USF Libraries
USF Digital Collections

Spelling errors in children with autism

MISSING IMAGE

Material Information

Title:
Spelling errors in children with autism
Physical Description:
Book
Language:
English
Creator:
Wiggins, Khalyn
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Phonology
Orthography
Autism
POMAS
Literacy
Dissertations, Academic -- Communication Sciences & Disorders -- Masters -- USF   ( lcsh )
Genre:
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: The goal of this study was to examine the spelling errors of children with Autism Spectrum Disorder (ASD) when asked to spell morphologically complex words. Specifically, this study sought to determine if percent accuracy across morphological areas would be similar to patterns noted in typical developing children, correlate with participant age, and correlate to performance on standardized measures of achievement. Additionally, the study wanted to highlight the types of errors made by children with ASD on homonyms and the specific linguistic patterns noted when spelling derivational and inflectional word types. Participants included 29 children diagnosed with Autism, PDD-NOS, and Asperger's Disorder, ages 8-15 years. The spelling protocol consisted of 36 words differing in morphological complexity, including homonyms, inflections and derivations. The derivational categories included: no shift, orthographic shift, phonologic shift, and orthographic + phonologic shift words (Carlisle, 2000). Spelling errors were analyzed both quantitatively and qualitatively. The qualitative analysis used a unique coding system, the Phonological, Orthographic, and Morphological Analysis of Spelling (POMAS; Silliman et al., 2006), which identified both the linguistic category of an error, as well as the specific linguistic feature in error. Results indicated that the spelling errors of children with ASD seemed to follow a developmental pattern that was similar to typically developing children (Carlisle, 1988; 2000). To be specific, phonologic and orthographic+phonologic shift categories evidenced significantly more errors than the no shift, orthographic shift, and inflections categories, which were not significantly different from each other. As expected, academic achievement, as measured by letter-word decoding, spelling, and age, were correlated with morphological spelling ability. Findings supported the use of the POMAS as a coding measure sensitive to spelling error patterns found in children with ASD. Several common feature errors emerged including: 1) vowel errors, 2) consonant deletions, 3) letter doubling, 4) derivational suffix errors, and 5) whole word substitutions. Overall, this heterogeneous group of spellers fit into three profiles of spelling ability: 1) competent spelling ability, 2) morphologically challenged spellers, and 3) generally challenged spellers. Hence, qualitative investigations of spelling errors play a crucial part in the characterization of spelling skill in children with ASD.
Thesis:
Thesis (M.S.)--University of South Florida, 2010.
Bibliography:
Includes bibliographical references.
System Details:
Mode of access: World Wide Web.
System Details:
System requirements: World Wide Web browser and PDF reader.
Statement of Responsibility:
by Khalyn Wiggins.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains X pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
usfldc doi - E14-SFE0003363
usfldc handle - e14.3363
System ID:
SFS0027679:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 22 Ka 4500
controlfield tag 007 cr-bnu---uuuuu
008 s2010 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0003363
035
(OCoLC)
040
FHM
c FHM
049
FHMM
090
XX9999 (Online)
1 100
Wiggins, Khalyn.
0 245
Spelling errors in children with autism
h [electronic resource] /
by Khalyn Wiggins.
260
[Tampa, Fla] :
b University of South Florida,
2010.
500
Title from PDF of title page.
Document formatted into pages; contains X pages.
502
Thesis (M.S.)--University of South Florida, 2010.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
3 520
ABSTRACT: The goal of this study was to examine the spelling errors of children with Autism Spectrum Disorder (ASD) when asked to spell morphologically complex words. Specifically, this study sought to determine if percent accuracy across morphological areas would be similar to patterns noted in typical developing children, correlate with participant age, and correlate to performance on standardized measures of achievement. Additionally, the study wanted to highlight the types of errors made by children with ASD on homonyms and the specific linguistic patterns noted when spelling derivational and inflectional word types. Participants included 29 children diagnosed with Autism, PDD-NOS, and Asperger's Disorder, ages 8-15 years. The spelling protocol consisted of 36 words differing in morphological complexity, including homonyms, inflections and derivations. The derivational categories included: no shift, orthographic shift, phonologic shift, and orthographic + phonologic shift words (Carlisle, 2000). Spelling errors were analyzed both quantitatively and qualitatively. The qualitative analysis used a unique coding system, the Phonological, Orthographic, and Morphological Analysis of Spelling (POMAS; Silliman et al., 2006), which identified both the linguistic category of an error, as well as the specific linguistic feature in error. Results indicated that the spelling errors of children with ASD seemed to follow a developmental pattern that was similar to typically developing children (Carlisle, 1988; 2000). To be specific, phonologic and orthographic+phonologic shift categories evidenced significantly more errors than the no shift, orthographic shift, and inflections categories, which were not significantly different from each other. As expected, academic achievement, as measured by letter-word decoding, spelling, and age, were correlated with morphological spelling ability. Findings supported the use of the POMAS as a coding measure sensitive to spelling error patterns found in children with ASD. Several common feature errors emerged including: 1) vowel errors, 2) consonant deletions, 3) letter doubling, 4) derivational suffix errors, and 5) whole word substitutions. Overall, this heterogeneous group of spellers fit into three profiles of spelling ability: 1) competent spelling ability, 2) morphologically challenged spellers, and 3) generally challenged spellers. Hence, qualitative investigations of spelling errors play a crucial part in the characterization of spelling skill in children with ASD.
590
Advisor: Sylvia F. Diehl, Ph.D.
653
Phonology
Orthography
Autism
POMAS
Literacy
690
Dissertations, Academic
z USF
x Communication Sciences & Disorders
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.3363



PAGE 1

Spelling Errors in Children with Autism by Khalyn I. Wiggins A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science Department of Communicati on Sciences and Disorders College of Behavioral and Community Sciences University of South Florida Major Professor: Sylvia F. Diehl, Ph.D. Ruth Huntley Bahr, Ph.D. Elaine R. Silliman, Ph.D. Date of Approval: March, 16 2010 Keywords: phonology, orthography, autism, POMAS, literacy Copyright 2010, Khalyn I. Wiggins

PAGE 2

Table of Contents List of Tables iv List of Figures v Abstract vi Chapter One: Review of the Literature 1 The Linguistic Basis of Spelling 2 Models of Spelling Development 3 Late Model 3 Early Model 7 Spelling Development in the Special Populations 8 Dyslexia 9 Language Learning Disability (LLD) 10 Autism Spectrum 11 Spelling 14 Summary 15 Purpose and Research Questions 16 Chapter Two: Method 18 Participants 18 Inclusion Criteria 19 Materials 21 Experimental Measures 21 i

PAGE 3

Construction of the Experimental Spelling Measure 23 Spelling Measure 23 Presentation 25 Procedures 26 Administration 26 Session 1: Inclusion Measures 27 Session 2: Experimental Spelling Measures 27 Response Format 27 Data Reduction 28 Parents Educational Level 28 Spelling Analysis 28 Quantitative Analysis 28 Qualitative Analysis 29 Statistical Analysis 30 Chapter Three: Results 31 Inter-Examiner Reliability 32 Participant Performance Quantitative 33 Developmental Sequence of Inflecti onal and Derivational Errors 35 Correlations with Partic ipant Age, Parents Level of Education, and Standardized Measures 36 Participant Performance Qualitative 39 Homonym Patterns 39 Patterns of Misspellings on Infl ections and Derivations 41 ii

PAGE 4

Phonological Feature Errors 41 Orthographic Feature Errors 44 Morphologic Feature Errors 46 Combination Feature Errors 47 Summary of Results 49 Chapter Four: Discussion 52 Summary of Results 47 Developmental Sequence of Inflecti onal and Derivational Errors 53 Spelling Accuracy and Measures of Literacy Achievement 54 Spelling Accuracy and Age 55 Spelling Accuracy and Parents Level of Education 56 Homonym Patterns 57 Patterns of Misspellings on Infl ections and Derivations 58 Feature Error Analysis 59 Vowel Errors 60 Consonant Deletion 62 Letter Doubling 63 Derivational Suffix Errors 65 Whole Word Substitution 67 Implications for Literacy Education 69 Strengths and Limitation 70 Sample 70 Spelling Inventory 71 iii

PAGE 5

Scoring Systems 71 Possible Directions for Future Research 72 References 74 Appendices 84 Appendix A: Informed Consent 85 Appendix B: Word Frequency Category 87 Appendix C: Stimuli 91 Appendix D: Assent Forms 96 Appendix E: Recognition Rules 100 iv

PAGE 6

List of Tables Table 1. Overview of Spelling Stag es Posed by Late Model Theorists 5 Table 2. Demographic Characteristics and Subtest Scores of Included Participants 20 Table 3. Experimental Spelling Words by Category 25 Table 4. Total Number of Correct Spellings on the Experimental Spelling Measure 33 Table 5. Post Hoc Results for Phonologic Shift and Orthographic+Phonologic ShiftCategory 36 Table 6. Correlations for Number of Words Spelled Correctly, Age, Parents Level of Education, and Subtests from the Woodcock Johnson II Tests of Achievement 32 Table 7. Distributions of Phonological Feature Errors Across Morphological Categories 43 Table 8. Distributions of Orthographi c Feature Errors Across Morphological Categories 45 Table 9. Distributions of Morphological Feature Errors Across Morphological Categories 47 Table 10. Distributions of Combination Feature Erro rs Across Morphological Categories 49 v

PAGE 7

vi List of Figures Figure 1. Percent Accuracy across Linguistic Categories 35 Figure 2. Distribution of Homonym Category results 40

PAGE 8

Spelling Errors in Children with Autism Khalyn Iman Wiggins ABSTRACT The goal of this study was to examine the spelling errors of children with Autism Spectrum Disorder (ASD) when asked to spell morphologically complex words. Specifically, this study sought to determine if per cent accuracy across morphological areas would be similar to patterns noted in typical developing children, correlate with participant ag e, and correlate to performance on standardized measures of achievement. Additionally, the study wanted to highlight the types of errors made by children with ASD on homonyms and the specific linguistic patterns noted when sp elling derivational and inflectional word types. Participants included 29 children di agnosed with Autism, PDD-NOS, and Aspergers Disorder, ages 8-15 years. The spelling protocol consisted of 36 words differing in morphological comple xity, including homonyms, inflections and derivations. The derivational categorie s included: no shift, orthographic shift, phonologic shift, and orthographic + phonol ogic shift words (Carlisle, 2000). Spelling errors were analyzed both qu antitatively and qualitatively. The qualitative analysis used a unique codi ng system, the Phonological, Orthographic, and Morphological Analysis of Spelling (POMAS; Silliman et al., 2006), which identified both the linguistic category of an error, as well as the specific linguistic feature in error. vi

PAGE 9

vii Results indicated that the spelling erro rs of children with ASD seemed to follow a developmental pattern that was similar to typically developing children (Carlisle, 1988; 2000). To be specifi c, phonologic and orthographic+phonologic shift categories evidenced significantly more errors than the no shift, orthographic shift, and inflections categories, which we re not significantly di fferent from each other. As expected, academic achieveme nt, as measured by letter-word decoding, spelling, and age, were correlated w ith morphological spelling ability. Findings supported the use of the PO MAS as a coding measure sensitive to spelling error patterns found in children with ASD. Several common feature errors emerged including: 1) vowel er rors, 2) consonant deletions, 3) letter doubling, 4) derivational suffix errors, and 5) whole word substitutions. Overall, this heterogeneous group of spellers fit into three profiles of spelling ability: 1) competent spelling ability, 2) morphol ogically challenged spellers, and 3) generally challenged spellers. He nce, qualitative i nvestigations of spelling errors play a crucial part in the characterization of spelling skill in children with ASD.

PAGE 10

Chapter One Review of the Literature The number of children diagnosed with Autism Spectrum Disorders (ASD) has steadily increased. Currently, 1 in every 150 children will be diagnosed with an ASD (Center for Disease Control, 2007). The nature of ASD leaves children with deficits in ma ny social and cognitive areas. These deficits are often evident in their academi c performance, which at one time led to an education segregated from their peer s. However, it is no longer common practice to separate children with ASD fr om their typically developing peers. Initiatives, such as No Child Left Behind (2001) and the Individuals with Disabilities Education Act (U.S. Depart ment of Education 2004), have helped provide children with ASD more consistent access to the general education curriculum. These advancements are focused on improving the quality of education for this population. However, there is little research to guide the development of a curriculum that encomp asses all of the knowledge areas covered in general education, yet still takes in to account the developmental differences found in this special population. An area of much needed investigation is that of specific relationships between certain li nguistic skills and literacy knowledge in children with ASD (Mirenda, 2003). One way this information has been accessed in typically developing children is by investigating their spelling abilities. What follows is a review of the spelling literature as it relates to the use of linguistic knowledge (e.g., phonology, orthography and morphology) in 1

PAGE 11

the ASD population. First, information will be given about the academic importance of spelling roles and linguisti c knowledge areas. Next, models of spelling development will be discussed in both typically developing children and children in special populations, includ ing the ASD population. Finally, the purposes of this study will be discussed al ong with the research questions it seeks to address, The Linguistic Basis of Spelling Current research in spelling has re cognized the importance of linguistic abilities in spelling development (Car lisle, 2000; Nagy, Berninger, & Abbott, 2006; Walker & Hauerwas, 2006). This line of research specifically highlights the importance of investigating how phonological, orthographic, and morphological knowledge provide the li nguistic basis for spelling. This knowledge base has been investigated in populations of children who have typical development (Carslie, 2000), language lear ning disability (LLD ) (Silliman, Bahr, & Peters, 2006; Windsor, Scott, & Str eet, 2000), and dyslexia, (Bourassa & Treiman, 2008; Tsesmeli & Seymour, 2006) but not children with ASD. Most of the research regarding literacy in children with ASD relates to their reading ability without considering spelling ability or the specific linguistic domains of phonology, orthography, or morphology. Research has shown that children with ASD show patterns in read ing ability that differ from typical developing children (Nation, Clarke, Wri ght, & Williams, 2006). For example, sometimes reading comprehension defic its exist alongside he ightened decoding skills (Welsh, Pennington, & Rodgers, 1987), while other children with ASD can 2

PAGE 12

read real words easily, but cannot read nonsense words (N ation, et al., 2006). The available evidence related to literacy abilities in child ren with ASD suggests that their integration of the linguistic informa tion needed to develop proficient spelling skill may differ from typically developing children. Furthermore, present research with children with ASD provide s little information related to their development of spelling ability. Analyzi ng spelling errors in this population will provide information about how they ar e integrating their linguistic knowledge during spelling tasks. Models of Spelling Development Numerous theories of spelling development have been proposed by the literature. The theories primarily differ in relation to the timing of the childs application of morphology to their spelli ng. The two most popular theories are discussed here: the late mode l and the early model. Late Model Late model theorists present ordere d stages in which levels of spelling aptitude are characterized by the kinds of knowledge that are predominate at each stage (Ehri, 1986; Gentry, 1982; Henderson, 1985). While each of these theorists has proposed a different number of stages and differing labels for those stages, there are overarching commonalities in th e progression of the type of knowledge used in spelling development Consideration of Table 1 reveals that Ehri (1986) describes three overarching stages in which the child transitions from beginning to use phonological knowledge (semi-phonetic stage) to being able to represent most 3

