Skills 1. Visual Discrimination (TVPS) 2. Visual Memory(TVPS) 3. Visual Closure(TVPS) Dependent Variable Reading Fluency 1. Oral Reading Fluency I (DIBELS) 2. Oral Reading Fluency II (DIBELS) Abbreviations: Dynamic Indicators of Basic Early Literacy Skills (DIBELS); CTOPP: Comprehensive Test of Phonological Processing, J LRRT: Jordan Left-Right Reversal Test; DEM: Developmental Eye Movements Test; TVPS:NM-R: Test of Visual-Perceptual Skills (non-motor)-Revised.

PAGE 69

59 Data Collection Procedure Permission forms for participation in the study were obtained from at least one parent or guardian of each study participant. The tests were administered individually to the study participants, who returned signed parental permission slips and met the inclusion criteria previously presented, in a designated place. All test items from the research instruments were read and carefully explained to the participants so that those with reading difficulties could participate in the study. This researcher and six other research assistants administered the research tests to all participants. Six research assistants were graduate students in education and took two halfday training sessions in test administration. Each participant was administered the test battery from 28 March 2004 through 17 May 2004. Test Instruments Dynamic Indicators of Basic Early Literacy Skills (DIBELS). A subtest of DIBELS, Oral Reading Fluency (ORF), was conducted. It is a measure that assesses fluency with textthe ability to translate letters-to-sounds-to-words fluently and effortlessly. The fluent reader is one whose decoding processes are automatic, requiring no conscious attention and used as a dependent measure for assessing reading fluency. It is a set of standardized, individually administered measures of early literacy development. The tests are designed to be short (1 minute) fluency measures used to regularly monitor the development of prereading and early reading skills (DIBELS). Such capacity then enables readers to allocate their attention to the comprehension and meaning of the text. The measures were developed upon the essential early literacy domains discussed in both the National Reading Panel (2000) reports to assess student

PAGE 70

60 development of phonological awareness, alphabetic understanding, and automaticity and fluency with the code. Each measure has been thoroughly researched and demonstrated to be reliable and valid indicators of early literacy development and predictive of later reading proficiency to aid in the early identification of students who are not progressing as expected. When used as recommended, the results can be used to evaluate individual student development as well as provide grade-level feedback toward validated instructional objectives. The website http://dibels.uoregon.edu displays reliability and validity information for the DIBELS ORF: test-retest reliability for elementary students ranges from .92 to .97; alternate form reliability ranges from .89 to .94, and criterion-related validity ranges from .52 to .91. Spring benchmark goals for a trajectory of progress of words correct per minute in grade-level material (Good, Simmons, & Kame'enui, 2001) Comprehensive Test of Phonological Processing (CTOPP). In order to assess phonological awareness and rapid-naming, the Comprehensive Test of Phonological Processing (CTOPP) is used. In particular, two subtests including a test of phoneme elision and blending into words is utilized for assessing phonological awareness. The subtest of CTOPP, the Phonological Awareness Quotient (PAQ), measures an individuals awareness and access to the phonological structure of oral language. Specifically, these two measures were administered to determine students abilities to process and manipulate sounds. One subtest of CTOPP, the Rapid-Naming Quotient (RNQ) measures the examinees efficient retrieval of phonological information from long-term or permanent memory, as well as the examinees ability to execute a sequence of operations quickly and repeatedly (Wagner et al., 1999). A serial naming of digits and

PAGE 71

61 letters is used for assessing rapid naming. The CTOPP provides percentiles, standard scores, and age and grade equivalents. Subtest standard scores have a mean of 10 and a standard deviation of 3. The Phonological Awareness has a mean of 100 and a standard deviation of 15. Age and grade equivalents show the relative standing of individuals scores. The CTOPP was normed on over 1,600 individuals ranging in age from 5 through 24 and residing in 30 states. Over half of the norming sample came from children in elementary school (through Grade 5), where the CTOPP is expected to have its widest use. The demographic characteristics of the normative sample are representative of the U.S. population as a whole with regard to gender, race, ethnicity, residence, family income, educational attainment of parents, and geographic regions. The sample characteristics were stratified by age and keyed to the demographic characteristics reported in the 1997 Statistical Abstract of the United States. Reliability of the CTOPP was investigated using estimates of content sampling, time sampling, and scorer differences. Most of the average internal consistency or alternate forms of reliability coefficients (content sampling) exceed .80. The test/retest (time sampling) coefficients range from .70 to .92. The magnitude of the coefficients reported from all the reliability studies suggests that there is limited error in the CTOPP and that examiners can have confidence in the results (Wagner et al., 1999). Jordan Left-Right Reversal Test (J L-RRT). The Jordan Left-Right Reversal Test (J L-RRT) is used for assessing one visual skill regarding orthographic processing. It tests memory for spatial orientation of upper/lower case letters and numerals. Visual reversals have been commonly accepted as one of a constellation of symptoms of learning disabilities in children (Jordan, 1980). The J L-RRT is a normreferenced test

PAGE 72

62 that assesses visual reversals of letters, numbers, and words in students ranging in age 5 through adult. This test was normed on over 3,000 individuals ranging in age from 5 through 12. The normative sample includes all socioeconomic levels, and 10% of this sample came from nonwhite racial backgrounds. For this test, the child scans rows of upperand lower-case letters and numerals and crosses out any that are incorrectly oriented. The score is the number of errors. Test-retest reliability coefficients for Level I, given in the test manual, range from 0.87 (ages 8 and 9) to .0.98 (ages 6-6 to 6-11) (Jordan, 1980). Developmental Eye Movements Test (DEM). The DEM is used for assessing a visual skill regarding eye movement. The most direct way to measure the quality of fixations and eye movements with reading is by direct movements recording. However, for young children, indirect measure is more applicable (Watson et al., 2003); therefore, it was utilized for this study. The developmental eye movement test (DEM) was designed to determine if children with reading defects have defective eye movements by comparing horizontal and vertical movements in a number naming task. The DEM was normed with 534 individuals selected from elementary school populations. The sample was equally drawn from rural public schools, suburban public schools, and suburban parochial schools. Each age and grade group was composed of an equal number of males and females. This test incorporates two number-calling subtests, one in vertical array and one in horizontal array. The vertical subtest measures an individuals visual-verbal automatic calling skills, whereas the horizontal subtest assesses ocula-motor function. This is the only test that separates the two functions (Garzia, Richman, Nicholson, & Gaines, 1990). The test also includes a pretest component in which the child must read a

PAGE 73

63 row of numbers quickly to determine automatic number knowledge. If the child is unable to perform this pretest satisfactorily, the DEM cannot be meaningfully administered. Test-retest reliability coefficients for both vertical and horizontal DEM tests, given in the test manual, are .89 and .86. Test of Visual-Perceptual Skills (non-motor)-Revised [TVPS:NM-R]. The visual processing assessment instrument for this study was Test of Visual-Perceptual Skills (non-motor)-Revised [TVPS:NM-R] (Gardner, 1994). The TVPS: NM-R is a test of visual perceptual processing ability that does not incorporate motor involvement. Children aged 4 through 13 can be tested through this instrument. It is designed for psychologists, learning specialists, optometrists, occupational therapists, remedial specialists, educators, and other professionals. The TVPS: NM-R was designed to measure seven distinct areas of a childs visual-perceptual skills: Visual Discrimination, Visual Memory, Visual-Spatial Relationships, Visual Form Constancy, Visual Sequential Memory, Visual Figure-Ground, and Visual Closure. Each of the seven subtests contains 16 items, arranged in order of difficulty, and can be administered to individual children. The TVPS: NM-R is quick and easy to administer, requiring approximately 10 to 20 minutes of testing time and only 3 to 5 minutes to score. Standardization was achieved on 1,032 children, most of whom lived in the San Francisco Bay Area but some of whom came from various states in the United States. The ages of the subjects in the standardization sample ranged from 4 years to 12 years. The reliability coefficients for the total score ranged from .83 to .91. Test results from the field study have been converted into perceptual ages, scaled scores, and percentiles for each of the seven subtests, along with perceptual quotients, percentiles, and median perceptual ages for the

PAGE 74

64 combined scores of all subtests. It was chosen because its normative data are superior to those of any other test of visual processing currently available. Among seven subtests of visual perceptual skills, three subtests were selected, including visual discrimination, visual memory and visual closure based on previous research findings. For example, Kavale (1982) conducted a meta-analysis of the relationship between visual perceptual skills and reading achievement. In this meta-analysis, 1,571 correlation coefficients were collected from 161 studies. This analysis yielded that tests of visual memory, visual discrimination, and visual closure are most significantly related with reading achievement. Therefore, these three subtests were conducted to examine visual perceptual skills. The descriptions of subtests are as follows: 1. Visual discrimination: A subjects ability to match or determine exact characteristics of two forms when one of the forms is among similar forms. 2. Visual Memory: A subjects ability to remember for immediate recall (after four or five seconds) all of the characteristics of a given form, and being able to find this form from an array of similar forms. 3. Visual closure: A subjects ability to determine, from among four incomplete forms, the one that would be the same as the completed form (Gardner, 1994). Inter-coder Reliability In general, inter-coder reliability examines the extent to which different interviewers, observers, or coders using the same instrument or measure get equivalent results. For this study, all the assessments for 45 students (out of 210 participants, 21.5%) was conducted by two assessors. Actually, the main assessor administered all the tests, and the other assessor observed and recorded separately. The inter-coder reliability for this study was examined with Cohen's kappa by using SPSS. The range of coefficients on each test was quite high, ranging from .92 to 1.00. Therefore, they were acceptable.

PAGE 75

65 Data Analysis Overview of Statistical Analysis The SPSS and LISREL 8.30 (Joreskog & Sorbom, 1996) were used to conduct all statistical analyses. The criterion for significance tests for all a priori hypotheses was set at =.05. To test the first hypothesis, the relationship between reading fluency and four latent variables, structural equation modeling with latent variables was used. To test Hypotheses 2 through 5, a regression analysis with SEM was used to examine the relationship among four independent variables and one dependent latent variable. Finally, to test Hypotheses 6 through 11, a correlational analysis was conducted to determine significant relationships between four independent latent variables. Structural Equation Modeling Overview Structural Equation Modeling (SEM) was conducted by using LISREL 8.30. LISREL 8.30 program defaults to the maximum likelihood fitting function for estimating a models parameters (Joreskog & Sorbom, 1996). This method consists of a measurement model to define hypothetical latent constructs in terms of measured variables, and a structural model to depict relationships among latent constructs. The SEM is a multivariate method combining aspects of factor analysis and multiple regression in analyzing a set of interrelated relationships among manifest and latent variables simultaneously. The advantages of SEM compared to most other conventional statistical methods include the following capabilities. First, the basic statistic in SEM is the covariance. Covariance is generally a more appropriate statistic in SEM than correlation, and covariance statistics convey more information than correlations. A correlation matrix is not suggested since SEM assumes

PAGE 76

66 that the variables are unstandardized. While conventionally analyses such as ordinary least square regression attempt to minimize differences between observed and expected individual cases, SEM aims to minimize differences between observed and expected covariance matrices. In other words, SEM, based on the covariance statistic, attempts to understand patterns of correlations among a set of variables and to explain as much of their variances (Kline, 1998, pp. 10-11). Second, SEM allows separating measurement error from hypothesized relationship among constructs that cannot be operated by conventional analyses. For example, a regression coefficient is composed of two elements: structural coefficient between the independent and dependent variable and the reliability for the predictor variable. By distinguishing a structural model from a measurement model, SEM can examine the relationship among constructs that are not influenced by measurement errors (Newcomb & Bentler, 1988). Third, SEM allows incorporation of latent variables into analyses unlike conventional analyses that focus only on observed variables. Since SEM is not limited to relations among observed variables, it gives researchers more flexibility to study any combination of relations. Finally, SEM allows researchers to estimate very complicated multivariate relationships, while conventional analyses cannot accommodate multiple indicators of the same construct due to potential problems such as multicollinearity. Even conventional multivariate techniques that deal with multiple dependent variables cannot examine relations simultaneously and only examine a single relationship at a time. However, SEM

PAGE 77

67 allows examining a series of relationships of multiple dependent variables simultaneously. Therefore, during the last three decades, SEM has become a widely used methodology for specifying, estimating, and testing hypothesized interrelationship among meaningful variables in behavioral and social sciences (Joreskog & Sorbom, 1996). Applications of the SEM methodology in education research are also on the increase. In order to apply SEM in estimating the causal model of the present study, all structural modeling was performed with the LISREL program. Assumptions of SEM analysis SEM requires certain assumptions to be met such as continuous variables, normality, and linearity of all relationships. Therefore, it is important to examine whether the assumptions of SEM have been met. First, variables used in SEM are typically continuous (Kline, 1998). All the variables used in the present study are continuous variables. Another requirement of SEM is the linearity of relationship between exogenous and endogenous variables. In order to examine the linearity, residual plots against independent variables are closely examined. In case of a violation of the linearity assumption, the use of transformed independent variables was considered to uphold these requirements. For each variable, the present study evaluates the linearity assumption by plotting the studentized residuals against the predicted values. Also, the correlation matrix indicates that multicollinearity is not a problem for multivariate analysis as there is only one correlation among the independent variables that is higher than .50. Frequency distributions, skewness and kuutosis of the data, and normal probability plots are inspected to examine normality. In addition, a covariance matrix analyzed with the Maximum Likelihood method (ML) is applied for the present study because ML is found

PAGE 78

68 to be fairly robust with the data that violate multivariate normality (Bollen & Long, 1993). Finally, this study replaces missing values with the mean values of each variable based on valid responses because SEM is sensitive to the presence of missing variables. Analytic Procedures for SEM To assure that both models are properly specified, it is critical to follow the three steps of SEM: (a) identify a proposed model, (b) evaluate the model fit, and (c) modify the model if the original model fit is bad. A Proposed model First, SEM begins with what is a conceptual rather that a statistical component: causal model. Therefore, based on previous literature and prior discussions, a reading model was proposed. In general, the models were established through three distinct phases. First, a measurement model confirmed the validity of the constructs that were used in later structural models. The purpose of this model was to identify the constructs with reduced error variance, and the constructs could be established by a factor analysis. Second, a predictive structural model was created to test the direct effects of multiple predictors on reading. Third, a mediated reading model was structured to test hypotheses. Figure 3-1 presents the model for this study that includes three factors that can be related to reading levels. A causal model is a hypothesis about the field of variables that affect one or more dependent variables of interest, presented formally as a path diagram. The diagram in Figure 3-1 demonstrates that early elementary childrens level of reading is correlated with phonological awareness, rapid-naming, and visual skills. The model was overidentified to meet basic requirement for model identification. In other words, the number of parameters must be less than the number of observations. The overidentified model resulted in positive degrees of freedom, which allowed for the rejection of the

PAGE 79

69 model. For this study a model was identified based on previous findings and hypotheses in the literature. In a path diagram, observed or measured variables are represented as rectangles, whereas latent variables that have been derived are represented as circles or ovals. Relations between variables in SEM, whether the variables are latent or measured are represented by path diagram, with collections of variables connected by lines and arrows. Variables with lines having one-way arrows pointed to them are dependent variables in the model. In this model, one latent dependent variable, reading fluency, is indicated. Variables that are the sources of lines with arrows and have no arrow leading into them are independent variables. In this model, three latent independent variables are estimated: phonological awareness, rapid-naming and visual skills. The curved arrow between latent variables represents the covariance or correlation between the two independent variables in predicting the dependent variables. The overall model fit The SEM requires that models be formally specified for estimation and model testing (Hoyle, 1995). In terms of model testing, the first step estimates the goodness-offit of the hypothesized research model. This is typically done using a chi-square (2) to determine if differences between the data and the model are statistically significant and can evaluate the relative fit of models of differing complexity.

PAGE 80

70 Elision PhonologicalAwareness Visual Perception Rapid Naming Reading Fluency Eye-movement/ Orthographi Blending Digit RN Letter RN ORF I DEM-vertical ORF II DEM-horizontal Reversal Visual Discrimination Visual Memory Visual Closur e Figure 3-1. A Proposed Model Maximum likelihood generates an estimated chi-square statistic and as such provides a probability value indicative of the overall fit of the model since the distribution of chi-square value is known (Bollen, 1989). The value of the chi-square statistic, in relation to the models degrees of freedom, determines whether the fitted residuals, resulting from (), deviate from the population values of zero. The differences between the corresponding elements of the implied covariance matrix (theory) and the observed covariance matrix (reality) consist of the matrix of residuals.

PAGE 81

71 The lower the chi-square value gained from the accumulation of these residuals, the more likely the theory represents the reality. However, such tests are sensitive to sample sizes, and the probability of rejecting any model increases when a sample size increases even with minimal deviation from the model. Therefore, Bentler and Bonnett (1980) suggested that 2/df ratio (df: degree of freedom) as a more appropriate measure of model fit. This ratio should not be bigger than 5 for models with a good fit (Bentler, 1990). In addition to the chi-square statistic, the root mean square error of approximation (RMSEA) and the non-normed fit index (NNFI) are used for evaluation for the significance of the relations between the constructs of all models. According to Bentler and Bonett (1980), these two indices were determined to be robust with relatively small sample sizes and non-normal or skewed distributional form (see also, Browne & Cudeck, 1993; Neuhaus & Swank, 2002). The second step in model estimation is to examine the path significance of each association in our research model and variance explained (R2 value) by each path. The LISREL reports raw and standardized estimates for all specified paths, along with standardized errors and test statistics for each path.

PAGE 82

CHAPTER 4 RESULTS Research findings presented in this chapter consist of four parts. The first part presents descriptive and inferential statistics on the sample. The second part of this chapter discusses the results of correlational analysis using all of the available variables for this study. Third, the results of how this model fits with the measurement model are presented, and a structural model also is presented. Finally, another correlational analysis and regression anlaysis with SEM approaches are discussed with the finalized SEM model. The following hypotheses were tested in this study: H1 Phonological awareness, rapid-naming, and visual skills are not significantly associated with reading fluency. H2 Phonological awareness measures are not significantly associated with reading fluency. H3 Rapid-naming measures are not significantly associated with reading fluency. H4 Eye-movement/Orthographic measures are not significantly associated with reading fluency. H5 Visual perception measures are not significantly associated with reading fluency. H6 Phonological measures are not significantly correlated with rapid-naming measures. H7 Phonological measures are not significantly correlated with Eye-movement/Orthographic measures. H8 Phonological measures are not significantly correlated with visual perception measures. H9 Rapid-naming measures are not significantly correlated with eye-movement/orthographic measures. 72

PAGE 83

73 H10 Rapid-naming measures are not significantly correlated with visual perception measures. H11 Eye-movement/orthographic measures are not significantly correlated with visual perception measures. Descriptive and Inferential Statistics Demographic Characteristics Descriptive statistics for the studys demographic characteristics are presented in Table 4-1. The sample of this present study consisted of 210 secondand third-grade students: 97 male students (46.2%) and 113 female students (53.8%). Approximately 53.3 % of the children were Caucasian; 31.0% were African-American children; 10% were Hispanic; and 5.7% were Asian/Pacific Islanders. The mean age of the sample was 8.5 years. Additional descriptive information of the sample on observed variables is presented in Table 4-2. Table 4-1. Descriptive Statistics on Demographic Characteristics Variables Frequency % Mean SD Range Gender Male Female Ethnicity Caucasian African-American Hispanic Asian-Pacific Grade 2nd Grade 3rd Grade Age 97 113 112 65 21 12 111 99 46.2 53.8 53.3 31.0 10.0 5.7 52.9 47.1 8.5 .50 7~10

PAGE 84

74 Table 4-2. Means and Standard Deviations on All of the Observable Variables Latent Variables Observed Variables (scale) Mean SD Minimum Maximum Phonological Awareness Elision (correct items out of 20) Blending into words (correct items out of 20) Rapid Naming Digit naming (time in secs) Letter naming (time in secs) Eye Movement/Orthographic Skills DEM vertical test (time in secs) DEM horizontal test (time in secs) Reversal Test (reversal errors) Visual Processing Skills Visual Discrimination (correct items out of 16) Visual Memory (correct items out of 16) Visual Closure (correct items out of 16) Reading Fluency Fluency I (wpm*) Fluency II (wpm) 12.84 12.73 36.76 40.67 47.68 69.04 3.30 12.82 12.19 10.26 121.93 109.48 4.38 3.27 10.68 11.79 14.16 24.21 3.48 2.27 2.41 3.39 45.26 39.39 5 6 17 17 25 28 0 1 1 1 29 21 20 20 89 104 109 199 33 16 16 16 258 208 wpm: word per minute Means and Standard Deviations on All the Observable Variables The means and standards deviations for the subjects on each test are displayed in Table 4-2. Additionally minimum and maximum scores are presented in the same table. For phonological awareness, the mean scores on elision and blending into words of this sample were 12.84 and 12.73, respectively. According to this result, the average score of digit naming was approximately 4 seconds lower than that of letter naming. This result implies children take longer time to recall letters than numbers. The average raw score for the Jordan Left-Right Reversal test was 3.30 for this sample. This score indicates the error score of this sample. According to the manual of the Jordan Left-Right Reversal test, the raw score can be converted to percentiles for boys and girls separately. Jordan indicates that approximately a 25 percentile score for boys in the age level of 8.6 is 4 and

PAGE 85

75 for girls in the same age level is 3. The 50 percentile score for both boys and girls is approximately 2. For the Fluency I test, a child was requested to read a story that is rated as one grade below, and for the Fluency II test, another story that is rated as the same grade of each student was provided. Therefore, the average score of Fluency II was lower than that of Fluency I. Carver (1990) presented approximate reading rates for students in Grades 2-12. According to Carvers (1990) analysis, the average reading fluency (wpm) for 2.5 grade students is 121. The subjects in this current study are within a range of an average reading fluency score. Table 4-3. Criteria for Percentile Cutoff Criteria for Percentile Latent Variables Observed Variables (scale) > 25 %tile 25~50 %tile 50~100 % tile Phonological Awareness Elision (correct items out of 20) Blending into words (correct items out of 20) Rapid Naming Digit naming (time in secs) Letter naming (time in secs) Eye Movement/Orthographic Skills DEM vertical test (time in secs) DEM horizontal test (time in secs) Reversal Test (reversal errors)--Boys Reversal Test (reversal errors)--Girls Visual Processing Skills Visual Discrimination (correct items out of 16) Visual Memory (correct items out of 16) Visual Closure (correct items out of 16) Reading Fluency Fluency I +Fluency II (wpm) /2---2nd Graders Fluency I +Fluency II (wpm) /2---3rd Graders 0~ 9 0~ 10 >= 46 >= 50 >= 52 >= 64 >= 5 >= 4 1~ 6 1~ 6 1~ 4 <= 70 <= 80 10~13 11~14 37~45 40~49 46~51 55~63 4 3 7~9 7~9 5~ 7 70~ 90 81~ 110 13~20 15~20 <=36 <=39 <=45 <=54 0-3 0-2 10~ 16 10~ 16 8 ~ 16 >= 91 >= 111

PAGE 86

76 Descriptive Statistics on Percentile Scores Descriptive statistics on percentile scores are also considered and presented in Table 4-4. In terms of phonological awareness, children who belong to the lowest percentile score group on elision are 36.2%. However, the number of children who belong to the lowest group in terms of blending into words was much lower (17.6%) compared to elision (36.2%). Surprisingly, the number of children who acquired lower than the 25 percentile score on DEM-horizontal test was too many (51.0%) compared to DEM-vertical test. The criteria were determined by DEM manual. Therefore, this result implies that this criterion score should be reconsidered and adjusted to the logical level. Table 4-4. Descriptive Statistics on Percentile Scores Latent Variables Observed Variables <25 Percentile (at risk) 25-50 Percentile (some risk) >50 Percentile (low risk) Phonological Awareness Elision Blending into words Rapid Naming Digit naming Letter naming Eye Movement/Orthographic Skills DEM vertical test DEM horizontal test Reversal Test Visual Processing Skills Visual Discrimination Visual Memory Visual Closure Reading Fluency Fluency I + Fluency II n (%) 76 (36.2) 37 (17.6) 35 (16.7) 34 (16.2) 60 (28.6) 107 (51.0) 64 (30.5) 6 (2.7) 6 (2.7) 11 (5.2) 36 (17.1) n (%) 33 (15.7) 87 (41.4) 36 (17.1) 71 (33.8) 39 (18.6) 38 (18.1) 30 (14.3) 25 (4.8) 17 (8.1) 41 (19.5) 39 (18.6) n (%) 100 (48.1) 86 (41.0) 139 (66.2) 105 (50.0) 111 (52.9) 65 (31.0) 116 (55.2) 194 (92.4) 187 (89.2) 158 (75.2) 135 (64.3) Interestingly, the results of the three tests of visual perception skills are consistent with prior studies. Many researchers have argued that 3% to 5% of all children may have vision-related reading disabilities. For example, Watson and Willows (1993) reviewed 14

PAGE 87

77 clinical studies and indicated that approximately 20% of low-achieving readers have some type of visual problems such as visual-spatial problems, visual-perceptual deficits, or visual-memory problems. This sample includes 36 children who show very low reading fluency level (less than 25 percentile). Therefore, approximately 20% of children with low-reading fluency may have vision-related reading problems. Gender Differences Group means on all measures were calculated, and a one-way analysis of variance (ANOVA) was conducted to see if there are any gender differences. Table 4-5 provides a summary of the one-way ANOVA of all the measures. Table 4-5. ANOVA Summaries on Gender Differences Boys Girls ANOVA Latent Variables Observed Variables Mean SD Mean SD F p Phonological Awareness Elision Blending into words Rapid Naming Digit naming Letter naming Eye Movement/Orthographic Skills DEM vertical test DEM horizontal test Reversal Test Visual Processing Skills Visual Discrimination Visual Memory Visual Closure Reading Fluency Fluency I (wpm*) Fluency II (wpm) 13.12 12.60 37.95 42.00 49.10 70.45 3.39 12.79 12.16 10.12 119.64 108.13 4.47 3.45 11.89 13.61 16.42 26.49 2.82 2.43 2.69 3.55 45.89 40.27 12.61 12.85 35.74 39.52 46.45 67.84 3.23 12.86 12.21 10.38 123.89 110.63 4.32 3.11 9.45 9.88 11.83 22.12 3.98 2.04 2.15 3.25 44.82 38.76 .714 .308 2.239 2.321 1.836 .607 .112 .044 .020 .320 .460 .209 .399 .579 .136 .129 .177 .437 .738 .835 .887 .572 .498 .647 wpm: word per minute

PAGE 88

78 On all the assessment, the girls performance was superior to boys. However, no significant gender differences were found. Because gender difference is minimal, these findings of this study are consistent with pervious studies. Differences Between Second and Third Graders A one-way analysis of variance (ANOVA) was conducted to see if there were any differences between second and third graders on all the observed measures. Table 4-6 provides a summary of the one-way ANOVA of all the measures. Table 4-6. ANOVA Summaries on Group Differences (2nd/3rd Graders) 2nd Graders 3rd Graders ANOVA Latent Variables Observed Variables Mean SD Mean SD F p Phonological Awareness Elision Blending into words Rapid Naming Digit naming Letter naming Eye Movement/Orthographic Skills: DEM vertical test DEM horizontal test Reversal Test Visual Perception Skills Visual Discrimination Visual Memory Visual Closure Reading Fluency Fluency I (wpm***) Fluency II (wpm) 13.12 12.60 37.95 42.00 49.10 70.45 3.39 12.79 12.16 10.12 114.14 104.48 4.47 3.45 11.89 13.61 16.42 26.49 2.82 2.43 2.69 3.55 40.01 37.15 12.61 12.85 35.74 39.52 46.45 67.84 3.23 12.86 12.21 10.38 130.66 115.08 4.32 3.11 9.45 9.88 11.83 22.12 3.98 2.04 2.15 3.25 49.25 41.22 .950 1.909 15.541 14.332 9.391 23.821 9.249 4.257 1.771 5.850 7.171 3.844 .331 .169 .000** .000** .002** .000** .003** .040* .185 .016* .008** .051 Statistically significant at the 0.05 level (2-tailed test) ** Statistically significant at the 0.01 level (2-tailed test) *** wpm: word per minute On all of the assessments, the performance of third graders was superior to that of second graders. However, no significant group differences were found on phonological awareness measures. On the other hand, there were statistically significant differences on rapid-naming and eye movement/reversal measures between second and third graders.

PAGE 89

79 For visual perception skill measures, the results of VD and VC measures indicated that there are significant differences between second and third graders. However, the level of VM seems to be similar between second and third graders. For the Fluency I test, third graders seemed to read significantly faster than second graders. For the Fluency II test, p-value was 0.051, indicating there is a marginal difference between second and third graders. Differences Between Students From Schools in Mid and Low SES Neighborhoods Since data on lunch support of each participant were unavailable, the sample was divided into two groups. The first group of students were currently going to schools (School 3 and 4) in the lower SES neighborhood. More than 50 % of students in the lower SES neighborhood schools (within this study) get lunch support. A one-way analysis of variance (ANOVA) was conducted to determine if there are any differences between the two groups of students on all the observed measures. The results indicated that mean scores of students who go to schools in higher SES neighborhood (Schools 1, 2, and 5) are higher on all the measures. On most of the measures, significant mean differences were detected between the two groups except rapid-naming variables. Table 4-7 provides a summary of the one-way ANOVA of all the measures. Correlational Analysis A correlational analysis was performed using all of the available observed variables. The correlation matrix is presented in Table 4-8. Surprisingly, a closer examination reveals that some indicators of theoretical constructs (e.g., phonological factors) have very week correlations on reading fluency.

