THE EFFECTS OF MORPHOLOGICAL AWARENESS ON READING IN CHINESE AND ENGLISH AMONG YOUNG CHINESE CHILDREN: A LONGITUDINAL STUDY

by

Katie Yan Yan Lam

A thesis submitted in conformity with the requirements for the degree of Master of Arts Graduate Department of Human Development and Applied Psychology Ontario Institute for Studies in Education University of Toronto

© Copyright by Katie Yan Yan Lam 2009

THE EFFECTS OF MORPHOLOGICAL AWARENESS ON READING IN CHINESE AND ENGLISH AMONG YOUNG CHINESE CHILDREN: A LONGITUDINAL STUDY Master of Arts 2009 Katie Yan Yan Lam Department of Human Development and Applied Psychology University of Toronto

Abstract This thesis comprised two longitudinal studies examining the role of morphological awareness in Chinese and English reading among Chinese children. In Study 1, participants were 84 kindergarten and first grade Chinese-speaking English Language Learners (ELLs) from Canada. Children’s morphological awareness, vocabulary and reading comprehension in English were assessed at two measurement points spaced one year apart. Study 2 involved the Chinese-Canadian children from Study 1, and 98 kindergarteners and first graders from China. Their morphological awareness, vocabulary and reading comprehension in Chinese were measured at the beginning of two successive academic years. Study 1 showed that for the ELLs, morphological awareness explained increasingly large proportions of variance in English vocabulary and reading comprehension with age. In Study 2, compound awareness significantly predicted Chinese vocabulary for children from both countries. Taken together, the two studies substantiated that morphological awareness contributes to reading in Chinese and English across different languagelearning contexts.

ii

Acknowledgements There are many people towards whom I would like to express my gratitude for their contribution towards the completion of this thesis. First of all, I would like to express my sincere appreciation to my thesis supervisor, Dr. Xi Chen, for her guidance and encouragement in every step of the way, and for her thoughtful comments and constructive suggestions. I would also like to thank my committee member, Dr. Esther Geva, for sharing her expertise in the field and giving many invaluable insights. The completion of this thesis would not be possible without the work of all the project members, who had spent many hours collecting data in China and in Canada. Thanks also go to my fellow graduate students in the lab: Cathy, Gloria, and Heidi, for their constant support as researchers and as friends. Special gratitude is given to my mom, for her continuous and unconditional love, patience, and support, and for always believing in me. My appreciation also goes to my family at Redeemer, who have been my most enthusiastic cheerleaders from the very beginning. To Ka Ho, thank you for all your prayers for me and your affirmation of who I am as our Lord has made me. Finally, I give thanks to my gracious Father above, who has blessed me with everyone that I have mentioned and many more, who has been my Strength and my Courage, my Wisdom and my Guide, my Shield and my Fortress. Unto You all praise and glory shall be given.

iii

Table of Contents Abstract……………………………………………………………….…………….. Acknowledgements………....……………………………………….……………… Table of Contents…………………………………………………….……………... List of Tables……………………………………………………………………….. List of Figures……………………………………………….…………………........ List of Appendices………………………………………………….………………. Chapter 1 Introduction Overview……………………………………………………………………... Rationale of the Research…………………………………………………….. Chapter 2 Literature Review Development of Morphological Awareness Development of Morphological Awareness in English ….……………. Development of Morphological Awareness in Chinese ...…………….. Morphological Awareness and Reading in English Morphological Awareness and Learning to Read English as a First Language……………………………………………………………. Morphological Awareness and Literacy Skills in English Language Learners....………………………………..………………………..... Morphological Awareness and Reading in Chinese…………………………. Chapter 3 Learning to Read Chinese in Mainland China and in Canada Amount of Exposure to the Chinese Language………………………………. Formal Chinese Learning Systems…………………………………………… Chapter 4 The Present Research Statement of the Problem.……………………………………………………. The Present Research..……………………………………………………….. Research Questions….………………………………………………………... Hypotheses of the Studies.……………………………………………………. Chapter 5 Study 1 Method……………………………………………………………………….. Results ……………………………………………………………………….. Discussion...………………………………………………………………….. Chapter 6 Study 2 Method……………………………………………………………………….. Results ……………………………………………………………………….. Discussion...………………………………………………...…………….….. Chapter 7 Conclusion General Discussion………………………………………………………...…. Limitations and Future Directions……………………………………………. References…………………………………………………………………………...

iv

i ii iv v vi vii 1 1

7 9

11 14 16 20 21 23 24 25 26 27 31 52 61 67 87 94 96 112

List of Tables Table 1 Descriptive Statistics of Measures for Children in Canada ……………….. Table 2 Correlations among all Measures at Time 1 and Time 2 for the Younger Cohort …………………………………………………………………… Table 3 Correlations among all Measures at Time 1 and Time 2 for the Older Cohort …………………………………………………………………… Table 4 Hierarchical Linear Regressions Predicting Concurrent English Receptive Vocabulary ……………………………………………………………… Table 5 Longitudinal Hierarchical Linear Regressions Predicting Time 2 English Receptive Vocabulary …………………………………………………... Table 6 Hierarchical Linear Regressions Predicting Concurrent English Reading Comprehension ………………………………………………………...... Table 7 Longitudinal Hierarchical Linear Regression Predicting Time 2 English Reading Comprehension .……………………………………………….. Table 8 Descriptive Statistics of Measures for Children in China and Canada …..... Table 9 Correlations among all Measures at Time 1 and Time 2 for the Children in Canada …………………………………………………………………... Table 10 Correlations among all Measures at Time 1 and Time 2 for the Children in China …………………..………………………………………………... Table 11 Hierarchical Linear Regressions Predicting Concurrent Chinese Vocabulary ………………………………..…………………………….. Table 12 Hierarchical Linear Regressions Predicting Concurrent Chinese Receptive Vocabulary for the Younger Cohort at Time 1 …………………………. Table 13 Longitudinal Hierarchical Linear Regressions Predicting Time 2 Chinese Expressive Vocabulary…………………………………………………... Table 14 Hierarchical Linear Regression Predicting Concurrent Chinese Reading Comprehension for Children in China at Time 2 ……………………...... Table 15 Longitudinal Hierarchical Linear Regression Predicting Time 2 Chinese Reading Comprehension for Children in China …………………………

v

32 38 39 43 45 48 51 69 74 75 80 81 82 85 86

List of Figures Figure 1 Children’s performance on the morphological production task at Time 1 and Time 2 .……………….……………………………………………... Figure 2 Children’s performance on the morphological analogy task at Time 1 and Time 2 .…………………….…………………………………………….. Figure 3 Children’s performance on the compound structure task at Time 1 and Time 2 ………………………………………………………………….... Figure 4 Children’s performance on the compound analogy task at Time 1 and Time 2 …………..………………………………………………………..

vi

35 35 71 72

List of Appendices Appendix A Appendix B Appendix C Appendix D Appendix E Appendix F Appendix G Appendix H Appendix I

