EARLY GROWTH TRAJECTORIES OF NARRATIVE LANGUAGE DEVELOPMENT: SEX DIFFERENCES IN MICROSTRUCTURAL AND MACROSTRUCTURAL SKILLS By Alexa Meier A THESIS Submitted to Michigan State University for the degree of 2020 in partial fulfillment of the requirements Communicative Sciences and Disorders – Master of Arts ABSTRACT By Alexa Meier EARLY GROWTH TRAJECTORIES OF NARRATIVE LANGUAGE DEVELOPMENT: SEX DIFFERENCES IN MICROSTRUCTURAL AND MACROSTRUCTURAL SKILLS Understanding how young children develop language is essential to promoting strong language growth. Unfortunately, current research indicates that young children’s language development differs by sex (Bouchard et al., 2009; Eriksson et al., 2012). These findings, however, are based on work examining language in isolation, which may not reflect children’s full capacity for language. Examining language within a more naturalistic context, such as occurs with a narrative assessment, may identify a different trajectory for girls and boys. To understand the sex differences in language development identified by previous research (Bouchard et al., 2009; Eriksson et al., 2012), the present study utilized a contextualized assessment, the Narrative Assessment Protocol (Bowles et al., 2020), to examine children’s content, form, and use of language across time. This study identified how preschoolers produce narratives and examined whether or not narrative skills develop differentially for girls vs. boys across time. Results indicated that for microstructural narrative skills, girls have higher skills than boys early on and this female advantage persists across time. However, for macrostructural skills, it was discovered that while boys begin preschool with lower skills than girls, the difference narrows within a year’s time. ACKNOWLEDGEMENTS First, I would like to thank my advisor, Dr. Hope K. Gerde, for all of her support throughout the completion of this project. Her patience, knowledge, and encouragement are inspirational and motivating. She is an amazing mentor and I am incredibly thankful for all of the years we have worked together. I would also like to thank the members of my committee: Dr. Courtney E. Venker, Dr. Ryan P. Bowles, and Matthew Phillips, CCC-SLP. Their immense knowledge and expertise provided me with the feedback and support necessary to complete this project. I would like to thank Dr. Ryan P. Bowles, Dr. Laura M. Justice, Dr. Kiren S. Khan, Dr. Shayne B. Piasta, Dr. Lori E. Skibbe, and Dr. Tricia D. Foster for allowing me to use their data to complete this research project. I would also like to thank the children and families who participated in this study. The research reported here was supported by the Institute of Education Sciences, U.S. Department of Education, through Grant R305A110293 awarded to Michigan State University. The opinions expressed are those of the authors and do not represent views of the Institute or the U.S. Department of Education. The study was approved by the Michigan State University Institutional Review Board IRB# 11-085. I am also thankful to Arianna Pikus, M.S. as well as Dr. Sarah Goodwin for their guidance in the completion of my analyses. Finally, I would like to thank my fiancé, Evan, as well as my family and friends for their support throughout this process. I am extremely grateful for all of you and I could not have done it without you. iii TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES INTRODUCTION Theoretical Frameworks of Language Development Importance of Oral Language Skills on Literacy and Academic Success Importance of Oral Language Skills for Overall Well-Being The Development of Language The Development of Narratives Sex Differences in Language Development Critique of Methods that Have Been Previously Used The Benefits of Narrative Assessment Current Study METHODS Participants Measures Procedures Data Analysis RESULTS Description of Children’s Narrative Skills Exploring Linear Models of Microstructural Language Components Exploring Linear Models of Macrostructural Language Components Multigroup Analysis DISCUSSION Girls and Boys Demonstrate Varying Growth Trajectories for Language Narrative Assessments Capture Important Variation in Language Development Clinical Implications Conclusions APPENDICES APPENDIX A: Family Questionnaire APPENDIX B: Tables APPENDIX C: Figures REFERENCES iv v vi 1 1 3 4 6 7 8 12 14 18 20 20 21 23 24 25 25 26 27 27 30 30 34 37 38 40 41 43 46 47 Table 1. Narrative Assessment Protocol Codes Table 2. Children’s Use of Microstructural Language Components Table 3. Children’s Use of Macrostructural Language Components LIST OF TABLES 43 45 45 v Figure 1. Microstructural Language Components Use By Sex Across Time Figure 2. Macrostructural Language Components Use By Sex Across Time LIST OF FIGURES 46 46 vi Success in school and social situations relies on children’s development of strong language skills. It is important to understand how young children develop language in order to support strong language growth, yet current research shows that young children’s language development differs by sex (Bouchard et al., 2009; Eriksson et al., 2012). These findings, however, are based primarily on work examining language in a decontextualized (e.g., standardized testing context) manner, which may not reflect children’s full