PAGE 13

sounds in their spelling (phonetic stag e) and finally to recognizing the irregularities in spelli ng which leads to tradit ional spelling (morphophonemic stage). On the other hand, Henderson (1985) included very early development by delineating the knowledge of drawing vers us writing. Whereas, Gentry (1982) chose to start recognizing spelling abili ties when children begin to write using strings of letters. Henderson and Gentry merge in their sim ilar description of children beginning to incorporate phonemi c knowledge in order to communicate through writing simple messages. This is when the child begins to use more phonological knowledge in their writing and th e child actually matches grapheme to sound. All of the theorists recognize that being able to make the grapheme to sound connection gives way to the integr ation of more orthographic knowledge into the spelling application. However, before this occurs, Gentry (1982) noted that children tend to create spellings th at may not necessarily match conventional English. These types of spellings result be cause the child tries to use a letter for every speech sound, sometimes providing mo re or less letters than necessary. Then, children begin to understand that th ere are alternatives to represent the same sound. Repeated exposure to prin t solidifies th e knowledge of when to apply these alternatives (Wright & Ehri 2007). Spelling development culminates in the use of morphological knowledge in spe lling. In this stage, prefix and suffix knowledge is integrated along with how to attach these morphemes to root words. In the final stage, derivational knowledge continues to expand until adult-like spelling is achieved. In su mmary, late model theorists propose a progression of 4

PAGE 14

spelling development that moves from phonological to orthographic and finally integrates morphology into spelling. Table 1 Overview of Spelling Stages Pr oposed by Late Model Theorists Gentry (1982) Henderson (1985) Ehri (1986 Preliterate Stage Difference between writing and coloring Pre-Communicative Stage -Using strings of letters that lack meaning Letter Name Spelling Stage -Using writing to communicate with others Semi-Phonetic Stage Beginning comprehension of letter-sound correspondence Within Word Pattern Stage -Beginning to use phonological strategies to spell unknown words Semi-Phonetic StageChildren rely on their knowledge of letter sounds (phonology) to spell. Very few correct spellings. Phonetic Stage -Use a letter or group of letters to represent every speech sound. Phonetic stage Children represent all or most of the sounds in spelling. 5

PAGE 15

Table 1 (Continued) May not conform to conventional English spelling. Transitional Stage Assimilates conventional alternatives for representing sounds. Using orthographic information and beginning morphological awareness Syllable Juncture Stage Repeated exposure to print supports appropriate use of orthographic variations; Use consonant doubling and inflectional endings. Morphological awareness begins Correct Stage -Basically correct spellings; understanding of prefixes, suffixes, silent consonants, irregular spellings; Grows throughout life. Derivational Principle Stage -Expansion of knowledge to derivational relationships (e.g., revisement/revision); Grows throughout life. Morphemic StageRecognition of spelling irregularities leading to conventional spelling. 6

PAGE 16

Early Model Unlike the Late Model, the Early Model postulates that children use basic knowledge of orthography and morphology ear lier than the late model suggests (Treiman & Cassar, 1997). For example, through the examination of spelling errors, it was noted that common errors, such as the use of letter names and consonant doubling, revealed a child s understanding of orthography in kindergarten and not just their reliance on phonology alone (Treiman, 1993; Treiman, Weatherston, & Berch, 1994). Early morphological knowledge was also evident in a study by Treiman, Cassar, and Zukowski (1994), who inves tigated childrens spelling of words containing one or two morpheme s and medial flaps, as in, duty ( a one morpheme word with no root) or dirty ( a two morpheme word where the root word is dirt) In this study, children were only required to finish spelling the words by filling in the missing t or d. It was noted that children in kindergarten and first grade had more correct responses for two morpheme words, such as dirty than the monomorphemic words, such as duty. Hence, in these words, the phonological representation of the flap was that of a /d/, while the correct orthographic representation was a t grapheme. However, children were able to spell these words correctly by placing the t grapheme for the /d/ phoneme. Had the children spelled these words with a /d/ graphe me, it would have illustrated use of phonological knowledge over morphological kno wledge. The use of the correct /t/ grapheme in two morpheme words indica ted that the children were tapping into their morphological knowledge. Other st udies have documented similar findings 7

PAGE 17

where children use some degree of mo rphological knowledge (Deacon & Bryant, 2006; Reece & Treiman, 2001; Walker & Hauerwas, 2006). In summary, both the late and early models highlight how important the linguistic skills of phonology, orthography, and morphology are to spelling ability. Despite when these skills are first evidenced or master ed, it is through the integration of these skills that children are able to become proficient spellers. Thus, it is important to determine the way children use this k nowledge when they encode written language, as well as, the e ffect the application of this knowledge has on their spelling proficiency. Spelling Development in Special Populations Spelling development in typically developing children provides a reference point to compare populations that have disorders affecting spelling skill. However, when making comparisons betw een populations that have disorders affecting academic ability and typically developing children, it is important to make sure outside factors do not affect outcomes. For example, age can affect a childs skill level because as they progre ss in school, they will inevitably gain more knowledge. Another factor that has been noted to affect academic skill level is maternal education (Davis-Keen, 2 005). It is important when drawing comparisons about typically developing chil dren and children with disorders that these aspects are taken into considerati on. This allows researchers to better determine if difficulties are truly related to the nature of the disorder or other outside factors. 8

PAGE 18

As illustrated previously, the degree to which students can integrate use of phonological, orthographic, and morphol ogical domain is important in differentiating poor from proficient spellers For this reason, it is important to know how development is affected in popul ations that have difficulty developing proficient spelling skills. The following se ction reviews research on variations in spelling development for populations with dyslexia, language learning disabilities (LLD), and ASD. Dyslexia Dyslexia is a reading disability that manifests itself in difficulties in the areas of word recognition, decoding, and spelling (Lyon, Shaywitz, & Shaywitz, 2003). Comparisons of the spelling skill of children with dyslexia and typically developing children reveal that children with dyslexia perform less well than their age-matched peers (Bourassa & Treiman, 2008; Tsesmeli & Seymour, 2006). When comparing the spelling errors of children with dyslexia to spelling agematched peers, the number and type of errors were similar for the two groups (Bernstein, 2009; Bourassa & Treiman, 2003; Cassar, Treiman, Moats, Polo, & Kessler, 2005;). In some cases, it has b een noted that indivi duals with dyslexia had similar spelling error patterns to indi viduals who were three or more years younger, indicating a level of performance commensurate with a three year or more delay (Bruck, 1993; Cassar et. al 2005) These findings were interpreted as indicating that spelling ability is delayed in individuals with dyslexia, mirroring the abilities of younger less proficient spellers. 9

PAGE 19

Studies investigating the spelling abilities of children with dyslexia have noted problems with consonant clusters, vowel letter names, (Cassar et. al, 2005) omission of final consonants, and overgeneralization of ed, (Egan & Pring, 2004), as well as, issues with deriva tions (Tsesmeli & Seymour, 2006). In looking at the types of errors being made and the age of the in dividuals who seem to be making similar errors, it appears th at in children with dyslexia, spelling develops in a manner much like typical peers, only in a slower progression. Language Learning Disability (LLD) Children with LLD have difficulties similar to children with dyslexia in the area of written langua ge, which are rooted in early difficulties in the comprehension and production of oral la nguage (Roseberry-McKibbin, 2007). In the area of spelling, children with LLD are similar to those with dyslexia. These similarities include showing reduced perf ormance in spelling tasks, similar but delayed performance to spelling level ma tched peers (Windsor, et al., 2000), and evidence of challenges with tasks that contain more complex morphology (Apel & Masterson, 2001; Hauerwas & Walker, 2003). They perform below both age and spelling level matched peer s (Hauerwas & Walker, 2003). Many of the studies investigating the spelling abilities of children with LLD looked at their misspellings quantitati vely. In a study that also included a qualitative analysis of misspellings, Sill iman, Bahr, and Peters (2006) examined the spelling abilities of children with LLD, as well as age-matched and spelling level matched control groups. First, a dictated spelling task that targeted the spelling of specific linguistic features known to be difficult for young children, 10

PAGE 20

such as, letter doubling, short vowels, and suffixes was administered. They then analyzed the misspellings in terms of the categories and f eatures listed on the Phonological, Orthographic, and Mo rphological Assessment of Spelling (POMAS), a qualitative measure. This method of qualitative analysis made it possible to study specific lingu istic feature patterns that differed as compared to those exhibited by the two control groups. Frequency results indicated that the children with LLD made more errors th an their age-matched peers, but were similar in error frequency to the group matched for spelling ability. However, investigation into th e quality of the spelling errors made between the LLD group and the spelling ability matched group revealed noteworthy differences. The LLD group had considerably more trouble repr esenting the basic phonological structure of words when complexity increased. Th is group also showed more frequent omissions of inflected and derived mo rphological markers. Because this study investigated the patterns a nd not just the frequency of errors, more than just a general delay in spelling skill development in children with LLD was uncovered. Autism Spectrum Autism Spectrum Disorder (ASD) is a neurobehavioral developmental disorder defined by clinical manifest ations (Allik, Larsson, & Smedje, 2006; Center for Disease Control, 2007; American Psychiatric Association, 1994). It is a global term that is commonly used fo r children diagnosed with one of three conditions, Autistic Disorder, Aspergers Disorder, and Pervasive Developmental Disorder-Not Otherwise Specified (PDD-NO S). ASD is delineated as a spectrum 11

PAGE 21

disorder because of the continuum of ch allenges from mild to severe (Rutter, 2006). The body of research in the area of l iteracy development in children with ASD is small (Mirenda, 2003). There is, however, strong evidence that many children with ASD have some word dec oding and alphabet know ledge even in the presence of complex language challenges. Children with ASD are often able to decode and encode words but have difficulty using those words in functional communication (Nation et al., 2006; O Connor & Hermelin, 1994; OConnor & Klein, 2004). This difficulty could be related to the demonstrated reduced functional connectivity in brain activity, evidenced in children with ASD. Specifically, research using fMRI sugge sts that children with ASD have a tendency to process language visually and have difficulty making semantic connections (Bregman, 2005; Cherkass ky, Kana, Keller, & Just, 2006; Gaffrey, Klienhans, Haist, & Mller, 2007). Just Cherkassky, Keller, & Minshew (2004) found that children with ASD even process letters differently, at a lower cortical level, than their typically developing peers. In addition to differences in the wa y letters are processed, some children with ASD are also able to decode and encode words far above their measured intelligence level. This heightened skill in decoding and encoding is referred to as hyperlexia. It has been noted that 5-10% of children with ASD are also hyperlexic, and that there is a higher ev idence of hyperlexia in ASD disorders than in other developmental disorders; how ever, it is not a dia gnostic indicator of ASD (Grigorenko, Klin, Paul et al., 2002; Grigorenko, Klin, & Volkmar, 2003). 12

PAGE 22

Hyperlexia is related to spelling tasks since children with better than average decoding and encoding skills might be expected to have above average spelling ability. However, Welsh, Pennington, & Rodgers, (1987) have suggested that children with hyperlexia pr efer to rely on their phonological knowledge, as opposed to their lexical or morphologi cal knowledge, when performing language tasks. If this is also true for sp elling, it is possible th at children who are hyperlexic may spell some words with ease, but have difficulty with words containing complex morphological elements. Since there is very little spelling research in children with ASD, it is diffi cult to determine how this strength in phonological knowledge affects the in tegration of morphological and orthographic knowledge during spelling tasks. While researchers have not specif ically examined the phonological, orthographic or morphological knowledge base of children with ASD, they have investigated the reading comprehension difficulties seen in this population. Saldana and Frith (2006) suggested th at the difficulties with reading comprehension seen in children with ASD were not related to their ability to access their world knowledge, which is an important component of comprehension. Instead, there may be some other aspect of reading comprehension that is different in childr en with ASD. The exact reason for the disconnection of reading comprehensi on from decoding is unknown. Some research suggests that the deficits ma y be related to the ability to understand anaphoric references (i.e., pronouns) in connected reading. OConnor & Klein (2004) showed that providing added cues to support comprehension of the 13

PAGE 23

nonspecific words in a passage showed the greatest increase in their reading comprehension scores. In this particul ar study, anaphoric reference support was compared to cloze tasks and pre-read ing questions (OConnor & Klein, 2004). Another study seeking to gain insi ght into the reading comprehension abilities as well as decoding skills of children with AS D was Nation et al. (2006). The results from this study indicated he terogeneous reading abilities among their participants. The various skill levels s uggested four subgroups of readers: 1) children who demonstrate good decoding of both real and non words, as well as good comprehension; 2) children who ha d good decoding skills and poor reading comprehension; 3) children who had good d ecoding skills related to real words, but poor nonword decoding; and 4) child ren who had poor decoding of both real and nonwords (Nation et al., 2006). Spelling. There currently are no studies on the spelling development of children with ASD, much less the stat us of development in the phonological, orthographic, and morphological knowledge sources that support spelling. Only three single subject case studies have been conducted that in vestigated spelling development in children with ASD (Blis chak & Schlosser, 2003; Schlosser & Blischak, 2004; Schlosser, Blischak, Belfiore, Bartley, & Barnett, 1998). However, unlike the research in dyslexi a and LLD, these studies focused on the advantages of speech generating devices as tools to improve spelling ability in children with ASD. The goal was to di scern relationships between synthetic speech feedback versus orthographic fee dback as methods to enhance spelling skills. 14

PAGE 24

Another method used to investigate sp elling abilities in children with ASD was video modeling. Video modeling is a therapy method which requires the participant to view videos of others performing tasks so that the participant can replicate these tasks. Video modeling was noted to increase generative spelling abilities (Kinney, Vedora, & Stromer 2003). This particular case study also noted that increasing the childs spelling ab ility helped her ach ieve grade level equivalent literacy skills. Again, while this study does highl ight the importance of spelling skill for children with ASD, it does not address how their phonological, orthographic, and morphologi cal knowledge affects their spelling skill. Summary Currently, research has illustrated the importance of spelling as it relates to proficiency in literacy. Studie s have analyzed the spelling errors of typically developing children and children with dyslex ia and LLD to determine how knowledge about phonology, orthography, and morphology are integrated to proficiently encode words. However, fe w studies have specif ically investigated the linguistic sources of sp elling, including how particular aspects of spelling are affected by underlying linguistic features (Silliman et al., 2006). In addition, few investigations have been performed to determine which of these areas of linguistic knowledge are most difficult. Despite the strides made in this line of research, there is little information about how these skills are developed in the ASD population. More research is needed but findings suggest that child ren with ASD process language more 15

PAGE 25

visually, can have higher than average decoding and encoding skills co-existing with low level reading comprehension, and rely more on rote memorization and phonological knowledge when decoding and en coding words. However an area that has remains uninvestigated in children with ASD is the spelling of inflectional and derivational morphol ogy or word formation processes. Purpose and Research Questions The major purpose of this study was to identify misspelling patterns in the spellings of children with ASD on inflectional and derivational morphologically complex words. In order to accomplish this goal, a spelling measure that highlighted homonyms a nd complex morphology (inflections and derivations) was developed. Therefore, the study addressed three research questions. The first question concerned quantitative findings, while the final questions addressed qua litative outcomes. 1. Does the percent accuracy ac ross morphological categories on the spelling test follow a developmental sequence similar to children who are typically developing? a. Does the accuracy of performance on a morphologically-based spelling measure correlate with partic ipant age and the highest level of parent education? b. Does the spelling accuracy co rrelate with performance on a standardized measure of achievement in spelling? 2. What types of errors do child ren with ASD make when spelling homonyms? 16