Morphological Production Task (English).………………………… Morphological Analogy Task (English)..………………………….… Compound Structure Task (Chinese).………………………………. Compound Analogy Task (Chinese) ………………………………. Chinese Receptive Vocabulary (Sample Items) …………………… Chinese Picture Naming (Sample Items) …………………………... Chinese Reading Comprehension (Sample Items) ………………… Syllable and Phoneme Deletion Task (Chinese)…………………… Chinese Character Reading …………………………………………

vii

99 100 102 103 105 107 108 110 111

Chapter 1 Introduction Overview Studies on the relations between different facets of metalinguistic awareness and learning to read have mostly focused on phonological awareness. An aspect of metalinguistic awareness that has received less attention in research is morphological awareness, especially pertaining to its contribution to reading among children who are learning two languages simultaneously. The intent of the present study was to examine the effects of morphological awareness on vocabulary and reading comprehension among young Chinese-speaking English Language Learners (ELLs) who received concurrent instructions in Chinese and English within a bilingual context, in comparison to those who were monolingual speakers of these languages. This thesis consists seven chapters. In chapter 1, the rationale of the current work is laid out. Existing research relevant to the present research is reviewed in chapter 2. Chapter 3 discusses the differences between Canada and China with respect to the educational context for learning Chinese. Chapter 4 provides an overview of the two studies comprising this thesis and outlines the specific research questions asked. The two studies are then reported in chapters 5 and 6, respectively. Finally, chapter 7 presents a general discussion of the two studies and directions for future research. Rationale of the Research Practical and theoretical reasons underpin the undertaking of the present research. In the recent decades, international immigration has led to a surge in the number of children of Chinese descent being educated in English-dominant countries. In Canada, China has been the leading country of birth among the incoming immigrants since 1998. On average, 33,500 people have emigrated from China to Canada each year between 1

2 1998 and 2007, constituting approximately 10-15% of all immigrants to Canada (Citizenship and Immigration Canada, 2007). Most children from these immigrant families are faced with a challenging task upon entering school in Canada: to achieve academically in a language that they have yet acquired from home (i.e., English). In particular, given that learning to read is critical to academic success, these children are demanded to attain reading proficiency in their second language (L2) rapidly. At the same time, a majority of the parents from these families support their children in maintaining literacy skills in Chinese to enable effective communication with the Chinese community and to develop cultural identity (Lao, 2004). Given these academic and social benefits, identifying powerful underlying component skills that can influence Chinese-Canadian children’s L1 and L2 literacy development is both timely and important. From a theoretical standpoint, research examining children who are learning to read in a second language or in bilingual contexts are valuable in furthering our understanding on reading skills development and its relation to cognitive processes that underlie reading. Linguistic contexts change when learning a language as a second language, or when learning a language within a bilingual environment; for example, children learning two languages simultaneously may have relatively less exposure to each language, in comparison to monolingual speakers of these languages. As a result, children in different linguistic contexts may follow different developmental trajectories (Bialystok, 1996). Therefore, to the extent that research conducted with monolingual speakers has identified specific underlying processes that are associated with reading development, it remains unclear whether these results are generalizable to children who

3 speak more than one language. Research involving children from L2 and bilingual contexts can thus shed light on the relations between underlying processes and reading in specific languages across different linguistic contexts. The particular aspect of metalinguistic awareness that we were interested in examining was morphological awareness. In research exploring the various cognitive skills that have been proposed to underlie reading, considerable attention has been devoted towards understanding children’s development of phonological awareness (i.e., the awareness of units of sound) and its effects on children’s literacy outcomes. Such emphasis in research is reasonable given reading involves mapping one’s spoken language onto a writing system (Carlisle, 2003). Indeed, it has been shown that phonological awareness not only plays a key role in the learning of alphabetic languages such as English (e.g., Goswami & Bryant, 1990; Wagner & Torgesen, 1987), but also in the learning of non-alphabetic languages such as Chinese (e.g., Ho & Bryant, 1997; Li, Anderson, Nagy, & Zhang, 2002). However, comprehension of printed text also requires the child to map phonological forms onto semantic information. Not surprisingly then, over the recent decade, there has been a growing interest to examine children’s development of morphological awareness (Kuo & Anderson, 2006). Some characteristics of the English orthography are believed to make morphological awareness a crucial skill for children’s reading development. English orthography is morphophonemic, i.e., the spelling system represents both phonemes and morphemes (Carlisle, 2003). More critically however, is that there exists a difference in the extent to which phonemes and morphemes can facilitate reading in English. Specifically, English orthography is considered to be a “deep” orthography, wherein

4 grapheme-phoneme correspondences are often equivocal because spelling rules are not directly governed by the phonological syllable structure on a consistent basis. As a result, English readers cannot rely solely on phonology in word recognition (Durgunoǧlu & Öney, 1999; Verhoeven & Perfetti, 2003). In contrast, English orthography generally preserves the identity of morphemes even in the face of semantic and phonetic variation (Mattingly, 1984). Therefore, morphological awareness can influence significantly a child’s English reading acquisition beyond that of phonological awareness by allowing the child to appreciate the semantic relations between words despite their phonological distinction, e.g., electric/electricity, sign/signature, or heal/health (Carlisle, 1995; Elbro & Arnbak, 1996; Fowler & Liberman, 1995; Kuo & Anderson, 2006; Verhoeven & Perfetti, 2003). Researchers have also proposed that morphological awareness may be particularly important for learning to read Chinese because of its unique morphological features (Packard, 2000). First, Chinese is often regarded as a “morphographic language”, where the graphemes represent syllables that are morphemes rather than phonemes. Learning to read Chinese then, entails the acquisition of graphememorpheme correspondences (Wang, Cheng, & Chen, 2006). This clear graphememorpheme association may encourage children to focus on the meanings (as opposed to the sounds) of the language (McBride-Chang, 2004). Another important feature of Chinese is that it has a large number of homophones. For example, in Mandarin, there are approximately 7,000 regularly used words but only about 1,200 spoken syllables. This leads to an average of more than five words per syllable (Li et al., 2002). Consequently, the ability to distinguish words with identical pronunciation but different

5 meaning is vital for children’s acquisition of the language. Finally, Chinese is an analytic and relatively semantically transparent language, in that most of the more complicated vocabulary concepts are built by combining single morphemes to form compounds, where the meaning of each constituent morpheme contributes directly to the meaning of the compound. For instance, whereas in English different countries are referred to in different ways (e.g., France, Thailand, Germany), in Chinese, a large number of these labels would contain the morpheme 国 /guo2/ (country), thus forming compounds such as 法国 /fa3 guo2/ (France), 泰国 /tai4 guo2/ (Thailand), and 德国 /de2 guo2/ (Germany). The appreciation of morphology can therefore aid children in deciphering and acquiring the meanings of the polymorphemic vocabularies. Psycholinguistic studies involving high school students and adult skilled readers also support a relation between morphological awareness and reading development in both English and Chinese. Studies in native English speakers have consistently shown that mature mental lexicons are morphologically organized (e.g., Nagy, Anderson, Schommer, Scott, & Stallman, 1989; Napps, 1989; Niswander, Pollatsek, & Rayner, 2000; Sandra, 1994; Tyler & Nagy, 1990). That is, unlike novice readers who represent morphologically complex words as unanalyzed full forms, experienced readers decompose these words into their constituent morphemes prior to processing and storage (Marslen-Wilson, Tyler, Waksler & Older, 1994). In research involving speakers of Chinese, similar results in support of a morpheme-based mental lexicon have also been obtained (e.g., Zhang & Peng, 1992; Zhou & Marslen-Wilson, 1994; 1995). Thus, for both languages, children’s emerging morphological awareness may function as a

6 mechanism to facilitate efficient word storage, retrieval and processing, all of which support advancements in reading abilities.