PAGE 90

80 Table 4-7. ANOVA Summaries on Group Differences (Schools with Lower/Higher SES Students) Students from Schools in LowSES Neighborhood (n= 91) Students from Schools in MidSES Neighborhood (n=119) ANOVA Latent Variables Observed Variables Mean SD Mean SD F p Phonological Awareness Elision Blending into words Rapid Naming Digit naming Letter naming Eye Movement/Orthographic Skills DEM vertical test DEM horizontal test Reversal Test Visual Processing Skills Visual Discrimination Visual Memory Visual Closure Reading Fluency Fluency I (wpm) Fluency II (wpm) 11.71 12.05 37.75 42.43 48.29 69.91 4.15 12.37 11.41 9.54 103.43 95.37 4.29 2.93 11.03 13.36 14.46 25.78 3.26 2.45 2.60 3.33 40.68 36.41 13.71 13.25 36.01 39.32 47.21 68.39 2.65 13.17 12.78 10.82 136.08 120.26 4.27 3.42 10.38 10.28 13.98 23.09 3.52 1.98 2.07 3.34 43.58 45.26 11.264 7.119 1.370 3.633 .296 .204 9.948 6.896 17.872 7.655 30.640 22.728 .001* .008* .243 .058 .587 .652 .002* .009* .000** .006* .000** .000** Statistically significant at the 0.05 level (2-tailed test) ** Statistically significant at the 0.01 level (2-tailed test) For the most part, correlations measured within constructs are higher than those across constructs. Two variables of phonological awareness, elision and blending into words, are significantly related to each other (r= .373, p < .01), and two rapid-naming variables are highly correlated (r= .849, p< .01). However, unlike elision, the other variable, blending into words, has relatively low correlations with dependent observed variables, Fluency I and II (r= .110 and r=.081) and has no statically significant correlations. The highest correlation is detected in the relationship between Fluency I and Fluency II (r=. 936, p < .01). Since reversal test scores are calculated with error scores, this variable is negatively related to two phonological variable, elision and blending into

PAGE 91

81 words (r= .137, p < .05 and r= .011, p > .05). This result represents that children who score higher on the reversal test are less likely to have higher score on elision. However, it shows that there is no significant relationship between reversal and blending into words. Table 4-8. Correlation Matrix of Phonological Awareness, Rapid-Naming, and Visual Skills and Reading Fluency Elision Blending Digit RN Letter RN DEM-V DEM-H Elision 1.00 Blending .37** 1.00 Digit RN -.15* -.01 1.00 Letter RN -.12 -.02 .85* 1.00 DEM-V -.16* .01 .89* .81** 1.00 DEM-H -.20** -.10 .67** .66** .67** 1.00 Reversal -.14* -.01 .24** .23** .22** .22** VD .19** .17* -.27** -.32** -.27** -.39** VM .18** .18* -.23** -.27** -.19** -.32** VC .25** .19** -.24** -.24** -.24** -.33** Fluency I .31** .11 -.60** -.60** -.58** -.57** Fluency II .25** .08 -.61** -.60** -.57** -.59** Reversal VD VM VC Fluency I Fluency II Reversal 1.00 VD -.28** 1.00 VM -.27** .55** 1.00 VC -.36** .49** .56** 1.00 Fluency I -.39** .35** .37** .50** 1.00 Fluency II -.38** .32** .36** .46** .94** 1.00 Correlation is significant at the 0.05 level (2-tailed test) ** Correlation is significant at the 0.01 level (2-tailed test) As predicted, all of the indicators of visual perception skills are positively related to each other and statistically significant at the .01 level. Additionally, all of the indicators of visual perception skills are positively related to Fluency I and II and statistically significant at the .01 level. The correlations among constructs provide some preliminary evidence for the hypothesized model. Since rapid-naming scores are the time in seconds (seconds to name 40 letters/40 digits), they are negatively correlated with phonological

PAGE 92

82 awareness and reading fluency factors. However, they are positively correlated with eye-movement/reversal factors because eye-movement scores are also the time in seconds to read numbers vertically and horizontally, and reversal scores indicate error scores. Furthermore, all of the vision-related variables and reading fluency variables are significantly related. Surprisingly, one of the observed variables, blending into words, from a phonological factor has a very weak relationship with other reading-related variables. One important fact has been detected through the correlation analysis is that there are very high correlations between the two variables in rapid-naming and vertical/horizontal eye movement variables. For example, DEM vertical test scores are highly correlated with digit rapid naming and letter rapid naming (r= .849 and r= .805 respectively). Since the two constructs are significantly correlated with each other, if we assign these scores to different constructs, it would violate one of assumptions in SEM, multicollinearity. In some cases, multiple regression results may seem paradoxical. Even though the overall P value is very low, all of the individual P values are high. This means that the model fits the data well, even though none of the X variables has a statistically significant impact on predicting Y. How is this possible? When two X variables are highly correlated, they both convey essentially the same information. For the current study, a rapid naming variable is very highly correlated with an eye movement variable. In this case, neither may contribute significantly to the model after the other one is included, but together they contribute a great deal. If you removed both variables from the model, the fit would be much worse. So, the overall model fits the data well, but neither X variable makes a significant contribution when it is added to the model last. When this happens, the X variables are collinear, and the results show multicollinearity.

PAGE 93

83 Therefore, modification of the measurement model may be necessary. The modified measurement model is discussed in detail in the next section. SEM Analysis The SEM analysis provides comprehensive statistical techniques to examine and investigate the relationship between measured variables and latent constructs. SEM is a priori: a model makes theoretical sense and its statistical correspondence to the data is reasonable. SEM is contained of two components; the measurement model and the structural model and SEM approaches usually follow three steps. First, measurement model fit was evaluated by using confirmatory factor analysis and second, the structural model fit was evaluated. Finally, model modification is taken place if a model fit of a structural model is worse than one of a measurement model. However, in this study, the structural model fit was fairly good. In response, the last step was not necessary to perform. Measurement Model The purpose of the measurement model was to create the constructs that would be used in the later structural model. The validity of the constructs was determined by confirming the significant relations between the observed variables that were assumed to be theoretically related. The purpose of this model is to identify the constructs with reduced error variance, and the constructs can be established by a factor analysis. A confirmatory factor analysis was completed in order to determine whether the variables load on five factors that can be labeled as phonological awareness, rapid-naming, eye-movement/orthographic skills, visual skills, and reading fluency. Findings appear to have adequately validated the five-factor structural model of phonological awareness, rapid-naming, eye-movement/orthographic skills, visual processing skills, and reading fluency.

PAGE 94

84 That is, the model of this study includes the five latent variables with 12 observed variables. The variables loaded neatly onto five factors, as predicted. The following model is the estimation of the authors model. The pattern of fixed and free parameters in a structural equation model defines two components of the general structural equation model: the measurement model and the structural model. The measurement model is that component of the general model in which latent variables are prescribed (Hoyle, 1995). The results of goodness of fit on the measurement model are shown in Table 4-9. The fit of the original, hypothesized model (Figure 3-1) was adequate because the chi-square was statistically significant (2 (66) = 1708.237), and the additional fit indices showed that the model has fairly good fit (NFI= .937, NNFI= .943, and RMSEA= .079). However, as discussed in the previous section, there is a fatal problem in this proposed measurement model. There are very high correlations between the two variables in rapid-naming and vertical/horizontal eye movement variables. For example, DEM vertical test scores are highly correlated with digit rapid naming and letter rapid naming (r= .849 and r= .805). This is an evidence of multicollinearity in this model. Therefore, new constructs were re-identified. Table 4-9 presents a modified constructs and goodness of fit results for the newly identified measurement model. As predicted, a measurement model fit became much greater with this modification (2 (45) = 78.160, NFI= .954, NNFI= .970, CFI = .980, SRMR= .050, RMSEA= .057). Therefore, within the measurement model, all of the measured variables were significant indicators of their respective latent variables, and no modifications of the measurement model were needed since all of the fit indices met their criteria for good fit.

PAGE 95

85 Table 4-9. Factors for Variables in the Overall Model and Goodness of Fit Results of the Measurement Model Proposed Constructs Observed Variables Modified Constructs Observed Variables Factor 1 Phonological Awareness Elision Blending into words Phonological Awareness Elision Blending into words Factor 2 Rapid Naming Digit rapid naming Letter rapid naming Rapid Naming/ Eye-movement Digit rapid naming Letter rapid naming Eye-movement-vertical Eye-movement-horizontal Factor 3 Eye-movement/Orthographic Skills Eye-movement-vertical Eye-movement-horizontal Reversal Orthographic Skills Reversal Factor 4 Visual Processing Skills Visual Discrimination Visual Memory Visual Closure Visual Processing Skills Visual Discrimination Visual Memory Visual Closure Factor 5 Reading Fluency Reading Fluency I Reading Fluency II Reading Fluency Reading Fluency I Reading Fluency II Goodness of Fit Results 2 (66) = 1708.237 NFI= .937 NNFI= .943 CFI = .962 SRMR= .072 RMSEA= .079 90% Confidence Interval of RMSEA : (.059~ .099) 2 (45) = 78.160 NFI= .954 NNFI= .970 CFI = .980 SRMR= .050 RMSEA= .057 90% Confidence Interval of RMSEA : (.033~ ,079) Based on all the goodness of fit indices, this measurement model fit the data well. The results of the confirmatory factor analysis are presented in Table 4-10. The italicized numbers represent the t-value. If the t-value is greater than 1.96 with .05 level, it means that the observable variable is statistically significant for this model. All of the t-values on latent variables are greater than 1.96 in Table 4-10.

PAGE 96

86 Table 4-10. Results of Confirmatory Factor Analysis PA RN/EV Reversal VP RF R2 Elision 0.777 (0.139) 5.559* 0.604 Blending 0.476 (1.101) 4.710* 0.227 Digit RN 0.955 (0.052) 18.502* 0.913 Letter RN 0.888 (0.054) 16.317* 0.789 DEM-vertical 0.925 (0.053) 17.470* 0.856 DEM-horizontal 0.722 (0.060) 11.946* 0.521 Reversal 0.933 (0.052) 17.787* 0.870 VD 0.674 (0.068) 9.879* 0.454 VM 0.732 (0.067) 10.929* 0.537 VC 0.777 (0.066) 11.733* 0.604 Fluency I 0.984 (0.051) 19.321* 0.968 Fluency II 0.956 (0.052) 18.332* 0.913 In addition to t-value, squared multiple correlations (R2) are displayed in Table 4-10. Since most of them are quite high, I can conclude that reliability of observable measures is fairly satisfactory.

PAGE 97

87 Structural Model The next step in SEM is construction of a structural model that fits the data as well as the final measurement model with fewer estimated parameters. The structural model is that component of the general model that prescribes relations between latent variables and observed variables that are not indicators of latent variables. In general, the multiple regression model is a structural model without latent variables and limited to a single outcome. However, SEM that includes the combinations of the measurements and structural components allows a comprehensive statistical model that can be used to evaluate relations among variables that are free of measurement error (Hoyle, 1995). The chi-square statistic included in Table 4-11 provides a test of the null hypothesis that the reproduced variance matrix has the specified model structure (i.e., that the model fit the data). First, the chi-square test for the model was statistically significant, 2 (66)= 1708.237, p< .05, indicating a good fit between the model and data. However, several researchers (Bollen, 1989; Bollen & Long, 1993; Joreskog & Sorbom, 1996) noted that the chi-square statistic is influenced by a large sample size (for this data set N= 210). In fact, Joreskog and Sorgom (1996) proposed that 2 be used as badness rather than a goodness-of-fit measure in the sense that a small 2 value (relative to its degrees of freedom) is indicative of good fit, whereas a large 2 value reflects bad fit in SEM (Byrne, 1998). Therefore, four additional goodness of fit indices were considered for testing a goodness of fit of SEM model. The criteria for four other indices are listed: CFI shows a good fit at .95 or higher; RMSEA should be between .08 or lower to indicate good model fit (Hu & Bentler, 1999); NFI and NNFI should be close to 1.0 and show a good fit at .90 or higher. These indices except RMSEA range from 0.00 to 1.00, with

PAGE 98

88 values close to 1.00 being indicative of good fit. According to Hu & Bentler (1999), the Root Mean Square Error of Approximation (RMSEA) has only recently been recognized as one of the most informative criteria in covariance structure modeling, and it takes into account the error of approximation in the population and asks the question, How well would the model fit the population covariance matrix if it were available? (Browne & Cudeck, 1993, pp. 137-138). This discrepancy measured by the RMSEA indicated that values of less than .05 indicate good fit; values as high as .08 represent reasonable errors of approximation in the population; values from .08 to .10 indicate mediocre fit; and those greater than .10 indicate poor fit (MacCallum, Browne, & Sugawara, 1996). The fit of the original, hypothesized model (Figure 3-1) was adequate because the chi-square was statistically significant (2 (45) = 78.160), and the fit indices showed that the model has a moderate fit (NFI= .954, NNFI= .970 CFI= .980, SRMR= .050, RMSEA= .057). Table 4-11. Results of Goodness of Fit Goodness of Fit Indices SEM Model 2 df NFI NNFI CFI SRMR RMSEA 90% Confidence Interval of RMSEA 78.160 45 .954 .970 .980 .050 .057 (.033~ ,079) This SEM model, therefore, can be accepted due to a good model fit. As a consequence, these results on goodness of fit can examine the following hypothesis: H1 Phonological awareness, rapid-naming, and visual skills are not significantly associated with reading fluency.

PAGE 99

89 Based on the several results of indices on this SEM model, phonological awareness, rapid-naming, and visual skills are significantly associated with reading fluency. Therefore, Hypothesis 1 was rejected. Regression Analysis and Correlational Analysis with SEM Multiple regression analysis A multiple regression analysis was conducted by using SEM to examine Hypotheses 2 through 5. The regression coefficientsof four independent constructs and one dependent construct is displayed in Figure 4-1. Table 4-12 represents regressions as well as t-value. The italicized numbers represent t-value. If the t-value is greater than 1.96 with .05 level, it means that the regression between one dependent variable and one independent variable is statistically significant. All of the t-values on this regression matrix are greater than 1.96. Therefore, the following hypotheses were able to examined by this statistical approach: H2 Phonological awareness measures are not significantly associated with reading fluency. H3 Rapid-naming measures are not significantly associated with reading fluency. H4 Eye-movement/Orthographic measures are not significantly associated with reading fluency. H5 Visual perception measures are not significantly associated with reading fluency. Hypotheses 2 and 5 were rejected since the correlations are statistically significant. However, due to the adjustment of constructs, the original Hypothesis 3 and 4 could not be examined. Instead, rapid-naming (RN) measures and eye-movement (EM) measures were combined into the modified construct and examined. The results indicate that the RN/EM latent variable is significantly associated with reading fluency. In addition, the

PAGE 100

90 orthographic measure with a single indicator, reversal, is significantly related to reading fluency. Elision Visual Perception Rapid Naming / Eye-movement Fluency I .98 Fluency II .96 Reading Fluenc y Orthographic .78 -.50 -.15 .26 14 .48 .96 .89 .92 .72 .93 .67 78 Phonological Awareness .73 Blending Digit Naming Letter Naming DEM-vertical DEM-horizontal Reversal Visual Discrimination Visual Closur e Visual Memory Figure 4-1. SEM Reading Fluency Model with Correlation Coefficients

PAGE 101

91 Table 4-12. Regression Matrix of Four Independent Constructs on Reading Fluency PA RN/EM Reversal VP RF 0.142 (0.071) 1.996* -0.497 (0.059) -8.408* -0.149 (0.061) -2.447* 0.264 (0.076) 3.487* The correlation matrix on four independent latent variables is separately presented in Table 4-13. The italicized numbers represent the t-value. If the t-value is greater than 1.96 with .05 level, it means that the correlation between the two observed variables is statistically significant. All of the t-values except a correlation between phonological awareness and reversal on this correlation matrix are greater than 1.96. Therefore, the final SEM model should remove the arrow between phonological awareness and reversal. H6 Phonological measures are not significantly correlated with rapid-naming measures. H7 Phonological measures are not significantly correlated with eye-movement/orthographic measures. H8 Phonological measures are not significantly correlated with visual perception measures. H9 Rapid-naming measures are not significantly correlated with eyemovement/orthographic measures. H10 Rapid-naming measures are not significantly correlated with visual perception measures. H11 Eye-movement/orthographic measures are not significantly correlated with visual perception measures. These correlations were utilized to examine Hypotheses 6 through 11. Since rapid-naming (RN) measures and eye movement (EM) measures were combined into a modified construct and RN/EM and the reversal variable were left alone in the reversal construct, there may be some confusion in examining these hypotheses. However, Table 4-13 clearly presents the correlations between the independent latent variables.

PAGE 102

92 According to the correlation matrix displayed in Table 4-13, six combinations from four independent latent variables were examined to check their correlations. Five out of six combinations showed a statistically significant relationship. Between phonological awareness and reversal, however, there was no significant correlation. Table 4-13. Correlation Matrix on Independent Latent Variables PA RN/EM Reversal VP PA 1.000 RN/EM -0.181 (0.088) -2.051* 1.000 Reversal -0.166 (0.092) -1.791 0.270 (0.071) 3.805* 1.000 VP 0.394 (0.098) 3.999* -0.372 (0.071) -5.254* -0.451 (0.070) -6.401* 1.000 Phonological Awareness Rapid Naming/Eye Movement Reversal -.37 .39 -.18 .27 -.17 -.45 Visual Perception Figure 4-2. Independent Latent Variables with Correlation Coefficients

PAGE 103

93 Summary In Chapter 4, findings of this study were presented that were consisted of four parts. First, descriptive and inferential statistics on the sample based on ANOVA were presented. For examination of group differences, the sample was divided into two groups with three different perspectives. There was no significant group difference on gender. However, 3rd graders are significantly superior to 2nd graders on many measures. Furthermore, there were significant group differences between students who attend schools in low SES neighborhoods and those who attend schools in mid-SES neighborhoods on many measures. The second part of this chapter discussed the results of correlational analysis using all of the available variables for this study. For the most part, correlations measured within constructs are higher than those across constructs. For example, two variables of phonological awareness, elision and blending into words, are significantly related to each other (r= .373, p < .01), and two rapid-naming variables are highly correlated (r= .849, p< .01). Overall, most independent observed variables are correlated with dependent observed variables, Fluency I and II. Third, the results of how this model fits with the measurement model were presented, and a structural model also was presented. Due to the issue related to multicollinearity, new constructs were re-identified. Based on all the goodness of fit indices, this adjusted measurement model fit the data well. Due to a good model fit, this SEM model could be accepted. Finally, another correlational analysis and regression anlaysis with SEM approaches were discussed with the finalized SEM model. The regression matrix of four independent constructs and one dependent construct indicated that all of the t-values on

PAGE 104

94 this regression matrix are greater than 1.96. This means that all of the constructs that were suggested should be existed in this SEM model.

PAGE 105

CHAPTER 5 DISCUSSION In this chapter, I will provide an overview of the current study and summarize the results. A discussion of the studys strength and limitations also are presented. Based on the research findings, implications for future research and practice are addressed. The conclusions of this study are discussed in the final section. Introduction Approximately 20% of U.S. children have serious reading difficulties (Grossen, 1997). It is widely recognized that early reading difficulties tend to persist over time. Researchers demonstrated that early elementary grade students who struggle with reading were highly likely to have reading difficulties in the later elementary grades and even in high school (Juel ,1988; Fransis Shaywitz, Stuebing, Shaywitz, & Fletcher, 1996). As a consequence, there have been numerous efforts to search for reliable indicators for early identification as well as to develop early intervention programs for young children at risk for reading difficulties (e.g., Adams, Foorman, Lundberg & Beeler, 1998; Slavin, Madden, & Wasik, 1997; Torgesen & Bryant, 1993). In spite of these enormous efforts on early detection and intervention, some students fail to learn to read. Why do they fail to learn to read? Researchers searched for important factors that affect RD. Many researchers discovered that phonological awareness is a powerful indicator of RD (Catts et al., 1999; Torgesen, Wagner, & Rashotte, 1994; Togesen et al., 1997). However, other factors such as rapid naming and visual processing factors are 95

PAGE 106

96 controversial and somewhat overlooked by reading researchers. Therefore, this study was designed to clarify whether these factors affect childrens reading fluency. In this study, a reading fluency model was constructed to examine the effects of phonological awareness, rapid-naming, eye-movement/orthographic skills, and visual perceptual skills on childrens reading fluency. Two hundred ten subjects from second and third grades were recruited for this study. Finally, the identified reading model in this study provided a basis for understanding the influence of various factors including phonological awareness, rapid-naming, and visual skills on reading. Summary of the Results For this study, various measures were used to assess young childrens phonological awareness, naming speed, and three different types of visual skills including eye movement, reversal, and visual perceptions. Data from 12 assessments were analyzed using various methods including ANOVA, correlational analysis, and SEM. A discussion of the results is organized as follows: (a) descriptive and inferential statistics, (b) correlational analysis, and (c) structural equation modeling. Results from inferential statistics indicated that there were no significant gender difference on any of the measures. However, there were significant differences between second and third graders in several measures including rapid-naming, eye movement, reversal, and reading fluency rate measures. As expected, third graders were superior to second graders on all the measures. Furthermore, the results indicated that mean scores of students who attended to schools in higher SES neighborhoods were higher than those of students who attended to schools in lower SES neighborhoods on all measures. Consistent with prior research which has indicated that approximately 3% to 5% of all children may have visual problems that affect their reading (e.g., Watson & Willows,

PAGE 107

97 1993), the current research also revealed that 6 out of 210 children (2.7%) are at risk for visual discrimination and visual memory. The results of the correlational analysis revealed that there were significant relationships between reading fluency and four constructs including phonological awareness, rapid-naming, eye-movement/reversal, and visual perception skills. In particular, the rapid-naming construct had the highest correlation with reading fluency However, the results of correlational analysis showed that rapid naming variables and eye movement variables were too highly correlated to remain as two separate constructs. If two highly correlated constructs were included in the same model, it would violate one of assumptions of SEM which is multicollinearity. Therefore rapid naming and eye movement variables were combined for further analysis with SEM. Finally, two separate SEM analyses were conducted with and without an adjustment of constructs. The results showed that the adjusted reading fluency model had a better fit and that the measures used in this study were significant indicators of their respective latent variables. Therefore, the adjusted reading model was selected. Since all of the fit indices met the criteria for a good fit, no further modification of the measurement model was needed. Interpretations of the findings Rapid Naming As discussed in Chapter 2, ample research has revealed that phonological awareness is the most significant indicator in determining reading disabilities (e.g., Stanovich & Siegel, 1994; Torgesen et al., 1997). Although the present study shows that phonological awareness was significantly correlated with reading fluency, phonological awareness was not the most important indicator in predicting reading fluency problems.

PAGE 108

98 This study demonstrates that rapid naming variables are the most important variables to predict the level of reading fluency for the sample consisted of 2nd and 3rd graders. This result is consistent with several studies that highlight that the relationship between rapid naming and reading ability appears to be the strongest in the early period of reading development (i.e., Grade 4 or earlier; McBride-Chang & Manis, 1996) The results also indicate that phonological sensitivity is not the only skill necessary to predict problems on reading fluency. One skill in particular, rapid naming, has the highest correlation with reading fluency. Rapid naming factor on reading has been studied in length and found to be of equal importance for predictability of childrens reading difficulties. In particular, the double-deficit theory of reading disability (Bowers & Wolf, 1993; Wolf & Bowers, 1999) contends that a combination of both phonological and rapid naming deficits result in performances at the lowest level while problems with only one skill or the other do not affect performance as drastically. The results add to the growing evidence of a relation between naming speed and reading fluency (de Jong & van der Leij, 2003; Manis et al., 1999; Schatschneider, Carlson, Francis, Foorman, & Fletcher, 2002; Wolf & Bowers, 1999). To date, there has been some debate about the relationship between phonological awareness and naming speed. Wagner and Torgesen (1987) considered them both a reflection of a unitary phonological process; however, other researchers (Badian, 1993; Felton & Brown, 1990) found no correlation between the two skills. By using correlational analysis, the current study discovered that there is a minimal relationship between phonological awareness and rapid-naming skills. For example, in the study scores on phonological awareness, blending into words and elision, were not

PAGE 109

99 significantly correlated with letter rapid naming (r= -.02 and r= -. 12, respectively). This study is consistent with these research findings highlighting the predictive capacity of naming speed tasks for later reading ability (Bowers & Swanson, 1991; Catts, 1991; Felton, 1992). Therefore, this study suggests that naming speed should be independently and significantly considered when discussing factors related to RD. Furthermore, the results also provide evidence that the double-deficit hypothesis suggested by Wolf and Bowers (1999) is a reasonable hypothesis to accept in reading research. The Relationship between Rapid Naming and Eye Movement One important fact has been detected through the correlation analysis is that there are very high correlations between the two variables in rapid-naming and vertical/horizontal eye movement variables. For example, DEM vertical test scores are highly correlated with digit rapid naming and letter rapid naming (r= .849 and r= .805 respectively). This result is consistent with a study conducted by Wolf, Bowers, and Biddle (2000). They argued that rapid naming tasks are consisted of intentional, visual, lexical, temporal, and recognition subprocesses that all contribute to naming speed performance. Since two factors are highly correlated if a child struggles with naming speed and seems to have no other linguistic impairment, problems on visual-motor skills should be seriously considered. The Relationship between Visual Factors and Reading Fluency The regression coefficients of four independent constructs on reading fluency are displayed in table 4-12. According to this regression matrix, the combination of rapid naming and eye movement variables has the strongest predictability of 2nd and 3rd graders reading fluency. Moreover, this result demonstrates that the second strongest construct to predict the level of reading fluency is the visual perception skills. This

PAGE 110

100 implies that children who perform poorly in the tests of visual perception skills they have higher tendency to acquire poor scores on reading fluency. Fortunately, there are only about 3% of the sample show significant impairments on visual perception skills. However, on top of severe problems in visual perception skills, their overall scores on these assessments should be carefully pointed out. Overall, these students showed severe reversal problems and performed profoundly poor at the tests on oral reading fluency as well as phonological awareness. Only one out of six students performed in an average range on phonological awareness and poorly on any other areas. Conventional intervention based on phonological awareness may not be sufficient for this group of children. Recently many measures have been developed for early pre-literacy screening in order to identify children at risk of reading difficulties. In turns, early intervention has been implemented for at risk children in reading. However, researchers have indicated that intervention is not successful for everyone. They describe children who fail to benefit from intervention as 'Treatment Resisters' (e.g., Blachman, 1994; Togesson et al, 1994). This group of students may be treatment resisters who may fail to respond adequately to interventions that are sufficient for most poor readers. Therefore, in someway, this study identified the characteristics which differentiate treatment resisters from those children who benefit from early phonological based intervention and from children not considered to be at risk of difficulties. The results of SEM suggest that four independent constructs including phonological awareness, rapid naming/eye-moment, reversal, and visual perception skills are significantly correlated with early elementary childrens reading fluency. In particular, the multiple regression analysis with SEM indicated that, among the four

PAGE 111

101 independent constructs, the rapid-naming/eye-movement construct has the highest correlation with reading fluency. These results imply that these four factors should be critically considered to indicate young childrens reading problems. Discussions on Study Limitations In this section, the limitations of the study related to measurement and generalizability issues will be discussed briefly. Measurement Issues One of the primary limitations of this study relates to the measures used to assess the areas of reading fluency, phonological awareness, rapid-naming, and visual skills. First, a limited number of assessment test batteries were used for this study. For example, phonological awareness was measured by using two subtests of CTOPP: (a) elision, and (b) blending into words. Even though these two subtests were carefully chosen to measure phonological awareness and it was believed that they were the two most applicable areas for second and third graders, other subtests of CTOPP (e.g., segmentation and phonological memory) should or could have been included in the study in order to have a more comprehensive score of phonological awareness. This would have allowed researchers to determine if phonological memory was related to other sources of variance such as visual memory and general processing time of rapid naming. In assessing orthographic skills, only one very basic orthographic task was administered to the participants, the Jordan Left-Right Reversal Test. This instrument tests memory for spatial orientation of upper/lower case letters and numerals. However, it may not have been sufficient to measure orthographic tasks comprehensively. Therefore, additional measures such as the Gardner Reversal Frequency Tests (Gardner, 1980) could have been added in order to get a more comprehensive score of reversals.

PAGE 112

102 In terms of determination of childrens visual perception skills, only three subtests, visual discrimination, visual memory, and visual closure, out of the seven subtests for testing visual perception skills (TVPS) were administered. Four other subtests were excluded; visual-Spatial relationships, visual form constancy, visual sequential memory, and visual figure-ground. The rationale for choosing the three subtests was provided in Chapter 3. Although three subtests may have been sufficient to determine the childrens perception skills, the use of all seven subtests to measure visual perception skills could have provided more information on the role visual perception played in reading fluency A second limitation was related to selection, a threat to internal validity. For example, for the blending into words measure, the percentage of children whose score was lower than the 25 percentile seems to be surprisingly low (17.6%) compared to the percentage on elision (37%). It is possible that low-achieving readers may have more instruction based practice on phonological awareness and more experiences with taking these types of tests. For example, I asked several students (who received low scores on reading fluency measures and relatively high scores on blending) if they had received reading instruction on blending into words. Most of them said they had received instruction on these skills. Moreover, some of the students were familiar with the CTOPP-Blending into words. This type of measurement error may cause problems in the overall results of the phonological awareness latent variable. Therefore, researchers should carefully consider validity issues in gathering precise data for future research. Generalizability Issues The participants in the current study were drawn from north central Florida. Even though children belonged to schools with a diverse population, the geographical factor may have affected a variety of factors that impact childrens reading outcome. As

PAGE 113

103 discussed earlier, the results are limited to secondand third-grade students and are not applicable to students younger or older. Therefore, the results of this study cannot be generalized to any graders other than 2nd and 3rd graders. Implication for Future Research and Practice The results of this study provide implications for future research. Based on the results, the logical next steps for research will be discussed in this section. Furthermore, several implications for future practice are discussed in a later section. Future Research Directions Selecting Reading Assessments Instruments for assessment of childrens reading levels should be clearly selected and used for detecting problems of reading. The IDEA and Public Law 105-17 require public school systems to evaluate children who are at risk for a reading disability. With legislative support, a growing body of research has established the necessary measurements for effective identification of early elementary students with reading disabilities. However, questions still remain about the most effective and reliable set of measurements for detecting reading disabilities in the early primary grades. Scarborough (1998) analyzed predictive studies and suggested that researchers use a multivariate approach to early and accurate screening of children. In her review, she indicated that such studies use a wide variety of variables that ranged from cognitive skills to family and school-based risk factors. In addition, the combination of variables used in each study varied greatly across studies. Although various types of measures were effective in detecting children with RD in the early ages, it is not feasible to give a full complement of the assessments to each student. This approach is not cost or time effective. Therefore, it is important to identify the most promising combination of predictors of early reading

PAGE 114

104 achievement that directly relate to the reading process (Goswami & Bryant, 1990). Specifically, this study suggests that the psychological assessment for determining a reading fluency problem should include tests for rapid naming, phonological awareness, visual perception, and ocular motor functions. The primary findings of this study revealed that visual perception skills are significantly correlated with childrens reading fluency. Interestingly, reading researchers have ignored the role of nonlinguistic factors (e.g., visual factors) in their research. The current study suggests that future research on reading and language disorders should be aimed at identifying and characterizing all nonlinguistic perceptual deficits and determining the impact of those deficits on reading and language processing. Effective reading intervention for subtypes of RD Many researchers have reported that reading programs, such as Success for All, the Winston-Salem Project, The Boulder Program and Reading Recovery, were significantly successful in improving childrens reading level. (Clay, 1985, 1993a, 1993b; Hiebert, Colt, Catto, and Gury, 1992; Greaney, Tunmer, & Chapman, 1997). For example, Hiebert, Colt, Catto, and Gury (1992) reported that 77% of the students with reading difficulties who participated in their intervention project were able to achieve primer level reading at the end of first grade. In contrast, only 18% of a comparison group who participated in a traditional Title I program achieved that level of reading proficiency at the end of first grade. While almost half (47%) of the students in the conventional Title I program remained nonreaders at the end of first grade, only 7% of the early intervention students remained nonreaders. Similarly, other research findings suggest that these programs are significantly effective in improving the participants reading level (add other citations here). Nonetheless, these programs have failed to take a close look at

PAGE 115

105 children who showed no gains in spite of their extensive and additional support. Presumably, we can conclude that general reading intervention programs can help improve the reading levels of most children at risk for RD. However, these programs may not be responsive to all types of RD. During the last two decades researchers and practitioners in the reading research community have focused on phonological awareness training for the general population of children with RD. Overall, these phonolocigcal awareness programs have improved the reading achievement of children with RD. However, children who have reading problems due to related factors such as a visual processing deficits and visual decoding problems, have not been adequately served when interventions were solely based on a phonological awareness. The current study detected that there was a group of children who showed higher performances on phonological awareness and lower visual perception scores (approximately 1.5% of the sample) and showed lower scores on reading fluency tests. These children may have problems with visual perception and need additional support other than general reading interventions. For a subtype of children who show visual perception problems, as revealed in this study, additional support such as comprehensive vision screening should be given. Unfortunately, there is no known visual cause for this type of reading disability and no know effective visual treatment (Romanchuk, 1995). More recently, a few researchers have searched for subtypes of RD (e.g., Korhonen, 1991; Wolf et al., 1999). For example, Korhonen (1991) used cluster analyses to identify subgroups among third graders with learning disabilities. One subgroup was

PAGE 116

106 characterized by slow naming speed. When follow-up evaluations of the sample were made in sixth grade, it was found that this naming subgroup had shown the least progress in reading achievement, whereas the other children with other cognitive profiles had improved over the three intervening years. It is important to note that only a few studies on subtypes of RD have been conducted to date. Clearly researchers need to replicate these studies in order to determine subtypes of RD. With knowledge of various subtypes of RD, specific conditions or characteristics of individuals with RD can be identified and effective reading interventions can be provided for these children. Advanced technology Advanced technology has given us the opportunity to look closely into brain images and eye movement during reading. Most recently, Functional Magnetic Resonance Imaging (fMRI) has added important piece of information about the anatomy of the brain. fMRI is a technique to image brain activity related to a specific task or sensory process. As discussed earlier, it is now known that an impairment of the M pathway might affect the generalized timing mechanisms of the cerebellum (Fawcett et al., 2001). This scientific evidence implies that abnormal function in the specialized brain area V5/MT explains visual problems in people with reading disabilities. Interestingly, studies using fMRI on phonological deficit revealed that the phonological deficit can be traced back to a more general auditory impairment, which has the same biological origin as the visual impairment, namely, a dysfunction of magno-cells in sensory pathways. This tells us that there seems to be an overlap of areas of the brain that have to do with sensitivity to visual motion and phonological segmentation. If the M pathway is impaired

PAGE 117

107 in dyslexic brains, their inability to see visual motion of images on a screen is somehow connected to their weak phonological awareness. In addition, the Visagraph is another example of advanced technology. It is an eye-movement recording system and provides an objective measure of the efficiency of a student's reading processwhich reflects his/her visual skills. Furthermore, this system provides detail information on eye movement such as the number of fixations, number of regressions, length of fixation, rate of reading with comprehension and the overall performance in relation to grade level norms. As discussed, advanced technology provides us more clear pictures on brain function and eye movements during reading. However, this current study only used paper-pencil tests for assessing such skills since it was believed that paper-pencil tests are convenient to conduct and cost-effective. However, it would be more advantageous if conventional methods as well as new technology were combined for reading research. Future research should provide further insights into the details of visual and language deficits and their effects on reading. SEM This study was complex using data from twelve different measures. Therefore, a more sophisticated methodology was required in order to analyze all the data. Using SEM became a solution to removing potential problems that result from using too many variables for regression studies. SEM can be utilized very effectively to address numerous research problems. However SEM is not commonly used in reading research to identify factors related to childrens reading difficulties. The use of this model represents a significant contribution to the literature. As discussed in Chapter 3, structural equation modeling has several benefits. According to Fornell (1982), SEM includes several

PAGE 118

108 desirable characteristics that depart from the older generation of multivariate procedures. First, SEM takes a confirmatory approach to the data analysis rather than exploratory analysis, and it lends itself well to the analysis of data for inferential purposes by demanding that the pattern of intervariable relations be specified a priori. Second, SEM provides explicit estimates of either assessing or even correcting for measurement error, while traditional multivariate procedures are incapable of these aspects. Finally, data analyses using SEM can incorporate both unobserved and observed variables with the analyses using the traditional multivariate methods based on observed measurements only. On the other hand, there are some consideration that have to be taken seriously when conducting research studies by using SEM. Brecker (1990) recommended the following for detecting some deficiencies in applications of structural equation modeling. First, the data should be inspected for potential violations of the multivariate normality assumption. Distributional assumptions are likely to be violated when variables are highly skewed or when extreme outliers are present. Second, authors must make every effort to identify equivalent models and to discuss weather such models offer plausible representations of the data. To reduce potential violations on skewness or outliers effects, a large sample should be guaranteed for future research. Taken together, if researchers seriously consider assumptions when analyzing data, SEM can address many problems effectively and provide more reliable results compared to other statistical methods. Therefore, it would be beneficial to continue to apply SEM as a research methodology for future research.

