1 VISUAL S IN THE MIND : INVESTIGATING GRAPH COMPREHENSION IN STUDENTS WITH DYSLEXIA By SUNJ UNG KIM A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2012
2 2012 Sunjung Kim
3 To my grandmother
4 ACKNOWLEDGMENTS First, I am ever grateful for God for His great provision His grace, and His mercy. I will Corinthians 12:9) helped me to stand up again and again. I wish to thank the students who volunteered to participate in t his stud y. Forty students with or without dyslexia were willing to spend their 3 hours to complete all tasks in spite of their busy schedule I would like to give a heartfelt, special thanks to my committee members, Dr. Altmann and Dr. Franks. Dr. Altmann is alwa ys providing numerous hours of advice and critiques not just for this dissertation but for my entire academic career. I am also humbly grateful to Dr. Franks for her constant support and encouragement. I extend my deepest thanks to my chairs, Dr. Lombardi no and Dr. Cowles. 6 years ago, I dreamt of pursuing a Ph.D in literacy and Dr. Lombardino is the very person who made my dream come t rue. Since then, she never stopped providing me with her excellent guidance, caring, and patience. Dr. Cowles intr oduced me to a n unfamiliar instrument, the eye tracker and, very patiently, trained me how to operate the instrument. Working in her lab led me to my current research goals. I am deeply appreciative of their support and input and personal cheering. Heartfelt gr atitude is also extended to my dear friends Camelia Raghinaru, Laurie Gauger, and Mario Mighty. Even though this was an unfamiliar topic to them, t hey read this dissertation several times a nd gave me valuable feedback. Without their help, it was not possib le to complete this project. My very special thanks to the persons whom I owe everything I am today, my father, Hyoungho Kim, and my mother, Jungnim Keum. Their love, support, and belief in me were a
5 treasure. My brother, Hanjun Kim called me every day to check how I was doing and listened to my talk very patiently. Finally, I wish to thank all my brothers and sisters Graduate Christian Fellowship for their constant support and en couragement. These acknowledgements woul d not be complete if I did not mention my forever roommates, Keshia, Harry, Jackie, and Alissa. They sacrificed their space and time for me and prayed for my work all the times.
6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ............... 4 LIST OF TABLES ................................ ................................ ................................ ........................... 8 LIST OF FIGURES ................................ ................................ ................................ ......................... 9 ABSTRACT 10 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .................. 12 2 LITERATURE REVIEW ................................ ................................ ................................ ....... 16 Definition of Graph Comprehension ................................ ................................ ...................... 16 What Affects Graph Comprehension? ................................ ................................ .................... 17 Pictorial Properties of a Graph ................................ ................................ ........................ 17 A Given Task ................................ ................................ ................................ ................... 18 ................................ ................................ ................................ ... 18 Language processing ability ................................ ................................ ..................... 19 Other cognitive processing abilities ................................ ................................ ......... 19 Graph familiarity ................................ ................................ ................................ ...... 22 Dyslexia and Graph Comprehension ................................ ................................ ...................... 23 Definition of Dyslexia ................................ ................................ ................................ ..... 23 How Individuals with Dyslexia Process Graphs ................................ ............................. 23 How to Me asure Graph Comprehension ................................ ................................ ................ 24 Reading Accuracy ................................ ................................ ................................ ........... 24 Eye Tracking ................................ ................................ ................................ ................... 25 Rational and Significance of the Study ................................ ................................ .................. 28 Study Objective s ................................ ................................ ................................ ..................... 33 3 METHODS ................................ ................................ ................................ ............................. 34 Introduction ................................ ................................ ................................ ............................. 34 Methods ................................ ................................ ................................ ................................ .. 34 Setting and Participants ................................ ................................ ................................ ... 34 Participants ................................ ................................ ................................ ...................... 35 Parti cipants with developmental dyslexia ................................ ................................ 36 Participants with typical reading skills ................................ ................................ ..... 36 Data Collection ................................ ................................ ................................ ....................... 36 Procedure ................................ ................................ ................................ ......................... 36 Instrumentation ................................ ................................ ................................ ................ 39 Verbal ability assessment ................................ ................................ ......................... 39 Phonological awareness assessment ................................ ................................ ......... 39
7 Phonological memory testing ................................ ................................ ................... 40 Rapid automatized naming testing ................................ ................................ ........... 40 Word reading fluency assessment ................................ ................................ ............ 40 Vocabulary assessment ................................ ................................ ............................ 41 Working memory assessment ................................ ................................ ................... 41 Executive function assessment ................................ ................................ ................. 41 Descriptive statistics for linguistic and cognition measures ................................ .... 42 Experimental graph comprehension assessment ................................ ...................... 43 4 RESULTS ................................ ................................ ................................ ............................... 48 Research Question 1: Comprehension Accuracy ................................ ................................ ... 48 Research Question 2: Eye Tracking Data Analysis ................................ ................................ 55 Research Question 3: Eye Tracking Data Analysis II ................................ ............................ 63 Research Question 4: Correlation among Graph Comprehension and Language and Cognition Measures ................................ ................................ ................................ ............ 70 Summary of Findings ................................ ................................ ................................ ............. 72 5 DISCUSSION ................................ ................................ ................................ ......................... 74 Group Comparisons on Graph Interpretation Accuracy ................................ ......................... 74 Group Comparison on Graph Viewing Times ................................ ................................ ........ 76 Findings from Language and Cognition Measures ................................ ................................ 78 Theoretical Implications ................................ ................................ ................................ ......... 83 Clinical Implications ................................ ................................ ................................ ............... 84 Limitations and Future Directions ................................ ................................ .......................... 86 A P PENDIX A RECRUITMENT FLYER ................................ ................................ ................................ ...... 88 B INFORMED CONSENT LETTER FOR PARTICIPANTS ................................ .................. 89 C QUESTIONAIRE FORM ................................ ................................ ................................ ....... 91 D SAMPLE OF SINGLE GRAPHIC DISPLAY GRAPH AND QUESTION .......................... 93 E SAMPLE OF DOUBLE GRAPHIC DISPLAY GRAPH AND QUESTION ........................ 94 L IST OF REFERENCES ................................ ................................ ................................ ............... 95 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ....... 110
8 LIST OF TABLES Table page 3 1 Test measurements used for graph comprehension, language, and cognitive skills .......... 38 3 2 Mean standard scores on the diagnostic reading tests for students with dyslexia ............. 45 3 3 Mean standard scores on the diagnostic reading tests for students with normal reading skills ................................ ................................ ................................ ...................... 46 3 4 Descriptive statistics for graphic familiarity, linguistic and cognition measures .............. 47 4 1 Descriptiv e statistics for DR and TR groups on graph accuracy ................................ ....... 51 4 2 Summary of mixed four way ANOVA on comprehension accuracy measures ................ 52 4 3 Descriptive statistics for DR and TR groups on overall viewing times ............................. 59 4 4 Summary of mixed four way ANOVA on overall viewing time measures ....................... 60 4 5 Descriptive statistics for DR and TR groups reading times on each graphic region ......... 67 4 6 Summary of between multivariate analysis of variance on interest areas of first pass times ................................ ................................ ................................ ................................ ... 68 4 7 Summary of between multivariate analysis of variance on interest areas of total viewing times ................................ ................................ ................................ ..................... 69 4 8 Students with dyslexia: correlations between comprehension accuracy, viewing times and cognition measures ................................ ................................ ................................ ...... 71 4 9 Typical readers: correlations between comprehension accuracy, viewing times and cognitio n measures ................................ ................................ ................................ ............. 71
9 LIST OF FIGURES Figure page 4 1 Two way interaction between question type and group ................................ .................... 53 4 2 Two way interaction between graphic display and graph type ................................ ......... 53 4 3 Three way interaction between graph type, question type, and group .............................. 54 4 4 Two way interaction between question type and group ................................ .................... 61 4 5 Two way interaction between graphic display and group ................................ ................. 61 4 6 Two way interaction between question type and graphic display ................................ ..... 62 4 7 Three way interaction between graphic display, question type, and graph type ............... 62 4 8 Six regions of the display: pattern, X Axis, Y Axis, legend, question, and answer .......... 6 4
10 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 VISUALS IN THE : INVESTIGATING GRAPH COMPREHENSION IN STUDENTS WITH DYSLEXIA By Sunjung Kim August 2012 Chair: Linda J. Lombardino Co chair: Heidi W. Cowles Major: Speech, Language, and Hearing Sciences Visual displays are commonly used to convey information. Graphic depictions of information are frequently recommended for individuals who have dyslexia because visual displays are thought to make information easier to understand. However, there is a dearth of research regarding the way individuals with dyslexia use visual representations or interpret information presented in the visuals. L ittle is known especially, about the effect of graphic factors (e.g., type of graph, type of information to be interpreted from the graph) and subject information shown in graphs. This project is designed to make progress toward ameliorating this problem. The primary goal of this project was to determine how individuals with dyslexia interpret graphic representations. The secondary goal was to investigate the extent to which cognitive abilities influence comprehension of graphically presented materials. College students with dyslexia were compared with age matched typical readers on graph comprehension tasks varying in graph type s, graph complexity, and graph interpretation questions. Students w ith dyslexia were significantly less accurate and took more time than age matched controls on graph
11 comprehension tasks and this difference was more robust for the more complex graphs and questions. Also, there was a high correlation between working memor y and graph comprehension for the students with dyslexia b ut not for the typical readers. These results underscore the necessity of conducting research on alternative and augmentative strategies for learning that are typically recommended for individuals w ho have reading disabilities and support the need for explicit instruction in graph interpretation for students with dyslexia.
12 CHAPTER 1 INTRODUCTION media society, people are incre asingly exposed to information that requires visu al processing, such as graphs, tables, charts, and diagrams in both academic and non academic settings. Visual displays are used, in large part, to make quantitati ve and qualitative information easy to under stand (Tversky & Schiano, 1989) Thus, graphs have been wid of data and examining trends in how data change over time (e.g., Guri Rozenblit, 1989; Levis & Lentz, 1982; Palmar, 1978) E ducators frequently encourage students with learning difficulties to use visual aids ( e.g., a graphic organizer) to assist them with text comprehension (Murphy, 20 05) There is, however, a dearth of research regarding whether the visual displays really facilitate learning. In fact, some researchers have reported that people do not obtain the expected advantages with visual representations (e.g., Mayer, 1993; Shah, Hegarty, & Mayer, 1999) Developmental dyslexia is the most common language based learning dis ability (Lyon, Shaywitz, & Shaywitz, 2003) The m ajority of research on dyslexia has focused on examining information processing presented in text formats (Shaywitz & Shaywitz, 2005; Vellutino, Fletc her, Snowling, & Scanlon, 2004 for reviews) However, recent research on the influence of specific cognitive processes on the performance of individuals with dyslexia on tasks presented in different modalities (e.g., auditory, visual), has encouraged re searchers to explore a wider range of processing skills in this population. In fact, several studies have identified cognitive functioning impair ment in persons with dyslexia in the areas of working memory and processing speed (e.g., Ackerman & Dykman, 1993; Berninger, et al., 2006; Cain, Oakhill, & Bryan t, 2004; Cohen, Netley, & Clarke, 1984; Jorm, 1983; Kail, Hall, & Caskey, 1999; Katzir, et al., 2006) Also, studies have shown that individuals with dyslexia have problems with both
13 maint enance /storage (e.g., Brunswick, McCrory, Price, Frith, & Frith, 1999; McDougall & Donohoe, 2002; Pennington, Cardoso Martins, Green, & Lefly, 2001) and manipulation requiring processing (e.g., Bronsnan, et al., 2002; Smith Spark, Fisk, Fawcett, & Nicolson, 2003; Swanson, 1999) of information. Findings from these numerous studies of individuals with dyslexia suggest that specific information processing abilities may be impaired across modalities (e.g., visual and/or auditory modalities) Hence, it is reasonable to hypot hesize that similar processing difficulties would be evident in their interpretation of graphically displayed information. Researchers have studied how we perceive meaning from graphs (e.g., Carpenter & Shah, 1998; Hegarty, Meyer, Narayanan, Freedman, & Shah, 2002; Pinker & Freedle, 1990) and c ollectively, suggest that processing a graph is a function of complex interaction s among the pictorial properties of the graph, task, and graph reader s ability (Peebles & Cheng, 2003) First, t he pictorial properties of graph s, including graph type and graphic complexity, have captured the attention of researchers. For example, there are different perspectives on the pictorial feature of graphs. Some researchers support the perceptual feature v iew which suggests that the representation of pictorial contents can be different depending on graph types such that each type of graph activates different representations (Lohse, 1993) Other researchers support the invariant structure v i ew which suggests that graphs share common characteristics (Peebles & Cheng, 2003) In this approach, graphs that shares features activate similar representation. Graphic complexity is another graphical property which plays a main role in graph processing. Carpenter and Shah (1998) used line graphs to examine the influence of graphic complexity on graph performance of college students. When the investigators added lin es to depict data, they expected that their participants would need increased processing time to
14 consider the third factor, z variable to interpret the relationship between x axis and y axis. College students were asked to interpret 12 pairs of line graph interpretation s were influenced by the number of lines (i.e., graphic complexity). Secondly, r esearchers have used two types of questions to examine individuals proficiency in understanding graphs (Curcio, 1982; Friel, Curcio, & Bright, 2001) The first type of question is referred to as a point locating question W ith this question, participants are asked to read a question to determine what a single point on the graph represents (e.g. How m uch tin was produced in 1980 ?) The second type is called a comparison question. I n this format, participants ar e required to compare two data points (e.g., W as more zinc or mor e cooper produced in 1982 ?). establish representations of data presented in the graph. Papafragou, Carruthers, Laurence, & Stich (2008) proposed the Salient Hypothes i s in which t he key linguistic features such as labels and keys in graphs bec ome privileged and attract viewers attention (Salience hypothesis). As noted by Gentner & Goldin Meadow (2003, p. 9) l anguage offers a l ens on non linguistic because attention to features that are relevant to linguistic encoding Additionally, c ognitive processes are central to graph comprehension. As o ur working memory temporarily stores only a limited amount of information (Baddeley & Hitch, 1974) multiple resources of graphs can potentially be a burden to readers (Hegarty & Steinhoff, 1997; Shah & Freedman, 2003) by taxing their memory resources and requiring them to split thei r attention (van Bruggen, Kirschner, & Joche ms, 2002) Furthermore, information must not on ly be retained but also regulated and controlled during processing The executive functions are
15 responsible for attentional control including regulating, controlling, monitoring, and suppressing information (Miyake, Friedman, Rettinger, Shah, & Hegarty, 2001) Because there are multiple resources in a graph, readers should inhibit irrelevant information, activate relevant information and update newly encod ed information (Berninger, Raskind, Richards, Abbott, & Stock, 2008) Thus, it is important to investigate cognitive factors in an attempt to study important processes involved in graph comprehension. To date, no scientific evidence has addre ssed ms to: (1) compare the accuracy of graph comprehension in young adults with and without dyslexia using comprehension tasks ; (2) c ompare overall reading time in young adults with and without dyslexia using eye tracking measurements; (3) compare eye gaze data for 6 specific graphic subregions ( i.e., pattern, x axis, y axis, legend, question, answer) in young adults with and without dyslexia; and (4 ) test which factors among graphical properties and complexity, question types, and subject factors, including working memory and executive functions affect their performance.
16 CHAPTER 2 LITERATURE REVIEW This chapter reviews the pertinent studies that have informed the research on graph comprehension of young adults and presents the specific experimental questions addressed Following a brief overview of various factors affecting graph processing, I discuss how dyslexia may affect graph comprehension. Next, I review methodology to measure graph performance of typical readers and present the significance for this study. Finally, the chapter concludes with an outline of the experimental research questions. Definition o f Graph C omprehension The Organization for Economic Co operation and Development (OECD; 1995) defined of literac y, document literacy refers to knowledge and skills required to comprehend information in various formats including tables and graphics (Murray, Kirsch, & Jenkins, 1997) Graphical literacy, also called graphicacy is the ability to process and comprehend data presented in graph format. According to the OECD (1995) the grap h is a core concept in literacy and needed for understanding social science s natural sciences, mathematics, etc (Kramarski, 2004) Simkin and Hastie (1987; 1988) suggested four main perceptual processes by which people process graphs: anchoring, scanning, projecting, and superimposition. In anchoring, people select a portion of the graph as a baseline (e.g., 50% of a bar, midpoint) In scanning, people move their eyes from the anchor to the other side (e.g., from the midpoint to its edge) In projecting, people draw an image b ased on the data they obtained through anchoring and scanning. Finally, in superimposing, people mentally move elements (e.g., size, angle) o f the graph to create overlap with another component of the image.