PAGE 26

3. What types of lingui stic patterns are noted in the misspellings of inflectional and derivational items? 17

PAGE 27

Chapter Two Method The purpose of this study was to inve stigate the spelling of inflections, derivations, and homonyms in children and adolescents with ASD using a spelling list comprised of real words differing in phonological, orthographic, and morphological features. In addition to sp ecific linguistic feat ures, derivations were varied along a continuum from more transparent to more opaque derived suffixes. Participants A total of 29 students with ASD (25 males and 4 females; mean age =10.79 years; SD = 2.26 years) were incl uded in the study. Participants were recruited through a listserve of constituents at the Ce nter for Autism and Related Disabilities (CARD), an organization created to provi de services for families of children with ASD located in an urban area near the University of South Florida in west central Florida. Participants were also recruited through use of letters presented to teachers at individual pub lic schools within the surrounding counties including Hillsborough, Pasco, and Pinellas c ounties. The letters were sent home by the teachers to the parents of poten tial participants. The students were currently enrolled in the equivalent of grades 2 through 9. A total of 41 students were initially recruited. Of these 41 students, 29 met the inclusion criteria. The parent of every child w ho participated in this study 18

PAGE 28

signed an informed consent approved by th e University of South Florida, Division of Research Compliance (see Appendix A) Additionally, all children signed an assent form before part icipating in the study. Inclusion Criteria A total of three inclusion criteria had to be satisfied to be included in the study. 1. The diagnosis of ASD (Autism, PDD-NOS, or Aspergers Disorder) was provided by and diagnosed by a certified professional, such as a clinical psychologist, psychi atrist, neurologist, or a phys ician. The diagnosis and the professional who made the diagnosis were verified by parent report on the informed consent. 2. A score within two standard deviations of the mean (70 or higher) on the Spelling Subtest of the Woodcock Johnson III Tests of Achievement (Woodcock, Mather, & Wendling, 2001). 3. Each participant had to pass a hearing screening on the frequencies of 1000, 2000, and 4000 Hz at a level of 30 dbHL. Information regarding the ethnicity and r ace of the individual participants was not formally collected. Demographic inform ation for all included participants is listed in Table 2. 19

PAGE 29

Table 2 Demographic Characteristics and Subtest Scores for Included Participants. Subject Age Gender Diagnosis LW SP PV PC 1 8 M PDD-NOS 120 122 87 94 2 8 M Aspergers Disorder 79 79 88 65 3 8 M Autism 123 122 108 88 4 8 M Autism 50 75 95 78 5 8 M PDD-NOS 87 77 88 65 6 8 M PDD-NOS 109 91 91 92 7 8 M PDD-NOS 98 92 101 92 8 9 M Autism 92 104 98 72 9 9 M Autism 97 95 108 82 10 9 M Autism 105 101 89 88 11 10 F Aspergers Disorder 120 130 103 99 12 10 M Aspergers Disorder 121 112 107 107 13 10 M PDD-NOS 91 81 86 84 14 10 M PDD-NOS 89 81 93 85 15 10 M Autism 76 85 85 68 16 11 M Autism 81 90 66 69 17 12 M PDD-NOS 87 95 82 84 18 12 M Aspergers Disorder 65 70 87 62 20

PAGE 30

Table 2 (Continued) 19 12 M Aspergers Disorder 91 79 94 96 20 12 M Autism 93 103 83 79 21 12 M Autism 85 95 67 50 22 13 M PDD-NOS 78 74 83 49 23 13 M Aspergers Disorder 95 84 70 80 24 13 F Aspergers Disorder 110 92 126 112 25 13 F Autism 113 122 110 100 26 14 M Aspergers Disorder 96 94 86 75 27 14 F Autism 73 80 74 55 28 14 M Autism 67 80 66 51 29 15 M Autism 64 78 64 55 M=Male; F=Female; LW=Letter Word Reading; SP=Spelling; PV= Picture Vocabulary; PC= Passage Comprehension Materials Experimental Measures The Woodcock Johnson III Test of Achievement (Woodcock et al., 2001) is a standardized battery for measuri ng academic achievement designed for individuals, ages 2 through 90+ years. Four subtests were individually 21

PAGE 31

administered to quantify the skill leve l of participants using standardized measures. These subtests are described below. 1. The Letter Word Reading Subtest provides information about word and letter recognition, as well as decoding ability. Words increase in complexity according to word length. Each participant was given several lists of words to read, with approximately six words to a pa ge to determine their skill level with decoding words. 2. The Spelling Subtest was administered orally in dictation format. The target word was read to the student followed by a sentence using the word in context and then the target word was pres ented again in isolation. Spelling ability was based on age level. According to the administration manual of the Woodcock Johnson III Test of Achievement (Woodcock et al., 2001), participants continued the test until they misspelled five word s in a row. This procedure caused the number of words presented to each student to vary slightly. Responses were written, on a form provided with the test ma terials. This subtest gave information about the participants spelling abilities. 3. The Picture Vocabulary Subtest provides global information about production vocabulary and consis ts of four colored pictur es per page. Participants were directed to look at the picture and th en were asked to give the name of the item. This subtest gave information a bout the participants vocabulary knowledge when word decoding was not a factor. 4. The Passage Comprehension Subtest was a cloze procedure with 1-3 items per page. Participants orally read a paragraph al oud and filled in the blank. 22

PAGE 32

This subtest gave information about the pa rticipants ability to read a passage and understand what they read. Construction of the Expe rimental Spelling Measure Spelling measure. The initial corpus of word s for the spelling measure was selected from words that were freque ntly misspelled by students in grades 5-9 in a study done by Bahr, Silliman, & Berninger (2009). These frequently misspelled words were further separated into two major types of words. The first type were inflections (n = 6, including the plural, past pa rticiple, regular past tense ed, and third person tense agreement es), chosen because these forms are typically mastered in writing no late r than age 10. The second word type consisted of derivations (n = 24), which represented an amalgam of meaning and form (Carlisle, 2004). These derivations were further delineated according to their degree of transparency (Carlisle, 198 8, 2000): 1) no shift in pronunciation or spelling (n = 6; e.g., friendship); 2) ort hographic shifts where there was a change in spelling, but not in pronunciation (n = 6; e.g., argument); 3) phonologic shifts in which pronunciation, but not spelling, cha nged (n = 6; e.g., majority); and 4) orthographic + phonologic shifts, the most opaque type since both pronunciation and spelling were altered (n = 6; e.g., pleasant). A third word type, homonyms, was added to investigate semantics-related spelling issues in this population. These three types of words (homonyms inflections, and derivations) were expected to challenge the participants sp elling abilities. However, in order to 23

PAGE 33

insure that the words were appropriate for the targeted age groups of this study, word frequency was also contro lled using the following methods. Word frequency was determined using The Educators Word Frequency Guide (Zeno, Ivens, Koslin, & Zeno, 1995; See Appendix B) for the word frequencies of the 36 experimental spelli ng items). This particular frequency guide is based on a corpus of 60,527 samples of text from 6,333 works of literature. Word frequency was report ed via U value and an SFI (Standard Frequency Index) value. The SFI value is a logarithmic version of the U value making it easier to use and understand. SFI values for the words ranged from 3.5, (about .0002 frequency of occurrence per million words) to 88.3 (about 67,500 frequency of occurrence per million words). For this study, words were chosen fo r the experimental spelling protocol that had an overall SFI of 40 or greater. This mean t that the frequency of occurrence of the selected words was at least one for every million words The value of 40 SFI falls towards the middle of the frequency range of very frequent words, 80.0 and above (e.g., the, and, & is) and infrequent words 20.0 and below (e.g.,acclimate, orthogonal, and speculate). Additionally, words had to have a U value higher than 0 for at least four of the eight grade levels present in this study because a U value of zero indicated that the word had little to no presence in the literature for that particular grade level. For example, the word convertible had U values of 0 for grades 2, 7, 8, indicating that it had very little to no presence in the literature sampled for these grades. 24

PAGE 34

When this process was completed, each category contained six words for a total of 36 words. (See Table 3 fo r the 36 target wo rds by category.) Table 3 Experimental Spelling Words by Category. H I N S O S P S P+O S Aloud Cries Dangerous Argume nt Different Pleasant Four Stirring Friendship A ttention Disappear Excellent Week Stopped Assignment Juicy Majority Student Clothes Kicked Smoothly Hungry Convertible Natural Sent Building Highest Easily Children Severity H= Homonym; I= Inflections; N S= No Shift; O S= Orthographic Shift; P S= Phonologic Shift; P+O S= P honologic + Orthographic Shift Presentation. Presentation of each target word was accompanied by a picture, depicting word meaning, foll owed by the spelling target, a sentence context for the spelling target, and then a carrier phrase, Spell the word_____. (See Appendix C for the pictures and sent ence contexts). All pictures were obtained from Boardmaker Plus, v6 or Mi crosoft Office clipart software. The pictures and verbal instructions were presented using a DELL Latitude laptop computer running EcosWin software. 25

PAGE 35

Procedures Administration Administration of the inclusion measures and the spelling protocol were completed over one to two sessions occurr ing approximately within one month of each other. When testing occurred in one session, the child was given a 15-30 minute break between the administration of the inclusion measures (approximately 1 hour in length) and the experimental measure (approximately 45 minutes in length). During the break, participants were permitted to play a game or complete an activity of their choice. Individual testing took place in a quiet area of the childs school or in the participants home. The participant wa s seated next to the examiner during testing procedures. Before each session, th e participant was read an assent form that explained the activities they would complete during that session (N= 2 forms per child; see Appendix D) and the ch ilds signature was obtained. During testing, the participant, examiner, and, sometimes, a parent were present. The parent was not permitted to help the child with the testing procedures, but parents were encouraged to observe and ask questions at the end of te sting. Parents were provided information about the scores their children received on the subtests form the Woodcock Johnson III Test of Achievement (Woodcock et al., 2001) 26

PAGE 36

Session 1: Inclusion Measures Only the Spelling Subtest from the Woodcock Johnson III Test of Achievement (Woodcock et al., 2001) was used to determine eligibility. The other three subtests, Letter Word Reading, Passage Comprehension, and Picture Vocabulary, were administered to obtain a profile of the pa rticipants general literacy abilities. Each subtest was admi nistered in a predetermined random order to avoid test order effects on the data. Subtests were administered and scored according to the guidelines in the administration manual of the Woodcock Johnson III Test of Achievement (Woodcock et al., 2001). Session 2: Experiment al Spelling Measure The 29 participants who met the inclus ion criteria were administered the experimental spelling protocol. Each was informed that they were going to spell some words and that they should do th eir best. The 36 spelling words were presented orally by a recorded human voi ce in random order acr oss participants using EcosWin software to st andardize the presentation. Response Format Research indicates that the mode of response does not affect spelling ability (Masterson & Apel, 2006). Base d on this finding, participants were allowed to choose the mode of respons e, whether typing or hand writing. If he/she chose to type, the pa rticipant was instru cted on how to use the keyboard to type their responses, erase mistakes, and re quest repeating of an item, if needed. This was completed through two practice items that could be repeated as many times as needed. All participants chose to type their responses. Some participants 27

PAGE 37

wrote their responses before typing in or der to keep them focused on the spelling task. During the administration, if a part icipant asked for help in spelling a word, he or she was instructed to give it his/her best try. It should be noted that one participant showed extreme anxiet y about writing th e answers. This participant was permitted to complete th e spelling portion by spelling the words verbally and having his answers recorded. Data Reduction The data obtained from participants was processed using the following procedures before it was analyzed statistically. Parents Educational Level Information about parents highest leve l of education was converted to an ordinal scale, from 1 to 5 as follows: Parents who did not have a high school diploma were assigned a 1; receipt of a high school diploma and/or participation in vocational school was assigned a 2; some college completion received a 3; an undergraduate degree was assigned a 4; and a 5 was given for any graduate level work, including doctoral level work. Spelling Analysis Quantitative analysis. General spelling performance was determined by determining the total number correct on the experimental spelling measure. The total number of misspellings for each li nguistic category (homonyms, inflections, no shift, orthographic shifts, phonologic sh ifts, orthographic + phonologic shifts), was tabulated for each participant to determ ine the proportion of errors in each of 28

PAGE 38

the morpheme categories. This number was divided by the total possible responses in each category to obtain a percentage incorrect for each of the word types. Qualitative analysis. The POMAS scoring system (Silliman et al., 2006) was applied to the qualitative analysis of misspellings in each category. This scoring system identifies the linguistic category of the spelling error as phonological, orthographic, morphological, or a combination of these categories and then describes the linguistic feature in error. An error was considered to be phonological in nature if the sound structure of the word was not fully represented or changed because of the deletion of phonemes. An orthographic error was considered if the sound structure of the word was fully represented with inappropriate graphemes. Errors that were considered to be morphological in nature consisted of two kinds: deletions or an incorrect grapheme representation of an inflection or a derivational suffix or roots that were misspelled in the process of adding derivational a nd/or inflectional suffixes. Following this first level analysis, erro rs were then analyzed based on their linguistic features according to the POMA S scoring guidelines. These features included codes developed from previous studies (Silliman et al., 2006). Additional codes were added specifically for this study to account for the morphologically balanced categories used in the current spelling protocol. (see Appendix E for the lingu istic feature codes). 29

PAGE 39

Statistical Analysis The independent variable in this study was the type of word spelled. The dependent variable was the percentage of errors in each morphological category. A repeated measures one-way ANOVA was used determine the differences among morphological categories. Post hoc related samples t-tests, using a Bonferroni Adjustment of .003 (.05/15), we re conducted to determine specifically where the differences were located. Pear son Product Correlations were applied to examine relationships among spelli ng accuracy and performance on the Woodcock Johnson III Test of Achievement Subtests (Woodcock et al., 2001). A qualitative analysis using the POMAS coding system (Silliman et al., 2006) was also performed to describe the nature of spelling errors made. This analysis looked at the different spel ling errors made by the participants and categorized them into broad linguistic categories a nd based on linguistic feature errors. 30

PAGE 40

Chapter Three Results This study investigated the error patte rns found in the spellings of children with ASD. The design of the investigat ion included the diag nostic categories of Autism, Aspergers Disorder, and PDD-NOS. An experimental spelling measure was designed and administered to investig ate the linguistic categories and features (phonological, orthographic, morphological, and combinations) of misspellings of these participants. The specific questions proposed were: 1. Does the percent accuracy ac ross morphological categories on the spelling test follow a developmental sequence similar to children who are typically developing? a. Does the accuracy of performance on a morphologically-based spelling measure correlate with partic ipant age and the highest level of parent education? b. Does the spelling accuracy corre late with subtests from the Woodcock Johnson III Tests of Achievement ? 2. What types of errors do children w ith ASD make when spelling homonyms? 3. What types of lingui stic patterns are noted in the misspellings of inflectional and derivational items? In order to answer question 1, a re peated measures one-way ANOVA and post hoc tests were run on the number of errors in each linguistic category (homonyms, inflections, no shift, ort hographic shift, phonologic shift, and 31