PAGE 119

109 Implications for Practices The earlier the better Early recognition of reading difficulties and effective intervention to promote literacy skills are important in preventing life-long educational and social struggles. Stanovich (1986) first used the label, Matthew Effects, to describe how the rich get richer and the poor get poorer in terms of reading. Many studies suggest that a child who fails to make good initial progress in reading finds it increasingly difficult to master reading (add citations). Prevention and early intervention with appropriate teaching methods in early elementary grades can decrease the numbers of special education referrals. As discussed previously, a study by Lyon, et. al, (2001) estimates that the number of children, who are typically identified as poor readers and served through either special education or compensatory education programs, could be reduced by up to 70% through early identification and prevention. If this intervention is delayed until the student is 9 years old, which is when most children are referred to special education, approximately 75% will continue to have difficulties learning to read through high school (Francis et al., 1996). This implies that early identification and intervention of children at risk for RD are crucial for later reading success. The current study suggests that only one grade level difference makes a huge difference on skills related to reading. On all the measures that were used for the current study, third-grade students were superior to second graders. In addition, average scores on reading Fluency I and II of third graders were significantly higher than second graders. It implies that children develop their reading skills during early elementary years, and those years are a critical time for improving their knowledge of reading.

PAGE 120

110 In addition to attention in literacy instruction in kindergarten though third grade, ensuring high-quality preschool and kindergarten program and literacy environment is significantly important. Children who will probably need additional support for early language and literacy development should receive it as early as possible. Early literacy with at risk children Unfortunately, the current study demonstrated that, participants from two schools (School 4 & 5) in low-SES neighborhoods showed significantly lower scores on reading fluency skills compared to children from mid-SES neighborhoods (School 1, 2, & 5). Through reviewing the current and previous reading research, it was recognized that there are still many students who enter school without a strong literacy background. In particular, reading researchers have demonstrated that children from low-income families are less likely to have opportunities to build their basic literacy skills (Hart & Risley, 1995; Smith & Dixon, 1995; Whitehurst, 1996) and have higher risk for RD. Therefore, researchers and practitioners have recognized that low socio-economic status (SES) may be one of risk factors of RD, which is related to a lack of opportunity to learn. Hochenbergher, Goldstein, and Hass (1999) investigated whether teaching mothers with low socioeconomic status (SES) to comment while reading to their children would increase communicative interaction in mother--child dyads and improve the children s emerging literacy skills. They reported that most mothers also increased other conversational acts that set the occasion for positive verbal interactions and increased their responsiveness to their children through the education sessions. In turns,all children demonstrated more assertive and responsive utterances during the commenting intervention and four out of the seven children showed improved emerging literacy skills on a standardized. One way to achieve early literacy would be parent education for low

PAGE 121

111 SES families in order to increase their childrens reading skills and reduce children at risk for reading disabilities. Careful observations of visual problems This study suggests that some students may have problems with visual perception and processing. In general, many professionals and parents have a limited view regarding vision and tend to perceive that visual acuity is the only component of vision. Problems in identifying children with reading-related vision problems arise when such a limited definition of vision is employed. People who have a limited view of vision are more likely to exclude concerns on visual efficiency and visual processing. However, three interrelated areas of vision need to be incorporated in order to understand how letters or words are seen and perceived in the reading process (Scheiman, 2002). The two other areas are visual information processing (i.e., visual search, spatial awareness, figure-ground discrimination, visual memory) and visual efficiency (i.e., focusing, teaming, eye movements). To illustrate, although a child may have a normal range of visual acuity, he or she also may have visual problems related to other components of vision. In other words, 20/20 visual acuity does not tell us the childs vision problems such as near vision for near tasks (e.g., reading) and visual processing deficits. It merely provides information about far vision. Hence, it is beneficial for researchers, educators, and parents to have an extended definition of vision in order to identify children with reading-related visual problems. Some children with RD may show abnormal behaviors such as head tilting during near vision tasks (e.g., reading), a lack of eye-hand coordination during a ball game, having a difficulty to copy down the board, etc. Moreover, children who have problems in visual perception skills may complain of visual problems related to reading. For example, they may say that words jump around on the page, that they cannot

PAGE 122

112 keep their place, or that their eyes get tired. In some cases, these complaints may be avoidance behaviors or just symptomatic of task difficulty. However, there may be some cases where RD has a visual component, and such complaints should be carefully evaluated. In terms of reversal errors, it is not believed that children who demonstrate frequent reversals in the first year of reading instruction go on to become poor readers. However, if reversals persist even in the face of systematic instruction, they may be an indication for expert assessment. The current study discovered that approximately 30% of the sample is at risk for reversals, which may be a normal condition for secondand third-grade students. However, if such a problem consistently exists when they become fourth and fifth graders, this would be a strong indicator of a vision-related reading disability. Therefore, teachers, parents, and caregivers should carefully observe these behaviors to detect visual problems. Comprehensive vision screening policy The research findings in the current and previous literature provide evidence that at least a certain percentage of children suffer from RD related to a visual deficit. This study revealed that 3% to 5% of the sample showed problems on visual perception skills. It is possible to mention that a certain group of children who have visual deficits may not be respond to intervention programs that are solely based on linguistic components. The ideal solution is to assess the visual deficits that interfere with reading development through comprehensive vision screening checklist that could be simply administered by the classroom teacher. The implementation of comprehensive vision screening, covering visual processing and visual efficiency beyond visual acuity, would yield a score that predict the presence or absence of reading-related visual problems. However, current vision screening for young children is administered insufficiently as well as

PAGE 123

113 inconsistently across the states. A state-by-state survey, conducted by Cincer et al. (1999), revealed that there are still 15 states that do not have guidelines for visual screening prior to a childs entrance into formal education. Nevada and Wyoming do not have guidelines even for kindergarten vision screening. In addition, the areas tested in the vision screening vary from state to state. Mostly, they only include a distance acuity test and a limited version of a binocular vision test. These tests may be helpful to detect some refractive errors such as myopia and astigmatism. However, they obviously cannot detect many other vision problems including reading-related visual problems. The current study discovered 3% to 5% of 210 students showed problems on the tests of visual perception skills. Moreover, the results demonstrated that 30% of the sample are at risk of reversal problems with letters and numbers. Comprehensive vision screening during preschool and kindergarten years will be helpful to detect these types of visual deficits in early years. Therefore, it is essential to develop an effective screening test for detecting reading-related visual problems and other types of visual problems. Conclusions Reading is at the core of the educational experience and its ability is an essential key to success in our society. Our socially, economically, and technologically advanced society highly values the ability to read. As our society develops, there must be rising demands for literacy. In a technological society, the demands for higher literacy are further increasing. Therefore, a number of people, who do not read well enough to meet the high demands, no longer to have competitive careers. Of course, the current study revealed most children were able to read fairly well. However, some children are struggling with reading. Why do some children in our society have problems in reading? Ample research has suggested that phonological processing is a powerful cognitive

PAGE 124

114 indicator determining later reading ability. On the other hand, some other researchers argue that the phonological awareness factor may not be sufficient to explain all types of reading disabilities. For example, some researchers pointed out that rapid naming might be another powerful indicator in determining problems in reading (Wolf et al., 1999; Bowers et al., 2000) and some other researchers suggested that visual processing deficit could be one of important indicators of vision related reading problems (e.g. Eden et al, 1993). Therefore, along with phonological awareness, the current study searched for other significant variables that may be associated with reading fluency. Many connectionist theories, including Adams reading model, were addressed in order to discuss multidimensional components of reading. Moreover, empirical research provided evidence that there may be other aspects of reading that relate to RD. Through a review of the literature, it was noted that rapid-naming and visual deficits were significant variables in determining later reading development. However, there are some remaining issues and questions indicating key risk factors of RD surrounding rapid-naming and visual deficits. This review of the literature leaded me to further study prominent factors related to reading. Since components of reading are quite broad, this research narrowed down and focused only on factors related to reading fluency. For this study, twelve measures were administered, and data were analyzed by using descriptive and inferential statistics, correlational analysis and SEM to determine the effects of phonological awareness, rapid naming, visual factors on reading fluency. Taken together, this study provides the following conclusions. A Comprehensive Set of Indicators of RD Searching for comprehensive indicators of reading problems is important for early identification and early intervention. We have learned a great deal about the factors that

PAGE 125

115 interfere with the process of learning to read (Snow et al., 1998). Perhaps the greatest contribution in the past two decades is accumulating evidence of a phonological processing deficit as the core problem leading to poor reading (Stanovich & Siegel, 1994). Moreover, ample evidence suggests that this deficit can be addressed effectively through instruction, particularly if the problem is identified early (Foorman, Francis, Fletcher, Schatschneider, & Mehta, 1998). Consistent with prior research, the current study also revealed that phonological awareness is significantly related to reading fluency. Other than phonological awareness, naming speed and visual factors were taken into consideration in this study, and the results indicated that these two variables are also significant indicators of problems in reading fluency. The results particularly demonstrate that a combination of naming speed and eye movement measures is the most effective indicator to determine a problem with reading fluency. Surprisingly, the visual perception factor was the second most important factor affecting reading fluency. As discussed previously, prior literature indicated mixed findings on the relationship between visual deficits and reading. However, this study recognized that visual skills including eye movement, orthographic-reversal, and visual perceptions are significantly related to childrens reading fluency level. Therefore, a comprehensive set of indicators of RD should include consideration of phonological awareness as well as rapid-naming and visual factors. A Balanced Approach No single method will be successful with all children; a combination of many approaches may achieve the best results. Therefore, the most effective way to detect childrens reading problems and to provide reading instruction to all students can be through a balanced approach. The current study indicated that three critical components

PAGE 126

116 that affect childrens reading fluency are phonological awareness, rapid-naming, and visual factors. The results showed that each of these components noticeably affect reading ability, especially reading fluency. First, rapid naming was found to be the most critical factor affecting the rate of reading fluency. It is the ability to name visual symbols rapidly. Children must be able to apply the sound and letter knowledge at a quick rate while decoding printed words. Second, the auditory skills of blending, segmenting, and manipulating sounds, as well as perceiving the number and order of sounds within spoken words, are important phonological awareness skills that are necessary for reading. Finally, the ability to visually perceive the sequences and patterns of individually printed shapes, letters, and numbers is also essential for success in reading. In Chapter 2, connectionism was discussed as a theoretical framework of this current study. Philosophers have been interested in neural networks because they may provide a new framework for understanding the nature of the mind and its relation to the brain (Rumelhart and McClelland, 1986). Connectionist models seem significantly well matched to what we know about neurology. The brain is a neural net, formed from massively many units (neurons) and their connections (synapses). With the core of connectionism, several reading models have been suggested. In particular, Adams connectionist reading model was discussed in Chapter 2. According to Adamss connectionist reading model, all four processors need to perform a balanced work and operate simultaneously for skillful reading. If any one processor does not work effectively, the connected system would not operate successfully. The results of this study support connectionist theories in terms of phonological processing and orthographical processing. Phonological awareness is required for a phonological

PAGE 127

117 processor, and rapid-naming and visual perception skills are needed for an orthographic processor. Each processor should work cooperatively and simultaneously for successful reading. Beyond the basic level of processing, the upper level of processors on skillful reading also should be carefully considered. Early reading instruction focuses on teaching a child how to read single words that is related to the basic level of processing. But being good at reading single words is not the only skill your child needs. Being able to read fluently is the next critical developmental step for their later grades. Once they have become good at decoding single words, they need to learn to read easily, quickly and accurately, becoming fluent readers. They tend to read many individual words automatically and understand what they are reading. Needless to say, this is the ultimate goal for reading. It is hard to understand what you are reading when you are not a fluent reader because we need to use a lot of attention to decode single words and there is little left over to use understanding what we have read. Researchers have indicated that children who suffer from severe reading difficulties may not be fluent readers (e.g., Fawcett, Nicholson & Dean, 1996 & Nicholson & Fawcett, 1999). 'Dyslexic Automatization Deficit' (D.A.D.) is suggested by Nicholson & Fawcett at Sheffield University in the North of England (Fawcett, Nicholson & Dean, 1996 & Nicholson & Fawcett, 1999). Automatization is the process by which well learned skills become more and more fluent until the child no longer needs to think about them intentionally. They suggested that the two conditions are required for automatic processing. The first one is that it should be fast and the second one is that it needs little conscious effort (Nicholson & Fawcett, 1999). Nicholson and Fawcett (1990) believed that if conscious effort were involved performance would be decreased under dual-task conditions. Under this

PAGE 128

118 hypothesis, they examined the performance of twenty-three thirteen-year-old dyslexic children on a test of motor balance, and compared it to that of the same age controls. To test the "conscious compensation hypothesis" they introduced a secondary task to switch attention away from the primary task. Under single conditions the children performed just the motor balance (primary) task. Under dual conditions they performed the primary task as well as a concurrent secondary task. Dyslexic subjects showed problems on the dual task as opposed to the single task, but there was no such decrease in score for the control subjects. Fluency is essential because it plays an important role as a bridge between word recognition (the basic level of processing for reading; the phonological processor and the orthographical processor) and comprehension (the upper level of processing for reading; meaning processor and context processor). Therefore, a balanced approach based on connectionist theories is necessary to determine early detection of RD and choosing intervention strategies for children with reading problems.

PAGE 129

APPENDIX PARENT INFORMED CONCENT FORM Department of Special Education G-315 Norman Hall P.O. Box 117050 Gainesville, FL 32611-7050 Dear Parent/Guardian, I am a graduate student in the Department of Special Education at the University of Florida, conducting research on the effects of phonological awareness, rapid-naming, and visual skills on early elementary students reading under the supervision of Dr. Vivian Correa. Phonological awareness is the awareness that language is composed of sounds and the understanding of the relationship of these sounds. For assessing phonological awareness, manipulating sounds including deleting, adding, and substituting syllables/sounds will be assessed. For rapidnaming, the ability to efficiently retrieve verbal labels for familiar visual stimuli (e.g., numbers and letters) will be assessed. Within this study, visual skills including visual perceptual skills such as visual memory and visual discrimination and visual motor skills such as eye movement will be assessed. The purposes of this study are 1) to examine reading fluency, phonological awareness, rapid-naming, and visual skills, 2) to detect problems associated with these factors, and finally 3) to determine if phonological awareness, rapid-naming, and visual skills are related to childrens reading level. The test results will directly help your child. Through assessments of reading efficiency, phonological awareness, rapid-naming, and visual skills, problems in these areas will be detected. In the long run, early identification of problems in these areas is very important since they can be more effectively treated in the early stages. Totally, 20-25 minutes will be taken to accomplish all assessments. With your permission, I would like to ask your child to volunteer for this research. I will replace their names with code numbers. Your childs identity will be kept confidential to the extent provided by law. Results will only be reported in the form of group data. However, if you request the individual results of your childs assessments, they will be provided. Participation or non-participation in this study will not affect the childrens grades or placement in any programs. You and your child have the right to withdraw consent for your childs participation at any time without consequence. There are no known risks. Group results of this study will be available in August 2004. If you have any questions about this research protocol, please contact me at (352) 392-0701 ext.262 or my faculty supervisor, Dr. Vivian Correa, at (352) 392-0701. Questions or concern about your childs rights as research participant may be directed to the UFIRB office, University of Florida, Box 112250, Gainesville, FL 32611, (352) 392-0433. Thank you for your consideration. Sincerely, 119

PAGE 130

120 Hoyeon Kim I have read the procedure described above, I voluntarily give my consent for my child, _________________, to participate in H oyeon Kim’s study of the effects of phonological awareness, rapid-naming, and visual skills on early elementary students’ reading. I have received a copy of this description. __________________________________ _________________________ Parent/Guardian Date __________________________________ _________________________ 2nd Parent /Witness Date

PAGE 131

121 LIST OF REFERENCES Aaron, P. G., & Joshi, R. M. (1992). Reading problems : Consultation and remediation New York: The Guilford Press. Adams, M. J. (1990). Beginning to read: Thinki ng and learning about print. Cambridge, MA: MIT Press. Adams, M. J., Foorman, B. R., Lund berg, I., & Beeler, T. (1998). Phonemic awareness in young children. Baltimore: Brookes. Anderson, M., Hiebert, E. H., Scott, J. A., & Wilkinson, A. G. (1985). Becoming a Nation of Readers: The Repor t of the Commission on Reading. Washington, DC: National Institute of Education, U.S. Department of Education. Badian, N. A. (1994). Preschool predicti on: Orthographic and phonological skills, and reading. Annals of Dyslexia, 44, 3-25. Badian, N. A. (1997). Dyslexia and the double deficit hypothesis. Annals of Dyslexia, 47 69-87. Badian, N. A. (1998). A validation of the role of preschool phonological and orthographic skills in th e prediction of reading. Journal of Learning Disabilities, 31 472-481. Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin 107 238-246. Bentler, P. M., & Bonett, D. G. (1980). Significance tests and goodness of fit in the analysis of covariance structures. Psychological Bulletin 88 588-606. Bentler, P. M., & Chou, C. P. (1987). Pr actical issues in structural modeling. Sociological Methods and Research, 16 78-117. Blachman, B. (1991). Phonological awareness: Implications for pre-reading and early reading instruction. In S. A. Br ady & D. P. Shankweiler (Eds.), Phonological processes in literacy: A tri bute to Isabelle Y. Liberman (pp. 37-45). Hillsdale, NJ: Lawrence Erlbaum Associates. Blachman, B. A. (2000). Phonological awareness In M. L. Kamil, P. B. Mosenthal, P. D. Pearson, & R. Barr (Eds.), Handbook of reading research (pp. 251-284). Mahwah, NJ: Lawrence Erlbaum Associates.

PAGE 132

122 Boden, C., & Brodeur, D. A. ( 1999). Visual processing of verb al and nonverbal stimuli in adolescents with reading disabilities. Journal of Learning Disabilities, 32 58-71. Bollen, K. A. (1989). Structural equations with latent variables New York: John Wiley & Sons. Bollen, K. A., & Long, J. S. (1993). Testing structural equation models Newbury Park, CA: Sage Publications. Bowers, P. G., & Swanson, L.B. (1991). Nami ng speed deficits in reading disability: Multiple measures of a singular process. Journal of Experimental Child Psychology, 51 195-219. Bowers, P. G., & Wolf, M. (1993). Theoreti cal links among naming speed, precise timing mechanisms and orthographic skill in dyslexia. Reading and Writing: An Interdisciplinary Journal 5 69-85. Browne, M. W., & Cudeck, R. (1993). Altern ative ways of asse ssing model fit. In: Bollen, K. A., & Long, J. S. (Eds.) Testing structural equation models (pp. 136 162). Beverly Hills, CA: Sage. Byrne, B. (1992). Studies in the acquisition pr ocedure for reading: Rationale, hypotheses, and data. In P. Gough, L, Eh ri, & R. Treiman (Eds.), Reading acquisition (pp. 134). Hillsdale, NJ: Lawrence Eflbaum Associates. Byrne, B. M. (1998). Structural equation modeling w ith Lisrel, Prelis, and Simplis New York: Lawrence Erlbaum. Carver, R. P. (1990). Reading rate: A review of research and theory San Diego, California: Academic Press, Inc. Casco, C., Tressoldi, P. E., & Dellantonio, A. (1998). Visual selective attention and reading efficiency are related in children. Cortex, 34 531-546. Catts, H. W. (1991). Early identification of reading disabilities. Topics in Language Disorders, 12 1-16. Catts, H. W., Fey, M.E., Zhang, X., and Tombli n, J. B. (1999). Language basis of reading and reading disabilities: Evidence from a longitudinal investigation. Scientific Studies of Reading, 3 331-361. Chan, K. S., & Cole, P. G., & Barfett, S. (1987). Comprehension monitoring: Detection and identification of text inconsis tencies by LD and normal students. Learning Disability Quarterly, 10 114-124. Chaney, C. (1992). Language development, meta linguistic skills, and print awareness in 3-year-old children. Applied Psycholinguistics. 13 485-514.

PAGE 133

123 Chard, D., & Dickson, S. V. (1999). Phonol ogical awareness: Instructional and assessment guidelines, Intervention in School and Clinic, 34 261-270. Ciner, E., Dobson, V., Schmidt, P., Allen, D ., Cyert, L., Maguire, M., Moore, B., OrelBixler, D., Schultz, J. (1999). A survey of vision screening policy of preschool children in the United States. Survey Ophthalmology, 43, 445-457. Clark, D. B., & Uhry, J. K. (1995). Dyslexia: Theory and practice of remedial instruction Baltimore, MD: York Press. Clay, M. M. (1985). The early detection of reading difficulties Portsmouth, NH: Heinemann. Clay, M. M. (1993a). An observation survey of early literacy achievement. Portsmouth, NH: Heinemann. Clay, M. M. (1993b). Reading recovery: A guidebook for teachers in training. Portsmouth, NH: Heinemann. Deeney, T., Gidney, C., Wolf, M., & Morris, R. (1999). Phonological skills of AfricanAmerican reading-disabled children. Paper presented at 6th Annual Meeting of the Society for the Scientific Study of Reading. Montreal, Canada. De Jong, P. F., & van der Leij, A. (2003). Deve lopmental changes in the manifestation of a phonologicaldeficit in dyslexic children le arning to read a regular orthography. Journal of Educational Psychology, 95 22. Denckla, M. B., & Rudel, R. G. (1976). Ra pid automatized naming (R.A.N.): Dyslexia differentiated from other learning disabilities. Neuropsychologia, 14 471-479. Denckla, M. B., & Rudel, R. G. (1974). Ra pid automatized maing of prictured objects, colors, letters, and numbers by normal children, Cortex, 10 186-202. Di Lollo, V., Hanson, D., & McIntyre, J. S. ( 1983). Initial stages of visual information differences in the acquisition of literacy. Reading Research Quarterly, 21 360-406. Dole, J., Duffy, G., Roehler, L., & Pearson, P. (1991). Moving from the old to the new: Research on reading comp rehension instruction. Review of Educational Research, 61, 239-264. Duke, N. (2000). 3.6 minutes per day: The scarc ity of informational texts in first grade. Reading Research Quarterly, 35 202-224. Eden, G. F., Stein, J. F., & Wood, F. B. (1993). Visuospatial ability and language processing in reading disabled and normal children. In S. F. Wright & R. Croner (Eds.), Facets of dyslexia and its remediatio n: Studies in visual information processing Amsterdam: North-Holland.

PAGE 134

124 Eden, G. F., Stein, J. F., Wood, M. H., & W ood, F.B. (1995). Verbal and visual problems in reading disabilities. Journal of Learning Disabilities, 28 272-290. Eden, G.F., & Zeffiro, T. A. (1996). The visual deficit theory of de velopmental dyslexia. NeuroImage, 4 108-117. Eden, G. F., & Zeffiro, T. A. (1998). Looking beyond the reading difficulties in dyslexia, a vision deficit. Biopsychology 98/99 B. M. Jubilan. Guilford, Dushkin/McGrawHill. Ehri, L. C. (1991). Learning to read and spel l words. In L. Rieben & C.A. Perfetti (Eds ), Learning to read: Basic rese arch and its implications (pp. 57-73). Hillsdale, NJ: Lawrene Erlbaum Associates. Ehri, L. C. (1992). Reconceptualizing the de velopment of sight word reading and its relationship to recoding. In P. G ough, L. Ehri, & R. Treiman (Eds.), Reading acquisition (pp. 107-143). Hillsdale, NJ: Lawr ence Erlbaum Associates. Ehri, L. C. (1995). Phases of developm ent in learning to read words by sight. Journal of Research in Reading, 18 116-125. Ehri, L. C. (1998). Word reading by sight a nd by analogy in beginning readers. In C. Hulme & M. Joshi (Eds.), Reading and spelling: Development and disorders (pp. 87-112). Mahwah, NJ: Erlbaum. Ehri, L. C. (1999). Phases of development in learning to read words. In J. Oakhill & R. Beard (Eds.), Reading development and the teachi ng of reading: A psychological perspective (pp. 79-108). Oxford, UK: Blackwell Publishers. Ehri, L. C., & Saltmarsh, J. (1995). Beginning readers outperform older disabled readers in learning to read words by sight. Reading and Writing, 7, 295-326. Facoetti, A., Paganoni, P., Turatto, M., Marzol a, V., & Mascetti, G. G. (2000). Visualspatial attention in developmental dyslexia. Cortex, 36 109-123. Farmer, M., & Klein, R. (1995). The evidence for a temporal proces sing deficit linked to dyslexia: A review Psychonomic Bulletin and Review, 2, 460-493. Fawcett, A. J., Nicolson, R. I., & Maclaga n, F. (2001). Cerebellar tests may differentiate between poor readers with a nd without IQ discrepancy. Journal of Learning Disabilities, 24, 119-135. Feagans, L. V., & Merriwether, A. (1990). Vi sual Discrimination of letter-like forms and its relationship to achievement over time in children with learning disabilities. Journal of Leaning Disabilities, 23, 417-425. Felton, R. H. (1992). Early identification of children at risk for reading disabilities. Topics in Early Childhood Special Education, 12 212-229.

PAGE 135

125 Felton, R. H., & Brown, I. S. (1990). Phonologi cal Processes as Predictors of Specific Reading Skills in Children at Risk of Reading Failure. Reading and Writing: An Interdisciplinary Journal 2 39-59. Fisher, B., & Weber, H. (1990) Saccadic r eaction times of dyslexic and age-matched normal subjects. Perception 19, 805-818. Fletcher, J. M., & Lyon, G. R. (1998). Readi ng: A research-based approach. In W.M. Evers (Ed.), Whats gone wrong in Americas classrooms (pp. 49-90) Stanford, CA: Hoover Institute Press. Foorman, B. R., Francis, D. J ., Fletcher, J. M., Schatschne ider, C., & Mehta, P. (1998). The role of instruction in learning to read: Preventing reading failure in at-risk children. Journal of Educational Psychology, 90 37-55. Fountas, I. C., & Pinnell, G. S. (1996). Guided reading: Good first teaching for all children. Portsmouth, NH: Heinemann. Francis, D. J., Shaywitz, S. E., Stuebing, K. K., Shaywitz, B. A., & Fletcher, J. M. (1996). Developmental lag versus defic it models of reading disability: A longitudinal, individual growth curves analysis. Journal of Educational Psychology, 88, 3-17. Fuchs, L., Fuchs, D., & Maxwell, L. (1988). Th e validity of informal measures of reading comprehension. Remedial and Special Education 9 20-28. Gardner, M. F.(1994). Manual for the test of visual -perceptual ski lls-upper level. Hydseville, CA: Psychological and Educational Publications Inc. Garzia, R. P., Richman, J. E., Nicholson, S. B., & Gaines, C. S. (1990). A new visualverbal saccade test: The Developmental Ey e Movement test (DEM). Journal of the American Optometric Association, 2, 124-125. Getman, G. N. (1981). Vision: Its role and integrations in learning processes. Journal of Learning Disabilities 14, 577580. Good, R. H., Simmons, D. C., & Kame'enui, E. J. (2001). The importance of decisionmaking utility of a continuum of fluenc y-based indicators of foundational reading skills for third-grade high stakes outcomes. Scientific Studies of Reading 5, 257288. Goswami, U. (1993). Toward an interact ive analogy model of reading development: vowel graphemes in beginning reading. Journal of Experimental Student Psychology, 54 443-475. Goswami, U., & Bryant, P. (1990). Phonological skills and learning to read. Hillsdale, NJ: Lawrence Erlbaum Associates.

PAGE 136

126 Gough, P. B., Juel, C., and Griffith, P. L. (1992). Reading, spelling, and the orthographic cipher. In P. B. Gough, L. C. Ehri, & R. Treiman (Eds.), Reading acquisition (pp. 35-48). Hillsdale, NJ: Lawrence Erlbaum Associates. Greaney, K., Tunmer, W., & Chapman, J. W. (1997). The use of rime-based orthographic analogy training as an interv ention strategy for reading disabled children. In B. Blachman (Ed.), Foundations of reading acquisition and dyslexia: Implications for early intervention (pp. 327-345). Hillsdale, NJ: Lawrence Erlbaum Associates. Grossen, B. (1997). Thirty years of research: What we know about how children learn to read: A synthesis of research on readi ng from the National Institute of Child Health and Human Development. Santa Cruz, CA: Center for the Future of Teaching and Learning. Habib, M. (2000). The neurological basis of developmental dyslexia. Brain 123 2373 2399. Hart, B., & Risley, T. (1995). Meaningful differences in everyday parenting and intellectual development in young American children Baltimore: Brookes. Heinemann. Hiebert, E. H., Colt, J. M., Catto, S. L., & Gury, E. C. (1992). Reading and writing of first-grade students in a restructured Chapter I program, American Educational Research Journal, 29, 545-572. Hochenbergher, E. H., Goldstein, H., & Ha ss, L. S. (1999). Effects of Commenting During Joint Book Reading by Mothers with Low SES, Topics of Early Childhood Special Education, 19, 15-27. Hoover, W., & Gough, P. (1990). Th e simple view of reading. Reading and Writing: An Interdisciplinary Journal, 2 127-160. Hoyle, R. H. (1995). Structural equation modeling: C oncepts, issues, and applications. Thousand Oaks: Sage Publications. Hu, L., & P. M. Bentler (1999). Cutoff criteri a for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling 6, 1-55 Johnson, R. A., Nottinghamn, D., Stratton, R., & Jaba, J. N. (1996). The vision screening of academically and behaviorally at risk pupil, Journal of Behavioral Optometry 7, 67-71. Jordan, B. T. (1980). Jordan left-right reversal test, Novato, CA : Academic Therapy Pub. Joreskog, K. G., & Sorbom, D. (1996). LISREL 8: User's Reference Guide Chicago, IL: Scientific Software International.