17 What Affects Graph C omprehension? Researchers have studied how we perceive meaning from graphs (e.g., Carpenter & Shah, 1998; Hegarty, et al., 2002; Pinker & Freedle, 1990) and suggest that processing a graph is a function of a complex interaction among the pictorial properties of the graph, the given task, and the graph viewer s ability to process graph (Peebles & Cheng, 2003) A review of each of these factors affecting graph processing follows. Pictorial P roperties of a G raph Fry (1983) the position of points, line s or area s on a two dimensional surface (p.5) Upon initial cons ideration, graphs might seem be the simplest ways to co nvey information visually, yet they are actually quite complex in terms of their structure, employing several structural components (Kosslyn, 1989; Parmar & Signer, 2005) Friel et al. (2001) summarized four structural components of graphs. The first component is a frame work (e.g., axes). The frame work is used to represent the kinds of measurements and the data being measured. The second component encompasses the visual dimensions, called specifiers These specifiers, such as lines on line graph s or bars on bar graph s, provide the data values of graph s The third component is labels such as titles and legends used to identify the type measurements being made. The last component is a background which includes the us e of colorin g or grid formats. A well designed graph uses these features to clearly identify the interpretation of information. As mentioned in chapter 1, there are two prom inent paradigms for discussing how features of graphs are processed. The Perceptual Feature v iew suggests that the representation of pictorial contents can be different depending on the graph types and that each type of graph activates different representa tions (e.g., Lohse, 1993; Ratwani & Trapton, 2008) For example,
18 line graphs and bar graphs have different perceptual characteristics and they, consequently, activate a different schema. On the other hand, the Invariant Structure v iew suggests that graphs share some common characteristics (e.g, Peebles & Cheng, 2003) such as the x and y axes (i.e., Cartesian coordinate system ) on line and bar graphs and that th ese similar frameworks share same underlying mental representation s A Given T ask Graph viewers extract information from a graph to complete specific task(s) under various conditions. Questioning is a fundamental cognitive mechanism used to assess graph comprehension (Friel, et al., 2001) A s noted previously, r esearchers have used two types of questions to examine individuals proficiency in understanding graphs : point locating and comparison (Curcio, 1982; Friel, et al., 2001) The point locating question is characterized by extracting data from a graph and the comparison question focuses on interpolating and finding relationships of data from the graph (Friel, et al., 2001) The level of complexity of the two question types appear s to differ. The information asked in the comparison question s is not explicitly represented in the graph requiring more cognitive processing time/effort than the point locating questions (Duesbery, Werblow, & Yovanoff, 2011) Participants show more difficulty on comparison tasks compared to point locating tasks (e.g., Friel, et al., 2001; Monteiro & Ainley, 2004; Wainer, 1992) providing evidence that per formance on graph comprehension can differ due to the question types used to elicit an interpretation from a graphic display G bility The third factor a ffecting graph comprehension is the internal representations of data presented in a graph. Rese archers have acknowledged that subject differences may have as much of an influence on comprehension processing as do the
19 pro perties of the graph itself (Friel, et al., 2001) abilities, few studies have investigated the interpretatio n. However, researchers have suggested that subject factors associated with language and other cognitive abilities that may play an important role in visual comprehension (e.g., Berg & Phillips, 1994; Boden & Brodeur, 1999; Carpenter & Shah, 1998) Language processing ability The role of language ability in graph compre hension has not bee n widely appreciated. However, given that graph interpretation is typically accompanied by written text, both linguistic decoding and encoding abilities are relevant to visual processing (Gentner & Goldin Meadow, 2 003) Only very recently have researchers begun to focus on domains of language, such as relationships of vocabulary and visual comprehension. For example, Yang (2012) inves taged performance on graph interpretation was closely related to their lexical knowledge. A s mentioned earlier, very little is know about the role of language i n graph processing. Therefore, it is important to understand the role of language in graph comprehension. Other cognitive processing abilities Graph reading is a cognitive activity that requires information processing from several sources (e.g., axes, lab els, values). Thus, it is important to understand the contribution of cognitive process ing in visual interpretations (Huang, Hong, & Eades, 2006) Working memory serves the function of storing and manipulating verbal and vi suospatial information (Baddeley & Hitch, 1974) and supports higher abilities such as comprehension and reasoning (e.g., Engle, Tuholski, Laughlin, & Conway, 1999; Shelton, Elliott, Matthews, Hill, & Gouvier, 2010) Several studies have attempted to explain the nature of working memory and the potential impact of working memory on cognitive processing (Duff & Logie, 2001) For example, the resource
20 sharing model posits that working memory relies on one domain general system (Cowan, et al., 2005) In this model, a single cognitive resource is used for both storage and processing (Just, Carpenter, & Keller, 1996) On the other hand, the multiple resource model posits that working memory taps several domain specific and domain general processes (Miyake, et al., 2001) In this multi ple source framework, there is more than one subsystem acting concurrently to perform a specific task (Duff & Logie, 2001) Although these two models view the nature and function of working memory from different perspectives, both models distinguish between verbal and visual working memory systems. Considerin g the distinguishable characteristics of visual and verbal working memory and the close relationship between working memory resources and comprehension, it is necessary to investigate the features of working memory in both visual and verbal modalities (Tanabe & Osaka, 2009) Working memory capacity is commonly measured in an item span format quences of words, digits, and objects (Waters & Caplan, 1996) In addition to working memory capacity, executive function skil ls, such as regulating, controlling, monitoring, and suppressing information (Miyake, et al., 2001) may influence an legends, graphic patterns) in a graph, it is important to measure executive function s that subserve activating relevant information, inhibiting irrelevant information, and shifting between information (Martin & Allen, 2008) Executive function s are closely rela ted to various cognitive abilities Sesma, Mahone, Levine, Eason, and Cutting (2009) examined the contrib ution of executive functions in the reading comprehension of sixty children, ag ed 9 15 years. After controlling for attention, decoding skills, fluency, and vocabulary executive functions were still an important factor in understanding written text. Busch, et al. (2005) investigated the
21 relationship between executive functions and visual memory in patients with a history of traumatic brain injury. In a one year follow up test, executive functions still played a key role in visual memory. Agostino, Johnson, and Pascual Leone (2010) examined if executive functions were associated with word problem solving skills in 155 children in Grades 3 6. T he in executive functions, especially in the case of multi step problems. Research on executive functions has lead to better understa nding of the development of various cognitive abilities (Miyake, Friedman, Emerson, Witzki, & Howerter, 2000) Because executive function is believed to consist of diverse functions, different sets of tasks have been administered to measure them (McCabe, Roediger, McDaniel, Balota, & Hambrick, 2010) For example, Ashendorf and McCaffrey (2008) used the Wiscon sin C ard Sorting t asks (Grant, Jones, & Tallantis, 1949) to measure mental flexibility in 25 younger adults aged 19 22 and 19 older adults, aged 63 89. Participants were asked to match stimulus cards by color, design, or quantity and verba lly state their reason for each match In another study conducted by Mah one, Koth, Cutt ing, Singer, and Denckla (2008) verbal fluency tasks were used to measure organizing and inhibitory functions in 46 children with either Tourette syndrome or Attention Deficit / H yperactivity The children were required to name as many examples of animals and food s (semantic word fluency) within a given period of time. McCabe, et al. (2010) used mental control tasks to evaluate task maintenance abilities in 200 participants between 18 and 90 years of age Mental control requires subjects to say well known categories of information (e.g., the day of the week; months of the year) in forward and backward orders (Wechsler, 1997a)
22 Graph familiarity Gr is potentially influenced by their familiarity and experience with graphs (Shah & Hoeffner, 2002) Knowledge about graphs is referred to as having graph schemata (Pinker & Freedle, 1990) or graph sense (Fr iel, et al., 2001) Xi (2010) investigated whether graph familiarity of English language learners in US graduate school s influenced their performance on line graph ta sks. It was found that participants who were more familiar with graphs obtained reliably higher content and organization scores in the line graph tasks than those with less familiar ity with graphs. The study found that familiarity with graphs in addition to familiarity with the contents d epicted in the graphs influenced knowledge in this graph study echoes the literature regarding the role of prior knowledge in learning (e.g., comparison of novice and expert). Langrall, Nisbet, Mooney, and Jansem (2 011) selected middle school students from Australia, United States, and Thailand who had expert knowledge of a par ticular topic and those who had no great interest or knowledge in the topic and compared their performance on mathematical data analysis ta sks. They found that students with context expertise were more likely to identify useful data for the task, and analyze and interpret the data compared to their peers. Cursio (1982) cautioned that the content in graphs and the vocabulary words used as specifiers may be factors that affect t comprehend relationships expressed in the graph. Therefore, it is important to ensure that graph r graph processing. Especially for people who have learning dis abilities, knowledge on graphs and content can make a significant difference on their performance on graph tasks.
23 D yslexia and Graph Comprehension De finition o f D yslexia The m ost current definition of dyslexia adopted by the International Dyslexia Associa tion neurological in origin. It is characterized by difficulties with accurate and/or fluent word recognition and by poor spelling and decoding abilities (Lyon, et al., 2003) Over the last two decades, the research literature on developmental dyslexia has focused on examining the word level phonological processing deficits that are believed to lie at the core of th is specific reading disability More recently specific cognitive deficits that are less directly associated with language such as executive fun ctions (Bronsnan, et al., 2002) and working memory (Beneventi, Tnnessen, Ersland, & Hugdahl, 2010) deficits have been observed in the behavioral profiles of many individuals diagnosed with dyslexia. How Individuals w ith Dyslexia Process G raphs U nfortunately, little is known about the graph comprehension abilities of individuals with dyslexia. To date, only one study by Parmar and Signer (2005) investigated line graph interpretation in children with learning disabilities. Seventeen fourth grade and 25 fifth grade students with learning disabilities and 23 fourth grade and 26 fifth grade students without learning disabilitie s were asked to complete four line graph tasks that include interpreting and constructing graphs. The students without learning disabilities significantly outperformed their peers with learning disabilities on all tasks. The authors emphasized the need to establish how individuals with language learning difficulties process graphically presented information so that academic supports can be used, if necessary, to facilitate their comprehension in this area. In spite of a lack of research in graph processing associated with dyslexia, it is likely that individuals with dyslexia will exhibit deficits in graph comprehension for two reasons; (1) A
24 large number of studies have shown cognitive function impair ments in persons with dyslexia in the areas of working me mory (e.g., Ackerman & Dykman, 1993; Berninger, et al., 2006; Cain, et al., 2004; Cohen, et al., 1984; Jorm, 1983; Kail, et al., 1999; Katzir, et al., 2006) ; and (2) Studies have shown that some individuals with dyslexia exhibit executive function difficulties with tasks that involve the maintenance/storage (Brunswick, et al., 1999; McDougall & Donohoe, 2002; Pennington, et al., 2001) and manipulation of information (Bronsnan, et al., 20 02; Smith Spark, et al., 2003; Swanson, 1999) It is quite plausible, given the results of previous studies, that g raphs may tax the memory capacity of individuals with dyslexia because they need to hold and integrate many elements (e.g., axes, labels, values, etc.) in memory in order to interpret graphs. Furthermore individuals with dyslexia have difficulty connecting verbal labels to visual images (e.g., Boden & Brodeur, 1999; Gang & Siegel, 2002) T his cross modal skill is used cont inuously when interpreting graphs (Ratwani, Trafton, & Boehm Davis, 2008) How t o Measure Graph C omprehension Reading Accuracy answers to specific question s Question a nswering is considered a fundamental element of cognition and plays a major role in comprehension (Friel, et al., 2001; Graesser, Swamer, Baggett, Sell, & Britton, 1996) Carcio (1987) compared the graph comprehension of 20 4 fourth grade and 185 seventh grad e students. The graph test consisted of three bar graphs, three line graphs, three circle graphs, and three pictographs. Six multiple choice questions followed each graph. Two questi ons were literal items (e.g., point loc ating), two questions were comparison items (i.e., requiring comparison of two data points), and two questions were extension items (i.e., predicting a trend). The seventh grade students reliably outperformed the fourth grade students on the graph test; ho wever, the differences in accuracy for the three question types were
25 not addressed. Kramarski (2004) used multiple choice question s to evaluate the effects of metacognitive instruct ion on the graph comprehension of 196 eighth grade students. In the metacognitive instruction s students were guided to generate questions themselves and develop a strategy to answer the questions. Students who were exposed to metacognitive instruction s pe rformed more accurately on the multiple choice questions. While multiple choice questions high reliability, it has been recommended t hat other techniques be use d in addition to the multiple choice format to increase the validity of assessing comprehension. (Levine, McGuire, & Nattress, 1970) Eye T racking We ceaselessly move our eyes to take in visual information ( Richardson, et al., 2007) Of most interest to researchers are saccades the rapid eye movements around the visual field, and fixation s the places where our eyes stay fixed between the saccades (Rayner, 1998) Researchers have shown t hat if the available visual environment is relevant to the task at hand, the location of eye fixations indicate s whether attention is being directed a phenomenon called the eye mind hypothesis (Just & Carpenter, 1980; van Gog & Scheiter, 2010) Additionally researcher s have reported that e ye movement s reflect underlying mechanism s that contribute to difficulty with performance (Rayner & Slattery, 2009) For example, fixation times are reliably longer and saccades are more frequent when reading irregular words (i.e., words which do not follow one to one correspondence between gr aphemes and phonemes ) (Jones, Kelly, & Corley, 2007) listening to ambiguous sentences (Spivey, Tanenhous, & Ed erhard, 2002) viewing incongruous objects (Henderson, Weeks, & Hollingworth, 1999) and detecting unfamiliar faces (e.g., Hirose & Hancock, 2007) Moreover, fixation times and frequency of saccades can distinguish be tween younger and older re aders (Eurich, 1933a) unsuc cessful and successful college students (Eurich,
26 1933b) physics experts and novices (Feil, 2010) beginning and experienced drivers (Underwood, Chapman, Brocklehurst, Underwood, & Crundall, 2003) and s econd language learners and native speakers (Keating, 2009) For examining individuals with dyslexia eye tracking is also a very effective tool (Rayner & Slattery, 2009) S ome eye care specialists even assume that dyslexia i s mainly due to oculomotor problems. They use simple eye movement test s to diagnose dyslexia and oculomotor training to treat it (e.g., Pavlidis, 1981; Solan, Feldman, & Tujak, 1995) However, Rayner (1998) emphasized that attempts to replicate the se studies have failed (e.g, Olson, Conneis, & Rack, 1991; Stanley, 1994) In his revie w paper, Rayner (1998) concluded that eye movement diff iculty is not a cause of dyslexia, but a symptom reflecting some underlying impaired mechanism s Most commonly, researchers today regard dyslexia as a language processing deficit and study eye movement as a reflection of a language processing disorder Individuals with dyslexia have been found to show more frequent and longer fixations and saccades on tasks of phonological awareness (Eden, Stein, Wood, & Wood, 1994) rapid automatized naming tasks (Jones, Obregn, Kelly, & Branigan, 2008) word reading (De Luca, Borrelli, Judica, Spin elli, & Zoccolotti, 2002) and text reading (Prado, Dubois, & Val dois, 2007) than their peers who do not have dyslexia. The rational for using eye tracking in this study is that it provide s a systematic method for examining how speed and accuracy influence the performance of individuals when interpreting graphs. P rev ious research ha s shown that individuals with dyslexia reliably take more time to respond to simple pure tone and lexical decision tasks than their nondyslexic peers although their responses are often accurate (Nicolson & Fawcett, 1994) Specifically, the tracking of eye movement allow s for (a) determining the amount of time participants spend looking at specific
27 areas (i.e., X axis, Y axis, and patterns) and (b) compar ing the eye movement data (i.e., saccades, fixation) of typical and impaired readers while examining graphs and answer ing written questions in response to the data shown in the graphs. Recently, eye tracking ha s been frequently used to understand graphics (e.g., photos, graph s, animations) (Hyn, 2010; Jarodzka, Scheiter, Gerjets, & van Gog, 2010) Holsanova, Holmberg, and Holmqvist (2009) investigated the influence of distance between text and graphics in eye movement behavior of thirty one typical readers to test text graphic integration. Participants were provided with a newspaper in two formats, a separated condition, in which text and graph ic box were far from each other, and an integrated condition, in which the graphic was close to the text where the re ferences needed to b e explained Participants more effectively integrated text and graphics when text and illustrations were physically close r. Schmidt Weigand, Kohnert, & Glowalla (2010) had 90 university student s read/listen to multimedia instruction accompanying visualizations in a system paced condition and another 31 students to read/listen to the instruction in a self paced condition. Following an instruction, the students were tested on the content in the co ntexts of retention, transfer of information, and visual memory. Students in the self paced condition spent more t ime on the important parts to complete tasks, compared to those in the system paced condition. Carroll, Young, and Guertin (1992) analyzed eye movements of participants when they were asked to examine drawings with captions. They found that participants briefly scanned the picture first, then looked at the text, and then rescanned the picture more carefully. These previous eye trac ki ng studies have shown that meas u r ing eye movements provides us with greater understand ing of how viewer s approach learning information that requires them
28 to integrate t extual and graphic information. Mayer (2010) suggested that eye tracking can contribute to the study of learning with graphics in both theoretical and practical ways. Specifically, eye fixation measures provide an online observation of attentional control, linking visual per ception and language. Also, eye fixation measures provide information on instructional design for online learning. However, eye tracking methods have their limitations, as they do not always give researchers accurate information on the success or failure o f learning (Hyn, 2010) The viewer may spend a long time look ing a t a given stimuli without adequ ate comprehension of its contents For example, De Koning, Tabbers, Rikers, and Paas (2010) provided 40 psychology undergraduate students with an animation of the human cardiovascular system with and without included the number of fixations and the proportion of fixed time in specific locations. The findings revealed that students looked at cued as c ompared to non cued content more frequently and for longer periods of time. However, the learning outcome measures, that included a comprehension test and verbal report, did not show cued content was learned more effectively. That is, students who looked at the cued animation did not necesarily comprehend the crucial relations hips between different component of the cardiovascular system at a higher rate of accuracy. The investigators surmised that students with low knowledge might have difficulty understanding th e content accuratel y and viewed the contents longer. They suggested that other factors that affect learning should be considered, too. Thus, the eye tracking data should be supplemented with other performance measures ( Co wles & Kim 2011; Hyn, 2010) Rational and Significance o f t he S tudy We are commonly exposed to visual representation of information. Children look at illustrations in storybooks when they are read to (Dockett, Whitton, & Perry, 2003) students are exposed to visual displays in textbook s (Cook, Wiebe, & Carter, 2008) and people read
29 newspapers and magazines containing graphs and diagrams (Underwood, Jebbett, & Roberts, 2004) In addition, there is a widespread belief that visual aids facilitate comprehension (e.g., Guri Rozenblit, 1989; Levis & L entz, 1982; Palmar, 1978) so educators frequently encour age the m to use visual aids ( e.g., graphic organizer s ) (Murphy, 2005) There is, however, a dearth of scientific evidence regardin g whether the visual displays actually facilitate or impede learning There are numerous types of visual displays of data such as graphs, charts, maps, and diagrams (Kosslyn, 1989) G raphs have well organized symbolic structure (Parmar & Signer, 2005) in which c ommon properties and configurations of lines, colors, regions, etc. are all integ rated to convey meaning (Habel & Acarturk, 2009) Thus, they can be a good start ing point for investigating h ow visual entit ies affect comprehension. Graphs serve to map perceptual features and co nceptua l messages; connections of these messages may be challenging for some people In fact, investigators have acknowledged this possibility. For example, Freedman and Shah (2002) noted that while graphs are ubiquitous because they are perceived by educators, journalists, and other authors as more effective and less demanding in deliver ing information than text formats, comprehension of graphs can be very effortful and lead to incorrect decision making Also, Keller (2009) addressed the likelihood that graphing enables students to develop abstract thinking, yet some students are unable to correctly interpret graphs and have trouble developing the graphing skill. Further, Shah and Hoeffner (2002) concluded that school aged children and even adults without learning difficulties commonly make an error in visual reasoning underlying external representational media like graph s (also see Gattis & Holyoak, 1996; Guthrie, Weber, & Kimmerly, 1993)
30 Considering that typical readers sometimes have diffi culty interpreting graphs, searching for main idea s in graphs can be particularly challenging for people who have learning disabilities or a lack of experience with visual representations This dimension of comprehension offers many opportunities for exploration. As mentioned before, o nly one study conduct ed by Parmar and Singer (2005) examined graph comprehension in students with learning disabilities and they found that they performed at a significantly lower level than their peers without learning disabilities suggesting that the lack of the graphic comprehension skills in children with reading disabilities may be contributing to their failure to com prehend the textual information that is augmented by graphs. The authors stressed the lack of lack of research on graph comprehension of people with learning disabilities This lack of data on the effects of graphs in the reading comprehension of individua ls who have reading disabilities is somewhat surprising given that graphic displays are frequently used to assist students with learning difficulties to more easily access text contexts (e.g., Dexter, Park, & Hughes, 2011; Kim, Vaughn, Wanzek, & Wei, 2004; Nesbit & Adesope, 2006) The importance of research on the exact nature of graph processing in individuals with learning disabilities is specifically emphasized for educational purpose s In the intermediate an d secondary grades, graphing has been frequently used to present complex materials and abstract concepts. As a result, lack of graphic comprehension skill has caused students to miss immerse learning of concepts (Parmar & Signer, 2005) In addition, most cognitive abilities on graph comprehension. To date, researchers studying graph comprehension have focused primarily on characteristics of graphic properties and questioning to determine the interpretation of graphed information. The nearly exclusive focus on graphical properties has limited our understanding of graph processin g, particularly, in the population of individuals who
31 have reading disabilities and oft en rely on contextual cues for comprehension. Furthermore, it is particularly important that subject factors such as working memory and executive functions, often assoc iated with reading disabilities, thus the factors need be explored for their potential role in graph reasoning (Huang, et al., 2006) Furthermore most studies examining learning disabilities have exclusively focused on off line assessment such as comprehension measurement s On line information such as eye tracking measurements are beginning to be use d more frequently to supplement off line information by providing moment by moment processing information, especially with people who are challenged in the skill of reading (Whitney & Cornelissen, 2007). This study was designed to address whether young adults with dyslexia are impaired in graph comprehension in comparison with their peers without dyslexia by investigating both comprehension accuracy and eye tracking data. I n the current study, the following experimental questions were addressed: RQ1. H ow does the performance of college stu dents who have dyslexia compare with their non dyslexic peers for comprehension accuracy on graph interpretation tasks ? a. It was expected that students with dyslexia would perform more poorly than their peers and that difference would be most prominent on items with the highest level of the graph complexity and the most difficult types of co mprehension question. RQ2. How does the performance of college students who have dyslexia compared with their non dyslexic peers for response time on graph interpretation tasks? a. It was expected that the students with dyslexia would be slower than their peers yielding longer eye fixations especially on the more complex graphs and more difficult comprehension questions.