PAGE 41

orthographic +phonologic shift). Additio nally correlations were used to determine the relationships between other va riables. Finally, qua litative data were analyzed to note the frequency of specific linguistic errors. InterExaminer Reliability Agreement was conducted for the error categorization of the 36 words. A total of 29 participants were includ ed, and 20 percent of this sample (N =6) was randomly selected for agreement analysis A second examiner, trained in the POMAS coding system, was asked to r ecode the spelling errors of these participants. Training consisted of re viewing specific wri tten rules regarding when to use certain POMAS codes (See Appendix E). When the second examiner felt comfortable with the coding system she was given six randomly selected samples to code. The complex nature of th e coding system required that the errors for linguistic features be reduced to the linguistic categories (phonological, orthographic, morphological, or a combination) for agreement analysis. Interexaminer reliability was determined using the formula for Cohens Kappa (Cohen, 1960). This resulted in =.73. A Cohens Kappa of .70 indicates that satisfactory reliability between two coders. A value below this level indicates unsatisfactory reliability between coders. Therefore, a of .73 is acceptable. This reliability value was expected due to the complex nature of this coding system and the sometimes indistinct na ture of the errors. For example, different spelled as differant could be interpreted as an or thographic error by choosing the wrong vowel or as a morphological error wh ere the child had difficulty with the spelling of the suffix. In these situations where errors could be dually coded, the 32

PAGE 42

overall common patterns of the childs ot her misspellings were examined and a decision was made that was consistent with these patterns. Participant Performance: Quantitative Analysis The total number of words spelled correctly was determined for all participants (see Table 4 for the total number of correct spellings for each participant). Additionall y, the percentage of words spelled incorrectly was calculated for each of the six linguistic categories (See Figure 1 for percentages). The data were then analyzed further to answer the specific questions related to this study. These analyses are discussed in detail in the sections that follow. Table 4 Total Number of Correct Spellings (Maximum Score = 36) on the Experimental Spelling Measure. Age Gender Diagnosis # Correct Spellings/36 10 M Aspergers Disorder 34 13 F Autism 33 8 M PDD-NOS 28 13 F Aspergers Disorder 27 12 M Aspergers Disorder 27 11 M Aspergers Disorder 24 12 M PDD-NOS 24 8 M Autism 24 10 F Autism 24 33

PAGE 43

Table 4 (Continued) 9 M Autism 19 9 M Autism 18 14 M Autism 18 8 M Autism 17 13 M Aspergers Disorder 16 12 M Autism 16 12 M Autism 14 14 F Autism 13 14 M Aspergers Disorder 12 10 M Autism 12 10 M PDD-NOS 11 12 M Autism 9 8 M PDD-NOS 8 8 M Aspergers Disorder 7 9 M PDD-NOS 6 8 M PDD-NOS 5 8 M Aspergers Disorder 1 10 M PDD-NOS 1 13 M PDD-NOS 1 15 M Autism 0 34

PAGE 44

73.56 72.41 57.47 51.72 49.43 38.5 0 10 20 30 40 50 60 70 80 90 100 Ortho+Phon o Phonologic Inflections Orthographi c No Shift Figure 1. Percent accuracy across linguistic ca tegories on experimental measure Developmental Sequence of Inflectional and Derivational Errors The first question addressed if inflectional and derivational spelling errors, including derivationa l shifts in phonology and orthography, followed a developmental sequence. The repeated measures one-way ANOVA revealed a statistically significan t main effect for error type ( F (5,140) =22.482, p< .001) with a moderate effect size ( p .45; Cohen 1988). Post hoc testing indicated that the phonologic shift and phonologic +orthographical shift 35

PAGE 45

categories were significantly more difficult than the other five categories, which were not significantly different from one another (See Figure 1) The results of the post hoc related-samples t-test are listed in Table 5. Table 5 Post Hoc Results for Phonologi c Shift and Orthographic+Phonologic Shift Category Inflections No Shift Orthographic Shift Phonological Shift t (28) = -3.588 p < .001 t (28) =-6.308 p < .001 t (28) = -4.446 p < .001 Orthographic + Phonological Shift t (28) = -3.854 p < .001 t (28) =-6.769 p < .001 t (28) = -4.456 p < .001 Correlations with Participant Age, Parents Level of Education, and Standardized Measures This part of the first question specifi cally addressed how age, the parents level of education, and the standardized measures corre lated with accuracy on the experimental spelling measure. The variables included the participants age, the mother and fathers highest level of e ducation, the standard scores on the four subtests from the Woodcock Johnson III Test of Achievement (Woodcock et al., 2001), and the total number of items sp elled correctly on the experimental spelling measure (See Table 5). Pears on Product correlations were conducted to 36

PAGE 46

determine the relationships among these vari ables. Results revealed that correct responses on the experimental spelling measure, (M = 15.48, SD = 9.8, N = 29) and age ( M = 10.87, SD = 2.25) were signifi cantly correlated, r (27) = .37, p = .041, r2 = .14 (See Table 5). The correlation was small and accounted for approximately 14% of the variance in the data (Cohen, 1988). This indicated that as the participants age increased, they spelled more words correctly on the experimental spelling measure. Table 6 Correlations for Number of Words Spelled Correctly, Age, Parents level of education, and Subtests from the W oodcock Johnson III Tests of Achievement Age #R LW PC PV SP ME DE Age .373* -.302 -.329 -.384* -.253 -.458 -.215 #R .586** .345 .243 .727** -.105 .184 LW .763** .595** .829** .232 .263 PC .775** .573** .096 .239 PV .445* .155 .205 SP .191 .350 ME .526* DE p < .05, ** p < .01; Age= Participants age; #R= Number correct on Experimental Spelling Measure; Woodcock Johnson III Test of Achievement; LW= Letter Word Reading Subtest; PC = Passage Comprehension Subtest; PV= 37

PAGE 47

Picture Vocabulary Subtest; SP= Spelling Su btest; ME= Mothers highest level of education; DE= Fathers hi ghest level of education There was not a significant correlat ion between participant performance and the parents level of education. However, the Pearson Product correlation conducted for the relationship between th e number of words spelled correctly ( M = 15.48, SD = 9.8, N = 29) and performance on the Letter Word Reading Subtest (Woodcock et al., & 2001) ( M = 91.55 SD = 18.69) indicated a significant correlation, r (27) = .59, p = .001, r2 = .34, with a moderate effect that accounted for approximately 34% of the variance. This finding indicated that performance on the experimental spelling measure was re lated to the participants ability to decode words. Additionally, a Pearson Product correla tion revealed that the correlation between the number of words spelled correctly on the expe rimental spelling measure (M = 15.48, SD = 9.8, N = 29) and the Spelling Subtest of the ( M = 92.93, SD = 16.27) was significant, r (27) = .73, p = .0001, r2 = .56. A moderate correlation was noted that accounted fo r 56% of the variance. The other two standardized measures from the Woodcock Johnson III Tests of Achievement, Reading Comprehension and Pi cture Vocabulary, were not correlated with performance on the experi mental spelling measure. This result can be expected as these two subtests di d not directly examine encoding and were more related to decoding and genera l vocabulary knowledge, respectively. Furthermore, all standard scores on the four subtests from the Woodcock Johnson III Tests of Achievement (Woodcock et al., 2001) areas were found to be 38

PAGE 48

significantly correlated with each other. This result can be expected since the subtests were developed to be given as part of a battery that pr ovides a profile of a students academic achievement (See Ta ble 6 for Pearson Product r values). Participant Perform ance: Qualitative Homonym Patterns This question addressed the participants ability to identify the homonym that semantically fit the dictated se ntence and accompanying picture prompt. Errors were qualitatively analyzed usi ng the POMAS coding systems (Silliman et al., 2006) to determine if th e spelling error was due to an incorrect choice in graphemes or an incorrect semantic c hoice. The code, morphological homonym error (MHOM), was used when an incorrect semantic choice was made. This code appeared 28 times in the POMAS error analysis. In the homonym category, six words were presented to each of the 29 participants for a total of 174 homonyms spelled in this investigation. Of these 174 homonyms 38.5% (67/174) of the words were spelled incorrectly (See Figure 2). Half of the participants ( n=14) made an error that resulted in a MHOM code. This error code accounted for 16.1% (28/174 ) of the total erro rs in the homonym category. Furthermore, several participants ( n= 5) made the MHOM error on multiple words. All but two of the other errors in this category were related to problems with phonological and orthographi c features of the words and not morphology Interestingly, 36/174 (20%) of the total homonym errors were on the words clothes and aloud. Examples of spelling errors included: clothes 39

PAGE 49

spelled as clos, cloths, croes close and aloud spelled as alode alloue alloud allowed. 61.5% 16.1% 22.4% No Errors Semantic Error s (MHOM) Non Semantic Errors Figure 2. Distribution of errors in the homonym category When the error was not related to an incorrect semantic choice, different feature error codes from the phonological ( n=16), orthographic ( n=31) and combination ( n=5) broad categories were used. Examples of these codes included phonological consonant deletion (PCD; n=3) (e.g. easily spelled as eaily ), phonological long vowel error (PLV; n=5) (e.g. teaches spelled as techs ), orthographic consonant doubling (OCD; n=4) (e.g. discussion spelled as disscussion) orthographic long vowel pattern error (OLVP; n=4) (e.g. clothes spelled as cloaths) orthographic whole word (OWW; n=8) (e.g. argument spelled 40

PAGE 50

as areyoument or discussion spelled as disguession) and phonological orthographic reversal error (POR; n=4) (e.g. building spelled as bliuding). Patterns of Misspellings on Inflections and Derivations The final question was addressed th rough a qualitative analysis to determine common errors according to the inflectional or derivational category of the word. Using the POMAS coding system, errors were divided into the broad categories of phonological, orthographic, morphological, and combination codes. Combination codes were codes which fe ll under two of the three broad POMAS categories. In total, there were 180 phonological error codes, 432 or thographic error codes, 37 morphological error codes, a nd 395 combination codes. The specific results of the qualitative feature analysis by morphological category are represented in Tables 6 9. Phonological feature errors. There were a total of 180 phonological feature errors. Five linguistic errors dominated here, as seen in Table 7. The error with the highest number of occurre nces was consonant de letion (code: PCD; n = 31 or 17.2% of the total number of phonological errors) (e.g., easily spelled as eaily), with errors occurring most fre quently in orthographic + phonologic shift but also distributed across the no sh ift and phonologic shift delineations. Epenthesis (code: PEP; n = 27 or 15.0% of the total number of phonological errors) (e.g., student spelled as stunden t), reducing a sonorant cluster (code: PSONC; n = 27 or 15.0% of the total phonological errors) (e.g., windy spelled as widy ), and turning long vowels in to short vowels (code: PLV; n = 26 or 14.4% of 41

PAGE 51

the total number of phonological errors) (e .g., teaches spelled as teches) followed. These errors primarily occurred in the phonologic shift delineation. Reducing syllables (code: PSR; n = 21 or 11.7% of the total number of phonological errors) (e.g., easily spelled as easly ), and vocalic r errors (code: PVOCR; n = 13 or 7.2% of the total number of phonological errors) (e.g., natural spelled as natual ) were third and fourth in frequency with er ror distributions found primarily in the orthographic and orthographic + phonologic shift categories. 42

PAGE 52

Table 7 Distribution of Phonological Feature Errors Across Morphological Categories Inflections No Shift Ortho Shift Phono Shift Ortho+ Phono Shift Total Consonant Deletion (PCD) 5 8 3 7 8 31 Sonorant Cluster Reduction (PSONC) 20 2 4 1 27 Epenthesis (PEP) 3 2 6 11 5 27 Long Vowel (PLV) 8 7 11 26 Syllable Reduction (PSR) 2 12 1 6 21 Vocalic r (PVOCR) 3 1 5 4 13 43

PAGE 53

Orthographic feature errors. There were a total of 432 orthographic feature errors. There were several linguistic f eature errors noted, but five errors dominated and are presented in Table 8. The most common orthographic error was letter doubling (code: OLD; n = 91 or 21.1% of the total number of orthographic errors) (e.g., stoped spelled as stoped ) with errors occurring most frequently in the inflections category, but also distributed across the phonologic shift, no shift, and orthog raphic shift delineations. Interestingly, there were instances where doubling occurred in nonobligatory orthographic contexts called orthographic consonant doubling (code: OCD; n = 25 or 5.8% of the total number of orthographic errors) (e.g.,, discussion spelled as disscussion). Using an incorrect consonant (code: OCE; n = 61 or 14.1% of the total number of orthographic errors) (e.g., confidence spelled as confinence ) was the next most frequently occurring error. These er rors primarily occurred in the phonologic shift delineation. Using an incorrect vowel (code: OVE; n = 53 or 12.3% of the total number of ort hographic errors) (e.g., severity spelled as suverity) and misspelling a rhotic vowel (code: OVr; n = 50 or 11.6% of the total number of orthographic errors) (e.g., stirring spelled as sturing ) were third and fourth in frequency of errors. These errors we re dominant in the orthographic+phonologic shift category, but were also distributed across the other three shift categories. The fifth most frequently occurring erro r was replacing part of a word with a phonologically similar whole word (code OWW; n = 33 or 7.6% of the total number of orthographic errors) (e.g., argument spelled as areyoument ). This error type dominated in the orthogr aphic + phonologic shift category. 44

PAGE 54

Table 8 Distribution of Orthographic Feature Errors Across Morphologic Categories Inflections No Shift Ortho Shift Phono Shift Ortho+ Phono Shift Total Letter Doubling (OLD) 5 14 6 16 24 65 Consonant Doubling (OCD) 2 12 1 6 21 Consonant Error (OCE) 8 7 11 26 Vowel Error (OVE) 3 2 6 11 5 7 Rhotic Vowel (OVr) 3 1 5 4 13 Whole Word (OWW) 2 5 6 8 12 33 45

PAGE 55

Morphological feature errors There were a to tal of 37 morphologic errors. Of these, four linguistic featur e errors predominated, as illustrated in Table 9. The most common error was corr ectly spelling the suffix with two or more errors in the root word (code: MDER; n = 14 or 38.9% of the total number of morphological errors) (e.g., juicy spelled as jaicey ). These errors occurred most frequently in the orthographic shif t category. The second most frequently occurring feature error was leaving o ff an inflectional ending (code: MINF; n = 11 or 30.6% of the total numb er of morphological errors) (e.g., kicked spelled as kick). This error only occurred in the inflections category. The third most frequently occurring morphologic featur e error was leaving off a derivational ending (code: MSUF; n = 6 or 16.7% of the total num ber of morphological errors) (e.g., juicy spelled as juice). This erro r also occurred most in the orthographic shift category. The final morphologic f eature error that occurred was overgeneralizing a derivational form (code: MDVM; n = 5 or 13.9% of the total number of morphologi cal errors) (e.g., disappear spelled as dissapered ), with errors occurring most frequently in phonologic shift category. 46

PAGE 56

Table 9 Distribution of Morphologic Feature Errors Across Morphologic Categories Inflections No Shift Ortho Shift Phono Shift Ortho+ Phono Shift Total Derivation root word (MDER) 8 3 3 14 Inflectional (MINF) 11 11 Suffixes (MSUF) 5 1 6 Derivational (MDVM) 5 5 Combination feature errors There were a total of 395 combination errors. The five most common errors are repr esented in Table 10. The most common error was misspelling the derivational suffix (code: OSUFD; n = 224 or 56.7 % of the total number of combination errors) (e.g., smoothly spelled as smoothle ). These errors occurred most frequen tly in the phonologic and orthographic + 47