PAGE 137

127 Jorn, A., Share, D., McLean, R., & Matth ew, R. (1984). Phonological recoding and learning to read: A longitudinal study. Applied Psycholinguistics, 5, 210-207. Juel, C. (1988). Learning to read and write : A longitudinal study of 54 children from first through fourth grades. Journal of Educational Psychology, 80 437-447. Kavale, K. (1982). Meta-analysis of the relati onship between visual pe rceptual skills and reading achievement, Journal of Learning Disabilities, 15 42-51. Keogh, B. D., & Pelland, M. ( 1985). Vision training revisited. Journal of Learning Disabilities, 18 228-236. Kirby, J. R., & Parrila, R. K. (1999). Th eory-based predicti on of early reading. Alberta Journal of Educational Research, 45 428-447. Kline, R. B. (1998). Principles and practice of st ructural equation modeling NY: Guilford Press. Korhonen, T. T. (1991). Neuropsychological stability and prognos is of subgroups of children with learning disabilities Journal of Learning Disabilities, 24 48-57. Leu, D. J., Jr., & Kinzer, C. K. (1995). Instructional practices in reading: A video to accompany the book, Effective Reading Instruction: K-8 (3rd ed.). Englewood Cliffs, NJ: Prentice Hall, Inc. Liberman, I. Y. (1973). Segmentation of the spoken word. Bulletin of the Orton Society 23: 65-77. Liberman, I. Y., Shankweiler, D., Fischer, F. W., & Carter, B. (1974). Explicit syllable and phoneme segmentation in the young child. Journal of Experimental Child Psychology, 18, 201-212. Livingstone, M. (1993). Parallel processing in the visual sy stem and the brain: Is one subsystem selectively affected in dys lexia? In A. M. Galaburada (Ed.), Dyslexia and development: Neurobiological aspects of extra-ordinary brains (pp. 237-256). Cambridge, MA: Harvard University Press. Loehlin, J. C. (1992). Latent variable models: An introduction to factor, path, and structural analysis. Hillsdale, NJ: Lawrence Erlbaum. Lovegrove, W. J. (1992). The visual deficit hypothesis, In N. Singh & I. Beale (Eds.), Learning Disabilities: Nature, Theory and Treatment New York: Spinger-Verlag. Lovegrove, W., Garzia, R., & Nicholson. S. ( 1990). Experimental ev idence for a transient system deficit in specific reading disability. Journal of American Optometry Association, 2, 137.

PAGE 138

128 Lovett, M. W., Steinbach, K. A., & Frijters, J. C. (2000). Remediating the core deficits of developmental reading disability : A double-deficit perspective. Journal of Learning Disabilities, 33 334-358. Lundberg, I., Olofsson, A., & Wall, S. (1980). R eading and spelling skill in the first school years predicted from phonemic aw areness skills in kindergarten. Scandinavian Journal of Psychology, 21 159-173. Lyon, R., & Fletcher, J. M. (2001). Early warning system Education Matter 23-29. Lyon, G. R., Fletcher, J. M., Shaywitz, S. E ., Shaywitz, B. A., Torgesen, J. K., Wood, F. B., Schulte, A., & Olson, R. (2001). Rethinki ng learning disabilities. In C. E. Finn, A. J. Rotherman, & C. R. Hokanson (Eds.), Rethinking special education for a new century (pp. 259-287). Washington, DC: Thomas B. Fordham Foundation and the Progressive Policy Institute. MacCallum, R., Browne, M., & Sugawara H. M. (1996). Power analysis and determination of sample size fo r covariance structure modeling, Psychological Methods, 2 130-149. Manis, F. R., Doi, L. M., & Bhadha, B. (2000). Naming speed, phonological awareness, and orthographic knowledge in second graders. Journal of Learning Disabilities, 33 325-333. Manis, F. R., Seidenberg, M. S., & Doi, L. M. (1999). See Dick RAN: Rapid naming and the longitudinal prediction of reading subskills in first and second graders. Scientific Studies of Reading, 3, 129-158. Mann, B. (1991). Language problems: A key to early reading problems, In B. Y. L. Wong (Ed.), Learning about lear ning disabilities (pp. 129-162). New York: Academic Press. Mann, V. A. (1993) Phonological skills and the prediction of early reading performance, In N. Jordan & J. Goldsmith-Philips (Eds.) Learning disabilities: New directions for assessment and intervention (pp. 67-83). Bosto n, MA: Allyn & Bacon. Mann, V. A., & Brady, S. B. (1988) Read ing disability: The role of language deficiencies. Journal of Consulting and Clinical Psychology, 56 811-816. Manzo, A. V., & Manzo, U. C. (1993). Literacy disorders: Holistic diagnosis and remediation. Fort Worth: Harcourt Brace Jovanovich College Publishers. Maria, K. (1990). Reading comprehension instru ction, issues and strategies Parkton, MD: York Press

PAGE 139

129 McClelland, J. L., Rumelhart, D. E., & Hi nton, G. E. (1986). The appeal of parallel distributed processing. In D. E. Ru melhart & J. L. McClelland (Eds), Parallel Distributed Processing: Explorations in the Microstructure of Cognition (p. 3-44). MIT Press. McGurk H., & McDonald, J. W. (1976). Hearing lips and seeing voices. Nature, 264 746-748 McKinney, J. D., Short, E. J., & Feagans, L. (1985). Academic consequences of perceptual-linguistic subtypes of learning disabled children. Learning Disabilities Research, 1, 6-17. Meyer, M., & Felton, R. (1999). Repeated read ing to enhance fluency: Old approaches and new directions. Annals of Dyslexia. 49 283-306. Meyer, M. S., Wood, F. B., Hart, L. A., & Felton, R. H. (1998). Selective predictive value of rapid automatized naming in poor readers. Journal of Learning Disabilities, 32, 106-117. Moats, L. C., & Lyon, G. R. (1993). Learning disabilities in the United States: Advocacy, science, and the future of the field. Journal of Leaning Disabilities, 26 282-294. Morad, Y., Lederman, R., Avni, I., Atzmon, D., Azoulay, I., & Segal, O. (2002). Correlation between reading skills and di fferent measurements of convergence amplitude, Current Eye Research, 25, 117-121. Morris, R. D., Stuebing, K. K., Fletcher J. M., Shaywitz, S. E., Lyon, G. R., Shankweiler, D. P., Katz, L., Francis, D. J., & Shaywitz, B. A. (1998). Subtypes of reading disability: Variabili ty around a phonological core. Journal of Educational Psychology, 90 347-373. National Assessment of Educati onal Progress. (NAEP). (2001). The NAEP reading report card Washington, DC: US Department of Education, Office of Educational Research and Improvement, National Ce nter for Educational Statistics. National Reading Panel. (2000). Teaching children to read. Retrieved Octover, 3, 2003, from http://www.nationalreadingpanel.or g/ publications/summary.htm Neuhaus, G. F., & Swank, P. R. (2002). Unde rstanding the relations between RAN letters subtest components and word read ing in first grade students. Journal of Learning Disabilities, 2, 76-89. Newcomb, M. D., & Bentler, P. M. (1988). Im pact of adolescent drug use and social support on problems of young adults: A longitudinal study. Journal of Abnormal Psychology 97, 64-75. Nicholson, R. I. & Fawcett, A.J. (1990). A New Framework for Dyslexia Research? Cognision, 35, 159-182.

PAGE 140

130 Nicholson, R. I. & Fawcett, A.J. (1999). Developmental Dyslexia: The Role of the Cerebellum. Dyslexia, 5 155-177. OConnor, R. E., Jenkins, J. R., Leicester, N., & Slocum, T. A. (1993). Teaching phonological awareness to young children with learning disabilities. Exceptional Children, 59, 532-546. Olsen, R., Forsberg, H., & Wise, B. (1994). Ge nes, environment, and the development of orthographic skills. In V. W. Berninger (Ed.), The varieties of orthographic Knowledge I: Theoretical and developmental issues (pp. 27 -71). Dordrecht, the Netherlands: Kluwer Academic. Pearson, P. D., & Fielding, L. (1991). Comprehe nsion instruction. In R. Barr, M. Kamil, P. Mosenthal, & P. D. Pearson (Eds.), Handbook of reading research: Volume II (pp. 815-860). White Plains, NY: Longman. Perry, R. J., & Zeki, S. (2000). The neurology of saccades and covert shifts in spatial attention: An event-related fMRI study. Brain, 123, 2273-2288. Pollatsek, A. (1993). Eye movement in reading. In D. M. Willows, R. S. Kruk, & E. Corcos (Eds.), Visual process in reading and reading disabilities (pp. 191-214) Hillsdale, NJ: Lawrence Erlbaun Associates. Pressley, M. (2000). What shoul d comprehension instruction be the instruction of? In M. L. Kamil, P. B. Mosenthal, P. D. Pearson, & R. Barr (Eds.), Handbook of reading research : Volume III (pp. 545-561). Mahwah, NJ: Erlbaum. Pugh, K., Mencl, E. W., Shaywitz, B. A., Shay witz, S. E., Fulbright, R. K., Skudlarski, P.,Constable, R. T., Marchione, K., Jenner A.R., Shankweiler, D. P., Katz, L., Fletcher, J., Lacadie, C., & Gore, J. C. (2000). The angular gyrus in developmental dyslexia: Task-specific differences in f unctional connectivity in posterior cortex. Psychological Science, 11, 51-56. Rayner, K. (1995). Eye movements and cogniti ve processes. Review of perception and cognition: Advances in eye movement research. American Journal of Psychology, 66, 460-464. Rayner, K. (1997). Understandi ng eye movements in reading. Scientific Studies of Reading, 1 317-339. Rayner, K. (1998). Eye movements in readi ng, visual search and scene perception: 20 years of research. Psychological Bulletin, 124, 372-422. Richards, G. P., Samuels, S. J., Turnure, J. E., & Yselldyke, J. E. (1990). Sustained and selective attention in children with learning disabilities. Journal of Learning Disabilities, 23, 129-136.

PAGE 141

131 Romanchuk, K.G. (1995) Skepticism about Irlen filters to treat learning disabilities. Canadian Medical Association Journal, 153 397-398 Rubin, H., Rotella, T., Schwartz, L., & Bern stein, S. (1991). The effect of phonological analysis training on naming performance. Reading and Writing: An Interdisciplinary Journal, 3, 1-10. Scarborough, H. S. (1998). Predicting the future achievement of second graders with reading disabilities: Contributions of phonemic awareness, verbal memory, rapid naming and IQ. Annals of Dyslexia, 48 115-136. Scarborough, H. S., Dobrich, W.., & Mhager M. C. (1991). Pr eschool literacy experiences and later reading achievement. Journal of Learning Disabilities. 24 508-511. Schatschneider, C., Carlson, C.D. Francis, D.J., Foorman, B.R., & Fletcher, J.M.(2002) Relationship of Rapid Automatized Nami ng and Phonological Awareness in Early Reading Development: Implications for the Double Deficit Hypothesis. Journal of Learning Disabilities 35, 245-256 Scheiman, M. (2002). Understanding and Managing Visual Deficits: A guide for occupational therapists (2nd ed.). Slack Inc, Philadelphia, PA. Seidenberg, M. S. (1992). Dyslexia in a co mputational model of word recognition and reading. In P. Gough, R. Treiman, & L. Ehri (Eds .), Reading acquisition (pp. 243274). Hillsdale, NJ: Erlbaum. Seidenberg, M. S., & McCelland, J. (1989). A distributed developmental model of word recognition and naming. Psychological Review, 96 35-49. Shaywitz, B. A., Shaywitz, S. E., Pugh, K., Menel, E., Fulbright, R., Skudlarski, P., Constable, R., Marichone, K., Fletcher, J. M., Lyon, G. R., & Gore, J. C. (2000). Disruption of posterior brain systems for reading for children with developmental dyslexia. Society of Biological Psychiatry, 52, 100-110. Shaywitz, S. (1996). Dyslexia. Scientific American 412. 98-104. Shinn, M. R., Good, R. H. Knutson, N., Tilly, W.D., & Collins, V. L. (1992). Curriculum-based measurement of oral readi ng fluency: A confirmatory analysis of its relation to reading. School Psychology Review 21 459-479. Siegel, L. S. (1989). IQ is irrelevant to the definition of learning disabilities. Journal of Learning Disabilities, 22, 469-479. Slavin, R. E., Madden, N. A., Dolan, L., & Wasik, B. A. (1996). Every child, every school: Success for all Thousand Oaks, CA: Corwin Press.

PAGE 142

132 Slavin, R. E., Madden, N. A., & Wasik, B. A. (1996). Success for all and roots and wings: Summary of research on achievement outcomes. Baltimore, MD: Johns Hopkins University, Center for Research on the Education of Students Placed at Risk. Smith, S. B., Simmons, D. C., & Kameenui E. J. (1998). Phonological awareness: Instructional and curricular basics and implications. In D. C. Simmons & E. J. Kameenui (Eds.), What reading research tells us about children with diverse learning needs (pp. 129-140). Mahwah, NJ: Lawrence Erlbaum Associates. Smith, S. S., & Dixon, R. (1995). Literacy conc epts that lowand middle-class four-yearolds bring to preschool. Journal of Educational Research 88 243-253. Snow, C., Burns, S., & Griffin, P. (1998). Preventing reading difficulties in young children Washington, DC: National Academy Press. Stanovich, K. E. (1982). Indivi dual differences in the cogni tive processes of reading: Word decoding. Journal of Learning Disabilities, 15, 485-493. Stanovich, K.E. (1986). "Matthew effects" in reading: Some consequences of individual differences in acquisition of literacy. Reading Research Quarterly 4 360-407. Stanovich, K. E. (1988). The right and wrong places to look for the cognitive locus of reading disability. Annals of Dyslexia, 38 154-157. Stanovich, K. E. (1991). Discrepancy definitions of reading disabili ty: Has intelligence led us astray? Reading Research Quarterly, 26 7-29. Stanovich, K.E. (1992). Speculations on the cau ses and consequences of individual differences in early reading acquisition. In P. B. Gough, L. C. Ehri, & R. Treiman (Eds.), Reading acquisition (pp. 307-342). Hillsdale, NJ: Erlbaum. Stanovich, K. E. (2000). Progress in understanding reading New York: Guilford Press. Stanovich, K. E., & Siegel, L. S. (1994). Phe notypic performance profile of children with reading disabilities: A regr ession-based test of th e phonological-core variabledifference model. Journal of Educational Psychology, 86, 24-53. Stevens, J. (1996). Applied multivariate statistics for the social sciences (3rd ed.). Hillsdale, NJ: Lawrence Erlbaum Associates. Stringer, R., & Stanovich, K. E. (2000). The connection between reaction time and variation in reading ability: Unravelli ng covariance relationships with cognitive ability and phonological sensitivity, Scientific Studies of Reading, 4 41-53.

PAGE 143

133 Tallal, P., Miller, S., & Fitch, R. H. (1993). Neurobiological basis of speech: A case for the preeminence of temporal processing. In P. Tallal, A. M. Galaburda, R. Llinas, & C. von Euler (Eds.), Temporal information processing in the nervous system (pp. 27-47) New York: The New York Academy of Sciences. Temple, E., Deutsch, G., Poldrack, R., Miller, S., Tallal, P., Merzenic h, M., & Gabrieli, J. (2003). Neural deficits in children w ith dyslexia ameliorated by behavioral remediation: Evidence from functional MRI. Proceedings from the National Academy of Sciences, 100 2860-2865. Temple, E., Poldrack, R. A., Klingberg, T ., Hedehus, M., Moseley, M., Merzenich, M. M., Tallal, P., & Gabrieli, J. D. (2000) Functional and structural brain differences in dyslexia: Evidence for a biological basis. Presentation at the International Dyslexia Association Annual Conferen ce, November 2000, Washington, DC. Torgesen, J. K. (2000). Individual responses in response to early interv entions in reading: The lingering problem of treatment resisters. Learning Disabilities Research and Practice, 15 55-64. Torgesen, J. K., & Bryant, B. R. (1993). Phonological awareness training for reading. Austin, TX: ProEd. Torgeson, J., & Bryant, B. (1994). Test of phonological awarenes s, Austin, TX: Pro-Ed. Torgesen, J. K., & Burgess, S. R. (1998). Consistency of readi ng-related phonological processes through early chil dhood: Evidence from longitudi nal studies. In J. L. Metsala & L. C. Ehri (Eds.). Word Recognition in Beginning Reading (pp. 161188) Mahwah, NJ: Lawrence Erlbaum Associates. Torgesen, J. K., Wagner, R. J., & Rashotte C. A. (1994). Longitudinal studies of phonological processi ng and reading. Journal of Learning Disabilities, 27 276286. Torgesen, J. K., Wagner, R. K., & Rashotte, C. A. (1997). Prevention and remediation of severe reading disabilities : Keeping the end in mind, Scientific Studi es of Reading, 1, 217-234 Torgesen, J. K., Wagner, R. K., Rashotte, C. A., Burgess, S., & Hecht, S. (1997). Contributions of phonological awareness and automatic naming ability to the growth of word-reading skills in secondto fifth-grade children. Scientific Studies of Reading, 1, 161-185. Torgesen, J. K., Wagner, R. K., Simmons, K., & Laughon, P. (1990). Identifying phonological coding problems in disabled readers: Naming, counting, or span measures? Learning Disability Quarterly 13 236-243. Vacca, R. T., & Vacca, J. L. (1996). Content area reading (5th ed.). New York: Harper Collins.

PAGE 144

134 Vellutino, F. R. (1979). Dyslexia: Research and theory Cambridge, MA: MIT Press. Vellutino, F. R. (1983). Dyslexia; Perceptual defi ciency of perceptual inefficiency. In J. P. Kavanagh & R. L. Venezky (Eds.). Orthography, reading, and dyslexia Baltimore: University Park Press. Vellutino, F. R. (1987). Dyslexia. Scientific American, 256, 34-41. Vellutino, F. R., & Scanlon, D. M. (1987) Phonological coding, phonological awareness and reading ability: Evid ence from a longitudinal and experimental study. MerrillPalmer Quarterly 33 (3), 321-363. Vellutino, F. R., Scanlon, D. M., & Tanzma n, M. S. (1994). Components of reading ability: Issues and problems in operati onalizing word identi fication, phonological coding, and orthographic codi ng. In G. R. Lyon (Ed.), Frames of reference for the assessment of learning disabilities : New views on measurement issues (pp. 279324). Baltimore: Brookes. Wagner, R. K., & Torgeson, J. K. (1987). Th e nature of phonologica l awareness and its causal role in the acquisi tion of reading skills. Psychological Bulletin, 101 192212. Wagner, R. K., Torgeson, J. K., & Rashotte, C. A. (1999). The comprehensive test of phonological processing: Examiners manual. Austin, Texas: Pro-Ed. Walker, B. J. (1996). Diagnostic teaching of reading: Techniques for instruction and assessment. Englewood Cliffs, NJ: Merrill. Watson, C., & Willows, D. M. (1993). Evidence for a visual-processing-deficit subtype among disabled readers. In D. M. Willows, R. S. Kruk, & E. Corcos (Eds.), Visual processes in reading and reading disabilities (pp. 287-309) Hillsdale, NJ: Lawrence Erlbaum. Watson, C. (1990). Specific processing difficulties in developmental reading disability among elementary school children. Unpublished doctoral dissertation, OISE, University of Toronto. Watson, C. S., Kidd, G. R., Horner, D. G., C onnell, P. J., Lowther, A., Eddins, D.A., Krueger, G., Goss, D. A., Rainey, B. B ., Gospel, M. D., & Watson, B. U. (2003). Journal of Learning Disabilities, 36, 165-197. Whitehurst, G.J. (1996). A structural equation model of the role of home literacy environment in the development of emergent literacy skills in children from lowincome backgrounds. Paper presented at the annual convention of the American Educational Research Association, New York.

PAGE 145

135 Wimmer, H., Mayringer, H., & Landerl, K. (2000). The double deficit hypothesis and difficulties in learning to read a regular orthography. Journal of Educational Psychology, 92 668-680. Wolf, M. (1991). Naming speed and reading: The contribution of the cognitive neurosciences. Reading Research Quarterly 26 123-141. Wolf, M. (1999). What time may tell: towards a new conceptualization of developmental dyslexia. Annals of Dyslexia, 49 3-28. Wolf, M., Bally, H., & Morris, R. (1986). Auto maticity, retrieval processes, and reading: A longitudinal study in aver age and impaired readers. Child Development 57, 9881000. Wolf, M., & Bowers, P. G. (1999). The doubledeficit hypothesis for the developmental dyslexia. Journal of Educational Psychology, 91, 415-438. Wolf, M., & Bowers, P. G. (2000). The question of naming-speed deficits in developmental reading disability: An introduction to the Double-Deficit Hypothesis. Journal of Learning Disabilities, 33, 322-324. Wolf, M., Bowers, P. G., & Biddle, K. (2000) Naming-speed processes, timing and reading; A conceptual review. Journal of Learning Disabilities, 33 387-407. Wolf, M., & Katzir-Cohen, T. (2001). R eading fluency and its intervention. Scientific Studies of Reading. (Special Issue on Fluency. Editors: E. Kameenui & D. Simmons), 5 211-238. Wolf, M., & Obregn, M. (1992). Early naming deficits, developmental dyslexia, and a specific deficit hypothesis. Brain and Language 42 219-247. Wolf, M., & Obregon, M. (1997). Naming-speed deficits and the DoubleDeficit Hypothesis: Implications for diagnosis and practice in reading di sabilities. In L. Putnam and S. Stahl (Eds.), Readings on language and literacy (pp. 177-210). Cambridge, MA: Brookline Books. Wood, F., Relton, R., Flowers, L., & Naylor, C. (1991). Neurobehaviroal definition of dyslexia. In D. D. Duane & D. B. Gray (Eds.), T he reading brain: The biological basis of dyslexia Parkton, MD: York Press. Zeffiro, T., & Eden, G.E., (2000). The neur al basis of developmental dyslexia. Annals of Dyslexia, 5 3-30.

PAGE 146

BIOGRAPHICAL SKETCH Hoyeon Kim was born and raised in Sunchon, Jeon-Nam, Korea. She earned her Bachelor of Arts degree with a concentration in special education for people with visual impairments from Dankook University in 1992. She earned a Master of Arts degree from the graduate program in special education at Dankook University in 1994. Her thesis focused on inclusion for children with visual impairments. She earned a certificate of orientation and mobility specialist from the graduate program in special education, school of education, at California State University at Los Angeles in 2000. Her research includes the effects of visual factors on reading as well as visual screening for young children. 136


Permanent Link: http://ufdc.ufl.edu/UFE0006664/00001

Material Information

Title: The Effects of Phonological Awareness, Rapid-Naming, and Visual Skills on Early Elementary Students' Reading Fluency
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0006664:00001

Permanent Link: http://ufdc.ufl.edu/UFE0006664/00001

Material Information

Title: The Effects of Phonological Awareness, Rapid-Naming, and Visual Skills on Early Elementary Students' Reading Fluency
Physical Description: Mixed Material
Copyright Date: 2008

Record Information

Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
System ID: UFE0006664:00001


This item has the following downloads:


Full Text












THE EFFECTS OF PHONOLOGICAL AWARENESS, RAPID-NAMING, AND
VISUAL SKILLS ON EARLY ELEMENTARY STUDENTS' READING FLUENCY















By

HOYEON KIM


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENT FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2004















ACKNOWLEDGEMENTS

I would like to express my sincere gratitude to all those who encouraged me go

forward and helped me complete this dissertation.

First of all, I would like to thank my advisor, Dr. Vivian Correa. She provided a

motivating, enthusiastic, and critical atmosphere and endless support and encouragement

in difficult times. She always kept an eye on the progress of my work and always was

available whenever I needed her advice. Her warmth, enthusiasm, and dedication have

been truly inspirational. It has been my great pleasure to conduct this dissertation under

her guidance.

Furthermore, I owe much gratitude to the other members of my Ph. D committee,

Dr. Martha League, Dr. Holly Lane, and Dr. David Miller who have shown me the way

of research. They also monitored my work and took a tremendous effort in reading my

dissertation and providing me with valuable comments. They could not even realize how

much I have learned from them.

I would like to express my special thanks to Dr. Algina. He let me audit his SEM

course and provided me a possibility to change my methodology into a more challenging

method. He even took a great effort in providing input for data analysis for this study.

I am deeply indebted to the department of special education for providing an

opportunity to work with undergraduate students as a field advisor for their internship.

I've learned so much from this experience as a teacher and researcher. Especially,









working under a guidance of Dr. Martha League was grateful. I specially thank her

wholeheartedly for the many things she has done for me.

My friends have helped me not to get lost during the painful and long journey of

this dissertation emotionally. Without their warm smile and kind words, my journey

wouldn't have been easy.

Needless to say, I am grateful to all of my colleagues, professors, and staff

members at the Department of Special Education. They always made my days in UF

much brighter and easier in so many ways. Thanks a lot!

I feel a deep sense of gratitude for my father and mother who have formed my

vision and taught me the good things that really matter in my life. Without their faith in

me, never-ending love, and great support, I wouldn't have lived such a happy life with

lots of dreams. I also would like to thank my sister, brother, and relatives who have

shown their concerns and love during the critical time in my life.

Last and foremost, my special thanks go to my husband, Jaemin Jung, whose

endless love, encouragement, and concerns enabled me to complete this work. I can't

imagine my life without him. Since 1977, when we both were 8 years old, we've been

together as each other's best friend and husband and wife. One of the best experiences

that we lived through in this period is that we both got Ph. D. degrees from the University

of Florida. My journey seems to be very easy because we travel together. Jaemin, you are

the best and the one who completes me.
















TABLE OF CONTENTS

page

A C K N O W L E D G E M E N T S ................................................................................................ ii

LIST OF TABLES ............... ...................... ..................... ...... .. ........ vii

LIST O F FIG U RE S .................. ....................................... ................ ............. viii

ABSTRACT................................................ ix

CHAPTER

1 INTRODUCTION TO THE PROBLEM ..........................................1

R rationale for the Study ..................................... .. ........... .. ..5
Importance of Early Identification and Intervention...........................................5
Ongoing Debate.................. .. ...... .................................
A Continuing Attention on Phonological Awareness Research............................6
New Trends of Reading-Related Vision Research..............................................8
A Rising Attention on Rapid-Naming Skills................................. .............. 10
Problem s w ith Prior R research ........................................................................ 11
W hy R leading F lu ency ? .................................................................................. 13
Scope of the Study .................. .......................... ..................................... 13
Delimitations .......................... ............... .........14
Lim stations ...................................... ..... ........... .14
Brief Definition of Term s ................................. ......................... .. ............ 14

2 LITERATURE REVIEW .............................................................................16

Theoretical Fram ew ork.............................................. 17
Connectionism ............... ............ ....... .. .............. .................. 17
Seidenberg and McClelland's Connectionist Model................. ..... ..........18
Seidenberg and McClelland' s Model on Orthographic Deficits.........................19
A dam s' Reading Process M odel ................................................... .... 21
Ehri's Theory of the Development of Word Recognition............... ...............23
Prealphabetic phase ................................... ......................24
P artial alph ab etic ph ase .................................................................................. 24
Full alphabetic phase .............................. ..............................25
Consolidated alphabetic phase ............................... ............... 25
Deficit Theories Involving Reading Disabilities .................. ................ ....26










A Core Deficit of Phonological Awareness .................. ............26
Double-Deficit Hypothesis .................. ................. ........28
A Visual Deficit Hypothesis.................................................. ...29
Tem poral Visual Processing Theory ..................................... ......... ......33
Components of Successful Reading ............................... ...............34
Reading Fluency .....................................................34
Definition....................................... .........34
Component ....................................................35
Factors of reading dysfluency .......... .................. .. ................. 36
Reading Comprehension ............................... ...................37
Definition............... .... ................................ 38
Level of reading comprehension ............................. ............... 38
Factors of reading comprehension............. ... ................................ ..................38
Significant Predictors of Reading Success ............. .. ....... ...............39
Linguistic Predictors............... .............. .... ........... .......40
Phonological aw areness.................................................. 40
Rapid naming..................... .............. ......... 43
V ision-Related Predictors............................................................. ..............46
Eye m ovem ent............... ........................................................................ ......... 46
V isual processing ............................. .......................48
Implication for Research................................................................. .. ........ 50
Conclusions............................ ........... .......... 52

3 METHODS .........................................................54

Introduction......................................................... ..........................54
Research Questions ..................... ................................54
Hypotheses ................................................55
M methods ......................................................... ........56
Settings ................................................56
Participants .................................................56
M easures..................................................................................... 58
Data Collection ................................................59
Procedure...................................... .................. .............. ........59
Test Instruments .................................................... ........ 59
Inter-coder Reliability ............................................... ........ 64
Data Analysis................. ...................................... 65
Overview of Statistical Analysis ............................. ........65
Structural Equation Modeling ................................. ........65
Overview ............... ...... ............ ................65
Assumptions of SEM analysis................................ ...... ........67
Analytic Procedures for SEM ......................................................... .........68
A Proposed model ................................................... ........68
The overall model fit ................................................. ........69

4 RESULTS .................................................72



v










D escriptive and Inferential Statistics............. ................................................ 73
D em graphic Characteristics........................ .......... .... ...............73
Means and Standard Deviations on All the Observable Variables...................74
Descriptive Statistics on Percentile Scores.............................. ............ ...76
Gender Differences ............... ................... .. .. .................... 77
Differences Between Second and Third Graders ..........................................78
Differences Between Students From Schools in Mid and Low SES
Neighborhoods ............ ..... .. ..... ..............................79
Correlational Analysis ...................... ........ ....... .. ...... 79
SEM Analysis ...................................... ..... ........... .83
M easurement M odel ................................................................ .. ....... 83
Structural M odel ....................... ... .. ...... .. .............. ..... ... .............87
Regression Analysis and Correlational Analysis with SEM .............................89
M multiple regression analysis............................................... 89
Summary ............... ...... .............................. ......... 93

5 DISCUSSION ...................................... ....... ............ .95

Introduction................. ............................................................... 95
Summary of the Results................................................................ .. ...... 96
Interpretations of the findings.............................................................. .............97
Rapid Naming.................. .. ... ..... ................................................97
The Relationship between Rapid Naming and Eye Movement...........................99
The Relationship between Visual Factors and Reading Fluency .....................99
D discussions on Study Lim stations ....................................................... 101
M easurem ent Issues. .................. ............................ ....... .. .... .. .... ........ 101
Generalizability Issues.........................................102
Implication for Future Research and Practice ...............................................103
Future R research D directions ........................................................ 103
Selecting R leading A ssessm ents ..................................... ............... 103
Effective reading intervention for subtypes of RD.............. .. .............104
A dv anced technology ................................................................................... 106
SEM ................ ............. .......................... ......... 107
Im plications for Practices ............................................................ ...............109
The earlier the better ........................ .... ........................ .. ... ........ .... 109
E arly literacy w ith at risk children .............................................................110
Careful observations of visual problem s....................................................111
Com prehensive vision screening policy ............................................................112
Conclusions.......................... ... .. ........ ...... .......... 113
A Com prehensive Set of Indicators of RD ........................................................114
A B alan ced A approach ................................................................................... 115

APPENDIX PARENT INFORMED CONCENT FORM..............................................119

L IST O F R E F E R E N C E S ............................................................................................ 12 1

BIOGRAPHICAL SKETCH .............................................................................136
















LIST OF TABLES

Table page

3-1 R research Q questions .............................................................54

3-2 Hypotheses .......................................................55

3-3 Descriptive Information for Schools.......................................... ............... 56

3-4 Demographic Data............... .............................57

3-5 Test Variables.......................... ....... .............. 58

4-1 Descriptive Statistics on Demographic Characteristics ........................................ 73

4-2 Means and Standard Deviations on All of the Observable Variables....................74

4-3 Criteria for Percentile Cutoff............................................................... ............75

4-4 Descriptive Statistics on Percentile Scores .................................... ....76

4-5 ANOVA Summaries on Gender Differences....................................................77

4-6 ANOVA Summaries on Group Differences ................. ................. ............78

4-7 ANOVA Summaries on Group Differences............................................ 80

4-8 Correlation Matrix of Phonological Awareness, Rapid-Naming, and Visual Skills
and Reading Fluency ................. ............................................ ....8 1

4-9 Factors for Variables in the Overall Model and Goodness of Fit Results of the
M measurement M odel ................... ........... .. ........ ........... .... ........ 85

4-10 Results of Confirmatory Factor Analysis....................... ... ...............86

4-11 R results of G oodness of Fit .......................................................................... ... ..........88

4-12 Regression Matrix of Four Independent Constructs on Reading Fluency ............91

4-13 Correlation M atrix on Independent Latent Variables ............................................92
















LIST OF FIGURES


Figure page

2-2 Adam s's Reading Process M odel.................................................................. 22

2-3 Adams's Reading Process Model with Breakdowns................... ............... 37

3-1 A Proposed M odel ............... ............................ ......... ........................ 70

4-1 SEM Reading Fluency Model with Correlation Coefficients ...............................90

4-2 Independent Latent Variables with Correlation Coefficients............... ...............92
















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

THE EFFECTS OF PHONOLOGICAL AWARENESS, RAPID-NAMING, AND
VISUAL SKILLS ON EARLY ELEMENTARY STUDENTS' READING FLUENCY


By

Hoyeon Kim

August 2004

Chairperson: Vivian L. Correa
Major Department: Special Education

Reading is at the core of the educational experience and its ability is an essential

key to success in our society. Unfortunately, 20% of U.S. children have serious reading

difficulties. As a consequence, there have been numerous efforts to search for reliable

indicators for early reading problems as well as to develop early intervention programs

for young children at risk for reading difficulties. In spite of these enormous efforts on

early detection and intervention, some students failed to learn to read. In addition, some

of those study findings regarding factors that affect reading have revealed inconsistent

outcomes. In response, this current study was designed to clarify factors that have been

controversial and inconsistent. Since components of reading are quite broad, this focused

only on factors related to reading fluency. Two hundred and ten early elementary students

participated in this study. Twelve measures were administered in order to determine the

relationships of phonological awareness, rapid-naming and visual skills on reading









fluency. Data were analyzed in various methods including ANOVA, correlational

analysis, and SEM.