Chapter 2 Literature Review In this chapter, the development of morphological awareness in the two languages is first discussed. Research examining the influence of morphological awareness on reading among monolingual English speakers as well as ELLs is then reviewed. Finally, literature on morphological awareness and reading in Chinese is considered. Development of Morphological Awareness Development of Morphological Awareness in English Studies on children’s acquisition of morphology and morphological awareness in English have focused on the acquisition of three types of linguistically complex words: inflections, derivatives, and compounds (Kuo & Anderson, 2006). Children experience substantial growth in their awareness of these three types of morphology beginning at a very young age. A number of studies have reported that children demonstrate incipient understanding of inflectional morphemes by age 2, and acquire most of the regular inflectional principles by the early elementary grades (e.g., Akhtar & Tomasello, 1997; Anisfeld & Tucker, 1968; Berko, 1958; Carlisle, 1995; Derwing & Baker, 1977). In comparison to inflectional morphemes, the developmental trajectory of derivational morphemes begins later in childhood and extends over a longer period of time, possibly into adulthood (Derwing & Baker, 1979, 1986; Windsor, 1994). There is some evidence suggesting that 3- to 5-year-old children are proficient in generating words with highly productive derivational suffixes (the agentive –er and the instrumental –er) (Clark & Hecht, 1982). Carlisle and her colleagues (Carlisle & Fleming, 2003; Carlisle & Nomanbhoy, 1993) showed that first graders’ awareness of the morphological structure of derivations is largely limited to phonetically transparent and common forms (e.g.,

7

8 quiet  quietly, still  stillness). However, by grade 3, monolingual children are adept in decomposing derived words that are less transparent or that contain less familiar suffixes (e.g., long  length, tree  treelet). Tyler and Nagy (1989) found that grade 4, 6, and 8 students are aware of the morphological relations between two words (i.e., relational knowledge). However, knowledge that derivational suffixes usually mark words for parts of speech (i.e., syntactic knowledge) and the realization that derivational affixes are generally constrained to specific stems to which they attach (i.e., distributional knowledge) are not fully mastered even by grade 8. Relatively few studies have examined children’s acquisition of compounds. By asking children to produce labels for novel objects (e.g., balloon-tree), Clark and her colleagues (Clark, 1981; Clark, Gelman & Lane, 1985) found that children understand the modifier-head relation in compounds by 2 years of age. Nonetheless, until they are in grade 4, children may not be able to explicitly explain that meanings of compound words predicate upon the meanings of the more basic words that had formed the compounds (Silvestri & Silvestri, 1977). In a recent study, using recognition, interpretation and discrimination tasks, Ku and Anderson (2003) observed a steady increase in the compound knowledge of second, fourth, and sixth graders as a function of age. Taken together, the existing evidence suggests that English-speaking monolingual children’s implicit knowledge of the compound structure emerges during preschool and gradually becomes more explicit over the elementary years.

9 Development of Morphological Awareness in Chinese The three types of morphologically complex words in English (i.e., inflections, compounds, and derivatives) are similarly found in Chinese (Packard, 2000). However, the two languages diverge in terms of the prominence of each type of morphological structure. In Chinese, lexical compounding is the primary method of word formation; it is estimated that over 75% of the words used in Modern Chinese are two- or threemorpheme compounds (Sun, Sun, Huang, Li & Xing, 1996). In contrast, there are far fewer inflectional and derivational affixes. Research on the acquisition of the three types of morphology in Chinese corresponds to the differences in their prevalence. To date, very little research has been conducted examining the acquisition of inflectional morphology in Chinese. Chang (1992) has reported that the acquisition of the aspect marker 了/-le/ develops over a long period, partly because of its multiple semantic functions. Other than Chang’s study, little is currently known about the mastery of other common Chinese inflectional morphemes such as 着 /-zhe/ (the progressive marker) and 过 /-guo4/ (the marker of past experience), and the extent to which children can manipulate these morphemes (Kuo & Anderson, 2006). The majority of morphological awareness research in Chinese children has focused on tracking the development in their understanding of the meanings and structures of compounds as a combination of constituent morphemes (e.g., Chen, Hao, Geva, Zhu, & Shu, 2009; Chow & Chow, 2005; Ku & Anderson, 2003; Li et al., 2002; McBride-Chang, Shu, Zhou, Wat, & Wagner, 2003; McBride-Chang et al., 2008; Wang, 2000). Across several studies, it has been consistently shown that by age 5, Chinese children are able to use the compounding rules in Chinese to form new compound words,

10 which indicates that they already have some insights into both the compound structure and the meanings of the constituent morphemes at this time. Notably, performance on this compound-construction task improves between kindergarten and grade 2 (Chen et al., 2009; Chow & Chow, 2005; McBride-Chang et al., 2003; 2008). Between grades 1 and 2, studies have also found significant increases in children’s abilities to select and use appropriate morphemes for specific word contexts (Chen et al., 2009; Wang, 2000). In a study involving children in grade 2, 4, and 6, Ku and Anderson (2003) reported that middle and upper elementary school children become progressively more proficient at interpreting low-frequency compounds composed of high-frequency base words, and at distinguishing between well- and ill-formed compounds. Taken together, these studies have shown that compound awareness is acquired by Chinese children quite early on, and develops continuously in the elementary school years. Only one study has been completed so far to examine Chinese children’s development of derivational awareness as an independent construct. Ku and Anderson (2003) tested second, fourth, and sixth graders in their abilities to select the best interpretation of low-frequency derivatives composed of high-frequency base words, and to distinguish between well- and ill-formed derivatives. They found that children’s performance on both of these tasks improved with age. While these results suggest that Chinese children’s derivational awareness increases during the elementary school years, it is clear that much more research is necessary to delineate the developmental trajectory of derivational awareness in Chinese children.

11 Morphological Awareness and Reading in English Morphological Awareness and Learning to Read English as a First Language School-age children may encounter up to 3,000 unfamiliar words each year when reading in English, with an increasing proportion of the words being comprised of more than one morpheme as children advance in age (Nagy & Anderson, 1984). More remarkably, Anglin (1993) estimated that while knowledge of root words doubled, there was almost a ten-fold increase in the number of multimorphemic words that children know from grade 1 to grade 5. Given the proliferation of morphologically complex vocabulary that children are exposed to and are learning over the elementary school years, some researchers have postulated and provided evidence to suggest that morphological analysis – the ability to decompose morphologically complex words into their constituent morphemes and to derive meanings of the whole words from these morphemes – is a critical underlying mechanism that facilitates vocabulary learning. For instance, by analyzing children’s definitions of multimorphemic words, Anglin (1993) reported the percentage of words known by school-age children for which there was evidence of morphological analysis significantly increased from 40% in grade 1 to 51% in grade 5. Freyd and Baron (1982) compared grade 5 high academic ability students (who were assumed to be faster vocabulary learners) and grade 8 average students. They concluded that the younger, grade 5 students who were able to analyze derived words into morphemes and utilized derivational rules when learning words acquired new vocabulary more quickly than the grade 8 students. In all, children with more developed morphological awareness appear to have an advantage in acquiring and retaining morphologically complex words.

12 Recent correlational research confirmed that awareness of the structures and meanings of morphologically complex words accounts for an increasing proportion of variance in predicting children’s vocabulary knowledge from kindergarten till grade 5 (e.g., Carlisle, 2000; Carlisle & Fleming, 2003; McBride-Chang, Wagner, Muse, Chow, & Shu, 2005a). In a longitudinal study, Carlisle and Fleming (2003) measured grade 3 children’s morphological awareness by asking them to distinguish two-morpheme from one-morpheme words (e.g., hilly vs. silly), and to define multimorphemic words (inflections, derivations and compounds). Children’s performance on the morphological awareness tasks in grade 3 predicted scores on a vocabulary test two years later. McBride-Chang and her colleagues (2005a) assessed children’s abilities to distinguish the different meanings of homophones and to produce multimorphemic words for novel objects or concepts. They found that performance on these morphological awareness measures explained a unique proportion of variance in kindergarten and grade 2 students’ expressive vocabulary, above and beyond other reading related skills. Overall, results from these studies underscore the distinct role of morphological awareness in facilitating children’s vocabulary gains over the elementary years. On the other hand, the unique contribution of inflectional, derivational and compound awareness on reading acquisition has never been addressed directly. Given that different types of morphological awareness skills have differential effects on literacy development (Kuo & Anderson, 2006; McBride-Chang et al., 2005a), discriminating between the effects of these three types of morphological awareness on vocabulary acquisition is the next important step in morphology research.