32 RQ3. If there would be a group effect on response time, where the effect would be driven by? a. It was expected that the effect would be due to particular parts of the display (e.g., text areas). RQ4 W hat is the nature of the relationships between the viewers graph comprehension and their cognitive abilities as measures on tasks of vocabulary, working m emory, and executive functions? a. It was hypothesized that vocabulary would not highly correlate with performance o n graph comprehension for both groups of participants because th level in the stimuli was controlled in terms of difficulty b. It was hypothesized that working memory and executive functions would correlate highly with n graph comprehension for both groups of participants. This study has theoretical and practical implications for dyslexia research. Firs t, it will allow us to explore the nature of visual comprehension in students with dyslexia which has been largely overlooked. Second, it may shed light on how we can support students with reading difficulties when graphs are involved in texts and other r eading materials In this study, we expect to see whether or not we should recommend graphs to dyslexics and whether or not they need explicit and additional instruction for interpreting this kind of information. Educators can use this information to plan what they should focus on for remediating comprehension problems. Thus, t his project will provide a foundation for research based instruction and intervention for comprehension of visual materials
33 Study Objectives The purpose of this study was to explore the effect of reading skill on graph comprehension using combined comprehension accuracy and eye tracking data. A review of the literature review motivated the following two study goals and predictions. The first goal of this study aimed to establish how students with dyslex ia perform on a graph comprehension task compared to typical readers To achieve this goal, I examine d the effect s of graph type ( i.e., line graphs, bar graphs, and horizontal bar graphs), graph complexity (i.e., sing le vs. double graph ic patterns: single line graphs vs. double line graphs, single bar graphs vs. double bar graphs, single horizontal bar graphs vs. double horizontal bar graphs), and question type ( i.e., questions asked to extract data from a single value vs. questions asked for compari ng two values) on their comprehension performance. I predicted that performance would be influenced by the type of graph and task and graph complexity Specifically, s tudents with dyslexia will perform less accurately than typic al readers on the more complex graphs and on the more complex questions for data interpretation Also, I e xpect ed that dyslexics w ould show longer eye fixations on graphs indicating difficulties in search and processing and this difference w ould become more robust with the increasing complexity of graphs and tasks The second goal of this study was to examine the extent to which linguistic and cognitive factors influence comprehension of graphs For this goal, I e xamine d one linguistic skill (i.e. receptive vocabulary) and two cognitive abilities ( i.e., working memory and executive functions ) that have been hypothesized to impact dyslexia and investigate d relationships between those factors and graph comprehension measured by accuracy and eye move ment data. I e xpect ed that both working memory and executive functions w ould have an impact on graph comprehension because of the verbal/ visu al characteristics of graph s
34 CHAPTER 3 METHODS Introduction The purpose of this study is to compare response accuracy and eye gaze pattern in young adults with and without dyslexia on different levels of complexity of graph types and graph interpretation questions. College students completed a graph co mprehension task while their eye movements were recorded Reading and reading related skills, working memory, and executive function skills were also assessed. In this chapter, the methods of the study are described. Methods Setting a nd P articipants Two g roups of college students aged 18 to 35 years of age were recruited with the approval of the University of Florida Behavioral/NonMedical Institutional Review Board (UFIRB #2011 U 0643): (1) dyslexic group: students who had been diagnosed previously with a specific reading disability or who reported lifelong difficulties with reading, writing, and/ or spelling but had not been formally diagnosed with reading disability; (2) typical reading group: students who had no history of neurological or sensory deficits or speech, language, reading, or general academic problems. Participants were required to be native English speakers. Recruitment flyers were distributed to University of Florida Disability Resource Center and Santa Fe College Disability Resource Center ( Appendix A ). Participants were also recruited through the UF LIN CSD Research Participant Pool website and Educational Research Participant System The flyers included brief descriptions of the study, benefits from participation, and contact information for potential participants. All participants were compensated in the form of 2 hours of research credit for applicable courses. Specifically, participants with reading disability were provided with a three page brief report of their test
35 findings when the y chose to be informed of their standardized test scores for documentation of their learning disability. Data were collected from August, 2011 through March, 2012. Participants A total of forty college students participated in the present study: 17 students with developmental dyslexia and 23 students with typical reading skills. Written informed consent was obtained from all participants prior to the onset of the research activity (Appendix B). All participants completed a graph familiarity questio nnaire about prior experience using graphs (Appendix C) Participants were asked to self estimate their graph familiarity in graphs on a 7 point scale from 1 (never or strongly disagree) to 7 (very often or strongly agree) To confirm the diagnosis of dysl exia and clarify the classification of two groups, spoken and written language tests were given to all participants. All tests were administered by the author. Participants were diagnosed with developmental dyslexia if they (1) showed deficits on standardi zed tests of phonological and/or orthographic processing that include phonological awareness, rapid naming, word decoding, word reading, and/or reading fluency unexpected for their other cognitive abilities, educational levels, and socio cultural opportuni ties; (2) reported having persistent difficulties and/or remarkable lack of progress in reading, spelling, and/or writing along with a positive family history for reading disabilities; (3) obtained relatively high scores on standardized test of comprehensi on despite poor word decoding, word recognition, and/or spelling scores; (4) obtained relatively high scores on standardized t e st of oral language; and (5) had no developmental history of diagnosis and therapy related spoken language impairment with the ex ception of minor articulation difficulties Data from 35 participants were used in the final analysis: 15 students with developmental dyslexia and 20 students with normal reading skills. The two group s did not differ significantly i n chronological age (F(1 ,33) = 2.44, p >.05 ).
36 Participants with developmental d yslexia Among seventeen participants with developmental dyslexia, two students were excluded from the study: (1) O ne did not meet the age requirement of 18 to 35 years, and (2) the other had difficulty with both spoken and written languages. A total of fifteen college students with developmental dyslexia were included in the final data analysis. A group with dyslexia was 67% female (Male = 5, Female = 10). T hey ranged in age from 18 to 35 years (Mean = 20.89, SD = 2.69). Table 3 2 shows mean standard scores on the diagnostic reading tests for students with dyslexia Participants with t ypical r ea ding sk ills Among twenty three participants who were initially placed into the typical reading control group, three students were excluded from the study for the following reasons : (1) there were technical problems with the eye tracking data of two students, and (2) one student scored below average on the norm referenced word reading test ( TOWRE Test) A total of twenty college students with typical reading skills were included in the final data analysis. A group with normal reading skills was 90% female (Male = 2, Female = 18). They ranged in age from 18 to 35 years (Mean = 19.71, SD = 1.44) Table 3 3 shows mean standard scores on the diagnostic reading tests for students with normal reading skills. Data Collection Procedure All testing was performed individually in the Language and C ognition R esearch laboratory in Turlington Hall at the University of Florida After t he author provided g eneral information about the study and the test procedures, the participant was asked to sign two copies of informed consent form s and then to complete a background/graph questionnaire. One copy of the consent form was given to the participant and the other copy was
37 The experimental protocols were administered by the author in approximately 2.5 3 hours. Each subject was completed all tests in the fixed order that follows : ( 1 ) general background / graph familiarity questionnaires and questionnaires of a graph familiarity ; ( 2) reading diagnostic session ; (3) an experimental session (a graph comprehension assessment ); 4 ) linguistic and cognitive assessments
38 Table 3 1. Test measur ements used for graph comprehension, language, and cognitive skills Skill Name of Test Subtest Verbal ability W oodcock Johnson III tests of cognitive abilities Verbal Comprehension Phonological processing Comprehensive Test of Phonological Processing Elision Blending Words Phonological Awareness Composite Memory for Digits Nonword Repetition Phonological Memory composite Rapid Digit Naming Rapid Letter Naming Rapid Naming Composite Word reading fluency Test of Word Reading Efficiency Sight Word Efficiency Phonemic Decoding Efficiency Total Word Reading Efficiency Vocabulary Shipley vocabulary test Working memory Verbal working memory Digit forward Digit backward Digit ordering Visual working memory Block recall Executive function Color stroop Color word stroop Trail Making A Trail Mak ing B D igit Symbol S ubstitution
39 Instrumentation A description of the seven psycho educational tests and the graph interpretation tasks used to test each participants follows. Table 3 1 shows all test measurements used in this study. Verbal ability assessment The Picture Vocabulary, Synonyms, Antonyms, and V erbal Analogies subtest s from the Woodcock Johnson III Test of Cognitive Abilities (WJ III COG; Wo odcock, McGrew, & Mather, 2002) were The Picture Vocabulary task required the participant to identify picture of objects. The Synonyms task required the participant to provide a synonym of a given word ( car required participants to create a n floor The Verbal Analogies task required coat is to wear as apple Two to t hree training items were administered prior to each subtest and testing was completed when a student missed three items in a row. Phonological awareness assessment The Elision and Blending Words subtest s from the C omprehensive Test of Phonological Processing (CTOPP; Wagner, Torgesen, & Rashotte, 1999) were administered to assess each phonological awareness skil ls. The Elision subtest required the participant to listen word spider without saying der ; say the word split without saying /p/). The Blending Words subtest required the participant to put separate sounds together to make a whole word (e.g., Testing was discontinued when the participant missed three items in a row. The score was recorded as th e total number of all items answered correctly. The Phonological awareness composite score derived from the Elision and Blending Words subtest scores.
40 Phonological memory testing Memory for digits and Nonword repetition subtests from the Comprehensive Test of Phonological Processing (CTOPP; Wagner, et al., 1999) was administered to measure each y. The Memory for digits subtest require d the participant to repeat a series of numbers ranging in length from two to eight digits. The Nonword repetition subtest required th e participant to repeat a made up word. T esting was discontinued when the particip ant missed three items in a row. The score was recorded as the total number of all items answered correctly. The Phonological memory composite score derived from the Memory for digits and Nonword repetition subtest scores. Rapid automatized n aming testing Rapid number naming and Rapid letter naming subtests from the Comprehensive Test of Phonological Processing (CTOPP; Wagner, et al., 1999 ) were administered to measure each The participant was given an 8 x 12 card showing the 6 items in 4 rows of 9 randomly repeated items and asked to name each stimulus item as quickly as possible without making any mi stakes. The total time taken to name the stimulus set was timed with a hand held digital stop watch The Rapid automatized naming composite score derived from the Rapid number naming and Rapid letter naming subtest scores Word reading f luency assessment The Sight Word Efficiency (SWE) and Phonemic Decoding Efficiency (PDE) subtests from the Test of Word Reading Efficiency (TOWRE; Torgesen, Wagner, & Ras hotte, 1999) were words fluently at the single word level. On the each of these subtests, practice items were presented and then the participant was given a series of real (SWE) or pseu do words (PDE) and asked to read aloud as many as
41 possible in 45 seconds. Inaccurate responses were deducted so that the final score reflected the total number of words correctly read within the given time frame. Vocabulary assessment The Shipley vocabulary test (Shipley, 1940) wer e administered. This multiple choice vocabulary test comprises 40 words, presented in ascending order of difficulty. The developer reported a split half reliability in which a test is randomly split into half and scored separately, then the score of one half is compared to the score of the other half (Field, 2009) is .87. Working m emory assessment To measure Verbal memory ability of the participant, Digit span for ward and D igit span backward from the Wechsler Memory Scale (Wechsler, 1987) and D igit ordering task (Hoppe, Muller, Werheid, Thone, & von Cramon, 2000) were used. In the D igit span forward measuring verbal short term memory the participant repeat ed verbatim increasingly long lists of numbers. In the D igit span backward measuring verbal working memory the participant repeat ed incr easingly long lists of numbers in the reverse order of presentation. For the D igit ord ering task, the participant was asked to repeat back in numerical order an increasingly long list of digits. v isual working memory ability, Block recall subtest from the Working Memory Test Battery for Children (WMTB C, Pickering & Gathercole, 2001) wa s used. In the Block recall subtest, participants saw a board with nine cubes, being tapped in random order by the author. The participant was then asked to reproduce the sequence in the correct order by tapping on a picture of the blocks. Executive function assessment C olor word Stroop (Golden, 1978) Trail s Making (the A and the B versions) (Reitan & Wolfson, 1993) and D igit Symbol S ubstitution (Wechsler, 1997b) test s were administered to fun ction skill In the Stroop test measuring inhibition the
42 participant was asked to name the lists of the color of ink and then name the color of the ink for a series of non congruent color words. The Trail Making test w as used to measure visual attention and switching and consists of two subtest. In the Trails A, the participant was presented with a sheet of paper showing 25 circles numbered 1 25 and asked to draw lines to connect the numbers in ascending order. In the Tr ails B, circles include d numbers and letters. The Participant was asked to connect the circles in the alternating number letter order of 1 A, 2 B, 3 C, and so forth. In the Digit symbol test measuring a processing speed the participant was present with a sheet of paper showing nine numbers pa ired with symbols. The p articipant was then asked to copy the corresponding symbols in boxes adjoining to the numbers in 90 seconds. Descriptive statistics for linguistic and cognition measures As shown in Table 3 4 there was no difference in graphic familiarity between dyslexics and typical readers ( F (1, 33) =.002, p >.05). A main effect for group was found on the vocabulary measure ( F (1, 33) =5.00, p < .05 p 2 = .58 ) showing that DR had significantly lower scores on the Shipley vocabulary test than TR ( M D R =27.13, SD D R =.6.71, M TR =31.00, SD TR =5.17). Similarly the univariate ANOVAs showed significant group effects for all four working memory variables : Di git span forward ( F (1, 33) =7.27, p < .05 p 2 = .18 ); D igit span backward ( F (1, 33) =4.93, p < .05 p 2 = .13 ); D igit ordering ( F (1, 33) =11.12, p < .01 p 2 = .25 ); and B lock recall ( F (1, 33) =7.27, p < .05 p 2 = .18 ). Lastly, the univariate ANOVAs showed group differences for all five executive function variables: S troop color ( F (1, 33) =10.70, p < .01 p 2 = .25 ); S troop word ( F (1, 33) =24.10, p < .001 p 2 = .42); T rail M aking A ( F (1, 33) =4.53, p < .05 p 2 = .12 ); T rail M aking B ( F (1, 33) =4.60, p < .05 p 2 = .12 ); and D igit symbol s ubstitution ( F (1, 33) =18.81, p < .01 p 2 = .30 ).
43 Exper imental graph comprehension assessment E xperimental graph comprehension test was created using three graph types: (1) line graphs, (2) bar graphs, and (3) horizontal bar graphs and two graphic display s of varying complexity: (1) single graphic display s (2) double g raphic display s (i.e., single line vs. double line ; single bar vs. double bar; single horizontal bar vs. double horizontal bar). Each graph was accompanied by two types of questions The fi rst type of qu estion required the participant to interpret data by examining the intersection of a point on the Y axis and a point on the X axis (e.g., H ow many cats does Harry have?). The second type of questions required the participant to interpret dat a by comparing values at two points on the graph oes Harry have more cats Seventy two graphs were created based on a study conducted by Ratwani and Trapton (2008) and 72 graphs with double graphic display s (i.e. double line graphs, double bar graphs, and double horizontal bar graphs) were added to create ano ther level of graph complexity. Each g ra ph included the following five major components (Carpenter & Shah, 1998; Renshaw, Finlay, Tyfa, & Ward, 2004) (1) a n X Axis consisting of 3 common objects ( e.g., cats, rabbits, dogs); (2) a n Y Axis with numbers ranging from 1 to 10 E ach number between 1 and 10 was assigned to each segment in order to ndicate the number of objects that it represented. In the horizontal bar graphs, the axes were reversed; (3) p attern of lines in the line graphs and bars in the bar an d the horizontal bar graphs to present the data ; (4) legend that informed the participant the referents of graphs features. The legend and coded by different colors; and (5) t wo types of question that appeared to the right of the graph one by one. For the p oint locating que stion, the participant was asked for a literal reading for what a point represents. The answer was one of the ten numbers representing the Y value presented in a given graph. For the comparison question, the participant was asked to compare two data sets represented on the
44 g raph and to answer as yes/no. Participants were asked to respond to the questions by pressing one of two buttons and receive d a score of either 1 (correct answer) or 0 (incorrect answer) The maximum total score was 144. Sample gr aph s and comprehension questions are shown in Appendices D and E.