PAGE 57

phonologic shift category, but we re distributed across all categories. The second most frequently occurring feature erro r was an error with a short vowel sound (code: POSV; n = 48 or 12.2% of the total numbe r of combination errors) (e.g., confidence spelled as confedence ). This error was almost evenly distributed across the no shift, phonologic shift, or thographic shift, and orthographic + phonologic shift categories. The third most frequently occurring combination feature error was reversing lett ers in a word (code: POR; n = 42 or 10.6% of the total number of combination errors) (e.g., building spelled as bliuding ). This error occurred most frequently in the phonol ogic shift category, bu t also was almost evenly distributed across the other word categories. The final combination feature error codes were missing vowels from a word (code: POVM; n = 40 or 10.1% of the total number of combination errors) (e.g., stirring spelled as strring) and misspelling the inflectional endi ng of a word (code: OSUFI; n = 34 or 8.6% of the total number of combination errors) (e.g., kicked spelled as kickt). Missing vowels were most noted in the spelling in the phonologi c shift and orthographic + phonologic shift categories. Misspellings of inflectional endings were only noted in the inflections category. 48

PAGE 58

Table 10 Distribution of Combina tion Feature Errors Across Morphologic Categories Inflections No Shift Ortho Shift Phono Shift Ortho+ Phono Shift Total Suffix Derivation (OSUFD) 0 46 68 75 65 Short Vowel (POSV) 4 8 12 13 48 Reversal (POR) 5 10 12 9 42 Vowels Missing (POVM) 5 4 16 12 40 Suffix Inflectional (OSUFI) 34 34 Summary of Results In summary, the participants in th is study had the most difficulty with spelling words that were the most comp lex in terms of derivational morphology. 49

PAGE 59

These were the words in the phonologic shift and orthographic + phonologic shift categories. Other word categories had a high error rate, but the number of errors was not significantly different between th ese categories. The performance of the participants was noted to correlate with th eir age in that as age increased, so did the number of words spelled correctly. Ho wever, the parents level of education was not correlated with participant pe rformance. Additionally, there were significant correlations between two of the four subtests (Letter Word reading and Spelling) from the Woodcock-Johnson Tests of Achievement III (Woodcock et al., 2001) and the number of words spelle d correctly by participants. The homonym category had the least numb er of errors of the six word categories represented in this study. Despite this, almost half of the participants in the study (n=14) had difficulty choosing the ho monym that semantically fit with the picture and sentence prompt. Th ese two observations taken together may indicate that there are so me individuals who have difficulty making semantic choices regarding homonyms. However, these types of errors may be more prevalent as homonym forms become more difficult than those used in this study. The POMAS coding system proved to be sensitive to the sp elling errors of children with ASD. Through use of the POMAS, feature codes several error patterns emerged from this sample. The br oad error analysis indicated that most errors occurred in the orthographic and combination categories. Further analysis using the category features indicated that specific erro rs were related to five patterns: 1) letter doubling (OLD), 2) using phonologica lly similar whole words to spell part of a word (OWW), 3) spel ling the prefix or suffix correctly but 50

PAGE 60

misspelling the root word (MDER), 4) l eaving off an inflectional ending (MINF), and 5) spelling the affix of a derivati on or inflection inco rrectly (OSUFD or OSUFI). 51

PAGE 61

Chapter Four Discussion Summary of Results The goal of this study was to examin e the error pattern s that occurred when children with ASD, ages 8-15 years, were asked to spell morphologically complex words. Specifically, this study s ought to determine if percent accuracy across morphological areas would be si milar to patterns noted in typical developing children, correlate with particip ant age, and correlate to performance on standardized measures of achievement. Additionally, the study wanted to highlight the types of errors made by children with ASD on homonyms and the specific linguistic patterns noted when sp elling derivational and inflectional word types. In this study, 29 children diagnos ed with ASD, ages 8-15 years, were asked to spell a total of 36 words differ ing in morphological complexity. Results indicated that spelli ng errors, made by the children th at participated in this study, seemed to follow a developmental pattern that was similar to typically developing children (Carlisle, 1988 ; 2000). Furthermore, as expected, academic achievement as measured by letter-word decoding, spelling, and age were correlated with spelling ability on a morphologically based measure. Most importantly, findings supported the use of the POMAS (Silliman et al., 2006) as a coding measure sensitive to spelling error patterns found in children with ASD. Through the use of category and feature analysis provid ed by this coding system, several common 52

PAGE 62

feature errors emerged including: (1) vow el errors, (2) cons onant deletion, (3) letter doubling, (4) derivational suffix e rrors, and (5) whole word substitution. A discussion of the results follows in the order of the studys research questions. In addition, the error patterns th at emerged will be highlighted as they relate to current literature. Finally, the strengths and limitations of the study and further areas for research will be addressed. Developmental Sequence of Inflectional and Derivational Errors The first question addressed whether the differences between the spelling categories followed normal developmental patterns. Results of the repeated measures ANOVA and post hoc paired samples t-tests indicated that the categories of phonologic shifts and or thographic + phonologic shifts were significantly more difficult th an the other categories presented. Hence, children with ASD experienced difficulty with spelling morphologically complex words. 53 These findings were both similar a nd dissimilar to the developmental profile noted in Carlisle (2000). Carlisle classified words that contained shifts in phonology and orthography as less transparen t than words that did not contain shifts. Since her participants made mo re errors on words containing shifts in orthography and phonology, she concluded that the opaque nature of these shifts made it more difficult to identify the root word, and thus more difficult to spell the word. Like her study, the less tr ansparent phonologic (e .g., majority or children) and orthographic + phonologic sh ift (e.g., excellent or severity) categories in this study proved to be more difficult for the participants. This finding paralleled the pattern of developmen t observed in the children with typical

PAGE 63

development from Carlisles study. Unlik e Carlisles findings, the present study noted that words containing only orthogra phic shifts (e.g., scary or easily) were seen as being equal in difficulty to inflections (e.g., teaches or building) and no shift words (e.g., smoothly or highest). Ca rlisle (1988) noted that children learn inflections and derivations around the same time and master derivations later than inflections. Examination of the stimulus items used in the present study may provide a possible explanation for this devi ation. Two of the stimulus words in the inflection category also contained sh ifts in orthography related to letter doubling (e.g., stirring and stopped). For th e participants in this study, letter doubling proved to be an especially difficu lt task and was found to be one of the more common feature errors noted duri ng the POMAS analysis. This added difficulty may have inflated the spel ling complexity of the words in the inflectional spellings category making them as challenging as or thographic shifts. Spelling Accuracy and Measures of Literacy Achievement Positive correlations between performances on the experimental spelling protocol were revealed for both the word decoding and spelling subtests from the Woodcock Johnson III Test of Achievement (Woodcock, et al., 2001). This finding suggested a possible connect ion between word decoding and morphological spelling abilities in children w ith ASD. Research investigating the spelling abilities and the reading abilitie s of typically developing children has noted connections between reading and spelling skill (Caravolas, Hulme, and Snowling, 2001; Carlisle, 2000; Carlisle & Katz, 2006; Nagy et al., 2006). Specifically, that word decoding and comp rehension skills increase with increased 54

PAGE 64

morphological knowledge. For this reason, it was not surprising that the children with ASD in this study also showed connections between these skills. There was a correlation between the spelling performance on the standardized spelling subtests and the experimental spe lling measure. Traditional spelling tests, such as this, give inform ation about whether children are able to correctly spell common age appropriate word s. This may indicate that spelling assessments based on systematically sele cted morphological features yield some of the same information about the spel ling abilities of children with ASD as traditional spelling protocols. However, the analysis technique used in the present study also gave linguistic feature informa tion not available on traditional spelling measures. The examination of the qual ity of spelling errors along with the quantity of spelling errors provided more complete information about how children used their linguistic systems to spell. This knowledge could be used to further enlighten spelling instruction, by using specific error information to guide curriculum development. Spelling Accuracy and Age There was a small correlation (r = .373; r2 =.14) between age and the number of words spelled correctly on th e experimental spelling measure. This connection was expected since phonological, orthographic, and morphological awareness abilities increase as children age. As children grow, these processes become better assimilated resulting in more and more conventional spelling (Berninger & Fayol, 2008). The design of the current spelling protocol may provide an additional factor linking spelli ng ability and age since the experimental 55

PAGE 65

spelling measure was specifically designed to test morphologi cal abilities in spelling. Berninger & Fayol (2008) re ported in their longitudinal study of children in grades 1 through 7 that phonological and ort hographic awareness showed growth until grade 3. However, morphological awareness continued to grow after grade 3. Therefore, it app ears that morphological awareness spans a broader age range. Since this study focused on morphological skill, the design would possibly strengthen the connection be tween age and spelling ability in this study. Spelling Accuracy and Parents Level of Education The correlation between the parents level of education and performance on the spelling protocol was not significan t for fathers or mothers level of education. This finding was unexpected since research has indicated that maternal education level is often correlated with the childs academic test scores and reading ability (Magnuson, 2007; Da vis-Kean, 2005; Sirin, 2005). Since maternal education did not correlate with either the standardized subtests or the morphological spelling ability of the participant, it is possible that factors related to the language differences evidenced in children with ASD causes the maternal education to have less of an effect on the academic abilities of the child. There are two factors that could affe ct the parental influence on children with ASD. First, children with complex disabilities spend much less time engaged in literacy experiences (Mirenda, & Iacono, 2008). Second, the challenges in joint attention experien ced by children with ASD (Mundy & Newell, 2007), may serve to diminish the benefits from parental led literacy experiences. Therefore, it 56

PAGE 66

is possible that the benefits of parental e ducation, as it applies to literacy, may not be fully appreciated by the child with ASD because of lessened opportunities related to reduced exposure a nd internal factors associated with joint attention. Both a qualitative and quantitative an alysis was used when looking at the specific word categories presented to partic ipants in this study Use of both types of analysis made it possible to gain so me traditional information about spelling ability as well as some more uni que information about spelling skill. Homonym Patterns The goal in evaluating homonyms was to determine the participants ability to process semantic information. The category of homonyms contained the lowest frequency of errors. While it was the lowest frequency, this category still had a significant number of errors. Pa rticipants had 38.5% incorrect in the homonym category with 17 participants having difficulty with words in this category. The fact that the participants had difficulty was not surprising given the documented issues that children with ASD show integrating semantic information (Gaffrey, Kleinhaus, Haist, & Muller, 2007). In order to correctly spell words in the homonym category, participants had to integrate the semantic information provided by the oral and picture prompts with their ow n knowledge of encoding. An example of this task is to correctly spell the target clothes as in The boy went shopping for new clothes as close. This could illustrate less than adequate semantic integration and coor dination during sp elling tasks. While integrating semantic knowledge has been shown to be difficult for children with ASD, there were, in fact, several issu es that may have ac tually supported the 57

PAGE 67

students performance in this category. Fi rst, item selection for the experimental spelling protocol may have influenced the outcome. When examining homonym pairs, the researchers chose the less complex word in the pair (i.e., aloud, allowed ) for inclusion in the protocol. These words were also the most frequently used word in the pair. These factors may indicate that these words are easier to spell. If this is the case it is possible that more sema ntic errors would surface if more difficult homonyms were chosen as the stimuli. Finally, the inclusion of picture prompts in the pr esentation might have also facilitated performance. By using picture prompts, participants were given additional semantic information about which homonym correctly corresponded to the sentence prompt. This information ma y have been specifically helpful when deciding which homonym form to spell if the participant had two spelling forms in their repertoire. Additionally, res earch has shown that multimedia methods beyond language input increases comprehens ion in children with ASD (Chiang & Lin, 2007). Thus, it is possible that use of the computer combined with the use of picture prompts may have increased partic ipant attention and thus comprehension of the presented homonyms causi ng more correct answers. Patterns of Misspellings on Inflections and Derivations Examination of the participants performance showed diverse spelling abilities which could be subdivided into three patterns: (1) competent spellers making very few errors (n= 5, spelling 25% or less in correct), (2) those who had challenges with morphological concepts ( n =15, spelling 26%-74% of the words 58

PAGE 68

incorrectly), and (3) children who seemed to have challenges with spelling as a whole ( n= 9, spelling 75% or more of the word s incorrectly). These profiles are similar to other studies i nvestigating language and liter acy abilities in children with ASD which describe heterogeneous profiles (Anderson, et al., 2007; Nation et al., 2006). For instance, Nation et al (2006) noted four subgroups of reading ability in students with ASD: (1) gene rally good reading ability, (2) difficulty with reading comprehension, (3) difficulty reading non-words or nonsense words, and (4) children who have difficulty with reading non-words, real words, and reading comprehension. The patterns noted in this study are sim ilar to the profiles noted by Nation et al. (2006) in that there are students w ho have little difficulty, students who evidence specific challenges, and students who have overall difficulty. These profiles highlight the he terogeneity found in AS D. The fact that there is diversity on linguistic-based tasks would seem logical since ASD is a spectrum disorder, where symptoms can range from mild to seve re (Rutter, 2006). A control group using these same morphol ogically complex stimuli will provide further information as to whether these groupings are specific to children with ASD or possibly more analogous to those s een in typically deve loping children. Feature Error Analysis Several feature errors were dominant in this sample, including vowel errors, consonant deletion, letter doubli ng, derivational suffix errors and whole word substitutions. It was possible to isolate these specific features using the POMAS coding system. Some featur e errors were similar to normal developmental error patterns seen in othe r spelling investigations (e.g., Hauerwas 59

PAGE 69

& Walker (2003). For example the partic ipants demonstrated an understanding of how to spell the suffix before learning how to attach the ending to the root word (e.g., juicy spelled as juicey or attention spelled as attension ). Other patterns seem to exhibit the challenges with more complex morphology found in children with LLD, such as deleting morphological markers (e.g., kicked spelled as kick and juicy spelled as juice ). Other errors seemed more specific to this population (e.g., using whole word strategies such as argument spelled as areyoument or dangerous spelled as dayjres ). A discussion follows of the common feature errors noted in this study as they re late to current literature. Vowel errors. The most common feature errors produced by the participants in this study were vowel e rrors. This was easily foreseen. Unlike other languages, the English language doe s not have a one to one letter-sound correspondence with its vowels (Sun-Alperin & Wang, 2008). This makes learning to spell vowels a complex task (T reiman, 1993). When a speller attempts the complex task of spelling vowels, re searchers have noted the following to influence their selections: vowel interd ependency, the type and number of consonants used within a word (Tre iman, Kessler, & Bick 2002), and the frequency in which a vowel pattern occurs (Caravolas et al., 2001). Vowel errors in this study were identified using several feature codes from the POMAS coding system that were phonologi cal and or orthographic in nature. The six main type of vowel erro rs were: 1) long vowel errors ( n=26 code: PLV or 14.4%; easily spelled as esily), 2) phonological vocalic r errors ( n= 13 code: PVOCR or 7.2%; natural spelled as natual ), 3) incorrect representation for a 60

PAGE 70

schwa vowel ( n=27 code: OVE or 6.2%; severity spelled as suverity), 4) orthographic rhotic vowel errors ( n =13 code: OVr or 3.0%; stirring spelled as sterring ), 5) short vowel errors ( n= 48 code: POSV or 12.2%; confidence spelled as confedence ), and 6) missing vowel errors ( n= 40 code: POVM or 10.1%; building spelled as blden scary spelled as scrry or majority spelled as mgd ). While word frequency was controlled in the experimental protocol, the stimulus items were chosen to stress ch ildrens morphological sk ills. Because of this process, more complex and possi bly less frequently used words were included. Since the spelling of vowels is sensitive to word frequency, this selection process could have increased the errors in this categ ory. For example, errors coded as OVE (e.g. severity spelled as suverity ) and POSV ( confidence spelled as confedence ) could have occurred because the student was unfamiliar with how the particular vow el or vowel pattern was supposed to be represented. This unfamiliarity may be due to the fre quency of occurrence or their exposure to that particular vowel spe lling. The changes in pronunciation from the root word to the derived word added another facet of complexity to these examples and others. Other vowel errors involved the incorrect spelling of the rhotic r. In some cases, an incorrect vowel-consonant combina tion was used to re present the rhotic r, as in stirring spelled as sterring. According to research, the most common way that typically developing children sp ell the rhotic vowel is by using er, which would indicate that this error is simila r to spelling errors seen in typically developing children (Reece & Treiman, 2001) The participants showed an 61