Results from descriptive statistics indicated that there was no significant gender

difference on all the measures. However, there were significant differences between

second and third graders in several measures including rapid-naming, eye movement,

reversal, and reading fluency rate measures. As predicted, third graders were superior to

second graders on all the measures. The results indicated that mean scores of students

who go to schools in higher SES neighborhoods are higher than those of students who

attend to schools in lower SES neighborhoods on all the measures. Specifically, except

the rapid-naming variables, there were statistically significant mean differences between

these two groups on most of the measures. The results of correlational analysis revealed

that there are significant relationships between reading fluency and four constructs

including phonological awareness, rapid-naming/eye-movement, reversal, and visual

perception skills. Finally, SEM results showed that the suggested reading fluency model

has a good fit, and measures used in this study were significant indicators of their

respective latent variables. No modification of the measurement model was needed since

all of the fit indices met the criteria for a good fit. Furthermore, the multiple regression

analysis indicated that among the four independent constructs, the rapid-naming/eye-

movement construct has the highest correlation with reading fluency. These results imply

that these factors should be importantly considered to indicate young children's reading

problems.














CHAPTER 1
INTRODUCTION TO THE PROBLEM

The ability to read is the gateway to knowledge. Worldwide, today's economy

demands a universally higher level of literacy achievement than at any other time in the

past. Many service-related and information-based jobs have been developed and

increased, and these jobs requires higher level of literacy. Moreover, advanced

vocational or academic training is required now for a wide variety of positions that

previously high school dropouts might be able to take. Therefore, ensuring advanced

literacy achievement for all students is indispensable for an economic necessity. More

recently, using computers and accessing the Internet to get information have become a

part of our life. Today, the ability to read thus has become a basic means to live a

successful life.

As children progress through school, they are expected to be able to read in order to

succeed academically. However, some children experience difficulties with reading and

constantly experience frustration and failure. Approximately, 40% of fourth-grade

students performed below basic levels on the National Assessment of Educational

Progress (NAEP, 2001). This longitudinal national assessment revealed that since 1992

the proportion of students who performed below basic levels of reading chronically

maintained about the 40% rate. In particular, researchers estimated that approximately

15% to 20% of all children have severe reading difficulties (Fletcher & Lyon, 1998; Lyon

& Fletcher, 2001) and need special education services due to their reading-related

learning disabilities. In the field of special education, 50% of all students with disabilities









are identified as having learning disabilities. Further, it is estimated that approximately

85% of students with learning disabilities have severe reading problems (Lyon &

Fletcher, 2001). If these children continuously experience reading difficulties in school,

frustration and failure can affect their motivation and their approach to the task of

learning (Smith, Simmons, & Kameenui, 1998). Eventually, reading difficulties may also

affect all aspects of a person's life. Therefore, many reading researchers and the NAEP

(2001) have warned that reading failure in the schools of the United States has reached

alarming proportions. Early reading failure in schools may also continue to affect

Americans' literacy rates. The National Adult Literacy Survey (National Center for

Educational Statistics, 1999) reported that nearly one-fourth of adult Americans were

functionally illiterate. With this alarming caution, Congress has finally reauthorized the

Elementary and Secondary Education Act (ESEA) after a 3-year period of debate and

consideration. In addition, the No Child Left Behind (NCLB) Act of 2001 was signed

into law by President Bush in January 2002. Since the passage of this law, reading

programs for young children have been more actively implemented. Under this legislative

support, we are hoping to have more positive outcomes in the near future.

Preventing the problem is the most effective in many ways. However, preventing

reading disabilities is dependent on identifying those factors that contribute to children's

reading disabilities. Fortunately reading disabilities (RD) have been studied from various

perspectives over the past two decades, and research has provided cumulative insight into

the factors that contribute to RD. There is a general consensus in the literature that

deficits in phonological processing of reading are correlated with RD. For example, a

large number of research studies have indicated that a child's lack of understanding of the









sound structure of language (phonological awareness) is one of the strong contributors of

RD (Adams, 1990; Torgesen, Wagner, & Rashotte, 1994; Torgesen, Wagner, & Rashotte,

1997). Most reading researchers widely believe that phonological processing abilities are

crucially related to reading and that phonological awareness is the core deficit of children

of RD. However, recent research has built on this consensus to extend beyond it (Stringer

& Stanovich, 2000) with a multi-dimensional view of reading. As a result, other linguistic

factors have been studied beyond phonological awareness. Specifically, a rapid-naming

deficit received significant consideration as one of the key indicators of RD (Wolf &

Bowers, 1999).

Compared to linguistic factors, there is much less agreement on other factors that

influence reading skills, such as vision-related factors. However, studies on factors that

impact reading are continuously presented in the research literature because there is a

strong belief that reading is not only a linguistic process but also dependent upon visual

input (i.e., deciphering written language) (Boden & Brodeur, 1999). In particular, many

connectionist theories that rely on the combined activity of numerous simple, interlinked,

neuron-like processing units, suggest that knowledge is constructed by the strengthening

of connections among processing units for reading (Adams, 1990; McClelland,

Rumelhart, & Hinton, 1986; Seidenberg & McClelland, 1989). For example, Adams'

reading process model (1990) contains a system of four processing units, including the

phonological processor, the orthographic processor, the meaning processor, and the

context processor. The model suggests that reading is successfully processed by the

connection of these four different units. Therefore, it is understandable to examine other

process units rather than focusing only on the phonological process unit. In addition,









there is another theory, the double-deficit theory, which was recently acknowledged in

the field ofRD research, that proposes that developmental reading disability can be

characterized with respect to two specific deficits in phonological awareness and visual

naming speed (Bowers & Wolf, 1993; Wolf, 1991). The second deficit concerns

deficiencies in rapidly accessing and retrieving names for visual symbols (Bowers &

Swanson, 1991; Bowers & Wolf, 1993). Recently, Functional magnetic resonance

imaging (fMRI) added important information about reading processes. For example,

studies on visual processes highlighted that the M pathway's projections are to post

parietal and frontal cortex, involved in the generation of saccades and attention shift

(Perry & Zeki, 2000). The results of this study suggest that there is scientific evidence on

the relationship between visual processing and reading.

Although reading researchers and practitioners have long believed that

phonological awareness is a major facet of the problem, recent neurological findings

(e.g., Perry & Zeki, 2000) on dyslexia have provided additional evidence on a visual

deficit as one of the important factors of RD. At the same time, some researchers

constantly suggest that there may be a certain percentage of children who have RD and

visual deficits affecting reading development (Eden, Stein, Wood, & Wood, 1995;

Rayner, 1998). Due to empirical research findings and scientific evidence, a

multidimensional viewpoint on factors related to RD has more widely been accepted

recently. Therefore, it is appropriate to investigate visual aspects of reading to ascertain

whether they affect RD or at least certain types of RD. In the next section, the rationale

for the current study will be provided in detail with various perspectives.









Rationale for the Study

Importance of Early Identification and Intervention

Unfortunately, children who perform poorly in reading in first and second grades

have a tendency to continue experiencing difficulties throughout their schooling

(Blachman, 2000). A large number of studies revealed that early reading failure seriously

affects future reading ability. For example, Francis, Shaywitz, Stuebing, Shaywitz, and

Fletcher (1996) found that approximately 75% of the children who had difficulties in

reading in the third grade would still be struggling with reading at the end of high school.

As these children grow, they face increasing obstacles to reading development (Torgesen

& Burgess, 1998). As a result, the importance of early identification and intervention on

reading has been recognized among reading researchers. A review study estimates that

the number of children who are typically identified as poor readers and served through

either special education or compensatory education programs could be reduced by up to

70% through early identification and prevention (Lyon et al., 2001). Effective prevention

efforts, however, depend on the effective early identification of students with reading

difficulties. Hence, there is a vital need to identify key indicators of RD for young

children to reduce the impact of reading difficulties and prevent the problems in the

earliest stage.

Ongoing Debate

As a consequence, much research has been conducted in order to identify the major

factors associated with reading disabilities. Overall, there are two parallel perspectives on

this issue. A foremost source of debate among researchers is whether reading difficulties

are the result of one or more than one type of processing problem. To date, many

researchers argue strongly that most reading difficulties are a result of verbal/linguistic









processing deficits (e.g., Mann, 1991; Velluntino, 1979). Some other researchers,

however, have demonstrated evidence that reading involves a combination of multiple

factors related to an individual's experiences, ability, and neurological functioning. Some

of the latter researchers argue that visual abilities contribute to children's reading abilities

(e.g., Eden et al.,1995; Lovegrove, Garcia, & Nicholson,1990; Rayner, 1997; Watson,

1990 ). However, such studies of reading-related visual skills yield inconsistent findings.

Some researchers were able to find that there is a significant difference in visual skills

between children with RD and non-RD (e.g., Morad, Lederman, Avni, Atzmon, Azoulay,

& Segal, 2002) while some were not (e.g., Velluntino, 1987).

In the context of discussions of visual deficits, there is an increasing consensus on

the hypothesis that these processing problems may not be general but may characterize a

small group of severely affected children. It means that not all children with RD have

visual processing deficit but some of children may have visually related reading

problems. If all children with reading disabilities receive the same type of instruction,

which is solely based on phonological awareness, the programs may not be beneficial for

a group of children who have reading disabilities that are associated with other factors

such as a rapid-naming deficit and a visual deficit. Therefore, we should be aware of

subgroups of reading disabled children for effective early identification and intervention.

A Continuing Attention on Phonological Awareness Research

Over the past two decades, no area of reading research has caught as much

attention as phonological awareness. The term "phonological awareness" began

appearing in the research literature in the late 1970s and early 1980s (Clark & Uhry,

1995). Early research on phonological awareness recognized that most poor readers have

difficulty tapping out the number of phonemes in a word as well as in adding or deleting









a phoneme or reserving the order of phonemes in a word (Liberman, 1973; Liberman,

Shankweiler, Fischer, & Carter, 1974; Lundberg, Olofsson, & Wall, 1980). Since the

1980s, there has been increasing agreement among reading researchers that problems in

phonological processing are a major factor in reading difficulties. Additionally, numerous

prediction studies have demonstrated that understanding the sound structure of language

(phonological awareness) is a strong predictor of later reading success. For instance,

Catts and his colleagues observed 604 young children longitudinally and indicated over

70% of poor readers had a history of language deficits including phonological awareness

and oral language in kindergarten (Catts, Fey, Zhang, & Tomblin, 1999).

During the 1990s, research demonstrating the importance of phonological

awareness to the reading process grew rapidly and become widely adapted in the field of

education. In response, a huge number of research studies were actively conducted. On

the basis of such empirical research studies, reading intervention studies, focusing on

instruction of phonological awareness, have continuously been implemented (Clark &

Uhry, 1995). Based on the findings, researchers (e.g., Adams, 1990; Catts et al., 1999;

Torgesen, Wagner, & Rashotte, 1994; Togesen et al., 1997) strongly agree that the

problem associated with phonological awareness is a confirmed factor in reading

difficulties. Recently, functional magnetic resonance imaging (fMRI) revealed the brains

of dyslexics show less activation in the language centers of brain (Broca's area) during

phonological awareness tasks, and researchers indicated that letter identification lights up

the extrastriate cortex in the occipital lobe and that the superior temporal gyrus and parts

of the middle temporal and supramarginal gyri are activated during word meaning

assessment (Shaywitz, 1996). Therefore, assessing skills on phonological awareness









should be continuously considered for improving the predictability of identifying children

at risk for RD as well as developing effective early intervention programs.

New Trends of Reading-Related Vision Research

Compared to phonological awareness research, studies of reading-related visual

skills have a long history with three eras (Rayner, 1998). More than century ago,

Hinshelwood (1895) concluded that poor reading ability was probably due to impaired

visual memory for words and letters (as cited in Eden et al., 1995). Since his study,

research has been implemented vigorously. Recently, Rayner (1998) did an extensive

review of studies focusing on reading-related vision-especially eye movements. In his

review he recognized the trends of prior research and divided research studies from the

late 1890s up to the early 1990s into three eras. According to his review, the first era

extended from the late 1890s up until the 1920s. During this first era, the basic facts of

eye movements were discovered. These included saccadic suppression (i.e., do not

perceive any visual information during an eye movement), saccadic latency, and the size

of the perceptual span (Rayner, 1998). Findings from a variety of studies in the first era

were the foundation of eye-movement research. Contrary to the first era, the second era

(from 1920s until the mid-1970s) produced little research since researchers believed they

had learned almost everything on eye movements, given the technology at that time

(Rayner, 1998). However, the third era (since the late 1970s) is drawing a different

picture. An advanced technology led to an enormous amount of research being

implemented to investigate the possibility of a relationship between visual skills and

reading (Rayner, 1998). The technological improvement on measurements allowed vision

research to be more accurately and more easily implemented.









A variety of experimental research studies revealed differences between disabled

and nondisabled readers on visual skills in the area of reading. However, research on

relationship between visual problems and reading was criticized by reading researchers

due to technical problems of the research designs. Specifically, methodological problems

in the experimental approaches were detected. Along with such criticisms, some

researchers believed that only language factors were involved in reading (e.g., Vellutino,

1987). Therefore, studies in this area were somewhat neglected among reading

researchers. Furthermore, since 1980, increased attention on studies related to

phonological awareness may have contributed to underestimating research in other

aspects of reading. Despite heavy criticism and underestimation, many reading

researchers revealed that visual skills play an important role in reading (e.g., Eden et al.,

1993, 1995; Watson et al., 2003). Recently, methodological problems were reduced

through using more advanced measurements with sufficient evidence of validity and

reliability and through carefully selecting appropriate research designs. Moreover, the

multidimensional viewpoint on reading is beginning to be more widely accepted among

reading researchers. Most recently, Habib, a neuroscientist, offered scientific evidence on

the roles of a visual deficit on RD (2000). Functional magnetic resonance imaging

(fMRI) has provided new information concerning the neuroanatomical location of the

system affected in RD (Eden & Zeffiro, 1998). By utilizing fMIRI, researchers have found

that both dyslexic and normal readers can see stationary dots on a screen but finding

moving dots was more difficult to detect in dyslextic readers compared to normal

readers. Furthermore, dyslexic participants show no activation of the MT (V5) visual

cortex, which detects motion, during reading and seeing these moving dots while normal









readers do have activation of the MT (V5) cortex (Eden & Jeffiro, 1996; Zeffiro & Eden,

2000). An impairment of the magnocellular (M) pathway might even affect the

generalized timing mechanisms of the cerebellum (Fawcett, Nicolson, & Maclagan

2001). The visual cortex is divided into 6 different areas; VI, V2, V3, V4, MT/V5. In the

visual cortex, there are two distinct pathways call Parvocellular Pathways (P Pathways)

and Magnocellular Pathways (M Pathways) that carry information about what and where

objects are. In particular, M pathway runs from area VI to V2 to V3 and MT/V5 and

represents information about dynamic properties such as motion and spatial relations

among objects.

With such new trends, reading-related vision research has been more actively

undertaken. However, since the findings are still contradictory and inconclusive, there is

a critical need to clarify whether visual factors are correlated to reading abilities of a

certain group of children.

A Rising Attention on Rapid-Naming Skills

There is a growing consensus that rapid-naming skills are significantly related to

reading ability (e.g., Swanson & Wolf, 1999; Wolf, 1997). Originally, research from the

field of cognitive psychology provided evidence that individuals with RD have a deficit

in naming speed (Wolf, 1997; Wolf& Obreg6n, 1997). Basically, a naming-speed deficit

refers to the inability to retrieve rapidly the spoken referent for a visual stimuli (Badian,

1997). Traditionally, rapid-naming skills have been considered as a part of phonological

skills (Wagner & Barker, 1994; Wagner & Torgeson, 1987). However, Wolf and her

colleagues provided evidence that a naming-speed deficit is another core deficit, along

with a phonological processing deficit, by showing a modest correlation between

phonological awareness and rapid-naming variables. Researchers who argue for a double-









deficit theory also have concerns about two certain types of group (e.g., Wolf, 1991;

Wolf & Bowers, 1999). Actually their research findings (Wolf & Bowers, 1999) suggest

that there are three types of RD based on processing deficits: a group of children with

mostly a phonological processing deficit, a group who have mostly a naming deficit, and

a group who have both phonological and rapid-naming deficits. Obviously, the last two

groups of children who have a rapid-naming deficit would not be well served if early

intervention programs that are mainly based on phonological awareness skills were

provided. In sum, although a growing body of research supports the importance of rapid-

naming skills as a separate indicator from phonological awareness of RD, it is still

inconclusive. Hence, there is a need to replicate such studies to clarify this issue.

Problems with Prior Research

While phonological awareness has captured researchers' attention, reading-related

vision studies have received less attention. As a result, there has been a lack of research,

particularly in the field of education. Specifically, only a few correlation studies

involving both visual skills and verbal skills in the regression models have been

undertaken (e.g., Eden et al., 1993, 1995). Moreover, a lack of research has revealed

inconsistent results. To illustrate, while McKinney, Short, and Feagans (1985) reported

that only 7% of a sample with learning disabilities showed severe visual problems,

Watson (1990) reported that 52% of a sample of students with reading disabilities

exhibited some type of visual processing problems. In addition to inconsistency, the

findings were conflicting and yielded mixed outcomes. For example, some researchers

were able to indicate a significant difference in the visual abilities between poor readers

and nondisabled readers (e.g., Watson, 1990). In contrast, others found no such

differences between the two groups (e.g., Vellutino, 1987).









Disagreement is also recognized in studies on a rapid-naming deficit. As mentioned

before, a large body of evidence has accumulated in favor of the importance of a rapid-

naming factor in reading development and a key indicator of reading disability (e.g.,

Wimmer, Mayringer & Landerl, 2000; Wolf & Bowers, 1999). However, the evidence is

somewhat inconsistent, and there are gaps in the evidence. Moreover, some issues still

remain. Although many studies have found phonological awareness and naming speed to

have significant independent unique effects on current or later reading (e.g., Badian,

1997, 1998; Kirby & Parrila, 1999; Manis, Doi, & Bhadha, 2000; Scarborough, 1998;

Torgesen, Wagner, Rashotte, Burgess & Hecht, 1997), the outcomes varied based on

when assessment occurred, how many indicators were used in their models, and what

other variables were used in their equations. Moreover, some other studies indicated that

one or the other individual effect disappears when prior achievement, which is an

autoregressor, is included in the regression model (e.g., Torgesen et al., 1997). This issue

is discussed in further detail later in Chapter 2.

Another problem is that the intervention studies are mostly limited to investigating

the effectiveness of their certain programs (e.g., Torgeson et al, 1997). Almost all reading

programs appeared to be very effective and helpful in improving children's reading

acquisition levels. However, the intervention researchers failed to take a close look at a

group of students who have shown no gains despite an intensive reading support with

specially designed approaches based on linguistic components. Among children who do

not show any improvement, some children may suffer from other problems such as visual

deficits that affect their reading. Hence, there is a need to clarify possible factors that are

associated with people with no gains after reading interventions.









Why Reading Fluency?

Reading fluency encompasses the speed or rate of reading as well as the ability to

read materials with expression. Meyer and Felton (1999) defined fluency as "the ability

to read connected text rapidly, smoothly, effortlessly, and automatically with little

conscious attention to the mechanics of reading, such as decoding" (p. 284). Children are

successful with decoding when the process used to identify words is fast and nearly

effortless or automatic. The concept of automaticity refers to a student's ability to

recognize words rapidly with little attention required to the word's appearance. The

ability to read words by sight automatically is the key to skilled reading (Ehri, 1998).

Researchers have utilized reading rate for their studies since a student's reading rate is

simply calculated by dividing the number of words read correctly by the total amount of

reading time. Furthermore, Fuchs, Fuchs, and Maxwell (1988) found that students'

fluency was more highly correlated with their scores on a standardized reading

comprehension test than were oral and written retellings, question answering, or cloze

tests. Shinn, Good, Knutson, Tilly, and Collins (1992) provided additional support for the

use of fluency measures to monitor progress in reading. Therefore, reading fluency rate

was selected as a dependent variable measure for this study.

In response, this research was designed to clarify the relationship among reading

fluency and phonological awareness, naming speed, and visual factors simultaneously

with structural equation modeling.

Scope of the Study

This study was conducted within a limited scope. Therefore, there were

delimitations and limitations in conducting the present study. They are described in this

section.









Delimitations

This study was geographically limited to one school district, Alachua County, in

north-central Florida. The 210 subjects were selected from five schools in this medium-

sized district. The study was conducted with early elementary students, especially with

second- and third-grade students since people's visual skills are fully developed by ages 6

to 7. Therefore, kindergarten and first-grade students were excluded from this study.

Subject selection did not include consideration for gender or ethnicity.

Limitations

This study was conducted with second- and third- grade students. Therefore, the

findings of this study are generalized to older or younger students. Due to time

limitations, some assessment subtests were excluded for this study. For example, a

subtest of CTOPP (Comprehensive Test of Phonological Processing), segmentation, for

assessing phonological awareness was not included as an observed variable.

Brief Definition of Terms

An understanding of terminology that was applied in this study is necessary to the

interpretation of this examination. The following section defines relevant terms as they

apply to this study.

Phonological awareness refers to a conscious sensitivity to the sound structures of

language. It includes the ability to detect, isolate, blend, segment, and otherwise

manipulate the sounds in language (Adams, 1990; Snow, Burns, & Griffin, 1998;

Torgesen & Bryant, 1994).

Phoneme refers to the smallest unit of spoken language. For example, the word

'chin' contains three phonemes (/ch/ /i/ /n/).









Rapid naming is the ability to recall quickly and to verbalize the name of a

presented object. Naming speed is typically assessed by the Rapid Automatic Naming

(RAN) Test (Denckla & Rudel 1974, 1976), which requires children to name familiar

colors, pictured objects, digits, and letters (Wagner & Togeson, 1987, as cited in Manis,

Seidenberg, & Doi, 1999) and subtests of CTOPP Comprehensive Test of Phonological

Processing (CTOPP) (Wagner, Torgesen, & Rashotte, 1999).

Orthographic processing refers to the use of printed symbols (i.e., orthographic

information) to process written language (Wagner & Marker, 1994).

Eye movement is one of the visual efficiency skills that include accommodation,

binocular vision, and ocular mobility. Eye movement in reading-related research is

involved in abilities of tracking and fixation.

Visual processing skills refer to a group of visual cognitive skills used for

extracting and organizing visual information from the environment and integrating this

information with other sensory modalities and higher cognitive functions (Scheiman,

2002).














CHAPTER 2
LITERATURE REVIEW


In order to have a better understanding of reading development based on roles of

cognitive processing, this section reviews three areas of the literature regarding

theoretical background based on connectionist theories: key deficit theories underlying

children's reading disabilities, the components of successful reading, and significant

predictors of reading success in the previous and current literature. As a theoretical

framework for this study, connectionism and connectionist reading models were

employed and are discussed in Chapter 2. The latter section of Chapter 2 includes four

general deficit theories including core deficit of phonological processing, a double-deficit

hypothesis, a visual deficit theory, and a visual temporal processing theory, identifying

causes of RD in the literature to date. For early identification of reading disabilities, two

significant linguistic indicators-phonological awareness and rapid naming-and two

visual related indicators-eye movement and visual processing-are discussed in the

final section of Chapter 2 based on prior literature that includes findings of empirical

studies. Currently, researchers in diverse fields as neuroscience, cognitive psychology,

educational psychology, special education, and reading have influenced this topic. The

present study reviewed journal articles and books from many of these disciplines.

Therefore, within this literature review, the terms "reading disability" (RD) and

"dyslexia" were used interchangeably to describe a person who has "a specific and

significant impairment in reading abilities, unexplainable by any kind of deficit in general









intelligence, learning opportunity, general motivation or sensory acuity" (Habib, 2000, p.

2374). The current research provides a reliable set of predictors of reading achievements

focusing on both verbal and visual factors.

Theoretical Framework

The theoretical framework of understanding children's reading process in this

literature review section is primarily grounded in connectionist theories that were

originally based on interaction models of the information processing theory. In this

section, general characteristics of connectionism, Seidenberg and McClelland's

connectionist model for visual word recognition, and Adams' reading acquisition model

are briefly discussed in order to explain the connections between reading behaviors and

the brain's cognition and perception. In addition, Ehri's theory of reading development is

also explained to understand the procedures of reading developments of young children.

Connectionism

Simply stated, connectionism is a theory of information processing. Connectionists

originally adopted the view that the basic building block of the brain is the neuron.

Connectionism offers a new approach to explaining human cognitive activities. While

classical cognitive theories use explicit, often logical, and rule-based theories in a serial

manner, connectionist systems rely on the combined activity of numerous simple,

interlinked, neuron-like processing units. In order to begin the reading process,

Seidenberg and McClelland (1989) developed a connectionist model based on an

interactive theory in an information-processing theory. Commonly, all connectionist

models, including a Seidenberg and McClelland word recognition model, have several

similar properties. They all have a set of interconnected simple processing units. The

types of units include input units, output units, and hidden units. Input units accept the









information either from the sensory channels or from other parts of the network. Output

units put out the information. They either control the information directly or send those to

other parts in the network. Hidden units communicate with the input and output units

internally. Most of the important work in a connectionist model occurs in the units and

their connections. Furthermore, there is no central control within the connections, and the

processing is spread across the entire network. In fact, all units have their own activation

value. However, each unit itself is meaningless and not active since the cognitive

processing produces output when the connections between units occur. In particular, to

understand the act of reading as connectionism, we should carefully consider Seidenberg

and McClelland's parallel distributed processing model.

Seidenberg and McClelland's Connectionist Model

Seidenberg and McClelland (1989) developed a connectionist word recognition

model by describing a parallel distributed processing (PDP) system consisting of four

different processors: context, meaning, orthographic, and phonological. In their PDP

model, they developed a general connectionist framework for thinking about how lexical

knowledge is obtained, represented in the brain, and used in processing (Manis et al.,

1999). Simply stated, their reasoning is that reading involves a series of associations or

connections resulting in accumulated lexical knowledge (Clark & Uhry, 1995). Actually,

the concept of PDP is based on the belief that "intelligence emerges from the interaction

of a large number of simple processing units, each sending excitatory and inhibitory

signals to other units" (McClelland et al., 1986, p. 10). In addition, their PDP model

involves connections in both directions between context and meaning and then additional

two-way connections between meaning, orthography, and phonology (see Figure 2-1).









That is, connections between any two processors can trigger other associations in any

direction and contribute to the overall reading process.

In order to increase the processing capabilities inherent in networks, they include a

set of hidden units in their model. Hidden units are considered important since they

mediate between the pools of representational units (Seidenberg & McClelland, 1989).

More interestingly, Seidenberg and McClelland used computational simulations of word

learning to provide evidence of the interactive nature of the relationship between

phonology and orthography. By reducing the number of hidden units, Seidenberg and

McClelland were able to simulate poorer performance in fluency, naming, and visual

processing, which are characteristics found in people with dyslexia. The following

section of their model focuses on the role visual processing plays within their model. It

provides the reason for poor reading when visual processing is not working properly in

this model.

Seidenberg and McClelland's Model on Orthographic Deficits

Seidenberg (1992) schematically illustrated the main component of the existing

model. One component considers the orthographic representations. "The model starts by

encoding a letter string as a pattern of activation across the orthographic units"

(Seidenberg, 1992, p. 262). Letter recognition is a computational process. However, the

process may fail to activate appropriated orthographic representation when this

component fails to operate. This could occur due to abnormalities in visual processes or

damage to the orthographic representations themselves (Seidenberg, 1992).













Context
Processor





Meaning
Processor








Hidden units Hidden units








Orthographic Phonological
Processor Processor


Input
MAKE


Output
/mAk/


Figure 2-1. Seidenberg and McClelland's Word Recognition Model

The model performed as in the simulations with 100 hidden units. Actually, the

"normal" model was configured with 200 hidden units, but for this simulation they

reduced a number of hidden units for the simulation. They reported that "with fewer

computational resources, the model was still able to learn; however, it performed poorly

on irregular words and non-words even after extensive training" (Seidenberg, 1992, p.









263). For another simulation, they damaged the input orthographic representation and

found that the outcomes of two damage simulations produced similar results. Through

these findings, they argued that two very different types of impairments yielded a

"dyslexic" pattern of performance; only one of these corresponds to a visual processing

deficit (Seidenberg, 1992, p. 265). Hence, Seidenberg (1992) noted that the simulations

confirm that there must be more than one source causing RD. He also described that "this

might go some way toward explaining why only some children whose behavior fits the

dyslexic profile exhibit visual processing deficits" (Seidenberg, 1992, p. 265).

Adams' Reading Process Model

With a thorough review of reading research, Adams (1990) framed her view of the

reading process based on connectionist theory. Since 1990, Adams's connectionist model

has been used as a foundation for understanding reading processes. Basically, Adams'

reading model includes four interrelated processors: orthographic, phonological,

meaning, and context (see Figure 2-2). She accentuated the connectivity among four

processors that are crucial for a skilled reader. Although each processor functions

independently in reading, the four units should work simultaneously for successful

reading.