13 Another prominent aspect of literacy development that has been associated with morphological awareness is reading comprehension. A series of studies conducted by Carlisle and her colleague (Carlisle 1995; Carlisle, 2000; Carlisle & Fleming, 2003) involving young native English speakers between kindergarten and grade 5 indicated that children’s emerging morphological analysis skills play a significant role in their development of reading comprehension skills. These relations are evident in children as young as 6 years of age, and become more robust with time. In two of their studies (Carlisle, 1995; Carlisle & Fleming, 2003), it was demonstrated that children’s performance on morphological awareness tasks in the early elementary years significantly predicted reading comprehension in the two years ensuing, with the relationships being notably stronger in the older age groups. Likewise, Deacon and Kirby (2004) reported that children’s morphological awareness in grade 2 predicted increasingly greater proportions of variance in reading comprehension in each of the subsequent three grades (i.e., grades 3, 4, and 5). Most importantly, this study is one of the first to demonstrate that the relations between morphological awareness and reading comprehension remain robust after accounting for the contributions of verbal and nonverbal reasoning abilities, as well as phonological awareness. Some research has shown that there is a direct relationship between morphological awareness and reading comprehension, beyond the mediating effect of vocabulary. Ku and Anderson (2003) found that, after removing variances attributable to children’s vocabulary, morphological awareness still made a significant contribution to reading comprehension in grade 2, 4, and 6 children. However, the effects of phonological awareness were not taken into consideration in that study. Nagy and his

14 colleagues (2006) explored various pathways by which morphological awareness, phonological memory and phonological decoding contribute to reading vocabulary and reading comprehension in students between grades 4 and 9. In that study, morphological awareness predicted a significant amount of variance in reading comprehension over and above vocabulary and other reading-related factors at all grade levels. Together, the two studies provide substantial support that morphological awareness is associated with comprehension not only through its contribution to vocabulary growth. Morphological Awareness and Literacy Skills in English Language Learners To date, few studies (Kieffer & Lesaux, 2008; Siegel, 2008; Wang, Cheng, & Chen, 2006) have examined the relation between morphological awareness and reading comprehension in ELL children from diverse linguistic backgrounds. In their two-year longitudinal study of grade 4 Spanish-speaking ELL students, Kieffer and Lesaux (2008) found that the magnitude of the relation between derivational morphological awareness and reading comprehension increased over the two years. Performance on the morphological awareness tasks in grade 4 also predicted reading comprehension scores in grade 5; however, this relationship was no longer significant after taking into account other reading skills measured in grade 4. Wang and her colleagues (2006) focused on grade 1 to grade 5 ELL students of Chinese descent. After taking age, oral vocabulary, and phonological awareness into consideration, these children’s awareness of compounds (but not derivations) concurrently predicted unique variances in reading comprehension. Siegel (2008) examined the relation between sensitivity to derivational morphemes and reading amongst grade 6 ELL children of diverse linguistic backgrounds and their English monolingual counterparts. Morphological awareness predicted significantly

15 variances in reading comprehension over and above the contribution of phonological awareness and, in a separate analysis, of syntactic awareness. However, analyses were not conducted separately for the two language groups (ELLs vs. English monolinguals) to compare the relative strength of the effects of morphological awareness on reading comprehension between the two groups. Taken together, there is some preliminary evidence that, similar to native English-speaking children, different aspects of English morphological awareness can predict concurrent reading comprehension in ELL children in the elementary school years, independent of their associations with vocabulary and with phonological awareness. Nonetheless, research to date is limited in its consideration of the various aspects of morphological awareness. Two of the studies (i.e., Kieffer & Lesaux, 2008; Siegel, 2008) focused exclusively on derivational morphology. In the study in which both compound and derivational morphology were explored (i.e., Wang et al., 2006), analyses were conducted with children spanning five grades, who would have considerable variances in their morphological awareness. Another issue is that the linkage between morphological awareness and vocabulary has only been assessed indirectly as a covariate in studies that focused on reading comprehension development. The extent to which morphological awareness contributes to vocabulary acquisition in comparison to other reading-related skills such as phonological awareness has not been systematically explored in ELL children. Clearly, more research involving ELLs is required that examines various components of morphological awareness and literacy skills.

16 Morphological Awareness and Reading in Chinese Although many scholars have suggested that morphological awareness and morphological analysis plays an important role in Chinese reading (e.g., Hoosain, 1992; Nagy & Anderson, 1998; Shu, Anderson, & Zhang, 1995), systematic investigations of the impact of Chinese children’s emerging morphological awareness on their reading have only begun recently. In particular, given lexical compounding is the most common word formation process in Chinese, research examining the relations between morphological awareness and reading outcomes in Chinese children has mainly focused on their awareness of the morphology of compounds. Because Chinese morphemes are combined in predictable ways to form compounds words, children’s abilities to reflect on and to manipulate compound structures as well as the constituent morphemes of words are believed to serve as a guiding principle for comprehending and producing vocabularies in Chinese (Chen et al., 2009; McBride-Chang et al., 2007, 2008). Several correlational studies conducted among kindergarten and elementary school children have demonstrated that children’s compound awareness is significantly associated with their vocabulary knowledge in Chinese (Chen et al., 2009; Chung & Hu, 2007; Ku & Anderson, 2003; Li et al., 2002; McBride-Chang et al., 2005b, 2006, 2008; Wang, 2000). McBride-Chang, Cheung, Chow, Chow, & Choi (2006) reported that measures of receptive and expressive compound awareness jointly accounted for unique variance in Cantonese-speaking kindergartner’s receptive vocabulary knowledge, after controlling for phonological awareness, reading skill, and general reasoning abilities. Similarly, among grade 1 and 2 Mandarin-speaking children, Chen et al. (2009) found that the students’ ability to

17 understand the structure and meaning of a compound word as a combination of two or more constituent morphemes significantly explained unique variance in their expressive vocabulary, above and beyond the effects of several reading related skills. A possible causal link between morphological awareness and vocabulary development is substantiated in a recent longitudinal study involving kindergarten children from Hong Kong and Beijing (McBride-Chang et al., 2008). In that study, McBride-Chang et al. (2008) found that kindergarten children’s ability to produce novel compound words using familiar morphemes predicted unique variance (2% in the Hong Kong sample and 4% in the Beijing sample) in their vocabulary knowledge one year later. Importantly, these predictions remained significant after controlling for children’s individual differences in nonverbal reasoning, phonological processing skills, morphological awareness, and vocabulary (the autoregressor) measured at the first testing phase. McBride-Chang et al.’s study is the first to show that morphological awareness impacts vocabulary development across time in Chinese children. More longitudinal studies with converging results will be critical in providing further support for a causal link between morphological awareness and vocabulary acquisition. In addition to its influence on vocabulary development, morphological awareness has been conjectured to facilitate reading comprehension in Chinese children. When children encounter novel characters in reading, their knowledge of the surrounding morphemes and their understanding of the morphemic structures may aid them in deriving approximate meanings for the unfamiliar characters and the overall text (Shu, McBride-Chang, Wu, & Liu, 2006). A handful of studies have preliminarily demonstrated that morphological awareness contributes to reading comprehension among

18 Chinese children (Ku & Anderson, 2003; Li et al., 2002; McBride-Chang et al., 2007; Shu et al., 2006; Wang, 2000). For instance, Wang (2000) found that morphological awareness predicted unique variance in reading comprehension among first and second graders, after controlling for the effects of phonological awareness. Likewise, when several reading related skills have been taken into consideration, McBride-Chang et al. (2007) found that morphological awareness was able to explain unique variance in grade 3 children’s reading comprehension. However, the effects of morphological awareness were no longer significant once character recognition was included as a control variable. The strongest evidence underscoring the importance of morphological awareness in Chinese reading comprehension comes from Shu et al.’s (2006) study. Using path analyses, Shu et al. compared the contribution of several reading related skills to literacy outcomes among fifth and sixth grade students. They reported that performance on a morphological production task was the strongest cognitive correlate of reading comprehension, even after taking into account the effects of vocabulary. In sum, these studies have provided some evidence supporting the facilitative role of morphological awareness in school-age children’s reading comprehension. However, most of these studies were conducted with children who would have had at least two years of literacy experience; the influence of morphological awareness on comprehension in beginners has scarcely been explored. In the lone study involving first and second graders (i.e., Wang, 2000), the only control variable considered was phonological awareness. While phonological awareness is considered to be one of the most foundational skills for early readers (McBride-Chang et al., 2008), other reading related skills such as rapid automatized naming, should also be taken into account when determining the unique

19 effects of morphological awareness on reading. Finally, as none of the previous studies were conducted longitudinally, the extent to which morphological awareness impacts children’s development in reading comprehension over time remains unclear.