45 Table 3 2 Mean standard scores on the diagnostic reading tests for students with dyslexia Diagnostic Reading Tests ( mean standard score) Mean SD Min Max WJ III COG Verbal Comprehension (100) 9 6.00 13.31 78 119 Verbal Ability Composite (100) 9 6.00 13.31 78 119 CTOPP Elision (10) 7.97 3.65 2 12 Blending Words (10) 8.05 3.18 4 14 Phonological Awareness Composite (100) 86.71 18.63 61 115 Memory for Digits (10) 8.28 2.84 3 13 Nonword Repetition (10) 7.36 1.15 6 9 Phonological Memory composite (100) 86.92 9.31 73 103 Rapid Digit Naming (10) 7.29 2.81 2 11 Rapid Letter Naming (10) 5.57 2.34 2 10 Rapid Naming Composite (100) 76.71 15.34 52 100 TOWRE Sight Word Efficiency (100) 81.50 10.30 67 109 Phonemic Decoding Efficiency (100) 77.29 8.70 55 95 Total Word Reading Efficiency (100) 75.21 9.41 53 92 (1) WJ III COG: Woodcock Johnson III Test of Cognitive Abilities, (2) WJ III ACH: Woodcock Johnson III Tests of Achievement, (3) CTOPP: Comprehensive Test of Phonological Processing, (4) TOWRE: Test of Word Reading Efficiency SD: standard deviation, Min: minimum, Max: maximum
46 Table 3 3 Mean standard scores on the diagnostic reading tests for students with normal reading skills Diagnostic Reading Tests ( mean standard score) Mean SD Min Max WJ III COG Verbal Comprehension (100) 101.55 8.57 85 121 Verbal Ability Composite (100) 101.55 8.57 85 121 CTOPP Elision (10) 9.65 2.56 4 12 Blending Words (10) 10.70 2.18 4 14 Phonological Awareness Composite (100) 101.05 11.32 70 115 Memory for Digits (10) 10.80 2.12 7 15 Nonword Repetition (10) 9.25 1.97 6 13 Phonological Memory composite (100) 100.15 10.27 82 118 Rapid Digit Naming (10) 11.20 1.94 6 14 Rapid Letter Naming (10) 11.00 2.79 6 16 Rapid Naming Composite (100) 106.60 12.93 76 127 TOWRE Sight Word Efficiency (100) 101.60 12.17 84 113 Phonemic Decoding Efficiency (100) 100.65 7.82 90 120 Total Word Reading Efficiency (100) 101.45 9.25 84 120 (1) WJ III COG: Woodcock Johnson III Test of Cognitive Abilities, (2) WJ III ACH: Woodcock Johnson III Tests of Achievement, (3) CTOPP: Comprehensive Test of Phonological Processing, (4) TOWRE: Test of Word Reading Efficiency SD: standard deviation Min: minimum, Max: maximum
47 Table 3 4 Descriptive statistics for graphic familiarity, linguistic and cognition measures Measures (maximum scores) DR Group ( N=15 ) T R Group (N=20 ) F value (1,33) p 2 Observed power M SD M SD Graph familiarity (133) 93.27 17.19 93.08 11.56 .002 .969 Shipley vocabulary (40) 27.13 6.71 31.00 5.17 5.00 .032* .131 .582 Working Memory Verbal Working Memory Digit Span Forward (14) 6.67 1.80 8.35 2.41 7.26 .011* .180 .744 Digit Span backward (14) 5.60 1.92 7.05 2.31 4.93 .033* .130 .578 Digit Ordering (24) 13.73 3.43 16.95 2.50 11.12 .002** .252 .899 Visual Working Memory Block Recall ( 54 ) 27.27 6.05 31.57 3.57 7.27 .011* .181 .745 Executive Functions Stroop Color (100) 71.00 14.55 84.15 9.82 10.70 .003** .245 .888 Stroop Word (100) 46.67 9.86 65.00 11.55 24.10 .000*** .422 .997 Trail Making A (sec) 79.93 25.24 65.35 17.75 4.53 .041* .121 .542 Trail Making B (sec) 94.80 25.60 78.60 19.06 4.60 .039* .122 .549 Digit Symbol Substitution (93) 62.20 8.18 71.60 8.23 13.81 .001** .295 .950 M: mean, SD: standard deviation p < .05, ** p < .01, *** p < .001
48 CHAPTER 4 RESULTS DATA ANALYSIS All data for demographic characteristics, diagnostic reading assessments, and research questions were analyzed using SPSS version 17.0 for Windows. Results are presented below as they related to each of the research questions followed by a summary of key findings. Research Question 1: Comprehension Accuracy The first goal of this study was to investi gate how dyslexic readers (DR) perform on a graph comprehension task compared to typical readers (TR). Table 4 1 shows the performance of the two groups on the graph comprehension tasks. A mixed 2 (between subjects reader type: DR vs. TR ) 3 (within s ub jects graph types : line vs. bar vs. horizontal bar graphs ) 2 (within subjects graphic display : single vs. double graphic displays ) 2 (within subjects question types: point locating vs. comparison questions ) an alysis of variance (ANOVA) and post hoc Bonferroni corrected pairwise comparisons were used to examine whether any significant differences exist in graph comprehension accuracy between young adult with DR and TR age matched controls This four way ANOVA yielded significant findings for (1) four main effects, (2) two two way interactions, and (3) one three way interac tions. A summary of results is shown in Table 4 2 ANOVAs with within subject variables were calculated under the assumption of sphericity, a condition where the variances of the differences between levels of the within subject variables (Maunchly, 1940) result indicated that the assumption of sphericity has not been violated (all p >.05). Therefore, sphericity assumed F values were reported for all measures. Total number of correct response for each of the graph interpretation questions (12 items per each condi tion) was used for data analysis. Effect sizes, using partial eta square, were
49 also reported to evaluate the size of mean differences. Cohen (1988) characterized the value of partial eta squared of .01 as small, .06 as moderate, and .14 as a large effect size. Significant main effects were found for all variables : group, graph type, question type, and graphic display patter n. Altho ugh differences in mean performance on comprehension accuracy were very small DR were significantly less accurate in interpreting graphs than TR (M D R = 11.33, SE D R = .064, M TR = 11.56, SD TR = .056; F (1,33) = 7.28, p =.011, p 2 = .181). Significant differences in comprehension accuracy were found for graph types (M line = 11.60, SE line = .056, M bar = 11.54, SD bar = .058; M horizontal bar = 11.20, SE horizontal bar = .060, F (2,32) = 19.87, p < .001, p 2 = .376). Bonferroni corrected pairwise comparisons between the three graph types revealed that all participants were less accurate on horizontal bar graphs items than on line graphs ( p < .001) or bar graph items ( p < .001). Similarly, all participants were less accurate in interpreting graphs w ith double graphic displays than those with single graphic displays (M single = 11.63, SD single = .053; M double = 11.27, SE double = .058, F (1,33) = 25.85, p < .01, p 2 = .439). Again, all participants were less accurate in interpreting graphs presented with comparison questions than graphs with point locating questions (M Point locating = 11.68, SD Point locating = .033; M comparison = 11.22, SE comparison = .066, F (1,33) = 55.80, p < .001, p 2 = .628). The ANOVA revealed two significant two way interac tions: group question type and graph type graphic display. First, a significant interaction between group and question type was found ( F (1,33) = 9.08 p < .01, p 2 = .214). To interpret the nature of the i nteractions, each main effect ( group) at different levels of the second main effect (question type) was compared, using Bonferroni correlations to control for alpha inflation. For the comparison question type, DR were significantly less accurate than TR ( t (33) = 3.11, p < .01). However, group di fference was not found for the point locating question type ( p > 0 5). Thus,
50 decreased when the complexity of question increased. Figure 4 1 displays graphical representations of this information. Further, a significant inte raction for graph type graphic display ( F (2,32) = 23.91, p < .001, p 2 =.599) was found. The interaction was explored by comparing the comprehension accuracy between line, bar, and horizontal bar graph across graphic display (single and double). compreh ension accuracy for line ( t (32) = 7.25, p < .001) or bar graphs ( t (32) = 8.57, p < .001) on the double graphic display. This pattern was not found on the single graph ic display (all p > .05 ). This interaction reveals that when the complexity of graph inc reased, part icipants have more difficulty process ing horizontal graphs compared to line or bar graphs Figure 4 2 displays graphical representations of this information. Lastly, one significant three way interaction was found for group graph type ques tion type ( F (2,32) = 5.94 p < .01, p 2 = .271 ). The interaction was explored by comparing the comprehension accuracy between groups across question types (i.e. point locating and comparison) in each graph type (i.e., line, bar, and horizontal bar graphs). The finding indicates that the group and question type interacti on was different across graph types (i.e., line, bar and horizontal bar graphs). Specifically, in the line ( t (32) = 3.70, p < .01) and the bar graphs ( t (32) = 2.53, p < .05) DR responded less accurately than TR for comparison questions, while there was no significant difference between the two groups for point locating questions (all p > .05). In contrast, for the horizontal bar graphs, this pattern was not found (all p > .05 ). Figure 4 3 displays graphical representations of this information.
51 Table 4 1. Descriptive statistics for DR and T R groups on graph accuracy DR Group (N=15) T R Group (N=20) Measures (milliseconds) P oint locating question C omparison question P oint locating question C omparison question Mean SD Mean SD Mean SD Mean SD Line graph S ingle pattern 12.00 .00 11.20 .68 11.85 .49 11.75 .55 D ouble pattern 11.98 .26 10.67 1.23 11.75 .44 11.65 .81 B ar graph S ingle pattern 11.87 .35 10.93 .96 11.85 .37 11.50 .76 D ouble pattern 11.47 .74 11.33 .72 11.80 .52 11.60 .60 Horizontal bar graph S ingle pattern 11.87 .52 11.47 .74 11.85 .37 11.35 .99 D ouble pattern 10.80 .41 10.47 .64 11.10 .31 10.47 .571 M: mean, SD: standard deviation, Min: minimum, Max: maximum
52 Table 4 2. Summary of mixed four way ANOVA on comprehension accuracy measures Effects F value p Partial 2 Observed power Main Effects Group F (1,33) = 7.28 (A ) .011* .181 .745 Graph type F (2, 32 ) = 19.87 (A ) .000*** .376 1.00 Question type F (1 ,33) = 55.80 (A ) .000*** .628 1.00 Graphic display F (1,33) = 25.85 (A) .001** 439 1.00 Two way Interaction Group x Graph type F (2,32 ) = 1.33 (A) .272 .039 .278 Group x Question type F (1, 33) = 9.08 (A) .005** .214 .830 Group x Graphic display F (1,33) = .072 (A) .663 .006 .071 Graph type x Question type F (2,32 ) = .815 (A ) .447 .024 .184 Graph type x Graphic display F (2,32 ) = 23.91 ( A ) .000*** .599 1.00 Question type x Graphic display F (1,33) = 2.85 (A) .101 .079 .374 Three way Interaction Group x Graph type x Question type F (2,32) = 5.94 (A ) .006** .271 .887 Group x Graph type x Graphic display F (2,32) = 1.54 (A ) .220 .045 .318 Group x Question type x Graphic display F (1,33) = 2.01 (A) .116 .057 .280 Graph type x Question type x Graphic display F (2,32) = .898 (A ) .412 .026 .199 Four way Interaction Group x Graph type x Graphic display x Question type F (2,32) = .719 (A ) .491 .021 .167 A: Sphericity assumed F values, B: Greenhouse Geisser corrected F values p < .05, ** p < .01, *** p < .001
53 Figure 4 1. Two way interaction between question type and group Figure 4 2 Two way interaction between graphic display and graph typ e 10.00 10.40 10.80 11.20 11.60 12.00 point locating comparison Comprehension accuracy Question type DR TR 10.00 10.40 10.80 11.20 11.60 12.00 Single Double Comprehension accuracy Graphic display Line Bar Horizontal Bar
54 Figure 4 3 Three way interaction between graph type, question type, and group 10.00 10.40 10.80 11.20 11.60 12.00 point locating comparison Line graph 10.00 10.40 10.80 11.20 11.60 12.00 point locating comparison Bar graph 10.00 10.40 10.80 11.20 11.60 12.00 point locating comparison Horizontal bar graph Comprehension accuracy
55 Research Question 2: Eye Tracking Data A nalysis The second goal of this study was to compare eye gaze data of DR and TR to see whether their performance would be influenced by graph type, graphic display, and question type. The overall viewing times (i.e., whole trial times) were analyzed using a mixed 4 way multivariate analysis of variance (ANOVA), with group (i.e., DR v s. TR) as a between subject factor and graph type (i.e., line vs. bar vs. horizontal bar graph), graphic display (i.e., single vs. double ), and question type (i.e., point location vs. comparison questions) as within subject factors. Post hoc pairwise compa risons using Bonferroni corrections to control for alpha inflation were conducted to interpret main and interaction effects. Table 4 3 shows the overall viewing times of the two groups on the graph comprehension tasks. The four way ANOVA yielded significa nt findings on overall viewing time for (1) four main effects, (2) three two way interactions, and (3) one three way interaction. These ANOVA results are shown in Table 4 4 Within subject variables were calculated under the assumption of sphericity, a condition where the variances of the differences between levels of the within been v iolated in graph type x graphic display interaction and group x graph type x graphic display x question type interaction ( p < .05). Thus, relatively conservative Greenhouse Geisser corrected F values were reported for the two interactions. Significant mai n effects were found for all variables : group, graph type, graphic disp lay, and question type. DR viewed graphs for significantly longer time than TR (M D R = 7197.04, SE D R = 352.07 ; M TR = 5371.79, SD TR = 304.90; F (1,33) = 15.36, p < .001, p 2 = .318). Viewing times were different depending on graph type (M line = 6411.39, SE line = 240.70, M bar = 6128.15, SD bar = 232.21; M horizontal bar = 6313.72, SE horizontal bar = 240.29, F (2,32) = 5.98, p < .01, p 2
56 = .266). Bonferroni corrected pairwise comparisons between three graph types revealed that all participants spent a significantly longer time processing line graphs than bar graphs ( t (32) = 3.23, p < .01). There was no time differences between line and horizontal bar graphs ( p > .05 ). Similarly all participants spent significantly longer time in processing graphs wit h double graphic displays than graphs with single graphic displays (M single = 5359.68, SE single = 193.70; M double = 7209.16, SE double = 280.88, F (1,3 3) = 215.06, p < .001, p 2 = .867). Finally, all participants spent significantly longer time in processing graphs with comparison questions than graphs with point locating questions (M Point locating = 5527.88, SE Point locating = 204.56; M comparison = 7040.96, SE comparison = 269.01, F (1,33) = 198.97, p < .001, p 2 = .858). The ANOVA revealed three significant two way interactions: group question type, group graphic display, and question type graphic display. First, a significant interaction betwe en group and question type was found ( F (1,33) = 9.46, p < .01, p 2 = .223). To interpret the nature of the interactions, each main effect at different levels of the second main effect, using Bonferroni correlations to control for alpha inflation, was examined. For both groups, the overall viewing times for comparison questi ons were significantly longer tha n for point locating questions ( t (33) = 11.37, p < .001 for DR, t (33) = 8.45 p < .001 for TR ) ; however, the overall viewing times in the two question types showed a greater difference in D R group rather than in TR group, as shown f igure 4 4. This finding suggests that time dyslexic readers spent to process graphs significantly increased when the complexity of question increased. Further, a significant interaction between group and graphic display was found ( F (1, 33) = 4.95, p < .05, p 2 = .131). The interaction was analyzed by using Bonferroni correlations to control for alpha in flation. For both groups, the overall viewing times for graphs with double displays were significantly longer than for graphs with single displays( t (33) = 11.21, p < .001
57 for DR and t (33) = 3.99 p < .0 1 for TR ) ; however, t he overall viewing times in the two graphic displays showed a greater difference in D R group rather than in TR group, as shown f igure 4 5 This finding suggests that time dyslexic readers spent to process graphs significantly increased when the complexity of graphic displays increased. L astly, a significant interaction for graphic display question type ( F (1,33) = 17.04, p < .001, p 2 = .341) was found. To interpret the nature of the interactions, each main effect at different levels of the second main effect was examined, using Bonferroni correlations to control for alpha inflation. For both graphic displays, the overall viewing times fo r graphs with comparison questions were significantly longer than for graphs with point locating questions ( t (32) = 12.08, p < .001for single displays and t (32) = 17.16, p < .001for double displays); however, th e overall viewing times in the two questi on types showed a greater difference for double graphic displays rather than for single graphic displays, shown figure 4 6. The ANOVA also yielded one significant three way interaction for graph type x graphic display x question type ( F (2,32) = 6.00, p < .01, p 2 = .154). The interaction s between graphic display and question type were compared at each type of graph (line, bar, and horizontal bar graphs). In the line graph, the viewing time difference between single and double graphic displays was significa ntly larger in graphs with comparison questions than graphs with point locating questions ( t (32) = 4.64, p < .001). That is, participants needed significantl y longer time to process double graphic graphs with comparison questions. Similarly, in the bar gr aph, the viewing time difference between the two graphic displays were significantly larger in the comparison question condition than in the point locating question condition ( t (32) = 3.46, p < .01). However, this different pattern between single and dou ble graphic displays across types
58 of questions was not found in the horizontal graphs. Figure 4 7 displays graphical representations of this information.
59 Table 4 3. Descriptive sta tistics for DR and TR groups on overall viewing time s DR Group (N=15) T R Group (N=20) Measures (milliseconds) P oint locating question C omparison question P oint locating question C omparison question Mean SD Mean SD Mean SD Mean SD Line graph S ingle pattern 5733.73 1317.12 6832.47 1226.65 4187.40 1122.98 5134.03 1217.07 D ouble pattern 7195.24 1180.96 9616.16 2069.06 5500.63 1373.14 7091.42 2464.90 B ar graph S ingle pattern 5063.78 1102.89 6905.02 1470.64 3881.83 1112.15 4928.30 1417.46 D ouble pattern 6950.64 1093.74 9275.73 1783.54 5337.07 1600.95 6682.80 2044.42 Horizontal bar graph S ingle pattern 5299.16 916.96 6957.64 1099.43 4210.68 1203.55 5182.12 1540.36 D ouble pattern 7410.56 2044.69 9124.40 1861.74 5563.85 1410.35 6761.37 2011.84 M: mean, SD: standard deviation
60 Table 4 4. Summary of mixed four way ANOVA on overall viewing time measures Effects F value p Partial 2 Observed power Main Effects Group F (1,33) = 15.36 (A ) .000*** .318 1.00 Graph type F (2,32) = 5.98 (A) .005** 266 .867 Question type F (1,33) = 198.97 (A) .000*** .858 1.00 Graphic display F (1,33) = 215.06 (A) .000*** .867 1.00 Two way Interaction Group x Graph type F (2,32) = .186 (A) .831 .006 .078 Group x Question type F (1, 33) = 9.46 (A) .004** .223 .847 Group x Graphic display F (1,33) = 4.95 (A) .033* .131 .579 Graph type x Question type F (2,32) = 1.20 (A) .308 .035 .253 Graph type x Graphic display F (1.69,55.08) = .19 (B) .824 .006 .079 Question type x Graphic display F (1,33) = 17.04 (A) .000* .341 .980 Three way Interaction Group x Graph type x Question type F (2,66) = .774 (A) .466 .023 .176 Group x Graph type x Graphic display F (2,66) = .194 (A) .824 .006 .079 Group x Question type x Graphic display F (1,33) = .888 (A) .353 .026 .150 Graph type x Question type x Graphic display F (2,32) = 6.00 (A) .004* .154 .868 Four way Interaction Group x Graph type x Graphic display x Question type F (1.90,62.84) = 1.47 (B) .240 .042 .301 A: Sphericity assumed F values, B: Greenhouse Geisser corrected F values p < .05, ** p < .01, *** p < .001
61 4,000 4,900 5,800 6,700 7,600 8,500 DR TR Overall viewing time (ms) Graphic display single double 4,000 4,900 5,800 6,700 7,600 8,500 DR TR Overall viewing time (ms) Question type point locating comparison Figure 4 4 Two way interaction between question type and group Figure 4 5 Two way interaction between graphic display and group
62 4,000 4,900 5,800 6,700 7,600 8,500 single double Overal viewing time (ms) Question type point locating comparison Figure 4 6 Two way interaction between question type and graphic display Figure 4 7 Three way interaction between graphic display, question type, and graph type 4,000 5,000 6,000 7,000 8,000 9,000 point locating comparison line graph 4,000 5,000 6,000 7,000 8,000 9,000 point locating comparison Bar graph 4,000 5,000 6,000 7,000 8,000 9,000 point locating comparison Horizontal bar graph Overall viewing times (ms)
63 Research Question 3 : Eye Tracking Data A nalysis II DR took longer to visually process this information overall. The following question was whether this effect would be driven by particular parts of the display or is due to longer looking times at a ll of the parts of the display. The third goal of this study was to compare eye gaze data of D R and TR in order to determine viewing times differ for specific six graphic subregions (i.e., pattern, x axis, y axis, legend, question, answer). A schematic of these subregion is shown in f igure 4 8 A multivariate analysis of variance (MANOVA) and subsequent univariate analyses of variances (ANOVAs) were conducted, using Table 4 5 show s the performance of the t wo groups in two standard time measures of eye gaze which have been adapted from key measures of reading text without graphs (i.e., first pass time and total viewing time). According to Filik et al. (2005) and Hutzler & Wimmer (2004) first pass time is the sum of all fixations in a re gion of the sentence prior to exiting that region and total viewing time is the sum of all fixations in a region. In reading research, regression path time is one of the key eye tracking measures. However, the regression path time was excluded in this study because the regression path method (i.e., looking back at text) was developed for examining text reading. Given that information on a graph is not likely to be viewed in the linear manner as in reading text, the use of this procedure has no intrinsi c value for graph interpretation.
64 Figure 4 8 S ix regions of the display: pattern, X Axis, Y Axis, legend, question and answer Three assumptions for MANOVA were checked prior to the analysis: (1) multivariate normal distribution, (2) homogeneity of covariance matrices, and (3) homogeneity of error variance. First, to check the multivariate normal distribution, skewness and kurtosi s of dependent variables were calculated. For the first pass time, skewness and kurtosis values for all regions except graph pattern fell within the range of 1 to + 1. For the total viewing time, the values for all regions fell within the range of 1 to + 1. The skewness and kurtosis values of 2 out of 18 variables were outside of the 1 range, which may indicate the violation of normality. However, the Multivarite F test is robust in accounting for non normality, if the non normality is caused by
65 skewn ess or kurtosis (French, Macedo, Poulsen, Waterson, & Yu, 2008) For the second (Box, 1949) for the homogeneity of variance covarianc e matrices across design cells were estimated. This assumption is especially important for unbalanced designs (i.e., unequal number in each group). both viewing time measures at the p =.001 level: first pass time ( F (45, 2995.10)=.778, p >.001 ) and to tal viewing time ( F (45, 2995.10) = 1.51, p > .001 ). These findings indicate that the assumption of homogeneity of the variance covariance matrix of dependent variables across design cells was not violated. For the third assumption, homogeneity of variance, (Levene, 1953, 1960) were conducted for each of the reading time measures. None of the variables had significant F value (p < .001). Thus, the assumption of homogeneity of variance is supported for all variables. In the most cases, the three assumptions for using the MANOVA were not violated, however, because of the limited number of d to estimate the multivariate F statistical inference. For both measures, DR spent significantly longer time than TR. Specifically, f irst pass times ( F (1, 33) = 12.79, p < .01, p 2 = .28) and total viewing times ( F (1, 33) = 21.07, p < .001, p 2 = .39) were reliably longer for DR than for TR. Next, multivariate analysis of variance (MANOVA) and subsequent univariate analyses of variances (ANOVAs) were conducted to compare first pass times and total viewing times on each subregion (i.e., pattern, x axis, y axis, legend, question, answer ). The analyses for graphics (pattern, x axis, y zaix and legend) and for text (question and answer) were conducted separately to investigate the etiology of group differences.