Context
Processor





Meaning
Processor




Orthographic Phonological
Processor Processor




Print Speech


Figure 2-2. Adams's Reading Process Model


As seen in Figure 2-1, each processor is systematically connected and works

cooperatively to obtain meaning from reading. According to Adams (1990), the system

begins its work when the orthographic processor obtains visual information from printed

materials and delivers visual input to the meaning processor. If the printed word was

recognized and connected to a reader's lexical memory, the meaning processor

automatically interprets the word. However, if the word is unfamiliar to a reader, the

phonological processor begins to compensate for the orthographic processor. At that

time, a reader simply tries to sound out the word and determines whether the word is

unfamiliar or familiar (i.e., in the reader's listening vocabulary). Based on the

cooperation of these two processors, the activities of these two processors can connect to

the meaning processor. In addition, the context processor helps a reader to understand









context and predict future developments. Thus, the context processor reinforces the

connectivity among these three processors for a reader. As a result, its work increases the

processing speed of the other three units. Furthermore, the context processor helps a

reader to verify and choose the real meaning of a word based on its context, which also

may improve comprehension. For a skilled reader, each of the four reading processors

should play a unique role but also function simultaneously, collaboratively, and

automatically with little conscious effort (Adams, 1990).

Ehri's Theory of the Development of Word Recognition

Understanding how normal reading develops is crucial to the understanding of

what goes wrong during the development of reading in children with RD (Clark & Uhry,

1995). Many researchers in the reading community have proposed theories to explain

how readers develop reading abilities. In particular, Ehri's theory proposes the four

developmental stages of word recognition by sight word. Adams (1990) mentioned that

the ability to read words by sight is dependent upon the strength of the connections that

have been developed between the orthographic and phonological processors. Ehri's

theory highlights such connections between the orthographic and phonological processors

(Adams, 1990). According to Ehri and Saltmarsh (1995), "readers read sight words by

accessing connections that they have formed between letter in the spellings specific

words and phonemes detected in pronunciation" (p. 298). Eventually, her viewpoint was

developed under the connectionist view (Ehri, 1992). She described that a reader's ability

to read progresses through a series of phases, and "each phase is defined by the kind of

connections that are formed between visual cues seen in print and information about a

specific word stored in memory" (Ehri, 1992, p. 123). The four phases are described

below.









Prealphabetic phase

During this first phase, children begin to recognize words in their environment

(Ehri, 1991, 1992, 1995). Since children under this phase have little knowledge about

letters, they create connections between salient features of a word and the meaning of the

word (Ehri, 1991, 1992, 1995). The visual cues selected by children in this stage might be

part of the spelling, shapes, or colors. Simply, a large yellow arch shape logo of

McDonald can be a typical example of a visual cue (Ehri, 1992). Therefore, Ehri (1991)

indicated that the identification of the word occurs at the semantic rather than the

phonological level, and the connections are visual and nonphonetic in this first stage of

word recognition. She also described that because most connections formed between

spellings and meaning are arbitrary, they are easily forgotten unless practiced frequently"

(Ehri, 1992, p. 125). However, to become a skilled reader, children should be able to

form the connections systematically between the printed word and the meaning. In the

second phase of Ehri's theory the reader finally begins to use letter-sound relations.

Partial alphabetic phase

During this second phase, "learners begin using their letter knowledge to form

visual-phonetic connections between letters seen in spellings and sounds detected in

pronunciation of words they are learning to read" (Ehri, 1992, p. 126). Compared to the

first phase where connections are arbitrary, those in the second phase are systematic.

Moreover, if only one connection occurs in the first stage, more connections are formed

in the second phases since a reader uses letters. Another main difference between two

phases is that the primary connection in Phase 1 is formed between a spelling and its

meaning, whereas the main connection in Phase 2 is between a spelling and its

pronunciation (Ehri, 1992). They use more reliable and systematic information, but they









are not yet able to use the full cipher in this stage. Therefore, this stage is referred to as

the partial alphabetic phase (Ehri, 1991). Novice readers in this stage use incomplete

letter-sound correspondences to access words in memory. They may use the first or last

letter sounds, ignoring the medial phonemes in the work (Ehri, 1992) such as recognizing

the word 'book' by the Ibl and the Xkl. However, the readers should be able to utilize the

full cipher for their skillful reading.

Full alphabetic phase

Ehri (1992) described that "cipher sight word reading refers to the process of

reading sight word by setting up connections in memory between the entire sequence of

letters in spelling and phonemic constituents in the word's pronunciation" (p. 132). In

contrast to readers in Phase 2 who may use some letters of spellings, readers in the third

phase can form connections out of the "complete array of letters connected to the

phonemic structure of the word" (Ehri, 1992, p. 134). For example, in a word with four

phonemes, 'desk,' a reader at the cipher reading phase would be able to create

connections out of all four letters and correspond phonemes to determine the correct

pronunciation (Ehri, 1999). Children who gain the ability to utilize the cipher can become

proficient readers (Gough, Juel, & Griffith, 1992).

Consolidated alphabetic phase

Finally, mature readers can analyze the common spelling patterns such as '-ing,' '-

cal,' and 'com-.' It means that they are able to recognize these patterns as a chunk. If

children encounter such combinations of letters more often, they may have stronger

concepts of such combinations and the speed of orthographic processing will be faster.

Therefore, connections to new words become more efficient and occur more rapidly. As a

result, readers in this final stage can improve their ability of word recognition









enormously. Consequently, Ehri's theory of word recognition implies why some children

are not very skillful at reading words (Ehri, 1992). They may not have opportunities to

put through the sequential processes. More specifically, the following section discusses

predominant deficit theories that explain a causal role ofRD.

Deficit Theories Involving Reading Disabilities

Different theories are often used to explain children's reading disabilities. These

theories serve as highlights of the different processing systems involved in reading.

Through the huge amount of prior reading research, everyone probably agrees with the

fact that all children with reading disabilities are not the same. Therefore, it is essential to

know the basis of the RD and to match the program of instruction to the inefficient

information processing system. Here are four general deficit theories identifying causes

of RD in the literature to date. Recently, studies on process factors involving reading

have been extensively conducted underlying the progress of neuroscience

A Core Deficit of Phonological Awareness

A consensus among reading researchers is that phonological processing is the

core deficit in RD. During the last 25 years, researchers and educators have been looking

for valid and reliable predictors of reading difficulties in order to prevent problems and

provide effective early intervention. Moreover, based on findings from a large amount of

reading research, many intervention programs for children with reading difficulties have

been developed over the years. One intervention that has attracted much recent attention

involves teaching phonological awareness skills to children. One important reason why

educators are so interested in phonological awareness is that research indicates that it is

the best predictor of early reading acquisition, better even than IQ, vocabulary, or

listening comprehension (Stanovich, 1988). In general, phonological awareness often









refers to the conscious awareness of or sensitivity to the sounds of speech apart from their

meaning (Snow et al., 1998; Torgeson, 2000).

Many researchers have specifically indicated that symptoms such as slow rate of

reading, spelling errors, and errors on function words in reading and writing could be

accounted for poor phonological skills (e.g., Aaron & Joshi, 1992). In addition, many

reading researchers indicated that phonological awareness interventions have been

successful. For example, Torgeson et al. (1997) reported that two types of research

support the phonological deficit perspectives as the core-causing factor of reading

difficulties. First, predictive research studies have determined that a measure of

phonological awareness is a major predictor of later success in reading. Torgeson et al.

(1997) used a set of predictors for their logistic regression model and indicated that letter-

name knowledge and phonological awareness predicted, with 91% accuracy, which

children would be in the bottom 10% when they become second graders. Second,

intervention studies have shown that instruction in phonological awareness improve

reading acquisition. Researchers also have highlighted that a group of children who were

assigned to have instructions in phonological awareness have a significantly low

retention rate in special classes (9%) compared to two other groups with different

instructions and a group with non-treatment (25%, 30%, and 41%, respectively)

(Torgeson et al., 1997).

Through vigorous research during the past two decades, it has become obvious that

a lack of ability in phonological awareness is the most widespread barrier to reading

acquisition (Adams, 1990; Goswami, 1993; Torgeson et al., 1997). Until recently,

however, researchers were unable to map cognitive functioning in the brain during









cognitive processes. Now, researchers are able to identify regions of the brain that are

activated during cognitive processes such as reading through functional magnetic

resonance imaging (fMRI). Anatomical work and functional brain imaging studies

support the notion of a left perisylvian dysfunction as a basis for the phonological deficit

(e.g., Pugh et al., 2000; Temple et al., 2003; Shaywitz et al. 2000). The phonological

theory, however, does not explain the other deficits of dyslexia. For example, there is a

limitation of explaining dyslexia with the phonological theory because the visual system

on reading process cannot be accounted for by phonological theory.

Double-Deficit Hypothesis

The double-deficit hypothesis of reading disabilities is currently popular among

those who investigate the neurological components of reading disabilities. The double-

deficit hypothesis was developed by Wolf and Bowers (2000) based on their research

findings showing that there are only modest correlations between phonological awareness

measures and Rapid Automatized Naming (RAN) performance in groups of dyslexic

children. They also found that dyslexic children as a group exhibit both phonological

awareness deficits and naming-speed deficits. Originally, naming speed deficits were

considered by some to be part of a phonological deficit (Wagner & Torgesen, 1987).

However, through such findings, Wolf and her colleagues argued that a naming speed

deficit is a separate deficit from a phonological deficit and affects reading acquisition

independently from phonological processing (Olson, Forsberg, & Wise, 1994). Further,

they suggested that there are different degrees of reading disability and different types of

processing impairment.

More specifically, there are three subtypes of dyslexic children: one having a

dominant phonological deficit but little naming-speed deficit, one exhibiting the naming-









speed deficit but little phonological awareness deficit, and one with both deficits (Wolf &

Bowers, 1999). Obviously, the second and the third groups of children are important in

this theory because the two subgroups with only a naming-speed deficit and both deficits

may not be well supported if reading intervention, which is only based on phonological

awareness approaches, is provided (Wolf & Bowers, 1999). Wolf and Bowers (1999)

mentioned that researchers have found that about a fifth of dyslexic children mainly

suffered from naming-speed deficit and not from a phonological awareness deficit. More

interestingly, Lovett, Steinbach, and Frijters (2000) found that more than half of the

participants suffered from both deficits. Furthermore, extensive data replicated the

existence of these three subtypes of reading disabilities in several language systems (e.g.,

German, Dutch, and Finnish), and they provided various evidence on the existence on the

subtypes based on double deficit hypothesis (e.g., Deeney, Gidney, Wolf, & Morris,

1999). It is interesting that speed of processing and semantic fluency are included in

some of the recently produced tests for dyslexic children.

Beyond a double deficit hypothesis, some researchers even argue for a triple

deficit hypothesis. For example, Badian (1997) purported evidence for a triple deficit

hypotheses implying that orthographic factors involving visual skills should also be

considered. The author found that there was a significant difference between the RD

group and the non-RD group on an orthography test, the Jordan Left-Right Reversal test,

which is a visual perception test.

A Visual Deficit Hypothesis

Another hypothesis is that problems in reading acquisition are linked to deficient

visual perception. More than a century ago, James Hinshelwood, an ophthalmologist

practicing in Scotland, called his patients with RD "congenital word blindness" and









believed that these problems were related to the left hemisphere of the brain, which

controls the storage of visual memory because these children seemed to have difficulty

remembering names of letters and words (Clark & Uhry, 1995). In 1928, Samuel Orton,

an American neurologist, theorized that people perceive visual inputs from both right and

left visual fields and relay this input as mirror images to the visual cortices (Clark &

Uhry, 1995). For most people, one side of brain becomes dominant over the other and

suppresses the image from the non-dominant hemisphere; however, he believed some

people who poorly perform 'dominance', may perceive backward images. He believed

this poor dominance might produce reading problems (Clark & Uhry, 1995). Therefore,

from the early 20th century, vision research related to RD has been continuously

undertaken. However, researchers revealed that reversal and sequencing errors did not

appear in a large proportion of people with RD, and most of people with RD showed

more oral reading errors. Moreover, Vellutino (1978, 1983) indicated that poorer readers

are as good as nondisabled readers on copying confusable letters and words from

memory. Rather, they showed problems at naming and pronouncing the items. With these

findings, he concluded that visual deficits could not explain RD. Moreover, several major

reviews of the evidence concluded that visual deficits are not a major cause of reading

disability (Stanovich, 1982; Velluntino, 1979). Since then, numerous studies have

demonstrated that understanding the sound structure of language (phonological

awareness) is a strong predictor of reading success.

In contrast, studies on visual processing factors have been under heavy attack by

reading researchers during the last three decades. As discussed earlier, inconsistent and

conflicting findings have fueled the debate and have hindered research on visual









processing within reading research. In addition, researchers pointed out that many of the

processing measures used in education lack reliability and validity (Moats & Lyon,

1993). Some even strongly stated that the case is completely closed (Stanovich, 1992).

Nevertheless, many researchers recently tried to revisit this area of reading and provide

reliable evidence on differences between disabled and nondisabled children in visual

recognition experiments.

More recently, one visual deficit theory, "saccadic-suppression deficit," emerged

surrounding eye movement and visual pathways. When people's eyes view visual stimuli

such as letters, words, and any stationary things, they fixate briefly (for roughly 200 to

300 ms). After viewing, it takes about 10 to 40 ms to construct an image in the brain

before the next fixation. It means visual information from the saccade is suppressed. If

this saccadic suppression, which insures a series of discrete images, is not achieved, the

images may be overlapped (Pollatsek, 1993). Since erratic eye movement has been

observed in children with RD, researchers have continuously indicated that poor readers

do have characteristic difficulties on visual ability measures (e.g., Eden et al., 1995). The

measurements used for recent studies have been more carefully selected, and most of

them are nationally standardized instruments (e.g., Developmental Eye-Movement test or

Visual- Motor Integration test). Since many reading researchers have begun to have a

multi-dimensional view of the reading process, the case against visual processing factors

seems to have become weaker, and in recent years more studies in this area have been

introduced in educational journals. Even though many reading researchers agree that a

phonological factor is a major predictor of reading difficulties, some of them also believe

visual processing may play an important role (Eden et al., 1993, 1995; Rayner 1998;









Temple et al., 2000). In particular, many studies in the medical field still focus on

determining the relationship between reading and visual skills. To illustrate, one study

found that children with reading difficulties had a specific disability in the shifting of

attention and were only able to maintain attention for a short period of time (Facoetti,

Paganoni, Turatto, Marzola, & Mascetti, 2000). As a result, researchers assumed that

deficits of visual attention may hinder efficient visual processing during reading (Facoetti

et. al., 2000). Casco, Tressoldi, and Dellantonio (1998) also found significant differences

between disabled and nondisabled readers on visual search tasks. These findings

suggested that children with the lowest performance on tasks that included searching for

a target letter in a set of background letters showed both a significantly slower reading

rate and a higher number of visual search errors in comparison to children with higher

performances on such visual search tasks. Other studies have compared visual and verbal

abilities as indicators for prediction of reading disabilities. One study compared the

phonological and visuo-spatial abilities of a selected group of 17 children with reading

disabilities and 21 without reading disabilities (Eden et. al., 1993). Interestingly, this

study recognized that there are some visual tasks such as eye movement and visual search

that are equally reliable at distinguishing between good and poor readers as phonological

tests (Eden et al., 1993). In sum, in recent years vision research has begun to be more

widely accepted by reading researchers. Due to advanced instruments and a multi-

dimensional viewpoint on the reading process, vision research has been vigorously

implemented and profoundly approached. Further, such research findings suggest there

may be a certain group of children who have RD and visual problems affecting reading

development.









Temporal Visual Processing Theory

The ability to process and retain pieces of information presented in a rapid

sequence is a vital part of learning to read. Specifically, this is related to a role of visual

processing in the reading process, the speed of process. Simply, there are two systems

that transmit information from the eye to brain. The sustained or 'parvocellular' system

operates during the fixation, and the transient or 'magnocellular' system works during

saccades or movements to the next fixation (Clark & Uhry, 1995) In order to explore this

theory, a number of researchers conducted experiments and demonstrated that the

transient system operates slowly in some poor readers in comparison with nondisabled

readers (Boden & Brodeur, 1999; Livingstone, 1993; Lovegrove, 1992). For example,

Boden and Brodeur (1999) hypothesized that adolescents with RD are slower at

processing visual information than normal readers and thus experience greater visual

persistence. Through the findings from two types of computer-based activities, the

authors concluded that the adolescents with RD demonstrated deficits in visual temporal

processing, which is the processing of rapidly presented visual information, due to visual

persistence. Previously, Di Lollo Hanson, and McIntyre (1983) stated that children with

RD show slower visual information processing and longer duration of visual persistence

compared to the control group. Similarly, Richards, Samuels, Turnure, and Ysseldyke

(1990) found that children with RD require longer response times on variations of visual

tasks; thus, the authors concluded that children with RD demonstrate slower information

processing. Based on the findings, we can assume there may be a relationship between

rapid visual naming and visual processing skills. Hence, further investigation should be

implemented to clarify the relationship.









Components of Successful Reading

What is reading? Reading is a comprehensive and active process in which readers

shift between sources of information (what they know and what the text says), elaborate

meaning and strategies, check their interpretation (revising when appropriate), and use

the social context to focus their response (Walker, 1996, p. 4). Other authors defined

reading as the act of simultaneously reading the lines, reading between the lines, and

reading beyond the lines (Manzo & Manzo, 1993, p. 5). The first part of their definition,

reading the lines, refers to the act of decoding the words in order to construct the author's

basic message. The next part, reading between the lines, refers to the act of making

inferences and understanding the author's implied message. Finally, reading beyond the

lines involves the judging of the significance of the author's message and applying it to

other areas of background and knowledge. In other words, reading is more than

translating print into the spoken word (decoding). Reading is getting meaning from print.

People who have not developed automatic word recognition skills may have

comprehension problems because their energy is focused on identifying words rather than

thinking about what they mean. Many of these children may read very slowly and often

have to read things more than once to understand them. Others have automatic word

recognition skills but cannot comprehend what they read. In that sense, the reading

process can be divided into two areas: reading fluency and comprehension.

Reading Fluency

Definition

Unfortunately, there are still no consensual definitions of reading fluency (Wolf &

Katzir-Cohen, 1999). However, it is closely related to the subset of time-related terms

and frequently related with other terms such as automaticity, speed of processing, reading









rate/speed, and word recognition rate/proficiency. During the last two decades, research

on reading fluency was actively undertaken, and many reading researchers suggested

definitions of reading fluency. Through a thorough review of previous fluency literature,

Wolf and Katzir-Cohen (1999) proposed their own working definition:

In its beginning, reading fluency is the product of the initial development of
accuracy and the subsequent development of automaticity in underlying sublecxical
processes, lexical processes, and their integration in single-word reading and
connected text. These include perceptual, phonological, orthographic, and
morphological processes at the letter, letter-pattern, and word levels, as well as
semantic and syntactic processes at the word level and connected-text level. After it
is fully developed, reading fluency refers to a level of accuracy and rate where
decoding is relatively effortless; where oral reading is smooth and accurate with
correct prosody; and where attention can be allocated to comprehension. (Wolf &
Katzir-Cohen, 2001, p. 219)

Wolf and Katzir-Cohen (2001) looked at fluency not so much as an outcome but as

a developmental process that is shaped and influenced by all the linguistic systems that

give us knowledge about words.

Component

Three major types of processes have been recognized as prominent in the

development of fluency: orthographic, phonological, and semantic. Interestingly, Adams

(1990) added emphasis to morphological and syntactic knowledge systems. For example,

Adams proposed that both orthographic and semantic information are necessary for

learning morphological knowledge and that both patterns and roots (e.g., Latin and Greek

roots) make up many words. Researchers emphasize the connections that link

orthographic, semantic, phonological, and morphological systems. In other words,

reading fluency involves every process and subskill involved in reading. Consequently,

language comprehension processes and higher-level processes affecting language

comprehension (the application of world knowledge, reasoning, etc.) do not become fully









operative in comprehending text until the child has acquired reasonable fluency (Adams,

1990; Hoover & Gough, 1990; Perfetti, 1985; Stanovich, 1991; Sticht & James, 1984;

Vellutino et al., 1991; Vellutino, Scanlon, & Tanzman, 1994).

Factors of reading dysfluency

Why do some people fail to read fluently? There must be specific factors that affect

reading fluency. During the last two decades, many reading researchers have provided

evidence on factors related to reading fluency (Wolf, 1991; Wolf, Bowers, & Biddel,

2000). They suggested that there is evidence that impaired readers break down across the

continuum of processes. Through a thorough review of previous fluency literature, Meyer

and Felton (1999) provided a summary of the existing research on explanations of

dysfluency. Three major areas were identified within their review:

1. A breakdown takes place in the lower level processes. In Figure 2-3,(1) represents
a breakdown in this lower level At this level, dysfluent reading results from
deficits in phonological processing and /or orthographic processing systems that
affect the timing and coordination of these systems. Researchers indicated that the
largest contribution to slowed word-reading rate among dyslexic readers in the
slowed speed of processing in perceptual stages (Farmer & Klein, 1995; Tallal,
Miller, & Fitch, 1993).

2. In this level, a breakdown takes places when people make higher order semantic
and phonological connections between words, meaning, and ideas.(2) represents a
breakdown in connections to meaning processor. In this level, dysfluent reading is
resulted from deficits that occur after perceptual identification has been completed.

3. This breakdown may involve the slowed retrieval of names, meaning, or both
(German, 1992; Wolf et al., 2000).

4. A breakdown in syntactic processing can be the last factor of dysfluency.(3)
represents the last level of breakdown. In this level, dysfluent readers may have
deficits exhibited in a lack of prosody and rhythm in oral reading and a lack of
sensitivity to prosodic cues.



























Orthographic Phonological
Processor Processor




Print Speech


Figure 2-3. Adams's Reading Process Model with Breakdowns


Reading Comprehension

As mentioned previously, reading is an interactive, holistic process involving the

reader, the text, and instruction (Maria, 1990). The skillful reader is able to orchestrate a

complex system of skills and knowledge. Since reading is a holistic process it is difficult

to divide the concepts into fluency and comprehension. To be an effective reader, a well-

developed system of decoding and comprehension skills is required (Rumelhart, 1977).

Readers who decode quickly and accurately coordinate this process with the processes

involved in the comprehension of the text (Anderson, Heibert, Scott, & Wilkinson, 1985).









Definition

Reading comprehension is the ability to understand or get meaning from text (any

type of written material). Fountas and Pinnell (1996) emphasized that reading

comprehension is not something that only happens after reading. He defined reading

comprehension as the thinking done before, during, and after reading. (Fountas & Pinnell,

1996, p. 323). Others researchers defined it as the interaction between the text being read

and the reader's existing prior knowledge, and expectations will generate meaning and

comprehension (Leu & Kinzer, 1995). Comprehension is the reason for reading and a

critical component of all content learning.

Level of reading comprehension

Specifically, Vacca and Vacca (1996) divided reading comprehension into three

levels:

1. Literal (reading the lines): People get information explicitly from the text.

2. Interpretive (reading between the lines): People put together information and use
information to express opinions and form new ideas.

3. Applied (reading beyond the lines): People perceive relationships and make
inferences.


Factors of reading comprehension

Broadly, many researchers identified three factors affecting reading

comprehension: reader, text, and activity. In addition, variations in the context in which

they occur play an important role. However, these factors are intertwined and do not

operate independently of the others in any authentic act of reading comprehension. First,

effective readers bring to the task of reading an array of capabilities and dispositions.

Reader differences in such capabilities as fluency in word recognition, oral language









ability (Pearson & Fielding, 1991; Pressley, 2000), and domain knowledge ((Dole, Duffy,

Roehler, & Pearson, 1991), along with differences in such dispositions as the reader's

motivation, goals, and purposes, are important sources of variability in reading

comprehension. Second, text is the second element affecting reading comprehension.

Duke (2000) documented that children who attend schools in poor districts have many

fewer texts available than do children who attend schools in richer areas; the availability

of texts in homes and libraries varies similarly. Texts that treat certain social issues or

that require an interpretation and appreciation of alternative perspectives may be

considered inappropriate by parents from some cultural or religious groups. Text factors

thus interact with reader, activity, and context in determining the difficulty of

comprehension. Third, activity refers to the acts a reader engages in with a text, and it

encompasses purpose, operations, and consequences. Reading comprehension can be

improved with strategies when readers themselves apply across a number of different

texts (National Reading Panel, 2000). For example, teachers may teach students to

activate their own background knowledge, to draw inferences as they read, or to restate

information in the text (Chan, Cole, & Barfett, 1987). The successful activation of two

areas of reading-fluency and comprehension-completes reading.

Significant Predictors of Reading Success

Many studies including longitudinal studies have been conducted to identify the

predictors of current and later reading ability. Such studies yielded many important

predictors of RD. Currently, phonological awareness and rapid-naming variables are very

frequently involved in prediction studies. Further, a critical role of phonological and

naming speed factors in classifying reading disabilities have been identified by many

reading studies (e.g., Morris et al., 1998). Within the field of reading-related vision









research, eye movement and visual processing factors are considered as key indicators of

reading difficulties. Therefore, these four individual predictors are discussed further in

this section.

Linguistic Predictors

Among various linguistic indicators, the two most frequently identified factors that

are significantly correlated to later reading ability are phonological awareness and rapid-

naming deficit. During the last 25 years, research has provided a greater understanding

that phonological awareness is a foundation of reading acquisition. As a consequence, an

assessment of phonological awareness skills always has been included in early

identification of RD. Along with phonological awareness, a wider variety of recent

studies has highlighted rapid-naming skills as being highly correlated with their later

reading abilities (e.g., Wolf, 1999; Wolf & Bowers, 1999). Therefore, the following

section provides a neutral overview of the two skills and evidence of empirical studies as

key linguistic predictors of RD.

Phonological awareness

In the 1970s, Isabelle Liberman and Donald Shankweiler introduced the significant

concept of phonological awareness to the field of reading (Liberman, 1973; Liberman et

al., 1974). They argued that an alphabetic system requires that readers have awareness

that words are made up of individual speech sounds (i.e., phonemes). Hence, if a child is

not sufficiently aware of the speech segments in spoken words, the child may have

difficulties in reading. These researchers and their colleagues studied young children and

poor readers and confirmed that they were lacking in awareness of the sound structure of

words (Liberman et al., 1974); that is, the children had difficulties performing activities

that require one to reflect on the speech sounds in words (e.g., counting the number of









phonemes in a word) (Liberman et al., 1974). Currently phonological awareness is the

widely accepted term and generally is used as an umbrella label for all levels of

conscious awareness of the sound elements in words. Simply, phonological awareness is

an ability to understand a word's sound structure, and it is critical for the efficient

decoding of printed words and the ability to form connections between sounds and letters

when spelling (Torgeson, 1997). In general, researchers indicate phonological awareness

includes five levels of language skills (e.g., Chard & Dickson, 1999). According to Chard

and Dickson (1999), these five levels of phonological awareness skills (rhyming,

phoneme identification, blending, segmentation, and manipulation) emerge following a

developmental hierarchy. The earliest developing phonological awareness skill was

rhyming and is the most challenging phonological awareness skill. It involves the ability

to manipulate phonemes (Chard & Dickson, 1999). More specifically, Adams (1990)

further described five levels of phonological awareness in terms of abilities to (a) hear

rhymes and alliteration as measured by knowledge of nursery rhymes, (b) do oddity tasks

(comparing and contrasting the sounds of words for rhyme and alliteration), (c) blend and

split syllables, (d) perform phonemic segmentation (such as counting out the number of

phonemes in a word), and (e) perform phoneme manipulation tasks (such as adding,

deleting a particular phoneme, and regenerating a word from the remainder). As

explained above, there are various dimensions of phonological awareness. A tapping task,

which was first used by Liberman (1973), is involved in developing children's

phonological awareness. Liberman had children use a wooden dowel to tap the number of

sounds they could hear in spoken sentences and then in words (Liberman, 1973). Over

the last 25 years, a huge amount of research has suggested that students who enter school









with little knowledge of phonological awareness experience less success in reading than

students who have sufficient knowledge of phonological awareness (Adams, 1990; Mann

& Brady, 1988; Stanovich, 1988). Overall, the studies of phonological awareness for

young children are twofold: First, the power of phonological awareness to predict

children's reading ability enables early identification of students at risk for RD (e.g.,

Mann, 1993; Torgeson et al., 1997; Watson et al., 2003), and, second, the effectiveness

on instruction, based on phonological awareness, has been recognized when compared to

other types of instruction or no instruction (O'Connor, Jenkins, Leicester, & Slocum,

1993; Torgeson et al., 1997; Vellutino & Scanlon, 1987). Research firmly established the

strong relationship between phonological awareness and reading acquisition. Along with

phonological awareness, some intervention studies suggest that naming speed may be a

critical component in successful reading, and it may be a reason why not all children

respond to phonological awareness (e.g., O'Connor et al., 1993; Torgeson Wagner,

Simmons, & Laughton, 1990). Although phonological awareness has been considered as

the strongest indicator of RD, until recently we were not able to form an image of the

brain of a living person engaged in reading (Clark & Uhry, 1995). The relationship

between phonological awareness and RD has been confirmed through recent studies

using fMRI (e.g., Pugh et al., 2000; Temple et al., 2000; Temple et al., 2003; Shaywitz et

al., 2000; Wood, Relton, Flowers, & Naylor, 1991). For example, Temple et al. (2000)

examined dyslexic and normal readers (age ranged from 8-12) by using fMIRI during

reading. They used fMRI in order to examine neurological differences between dyslexic

and normal readers during phonological and orthographic tasks of rhyming and matching

visually present letters. During letter rhyming for a phonological task, both dyslexic and









normal readers had activity in left frontal brain regions, whereas only normal readers had

activity in left temporo-parietal cortex. During letter matching performance for an

orthographic task, while normal readers showed activity throughout extrastriate cortex,

dyslexic readers had little activity in that region. Their findings suggest that children with

RD may be characterized by disruptions in the neural bases of both phonological and

orthographic processes (Temple et al., 2000).

Rapid naming

Rapid naming is the ability to recall quickly and to verbalize the name of a

presented object. Naming speed is typically assessed by Rapid Automatic Naming Test

(RAN; Denckla & Rudel 1974, 1976), which requires children to name familiar colors,

pictured objects, digits, and letters (Wagner & Togeson, 1987, as cited in Manis et al.,

1999). Originally, Denckla and Rudel (see Clark & Uhry, 1995) found that the length of

time it took to name these stimuli varied with age and reading ability. Older children

were able to name quicker than younger children, and children with higher reading ability

were quicker that poor readers. This hypothesis was confirmed through continuing

studies. Recently, there is a growing consensus that RAN is highly correlated with

reading ability (e.g., Bowers & Swanson, 1991; Wolf, 1991) and serves as an accurate,

early, time-efficient predictor of reading achievement (Bowers & Wolf, 1993; Wolf,

1991). Over the last two decades, an extensive body of research indicated that rapid-

naming skill is strongly related to reading development (Badian, 1993; Bowers, 1996;

Wolf, Bally, & Morris, 1986; Wolf & Bowers, 1999). Unfortunately, researchers have

provided mixed findings, and such findings vary greatly across studies. For example, the

relationship between rapid-naming and phonological deficits in RD becomes a subject of









controversy. More specifically, Manis and his colleagues (1999) indicated two remaining

issues in rapid naming:

1. Is a rapid-naming skill is simply a part of phonological skills?
2. What processing skills are involved in rapid naming?