Chapter 3 Learning to Read Chinese in Mainland China and in Canada Given the environment for learning Chinese in China and in Canada may be less familiar to many readers, this chapter explores the differences with respect to the cultural and linguistic contexts of learning to read Chinese in China and Canada in order to provide contextual information for Study 2. There are two major differences between the cultural and linguistic contexts of learning to read Chinese in Mainland China and in Canada: the amount of exposure to the Chinese language and formal Chinese learning systems. Amount of Exposure to the Chinese Language Although there are many different dialects throughout Mainland China, Mandarin Chinese (Putonghua) is the only official oral language in Mainland China. The official orthography consists of simplified Chinese characters. Mandarin Chinese and simplified Chinese characters are used in the majority of official publications, school materials, and the media. Meanwhile, foreign language learning is usually very limited inside and outside of schools. Therefore, China can be considered a predominantly monolingual learning environment in which children receive a large amount of exposure and instruction in both oral Chinese and Chinese orthography on a daily basis. By contrast, Chinese people immigrated to Canada usually live in a bilingual environment where Chinese is spoken at home, and English or French is used in most other occasions (e.g., daycare, school). Not surprisingly then, children from these immigrant families generally spend less time reading and writing Chinese compared to their counterparts in Mainland China. For instance, data collected from the home literacy practices questionnaire in the current study indicated that, whereas Chinese children in

20

21 Canada on average read in Chinese for less than 2 hours per week, a majority of their counterparts in China spend at least 2 to 5 hours per week reading in Chinese. On average, children in Canada spend less than 2 hours per week watching Chinese TV programs; in contrast, most children in China watch Chinese TV for 1 to 2 hours every day. At the same time, children in Canada on average spend 2 to 5 hours per week reading in English and almost 1 to 2 hours a day watching TV in English. Very few children in China reported to engage in these activities on a regular basis. Clearly, children of Chinese origin in Canada are immersed in an environment that is much more bilingual than the one experienced by their counterparts in China. Formal Chinese Learning Systems In terms of formal Chinese instruction, children in China learn to read Chinese in a vastly different setting compared to those in Canada, mainly due to the differential status that Chinese language learning receives in each educational system. In Mainland China, Mandarin Chinese is systematically taught to children in all elementary grades. All schools follow the National Curriculum Guide and more than 70% of the schools use national textbooks series. Preschool literacy curricula focuses on preparing children for formal Chinese instruction in elementary schools through promoting children’s interest in Chinese orthography and equipping children with basic reading and writing skills (Government of the People’s Republic of China, Ministry of Education, 2001). Beginning in grade 1, children receive explicit instruction in Chinese characters and Pinyin, a phonological system representing the pronunciations of Chinese orthography. The meanings and the order of strokes of Chinese characters are introduced and learned through mass practice. The objective of Pinyin learning for first and second graders is to

22 enable them to sound out a syllable according to its onset, rime, and tone. It is also expected that by the end of the second grade, children will be able to recognize and write about 1,250 familiar characters in Chinese (People’s Education Press, 2008). In Canada, many Chinese children gradually lose their Chinese when they start attending daycares or public schools because they use progressively more the primary language of instruction (i.e., English or French) in their academic learning and in interacting with their peers, while spending less time speaking and writing Chinese (Feuerverger, 1997). In order to maintain their heritage language, in addition to speaking Chinese at home, some Chinese parents send their children to Chinese heritage language classes or English-Chinese bilingual programs. In Ontario, programs for preschool children are typically incorporated into private daycare programs sponsored by individuals or organizations in the Chinese community. School-age children generally receive Chinese instruction for 2.5 hours each week in the International Language Program, implemented and financially supported by the government of Ontario since 1977. Alternatively, they can attend private Chinese heritage language classes organized by agencies within the Chinese community. However, to the extent that these language classes are available, it is important to note that participation in the programs is entirely voluntary. Moreover, there is currently no standard curriculum to guide Chinese language teaching and learning in Canada. As a result, pedagogical approaches in instructing Chinese may vary from teacher to teacher depending on the instructional goals. These contextual differences may lead to divergences in the Chinese learning processes in China and in Canada. Children from the two countries may develop diverse learning strategies; they may also be at different stages of the learning processes.

Chapter 4 The Present Research This chapter commences with a brief summary of the current state in morphological awareness research, followed by a description of the two studies included in the present research. The specific research questions of this thesis are then introduced. Lastly, the hypotheses of the two studies are stated. Statement of the Problem As reviewed in chapter 2, recent research has supported the central role that morphological awareness plays in reading among monolingual speakers of English and of Chinese. To date however, limited research has been conducted to examine the influence of morphological awareness on reading in children who are learning to read Chinese and English simultaneously, especially those who are at the initial stages of reading acquisition. It therefore remains unclear whether the associations between morphological awareness and reading in Chinese and English established previously among monolingual speakers can be generalized to beginner readers from bilingual contexts. By examining young Chinese-speaking ELL children in Canada, the present research seeks to extend the current understanding of the associations between morphological awareness and reading among monolingual speakers of Chinese and English to children who are learning the two languages simultaneously. Findings from this research will help determine whether the relations between morphological awareness and reading are fundamentally related within Chinese and English, independent of children’s linguistic contexts. This can enhance our theoretical conceptualization regarding the contribution of morphological awareness on children’s reading skills development across different language-learning environments.

23

24 The Present Research The present research builds upon the theoretical frameworks established from previous research conducted among monolingual Chinese- and English-speaking children, and explored the effects of morphological awareness on young Chinese-speaking ELL children’s reading development in Chinese and English over time. In Study 1, we were primarily interested in determining whether and to what extent morphological awareness for compound and derivations in English contribute to Chinese-speaking ELL children’s development of English vocabulary knowledge as well as reading comprehension, as demonstrated previously in monolingual English speakers. In particular, we aimed to establish the unique contribution of morphological awareness to these two reading skills beyond the effects of other cognitive processes that have been demonstrated previously to underlie reading, e.g., phonological awareness and general nonverbal reasoning abilities. To better delineate the developmental trajectory of the linguistic skills and to elucidate the directionality of the relations within each age group, we followed the children’s language development by measuring their morphological awareness and reading abilities in English over one year. The main objective of Study 2 was to compare the contributions of morphological awareness to reading among children who were learning Chinese within a monolingual (i.e., China) and a bilingual (i.e., Canada) context. These relations were considered while taking into account the effects of several other underlying processing skills important for reading. We included a group of monolingual Chinese children from China as our comparison group because, to our knowledge, few studies have examined the associations between morphological awareness and reading among monolingual Chinese

25 children who were as young as those considered in the current study, while controlling for other reading related factors. Similar to Study 1, to follow the developmental growth of the linguistic skills and to establish the directionality of the associations between morphological awareness and the two reading skills, we tracked children’s language development in Chinese over one year. In both studies, we assessed the relations between morphological awareness and reading among kindergarten, grade 1 and grade 2 children, a sample that is younger than many of the previous studies. This age group was selected for two reasons. Carlisle (1995) suggested that kindergarten and first grade is of particular interest to language development researchers because it is during this period of time when children start to shift from an implicit to a more explicit awareness of language use and language structures. Past research involving Chinese and English L1 speakers also indicate that there is an emerging association between morphological awareness and reading by this age. We therefore considered it important to substantiate these relations in children who are learning these languages within a bilingual context. Research Questions The specific research questions addressed in the two studies constituting this thesis were: 1. In comparison to monolingual English and Chinese speakers, do Chinesespeaking ELL children demonstrate similar levels of morphological awareness in English and Chinese in the early school years? a. Are there developmental changes in morphological awareness over the early school years for these children?