66 As shown in Table 4 6 in the graphic areas, a main effect for group was not found for first pass times ( F (4,30) = 1.38, p > .05, p 2 = .01). The following univariate ANOVA also did not reveal any significant group effect. In contrast, in the text areas, a main effect for group was found for first pas s times ( F (2,32) = 11.29, p < .001, p 2 = .41). Thus, univariate ANOVAs were used to examine which dependent variables were affected by group. Significant main effects were found for both subregion variables, question ( F (1, 33) = 17.23, p < .001, p 2 = .34 ) and answer ( F (1, 33) = 12.08, p < .01, p 2 = .268). The effect sizes suggest moderate to large effects for eye gaze on subregion variables with the largest effect for the question subregion As shown in Table 4 7 the multivariate analyse s of variance for total viewing time s of graphics and text were also conducted separately. In the graphics multivariate analysis of variance there was no significant main effect for group ( F (4, 30) = 2.52, p >.05, p 2 = .25) because there was only one significant uni variate main effect, for pattern ( F (1, 33) = 6.82, p < .05, p 2 = .171). In contrast, the multivariate analysis of variance for total viewing time for text regions showed significant main effects for group ( F (2, 32) = 16.66, p < .001, p 2 = .51). Using univariate ANOVAs, significant main effects were found for both dependent variables: question ( F (1, 33) = 26.25, p < .001, p 2 = .443); and answer ( F (1, 33) = 6.29, p < .05, p 2 = .16). The effect sizes suggest moderate to large effects of t otal viewing time with the largest effe ct shown for the question variable.
67 Table 4 5 Des criptive statistics for DR and T R groups reading times on each graphic region Measures ( millisecond ) DR Group ( N=15 ) T R Group (N=20 ) M SD Min Max M SD Min Max First Pass Time X axis 439.03 85.27 329.40 661.08 446.07 86.95 324.50 655.43 Y axis 205.16 53.31 107.52 288.26 162.56 74.18 25.20 308.49 Pattern 275.34 98.70 188.24 515.43 273.37 75.95 191.06 460.45 Legend 440.50 59.56 334.00 563.34 456.74 78.42 340.57 612.40 Question 2224.89 398.64 1484.21 2864.03 1633.06 430.67 935.86 2459.97 Answer 420.68 96.43 273.37 636.79 329.19 48.87 188.32 428.26 Total Viewing Time X axis 654.35 200.41 339.40 1074.72 611. 78 206.95 282.83 971.42 Y axis 294.92 41.53 227.14 386.18 296.16 60.72 218.93 450.52 Pattern 917.05 307.46 453.74 1413.54 705.83 166.46 455.32 1108.82 Legend 760.10 124.68 471.27 974.63 654.62 178.38 446.11 1112.11 Question 3783.86 593.49 2545.48 4734.54 2476.39 842.57 1392.89 4795.48 Answer 490.38 68.29 321.25 604.95 431.27 59.09 312.97 543.34 M: mean, SD: standard deviation, Min: minimum, Max: maximum
68 Table 4 6 Summary of b etween m ultivariate a nalysis of v ariance on interest areas of first pass times F value p 2 Observed power Graphics F (4,30) = 1.38 .967 .018 .075 X axis F (1,33) = .06 .812 .002 .056 Y axis F (1,33) = .01 .946 .000 .050 Pattern F (1, 33) = .00 .957 .000 .050 Legend F (1,33) = .45 .508 .013 .100 Text F (2,32) = 11.29 .000*** .414 .987 Question F (1, 33) = 17.23 .000*** .343 .981 Answer F (1,33) = 6.29 .017* .160 .682 p < .05, ** p < .01, *** p < .001
69 Table 4 7 Summary of b etween m ultivariate a nalysis of v ariance on interest areas of total viewing time s F value p 2 Observed power Graphics F (4, 30) = 2.52 .06 .252 .642 X axis F (1,33) = .31 .55 .01 .091 Y axis F (1,33) = 3.56 .07 .10 .45 Pattern F (1, 33) = 6.82 .013* .171 1.00 Legend F (1,33) = 3.83 .059 .104 .476 Text F (2,32) = 16.664 .000*** .510 .999 Q uestion F (1, 33) = 26.25 .000*** .443 1.00 Answer F (1,33) = 12.08 .001** .268 .921 p < .05, ** p < .01, *** p < .001
70 Research Q uestion 4 : Correlation a mong Graph C omprehension and Language a nd Cognition M easures Pearson's correlation coefficients were used to measure the strength of the relationships among comprehension accuracy, graphic (including x axis, y axis, legend, and pattern) view ing time, text (including question and answer) viewing time, receptive vocabulary, working memory, and executive function perf ormance, as shown in tables 4 8 and 4 9 The s ignificance level for these calculations was set at p < .05 ( two tailed). For DR, comprehension accuracy did not correlate significantly with any language or cognition measures (all p >.05). The graphic viewing time correlated significantly with two working memory measures (D igit forward : r = .561 p < .0 5 B lock recall: r = .541, p < .0 5) and one executive function measure ( D igit Symbol S ubstitution : r = .528, p < .0 5). The text viewing time correlated with three verbal working memory measures (D igit forward : r = .647 p <.01 D igit backward: r = .610 p <.0 5 D igit ordering: r = .577 p < .0 5 ) and one executive function measure ( Stroop word : r = .536 p < .0 5). For TR, comprehension accuracy correlated significantly with three executive function measures ( Stroop word : r = .592 p < .01 Trai ls A: r = .447 p < .0 5 Trail s B: r = .5 63 p < .0 1). The graphic viewing time c orrelated significantly with one working memory measure (B lock recall: r = .551 p < .05) and two executive function measure s ( Stroop word : r = .738 p <.01 D igit Symbol S ubstitution : r = .545 p < .05). The text viewing time correlated significantly with one working memory (B lock recall: r = .550, p < .05 ) and one executive function m easure ( Stroop word : r = 521, p < .05 ).
71 Table 4 8 Students with d yslexia: correlations betwe en comprehension accuracy, viewing times and cognition measures Vocabulary Working Memory Executive Function Shipley Digit f orward D igit b ackward Digit ordering Block recall Stroop Color Stroop Word Trail s A Trail s B D i git Symbol S ubstitution Comp. accuracy .293 .173 .313 .042 .285 .046 .141 .223 .078 .264 Graphic viewing time .096 .561* .050 .042 .541* .041 .346 .246 .378 .528* Text viewing time .209 .647** .610* .577* .346 .104 .536* .424 .051 .212 Table 4 9 Typical readers: correlations between comprehension accuracy, viewing times and cognition measures Vocabulary Working Memory Executive Function Shipley Digit f orward D igit b ackward Digit ordering Block recall Stroop Color Stroop Word Trail s A Trail s B D igit Symbol S ubstitution Comp. accur acy .222 .067 .068 .052 .434 .326 .592** .447* .563** .052 Graphic viewing time .418 .372 .092 .031 .551* .020 .738** .406 .409 .545* Text viewing time .159 .391 .037 .037 .550* .105 .521* .039 .336 .041 p < .05, ** p < .01, *** p < .001 Comp : comprehension Graphic viewing time s : viewing time s for x axis, y axis, pattern, and legen d Text viewing time s : viewing time s for question and answer
72 Summary o f Findings A summary of the comparative data for DR and TR on graph interpretation accuracy, graph interpretation processing time and correlations between graph comprehension variables, vocabulary, working m emory, and executive functions is shown below. R esults for questions 1 and 2 (the comprehension a ccuracy and eye tracking data ) are as follows : (1 ) Students with dyslexia performed less accurately and more slowly than typical re aders on interpreting graphs; (2) Participants overall performed less accurately and mo re slowly for graphs associated with double graphic displays than for graphs with single graphic displays; (3) Participants performed less accurately and more slowly for graphs presented with comparison questions than for graphs with point locating questio ns; (4) Participants were most accurate and efficient in answering questions associated with bar graphs. For the line graphs, participants were as accurate as for the bar graphs; however, they spent greater time processing the line graphs than the bar grap hs. Participants were least accurate and spent the longest amount of time processing the horizontal bar graphs ; (5) For both comprehension accuracy and viewing times, there was a significant group and question type interaction effect. Dyslexic participants took significantly longer to process graphs associated with comparison questions than typical readers. This finding suggests that dyslexics were more i nfluenced by the complexity of the question than typical readers ; (6) For the viewing times, there was a significant group and graphic display interaction effect. Dyslexic readers spent s ignificantly more time viewing graphs associated with double graphic displays (i.e., more difficult questions) than typical readers. Again, this finding suggests that the co mplexity of the graph influenced dysl exics more than typical readers; (7) There was no interaction between the group and the graph types for either comprehension accuracy or v iewing times. That is, dyslexic subjects did not performed differently from typic al readers across graph types; and (8) There was a three way interaction
73 between g roup graph type and question type for comprehension accuracy. For the line and bar graphs, the difference between dy slexics and typical readers was larger for comparison que stion s than for point locating question s. This pattern was not observed for the horizontal bar graphs. In addition, there was a three way interaction between graph type, graphic display and question type for the viewing times. For both line and bar graphs, the difference between single and double gra phic displays was larger for graphs associated with comparison questions than graphs associated with point locating questions. This pattern was not observed for horizontal bar graphs. Results for the question 3 are as follows : (1) The eye fixation data in the linguistic text regions (i.e., location of questions and answers ) revealed that students with dyslexia spent significantly longer times reading text than typical r eaders; and (2) The eye fixation data in th e graphic pattern regions revealed that when participants initially looked at the graph (first pass) there was no difference between the two groups. However, the entire time spent in the pattern area was longer for the students with dyslexia than for the typical readers. Results for the question 4 are as follows (1) interpretation correlated significantly with several working memory variables, while typical e correlated significantly with several executi ve function skills ; and (2) On vocabulary and other cognitive variables (i.e., working memory and executive functions),
74 CHAPTER 5 DISCUSSION This study was designed to investigate group differences between dyslexic and non dyslexic young adults on their interpretation of graphs and to determine if the groups differed when when graph complexity and question complexity were assessed. The specific aims of this study were 1) to compare comprehension accuracy of the two groups at different levels of com plexity for graph s and question s ; 2) to compare the eye gaze pattern s of the two groups at different levels of com plexity for graph and question; 3) to compare the two groups on the amount of time that they spent lookin g at specific subregion s of graph s (i.e., pattern, x axis, y axis, legend, question, answer); and 4) to explore relationship s between of graph interpretation, receptive vocabulary and cognitive abilities for working memory, attention, and in hibition. Fifteen college students with dys lexia and 20 control peers were asked to look at each graph presented with a ques tion on a computer screen and answer written questions while their eye movement s were being tracked. Three types of graphs (i.e., l ine, bar, and horizontal bar graphs ), two types of graphic displays (i.e., single and double graphic displays ), and two types of questions (i.e., point locating and comparison) were used to measure the effect of graph type and question type on the interpretation Group Comparisons o n Graph Interpretation Accuracy The results revealed differences between the two groups on comprehension accuracy even though the differences were numerically small Students with dyslexia performed sign ificantly more poorly than typical readers o n the graph interpretation tasks This finding is consistent with previous research on the graph com prehension skills of individuals with learning disabilities.
75 Parmar and Singer (2005) reported that elementary school children with learning disabilities had more difficulty interpreting line graphs tha n their p eers without learning disabilities and Curcio (1982) reported that reading achievement is a unique pred ictor of graph comprehension in fourth and seventh graders. Across the two groups, c omprehension accuracy was significantly lower o n the horizontal graph s than on the line or the bar graph s While the three graph types consists of an x axis and a y axis, the horizontal bar graph requires a different visual orientation because the axes are reversed. Shah and Hoeffner (2002) noted that people typically assume the x axis to be the independent variable an d the y axis to be the dependent variable. The horizontal bar graph rpretation scores. This finding, showing an association between task difficulty and specific graph features appears to support the perceptual feature view previously discussed, in which different graphic features activate different mental representation s (Lohse, 1993) as opposed to the invariant structure v iew, that posits all graphs are mentally represented in similar ways. In addition to differences found across grap h types, in the current study, differences were also found when a graph was made more complex by including additional data. For example, participants were less accurate in interpreting graphs with double graphic displays (e.g., double line graph) than with single graphic displays (e.g., single line graphs). This finding is con sistent with the result of several previo us studies of graphic displays (e.g., Carpenter & Shah, 1998; Shah & Carpenter, 1995; Shah & Hoeffner, 2002) The question type variable has been the most widely studied in the context of graph comprehension (e.g., Bright, Friel, & Lajoie, 1998; Carswell, 1992; Curcio, 1982; Curcio, 1987; Pereira Mendoza & Mellor, 1991; Wainer, 1992) In their review of the research on the effects of
76 graph comprehension, Friel, Curcio, and Bright (2001) characterized the point locating questions In the current study, interactions were found between groups and graph variables. There was a signi ficant difference between the students with dyslexia and typical readers for the comparison questions with dyslexic readers scoring less accurately. However, there was no significant difference between the two groups for point locating questions although t he accuracy scores for the dyslexic group were numerically lower. This finding suggests that graph comprehension difficulty in dyslexics deteriorates as the complexity of the stimulus increases. The effect of stimulus complexity has been consistently repor ted in children and adults with dyslexia (Fawcett & Nicolson, 2007; P ark, 2011) Furthermore, significant interactions between graph type (i.e., line, bar, horizontal bar graphs) and graphic display (i.e., single and double graphic displays) was found. This interaction suggests that when tasks involve both complex graphic stimuli and complex quest ions, individuals who have processing deficits, such as dyslexia, may be faced with much greater challenges than their peers who do not have a learning disability. Group Comparison o n Graph Viewing Time s Overall, the eye tracking data were consistent with the comprehension accuracy data Students with dyslexia needed significantly longer time than typical readers in completing graphic tasks. All participants took longer to interpret graphs associated with double patterns than graphs with single patterns and gr aphs presented with comparison questions than graph s with point locating questions. In addition, there were significant interactions between group and question t ype, group and graphic display, and question typ e and graphic disp lay. Dyslexic participants spent significantly
77 longer looking at graphs presented with comparison questions than graphs associated with point locating questions. Also, dyslexics spent longer time s looking at graphs with double graphic patterns than those w ith single graphic patterns. Again, these finding s underscores that level of complexity has a greater effect on the response times of individuals who have learning disabili ties than on typical learners. The interaction between question type and graphic dis play shows that when the complex stimuli were associated with complex questions, they were more likely to result in less accurate and slower responses. In exami ning the eye fixation times on graph comprehension tasks, visual display s were segmented into six areas of interest (i.e., x axis, y axis, patte rn, legend, question, answer). The e ye tracking data showed that the students with dyslexia fixated for significantly longer periods of time in the linguistic text region s (i.e., where quest ions and answers were located). The group difference was found both in the first pass time and the total viewing time. This finding suggests that the greatest time difference s between the groups in graph interpretation occurred while processing the text ar ea region s. Among the four graphic regions (x axis, y axis, pattern, legend), dyslexics spent longer time than typical readers in the pattern region only This longer period o f viewing time was only found in the total reading time not on the first pass. T he students with dyslexia needed a longer time to process the pattern after the initial viewing of that region. Impaired visual processing in dyslexia, in some form or another, has been suggested as a potential cause of dyslexia. Dyslexia was first descri bed in the literature in the 18 th century by Dr. Samuel Orton and referred to as wordblindness or strephosymbolia (Orton, 1928, 1943) Since then, many theories regarding the biological underpinnings of dyslexia have been posit ed For example, some researchers argue that there is a high correlation between visual analytical skills and reading achievement (Garzia, 2000; Grisham, Powers, & Riles, 2007; Stein, 2008) and
78 suggest th ree subtypes of dyslexia exist, the phonological processing deficit type, the visual processing deficit type, and the phonological and visual processing deficit type (Stein, Talcott, & Walsh, 2000) Others, however, argue that visual deficits are f ound only in small group of individuals who have dyslexia (Johannes, Kussmaul, Mnte, & Mangun, 1996; Victor, Conte, Burton, & Nass, 1993) and conclude that visual processing deficits are likely to be s ymptom s of linguistic deficiencies, rather than a cause of this disability (American Academy of Pediatrics, 2010) This study revealed that most of the between group differences were found in th e viewing time s associated with the text areas. Findings f rom Language a nd Cognition M easures performance on graph interpretation was more strongly related to e strongly related to their executive function skills. It is possible that the graph ic stimuli (including graphics and text) may have become overle arned (three types of question and graph were repeated 48 times respectively) for typical readers to the extent that they had fewer demands on verbal working memory. In contrast, the dyslexics may not have had this ad vantage and were required to re ly on working memory to a greater extent. correlated with all verbal working memory measures (i.e., digit forward, digit backward, and digit ordering subtests). Previous studies have consistently ding correlates with reduced verbal short term and working memory span. Griffiths & Snowling (2002) found that verbal memory span was the unique predictor of reading in 9 15 year old dyslexic children. Berninger and c olle a gues (2006) a nd Smith Spark and Fisk (2007) showed that verbal memory difficulties in dyslexia extend into adulthood and affect performa nce in reading and academic achievement Vocabulary was not ph interpretation because w ords
79 in this study were carefully cont rolled in terms of difficulty. The close association between ver bal working memory and text viewing time in dyslexic readers suggest s that, perhaps, it is not vocabulary per se, but the cognitive resources to actively maintain or store the meanings of the words in questions may be impaired in dyslexics. B lock recall subtest correlated significantly with graphic viewing times in both dyslexic and typical reader groups. Block recall measures the ability to analyze and synthesize the composite parts of visual patterns and to reproduce s demonstration (Meyler & Breznitz, 1998) Visual dimensions (i.e., shape, color, length, etc.) in a graph should be analyzed and synthesized to comprehend the graph and to answer to the related questions. Thus, significant correlation betwe en graphic viewing times and the block rec all test scores is predictable. In addition, the block recall correlated with text viewing time in typical readers but not in dyslexic readers It is currently not universally agreed that visuo spatial memory cont ribute s to verbal information processing. Hulme, Snowling, Caravolas, and Carroll (2005 ) found that visual memory did not play a significan t role in the reading and spelling of 153 elementary school children. Bayliss, Jarrold, and Gunn (2003) explored the relationship between visual working memory and reading and math performance in typical adults. The investigator s reported non significant correlation between the visual memory variabl e and reading and math ability. In contrast, oth er researchers have reported that visual memory is associated with reading as strongly as verbal memory. In their longitudinal study of pre reading children Meyler and Breznitz (1998) found that the influenc e of visual memory was stronger and more consistent than influence of verbal memory while both visual and verbal memory played an important role in the acquisition of reading. Visual memory predicted not only reading accurac y but also reading speed of seco nd grade children. Based on this result, they suggested that visual memory may be