PAGE 71

additional error pattern by deleting the r from the r-colored vowel combination, as in natural spelled as natual The deletion of the post-vocalic r may be related to the children using a phonologic strate gy in which the r-colored vowel is represented by one sound. In this case, wh ere the syllabic r is less stressed, the participants chose to represent the vowel. Research indicates that as spellers become more proficient and have more e xposure to print they recognize that more than one grapheme can represent a single sound (Reece and Treiman, 2001). This would indicate a spelling strategy similar to that of a typically developing less proficient speller. There were also cases where the participant deleted vowels from a word, which was coded as POVM. This error is common in the early spelling of kindergartenage children. Children just learni ng to spell will often represent car as cr (Gentry, 1982). This type of spelling shows a heavy reliance on phonological knowledge. In the same respect, some participants ma de long vowel errors that resulted in the vowel pattern changing from a long to a short vowel sound (i.e., easily spelled as esily). In these cases, the participants app eared to be using the vowels so that the single vowel says its name; ignoring the n eed for a vowel pattern. This is similar to using the consonant to say its name and deleting the vowels around it. These errors in the present stud y may indicate that some participants with ASD are displaying spelling proficiency commens urate with younger typically developing spellers. Consonant deletion. A prominent error noted in this study was the deletion of consonants from words. Ther e were 31 instances of participants 62

PAGE 72

deleting random consonants from words (n=31 with the code: PCD or 17.2%). Additionally, there were almost an equal number of instances where participants deleted consonants from sonor ant consonant clusters ( n=27 with the code PSONC or 15.0%). In these cases, the particip ants would delete the sonorant following the vowel and preserve the consonant that followed the sonorant. Research with typically developing children suggests th at, when the child has a difficult time identifying the sonorant, they blend this sound with the vowel omitting it in their spelling (Treiman, Zukowski, & Richm ond-Welty, 1995). The children in this study followed this pattern with fe w exceptions. They had difficulty discriminating the vowel and sonorant sounds and as a result deleted these sounds. Since this finding was also noted in typically developing children, it appears that the participan ts in this study are making some errors similar to typically developing students. Letter doubling. The spelling protocol cons isted of eight words that required the use of double consonants. In some cases, the root words contained double consonants (n= 6) and, in others, the consonant needed to be doubled according to the syllable juncture rule (n=2). Errors made regarding double consonants occurred both because particip ants failed to double a consonant (n=21 code: OCD or 4.9%) and/or inco rrectly doubled a consonant ( n=65 code: OLD or 15.0%). The appropriate use of double consonants in spelling requires integration of phonological knowledge to translate the phonemes to their correct grapheme counterparts, orthographic knowledge to de termine the appropriateness of using 63

PAGE 73

two of the same grapheme, and morphological knowledge to identify if the double consonant is included in the root word or needs to be added as part of a derivational or inflectional shift. These skills are developed through repeated exposure to print (Treiman & Cassa r, 1997). Moreover, when doubling consonants, children are more likely to follow legal letter dou bling patterns found in the English language even when they ar e taught illegal pattern s, such as initial consonant doubling (Wright & Ehri, 2007). Therefore, repeated exposure to words that have legal lett er doubling appears to trump the explicit teaching of the illegal letter doubling. The children in the present study appeared to have difficulty determining when it was a ppropriate to double consonants while encoding words. The significant challe nges with letter doubling when double letters occurred in the root portion of the word, in the present study could be influenced by a lack of experience with pr int. This would significantly decrease the frequency of exposure to words whic h feature letter doubling. Children with ASD often do not have the benefit of c onsistent scope and sequence in their literacy instruction (Mirenda & Iacono, 2007). As a resu lt, their performance may reflect challenges in integrating their phonological, orthographic, and morphological knowledge that results from repeated exposure to accurate use of letter doubling. While lack of exposure to words with double letters may also explain doubling errors that occurred at the sylla ble juncture, other studies have shown this type of errors to be developmenta l. A study that used the POMAS coding system also noted issues re lated to letter doub ling (Bahr, Silliman, & Berninger, 64

PAGE 74

2009). Taking this information into consid eration with the results of the present study, it is also possible that difficulty with letter doubling is part of typical development that occurs as children are learning to integrate orthographic knowledge. A study with an age-matched control group using the same corpus of words would give more information about these findings. Derivational suffix errors. As stated before, the corpus of words used in this study contained derivational shif ts that were both orthographic and phonologic in nature. That is, orthographi c shifts (alterati ons in spelling), phonologic shifts (changes in pronunciati on), and orthographic +phonologic shifts (changes in both spelling and pronunciation) occurred when a suffix was added to the root form of the word. A frequent error in this study dealt w ith the ability to correctly spell the derivational suffix. There were mi sspellings of both inflectional ( n = 34 code OSUFI or 7.9%) and derivational endings ( n = 65 code OSUFD or 15.0%). Errors were distributed widely across th e stimuli. It has been suggested that typically developing children first learn the suffix of a word and later learn how to attach the suffix to the root word (H auerwas & Walker, 2003; Nunes, Bryant, & Bindman, 1977). This pattern was frequently displayed by the participants in this study who both seemed to have knowledge of how to spell the root word and those who did not (e.g. scary spelled as scarey or argument spelled as rgyment) However, the opposite pattern featuring inco rrect spellings of the suffix was also displayed by participan ts in this study (e.g. discussion spelled as discushen or windy spelled as windie ). Larkin and Snowling (2 008) noticed this pattern 65

PAGE 75

(incorrect spelling or omission of suffixes), in the 5, 6, and 7 year old spellers in their study on morphology. In their study, ev en if a child misspelled the suffix, it was considered to be correct and counted as though the child represented the second morpheme of a two morpheme wo rd. The reason for this was the low number of children representing the second morpheme (i.e., incorrect or correct spelling of the second morpheme) as compared to the number of students that did not represent the second morpheme (i.e., omission of the suffix with the root spelled correctly or incorrectly). Added to this were comparisons to the spellings of one morpheme words. It was conjectur ed that children were using little to no morphological knowledge. Walker & Ha uerwas (2006) found a similar error pattern in their children with spelling defic its. The frequency of this type of error compared to typically developing children helped delineate the children with spelling deficits. In summary, some participants in the present study appear to have knowledge about how suffixes are spelled, but not about how to attach them to root words (e.g. scary spelled as scarey). Others appear to have difficulty with representing the suffixes and the root words in their spellings. Both patterns are suggestive of difficulties with complex morphology. In the case of being able to spell the root form and not the suffix, the participant illu strates difficulty with integrating their morphological knowledge with their ort hographic knowledge to aid them in appropriately attaching the morpheme to the root word. Conversely, in the case where participants have difficulty representing both parts, they maybe illustrating difficulty with manipulating morphemes a ll together. The qualitative analysis 66

PAGE 76

done here revealed that the errors di splayed do not solely represent younger developmental patterns. Whole word substitution. A small group of participants ( n=14) inserted whole words within their spelling of a word. The inserted word was often phonologically similar to part of the target word. An example of this was a child who spelled argument as areyoument and discussion as disguession. In using this pattern, the participants ma y be illustrating a lack of acknowledgement that that these are two separate words with two di fferent meanings. The participants may be relying on what they know matche s phonetically rather than integrating morphological information about what the inserted word means versus what the prompted word within the sentence shoul d mean. This suggests that some of these participants had problems integrating semantic information during the spelling task. Studies have noted that ch ildren with ASD have difficulties with semantic decision tasks and do not benefit from semantic priming like their typically developing peers (Kamio, Robi ns, Kelley, Swainson, & Fein, 2007). Additionally, children with ASD have b een noted to neglect the semantic relatedness of words. For instance, Just, Cherkassky, Keller, & Minshew (2004) showed more brain activation in Wernic kes area than in Brocas area during reading tasks. This pattern of activation is typically associated with a tendency to process single words and not the semantic relatedness of words. Almost half of the participants in this study (n=14) illust rated semantic difficu lties through use of the whole word strategy. 67

PAGE 77

It is also possible that the us e of known whole words reflect the associational challenges f ound in children with ASD (Cherkassky, Kana, Keller, & Just, 2006; Kamio, Robins, Kelley Swainson, and Fein, 2007). These challenges, related to functional connectiv ity, may make it difficult to integrate phonological, orthographic, and morphologic knowledge during ta sks that require higher semantic demand, such as enc oding morphologically complex words and derivational shifts. The result is a sp elling that illustrated less semantic connection to the actual target. As prev iously stated, there were several students ( n=17) who made incorrect semantic choices when spelling homonyms. The majority of this gr oup of participants ( n=14) were also the participants who used the whole word strategy when spelling. T hus, with the feature errors present in this study, it seems possible that some ch ildren with ASD show semantic neglect within the encoding process. In conclusion, one of the main goals of this study was to determine how children with ASD were applying kn owledge about phonology, orthography, and morphology to the task of encoding word s. Results revealed that their performance was heterogeneous across part icipants with thre e patterns emerging. Despite the numerous quantity and quality of errors present in this study, there was still a group of participan ts that made few errors. The quality of these errors was often noted to be very specific requi ring only one error code. There were also a group of participants that stru ggled with complex morphological issues. This group made more errors but they we re often related to issues with letter doubling, representing vowel errors in the presence of phonological shifts, and 68

PAGE 78

correctly manipulating morphemes in the pr esence of derivational and inflectional shifts. Lastly, there was a group of ch ildren who were generally challenged spellers and made the largest number of errors. Along with illustrating some of the challenges seen in the other two groups, this group tended to deleted consonants from words, omit suffixes, and use whole word substitutions. Further research is still needed to further defi ne the possibilities of spelling subgroups in this population. Implications for Literacy Education Based on the data presented here, several feature errors were noted to be difficult for children with ASD. Th ese feature errors included: vowel errors, consonant deletion, letter doubling, deriva tional suffix errors, and whole word substitution. While further research is still needed, this information aids in determining the areas of enc oding that are difficult for children with ASD. It is this type of information that can be used to develop instructio nal strategies that may help to bridge performance gaps for atypical populations. Examples of instruction that may benefit children w ith these difficulties could include teaching children how root words connect to morphological endings and how to orthographically represent this info rmation. Additionally, highlighting connections between words, such as how the word magic relates to the word magician, may not only increase spelling abili ty, but reading comprehension as well. Knowledge about how words are re lated and how the addition of suffixes 69

PAGE 79

and prefixes changes the words meaning may help children with ASD interpret and better comprehend words when reading. With only the information from this study, it is difficult to determine if these st rategies would be effective. However, with more research it may be possible to develop instructional programs and curriculum that can better illustrate the complex morphological concepts that children with ASD find difficult. Strengths and Limitations Sample One limitation concerning the study sample was the sample size (N = 29 participants). However, this sample size is larger than many ASD studies. A strength of the sample used in this study is that while information was collected concerning subtype of ASD, this information was not used in analysis of the data. Research has indicated that diagnosticians are generall y good at determining if a child has an ASD, but when it comes to determining the subtype category, Autism, Aspergers Disorder, or PDD-NOS there is significant disagreement between diagnosticians (Klin, Lang, Cicchet ti, & Volkmar, 2002). In light of these research findings, the investigator did not conduct an analysis based on subtype. Rather, all particip ants were included together broadly as having an ASD. Another addition that would strengthen the sample of this study is the use of a control group made up of typically de veloping age-matched peers. Currently, a companion study is being developed using age-matched control participants, adding the control group would make it possible to compare data and determine if the patterns seen here are specific to child ren with ASD or specific to the use of more morphologically complex words types. 70

PAGE 80

Spelling inventory. A strength of the spelling inve ntory is that it was balanced according to frequency of the words within the literature for the grade levels of the participants. Given the importance of word frequency, this was considered vital to the task. Additionally, the use of both sentence and picture context were included. This procedure gave children the opportunity to use more than one modality to determine the meaning of a word. This was especially salient in the homonym section. Children did not just have to rely on their hearing of the word, but could use visual knowledge as well. Ho wever, as previously noted, the picture prompts could have inflated performance on homonym stimulus items since pictorial semantic information was depicted pictorially. Another strength of the study was the way in which the spelling inventory was presented. By utilizing EcosWin software, it was insured that all participants received the words in the exact same form at. This kept presentation from being a factor that affected the outcome of the data. Scoring systems. One difficulty with the scori ng system used in this study was the complex nature of the error c odes. This made extensive training necessary for reliability. However, use of the POMAS scoring system strengthened this study because it allowed for an examination of error patterns and not just overall correct performance. Since the POMAS defines errors based on the linguistic features of the words, id entification of specific error patterns was facilitated. This gave the researcher more information than just how many words were spelled correctly and incorrectly, and allowed a d eeper understanding of the challenges related to mor phological complex words. 71

PAGE 81

Possible Directions for Future Research There are still many questions regard ing the spelling abilities of children with ASD and how they develop these skills. Future research should include a control group of age-matched typically developing children. This will make it easier to determine if patterns are specific to children with ASD or are seen across all children when they are faced with morphologically complex spelling tasks. While there is spelling res earch with typically devel oping children, these studies do not use the same morphological test items as presented in this study. Additionally, these studies ra rely test morphology as stri ngently as in the present study. A study using the same protocol would allow insight into how typically developing children handle the specific challenges presented by homonyms and derivations. Specifically, it w ould be possible to compare quality of errors. This would especially be helpful with errors th at appear to be related to difficulties with semantic integration. If typicall y developing students who are capable of making semantic connections made similar errors to children with ASD, it would suggest that difficulties with semantic in tegration at the word level are typical when testing spelling in this way or for these particular words. Since children with ASD are noted to have difficultie s with functional connectivity, an agematched control study would help to determ ine if limits in f unctional connectivity makes it difficult for children with ASD to spell morphologically complex words. Further studies should also incl ude more in-depth information about how children with ASD unders tand derivational and inflectional morphology. As discussed, the current study on ly asked that the children spell the 72

PAGE 82

derived forms of words. No testing was done to determine if the children were able to discern the root word or spell the root word in is olation. Deacon and Bryant (2006) found that children were better able to spell words in a sentence context when a clue indicating how to sp ell the root word was given, as opposed to when one was not given. This study indi cated that simple combinations of root words and suffixes were easily accomplished when children had knowledge of the root. This task became more difficult as morphological shifts in orthography and phonology were added to the process. These findings illustrate the impact that familiarity with the root forms of words has on spelling ability. In this investigation subgroups of perf ormance were noted including students who have little difficulty, students who evid ence specific challenge s, and students who have overall difficulty. In order to de termine if this type of heterogeneous performance is common among children with ASD it will be necessary to repeat this investigation with a different gro up of participants. It would also be necessary to incorporate a la rger group of participants to further identify any other differences between these groups. This larger group would allow for more specific identification of profiles of performan ce within these groups. 73