26 2. Can morphological awareness predict English vocabulary and reading comprehension among young Chinese-speaking ELL children? a. (How) do the effects of morphological awareness on these reading skills change with age? b. Do distinct aspects of morphological awareness contribute differently to Chinese-speaking ELL children’s English reading skills? 3. Does morphological awareness specifically for compounds play a role in Chinese vocabulary and reading comprehension similarly among monolingual Chinese speakers from China and Chinese-speaking ELL children from Canada? a. (How) does the role of morphological awareness in reading change with age for the two groups of children? Hypotheses of the Studies Based on previous findings in young native English speakers, it was hypothesized that in Study 1, compound and derivational awareness in English would concurrently and prospectively predict English vocabulary knowledge and reading comprehension skills among kindergarten and grade 1 Chinese-speaking ELL children. Relatedly, it was expected that these links would remain robust after mother’s education, nonverbal reasoning ability and phonological awareness have been taken into consideration. In Study 2, building on previous findings among monolingual Chinese children, compound awareness was expected to predict Chinese vocabulary knowledge and reading comprehension skills similarly for children from China and those from Canada. These relations were also expected to remain significant after considering the effects of mother’s education, nonverbal reasoning ability and phonological awareness.

Chapter 5 Study 1 The objective of Study 1 was to examine the contribution of morphological awareness on concurrent and prospective vocabulary and reading comprehension in English among Chinese-speaking ELL children. The methodologies used for the study is first described. The results are then presented, followed by a discussion of the results. Method Participants Participants were 84 Chinese immigrant children recruited from six public schools, three government-funded Chinese heritage language schools, and one private Chinese heritage language school located in a large Canadian metropolitan area. At the time of recruitment, 48 of the participants were in kindergarten (51% males and 49% females) and 36 were in grade 1 (60% males and 40% females). The kindergarten participants’ average age was 5 years 5 months (SD = 5.70 months); the average age of the first graders was 6 years 7 months (SD = 4.35 months). All children received instruction in English in public schools. Demographic information was collected through a family questionnaire designed by the researchers. Sixty-one percent of the children in kindergarten and 33% of the children in grade 1 were born in Canada; the remaining children were born in China, Taiwan, Japan, or the United States. For children born outside of Canada, the average age of immigration was 1 year and 7 months (SD = 19.68 months) for children in kindergarten, and 10 months (SD = 17.82 months) for the first graders. All but one child had exposure to Chinese at home to varying extents. Seventythree percent of the participants attended Chinese heritage language classes for an average of 2.5 hours each week, where they received instruction in both oral language

27

28 and literacy skills. Approximately 74% of the children read Chinese books at home during the week. The average level of maternal education was a university degree. Mother’s education was used as a proxy for family socioeconomic status (SES)1. Measures Children were tested at two measurement points spaced one year apart. At both measurement points (Time 1 and Time 2, respectively), participants received a battery of tests in English including phonological awareness, morphological awareness, word reading, and receptive vocabulary. A non-verbal reasoning measure was administered to all participants at Time 1. Task assessing reading comprehension was administered only to children in the older cohort at Time 1; at Time 2, the task was given to all children. Instructions for all tasks were given in English. To identify home literacy practices, a questionnaire on family background and literacy activities at home was sent to each participant’s family at Time 1 to be completed by the parents in English or Chinese. Measures of Morphological Awareness Test of Morphological Production. This task was adapted from Carlisle (2000) to examine children’s derivational morphological awareness. For this task, children were required to produce a derived word to complete a sentence. For example, “(Farm) My uncle is a _______ (farmer)”. Children were first given three practice trials with feedback to ensure that they had understood the task. This task contained 24 test items at Time 1 and 27 items at Time 2. Reliability coefficients for this task were .87 at Time 1 and .90 at Time 2. Appendix A presents the measure administered at Time 1.

1

Gottfried, Gottfried, Bathurst, Wright, Guerin, and Parramore (2003) have discussed the adequacy in using mother’s education level as an alternative to composite measures of socioeconomic status.

29 Morphological Analogy Test. Based on the task developed by McBride-Chang et al. (2005a), this test was designed to evaluate children’s compound morphological awareness. In each trial, children were presented with the definition of a compound word, and were then asked to create a compound of similar structure using newly presented concepts. For example, “Early in the morning, we can see the sun rising. This is called a sunrise. At night, we might also see the moon rising. What could we call this? (moonrise)”. There were two practice trials and 15 test items on this test. The reliability of this test was reasonable (Cronbach’s α = .83). See Appendix B for the items included in this task. Literacy Outcome Measures Vocabulary. A shortened version of the Peabody Picture Vocabulary Test, Third Edition, Form III A (PPVT-III A) (Dunn & Dunn, 1997) was administered to measure children’s receptive vocabulary. To maintain the same progression of item difficulty as the original task, every third item from the original task was selected to create the shortened version, with a total of 60 items. The modifications were made to allow for group administration and to shorten administration time. The experimenter read each item twice and the children circled in response booklets the picture that best described the word heard. The reliability coefficient for this task was .61. Reading Comprehension. Reading comprehension was assessed using the Reading Comprehension subtest of the Peabody Individual Achievement Test – Revised (PIAT-R) (Markwardt & Frederick, 1998). To facilitate group administration and to shorten administration time, we created a shortened version of the original test by including only every other item. The final number of items selected was 36. Each child

30 received booklets containing sentences and short paragraphs of increasing difficulty, along with stimulus pictures. Children were asked to silently read each sentence or short paragraph once. They were then to turn the page and select out of four stimulus pictures the one that best represented the sentence or short paragraph previously read, and to mark their answers on a response sheet provided by the examiner. The Cronbach’s α for this task was .90. Control Measures Non-verbal Reasoning. Non-verbal reasoning ability was measured using the Raven’s Standard Progressive Matrices (Raven, Raven, & Court, 1998). For each item children were asked to complete a visual-spatial matrix by choosing the missing piece from six or eight patterned segments. Phonological Awareness. Children’s phonological awareness was assessed using the Elision subtest from the Comprehensive Test of Phonological Processing (CTOPP) (Wagner, Torgesen, & Rashotte, 1999). Children were asked to delete individual sounds from words and to give the remaining part (e.g., “cat”, say it without “/k/”). The 20 items in this test included initial, middle and last phoneme deletion. Word reading. Children’s word reading ability in English was assessed using the Letter-Word Identification Subtest from the Woodcock Language Proficiency Battery (WLPB) (Woodcock, 1984). This test required children to identify 14 letters and to read 62 words of increasing difficulty. The test was discontinued if the child read 6 consecutive words incorrectly. The score was the total number of words read correctly.