80 more critical for reading speed, while v erbal memory may be important for later reading accuracy. In the current study, the correlation between block recall test scores and t e xt viewing time, not comprehension accuracy for the t ypical readers may support their argument. A mong all execution function test scores, Stroop word task correlated mostly highly with the graph viewing time. The Stroop task is a popular measurement of i nhibition (or resistance to interference) and suppression of automatic responses in order to process less automatic responses (Brocki & Bohlin, 2004) A diminished capacity fo r inhibition has been viewed as a key executive func tion deficit affecting reading skill (Smith Spark & Fisk, 2007) I n the present study, Stoop color word s cores correlated with text viewing time in both typical readers and dyslexic readers. The next largest correlation coefficient was found in the relationship between graphic viewing time and D igit Symbol S ubstitution s ubtest This task entails the ability to match a familiar stimulus (i.e., digit) with a novel stimulus (i.e., an abstract symbol that may look similar) in a timed condition (Venkatraman, et al ., 2011) Digit Symbol Substitution has some features in common with graph interpretation in that graphic task used involved making rapid connections between symbols and conceptual referents. Also, Digit symbol substitution was one of the only two executive function skills significantly correlated with graph viewing times in dyslexic readers. Even tho ugh it is not directly related to the research question, it is worth noting that d yslexic readers had significantly lower scores than typical readers on the receptive vocabulary measure and on all other measures of cognition In a previous study by Wisehea rt, Altmann, Park and Lombardino (2009) young adults with and without dyslexia did not differ in their scores on the Shipley vocabulary test. The average vocabu lary scores of dyslexic readers and typical readers in their experiment were 30.22 and 31.80, respectively, while in the current study, the
81 scores were 27.13 and 31.00, respectively. Thus, the scores of typical readers were similar between the two studies, but the scores for dyslexic readers in this study were lower than those in Wiseheart et al. (2009) It appears that dyslexic subjects in the current study had o verall more severe learning disabilities than the dyslexic subjects in the Wiseheart et al. (2009) Most of the dyslexic participants in this study were referred by the University of Florida Disability Resource Center because they were at risk of failure in their classes. G iven that the performance of the dyslexic group in this study was lower on the other cognitive measures (e.g., working memory scores) than in t he Wiseheart et al. (2009) study it may be that the present study included more severely impaired dyslexics. Another possible reason for the inconsistency betwe en the current d a ta and the Wiseheart et al. (2009) data is the dispersion of scores in the dyslexic readers. The difference of scores (i.e., subtracting the smallest score from the largest one) for typical reader s in the current study was 16 (the range wa s 37 and 21 ) while the difference for dyslexic reader s was 23 (the range was 37 and 14). T hree students in the dyslexic group had scores lower t han 19 (total score was 40), where as none in typical gro up had such low scores. O n all verbal working memory co gnitive measures, the dyslexic group in this study performed more poorly and slowly than their peers without dyslexia, supporting a considerable body of evidence of impaired verbal working memory in individuals who have dyslexia. (for a review see Snowling, 2000) Menghini, Finzi, Carlesimo, & Vicari (2011) summarized three hypotheses in an attempt to account for the association between poor verbal working memory and dyslexia: (1) an impairment of subvocal rehearsal mechanisms might be responsible for dyslexia (Baddeley, 2003) ; (2) A reduced speech rate may be associated with both impaired subvocal rehearsal system and reading defi cits (McDougall, Hulme, Ellis, & Monk, 1994) ; and
8 2 (3) an impairment of phonological store rather than subvocal rehearsal lies at the root of these verbal working memory deficits (Kibby, Marks, Morgan, & Long, 2004) Evidence is inconclusive regarding the involvement of visual working memory in dyslexia. Indeed, some studies have reported that visual working memory, in contrast to verbal working memory, is intact in dyslexic individuals. Jeffries and Everatt (2004) noted that the difference between dyslexic children and t ypical readers was found o n verbal working memory measures, but not o n visual working memory measures (using v is ual spatial sketch tasks). Similarly Kibby et al. (2004) reported that childr en with dyslexia have impaired phonological loop functions, but intact visual spatial sketch pad functions. In contrast, ot her researchers have reported impaired visual working memory in dyslexics. Smith Spark and Fisk (2007) compared performance of verbal and visual working memory measures in two groups of adults, dyslexics and age and IQ matched typical readers. They found t hat dyslexics were impaired on both verbal and visual tasks of simple and complex span. Based on these findings, some researchers suggest that the deficits found in individuals with dyslexia are do main general rather than domain specific (Cohen Mimran & Sapir, 2007) Based on processing speed data using linguistic and non linguistic stimuli presented in both the visual and motor modalities, Park (2011) found that visual working memory as well as verbal working memory is impaired in dyslexia, but the deficits are less severe for visual working memory than for verbal working memory. The difference between dyslexic and typical reader s was also apparent in the executive function tasks. Indeed, the effect sizes (i.e., a measure of the magnitude of the observed effect; Field, 2009) for the executive function s (partial 2 =.45) was larger than f or verbal working memory (partial 2 =.27) and visual working memory (partial 2 =.18). E xecutive function refers to cognitive abilities such as planning, sequencing, persistence, inhibition, and flexibility of
83 action, and is linked to the prefrontal corte x of the brain (Brocki & Bohlin, 2004) Smith Park and Fisk (2007) found that young adults with dyslexia showed significant deficits in executive function compared to typical readers, after controlling for working memory performance. They argued that the central executive impairments demonstrate d by dyslexics are a domain general processing problem, independent of verbal or visual working memory and that executive functions are the major cognitive factor s (1990) Dyslexic Automatization Deficit (DAD) hypothesis. The DAD hypothesis posits that dyslexics have impaired automatization abilities overall, which, in turn, place a burden on their attentional capacity for the task th ey are involved. Overall, Fawcett and Nicolson (1990; 2001) and Smith Spark and Fisk (2007; 2003) argued that ed in four types of skills: (1) complex skills requiring fluency; (2) time dependent skills; (3) multi modality skills; (4) vigilance tasks associated with concentration over time. Theoretical Implications There are several potentially important theoreti cal implications associated with the findings from this study. Due to the large number of analyses conducted, these implications are presented below: (1) G raph comprehension is influenced by various abilities the gra phical properties, and the specific requirements of the given task. In addition, these variables are closely related to each other and the influence of one variable is dependent upon the other ones. Therefore, it is important to consider the close interact ion among perceptual processes (e.g., interpreting the patterns/lines), conceptual processing (e.g., interpreting questions), and the when examining graph comprehension ; (2) As in many other studies, the present study showed that That is, the dyslexics needed longer
84 time to process not only text, but also pictures. However, it is worth noting that for the verbal stimuli (que stion and text), the difference between dyslexics and typical readers was larger than for the non verbal stimuli ; (3) apparent on the complex tasks (e.g., double horizontal bar graphs). This finding support s other data showing that dyslexics are more challenged when the stimulus is complex (Meyler & Breznitz, 2005; Nicolson & Fawcett, 1994; Stein & McAnally, 1995) ; (4) inter pretation correlated closely with several working memory measures including verbal and closely with several executive function skills ; and (5) The p resent study supports that dyslexia is charac terized by a range of cognitive deficits that are not limited to reading skills. Young adults with dyslexia performed significantly poorer than their peers without dyslexia not only on verbal working memory, but also on visual working memory and executive function tasks. Clinical Implications Findings from this study are very relevant to the education of students with dyslexia. Graphing is a fundamental part of the curriculum (Kramarski, 2004; National Council of Teachers of Mathematics, 2000) and the present study revealed that young adults with dyslexia perfo rmed significan tly less accurate ly and more slowly in interpreting graphs t han typical readers even when the stimuli were elementary level graphics. There is a need for testing nonverbal skills in dyslexics. I n a clinic setting, the evaluation of dyslexia appears to focus on language skills. However, this study is consistent with previous studies showing that can influence their processing of other stimuli such as graph s that include verbal and visu al variables. To my knowledge, there is currently no standardiz ed graphic skill measurement Some researchers and educators have used the Test of Graphing in Science (TOGS; McKenzie & Padilla, 1986) This 26 item multiple choice test is
85 developed to measure graphing skill The developers indicated that the test has content validity by showing that a panel of reviewers agreed over 94% of the time on assignment of items to objectives and 98% on the scoring o f items. The test is used to measure elementary level graph interpretation skills targeting upper elementary school or lower middle school students Another test, the Graphing Interpretation Skill Test (GIST; Svec, 1995) contains eleven multiple choice items This task was based on the TOGS by McKenzie & Padilla (1986) and was developed to measure high The results of this study also support the development of intervention programs that focus on visual information interp retation skills for dyslexics. T his study demonstrated that longer time does not guarantee better performance for dyslexics E ven when the dyslexic students spent significantly longer time to so lve the graphic questions, their accuracy scores were still significantly lower than their peers. The most common accommodation for students with learning disabilities is the provision of extra ti me to complete assignment s quizzes, and examinations. Data from this study indicate that while additional time may be a necessary accommodation, it is not a suff icient accommodation. Students with dyslexia c ould profit from more expli cit and intensive intervention in the area of graph interpretation. Parmar & Signer (2005) provided the following recommendation for te aching graphics to students: (1) S tuden ts need to explicitly learn information on the framework of the graph and its referents, such as how to label the axes, what scale is used for measure, and what ty pe of information is presented; (2) E ducators need to emphasize scale markings and making int er polations and extrapolations; (3) S tudents need to experience higher levels of thinking in interpreting graphic data, such as interpreting data from two or more variables concurrently o r comprehending trends; (4) S tudents should be encouraged to use the ma thematical terms related
86 ; and (5) E ducators need to encourage students to write stories based on graphs or construct graphs based on stories. In addition, Keller (2009) recommend ed that e ducators consider the importance of automaticity on graphing tasks as well as accuracy. A utomaticity should free up the graph viewer cognitive resources to be used for sophisticated graph interpretation. Keller (2009) also stated that students need to be encouraged to analyze graphed data based on their background knowledge or real life experience s instead o f interpreting data without using prior knowledge as a point of reference. Limitations a nd Future Directions This study demonstrated that and slower than typical readers by analyzing participants response accuracy and eye gaze data. Future research on graph comprehension should extend our knowledge base First ly even though this study showed differences between the two groups, the mean differences (especially on comprehension accur acy) were num erically small. The simplicity of the stimuli may played a role in this finding. T he stimuli were elementary grade level graphs and may have been too easy for college students. T he processing of simple graphs may not be comparable to the processing of more complex graphs (Ratwani & Trapton, 2008; Shah & Freedman, 2009) In meaningful contexts such as in social science, graphs can present rela tively complex data, requiring processing of information that is not explicitly presented (Shah, et al., 1999; Tricket & Trafton, 2006) and graphs may be more helpful when the data is more complex. Therefore, future resear ch should attempt to use more age appropriate stimuli Secondly, s was significantly poor er than typical readers, their comprehension of graphs might still be better than their comprehensio n of passage reading. To investigate this issue the same information in this
87 study could be provided in the form of text. A future study should be conducted to compare response accura cy and/or response time when information is presented in two formats, te xt and graphs. It is widely accepted that visual displays enhance the comprehension of data; however, recently, advantage of textual over graphical representation of information has been reported (Van Der Meulen, et al., 2010) Thus such research would further our understanding of graph processing. Third ly this study allowed participants to spend an unlimited amount of time to process the graphs. Future st udies should adopt the system paced time method in which the length of each scene is controlled by computer. T his methodology may reveal even clear er differences in the performance between dyslexics and typical readers. The comparison of system controlled and self controlled data processing has recently captured the attention of researchers, especially those involved in multimedia learning research (Schmidt Weigand, et al., 2010) Finally e xtending this research to include graph production w ould contribute to the understanding of the graph processing of dyslexics as well as typical readers. Leinhardt, Zaslavsky, & Stein (1990) pointed out that graph construction is differe nt from graph interpretation because i nterpretation comes from a give n piece of data while the construction requires the g eneration of new parts It is likely that dyslexics have more difficulty producing graph s than interpreting them bec ause construction relies on higher cognitive processing (Kramarski, 2004)
88 APPENDIX A RECRUITMENT FLYER Investigating graph comprehension of students Researchers: Linda J. Lombardino, PhD., Sunjung Kim, MA UF IRB: ____________ (For use through_______) Estimated duration of experiment: 3 hours In this study, you will see graphs on a computer screen and be asked to identify certain features to answer q position recorded. The experiment should take three hours or less. To participate in this study, you must be a native speaker of English, age 18 35, with normal or correct ed vision. Two groups of subjects needed: Group 1 Students who have no history of difficulty with reading or spelling Group 2 Students who have had reading or severe spelling difficulties since elementary school. These students may have no diagnosis or may have a diagnosis of a learning disability, reading disability, processing disability or dyslexia. you have a reading disability but suspect that you might, we will do a screening to find out! All participants with reading disability will be given a three page brief report of their test findings if they choose to be informed of their standardized test scores for documentation of their learning disability. or applicable courses, participants will be compensated in the form of 2 hours of research cred it. Sunjung Kim (email@example.com) Volunteers Needed To participate in a graph reading study
89 APPENDIX B INFORMED CONSENT LET TER FOR PARTICIPANTS Investigating graph comprehension in students Please read this consent document carefully before you decide to participate in this study. Purpose of the research project The purpose of this study is to examine how people comprehend graphs to answer comprehension questions. What you will be asked to do in the study If you agree to participate, then you will see graphs on a computer screen and b e asked to is monitored and their position recorded. The experiment should take three hours or less. We will provide you with short breaks between tasks, and you can ask for a longer break at any time. All testing will be carried out by the principal investigator, Dr. Lombardino, or trained research assistant s in her lab. Time required The study involves a one time visit that lasts about 3 hour. Risks and Benefits There are no known risks involved in this experiment, apart from those involved in everyday life. You will receive one half hour of course credit for each 30 minutes of participation (rounded up). There will be no direct benefit to you for parti cipating in this experiment (apart from the educational experience), although the experiment should help the scientific community and the public at large to gain an increased understanding of the human ability to use language. Compensation There will be a non research alternative for extra credit, if you choose the extra credit for compensation. Instructors preserve rights of determining the options for the course credit. If instructors want examiners to decide the alternative, you can read a short resear ch article related to graph comprehension and write a 1.5 2 page synopsis of it and receive the same amount of course credit. You can choose any articles related to graph comprehension research. The extra credit for participation is limited to no greater t han the equivalent of 2% of the student's overall grade in the course. Confidentiality
90 Your identity will be kept confidential to the extent provided by law. Your information and recordings will be assigned a code number for identification. The list conn ecting your name to completed and the data have been analyzed, the list will be destroyed. Your name will not be used in any report. Test results and recording s will be archived for research purposes, but your name will in no way be associated with these results. Voluntary participation Your participation in this study is completely voluntary. There is no penalty for not participating. If you choose not to pa rticipate, there are research and non research alternative options for this extra credit that will take relatively the same amount of time and effort as this opportunity. Right to withdraw from the study You have the right to withdraw from the study at an y time without penalty. Whom to contact if you have questions about the study Linda J. Lombardino, Ph.D. Department of School of Special Education, School Psychology, and Early Childhood Studies, University of Florida P.O. Box 117050; Gainesville, FL 3261 0 Office: (352) 273 4279; Email: firstname.lastname@example.org Whom to contact about your rights as a research participant. UFIRB Office, P.O. box 112250, University of Florida, Gainesville, FL32611 2250 352 392 0433. Agreement I have read the procedure described above. I voluntarily agree to participate in the procedure and have received a copy of this description. Participant: Date: Principal Investigator: Date: Experimenter: Date: KEEP THIS FORM TO CLAIM YOUR COURSE CREDIT Check your syllabus for details
91 APPENDIX C QUESTIONAIRE FORM Subject # Study Score Visit Group Tester 1. Age _____ 2. Gender _____ Female _____ Male 3. What is/are your majors (s)? _____ 4. Have you taken any math or statistics classes during your undergraduate and/or graduate studies? How many have you taken? Math _____ yes _____ no how many _____ Statistics _____ yes _____ no how many _____ Below are several statements concerning your experience using graphs. Please chose a number from 1 7 and write it next to each statement to indicate the frequency with which you perform each task. 1 2 3 4 5 6 7 Never Very often 1. ____ I need to interpret graphs as part of my job or field of study. 2. _____ I read graphs in the popular press (e.g., magazines, newspapers). 3. _____ I use a software package to produce graphs. 4. _____ I notice errors or misrepresentations in graphs presented in academic journals or the popular press (e.g., magazines, newspapers). 5. _____ When I look at a graph, I try to understand the main point the creator of the graph was trying to make. 6. _____ When I look at a graph, I try to identify the overall patterns or trends represented. 7. _____ When I look at a graph, I thin k about the likely reasons for the pattern(s) of data presented. Please chose a number from 1 7 and write it next to each statement describing your typical reaction to graphs. 1 2 3 4 5 6 7 Never Very often
92 8. ____ When I encounter a graph in a text, newspaper, or magazine, I tend to ignore it (skip it completely). 9. represents, but I do not study it in detail. 10. _____ I find graphs useful for rem embering information. 11. _____ I would prefer seeing a table of numbers rather than seeing a graph of the numbers. 12. _____ Graphs are generally a waste of space. 13. _____ Overall, on a scale of 1 to 6, how useful do you find graphs? Below are several questions con cerning your graph reading ability. Please chose a number from 1 7 and write it next to each statement to indicate the extent to which you agree or disagree with each statement. 1 2 3 4 5 6 7 Strongly disa gree Strong agree 14. ____ I am familiar with reading line graphs. 15. _____ I am familiar with reading bar graphs. 16. _____ I am familiar with reading pie charts. 17. _____ I am familiar with reading donut graphs. 18. _____ I am familiar with reading horizontal bar graphs 19. _____ Overall, on a scale of 1 to 7, how would you rate your ability to read graphs?