PAGE 83

References Allik, H. L., Larson, J., & Smedje, H. (2006). Sleep patterns of school-age children with aspergers disorder syndrome or high-functioning autism. Journal of Autism and Developmental Disorders, 36, 585-595. American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author. Anderson, D., Lord, C., Risi, S., DiLavor e, P., Shulman, C., Thurm, A., et al. (2007). Patterns of growth in verbal abilities among children with autism spectrum disorder. Journal of Consulting and Clinical Psychology 75(4), 594604. Apel, K., & Masterson, J. (2001). Th eory-guided spelling assessment and intervention: A case study. Language, Speech, and Hearing Services in Schools 32, 182-195. Bahr, R. H., Silliman, E. R., & Berni nger, V. (2009). What spelling errors have to tell about vocabul ary learning? In C. Wo od & V. Connelly (Eds). Reading and spelling: Contemporary perspectives (pp. 177-210) London: Routledge. Berninger, V., & Fayol, M. (2008). W hy spelling is important and how to teach it effectively. Encyclopedia of language and literacy development (pp. 1-13). London, ON: Canadian Language and Literacy Research Network. 74

PAGE 84

Bernstein, S. E. (2009). Phonology, decoding, and lexical compensation in vowel spelling errors made by children with dyslexia. Reading & Writing, 22 307-331. Blischak, D., & Schlosser, R. (2003). Use of technology to support independent spelling by students with autism. Topics in Language Disorders, 23, 293-304. Bourassa, D., & Treiman, R. (2003) Spelling in dyslexic children: Analyses from the Treiman-B ourassa Early Spelling Test. Scientific Studies of Reading, 7, 309-333. Bourassa, D., & Treiman, R. (2008). Mo rphological constancy in spelling: A comparison of children with dyslexi a and typically developing children. Dyslexia: An International J ournal of Research and Practice 14, 155-169. Bregman, J. D. (2005). Definitions and characteristics of the spectrum. In D. Zager (Ed.), Autism spectrum disorders: Id entification, ed ucation, and treatment (3rd ed., pp. 10-11). Mahwah, NJ: Lawrence Erlbaum & Associates. Bruck, M. (1993). Component spelling sk ills of college students with childhood diagnosis of dyslexia. Learning Disability Quarterly, 16 171-184. Caravolas, M., Hulme, C., & Snowli ng, M. J. (2001). The foundations of spelling ability: Evidence from a 3-year longitudinal study. Journal of Memory and Language, 45, 751-774. Carlisle, J. (1988). Knowledge of derivational morphology and spelling ability in fourth, sixth, and eighth graders. Applied Psycholinguistics, 9 247-266. 75

PAGE 85

Carlisle, J. (2000). Awareness of the structure and meaning of morphologically complex words: Impact on reading. Reading and Writing: An Interdisciplinary Journal, 12, 169-190. Carlisle, J. (2004). Morp hological processes that influence learning to read. In C. A. Stone, E. R. Silliman, B. J. Ehren, & K. Apel (Eds.), Handbook of language and literacy (pp. 318-339). NY: Guilford Press. Carlisle, J., & Katz, L. (2006). Effect s of word and morpheme familiarity on reading of derived words. Reading and Writing 19, 669-693. Cassar, M., Treiman, R., Moats, L., Curry Pollo, T., & Kessler, B. (2005). How do the spellings of children with dys lexia compare with those of nondyslexic children? Reading and Writing, 18 27-49. Center for Disease Control. (2007, February). CDC Releases New Data on Autism Spectrum Disorders (ASDs) from Multiple Communities in the United States. Retrieved January 2, 2009, from http://www.cdc.gov/media /pressrel/2007/r070208.html Cherkassky, V., Kana, R., Keller, T., & Just, M. (2006). Functional connectivity in a baseline rest ing-state network in autism. Neuroreport: For Rapid Communication of Neuroscience Research, 17(16), 1687-1690. Chiang, H., & Lin, Y. (2007). Reading comprehension instruction for students with autism spectrum disorders: A review of the literature. Focus on Autism and Other Developmental Disabilities, 22 259-267. 76

PAGE 86

Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and Psychological Measurement 20 37. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Mahwah, New Jersey: Lawrence Erlbaum. Davis-Kean, P. (2005). The influence of parent education and family income on child achievement: The indirect role of parent expectations and the home environment. Journal of Family Psychology, 19, 294. Deacon, S. H. & Bryant, P. (2006). This turnips not for turning: Childrens morphological awareness and thei r use of root morphemes in spelling. British Journal of Developmental Psychology, 24 567-575. Egan, J. & Pring, L. (2004). The processing of inflectional morphology: A comparison of children with and without dyslexia. Reading and Writing: An Interdisciplinary Journal, 17 567. Ehri, L.C. (1986). Sources of difficulty in learning to spell and read. Advances in developmental and behavioral pediatrics 7, 121-195. Gaffrey, M. S., Klienhans, N., Haist, & Mller, R. A. (2007). Reduced left inferior frontal and enha nce occipital activation during semantic decision in autism: An fMRI study. Neuropsychologia 45, 1672-1684. Gentry, J. (1982). An analysis of developmental spelling in GNYS AT WRK. Reading Teacher, 36, 192-200. Grigorenko, E. L., Klin, A., Pauls, D. L., Senft, R., Hooper, C., & Volkmar, F. (2002). A descriptive study of hyperlexia in a clinically referred 77

PAGE 87

sample of children with developmental delays. Journal of Autism and Developmental Disorders, 32, 3-12. Grigorenko, E. L., Klin, A., & Volkmar, F. (2003) Annotation: Hyperlexia: Disability or superability? Journal of Child Psychology and Psychiatry, 44, 1079-1091. Hauerwas, L., & Walker, J. (2003). Sp elling of inflected verb morphology in children with spelling deficits. Learning Disabilities Research & Practice 18, 25-35. Henderson, E. (1985). Teaching spelling. Boston: Houghton Mifflin. Just, M., Cherkassky, V., Keller, T ., & Minshew, N. (2004). Cortical activation and synchroni zation during sentence comprehension in highfunctioning autism: Evidence of underconnectivity. Brain: A Journal of Neurology 127, 1811-1821. Kamio, Y., Robins, D., Kelley, E., Swainson, B., & Fein, D. (2007). Atypical lexical/semantic processing in high-functioning autism spectrum disorders without early language delay. Journal of Autism and Developmental Disorders, 37 1116-1122. Kinney, E.M., Vedora, J., & Stromer, R. (2003). Computer-presented video models to teach generative spelli ng to a child with an autism spectrum disorder. Journal of Positive Be havior Interventions, 5, 22-29. Klin, A., Lang, J., Cicchetti, D. V., Volkmar, F. R. (2002). Inter-rater reliability of clinical diagnosis and DSM-IV criteria for autistic disorder: Results 78

PAGE 88

of the DSM-IV autism field trial. Journal of Autism and Developmental Disorder, 30, 163. Larkin, R. F. and Snowling, M. J. (2008). Morphological spelling development. Reading and Writing Quarterly 24 363 376. Lyon, G. R., Shaywitz, S. E., & Shaywitz, B. A. (2003). A definition of dyslexia. Annals of Dyslexia, 53 1-15. Magnuson, K. (2007). Maternal education and children's academic achievement during middle childhood. Developmental Psychology, 43 (6), 14971512. Masterson, J. & Apel, K. (2006). Effect of modality on spelling words varying in linguistic demands. Journal of Developmental Neuropsychology, 29, 261-277. Mirenda, P. (2003). 'He's not really a reader...': Perspectives on supporting literacy development in individuals with Autism. Topics in Language Disorders 23, 271-282. Mirenda, P. & Iacono, T. (2008). Autism spectrum disorders and AAC. Baltimore, MD: Brookes Publishing. Mundy, P. & Newell, L. (2007). Atten tion, joint attention, and social cognition. Current Directions in Psychological Science, 16 (5), 269-274. Nagy, W., Berninger, V. W., & Abbott, R. D. (2006) Contributions of morphology beyond phonology to literacy outcomes of upper elementary and middle-school students. Journal of Educational Psychology, 98 134-147. 79

PAGE 89

Nation, K., Clarke, P., Wright, B., & Williams, C. (2006). Patterns of reading ability in children w ith autism spectrum disorder. Journal of Autism and Developmental Disorders, 36, 911-919. No Child Left Behind Act of 2001, P.L. 107-110, 115 Stat. 1425 Nunes, T., Bryant, P., & Bindma n, M. (1977). Morphological spelling strategies: Developmental stages and processes. Developmental Psychology 33, 1001-1113. O'Connor, N., & Hermelin, B. (1994) Two autistic savant readers. Journal of Autism and Developmental Disorders 24, 501-515. OConnor, I. M. & Klein, P. D. (2004) Exploration of strategies for facilitating the reading comprehension of high-functioning students with autism spectrum disorders. Journal of Autism and Developmental Disorders, 34, 115127. Reece, C., & Treiman, R. (2001). Children's spelling of syllabic /r/ and letter-name vowels: Broadening th e study of spelling development. Applied Psycholinguistics 22, 139-165. Roseberry-McKibbin, C. (2007). Introduction to language disorders in children: A multicultural approach. Boston, MA: Allyn & Bacon. Rutter, M. (2006). Autism: Its reco gnition, early diagnosis, and service implications. Journal of Developmental & Behavioral Pediatrics, 27 54-58. Saldana, D. & Frith, U. (2006). Do readers with autism make bridging inferences from world knowledge? Journal of Experimental Child Psychology, 96, 310-319. 80

PAGE 90

Schlosser, R. W., & Blischak, D. M. (2004). Effects of speech and print feedback on spelling by children with autism. Journal of Speech, Language, and Hearing Research, 47 848-862. Schlosser, R.W., Blischak, D. M., Belf iore, P. J., Bartley, C., & Barnett, N. (1998). Effects of synthetic speech output and orthographic feedback on spelling in a student with autism: a preliminary study. Journal of Autism and Development Disorders, 28, 309-319. Silliman, E., Bahr, R., & Peters, M. (2006). Spelling patterns in preadolescents with atypical language skills: Phonological, morphological, and orthographic factors. Developmental Neuropsychology 29, 93-123. Sirin, S. R. (2005). Socioeconomic status and academic achievement: A meta-analytic review of research. Review of Educational Research, 75 417-453. Sun-Alperin, M., & Wang, M. (2008) Spanish-speaking childrens spelling errors with English vowel so unds that are repres ented by different graphemes in English and Spanish words. Contemporary Educational Psychology, 33, 932-948. Treiman, R. (1993). Beginning to spell: A study of first-grade children. New York: Oxford University Press. Treiman, R., & Bourassa, D. (2000). The development of spelling skill. Topics in Language Disorders 20, 1-18. Treiman, R., & Cassar, M. (1997) Spelli ng acquisition in English. In C. A. Perfetti, L. Rieben, & M. Fayol (Eds.), Learning to spell: Research, theory, and 81

PAGE 91

practice across languages (pp. 61-80). Mahwah, NJ: Lawrence Erlbaum & Associates. Treiman, R., Cassar, M., & Zukowski, A. (1994). What types of linguistic information do children use in spelling? The case of flaps. Child Development, 65, 1310-1329. Treiman, R., Kessler, B., & Bick, S. (2002). Context sensitivity in the spelling of English vowels. Journal of Memory and Language, 47 448-468. Treiman, R., Weatherston, S., & Berc h, D. (1994). The role of letter names in childrens learning of phoneme-grapheme relations. Applied Psycholinguistics, 15, 97-122. Treiman, R., Zukowski, A., & Richmond-Welty, E. D. (1995). What happened to the n of sink? Childrens spellings of final consonant clusters. Cognition, 55 1. Tsesmeli, S. N. & Seymour, P. H. K. (2006). Derivational morphology and spelling in dyslexia. Reading and Writing, 19 587-625. U.S. Department of Education. (200 4). Twenty-sixth annual report to Congress on the implementation of the I ndividuals with Disa bilities Education Act. Washington, DC: Author. Walker, J., & Hauerwas, L. ( 2006). Development of phonological, morphological, and orthographic knowle dge in young spellers: The case of inflected verbs. Reading and Writing, 19 819-843. Welsh, M. C., Pennington, B. F., & Rodgers, S. (1987). Word recognition and comprehension skills in hyperlexic children. Brain and Language, 32, 76-96. 82

PAGE 92

Windsor, J., Scott, C., & Street, C. (2000). Verb and noun morphology in the spoken and written language of childre n with language learning disabilities. Journal of Speech, Language, and Hearing Research 43, 1322-1336. Woodcock, R. W., Mather, K. S., & Wendling, N. (2001). Woodcock Johnson III Test of Achievement. Riverside Publishing Company Wright, D., & Ehri, L. (2007). Begi nners remember orthography when they learn to read words: The case of doubled letters. Applied Psycholinguistics 28, 115-133. Zeno, S., Ivens, S., Koslin, S., Zeno, B. (1995).Touchstone Applied Science Associates, National Institute of Child Health and Human Development (U.S.), & New York State Sc ience and Technology Foundation. The educators word frequency guide. United States: Touchstone Applied Science Associates. 83

PAGE 93

84 Appendices

PAGE 94

Appendix A Informed Consent 85 Appendix A: Continued 1. Our Study: The University of South Florida is allowing us to carry out a study that focuses on the spelling abilities of children on the Autism Spectrum. We will be considering how the level of difficulty of a word affects this group of childrens ability to spell. The results from this test could have a very important impact on what is known about spelling in children with autism. We plan to enroll a large group of children who have been diagnosed with Autism, PDD-NOS, and Aspergers Disorders syndrome. Your child will be asked to complete several short screenings to assess their speech/ language skills. They will also be given a hearing screening. Then, they will be asked to complete a spelling test via computer, and they will get a sticker or small toy after the session. We will complete these tests at a convenient location and time for you. It will take about 1.5 hours to be completed in two sessions. Your childs responses to these screening tests and activities will be kept strictly confidential. We hope you can help! We hope to enroll 50 children ages 8-15 and in grades 2nd thru 9th from several locations for this study. 3. Confidentiality: All information we have about your child will be completely confidential. Only the few people working on the study and authorized personnel, USF Institutional Review Board and its staff, the United States Dept. of Health and Human Services, the Florida Dept. of Health, and any other individuals acting on the behalf of USF, may inspect the records from this research project. The test material gathered from your child will be assigned a numerical code to make sure that your child cannot be identified. Any reports about our study will talk about groups of children and will not include the real names of any child or the school that your child attends. Any test scores resulting from this study will only be given to parents and only upon their request. No information about individual students will be given to the classroom teacher. Participation will in no way affect your childs academic standing. 2. Benefits of the Study: While your child will not directly benefit from participating in this study, there are benefits. The implications from this study could show important information about spelling abilities for the autistic population. Such information is currently limited. In addition academic information about your childs performance on the spelling test will be available to you by written report if requested. Participating in this study presents i k hild 4. Instructions: Please read and sign the agreement form on the reverse side and return it to your childs teacher by ________________________. We really appreciate your help in allowing your child to be part of this important study. If you have any questions, please call Khalyn Wiggins at 850-264-3884. If she is not there, please leave her a message. You can also email her at kwiggins@mail.usf.edu.