31 Procedure Participants were assessed in a quiet room at their schools within school hours. Experimenters were trained undergraduate and graduate research assistants who majored in psychology, linguistics, or human biology. For each data collection phase, testing was divided into two sessions of about 60 minutes each. At Time 1, all tasks were administered to the children individually except that for the grade 1 students, vocabulary and reading comprehension tasks were completed in small groups. At Time 2, all children completed the vocabulary and reading comprehension tasks in small groups; the remaining tasks were given individually. For tests that were administered individually, the order by which the tests were given was counterbalanced across participants. Results We first examined the data for univariate and bivariate outliers. There were no univariate outliers in the sample. One child in grade 1 was identified to be a bivariate outlier when the relation between the morphological production task and reading comprehension at Time 2 was considered; the child was excluded from subsequent analyses. The means and standard deviations of each measure administered at Time 1 and Time 2 are presented in Table 1. The mean scores show that, for every measure, performance by the older cohort was better than that of their younger counterparts. Further, with the exception of the receptive vocabulary task for the younger cohort, children improved on all linguistic tasks over the course of one year. Across groups, there was adequate variability on all measures. However, there was a possible ceiling

32 effect for the older cohort on the morphological analogy task at Time 2, where approximately 30% of the children achieved a perfect score. Table 1. Descriptive Statistics of Measures for Children in Canada. _____________________________________________________________________________________

Time 1 Time 2 ____________________________________ M SD M SD _______________________________________________________________________ Younger Cohort Non-verbal ability (Raven’s Matrices) Phonological awareness (Elision) Morphological productiona Morphological production (18 items)b Morphological analogy Word reading Receptive vocabulary (PPVT) Reading comprehension

0.44 0.23 0.23 0.20 0.37 0.31 0.54 ---

0.12 0.21 0.14 0.13 0.27 0.13 0.07 ---

--0.49 0.50 0.51 0.67 0.48 0.52 0.38

--0.24 0.20 0.21 0.24 0.14 0.07 0.20

Older Cohort Non-verbal ability (Raven’s Matrices) 0.60 0.16 ----Phonological awareness (Elision) 0.48 0.27 0.61 0.24 Morphological productiona 0.37 0.20 0.62 0.19 b Morphological production (18 items) 0.35 0.19 0.61 0.19 Morphological analogy 0.59 0.31 0.81 0.20 Word reading 0.49 0.16 0.58 0.14 Receptive vocabulary (PPVT) 0.56 0.09 0.57 0.09 Reading comprehension 0.41 0.19 0.53 0.17 _______________________________________________________________________ a

The mean percentage scores and standard deviations calculated using all items administered at each time point. b The mean percentage scores and standard deviations for the 18 common items in the task administered at Time 1 and Time 2.

We checked the data for normality for each predictor and outcome variable by inspecting skewness and kurtosis. Scores on the morphological analogy and morphological production tasks for the older group at Time 2 were significantly negatively skewed. Following the recommendations of Tabachnick and Fidell (2007), a

33 sequence of reflection, square root transformation, and re-reflection was performed on each variable to remove the skews. After the transformations were performed, the skews came within reasonable limits. We then compared correlational and linear regression analyses conducted with raw scores and with transformed scores but found no significant differences in results as a function of the set of scores used. Since analyses conducted with raw scores are more readily interpretable because the scores represent children’s actual performance levels, only results from analyzing raw data are included in the current report. In order to gain a better perspective of the children’s English language abilities in our study, their levels of performance on the predictor variables were further considered within the context of their native English-speaking peers. Specifically, children’s mean scores on the standardized measure of phonological awareness used in our study (i.e., CTOPP Elision) were compared to the age norms outlined in the test administration manual (Wagner et al., 1999). Children’s performance on this task fell in the range of 63rd to 80th percentile, suggesting that the Chinese-speaking ELLs in our sample were similar to or above their native-speaking peers in their phonological awareness. For the morphological analogy task where standardized norms are not available, children’s percentage scores from our study were compared with those achieved by their same-age native English-speaking counterparts in an earlier study (i.e., McBride-Chang et al., 2005a)2. This comparison indicated that the Chinese-speaking ELLs’ performance on the analogy task was similar to their English-speaking counterparts in kindergarten (33% vs. 40%), but was substantially better than the native English speakers at grade 2 (80% vs. 2

Mean score comparisons could not be conducted because the task used for our study differed from that administered by McBride-Chang et al. (2005a) with respect to the stimulus words and the number of items included.

34 65%). The morphological production task was administered to kindergarten children in two previous studies (Carlisle 1995; Carlisle & Nomanbhoy, 1993). However, given the descriptive data for the task were not reported for either study, we were not able to determine whether performance of our current sample is comparable to that of native English-speaking children. Development in Morphological Awareness A 2 (cohort; older versus younger) × 2 (time; time 1 versus time 2) repeated measures Analysis of Variance (ANOVA) was conducted for each morphological measure to examine children’s morphological awareness development over time. In each model, children’s scores on the specific morphological task of interest at Time 1 and Time 2 were entered as within-subject factors; their cohort level was entered as the between-subject factor. Given that at Time 2 some test items were added to or changed from the morphological production task administered at Time 1, only total scores calculated for the 18 test items common in both versions of the task were used in the current analysis. Figure 1 and 2 depict the changes in children’s performance across time on the morphological production and morphological analogy tasks, respectively. On both measures, significant main effects were found for cohort (for morphological production, F(1,52) = 8.04, p < .01, for morphological analogy, F(1,53) = 6.34, p < .05). Thus, across time, children in the older cohort performed significantly better on the morphological measures than those in the younger group. Main effects for time were also significant, F(1, 52) = 208.47, p < .01, for morphological production, and F(1,53) = 82.91, p < .01, for morphological analogy. This indicates that over time, all children

35 improved in their performance on the morphological tasks. The Cohort × Time interaction was not significant for either task (both p > .05). Figure 1. Children’s performance on the morphological production task at Time 1 and Time 2.

Scores (18)

Morphological Production 18 16 14 12 10 8 6 4 2 0

Younger Cohort Older Cohort

Time 1

Time 2

Figure 2. Children’s performance on the morphological analogy task at Time 1 and Time 2.

Morphological Analogy 14

Scores (15)

12 10

Younger Cohort Older Cohort

8 6 4 2 0 Time 1

Time 2

36 The Relations between Morphological Awareness, Vocabulary and Reading Comprehension Intercorrelations among all measures included in the present study are displayed in Tables 2 and 3 for the younger and older cohorts, respectively. Across groups and measurement points, the two morphological awareness measures were strongly associated with one another (r’s range from .49 to .76, all p’s < .01). For the younger cohort, as indicated in Table 2, the association between phonological awareness and vocabulary was significant only at Time 2. Similarly, correlations between morphological tasks and vocabulary were moderate at Time 1 but robust at Time 2. Phonological awareness and morphological awareness assessed at Time 1 were both significantly correlated with vocabulary measured at Time 2. Table 3 suggests that for the older cohort, the association between phonological awareness and vocabulary was significant only at Time 1. The morphological production task was significantly correlated with vocabulary concurrently at both time points and longitudinally, whereas the relations between the morphological analogy task and vocabulary were moderate. For the younger cohort, correlations between measures of phonological awareness, morphological awareness and reading comprehension were robust and significant at Time 2. Phonological awareness assessed at Time 1 was significantly associated with Time 2 reading comprehension. Likewise, Time 1 morphological analogy task and Time 2 reading comprehension were strongly correlated. For the older cohort, the associations between phonological awareness, the two morphological measures and reading comprehension were significant at both time points. Time 1 phonological awareness and

37 morphological awareness were also significantly related to reading comprehension tested at Time 2.

38 Table 2. Correlations among all Measures at Time 1 and Time 2 for the Younger Cohort. _____________________________________________________________________________________________________________________________ _ 1 2 3 4 5 6 7 8 9 10 11 12 13 _____________________________________________________________________________________________________________________________ _ 1. Age -2. Non-verbal ability .42** -3. Mother’s education -.08 .07 -4. 5. 6. 7. 8.