93 APPENDIX D SAMPLE OF SINGLE GRA PHIC DISPLAY GRAPH AND QUESTION Jackie loves her animals. How many cats does s he have? 4 5 Jackie loves his animals. Does she have more rabbits than dogs? No Yes Harry drinks a lot of things How many cups of teas did he drink? 5 8 Har ry drinks a lot of things. Did he drink more cups of water than cups of coffee? No Yes Alissa has to ride to work this week. How many train rides did s he take? 6 3 Alissa has to ride to work this week. Did s he take more bus rides than taxi rides? No Yes
94 APPENDIX E SAMPLE OF DOUBLE GRA PHIC DISPLAY GRAPH AND QUESTION Stella loves her animals. How many dogs does s he have? 9 7 Stella loves her animals. Does she have more cats than rabbits? No Yes Joe drinks a lot of things How many cups of water did he drink? 5 7 Joe drinks a lot of things. Did he drink more cups of tea than cups of coffee? No Yes Rachel has to ride to work this week. How many taxi rides did s he take? 9 6 Rachel has to ride to work this week. Did s he take more bus rides than train rides? No Yes
95 LIST OF REFERENCES Ackerman, P. T., & Dykman, R. A. (1993). Phonological processes, confrontational naming, and immediate memory in dyslexia. Journal of Learning Disabilities, 26 (9), 597 609. doi: 10.1177/002221949302600910 Agostino, A., Johnson, J., & Pascual Leone, J. (2010). Executive func tions underlying multiplicative reasoning: Problem type matters. Journal of Experimental Child Psychology, 105 (4), 286 305. doi: 10.1016/j.jecp.2009.09.006 American Academy of Pediatrics, S. o. O., Council on Children with Disabilities; American Academy of Ophthalmology; American Association for Pediatric Ophthalmology and Strabismus; American Association of Certified Orthoptists. (2010). Joint Statement Learning Disabilities, Dyslexia, and Vision. Pediatrics, 124 (2), 837 844. doi: 10.1542/peds.2009 1445 As hendorf, L., & McCaffrey, R. J. (2008). Exploring age related decline on the Wisconsin Card Sorting Test. The Clinical Neuropsychologist, 22 (2), 262 272. doi: 10.1080/13854040701218436 Baddeley, A. (2003). Working memory and language: an overview. Journal of Communication Disorders, 36 (3), 189 208. doi: 10.1016/S0021 9924(03)00019 4 Baddeley, A. D., & Hitch, G. J. (1974). Working memory. In G. A. Bower (Ed.), Recent advances in learning and motivation (Vol. 8, pp. 47 90). New York: Academic Press. Bayliss, D. M., Jarrold, C., Gunn, D. M., & Baddeley, A. D. (2003). The complexities of complex span: Explaining individual differences in working memory in children and adults. Journal of Experimental Psychology: General, 132 (1), 71 92. doi: 10.1037/0096 3445.132. 1.71. Beneventi, H., Tnnessen, F. E., Ersland, L., & Hugdahl, K. (2010). Working memory deficit in dyslexia : behavioral and fMRI evidence. International Journal of Neuroscience, 120 51 59. doi: 10.3109/00207450903275129 Berg, C. A., & Phillips, D. G. (1 994). An investigation of the relationship between logical thinking structures and the ability to construct and interpret line graphs. Journal of Research in Science Teaching, 31 (4), 323 344. Berninger, V. W., Abbott, R. D., Thomson, J., Wagner, R., Swans on, H. L., Wijsman, E. M., et al. (2006). Modeling phonological core deficits within a working memory architecture in children and adults with developmental dyslexia. Scientific Studies of Reading, 10 (2), 165 198. Berninger, V. W., Raskind, W. H., Richard s, T. L., Abbott, R. D., & Stock, P. (2008). A multidisciplinary approach to understanding developmental dyslexia within working memory architecture: Genotypes, phenotypes, brain, and instruction. Developmental Neuropsychology, 33 (6), 707 744. doi: 10.1080 /87565640802418662
96 Boden, C., & Brodeur, D. A. (1999). Visual processing of verbal and nonverbal stimuli in adolescents with reading disabilities. Journal of Learning Disabilities, 32 (1), 58 71. doi: 10.1177/002221949903200106 Box, G. E. P. (1949). A gener al distribution theory for a class of likelihood criteria. Biometrika, 36 317 346. Bright, G. W., Friel, S. N., & Lajoie, S. P. (1998). Graphical representations: Helping students interpret data Reflections on statistics: Learning, teaching, and assessme nt in Grades K??2. (pp. 63 88). Mahwah, NJ US: Lawrence Erlbaum Associates Publishers. Brocki, K. C., & Bohlin, G. (2004). Executive functions in children aged 6 to 13: A dimensional and developmental study. Developmental Neuropsychology, 26 (2), 571 593. doi: 10.1207/s15326942dn2602_3 Bronsnan, M., Demetre, J., Hamill, S., Robson, K., Shepherd, H., & Cody, G. (2002). Executive functioning in adults and children with developmental dyslexia. Neuropsychologia, 40 2144 2155. doi: 10.1016/S0028 3932(02)00046 5 Brunswick, N., McCrory, E., Price, C. J., Frith, C. D., & Frith, U. (1999). Explicitly and implicit processing of words and pseudowords by adult developmental dyslexics. Brain: A Journal of Neurology, 122 (10), 1901 1917. doi: 10.1093/brain/122.10.1901 Bus ch, R. M., Booth, J. E., McBride, A., Vanderploeg, R. D., Curtiss, G., & Duchnick, J. J. (2005). Role of Executive Functioning in Verbal and Visual Memory. Neuropsychology, 19 (2), 171 180. doi: 10.1037/0894 4126.96.36.199 Cain, K., Oakhill, J., & Bryant, P. (2004). Children's reading comprehension ability: concurrent prediction by working memory, verbal ability, and component skills. Journal of Educational Psychology, 96 (1), 31 42. doi: 10.1037/0022 06188.8.131.52 Carpenter, P. A., & Shah, P. (1998). A model o f the perceptual and conceptual processes in graph comprehension. Journal of Experimental Psychology: Applied, 4 (2), 75 100. doi: 10.1037/1076 898X.4.2.75 Carroll, P. J., Young, J. R., & Guertin, M. S. (1992). Visual analysis of cartoons: A view from the f ar side. In K. Rayner (Ed.), Eye movements and visual cognition: Scene perception and reading (pp. 444 461). New York: Springer Verlag. Carswell, C. M. (1992). Choosing specifiers: An evaluation of the basic tasks model of graphical perception. Human Facto rs, 34 (5), 535 554. Cohen Mimran, R., & Sapir, S. (2007). Deficits in working memory in young adults with reading disabilities. Journal of Communication Disorders, 40 168 183. doi: 10.1016/j.jcomdis.2006.06.006 Cohen, J. (1988). Statistical power analysi s for the behavioral sciences. NJ Erlbaum: Hillsdale.
97 Cohen, R. L., Netley, C., & Clarke, M. A. (1984). On the generality of the short term memory/reading ability relationship. Journal of Learning Disabilities, 17 (4), 218 221. doi: 10.1177/0022219484017004 06 Cook, M., Wiebe, E. N., & Carter, G. (2008). The influence of prior knowledge on viewing and interpreting graphics with macroscopic and molecular representations. Science Education, 92 (5), 848 867. doi: 10.1002/sce.20262 Cowan, N., Elliott, E. M., Sault s, J. S., Morey, C. C., Mattox, S., Hismjatullina, A., et al. (2005). On the capacity of attention: Its estimation and its role in working memory and cognitive aptitudes. Cognitive Psychology, 51 (1), 42 100. doi: 10.1016/j.cogpsych.2004.12.001 Cowles, W. & Kim, S. (2011). Comprehension and reading of canonical and non canonical sentences: an eye tracking study Paper presented at the Society for the Scientific Study of Reading St. Pete Beach, FL Curcio, F. R. (1982). The effect of prior knowledge, reading and mathematics achievement, and sex on comprehending mathematical relationships expressed in graphs. 42, ProQuest Information & Learning, US. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&db=psyh&AN=1982 74019 001&site=ehost live Curcio, F. R. (1987). Comprehension of mathematical relationships expressed in graphs. Journal for Research in Mathematics Education, 18 (5), 382 393. d oi: 10.2307/749086 de Koning, B. B., Tabbers, H. K., Rikers, R. M. J. P., & Paas, F. (2010). Attention guidance in learning from a complex animation: Seeing is understanding? Learning and Instruction, 20 (2), 111 122. doi: 10.1016/j.learninstruc.2009.02.010 De Luca, M., Borrelli, M., Judica, A., Spinelli, D., & Zoccolotti, P. (2002). Reading words and pseudowords: An eye movement study of developmental dyslexia. Brain and Language, 80 617 626. doi: 10.1006/brln.2001.2637 Dexter, D. D., Park, Y. J., & Hughes C. A. (2011). A meta analytic review of graphic organizers and science instruction for adolescents with learning disabilities: Implications for the intermediate and secondary science classroom. Learning Disabilities Research & Practice, 26 (4), 204 213. d oi: 0.1111/j.1540 5826.2011.00341.x Dockett, S., Whitton, D., & Perry, B. (2003). Children starting school: Images from picture storybooks. Paper presented at the International Education Research Conference Auckland, New Zealand. Duesbery, L., Werblow, J., & Yovanoff, P. (2011). Graphical literacy moderates the interaction of decorative dimensionality and cognitive demand in computer based graph comprehension. Journal of Educational Computing Research, 45 (1), 75 93.
98 Duff, S. C., & Logie, R. H. (2001). Proc essing and storage in working memory span. The Quarterly Journal of Experimental Psychology A: Human Experimental Psychology, 54A (1), 31 48. doi: 10.1080/02724980042000011 Eden, G. F., Stein, J. F., Wood, H. M., & Wood, F. B. (1994). Differences in eye mov ements and reading problems in dyslexia and normal children. Vision Research, 34 (10), 1345 1358. Engle, R. W., Tuholski, S. W., Laughlin, J. E., & Conway, A. R. A. (1999). Working memory, short term memory, and general fluid intelligence: A latent variabl e approach. Journal of Experimental Psychology: General, 128 (3), 309 331. doi: 10.1037/0096 34184.108.40.2069 Eurich, A. C. (1933a). Fourth and fifth grade standards for photographic eye movement records. The Pedagogical Seminary and Journal of Genetic Psycho logy, 43 466 471. Eurich, A. C. (1933b). The photographic eye movement records of successful and unsuccessful college students. Journal of Applied Psychology, 17 (5), 604 613. doi: 10.1037/h0071906 Fawcett, A. J., & Nicolson, R. I. (2007). Dyslexia, learn ing, and pedagogical neuroscience. Developmental Medicine & Child Neurology, 49 (4), 306 311. doi: 10.1111/j.1469 8749.2007.00306.x Feil, A. J. (2010). Attention and encoding in physics learning and problem solving. 71, ProQuest Information & Learning, US. Retrieved from http://search.ebscohost.com/login.aspx?direct=true&db=psyh&AN=2010 99130 233&site=ehost live Field, A. (2009). Discovering statistics using SPSS (Vol. 3). London: SAGE Publications LtD. Filik, R., Paterson, K. B., & Liversedge, S. P. (2005). Parsing with focus particles in context: Eye movements during the processing of relative clause ambiguities. Journal of Memory and Langua ge, 53 473 495. Freedman, E. G., & Shah, P. (2002). Toward a Model of Knowledge Based Graph Comprehension Diagrammatic Representation and Inference (Vol. 2317, pp. 59 141): Springer Berlin / Heidelberg. French, A., Macedo, M., Poulsen, J., Waterson, T., & Yu, A. (2008). Multivariate Analysis of Variance (MANOVA) Retrieved 06 04, 2008, from http://userwww.sfsu.edu/~efc/classes/biol710/manova/manovanewest.htm Friel, S. N., C urcio, F. R., & Bright, G. W. (2001). Making sense of graphs: Critical factors influencing comprehension and instructional implications. Journal for Research in Mathematics Education, 32 (2), 124 158. Fry, E. (1983). A theory of graphs for reading comprehe nsion and writing communication New Brunswick: NJ: Rutgers University.
99 Gang, M., & Siegel, L. S. (2002). Sound symbol learning in children with dyslexia. Journal of Learning Disabilities, 35 (2), 137 157. doi: 10.1177/002221940203500205 Garzia. (2000). Care of the patient with learning related vision problems. Gattis, M., & Holyoak, K. J. (1996). Mapping conceptual to spatial relations in visual reasoning. Journal of Experimental Psychology: Learning, Memory, and Cognition, 22 (1), 231 23 9. doi: 10.1037/0278 73220.127.116.11 Gentner, D., & Goldin Meadow, S. (2003). Language in mind: Advances in the study of language and thought Cambridge, MA US: MIT Press. Golden, C. J. (1978). Stroop Color and Word Test: A Manual for Clinical and Experiment al Uses Chicago, Illinois: Skoelting. Graesser, A. C., Swamer, S. S., Baggett, W. B., Sell, M. A., & Britton, B. K. (1996). New models of deep comprehension Models of understanding text. (pp. 1 32). Hillsdale, NJ England: Lawrence Erlbaum Associates, Inc. Grant, D. A., Jones, O. R., & Tallantis, B. (1949). The relative difficulty of the number, form, and color concepts of a Weigl type problem. Journal of Experimental Psychology, 39 (4), 552 557. doi: 10.1037/h0062126 Griffiths, Y. M., & Snowling, M. J. (200 2). Predictors of exception word and nonword reading in dyslexic chidlren: the Severity hypothesis. Journal of Educational Psychology, 94 (1), 34 43. Grisham, D., Powers, M., & Riles, P. (2007). Visual skills of poor readers in high school. Optometry Jou rnal of the American Optometric Association, 78 (10), 542 549. doi: 10.1016/j.optm.2007.02.017 Guri Rozenblit, S. (1989). Effects of a tree diagram on students' comprehension of main ideas in an expository text with multiple themes. Reading research quarte rly, 24 (2), 236 247. doi: 10.2307/747866 Guthrie, J. T., Weber, S., & Kimmerly, N. (1993). Searching documents: Cognitive processes and deficits in understanding graphs, tables, and illustrations. Contemporary Educational Psychology, 18 (2), 186 221. doi: 1 0.1006/ceps.1993.1017 Habel, C., & Acarturk, C. (2009). Eye tracking evidence for multimodal language graphics comprehension. Retrieved 08.01.2010, from Northern European Association for Language Technology (NEALT) http://hdl.handle.net/10062/9825 Hegarty, M., Meyer, B., Narayanan, N., Freedman, E., & Shah, P. (2002). Toward a Model of Knowledge Based Graph Comprehension Diagrammatic Representation and Inference (Vol. 2317, pp. 59 141): Springer Berlin / Heide lberg.
100 Hegarty, M., & Steinhoff, K. (1997). Individual differences in use of diagrams as external memory in mechanical reasoning. Learning and Individual Differences, 9 (1), 19 42. Henderson, J. M., Weeks, P. A., Jr., & Hollingworth, A. (1999). The effects of semantic consistency on eye movements during complex scene viewing. Journal of Experimental Psychology: Human Perception and Performance, 25 (1), 210 228. doi: 10.1037/0096 1518.104.22.168 Hirose, Y., & Hancock, P. J. B. (2007). Equally attending but stil l not seeing: An eye tracking study of change detection in own and other race faces. Visual Cognition, 15 (6), 647 660. doi: 10.1080/13506280601069578 Holsanova, J., Holmberg, N., & Holmqvist, K. (2009). Reading information graphics: The role of spatial co ntiguity and dual attentional guidance. Applied Cognitive Psychology, 23 (9), 1215 1226. doi: 10.1002/acp.1525 Hoppe, C. D., Muller, U. D., Werheid, K. D., Thone, A. D., & von Cramon, Y. D. (2000). Digit ordering test: Clinical, psychometric, and experiment al evaluation of a verbal working memory test. Clinical Neurophysiology, 14 651 665. Huang, W., Hong, S., & Eades, P. (2006). Predicting graph reading performance: A cognitive approach Paper presented at the Asia Pacific Symposium on Information Visuali zation, Tokyo, Japan. Hulme, C., Snowling, M., Caravolas, M., & Carroll, J. (2005 ). Phonological skills are (probably) one cause of success in learning to read: a comment on Castles and Coltheart. Scientific Studies of Reading, 9 (4), 351 365. Hutzler, F ., & Wimmer, H. (2004). Eye movements of dyslexic children when reading in a regular orthography. Brain and Language, 89 (1), 235 242. doi: 10.1016/S0093 934X(03)00401 2 Hyn, J. (2010). The use of eye movements in the study of multimedia learning. Learning and Instruction, 20 (2), 172 176. doi: 10.1016/j.learninstruc.2009.02.013 Jarodzka, H., Scheiter, K., Gerjets, P., & van Gog, T. (2010). In the eyes of the beholder: How experts and novices interpret dynamic stimuli. Learning and Instruction, 20 (2), 146 154. doi: 10.1016/j.learninstruc.2009.02.019 Jeffries, S., & Everatt, J. (2004). Working memory: its role in dyslexia and other specific learning difficulties. Dyslexia: An International Journal of Research and Practice, 10 (3), 196 214. doi: 10.1 002/dys.278 Johannes, S., Kussmaul, C. L., Mnte, T. F., & Mangun, G. R. (1996). Developmental dyslexia: Passive visual stimulation provides no evidence for a magnocellular processing defect. Neuropsychologia, 34 (11), 1123 1127. doi: 10.1016/0028 3932(96)0 0026 7
101 Jones, M. W., Kelly, M. L., & Corley, M. (2007). Adult dyslexic readers do not demonstrate regularity effects in sentence processing: evidence from eye movements Reading and Writing, 20 (9), 933 943. Jones, M. W., Obregn, M., Kelly, M. L., & Branig an, H. P. (2008). Elucidating the component processes involved in dyslexic and non dyslexic reading fluency: An eye tracking study. Cognition, 109 (3), 389 407. doi: 10.1016/j.cognition.2008.10.005 Jorm, A. F. (1983). Specific reading retardation and workin g memory: A review. British Journal of Psychology, 74 (3), 311 342. Just, M. A., & Carpenter, P. A. (1980). A theory of reading: From eye fixations to comprehension. Psychological Review, 87 (4), 329 354. doi: 10.1037/0033 295X.87.4.329 Just, M. A., Carpent er, P. A., & Keller, T. A. (1996). The capacity theory of comprehension: new frontiers of evidence and arguments. Psychological Review, 103 (4), 773 780. Kail, R., Hall, L. K., & Caskey, B. J. (1999). Processing speed, exposure to print, and naming speed. Applied Psycholinguistics, 20 (2), 303 314. doi: 10.1017/S0142716499002076 Katzir, T., Kim, Y., Wolf, M., Kennedy, B., Lovett, M. W., & Morris, R. (2006). The relationship of spelling recognition, RAN, and phonological awarenes to reading skills in older po or readers and younger reading matched controls. Reading and Writing, 19 (8), 845 872. Keating, G. D. (2009). Sensitivity to violations of gender agreement in native and nonnative Spanish: An eye movement investigation. Language Learning, 59 (3), 503 535. d oi: 10.1111/j.1467 9922.2009.00516.x Keller, S. K. A. (2009). Levels of line graph question interpretation with intermediate elementary students of varying scientific and mathematical knowledge and ability: A think aloud study. 70, ProQuest Information & L earning, US. Retrieved from https://search.ebscohost.com/login.aspx?direct=true&db=psyh&AN=2009 99130 010&site=ehost live Kibby, M. Y., Marks, W., Morgan, S., & Long, C. J. (2004). Specific Impairment in Developmental Reading Disabilities: A Working Memo ry Approach. Journal of Learning Disabilities, 37 (4), 349 363. doi: 10.1177/00222194040370040601 Kim, A. H., Vaughn, S., Wanzek, J., & Wei, S. (2004). Graphic organizers and their effects on the reading comprehension of students with LD: A synthesis of res earch. Journal of Learning Disabilities, 37 (2), 105 118. doi: 10.1177/00222194040370020201 Kosslyn, S. M. (1989). Understanding charts and graphs. Applied Cognitive Psychology, 3 (3), 185 225. doi: 10.1002/acp.2350030302
102 Kramarski, B. (2004). Making sense o f graphs: Does metacognitive instruction make a difference on students' mathematical conceptions and alternative conceptions? Learning and Instruction, 14 (6), 593 619. doi: 10.1016/j.learninstruc.2004.09.003 Langrall, C., Nisbet, S., Mooney, E., & Jansem, S. (2011). The role of context expertise when comparing data. Mathematical Thinking and Learning, 13 (1 2), 47 67. doi: 10.1080/10986065.2011.538620 Leinhardt, G., Zaslavsky, O., & Stein, M. K. (1990). Functions, Graphs, and Graphing: Tasks, Learning, and T eaching. Review of Educational Research, 60 (1), 1 64. doi: 10.2307/1170224 Levene, H. (1953). Genetic equilibrium when more than oneecological niche is available. American Naturalist, 87 331 333. Levene, H. (1960). Robust testes for equality of variances In I. Olkin (Ed.), Contributions to Probability and Statistics Palo Alto, CA: Stanford University Press. Levine, H. G., McGuire, C. H., & Nattress, L. W., Jr. (1970). The Validity of Multiple Choice Achievement Tests as Measures of Competence in Medicin e. American Educational Research Journal, 7 (1), 69 82. Levis, H. W., & Lentz, R. (1982). Effects of text illustrations: A review of research. Educational Communication and Technology, 30 195 232. Lohse, G. L. (1993). A cognitive model for understanding graphical perception. Human Computer Interaction, 8 (4), 353 388. doi: 10.1207/s15327051hci0804_3 Lyon, G. R., Shaywitz, S. E., & Shaywitz, B. A. (2003). A definition of dyslexia. Annals of Dyslexia, 53 1 14. doi: 10.1007/s11881 003 0001 9 Martin, R. C., & Allen, C. M. (2008). A disorder of executive function and its role in language processing. Seminars in Speech and Language, 29 (3), 201 210. Maunchly, J. W. (1940). Significance test for sphericity of a normal n variate distribution. The Annals of Mathematical Statistics, 11 204 209. Mayer, R. E. (1993). Comprehension of graphics in texts: An overview. Learning and Instruction, 3 (3), 239 245. doi: 10.1016/0959 4752(93)90007 M Mayer, R. E. (2010). U nique contributions of eye tracking research to the study of learning with graphics. Learning and Instruction, 20 (2), 167 171. doi: 10.1016/j.learninstruc.2009.02.012 McCabe, D. P., Roediger, H. L., III, McDaniel, M. A., Balota, D. A., & Hambrick, D. Z. (2010). The relationship between working memory capacity and executive functioning: Evidence for a common executive attention construct. Neuropsychology, 24 (2), 222 243. doi: 10.1037/a0017619
103 McDougall, S., Hulme, C., Ellis, A., & Monk, A. (1994). Learning to read: The role of short term memory and phonological skills. Journal of Experimental Child Psychology, 58 (1), 112 133. doi: 10.1006/jecp.1994.1028 McDougall, S. J. P., & Donohoe, R. (2002). Reading ability and memory span: Long term me mory contributions to span for good and poor readers. Reading and Writing, 15 (3 4), 359 387. doi: 10.1023/A:1015224830778 McKenzie, D. L., & Padilla, M. J. (1986). The construction and validation of the Test of Graphing in Science (TOGS). Journal of Resear ch in Science Teaching, 23 (7), 571 579. doi: 10.1002/tea.3660230702 Menghini, D., Finzi, A., Carlesimo, G. A., & Vicari, S. (2011). Working memory impairment in children with developmental dyslexia: Is it just a phonological deficity? Developmental Neurops ychology, 36 (2), 199 213. doi: 10.1080/87565641.2010.549868 Meyler, A., & Breznitz, Z. (1998). Developmental associations between verbal and visual short term memory and the acquisition of decoding skill. Reading and Writing, 10 (6), 519 540. doi: 10.1023/A :1007915517339 Meyler, A., & Breznitz, Z. (2005). Visual, Auditory and Cross Modal Processing of Linguistic and Nonlinguistic Temporal Patterns among Adult Dyslexic Readers. Dyslexia: An International Journal of Research and Practice, 11 (2), 93 115. doi: 0 .1002/dys.294 Miyake, A., Friedman, N. P., Emerson, M. J., Witzki, A. H., & Howerter, A. (2000). The unity and diversity of executive functions and their contributions to complex 'frontal lobe' tasks: A latent variable analysis. Cognitive Psychology, 41 (1) 49 100. doi: 10.1006/cogp.1999.0734 Miyake, A., Friedman, N. P., Rettinger, D. A., Shah, P., & Hegarty, M. (2001). How are visuospatial working memory, executive functioning, and spatial abilities related? A latent variable analysis. Journal of Experimen tal Psychology: General, 130 (4), 621 640. doi: 10.1037/0096 3422.214.171.1241 Monteiro, C., & Ainley, J. (2004). Interpretation of media graphs and critical sense: implications for teaching and teachers Paper presented at the the International Conference of Mathematics Educators in Topic Study Group (TSG 11): Research and development in the teaching and learning of probability and statistics, Copenhagen. Murphy, S. (2005). Accessibility of graphics in technical documentation for the cognitive and visually im paired Paper presented at the Proceedings of the 23rd annual international conference on Design of communication: documenting \ & designing for pervasive information, Coventry, United Kingdom. Murray, T. S., Kirsch, I. S., & Jenkins, L. B. (1997). Adu lt literacy in OECD countries: A technical report for the first international adult literacy survey Washington, DC: U.S. Government Printing Office.