PAGE 95

86 Appendix A (Continued) 5. Agreement Form: [ ] I give my permission for my child to be included in this study. If you sign, it means that you agree to let your child participate in the study of spelling abilities. You and your child are free to stop participating at any time, without question. Parents signature: ______________________ Parents printed name:____________________________ Date:_____________________________ Home phone number: __________________________ Childs name and date of birth:_____________________________ 6. Diagnostic information: Please circle your childs formal diagnosis and who gave this diagnosis: Diagnosis: Aspergers Disorders Autism PDD-NOS Diagnosed by: Physician Speech Language Pathologist Social Worker Psychologist Psychiatrist Other Please indicate parents highest level of education: Mother ________________ Father__________________ Guardian________________ 6. Institutional Approval of Study and Informed Consent This research project and informed c onsent form were reviewed and approved by the University of South Florida Institutional Review Board for the protection of human subjects. This approval is valid until the date provided below. The Board may be contacted at (813) 974-5638.

PAGE 96

Appendix B Word Frequency by Category (Zeno et al., 1995) Homonyms Word Page # U value SFI value 2 3 4 5 6 7 8 9 Aloud 30 18 52.7 1 24 25 23 18 17 18 17 Four 122 349 65.4 249 293 326 339 375 384 396 399 Week 294 141 61.5 133 193 125 132 149 141 139 133 Clothes 68 126 61 192 179 181 155 130 127 117 122 Sent 241 179 62.5 123 139 164 212 223 221 224 215 Wear 294 82 59.2 106 108 111 97 85 84 80 78 No Shift Word Page # U value SFI value 2 3 4 5 6 7 8 9 Dangerous 84 73 58.7 35 57 62 78 87 90 86 81 Friendship 123 15 51.9 5 8 9 13 17 17 16 17 Assignment 38 13 51.4 1 6 6 7 8 9 10 12 Smoothly 250 15 51.8 4 9 13 18 19 18 18 17 Highest 139 44 56.5 7 13 20 40 42 42 45 45 Windy 297 6 48.1 11 13 13 7 4 3 3 3 87

PAGE 97

Appendix B (Continued) Phonological Shift Word Page # U value SFI value 2 3 4 5 6 7 8 9 Different 91 703 68.5 413 554 599 670 713 704 711 716 Disappear 92 18 52.7 10 18 18 18 20 21 20 20 Majority 171 41 56.2 0 0 0 2 9 15 20 24 Convertible 77 1 41.9 0 2 2 1 1 0 0 0 Children 64 478 66.8 503 483 440 469 450 437 421 422 Discussion 93 44 56.5 1 5 6 11 16 19 23 23 Orthographic Shift Word Page # U value SFI value 2 3 4 5 6 7 8 9 Argument 36 31 55 5 7 12 11 13 15 16 16 Attention 39 127 61 43 58 74 89 113 120 126 123 Juicy 157 3 45.2 4 4 3 4 4 2 2 1 Hungry 143 62 57.9 166 127 112 79 63 56 49 45 Easily 100 134 61.3 43 67 88 104 123 122 129 146 Scary 237 5 47.2 22 8 7 6 5 5 3 2 88

PAGE 98

Appendix B (Continued) Inflections Word Page # U value SFI value 2 3 4 5 6 7 8 9 Cries 81 10 50.2 15 11 10 10 12 12 11 10 Stirring 260 9 49.7 5 5 10 14 12 10 10 9 Stopped 260 183 62.6 442 367 310 229 192 175 157 142 Kicked 159 12 51.1 31 28 23 18 17 14 12 10 Building 54 192 62.8 149 130 165 172 177 183 201 205 Teaches 269 5 47.3 4 4 4 5 4 5 5 5 Orthographic + Phonological Shift Word Page # U value SFI value 2 3 4 5 6 7 8 9 Pleasant 206 44 56.5 23 36 43 51 58 58 50 57 Excellent 109 36 55.6 2 7 15 26 31 36 40 43 Student 262 82 59.1 10 20 22 33 43 46 51 56 Natural 185 184 62.7 16 34 52 119 145 159 116 9 191 Severity 242 2 44.1 0 0 0 1 1 1 1 Confidence 73 25 54 3 6 10 13 16 17 18 21 89

PAGE 99

Appendix B (Continued) Type= number of different words; Tokens= wo rds in the corpus; U= the frequency of the type per million tokens weighted by the words use across different subject areas; SFI= a logarithmic transformation of U maki ng the U value easier to use (e.g., SFI of 88.3= frequency of 67,500 per million); 2-9= The U va lue per grade level text in the corpus 90

PAGE 100

Appendix C Stimuli The dog waited for his food. Spell the word dog. The cats played with the ball of yarn. Spell the word cats. The woman read her speech aloud. Spell the word aloud. The answer to the math problem was four. Spell the word four. Her calendar was full for the week. Spell the word week. The boy went shopping for new clothes. Spell the word clothes. Her mom sent her to but some fruit. Spell the word sent. She wanted to wear the new boots to school. Spell the word wear. 91

PAGE 101

Appendix C (Continued) The mother helped her son with his assignment. Spell the word assignment. The interview went smoothly. Spell the word smoothly. The Eiffel tower is the highest point in Paris. Spell the word highest. During a tornado it is very windy. Spell the word windy. A diamond and an oval look very different. Spell the word different. The girls had a very close friendship. Spell the word friendship. The magician made the rabbit disappear. Spell the word dissappear. The majority of the faces were smiling. Spell the word majority. 92

PAGE 102

Appendix C (Continued) The red convertible was for sale. Spell the word convertible. The children were playing ball. Spell the word children. The employees were having a discussion. Spell the word discussion. The two boys had an argument. Spell the word argument. When you are driving you have to pay close attention. Spell the word attention. The boy took a bite of the juicy apple. Spell the word juicy. At lunch time the girl was hungry. Spell the word hungry. The girl easily ate her lunch and worked at the same time. Spell the word easily. 93

PAGE 103

Appendix C (Continued) The boy wore a scary mask. Spell the word scary. The baby cries a lot. Spell the word cries. The girl is stirring the soup. Spell the word stirring. The car stopped at the light. Spell the word stopped. The girl kicked the ball. Spell the word kicked. The woman had the plans for the new building. Spell the word building. The man teaches the woman how to make hamburgers. Spell the word teaches. The nice weather at the beach made it a pleasant day. Spell the word pleasant. 94

PAGE 104

Appendix C (Continued) The girl received an excellent grade on her test. Spell the word excellent. The student raised her hand. Spell the word student. The forest has a lot of natural beauty. Spell the word natural. The criminal knew the severity of his actions. Spell the word severity. The woman had confidence when she gave her speech. Spell the word confidence. Sky diving can be dangerous. Spell the word dangerous. 95

PAGE 105

Appendix D Assent Forms Assent Form Part One Young My name is ____________________ and today if you want you can play some games with me. First we are going to check your hearing. Then we are going to play some games where I will ask you some questions. We will be playing these games so I can find out some of the things you know and some of the things you do not know for an experiment I am doing for school. While we are playing these games you may get tired and want to stop. It is okay if you decide you do not want to play anymore. Just tell me you want to stop and we will. If you would like to play these game s with me today please sign your name on the line on this paper. _________________________________ _____________________________ Participant Date Investigator Date 96

PAGE 106

Appendix D (Continued) Assent Form Part Two Young My name is ____________________ and today if you want you can do a spelling game. The game is on the computer and you can type or write your responses. We will be playing the spelli ng game so I can find out some of the words you know how to spell and some of the words you do not know how to spell for an experiment I am doing for school. While we are playing the game you may get tired and want to stop. It is oka y if you decide you do not want to play anymore. Just tell me you want to stop a nd we will. If you would like to play the game with me today please sign your name on the line on this paper. _________________________________ _____________________________ Participant Date Investigator Date 97

PAGE 107

Appendix D (Continued) Assent Form Part One Older My name is ____________________ and today if you want you can do some assessments with me. First we are going to check your hearing. Then we are going to play some games where I will ask you some questions. We will be doing these assessments so I can find out some of the things you know and some of the things you do not know for an experiment I am doing for school. While we are doing the assessments you may get tired and wa nt to stop. It is okay if you decide you do not want to participate anymore. Ju st tell me you want to stop and we will. If you would like to participate with me today please sign your name on the line on this paper. _________________________________ _____________________________ Participant Date Investigator Date 98

PAGE 108

Appendix D (Continued) Assent Form Part Two Older My name is ____________________ and today if you want you can do a spelling assessment. The assessment is on the computer and you can type or write your responses. We will be doing the spelling assessment so I can find out some of the words you know how to spell and some of the words you do not know how to spell for an experiment I am doing for school. While we are doing the assessment you may get tired and want to stop. It is okay if you decide you do not want to participate anymore. Just tell me you want to stop and we will. If you would like to participate with me today please sign your name on the line on this paper. ____________________________ ____________________________ Participant Date Investigator Date 99

PAGE 109

Appendix E Recognition Rules In order to describe the error, it was first determined which linguistic category the error was made in (e.g., phonological, orthographic, morphological, or a combination). After the linguistic categor y was determined, the error feature was coded according to the POMAS coding chart. Phonological category: An error wa s considered to be phonological in nature if the sound structure of the word was not fully represented or was changed as a result of the incorrect, additional or omitted letters. Orthographic category: An error was considered to be orthographic in nature if the sound structure was fully represented, but with incorrect or omitted letters. Morphological category: An error was considered to be morphological in nature if there were omitted or incorrect letters in the inflectional or the derivational affix. POMAS Codes with Examples Category Code Description Example P PCA Consonant addition When a consonant is added, resulting in a change in the phonologic structure assindment | assignment 100

PAGE 110

Appendix E (Continued) Category Code Description Example P PCD Consonant deletion When a consonant is deleted, resulting in a change in the phonologic structure beame | became P PDIP Diphthong When diphthong is reduced to a single vowel arond | around P PDV Devoicing When a voiced consonant is spelled with an incorrect voiceless consonant pusels | puzzles P PEP Epenthesis When a vowel is added that creates an additional syllable in a word tolid | told P PFCD Final consonant deletion When the final consonant is omitted kee | keep P PFLP Flaps (t/d; d/t) When a flap is spelled with the incorrect consonant pride | pretty P PFPV Final position voicing When a final consonant that should be voiced is represented as voiceless becus | because P PGLI Gliding When a glide (w/y) is substituted for a liquid (r/l) cawe | scary 101

PAGE 111

Appendix E (Continued) Category Code Description Example P PFR Fronting When a back consonant is replaced by a front consonant graphits | graphics P PNE Nasal error (n/m; m/n) When an n is substituted for m or vice versa junp | jump P PSC S-clusters When an s-cluster is reduced bes | best P PSE Silent -e patterns When the silent -e in a long vowel pattern is omitted lik | like P PSON Sonorant clusters (nasals, l, r, j) When a sonorant cluster is reduced ad | and P PSONC Sonorant substitution (r/l; l/r) When r is substituted with l or vice versa crever | clever P PSR Syllable reduction When a syllable is omitted from a word maroni | macaroni P PSRS Schwa reduced syllable When the syllabic schwa is omitted from a word anmols | animals 102

PAGE 112

Appendix E (Continued) Category Code Description Example P PST Stopping When a plosive is substituted for a fricative or affricate teel | feel P PSV Short vowels When a short vowel is spelled with a long vowel pattern kite | kit P PVOCR Vocalic r When the r following a vowel is omitted cos | curls O OAA Apostrophe added gets | gets O OCD Consonant doubling When a consonant is doubled unnecessarily terriffic | terrific O OCE Consonant error When the incorrect consonant is used sogt | soft O OCL Capital letter When a word is not capitalized that should be california | California O ODI Digraphs When a consonant digraph is reduced sip | ship O OHY Hyphen When a hyphen is removed or used incorrectly fortytwo | forty-two 103

PAGE 113

Appendix E (Continued) Category Code Description Example O OHSV c/k hard and soft velars When a hard velar such as k is replaced with a soft velar such as c mace| make O OLD Letter doubling When a consonant is supposed to be doubled, but is not (usually syllable juncture) triped | tripped O OLR Letter reversal (b/d, d/b) When a letter is reversed resulting in an incorrect letter bolls | dolls O OLS Letter sound (c/s, s/c, etc.) When an incorrect letter(s) is/are used, but represent(s) the same sound as the correct letter(s) sereal | cereal O OLVP Long vowel pattern When an incorrect long vowel pattern is used; must be phonologically similar to the correct vowel pattern keep | kipe O OLWF Long vowel word families (-old, -igh) A member of a long vowel word family is spelled with the incorrect long vowel pattern, but is phonologically plausible nite | night 104

PAGE 114

Appendix E (Continued) Category Code Description Example O ONA No apostrophe When an apostrophe is omitted somebodys | somebodys O OOW One word When one word is spelled as two some times | sometimes O OPA Phoneme addition When a letter is added, but does not change the phonological structure of the word graede | grade O OSJ Syllable juncture y to I When adding the suffix, the syllable juncture rule calls for a change in spelling but the original spelling is maintained cryed | cried O OSL Silent letter /h/ When the silent h in a word is omitted wen | when O OSR Syllable reduction When a syllable is reduced as a result of syncope diffrent | different O OSUFD Suffix error (derivation) When a derivational suffix is spelled incorrectly dangerios | dangerous O OSUFI Suffix error (inflection) When an inflectional suffix is spelled incorrectly stopt | stopped 105

PAGE 115

Appendix E (Continued) Category Code Description Example O OSY Syllabic l When syllabic l is spelled incorrectly but is phonologically plausible terdals | turtles O OUVP Unusual vowel pattern When an unusual vowel pattern is represented incorrectly, but is plausible phonologically cof | cough O OVDI Vowel digraph (short vowel digraph) When one part of a vowel digraph is omitted but the resulting word is phonologically the same hed | head O OVr Rhotic (r-colored) Vowel When rhotic vowel is spelled incorrectly sistr | sister O OVS Vowel error (short) When the incorrect short vowel is used stuped | stupid O OWB Word boundary (2 sep. words) When two separate words are combined eachother | each other O OWW Whole word (substitution) When a whole word is used to substitute a word in part of or in whole areyoument | argument 106

PAGE 116

Appendix E (Continued) M MDER Derivation (root word) When the root word of a derivation is spelled incorrectly, but the suffix is spelled correctly depasition | deposition M MDVM Derivational morphology When an incorrect derivational form is used brang | brought M MHOM Homonyms When the incorrect homonym is used there | their M MINF Inflectional morphology When the inflectional suffix is omitted bike | bikes M MPRE Prefixes When the prefix is omitted organize | reorganize M MSUF Suffixes When the derivational suffix is omitted normal | normally PO POR Reversals When letters in a word are reversed tis | its PO POVDS Vowel dependent spellings (tch, dge, ck/ch, ge) When short vowel dependent spellings use the incorrect vowel baitch| batch PO POVM Vowels missing/deleted When there one or more vowels are omitted from the word dble | double 107

PAGE 117

108 Appendix E (Continued) Category Code Description Example MO MCON Contraction Neglecting to add the apostrophe to a contraction wasnt | wasnt MO MOSP Misspelled root word The root word is misspelled but the suffix is spelled correctly; still represents a plausible phonological spelling edgeucation | education MO MOV Overgeneralization When an inflectional suffix is added to a word that is already in the inflectional form losted | lost MP MPVS Visually similar error are | car ILS Illegal letter strings When an error has is not representative of the target word and are not a legal string of letters within the constraints of the English language lkhbit | cabin CQ Started the word but neglected to finish it b | buy