As described previously, several researchers argued that rapid naming is correlated

with reading because these two are highly related to phonological coding skills (Share,

1995; Wagner & Togeson, 1987). Notably, Torgeson et al, (1997) pointed out that one

limitation of predictive studies is the autoregressive effects of prior reading skills. When

they did not control the effect of prior reading skills, they were able to indicate separate

contributions of phonological and rapid-naming factors. On the other hand, when the

autoregressive effect was factored out, rapid naming was no longer accounted for reading

outcomes, but phonological awareness was still accounted for variance in reading. In

contrast, Meyer, Wood, Hart, and Felton (1998) reported somewhat opposite findings.

They conducted a study with older children, Grades 5 and 8, and yielded different results.

They found that rapid naming was the only variable that significantly predicted later

word identification for children with RD (not for average readers) when the

autoregressive variables (Grade 3 word identification, IQ, and socioeconomic status)

were controlled. Further, when autoregressor was factored out, rapid naming still

accounted for variance in later reading development, but phoneme segmentation did not.

Researchers (Bowers & Wolf, 1993; Wolf, 1997; Wolf & Bowers, 1999) argued that even

though rapid naming may rely on and share variance with some phonological tests, some

aspects of rapid naming such as speed of processing and sensitivity to temporally ordered

information cannot be explained by phonological processing (Manis et al., 1999). Some

researchers have tried to provide evidence by examining correlations between









phonological skills and rapid-naming skills of study participants (Felton & Brown, 1990)

and identifying subgroups of children with RD (e.g., Korhonen, 1991; Lovett, 1995; Wolf

& Bowers, 1999). For example, Korhonen (1991) conducted cluster analyses to identify

subgroups among third graders with learning disabilities (LD). They were able to identify

one subgroup of children who were characterized by their slow rapid naming. They

further noted that this particular subgroup showed the least progress in reading

achievement. Similarly, Lovett (1995) was also able to identify three subtypes among

children with severe RD by assessing pretreatment diagnostic testing. Three subtypes

were identified according to whether they had a phonological deficit only, a naming

speed deficit only, or both deficits. Compared to children in a control group, all subjects,

in general, made significant progress after 35 hours of intervention, which involved

phonological awareness and word identification strategy training. However, the

subgroup with a phonological deficit only made much further gains compared to the other

two groups. This finding suggests that the remedial program, which is mainly based on

phonological awareness, is not very effective for children who mainly have a naming

speed deficit. In that sense, a rapid-naming predictor is a useful avenue of identifying

children who may need additional assistance above and beyond the kinds of remedial

approaches that are effective for poor readers with no naming deficit (Scarborough,

1995). Additionally, a reading-related vision research added a naming variable for the

multiple regression model (Eden et al., 1995) and indicated that naming to confrontation

on the Boston Naming Test proved to be significantly more difficult for the RD group

than for the non-RD group.









Vision-Related Predictors

Since many studies estimated that 70% to 85% (e.g., Getman, 1981; McGurk &

McDonald, 1976) of what we learn comes through vision. For example, McGurk and

McDonald (1976) addressed the degree of development of how people's visual skills

significantly determine how effectively one is able to learn. According to Getman (1981),

there is a significant relationship between vision and school achievement. Specifically,

empirical studies indicate that 80% of all learning during a child's first 12 years is

obtained through vision (Getman, 1981). One interesting study examined the link among

learning disabilities, delinquency, and visual ability and found that significant numbers of

juvenile offenders had one or more visual deficiencies (Johnson, Nottinghamn, Stratton,

& Jaba, 1996). Further, Johnson and his colleagues (1996) reported that 97% of students

with behavioral problems and learning difficulties failed at least one of the vision subtests

(tracking, fusion, acuity-near, acuity-distant, convergence, hyperopia, color vision, and

visual motor integration). Byrne (1992) described several concerns in the field of reading

research. One important viewpoint he highlighted was the "difficulties in visual

perception as underlying RD may not be a dead issue" (p. 2). Although the visual aspects

of RD were largely ignored until recently, a wider variety of literature indicated visual

and ocular motor abnormalities, that is, eye movement and a visual processing factor in

subjects with dyslexia (Fisher & Weber, 1990) as risk factors of reading difficulties.

Therefore, these two skills are discussed further in this section.

Eye movement

Specifically, visual efficiency skills include accommodation, binocular vision, and

ocular mobility. Some visual efficiency skills such as eye movement, tracking, and

fixation are frequently mentioned in reading-related research. Reading begins with the









input of information through the visual-motor system. Basically, a variety of eye-related

muscles are involved in eye movement. Measuring eye movement involves an

examination of the number of fixations, duration of fixation, and the sequence of picture

elements fixated. These eye fixations are called saccades and become more articulated as

a child grows. When a preschooler looks between different objects, he or she typically

uses head and neck movements rather than eye movements. By age 6 or 7, the child

should be able to read a book or a board with proper eye movements and without moving

his or her head or body (Schieman, 2002). Due to body or neck movement, a student who

has low functioning eye-movement skills may have difficulties in keeping his or her

place while reading or adding numbers in columns (Scheiman, 2002). Further, reading

speed may be reduced due to slow eye movements, recurrent loss of place, or due to a

small span of recognition (the width of text that is processed per fixation). Therefore, we

may conclude that underachieved or incorrectly developed eye movement may be

associated with reading difficulties. A study conducted with a selected group of 17

children with reading disabilities and 21 without reading disabilities found that visual

tasks such as eye movement and visual search are able to differentiate good and poor

readers as well as phonological tests (Eden et al., 1993). Specifically, Eden and his

colleagues (1995) found some visual deficits including vertical tracking and vergence

control in eye movement skills were stronger predictors than some other verbal skills.

Their findings strongly support the hypothesis that some children with reading difficulties

have visual problems. In summary, Rayner (1995) recapitulated the process of reading

and eye movements by introducing five principal issues: (a) the span of effective vision,

(b) integration of information across eye movements, (c) eye movement control (where to









fixate), (d) eye movement control (where to move), and (e) models of eye movement

control. His research confirmed that the average span of perception can be determined by

computing the number of words per fixation. As a result, slow reading speed may be

highly correlated with eye movement.

Visual processing

Visual processing skills refer to a group of visual cognitive skills used for

extracting and organizing visual information from the environment and integrating this

information with other sensory modalities and higher cognitive functions (Scheiman,

2002). Visual processing skills are classified as visual spatial skills, visual analysis skills,

and visual motor integration skills. As mentioned earlier, Adams (1990) stated that

reading is accomplished through complex and interrelated processes, one of which is the

visual process. Therefore, when determining the relationships between vision and

reading, tests of visual processing should be involved rather than limiting the evaluation

to eye health and visual acuity. For a skilled reader, the identification of a printed word

begins with a visual process that operates on the visual forms of letters that make up a

word and then instantaneously generates other processes including the phonological

decoding process, to which printed letters and the sounds of the language correspond

(Adams, 1990). Many researchers have considered visual processing one of several

factors that may affect children's reading abilities. There are some concrete examples.

Kavale (1982) used meta-analysis to establish the relationship between visual-perceptual

skills and reading. In his review, the 161 studies produced a total of 1,571 correlation

coefficients, of which 1.062 indicated the relationship. Across all visual perceptual skills

(visual discrimination, visual memory, visual closure, visual spatial relationship, visual-

motor integration, visual association, figure ground discrimination, and visual-auditory









integration) and reading achievement skills, the 1,062 coefficients yielded an average

correlation, 376 with a standard deviation of .139 and a standard error of .004. The

correlation ranged from .01 to .89 with a median of .408. Among the visual perceptual

skills, he identified that visual memory and visual discrimination skills were most

significantly related to reading achievement. More interestingly, even after IQ was

factored out, visual discrimination ability as measured in these studies was still highly

related to reading (Kavale, 1982). The findings of his extensive meta-analysis imply that

visual perceptual skills appear sufficiently associated with reading achievement to be

considered among the complex of factors to the prediction of reading ability (Kavale,

1982). In order to examine the proportion of children with RD, Watson and Willow

(1993) reviewed 14 clinical studies and indicated that approximately 20% of low-

achieving readers have some type of visual problem such as a visual-spatial problem, a

visual-perceptual deficit, or a visual memory problem. Furthermore, Aaron and Joshi

(1992) reported that children are more likely to process printed words as visual

information rather than as phonological information. They also found that some

participants tend to disregard the orientation of letters and words and then make reversal

errors. Vellutino and Scanlon (1988) also concluded that poor readers make more

reversal errors due to their weakness in decoding skills and that they are therefore less

coherent in their visual processing of printed words. After those poor readers were taught

to use strategies regarding visual processing, they made very few reversal errors. With

more wide perspectives, one research study revealed the relationship between general

academic performances and visual perceptional skills. Watson et al. (2003) administered

standardized sensory, perceptual linguistic intellectual, and cognitive tests to 470 children









who entered the first grade over a 3-year period. The results of 36 tests and subtests were

divided into four factors, including reading-related skills (phonological awareness, letter

and word recognition); visual cognition (visual perception, visual spatial perception,

visual memory); verbal cognition (language development, vocabulary verbal concepts);

and speech processing (the ability to understand speech under difficult listening

conditions). Not surprisingly, the profiles of factor scores showed that reading-related

skills have the biggest impact on reading and other academic performance. Notably, the

second strongest predictor of both reading and mathematics grades was the visual

cognition factor.

Overall, vision researchers addressed vision as not entirely innate but as a skill that

continues to develop after birth and is influenced by the visual environment and the

individual's experience. If a child's visual system has not developed to its optimal level,

visually demanding activities such as reading may appear to be inefficient or

uncomfortable for the child. Although a child has a good visual acuity (clarity of sight), it

does not necessarily assure that the child does not have any visual problems. An extended

view of the definition of vision may enhance the understanding of children with reading

difficulties in terms of visual skills.

Implication for Research

The review of the literature provided a theoretical basis and empirical evidence on

early identification of reading disabilities. Empirical studies to date have presented

knowledge of factors necessary for early identification of reading disabilities. Overall,

such evidence indicated that phonological awareness and other factors including visual-

naming skills and visual-processing abilities are associated with the development of

reading skills. However, those evidence are still not clear. Some researchers have argued









that questions and disagreements are still remained about the visual factors as one of the

contributors to reading disabilities. However, the findings of many other studies from

various perspectives allow us to assume that there may be a considerable relationship

between a visual deficit and RD, at least for a certain percentage of children. There are

the remaining issues that need to be further studied in the future.

First, this present review revealed that there are still some questions surrounding

the visual deficit factor, and it is worthy of further investigation. Although a limited

number of children with RD may have visual characters, we need to identify risk factors

at the earliest stage to prevent or reduce influences of RD. Although previous literature

indicated mixed findings on the relationship, there has been an increasing interest in the

area of visual skills since more recent empirical research provided concrete evidence.

Furthermore, there is a growing consensus that at least a group of children with RD have

vision-related RD. However, little research has been done to indicate the specific

proportion of children with RD who have visual deficits. Therefore, it is important to

understand better what proportion of children with RD have visual deficits and whether

these deficits are truly related to reading achievement (Feagans & Merriwether, 1990).

Second, the autoregressor effects are an unsolved issue in rapid-naming research.

As noted previously, Torgeson et al. (1997) discussed a limitation of predictive studies in

terms of the autoregressive effects of prior reading skills. Torgeson et al. (1997) argued

that rapid naming is not a separate predictor of RD and the rapid-naming skills are

included in phonological skills by providing evidence. When they did not control the

effect of prior reading skill, they were able to indicate a separate contribution of

phonological and rapid-naming factors. However, when the autoregressive effect was









factored out, rapid naming no longer accounted for variances in reading outcomes but

phonological awareness still accounted for variances in reading (Torgeson et al., 1997).

On the other hand, Meyer et al. (1998) found contradictive results. When autoregressive

variables were controlled, rapid naming still accounted for variance in later reading

development, but phoneme segmentation did not. Therefore, it is important that

autoregressor effects should be clearly considered when researchers conduct prediction

studies.

Finally, a growing body of research established the necessary measurements for

effective identification of early elementary students with reading disabilities. However,

questions still remain about the most effective and reliable set of measurements for early

primary graders of reading disabilities. Scarborough (1998) analyzed predictive studies

and suggested that researchers use a multivariate approach to early, accurate screening of

children. In her review, she indicated that such studies use a wide variety of variables that

ranged from cognitive skills to family and school-based risk factors. In addition,

combinations of variables used in each study vary greatly across studies. Although

various types of measures are effective to detect children with RD in the early ages, we

cannot give every possible assessment to each student. This approach is not cost and time

effective. Therefore, it is essential to identify the most promising combination of

predictors of early reading achievement that directly relates to the reading process

(Goswami & Bryant, 1990).

Conclusions

The review of literature adds support to two conclusions:

1. Phonological processing is the most powerful cognitive variable determining later
reading ability.









2. Phonological awareness is not enough to explain all types of reading disabilities.

Along with phonological awareness, the current review searched for other

significant variables that may be associated with RD. Many connectionist theories and

empirical researchers provided an insight that there may be other aspects of reading

related to RD. From this review process, rapid-naming and visual deficits were seriously

considered as significant variables determining later reading development. Overall,

studies of phonological awareness, naming speed, and visual deficits as predictors of

reading development have value in helping to understand the nature of the cognitive

processes underlying reading. Therefore, the purpose of this study was to determine the

effects of three variables including phonological awareness, rapid-naming, and visual

skills on the reading levels of early elementary students. Further, this study investigated

how the three variables are interrelated to their reading levels.














CHAPTER 3
METHODS


Introduction

The purpose of this study was to determine the effects of various variables

including phonological awareness, rapid-naming, and visual skills on the reading levels

of early elementary students. In addition, this study investigated how these variables are

interrelated to their reading levels. In Chapter 3, the research methods of the study are

described including the participants, research instruments, data collection procedures, and

methods of data analysis employed by this study.

Research Questions

Three research questions were investigated (Table 3-1).

Table 3-1. Research Questions
Research Questions
Question 1 Is the combination of phonological awareness, rapid-naming, eye-
movement/orthographic skills and visual perception skills significantly
associated with the level of reading fluency?

Question 2 Among phonological awareness, rapid-naming, eye-
movement/orthographic skills and eye-movement/orthographic skills,
what is the most strongly correlated with the level of reading fluency?

Question 3 Among phonological awareness, rapid-naming, eye-
movement/orthographic skills and visual perception skills what latent
variables should be included in this reading fluency model?









Hypotheses

Based on the research questions and prior discussion, the following null hypotheses

were generated for this study (Table 3-2).

Table 3-2. Hypotheses
Null Hypotheses


Hypothesis 1


Hypothesis 2


Hypothesis 3


Hypothesis 4


Hypothesis 5


Hypothesis 6


Hypothesis 7


Hypothesis 8


Hypothesis 9


Hypothesis 10


Hypothesis 11


Phonological awareness, rapid-naming, and visual skills are not
significantly associated with reading fluency.

Phonological awareness measures are not significantly associated
with reading fluency.

Rapid-naming measures are not significantly associated with reading
fluency.

Eye-movement/Orthographic measures are not significantly
associated with reading fluency.

Visual perception measures are not significantly associated with
reading fluency.

Phonological measures are not significantly correlated with rapid-
naming measures.

Phonological measures are not significantly correlated with Eye-
movement/Orthographic measures.

Phonological measures are not significantly correlated with visual
perception measures.

Rapid-naming measures are not significantly correlated with eye-
movement/orthographic measures.

Rapid-naming measures are not significantly correlated with visual
perception measures.

Eye-movement/orthographic measures are not significantly correlated
with visual perception measures.









Methods

Settings

Five school sites in a north central Florida district were selected for this study. The

schools were selected to represent a distribution of both low- and middle-income

populations. The socioeconomic level of the schools is based on the percentage of

students in the school receiving lunch support. The specific percentage of students who

participate in the free and reduced lunch program and percentage of nonwhite students

enrolled in each school are presented in Table 3-3.

Table 3-3. Descriptive Information for Schools

School Total Enrollment % of Nonwhite % of Students that
Students receive lunch
support
1 611 28 28

2 441 43 53

3 577 59 80

4 620 76 79

5 985 43 31


Participants

For this investigation, 210 second- and third-grade students from five public

schools in an urban area of north-central Florida participated in this study. Reading

researchers have come to believe that success or failure in reading is largely determined

by the student's experiences in Grades Kindergarten through Grade 4. As a consequence,

reading research on the determinants of processes of reading achievement has been

focused on this age range. Therefore, this study took second- and third-grade students as

study participants who were falling into the most commonly researched developmental









stage. Students with diverse academic achievement levels were recruited; however,

inclusion was limited to children who were primarily English speaking and currently had

no special education due to mental retardation, visual impairments, and hearing

impairments. The sample size was appropriate for an SEM model with three factors for

this study. According to Stevens (1996), a good rule of thumb is 15 cases per predictor in

a standard ordinary least squared multiple regression analysis. Since SEM is closely

related to multiple regression in some respects, 15 cases per measured variable in SEM

can be reasonable. Bentler and Chou (1987) suggested that researchers who use SEM

may go as low as five cases per parameter. More generally, Loehlin (1992) reported the

result after reviewing the literature; in the class of model with two to four factors, the

investigator should plan on collecting at least 100 cases, with 200 being better (if

possible). Demographic data in this study are shown in Table 3-4.

Table 3-4. Demographic Data
Variables Total N=210 Frequency Percentage(%)
Gender
Male 97 46.2
Female 113 53.8
Ethnicity
Caucasian 112 53.3
African-American 65 31.0
Hispanic 21 10.0
Asian-Pacific 12 5.7
Grade
2nd Grade 111 46.2
3rd Grade 99 53.8


Forty six percent (n=97) were male students and 54% (n= 113) were female

students. The sample of five public schools included a diverse population; 53% were

Caucasian; 31% were African-American; 10% were Hispanic; and 6% were Asian-









Pacific Islanders. Forty six percent were second graders, and 54% were third graders.

Their ages ranged between 7 and 10 with the majority being 8 or 9.

Measures

The early elementary students who participated in this study completed twelve

assessments with the instruments. Table 3-5 presents all the variables that were examined

in the study. In addition, descriptions of the test instruments are presented in detail in the

next section.

Table 3-5. Test Variables


Latent Variable:
Factor
Phonological
Awareness

Rapid- Naming


Independent
Variable


Orthographic and Eye
Movement Skills


Visual Processing
Skills


Observed Variable (Test Instrument)

1. Phoneme elision (CTOPP)
2. Blending phonemes into words
(CTOPP)
1. A serial naming of digits
(CTOPP)
2. A serial naming of letters
(CTOPP)

1. Orthographic (J L- RRT)
2. Eye movement (DEM)-Vertical
3. Eye movement (DEM)-
Horizontal
1. Visual Discrimination (TVPS)
2. Visual Memory(TVPS)
3. Visual Closure(TVPS)


Dependent Variable


Reading Fluency


1. Oral Reading Fluency I
(DIBELS)
2. Oral Reading Fluency II
(DIBELS)


Abbreviations:
Dynamic Indicators of Basic Early Literacy Skills (DIBELS); CTOPP: Comprehensive
Test of Phonological Processing, J L- RRT: Jordan Left-Right Reversal Test; DEM:
Developmental Eye Movements Test; TVPS:NM-R: Test of Visual-Perceptual Skills
(non-motor)-Revised.


Variable









Data Collection

Procedure

Permission forms for participation in the study were obtained from at least one

parent or guardian of each study participant. The tests were administered individually to

the study participants, who returned signed parental permission slips and met the

inclusion criteria previously presented, in a designated place. All test items from the

research instruments were read and carefully explained to the participants so that those

with reading difficulties could participate in the study. This researcher and six other

research assistants administered the research tests to all participants. Six research

assistants were graduate students in education and took two half- day training sessions in

test administration. Each participant was administered the test battery from 28 March

2004 through 17 May 2004.

Test Instruments

Dynamic Indicators of Basic Early Literacy Skills (DIBELS). A subtest of

DIBELS, Oral Reading Fluency (ORF), was conducted. It is a measure that assesses

fluency with text-the ability to translate letters-to-sounds-to-words fluently and

effortlessly. The fluent reader is one whose decoding processes are automatic, requiring

no conscious attention and used as a dependent measure for assessing reading fluency. It

is a set of standardized, individually administered measures of early literacy

development. The tests are designed to be short (1 minute) fluency measures used to

regularly monitor the development of prereading and early reading skills (DIBELS). Such

capacity then enables readers to allocate their attention to the comprehension and

meaning of the text. The measures were developed upon the essential early literacy

domains discussed in both the National Reading Panel (2000) reports to assess student









development of phonological awareness, alphabetic understanding, and automaticity and

fluency with the code. Each measure has been thoroughly researched and demonstrated to

be reliable and valid indicators of early literacy development and predictive of later

reading proficiency to aid in the early identification of students who are not progressing

as expected. When used as recommended, the results can be used to evaluate individual

student development as well as provide grade-level feedback toward validated

instructional objectives. The website http://dibels.uoregon.edu displays reliability and

validity information for the DIBELS ORF: test-retest reliability for elementary students

ranges from .92 to .97; alternate form reliability ranges from .89 to .94, and criterion-

related validity ranges from .52 to .91. Spring benchmark goals for a trajectory of

progress of words correct per minute in grade-level material (Good, Simmons, &

Kame'enui, 2001)

Comprehensive Test of Phonological Processing (CTOPP). In order to assess

phonological awareness and rapid-naming, the Comprehensive Test of Phonological

Processing (CTOPP) is used. In particular, two subtests including a test of phoneme

elision and blending into words is utilized for assessing phonological awareness. The

subtest of CTOPP, the Phonological Awareness Quotient (PAQ), measures an

individual's awareness and access to the phonological structure of oral language.

Specifically, these two measures were administered to determine students' abilities to

process and manipulate sounds. One subtest of CTOPP, the Rapid-Naming Quotient

(RNQ) measures the examinee's efficient retrieval of phonological information from

long-term or permanent memory, as well as the examinee's ability to execute a sequence

of operations quickly and repeatedly (Wagner et al., 1999). A serial naming of digits and









letters is used for assessing rapid naming. The CTOPP provides percentiles, standard

scores, and age and grade equivalents. Subtest standard scores have a mean of 10 and a

standard deviation of 3. The Phonological Awareness has a mean of 100 and a standard

deviation of 15. Age and grade equivalents show the relative standing of individuals'

scores. The CTOPP was normed on over 1,600 individuals ranging in age from 5 through

24 and residing in 30 states. Over half of the norming sample came from children in

elementary school (through Grade 5), where the CTOPP is expected to have its widest

use. The demographic characteristics of the normative sample are representative of the

U.S. population as a whole with regard to gender, race, ethnicity, residence, family

income, educational attainment of parents, and geographic regions. The sample

characteristics were stratified by age and keyed to the demographic characteristics

reported in the 1997 Statistical Abstract of the United States. Reliability of the CTOPP

was investigated using estimates of content sampling, time sampling, and scorer

differences. Most of the average internal consistency or alternate forms of reliability

coefficients (content sampling) exceed .80. The test/retest (time sampling) coefficients

range from .70 to .92. The magnitude of the coefficients reported from all the reliability

studies suggests that there is limited error in the CTOPP and that examiners can have

confidence in the results (Wagner et al., 1999).

Jordan Left-Right Reversal Test (J L-RRT). The Jordan Left-Right Reversal

Test (J L-RRT) is used for assessing one visual skill regarding orthographic processing. It

tests memory for spatial orientation of upper/lower case letters and numerals. Visual

reversals have been commonly accepted as one of a constellation of symptoms of

learning disabilities in children (Jordan, 1980). The J L-RRT is a norm- referenced test









that assesses visual reversals of letters, numbers, and words in students ranging in age 5

through adult. This test was normed on over 3,000 individuals ranging in age from 5

through 12. The normative sample includes all socioeconomic levels, and 10% of this

sample came from nonwhite racial backgrounds. For this test, the child scans rows of

upper- and lower-case letters and numerals and crosses out any that are incorrectly

oriented. The score is the number of errors. Test-retest reliability coefficients for Level I,

given in the test manual, range from 0.87 (ages 8 and 9) to .0.98 (ages 6-6 to 6-11)

(Jordan, 1980).

Developmental Eye Movements Test (DEM). The DEM is used for assessing a

visual skill regarding eye movement. The most direct way to measure the quality of

fixations and eye movements with reading is by direct movements recording. However,

for young children, indirect measure is more applicable (Watson et al., 2003); therefore,

it was utilized for this study. The developmental eye movement test (DEM) was designed

to determine if children with reading defects have defective eye movements by

comparing horizontal and vertical movements in a number naming task. The DEM was

normed with 534 individuals selected from elementary school populations. The sample

was equally drawn from rural public schools, suburban public schools, and suburban

parochial schools. Each age and grade group was composed of an equal number of males

and females. This test incorporates two number-calling subtests, one in vertical array and

one in horizontal array. The vertical subtest measures an individual's visual-verbal

automatic calling skills, whereas the horizontal subtest assesses ocula-motor function.

This is the only test that separates the two functions (Garzia, Richman, Nicholson, &

Gaines, 1990). The test also includes a pretest component in which the child must read a









row of numbers quickly to determine automatic number knowledge. If the child is unable

to perform this pretest satisfactorily, the DEM cannot be meaningfully administered.

Test-retest reliability coefficients for both vertical and horizontal DEM tests, given in the

test manual, are .89 and .86.

Test of Visual-Perceptual Skills (non-motor)-Revised [TVPS:NM-R]. The

visual processing assessment instrument for this study was Test of Visual-Perceptual

Skills (non-motor)-Revised [TVPS:NM-R] (Gardner, 1994). The TVPS: NM-R is a test

of visual perceptual processing ability that does not incorporate motor involvement.

Children aged 4 through 13 can be tested through this instrument. It is designed for

psychologists, learning specialists, optometrists, occupational therapists, remedial

specialists, educators, and other professionals. The TVPS: NM-R was designed to

measure seven distinct areas of a child's visual-perceptual skills: Visual Discrimination,

Visual Memory, Visual-Spatial Relationships, Visual Form Constancy, Visual Sequential

Memory, Visual Figure-Ground, and Visual Closure. Each of the seven subtests contains

16 items, arranged in order of difficulty, and can be administered to individual children.

The TVPS: NM-R is quick and easy to administer, requiring approximately 10 to 20

minutes of testing time and only 3 to 5 minutes to score. Standardization was achieved on

1,032 children, most of whom lived in the San Francisco Bay Area but some of whom

came from various states in the United States. The ages of the subjects in the

standardization sample ranged from 4 years to 12 years. The reliability coefficients for

the total score ranged from .83 to .91. Test results from the field study have been

converted into perceptual ages, scaled scores, and percentiles for each of the seven

subtests, along with perceptual quotients, percentiles, and median perceptual ages for the









combined scores of all subtests. It was chosen because its normative data are superior to

those of any other test of visual processing currently available. Among seven subtests of

visual perceptual skills, three subtests were selected, including visual discrimination,

visual memory and visual closure based on previous research findings. For example,

Kavale (1982) conducted a meta-analysis of the relationship between visual perceptual

skills and reading achievement. In this meta-analysis, 1,571 correlation coefficients were

collected from 161 studies. This analysis yielded that tests of visual memory, visual

discrimination, and visual closure are most significantly related with reading

achievement. Therefore, these three subtests were conducted to examine visual perceptual

skills. The descriptions of subtests are as follows:

1. Visual discrimination: A subject's ability to match or determine exact characteristics
of two forms when one of the forms is among similar forms.

2. Visual Memory: A subject's ability to remember for immediate recall (after four or
five seconds) all of the characteristics of a given form, and being able to find this
form from an array of similar forms.

3. Visual closure: A subject's ability to determine, from among four incomplete forms,
the one that would be the same as the completed form (Gardner, 1994).

Inter-coder Reliability

In general, inter-coder reliability examines the extent to which different

interviewers, observers, or coders using the same instrument or measure get equivalent

results. For this study, all the assessments for 45 students (out of 210 participants, 21.5%)

was conducted by two assessors. Actually, the main assessor administered all the tests,

and the other assessor observed and recorded separately. The inter-coder reliability for

this study was examined with Cohen's kappa by using SPSS. The range of coefficients on

each test was quite high, ranging from .92 to 1.00. Therefore, they were acceptable.









Data Analysis

Overview of Statistical Analysis

The SPSS and LISREL 8.30 (Joreskog & Sorbom, 1996) were used to conduct all

statistical analyses. The criterion for significance tests for all a priori hypotheses was set

at c =.05. To test the first hypothesis, the relationship between reading fluency and four

latent variables, structural equation modeling with latent variables was used. To test

Hypotheses 2 through 5, a regression analysis with SEM was used to examine the

relationship among four independent variables and one dependent latent variable. Finally,

to test Hypotheses 6 through 11, a correlational analysis was conducted to determine

significant relationships between four independent latent variables.

Structural Equation Modeling

Overview

Structural Equation Modeling (SEM) was conducted by using LISREL 8.30.

LISREL 8.30 program defaults to the maximum likelihood fitting function for estimating

a model's parameters (Joreskog & Sorbom, 1996). This method consists of a

measurement model to define hypothetical latent constructs in terms of measured

variables, and a structural model to depict relationships among latent constructs. The

SEM is a multivariate method combining aspects of factor analysis and multiple

regression in analyzing a set of interrelated relationships among manifest and latent

variables simultaneously. The advantages of SEM compared to most other conventional

statistical methods include the following capabilities.

First, the basic statistic in SEM is the covariance. Covariance is generally a more

appropriate statistic in SEM than correlation, and covariance statistics convey more

information than correlations. A correlation matrix is not suggested since SEM assumes









that the variables are unstandardized. While conventionally analyses such as ordinary

least square regression attempt to minimize differences between observed and expected

individual cases, SEM aims to minimize differences between observed and expected

covariance matrices. In other words, SEM, based on the covariance statistic, attempts "to

understand patterns of correlations among a set of variables and to explain as much of

their variances" (Kline, 1998, pp. 10-11).

Second, SEM allows separating measurement error from hypothesized relationship

among constructs that cannot be operated by conventional analyses. For example, a

regression coefficient is composed of two elements: structural coefficient between the

independent and dependent variable and the reliability for the predictor variable. By

distinguishing a structural model from a measurement model, SEM can examine the

relationship among constructs that are not influenced by measurement errors (Newcomb

& Bentler, 1988).

Third, SEM allows incorporation of latent variables into analyses unlike

conventional analyses that focus only on observed variables. Since SEM is not limited to

relations among observed variables, it gives researchers more flexibility to study any

combination of relations.

Finally, SEM allows researchers to estimate very complicated multivariate

relationships, while conventional analyses cannot accommodate multiple indicators of the

same construct due to potential problems such as multicollinearity. Even conventional

multivariate techniques that deal with multiple dependent variables cannot examine

relations simultaneously and only examine a single relationship at a time. However, SEM









allows examining a series of relationships of multiple dependent variables

simultaneously.

Therefore, during the last three decades, SEM has become a widely used

methodology for specifying, estimating, and testing hypothesized interrelationship among

meaningful variables in behavioral and social sciences (Joreskog & Sorbom, 1996).

Applications of the SEM methodology in education research are also on the increase. In

order to apply SEM in estimating the causal model of the present study, all structural

modeling was performed with the LISREL program.

Assumptions of SEM analysis

SEM requires certain assumptions to be met such as continuous variables,

normality, and linearity of all relationships. Therefore, it is important to examine whether

the assumptions of SEM have been met. First, variables used in SEM are typically

continuous (Kline, 1998). All the variables used in the present study are continuous

variables. Another requirement of SEM is the linearity of relationship between exogenous

and endogenous variables. In order to examine the linearity, residual plots against

independent variables are closely examined. In case of a violation of the linearity

assumption, the use of transformed independent variables was considered to uphold these

requirements. For each variable, the present study evaluates the linearity assumption by

plotting the studentized residuals against the predicted values. Also, the correlation

matrix indicates that multicollinearity is not a problem for multivariate analysis as there

is only one correlation among the independent variables that is higher than .50.