104 NCTM. (2000). Principles and standards for school mathematics. Paper presented at the National Council of Teachers of Mathematics, Reston, VA. Nesbit, J. C., & Adesope, O. O. (2006). Learning with concept and knowledge maps: A meta analysis. Review of Educational Research, 76 (3), 413 448. doi: 10.3102/00346543076003413 Nicolson, R. I., & Fawcett, A. J. (1990) Automaticity: A new framework for dyslexia research? Cognition, 35 (2), 159 182. doi: 10.1016/0010 0277(90)90013 A Nicolson, R. I., & Fawcett, A. J. (1994). Reaction times and dyslexia. The Quarterly Journal of Experimental Psychology A: Human Experimenta l Psychology, 47A (1), 29 48. Nicolson, R. I., Fawcett, A. J., & Dean, P. (2001). Developmental dyslexia: The cerebellar deficit hypothesis. Trends in Neurosciences, 24 (9), 508 511. doi: 10.1016/S0166 2236(00)01896 8 OECD. (1995). Literacy, economy, and society: Results of the first international adult literacy survey Ottawa: Statistics Canada Olson, R. K., Conneis, F. A., & Rack, J. P. (1991). Eye movements in dyslexia and normal readers. In J. F. Stein (Ed.), Vision and visual dy slexia London: Macmillan. Orton, S. T. (1928). Specific reading disability? trephosymbolia. Journal of the American Medical Association, 90 1095 1099. Orton, S. T. (1943). Visual functions in strephosymbolia. Archives of Ophthalmology (Chicago), 30 707 717. Palmar, S. E. (1978). Fundamental aspects of cognitive representation. In E. Rosch & B. B. Lloyd (Eds.), Cognition and categorization (pp. 492 546). Hillsdale, NJ: Lawrence Erlbaum. Papafragou, A., Carruthers, P., Laurence, S., & Stich, S. (2008). Sp ace and the language cognition interface The innate mind Volume 3: Foundations and the future. (pp. 272 289). New York, NY US: Oxford University Press. Park, H. (2011). Processing speed deficits in young adults with developmental dyslexia [electronic resou rce] / by Heeyoung Park [Gainesville, Fla: University of Florida. Parmar, R. S., & Signer, B. R. (2005). Sources of Error in Constructing and Interpreting Graphs: A Study of Fourth and Fifth Grade Students with LD. Journal of Learning Disabilities, 38 (3) 250 261. doi: 10.1177/00222194050380030601 Pavlidis, G. T. (1981). Do eye movements hold the key to dyslexia? Neuropsychologia, 19 (1), 57 64. doi: 10.1016/0028 3932(81)90044 0
105 Peebles, D., & Cheng, P. C. H. (2003). Modeling the effect of task and graphic al representation on response latency in a graph reading task. Human Factors, 45 (1), 28 46. doi: 10.1518/hfes.126.96.36.199225 Pennington, B. F., Cardoso Martins, C., Green, P. A., & Lefly, D. L. (2001). Comparing the phonological and double deficit hypothese s for developmental dyslexia. Reading and Writing, 14 (7 8), 707 755. doi: 10.1023/A:1012239018038 Pereira Mendoza, L., & Mellor, J. (1991). Students'c oncepts of barg raph Some preliminary findings Paper presented at the the third international conference on teaching statistics Voorburg,The Netherlands. Pickering, S. J., & Gathercole, S. E. (2001). Working Memory Test Battery for Children (WMTB C) Manual London: The Psychological Corporation. Pinker, S., & Freedle, R. (1990). A theory of graph comprehensio n Artificial intelligence and the future of testing. (pp. 73 126). Hillsdale, NJ England: Lawrence Erlbaum Associates, Inc. Prado, C., Dubois, M., & Valdois, S. (2007). The eye movements of dyslexic children during reading and visual search: Impact of the visual attention span. Vision Research, 47 (19), 2521 2530. doi: 10.1016/j.visres.2007.06.001 Ratwani, R. M., Trafton, J. G., & Boehm Davis, D. A. (2008). Thinking graphically: Connecting vision and cognition during graph comprehension. Journal of Experimen tal Psychology: Applied, 14 (1), 36 49. doi: 10.1037/1076 898X.14.1.36 Ratwani, R. M., & Trapton, J. G. (2008). Shedding light on the graph schema: Perceptual features versus invariant structure. Psychonomic Bulletin & Review, 15 (4), 757 762. doi: 10.3758/P BR.15.4.757 Rayner, K. (1998). Eye movement in reading and information processing: 20 years of research. Psychological Bulletin, 124 (3), 372 422. doi: 10.1037/0033 2909.124.3.372 Rayner, K., & Slattery, T. J. (2009). Eye movements and moment to moment comp rehension processes in reading. In R. K. Wagner, C. Schatschneider & C. Phythian Sence (Eds.), Beyond decoding: The behavioral and biological foundations of reading comprehension (pp. 27 45). New York, NY, US: Guildford Press. Reitan, R. M., & Wolfson, D. (1993). The Halstead Reitan Neuropsychological Test Battery: Theory and clinical interpretation Tucson, AZ: Neuropschology Press Renshaw, J. A., Finlay, J. E., Tyfa, D., & Ward, R. D. (2004). Understanding visual influence in graph design through temporal and spatial eye movement characteristics. Interacting with Computers, 16 (3), 557 578. doi: 10.1016/j.intcom.2004.03.001 Richardson, D. C., Dale, R., Spivey, M. J., Gonzal ez Marquez, M., Mittelberg, I., & Coulson, S. (2007). Eye movements in language and cognition: A brief introduction Methods in
106 cognitive linguistics. (pp. 323 344). Amsterdam Netherlands: John Benjamins Publishing Company. Schmidt Weigand, F., Kohnert, A., & Glowalla, U. (2010). A closer look at split visual attention in system and self paced instruction in multimedia learning. Learning and Instruction, 20 (2), 100 110. doi: 10.1016/j.learninstruc.2009.02.011 Sesma, H. W., Mahone, E. M., Levine, T., Eason, S. H., & Cutting, L. E. (2009). The contribution of executive skills to reading comprehension. Child Neuropsychology, 15 (3), 232 246. doi: 10.1080/09297040802220029 Shah, P., & Carpenter, P. A. (1995). Conceptual limitations in comprehending line graphs. Journal of Experimental Psychology: General, 124 (1), 43 61. doi: 10.1037/0096 34188.8.131.52 Shah, P., & Freedman, E. G. (2003). Visuospatial cognition in electronic learning. Journal of Educational Computing Research, 29 (3), 315 324. doi: 10.2190/QYVJ Q59L VE7C EHUV Shah, P., & Freedman, E. G. (2009). Bar and Line Graph Comprehension: An Interaction of Top Down and Bottom Up Processes. Topics in Cognitive Science no no. doi: 10.1111/j.1756 8765.2009.01066.x Shah, P., Hegarty, M., & Mayer, R. E. (1999). Gra phs as aids to knowledge construction: Signaling techniques for guiding the process of graph comprehension. Journal of Educational Psychology, 91 (4), 690 702. doi: 10.1037/0022 06184.108.40.2060 Shah, P., & Hoeffner, J. (2002). Review of graph comprehension re search: Implications for instruction. Educational Psychology Review, 14 (1), 47 69. doi: 10.1023/A:1013180410169 Shaywitz, S. E., & Shaywitz, B. A. (2005). Dyslexia (specific reading disability). Biological Psychology, 57 1301 1309. doi: 10.1016/j.biopsych .2005.01.043 Shelton, J. T., Elliott, E. M., Matthews, R. A., Hill, B. D., & Gouvier, W. D. (2010). The relationships of working memory, secondary memory, and general fluid intelligence: Working memory is special. Journal of Experimental Psychology: Learni ng, Memory, and Cognition, 36 (3), 813 820. doi: 10.1037/a0019046.W Simkin, D., & Hastie, R. (1987). An Information processing analysis of graph perception. Journal of the American Statistical Association 82 (398), 454 465. Simkin, D. K. (1988). Towards an information processing view of graph comprehension. 48, ProQuest Information & Learning, US. Retrieved from https://search.ebscohost.com/login.aspx?direct=true&db=psyh&AN=1989 50338 001&site=ehost live
107 Smith Spark, J. H., & Fisk, J. E. (2007). Working me mory functioning in developmental dyslexia. Memory, 15 (1), 34 56. doi: 10.1080/09658210601043384 Smith Spark, J. H., Fisk, J. E., Fawcett, A. J., & Nicolson, R. I. (2003). Investigating the central executive in adult dyslexics: Evidence from phonological a nd visuospatial working memory performance. European Journal of Cognitive Psychology, 15 (4), 567. doi: 10.1080/09541440340000024 Snowling, M. J. (2000). Dyslexia: A cognitive developmental perspective Reading and writing disorders in different orthographic systems. (pp. 1 23). Oxford: Blackwell. Solan, H. A., Feldman, J., & Tujak, L. (1995). Developing visual and reading efficiency in older adults. Optometry and Vision Science, 72 139 145. Spivey, M. J., Tanenhous, M. K., & Ederhard, K. M. (2002). Eye movements and spoken language comprehension: Effects of visual context on syntactic ambiguity resolution. Cognitive Psychology, 45 447 481. Stanley, G. (1994). Eye movements in dyslexic and normal children. In J. Ygge & G. Le nnerstrand (Eds.), Eye movements In reading Oxford: Pergamon Press. Stein, J. (2008). The neurobiological basis of Dyslexia. In G. Reid, A. J. Fawcett, F. Manis & L. S. Siegel (Eds.), The SAGE handbook of dyslexia Los Angeles: SAGE publications Stein, J. Talcott, J., & Walsh, V. (2000). Controversy about the visual magnocellular deficit in developmental dyslexics. Trends in Cognitive Sciences, 4 (6), 209 211. doi: 10.1016/S1364 6613(00)01484 4 Stein, J. F., & McAnally, K. (1995). Auditory temporal processing in developmental dyslexics. The Irish Journal of Psychology, 16 (3), 220 228. Svec, M. T. (1995). Effect of Micro Computer Based Laboratory on Graphing Interpretation Skills and Understanding of Motion. Pa per presented at the the annual meeting of the National Association of Research in Science Teaching, San Francisco, CA. Swanson, H. L. (1999). Reading comprehension and working memory in learning disabled readers: is the phonological loop more important th an the executive system? Journal of Experimental Child Psychology, 72 (1), 1 31. doi: 10.1006/jecp.1998.2477 Tanabe, A., & Osaka, N. (2009). Picture span test: Measuring visual working memory capacity involved in remembering and comprehension. Behavior Rese arch Methods, 41 (2), 309 317. doi: 10.3758/BRM.41.2.309. Torgesen, Wagner, R., & Rashotte, C. (1999). Test of Word Reading Efficiency Austin, TX: Pro ED.
108 Tricket, S. B., & Trafton, J. G. (2006). Toward a comprehensive model of graph comprehension: Making t he case for spatial cognition. In R. Cox & N. Swoboda (Eds.), Diagrammatic Representation and Inference (pp. 286 300): D. Barker Plummer. Tversky, B., & Schiano, D. J. (1989). Perceptual and conceptual factors in distortions in memory for graphs and maps. Journal of Experimental Psychology: General, 118 (4), 387 398. doi: 10.1037/0096 34220.127.116.117. Underwood, G., Chapman, P., Brocklehurst, N., Underwood, J., & Crundall, D. (2003). Visual attention while driving: Sequences of eye fixations made by experienc ed and novice drivers. Ergonomics, 46 (6), 629 646. doi: 10.1080/0014013031000090116 Underwood, G., Jebbett, L., & Roberts, K. (2004). Inspecting pictures for information to verify a sentence: Eye movements in general encoding and in focused search. The Qua rterly Journal of Experimental Psychology A: Human Experimental Psychology, 57A (1), 165 182. doi: 10.1080/02724980343000189 van Bruggen, J. M., Kirschner, P. A., & Jochems, W. (2002). External representation of argumentation in CSCL and the management of c ognitive load. Learning and Instruction, 12 (1), 121 138. doi: 10.1016/S0959 4752(01)00019 6 Van Der Meulen, M., Logie, R. H., Freer, Y., Sykes, C., McIntosh, N., & Hunter, J. (2010). When a graph is poorer than 100 words: A comparison of computerised natur al language generation, human generated descriptions and graphical displays in neonatal intensive care. Applied Cognitive Psychology, 24 (1), 77 89. doi: 10.1002/acp.1545 van Gog, T., & Scheiter, K. (2010). Eye tracking as a tool to study and enhance multim edia learning. Learning and Instruction, 20 (2), 95 99. doi: 10.1016/j.learninstruc.2009.02.009 Vellutino, F. R., Fletcher, J. M., Snowling, M. J., & Scanlon, D. M. (2004). Specific reading disability (dyslexia): what have we learned in the past four decad es? Journal of child Psychology and Psychiatry, 45 (1), 2 40. Venkatraman, V. K., Aizenstein, H. J., Newman, A. B., Yaffe, K., Harris, T., Kritchevsky, S., et al. (2011). Lower digit symbol substitution score in the oldest old is related to magnetization t ransfer and diffusion tensor imaging of the white matter. Frontiers in Aging Neuroscience, 3 doi: 10.3389/fnagi.2011.00011. Victor, J. D., Conte, M. M., Burton, L., & Nass, R. D. (1993). Visual evoked potentials in dyslexics and normals: Failure to find a difference in transient or steady state responses. Visual Neuroscience, 10 (5), 939 946. doi: 10.1017/S0952523800006155 Wagner, Torgesen, J., & Rashotte, C. (1999). Comprehensive Test of Phonological Processing Austin, TX: PRO ED. Wainer, H. (1992). Unders tanding Graphs and Tables. Educational Researcher, 21 (1), 14 23.
109 Waters, G. S., & Caplan, D. (1996). The measurement of verbal working memory capacity and its relation to reading comprehension. The Quarterly Journal of Experimental Psychology A: Human Exp erimental Psychology, 49A (1), 51 79. doi: 10.1080/027249896392801 Wechsler, D. (1987). Wechsler memory scale San Antonio: Psychological Corporation. Wechsler, D. (1997a). WAIS III/WMS III technical manual San Antonio, TX: Psychological Corporation. Wechs ler, D. (1997b). Wechsler Adult Intelligence Scale Third edition (WAIS III) San Antonio, TX: Psychological Corporations. Wiseheart, R., Altmann, L. J. P., Park, H., & Lombardino, L. J. (2009). Sentence comprehension in young adults with developmental dysl exia. Annals of Dyslexia, 59 (2), 151 167. Woodcock, R., McGrew, K., & Mather, N. (2002). Woodcock Johnson III tests of cognitive abilities Itasca, IL: Riverside Publishing Company. Xi, X. (2010). Aspects of performance on line graph description tasks: influenced by graph familiarity and different task features. Language Testing, 27 (1), 73 100. Yang, H. C. (2012). Modeling the relationships between test taking strategies and test per formance on a graph writing task: Implications for EAP. English for Specific Purposes, 32 (3). doi: 10.1016/j.esp.2011.12.004,
110 BIOGRAPHICAL SKETCH Sunjung Kim graduated from the psychology. In 2005, she received her Master of Arts degree in speech language pathology from the same Development of Phonological Processing Abilities of Children in the age of 3 to 6, was chaired by Dr. Young tae Kim. Sunjung has worked as a speech language pathologist since graduation. This clinic experience provided a foundation for literacy difficulties of dyslexia grounding this research project. In 2007, she came to the University of Florida to pursue doctoral studies in t he department of Speech, Language, and Hearing Sciences and have an interdisciplinary specialization through educational psychology in the area of reading comprehension. She received her Ph.D. from the University of Florida in the summer of 2012