T1 Phonological awareness T1 Morphological production T1 Morphological analogy T1 Word reading T1 Receptive vocabulary

.21 .26 .32* .23 .23

.36* .15 .47** .20 .31*

.27 .00 .14 .25 .26

-.59** .48** .53** .16

-.49** .59** .30

-.48** .14

-.17

--

9. T2 Phonological awareness .16 .29 .37* .57** .30 .47** .66** .23 -10. T2 Morphological production .41* .32 .39* .58** .68** .52** .68** .28 .57** -11. T2 Morphological analogy .26 .34 .50** .50** .56** .66** .49** .37* .64** .68** -12. T2 Word reading .21 .28 .38* .39* .28 .25 .80** .31 .72** .61** .50** -13. T2 Receptive vocabulary .11 .21 .53** .40* .44* .44* .38* .36* .45** .51** .64** .44** -14. T2 Reading comprehension .40* .43* .25 .41* .34 .38* .71** .32 .47** .58** .38* .76** .48** _____________________________________________________________________________________________________________________________ _ *p < .05, **p < .01

39 Table 3. Correlations among all Measures at Time 1 and Time 2 for the Older Cohort. _____________________________________________________________________________________________________________________________ ________ 1 2 3 4 5 6 7 8 9 10 11 12 13 14 _____________________________________________________________________________________________________________________________ ________ 1. Age -2. Non-verbal ability .28 -3. Mother’s education -.05 .39* -4. 5. 6. 7. 8. 9.

T1 Phonological awareness T1 Morphological production T1 Morphological analogy T1 Word reading T1 Receptive vocabulary T1 Reading comprehension

.11 .38* .35 .25 .20 .36*

.26 .23 .47** .22 -.01 .32

.45* .32 .45* .44* .13 .36*

-.57** .54** .69** .37* .66**

-.67** .60** .55** .64**

-.53** .28 .56**

-.39* .89**

-.42*

--

10. T2 Phonological awareness .00 .49* .38 .67** .31 .38 .43* -.02 .43* -11. T2 Morphological production -.06 .29 .61* .53** .70** .65** .59** .44* .62** .45* -12. T2 Morphological analogy -.03 .36 .72** .66** .49* .72** .64** .14 .61** .53** .76** -13. T2 Word reading -.11 .25 .49* .53* .71** .59** .75** .38 .69** .54** .89** .77** -14. T2 Receptive vocabulary .39 .10 .09 .33 .65** .37 .54** .14 .55** .40 .43* .33 .50* -15. T2 Reading comprehension -.01 .25 .58** .57** .74** .58** .72** .28 .68** .56** .84** .75** .92** .61** _____________________________________________________________________________________________________________________________ _________ *p < .05, **p < .01

40 The Role of Morphological Awareness in Vocabulary Hierarchical regression analyses were conducted to examine predictors of vocabulary. Prior to running the analyses, we computed a morphological awareness factor score from principal component analyses for each cohort at each measurement point. This procedure was undertaken for two reasons. First, strong correlations were found between the two morphological measures, suggesting that they share a significant degree of commonality. Including redundant variables in the same analysis can inflate the size of error terms, thus weakening the analysis (Tabachnick & Fidell, 2007). A second and related reason was the finding of negative suppressions in our initial regression analyses. Thompson and Levine (1997) had previously noted that suppression effects can occur spuriously when predictors in the regression models are highly correlated. In the present study, the sign of the regression weight of the morphological tasks was negative in several of the regression models, which was the opposite of what we expected on the basis of its positive correlation with the outcome measures, and considerably reduced the interpretability of our results. Conducting regression analyses with factor scores thus enabled us to more appropriately determine the overall effects of morphological awareness on the outcome variables. Two sets of hierarchical linear regression analyses were carried out separately for each cohort to examine the concurrent, independent contributions of morphological awareness on vocabulary at the two measurement points. In each set of the analyses, a baseline control model was created to control for the effects of child’s incoming characteristics (age, non-verbal reasoning ability, mother’s education), and phonological awareness. Given the inter-relationships among morphological awareness, the three incoming characteristics, phonological awareness, and vocabulary, it is essential to

41 examine the predictive power of morphological awareness over and beyond that provided by these known predictors of reading performance. Control variables were kept in the subsequent models regardless of their significance level in order to account for the collinearity they may share with morphological awareness in predicting vocabulary. This is a particularly important concern given that morphological awareness tasks tap into phonological knowledge. Thus, for each model, the three incoming characteristics were entered in step one, phonological awareness was entered in step two, and the morphological awareness factor score was entered in step three. In addition, to explore the unique variance explained by each of the morphological tasks separately, we conducted two alternative hierarchical regressions within each set of analysis; in these regressions, only one of the morphological tasks was entered in step three. Results from the regression analyses conducted for the two cohorts are summarized in Table 4. The evaluations of the unique contributions of the morphological awareness factor score, the morphological production task, and the morphological analogy task to vocabulary are presented as Model A, Model B, and Model C, respectively. As shown in the upper half of the table, the morphological awareness factor score explained about 5% (p > .05) of the unique variance in vocabulary for the younger cohort at Time 1. Jointly, all variables entered in Model A explained about 18.5% of the variance in vocabulary; however, none of them emerged as a unique predictor of vocabulary. Similar results were obtained when the morphological analogy task was entered alone (Model C). By contrast, the morphological production task contributed to 16% of the variance (p < .05) when entered independently, and was a unique predictor of vocabulary. Together, variables in Model B explained about 31% of

42 the variance in vocabulary. At Time 2, the incoming characteristics accounted for a significant portion (32.4%) of the unique variance in vocabulary. When entered in the final step of Model A, the morphological awareness factor score explained significantly an additional 10.5% of variance in vocabulary beyond the control variables. Final beta weights suggested that the morphological awareness factor score was the only unique predictor of vocabulary. Altogether, predictor variables in Model A explained approximately 47% of the variance in vocabulary at Time 2. The contributions of the morphological tasks approached significance when they were considered separately; each task was able to explain approximately 7% of unique variance in vocabulary. The lower half of Table 4 presents the results for the older cohort. The morphological awareness factor score was the only unique predictor of vocabulary at Time 1, significantly accounting for 13% of variance in vocabulary (p < .05). Altogether, the variables considered in Model A explained about 33% of the variance in vocabulary. As shown in Model B, the morphological production task accounted for a significant 22% of the unique variance in vocabulary and was the only unique predictor of vocabulary. In combination, variables in Model B explained about 41.5% of the variance in vocabulary. Model C indicated that the morphological analogy task did not contribute to vocabulary significantly at Time 1. At Time 2, neither phonological awareness nor any of the morphological awareness scores made a unique, significant contribution to vocabulary. While all variables considered in each regression model explained some portion of variance (19% to 30%) in vocabulary, final beta weights indicated that there was no unique predictor of vocabulary at this measurement point in any of the models.

43 Table 4. Hierarchical Linear Regressions Predicting Concurrent English Receptive Vocabulary. __________________________________________________________________________________________________ Time 1 General model Model summary A

Time 2 General

Model B

Model C

model

summary

Model A

Model Model B C

Step and predictors ∆R2 β β β ∆R2 β β β __________________________________________________________________________________________________ Younger Cohort 1. Age Non-verbal ability Mother’s education 2. Phonological awareness 3. Morphological awareness 3. Morphological production 3. Morphological analogy

.135 .001 .049 .164** .001

.068 .262 .163 -.193 .292

.031 .384* .232 -.382~

.128 .296 .156 -.022

.528** -.032

.324* .038 .105* .075~ .074~

-.021 .009 .309~ -.018 .499*

-.011 .034 .371* .062

.049 .005 .332~ .041

.376~ .395~

Older Cohort 1. Age .029 .013 .126 .261 .283 .247 Non-verbal ability -.233 -.175 -.238 .041 .049 .032 Mother’s education .076 -.005 .007 .047 .086 -.310 -.287 -.169 2. Phonological awareness .121~ .160 .117 .300 .082 .246 .277 .277 3. Morphological awareness .130* .510* .092 .457 3. Morphological production .218** .597** .133 .474 3. Morphological analogy .029 .240 .025 .246 __________________________________________________________________________________________________ Note. Model A = MA factor score; Model B = Morphological Production task only; Model C = Morphological Analogy task only. ~ p