Frequency distributions, skewness and kuutosis of the data, and normal probability plots

are inspected to examine normality. In addition, a covariance matrix analyzed with the

Maximum Likelihood method (ML) is applied for the present study because ML is found









to be fairly robust with the data that violate multivariate normality (Bollen & Long,

1993). Finally, this study replaces missing values with the mean values of each variable

based on valid responses because SEM is sensitive to the presence of missing variables.

Analytic Procedures for SEM

To assure that both models are properly specified, it is critical to follow the three

steps of SEM: (a) identify a proposed model, (b) evaluate the model fit, and (c) modify

the model if the original model fit is bad.

A Proposed model

First, SEM begins with what is a conceptual rather that a statistical component:

causal model. Therefore, based on previous literature and prior discussions, a reading

model was proposed. In general, the models were established through three distinct

phases. First, a measurement model confirmed the validity of the constructs that were

used in later structural models. The purpose of this model was to identify the constructs

with reduced error variance, and the constructs could be established by a factor analysis.

Second, a predictive structural model was created to test the direct effects of multiple

predictors on reading. Third, a mediated reading model was structured to test hypotheses.

Figure 3-1 presents the model for this study that includes three factors that can be related

to reading levels. A causal model is a hypothesis about the field of variables that affect

one or more dependent variables of interest, presented formally as a path diagram. The

diagram in Figure 3-1 demonstrates that early elementary children's level of reading is

correlated with phonological awareness, rapid-naming, and visual skills. The model was

overidentified to meet basic requirement for model identification. In other words, the

number of parameters must be less than the number of observations. The overidentified

model resulted in positive degrees of freedom, which allowed for the rejection of the









model. For this study a model was identified based on previous findings and hypotheses

in the literature.

In a path diagram, observed or measured variables are represented as rectangles,

whereas latent variables that have been derived are represented as circles or ovals.

Relations between variables in SEM, whether the variables are latent or measured are

represented by path diagram, with collections of variables connected by lines and arrows.

Variables with lines having one-way arrows pointed to them are dependent variables in

the model. In this model, one latent dependent variable, reading fluency, is indicated.

Variables that are the sources of lines with arrows and have no arrow leading into them

are independent variables. In this model, three latent independent variables are estimated:

phonological awareness, rapid-naming and visual skills. The curved arrow between latent

variables represents the covariance or correlation between the two independent variables

in predicting the dependent variables.

The overall model fit

The SEM requires that models be formally specified for estimation and model

testing (Hoyle, 1995). In terms of model testing, the first step estimates the goodness-of-

fit of the hypothesized research model. This is typically done using a chi-square (X2) to

determine if differences between the data and the model are statistically significant and

can evaluate the relative fit of models of differing complexity.






70















Reading
Fluency

















Figure 3-1. A Proposed Model

Maximum likelihood generates an estimated chi-square statistic and as such

provides a probability value indicative of the overall fit of the model since the

distribution of chi-square value is known (Bollen, 1989). The value of the chi-square

statistic, in relation to the model's degrees of freedom, determines whether the fitted

residuals, resulting from I 1(9), deviate from the population values of zero. The

differences between the corresponding elements of the implied covariance matrix

(theory) and the observed covariance matrix (reality) consist of the matrix of residuals.









The lower the chi-square value gained from the accumulation of these residuals, the more

likely the theory represents the reality.

However, such tests are sensitive to sample sizes, and the probability of rejecting

any model increases when a sample size increases even with minimal deviation from the

model. Therefore, Bentler and Bonnett (1980) suggested that X2/df ratio (df. degree of

freedom) as a more appropriate measure of model fit. This ratio should not be bigger than

5 for models with a good fit (Bentler, 1990). In addition to the chi-square statistic, the

root mean square error of approximation (RMSEA) and the non-normed fit index (NNFI)

are used for evaluation for the significance of the relations between the constructs of all

models. According to Bentler and Bonett (1980), these two indices were determined to be

robust with relatively small sample sizes and non-normal or skewed distributional form

(see also, Browne & Cudeck, 1993; Neuhaus & Swank, 2002). The second step in model

estimation is to examine the path significance of each association in our research model

and variance explained (R2 value) by each path. The LISREL reports raw and

standardized estimates for all specified paths, along with standardized errors and test

statistics for each path.















CHAPTER 4
RESULTS

Research findings presented in this chapter consist of four parts. The first part

presents descriptive and inferential statistics on the sample. The second part of this

chapter discusses the results of correlational analysis using all of the available variables

for this study. Third, the results of how this model fits with the measurement model are

presented, and a structural model also is presented. Finally, another correlational analysis

and regression anlaysis with SEM approaches are discussed with the finalized SEM

model.

The following hypotheses were tested in this study:

H1 Phonological awareness, rapid-naming, and visual skills are not significantly
associated with reading fluency.

H2 Phonological awareness measures are not significantly associated with reading
fluency.

H3 Rapid-naming measures are not significantly associated with reading fluency.

H4 Eye-movement/Orthographic measures are not significantly associated with reading
fluency.

H5 Visual perception measures are not significantly associated with reading fluency.

H6 Phonological measures are not significantly correlated with rapid-naming measures.

H7 Phonological measures are not significantly correlated with Eye-
movement/Orthographic measures.

H8 Phonological measures are not significantly correlated with visual perception
measures.

H9 Rapid-naming measures are not significantly correlated with eye-
movement/orthographic measures.









H10 Rapid-naming measures are not significantly correlated with visual perception
measures.

H11 Eye-movement/orthographic measures are not significantly correlated with visual
perception measures.



Descriptive and Inferential Statistics

Demographic Characteristics

Descriptive statistics for the study's demographic characteristics are presented in

Table 4-1. The sample of this present study consisted of 210 second- and third-grade

students: 97 male students (46.2%) and 113 female students (53.8%). Approximately

53.3 % of the children were Caucasian; 3l1.0% were African-American children; 10%

were Hispanic; and 5.7% were Asian/Pacific Islanders. The mean age of the sample was

8.5 years. Additional descriptive information of the sample on observed variables is

presented in Table 4-2.

Table 4-1. Descriptive Statistics on Demographic Characteristics
Variables Frequency % Mean SD Range
Gender
Male 97 46.2
Female 113 53.8
Ethnicity
Caucasian 112 53.3
African-American 65 31.0
Hispanic 21 10.0
Asian-Pacific 12 5.7
Grade
2nd Grade 111 52.9
3rd Grade 99 47.1
Age 8.5 .50 7-10









Table 4-2. Means and Standard Deviations on All of the Observable Variables
Latent Variables
Observed Variables (scale) Mean SD Mini- Maxi-
mum mum
Phonological Awareness
Elision (correct items out of 20) 12.84 4.38 5 20
Blending into words (correct items out of 20) 12.73 3.27 6 20
Rapid Naming
Digit naming (time in sees) 36.76 10.68 17 89
Letter naming (time in sees) 40.67 11.79 17 104
Eye Movement/Orthographic Skills
DEM vertical test (time in sees) 47.68 14.16 25 109
DEM horizontal test (time in sees) 69.04 24.21 28 199
Reversal Test (reversal errors) 3.30 3.48 0 33
Visual Processing Skills
Visual Discrimination (correct items out of 16) 12.82 2.27 1 16
Visual Memory (correct items out of 16) 12.19 2.41 1 16
Visual Closure (correct items out of 16) 10.26 3.39 1 16
Reading Fluency
Fluency I (wpm*) 121.93 45.26 29 258
Fluency II (wpm) 109.48 39.39 21 208

* wpm: word per minute

Means and Standard Deviations on All the Observable Variables

The means and standards deviations for the subjects on each test are displayed in

Table 4-2. Additionally minimum and maximum scores are presented in the same table.

For phonological awareness, the mean scores on elision and blending into words of this

sample were 12.84 and 12.73, respectively. According to this result, the average score of

digit naming was approximately 4 seconds lower than that of letter naming. This result

implies children take longer time to recall letters than numbers. The average raw score

for the Jordan Left-Right Reversal test was 3.30 for this sample. This score indicates the

error score of this sample. According to the manual of the Jordan Left-Right Reversal

test, the raw score can be converted to percentiles for boys and girls separately. Jordan

indicates that approximately a 25 percentile score for boys in the age level of 8.6 is 4 and









for girls in the same age level is 3. The 50 percentile score for both boys and girls is

approximately 2. For the Fluency I test, a child was requested to read a story that is rated

as one grade below, and for the Fluency II test, another story that is rated as the same

grade of each student was provided. Therefore, the average score of Fluency II was lower

than that of Fluency I. Carver (1990) presented approximate reading rates for students in

Grades 2-12. According to Carver's (1990) analysis, the average reading fluency (wpm)

for 2.5 grade students is 121. The subjects in this current study are within a range of an

average reading fluency score.

Table 4-3. Criteria for Percentile Cutoff
Latent Variables Criteria for Percentile
Observed Variables (scale) > 25 25-50 50-100
%tile %tile % tile
Phonological Awareness
Elision (correct items out of 20) 0- 9 10-13 13-20
Blending into words (correct items out of 20) 0- 10 11-14 15-20
Rapid Naming
Digit naming (time in secs) >= 46 37-45 <=36
Letter naming (time in secs) >= 50 40-49 <=39
Eye Movement/Orthographic Skills
DEM vertical test (time in secs) >= 52 46-51 <=45
DEM horizontal test (time in secs) >= 64 55-63 <=54
Reversal Test (reversal errors)--- Boys >= 5 4 0-3
Reversal Test (reversal errors)--- Girls >= 4 3 0-2

Visual Processing Skills
Visual Discrimination (correct items out of 16) 1- 6 7-9 10- 16
Visual Memory (correct items out of 16) 1- 6 7-9 10- 16
Visual Closure (correct items out of 16) 1- 4 5- 7 8- 16
Reading Fluency
Fluency I +Fluency II (wpm) /2---2nd Graders <= 70 70- 90 >= 91
Fluency I +Fluency II (wpm) /2---3rd Graders <= 80 81- 110 >= 111









Descriptive Statistics on Percentile Scores

Descriptive statistics on percentile scores are also considered and presented in

Table 4-4. In terms of phonological awareness, children who belong to the lowest

percentile score group on elision are 36.2%. However, the number of children who

belong to the lowest group in terms of 'blending into words' was much lower (17.6%)

compared to elision (36.2%). Surprisingly, the number of children who acquired lower

than the 25 percentile score on DEM-horizontal test was too many (51.0%) compared to

DEM-vertical test. The criteria were determined by DEM manual. Therefore, this result

implies that this criterion score should be reconsidered and adjusted to the logical level.

Table 4-4. Descriptive Statistics on Percentile Scores
Latent Variables <25 25-50 >50
Observed Variables Percentile Percentile Percentile
(at risk) (some risk) (low risk)
Phonological Awareness n (%) n (%) n (%)
Elision 76 (36.2) 33 (15.7) 100 (48.1)
Blending into words 37(17.6) 87(41.4) 86(41.0)
Rapid Naming
Digit naming 35 (16.7) 36 (17.1) 139 (66.2)
Letter naming 34(16.2) 71(33.8) 105 (50.0)
Eye Movement/Orthographic Skills
DEM vertical test 60 (28.6) 39 (18.6) 111 (52.9)
DEM horizontal test 107 (51.0) 38 (18.1) 65 (31.0)
Reversal Test 64(30.5) 30(14.3) 116(55.2)
Visual Processing Skills
Visual Discrimination 6 (2.7) 25 (4.8) 194 (92.4)
Visual Memory 6 (2.7) 17(8.1) 187 (89.2)
Visual Closure 11(5.2) 41(19.5) 158 (75.2)
Reading Fluency
Fluency I + Fluency II 36 (17.1) 39(18.6) 135 (64.3)


Interestingly, the results of the three tests of visual perception skills are consistent

with prior studies. Many researchers have argued that 3% to 5% of all children may have

vision-related reading disabilities. For example, Watson and Willows (1993) reviewed 14









clinical studies and indicated that approximately 20% of low-achieving readers have

some type of visual problems such as visual-spatial problems, visual-perceptual deficits,

or visual-memory problems. This sample includes 36 children who show very low

reading fluency level (less than 25 percentile). Therefore, approximately 20% of children

with low-reading fluency may have vision-related reading problems.

Gender Differences

Group means on all measures were calculated, and a one-way analysis of variance

(ANOVA) was conducted to see if there are any gender differences. Table 4-5 provides a

summary of the one-way ANOVA of all the measures.

Table 4-5. ANOVA Summaries on Gender Differences
Latent Variables Boys Girls ANOVA
Observed Variables Mean SD Mean SD F p
Phonological Awareness
Elision 13.12 4.47 12.61 4.32 .714 .399
Blending into words 12.60 3.45 12.85 3.11 .308 .579
Rapid Naming
Digit naming 37.95 11.89 35.74 9.45 2.239 .136
Letter naming 42.00 13.61 39.52 9.88 2.321 .129
Eye Movement/Orthographic
Skills
DEM vertical test 49.10 16.42 46.45 11.83 1.836 .177
DEM horizontal test 70.45 26.49 67.84 22.12 .607 .437
Reversal Test 3.39 2.82 3.23 3.98 .112 .738
Visual Processing Skills
Visual Discrimination 12.79 2.43 12.86 2.04 .044 .835
Visual Memory 12.16 2.69 12.21 2.15 .020 .887
Visual Closure 10.12 3.55 10.38 3.25 .320 .572
Reading Fluency
Fluency I (wpm*) 119.64 45.89 123.89 44.82 .460 .498
Fluency II (wpm) 108.13 40.27 110.63 38.76 .209 .647

* wpm: word per minute









On all the assessment, the girls' performance was superior to boys'. However, no

significant gender differences were found. Because gender difference is minimal, these

findings of this study are consistent with pervious studies.

Differences Between Second and Third Graders

A one-way analysis of variance (ANOVA) was conducted to see if there were any

differences between second and third graders on all the observed measures. Table 4-6

provides a summary of the one-way ANOVA of all the measures.

Table 4-6. ANOVA Summaries on Group Differences (2nd/3rd Graders)
Latent Variables 2nd Graders 3rd Graders ANOVA
Observed Variables Mean SD Mean SD F p
Phonological Awareness
Elision 13.12 4.47 12.61 4.32 .950 .331
Blending into words 12.60 3.45 12.85 3.11 1.909 .169
Rapid Naming
Digit naming 37.95 11.89 35.74 9.45 15.541 .000**
Letter naming 42.00 13.61 39.52 9.88 14.332 .000**
Eye Movement/Orthographic
Skills:
DEM vertical test 49.10 16.42 46.45 11.83 9.391 .002**
DEM horizontal test 70.45 26.49 67.84 22.12 23.821 .000**
Reversal Test 3.39 2.82 3.23 3.98 9.249 .003**
Visual Perception Skills
Visual Discrimination 12.79 2.43 12.86 2.04 4.257 .040*
Visual Memory 12.16 2.69 12.21 2.15 1.771 .185
Visual Closure 10.12 3.55 10.38 3.25 5.850 .016*
Reading Fluency
Fluency I (wpm***) 114.14 40.01 130.66 49.25 7.171 .008**
Fluency II (wpm) 104.48 37.15 115.08 41.22 3.844 .051
* Statistically significant at the 0.05 level (2-tailed test)
** Statistically significant at the 0.01 level (2-tailed test)
*** wpm: word per minute

On all of the assessments, the performance of third graders was superior to that of

second graders. However, no significant group differences were found on phonological

awareness measures. On the other hand, there were statistically significant differences on

rapid-naming and eye movement/reversal measures between second and third graders.









For visual perception skill measures, the results of VD and VC measures indicated that

there are significant differences between second and third graders. However, the level of

VM seems to be similar between second and third graders. For the Fluency I test, third

graders seemed to read significantly faster than second graders. For the Fluency II test, p-

value was 0.051, indicating there is a marginal difference between second and third

graders.

Differences Between Students From Schools in Mid and Low SES Neighborhoods

Since data on lunch support of each participant were unavailable, the sample was

divided into two groups. The first group of students were currently going to schools

(School 3 and 4) in the lower SES neighborhood. More than 50 % of students in the

lower SES neighborhood schools (within this study) get lunch support. A one-way

analysis of variance (ANOVA) was conducted to determine if there are any differences

between the two groups of students on all the observed measures. The results indicated

that mean scores of students who go to schools in higher SES neighborhood (Schools 1,

2, and 5) are higher on all the measures. On most of the measures, significant mean

differences were detected between the two groups except rapid-naming variables. Table

4-7 provides a summary of the one-way ANOVA of all the measures.

Correlational Analysis

A correlational analysis was performed using all of the available observed

variables. The correlation matrix is presented in Table 4-8. Surprisingly, a closer

examination reveals that some indicators of theoretical constructs (e.g., phonological

factors) have very week correlations on reading fluency.









Table 4-7. ANOVA Summaries on Group Differences (Schools with Lower/Higher SES
Students)


Latent Variables
Observed Variables


Phonological Awareness
Elision
Blending into words
Rapid Naming
Digit naming
Letter naming
Eye Movement/Orthographic
Skills
DEM vertical test
DEM horizontal test
Reversal Test
Visual Processing Skills
Visual Discrimination
Visual Memory
Visual Closure
Reading Fluency
Fluency I (wpm)
Fluency II (wpm)


Students from
Schools in
Low- SES
Neighborhood
(n= 91)
Mean SD

11.71 4.29
12.05 2.93

37.75 11.03
42.43 13.36


48.29
69.91
4.15

12.37
11.41
9.54

103.43


14.46
25.78
3.26

2.45
2.60
3.33

40.68


Students from
Schools in
Mid- SES
Neighborhood
(n=119)
Mean SD

13.71 4.27
13.25 3.42

36.01 10.38
39.32 10.28


47.21
68.39
2.65

13.17
12.78
10.82

136.08


13.98
23.09
3.52

1.98
2.07
3.34

43.58


95.37 36.41 120.26 45.26 22.728 .000**


* Statistically significant at the 0.05 level (2-tailed test)
** Statistically significant at the 0.01 level (2-tailed test)
For the most part, correlations measured within constructs are higher than those

across constructs. Two variables of phonological awareness, elision and blending into

words, are significantly related to each other (r= .373, p < .01), and two rapid-naming

variables are highly correlated (r= .849, p< .01). However, unlike elision, the other

variable, blending into words, has relatively low correlations with dependent observed

variables, Fluency I and II (r= .110 and r=.081) and has no statically significant

correlations. The highest correlation is detected in the relationship between Fluency I and

Fluency II (r=. 936, p < .01). Since reversal test scores are calculated with error scores,

this variable is negatively related to two phonological variable, elision and blending into


ANOVA


F

11.264
7.119

1.370
3.633


.296
.204
9.948

6.896
17.872
7.655

30.640


P

.001*
.008*

.243
.058


.587
.652
.002*

.009*
.000**
.006*

.000**










words (r= .137, p < .05 and r= .011, p > .05). This result represents that children who

score higher on the reversal test are less likely to have higher score on elision. However,

it shows that there is no significant relationship between reversal and blending into

words.

Table 4-8. Correlation Matrix of Phonological Awareness, Rapid-Naming, and Visual
Skills and Reading Fluency
Elision Blending Digit RN Letter RN DEM-V DEM-H
Elision 1.00


Blending
Digit RN
Letter RN
DEM-V
DEM-H
Reversal
VD
VM
VC
Fluency I
Fluency II


.37**
-.15*
-.12
-.16*
-.20**
-.14*
.19**
.18**
.25**
.31 **
.25**


1.00
-.01
-.02
.01
-.10
-.01
.17*
.18*
.19**
.11
.08


1.00
.85*
.89*
.67**
.24**
-.27**
-.23**
-.24**
-.60**
-.61**


1.00
.81**
.66**
.23**
-.32**
-.27**
-.24**
-.60**
-.60**


1.00
.67**
.22**
-.27**
-.19**
-.24**
-.58**
-.57**


1.00
.22**
-.39**
-.32**
-.33**
-.57**
-.59**


Reversal VD VM VC Fluency I Fluency II
Reversal 1.00
VD -.28** 1.00
VM -.27** .55** 1.00
VC -.36** .49** .56** 1.00
Fluency I -.39** .35** .37** .50** 1.00
Fluency II -.38** .32** .36** .46** .94** 1.00

* Correlation is significant at the 0.05 level (2-tailed test)
** Correlation is significant at the 0.01 level (2-tailed test)

As predicted, all of the indicators of visual perception skills are positively related to

each other and statistically significant at the .01 level. Additionally, all of the indicators

of visual perception skills are positively related to Fluency I and II and statistically

significant at the .01 level. The correlations among constructs provide some preliminary

evidence for the hypothesized model. Since rapid-naming scores are the time in seconds

(seconds to name 40 letters/40 digits), they are negatively correlated with phonological









awareness and reading fluency factors. However, they are positively correlated with eye-

movement/reversal factors because eye-movement scores are also the time in seconds to

read numbers vertically and horizontally, and reversal scores indicate error scores.

Furthermore, all of the vision-related variables and reading fluency variables are

significantly related. Surprisingly, one of the observed variables, blending into words,

from a phonological factor has a very weak relationship with other reading-related

variables. One important fact has been detected through the correlation analysis is that

there are very high correlations between the two variables in rapid-naming and

vertical/horizontal eye movement variables. For example, DEM vertical test scores are

highly correlated with digit rapid naming and letter rapid naming (r= .849 and r= .805

respectively). Since the two constructs are significantly correlated with each other, if we

assign these scores to different constructs, it would violate one of assumptions in SEM,

multicollinearity. In some cases, multiple regression results may seem paradoxical. Even

though the overall P value is very low, all of the individual P values are high. This means

that the model fits the data well, even though none of the X variables has a statistically

significant impact on predicting Y. How is this possible? When two X variables are

highly correlated, they both convey essentially the same information. For the current

study, a rapid naming variable is very highly correlated with an eye movement variable.

In this case, neither may contribute significantly to the model after the other one is

included, but together they contribute a great deal. If you removed both variables from

the model, the fit would be much worse. So, the overall model fits the data well, but

neither X variable makes a significant contribution when it is added to the model last.

When this happens, the X variables are collinear, and the results show multicollinearity.









Therefore, modification of the measurement model may be necessary. The modified

measurement model is discussed in detail in the next section.

SEM Analysis

The SEM analysis provides comprehensive statistical techniques to examine and

investigate the relationship between measured variables and latent constructs. SEM is a

priori: a model makes theoretical sense and its statistical correspondence to the data is

reasonable. SEM is contained of two components; the measurement model and the

structural model and SEM approaches usually follow three steps. First, measurement

model fit was evaluated by using confirmatory factor analysis and second, the structural

model fit was evaluated. Finally, model modification is taken place if a model fit of a

structural model is worse than one of a measurement model. However, in this study, the

structural model fit was fairly good. In response, the last step was not necessary to

perform.

Measurement Model

The purpose of the measurement model was to create the constructs that would be

used in the later structural model. The validity of the constructs was determined by

confirming the significant relations between the observed variables that were assumed to

be theoretically related. The purpose of this model is to identify the constructs with

reduced error variance, and the constructs can be established by a factor analysis. A

confirmatory factor analysis was completed in order to determine whether the variables

load on five factors that can be labeled as phonological awareness, rapid-naming, eye-

movement/orthographic skills, visual skills, and reading fluency. Findings appear to have

adequately validated the five-factor structural model of phonological awareness, rapid-

naming, eye-movement/orthographic skills, visual processing skills, and reading fluency.









That is, the model of this study includes the five latent variables with 12 observed

variables. The variables loaded neatly onto five factors, as predicted. The following

model is the estimation of the author's model. The pattern of fixed and free parameters in

a structural equation model defines two components of the general structural equation

model: the measurement model and the structural model. The measurement model is that

component of the general model in which latent variables are prescribed (Hoyle, 1995).

The results of goodness of fit on the measurement model are shown in Table 4-9. The fit

of the original, hypothesized model (Figure 3-1) was adequate because the chi-square

was statistically significant (X2 (66) = 1708.237), and the additional fit indices showed

that the model has fairly good fit (NFI= .937, NNFI= .943, and RMSEA= .079).

However, as discussed in the previous section, there is a fatal problem in this proposed

measurement model. There are very high correlations between the two variables in rapid-

naming and vertical/horizontal eye movement variables. For example, DEM vertical test

scores are highly correlated with digit rapid naming and letter rapid naming (r= .849 and

r= .805). This is an evidence of multicollinearity in this model. Therefore, new

constructs were re-identified. Table 4-9 presents a modified constructs and goodness of

fit results for the newly identified measurement model. As predicted, a measurement

model fit became much greater with this modification (X2 (45) = 78.160, NFI= .954,

NNFI= .970, CFI= .980, SRMR= .050, RMSEA= .057).

Therefore, within the measurement model, all of the measured variables were

significant indicators of their respective latent variables, and no modifications of the

measurement model were needed since all of the fit indices met their criteria for good fit.









Table 4-9. Factors for Variables in the Overall Model
Measurement Model
Proposed Constructs
Observed Variables


Factor 1


Phonological Awareness
Elision
Blending into words


Factor 2 Rapid Naming
Digit rapid naming
Letter rapid naming


Factor 3 Eye-movement/Orthographic
Skills
Eye-movement-vertical
Eye-movement-horizontal
Reversal
Factor 4 Visual Processing Skills
Visual Discrimination
Visual Memory
Visual Closure
Factor 5 Reading Fluency
Reading Fluency I
Reading Fluency II
Goodness x2 (66) = 1708.237
of Fit NFI= .937
Results NNFI= .943
CFI= .962
SRMR= .072
RMSEA= .079
90% Confidence Interval of RMSEA
: (.059- .099)


I and Goodness of Fit Results of the

Modified Constructs
Observed Variables
Phonological Awareness
Elision
Blending into words
Rapid Naming/ Eye-movement
Digit rapid naming
Letter rapid naming
Eye-movement-vertical
Eye-movement-horizontal
Orthographic Skills
Reversal



Visual Processing Skills
Visual Discrimination
Visual Memory
Visual Closure
Reading Fluency
Reading Fluency I
Reading Fluency II
X2 (45) =78.160
NFI= .954
NNFI= .970
CFI= .980
SRMR= .050
RMSEA= .057
90% Confidence Interval of
RMSEA
: (.033- ,079)


Based on all the goodness of fit indices, this measurement model fit the data well.

The results of the confirmatory factor analysis are presented in Table 4-10. The italicized

numbers represent the t-value. If the t-value is greater than 1.96 with .05 level, it means

that the observable variable is statistically significant for this model. All of the t-values


on latent variables are greater than 1.96 in Table 4-10.









Table 4-10.

Elision


Blending


Digit RN


Letter RN


DEM-
vertical


DEM-
horizontal

Reversal


VD


VM


VC


Results of Confirmatory Factor Analysis
PA RN/EV Reversal
0.777
(0.139)
5.559*
0.476
(1.101)
4.710*
0.955
(0.052)
18.502*
0.888
(0.054)
16.317*
0.925
(0.053)
17.470*
0.722
(0.060)
11.946*
0.933
(0.052)
17.787*


0.674
(0.068)
9.879*
0.732
(0.067)
10.929*
0.777
(0.066)
11. 733*


Fluency I


Fluency II


In addition to t-value, squared multiple correlations (R2) are displayed in Table 4-

10. Since most of them are quite high, I can conclude that reliability of observable


measures is fairly satisfactory.


R2
0.604


0.227


0.913


0.789


0.856


0.521


0.870


0.454


0.537


0.604


0.968


0.913


0.984
(0.051)
19.321*
0.956
(0.052)
18.332*









Structural Model

The next step in SEM is construction of a structural model that fits the data as well

as the final measurement model with fewer estimated parameters. The structural model is

that component of the general model that prescribes relations between latent variables

and observed variables that are not indicators of latent variables. In general, the multiple

regression model is a structural model without latent variables and limited to a single

outcome. However, SEM that includes the combinations of the measurements and

structural components allows a comprehensive statistical model that can be used to

evaluate relations among variables that are free of measurement error (Hoyle, 1995). The

chi-square statistic included in Table 4-11 provides a test of the null hypothesis that the

reproduced variance matrix has the specified model structure (i.e., that the model fit the

data). First, the chi-square test for the model was statistically significant, X2 (66)=

1708.237, p< .05, indicating a good fit between the model and data. However, several

researchers (Bollen, 1989; Bollen & Long, 1993; Joreskog & Sorbom, 1996) noted that

the chi-square statistic is influenced by a large sample size (for this data set N= 210). In

fact, Joreskog and Sorgom (1996) proposed that X2 be used as badness rather than a

goodness-of-fit measure in the sense that a small X2 value (relative to its degrees of

freedom) is indicative of good fit, whereas a large X2 value reflects bad fit in SEM

(Byrne, 1998). Therefore, four additional goodness of fit indices were considered for

testing a goodness of fit of SEM model. The criteria for four other indices are listed: CFI

shows a good fit at .95 or higher; RMSEA should be between .08 or lower to indicate

good model fit (Hu & Bentler, 1999); NFI and NNFI should be close to 1.0 and show a

good fit at .90 or higher. These indices except RMSEA range from 0.00 to 1.00, with









values close to 1.00 being indicative of good fit. According to Hu & Bentler (1999), the

Root Mean Square Error of Approximation (RMSEA) has only recently been recognized

as one of the most informative criteria in covariance structure modeling, and it takes into

account the error of approximation in the population and asks the question, "How well

would the model fit the population covariance matrix if it were available?" (Browne &

Cudeck, 1993, pp. 137-138). This discrepancy measured by the RMSEA indicated that

values of less than .05 indicate good fit; values as high as .08 represent reasonable errors

of approximation in the population; values from .08 to .10 indicate mediocre fit; and

those greater than .10 indicate poor fit (MacCallum, Browne, & Sugawara, 1996). The fit

of the original, hypothesized model (Figure 3-1) was adequate because the chi-square

was statistically significant (X2 (45) = 78.160), and the fit indices showed that the model

has a moderate fit (NFI= .954, NNFI= .970 CFI= .980, SRMR= .050, RMSEA= .057).

Table 4-11. Results of Goodness of Fit
Goodness of Fit Indices SEM Model
X2 78.160
df 45
NFI .954
NNFI .970
CFI .980
SRMR .050
RMSEA .057
90% Confidence Interval of RMSEA (.033- ,079)


This SEM model, therefore, can be accepted due to a good model fit. As a

consequence, these results on goodness of fit can examine the following hypothesis:

H1 Phonological awareness, rapid-naming, and visual skills are not significantly
associated with reading fluency.









Based on the several results of indices on this SEM model, phonological

awareness, rapid-naming, and visual skills are significantly associated with reading

fluency. Therefore, Hypothesis 1 was rejected.

Regression Analysis and Correlational Analysis with SEM

Multiple regression analysis

A multiple regression analysis was conducted by using SEM to examine

Hypotheses 2 through 5. The regression coefficientsof four independent constructs and

one dependent construct is displayed in Figure 4-1. Table 4-12 represents regressions as

well as t-value. The italicized numbers represent t-value. If the t-value is greater than

1.96 with .05 level, it means that the regression between one dependent variable and one

independent variable is statistically significant. All of the t-values on this regression

matrix are greater than 1.96. Therefore, the following hypotheses were able to examined

by this statistical approach:

H2 Phonological awareness measures are not significantly associated with reading
fluency.

H3 Rapid-naming measures are not significantly associated with reading fluency.

H4 Eye-movement/Orthographic measures are not significantly associated with
reading fluency.

H5 Visual perception measures are not significantly associated with reading fluency.

Hypotheses 2 and 5 were rejected since the correlations are statistically significant.

However, due to the adjustment of constructs, the original Hypothesis 3 and 4 could not

be examined. Instead, rapid-naming (RN) measures and eye-movement (EM) measures

were combined into the modified construct and examined. The results indicate that the

RN/EM latent variable is significantly associated with reading fluency. In addition, the







orthographic measure with a single indicator, reversal, is significantly related to reading
fluency.


7--=
~7(
A .48 f

V- .96:
4-E :1


w 9


A- ----


Figure 4-1. SEM Reading Fluency Model with Correlation Coefficients


A -=3 \