capacity for language. By examining language with a narrative assessment that establishes a more naturalistic context for capturing language, a different trajectory for girls and boys may be determined. Therefore, this study examined how preschoolers produce narratives and whether or not narrative skills develop differentially for girls vs. boys. Communicating is something all humans must learn to do in order to interact with others, yet it is one of the most complex skills to develop. The vast majority does this through language. Language development can be a challenging and long-term process; development begins in utero and continues throughout our adult lives (Gomez & Gerken, 2000). Understanding the main components of language is important for effective communication. One theory of language development suggests that language can be divided into three individual components: content, form, and use (Bloom & Lahey, 1978). Another valuable theory divides language into three different, but similar, components: semantic, syntactic, and phonological (Lindfors, 1987). Language content can be considered the semantics of language, or how meaning is portrayed through words and signs (Bloom & Theoretical Frameworks of Language Development INTRODUCTION 1 Lahey, 1978). Semantic language involves the meaning behind words and parts of words (Lindfors, 1987). This is similar to Bloom and Lahey’s component of language content. Language form comprises multiple features including phonology, morphology, and syntax (Bloom & Lahey, 1978). Phonology includes the small units of sound in a language (e.g., /b/ or /p/). These sound units are used to make up words and can be divided into a segmental component, which includes phonemes and syllables (i.e., small units of sounds), and a suprasegmental component, which includes intonation, stress, and pause (Bloom & Lahey, 1978). Intonation is also known as the pitch in which you say a word (Bloom & Lahey, 1978). Stress can be conveyed by putting emphasis either on a specific syllable of a multisyllabic word or on a specific word in a sentence (Bloom & Lahey, 1978). Pause is identified when the speaker takes a breath before saying a specific syllable of a multisyllabic word or before a specific word in a sentence (Bloom & Lahey, 1978). Morphology is “the units of meaning that are words or inflections” (Bloom & Lahey, 1978, p. 15). More specifically, morphology looks at how words are created in a given language. This includes how adding a prefix or a suffix can change the meaning of a word. Syntax is “the ways in which units of meaning are combined with one another,” or the ways in which words are combined together to form sentences using grammar (Bloom & Lahey, 1978, p. 15). Syntactic language involves the basic organization of putting words together to form sentences (Lindfors, 1987). This is similar to the syntax aspect of Bloom and Lahey’s component of language form. Phonological language involves the basic sounds of language (Lindfors, 1987). This is similar to the phonology aspect of Bloom and Lahey’s component of language form as well. 2 Importance of Oral Language Skills on Literacy and Academic Success The final component of language, language use, is considered to be the functions of language, as well as the context in which we use different forms of language (Bloom & Lahey, 1978). Language can be functional both intrapersonally (e.g., when problem solving) and interpersonally (e.g., obtaining information from someone, Bloom & Lahey, 1978). The context of language can be linguistic (e.g., verbalizing an idea) or nonlinguistic (e.g., adjusting to recognize the needs of the audience). The intersection and careful use of each of these features, thus, creates language. While the organization of language differs by theory, a general summary would agree, and will be applied to the present study, that language is divided into components of content, form, and use (Bloom & Lahey, 1978). While oral language skills include both expressive (i.e., the words a child produces) and receptive (i.e., the words a child understands) language, the present study focuses on expressive language. This is because when a child begins preschool, it is expected that they will be able to use their oral language skills to communicate for themselves to their teachers and peers (Cooper et al., 2002). Furthermore, children’s developing early language skills (i.e., vocabulary, syntax) are necessary for academic success as well as to develop literacy skills (Harris et al., 2011). Oral language skills include vocabulary (i.e., expressive and receptive), syntactic knowledge (i.e., grammar), and the process of narrating (i.e., comprehension and storytelling), each of which have an effect on children’s literacy and reading achievement, with both the early stages of decoding vocabulary and the later stages of reading comprehension (Whorrall & Cabell, 2015). Oral language skills are important for school readiness (Cooper et al., 2002). Research has found that by 24 months, children who have smaller oral vocabularies are at 3 Importance of Oral Language Skills for Overall Well-Being greater risk of starting school at lower levels academically and behaviorally (Morgan et al., 2015). Clearly, strong language skills are essential for early and later for academic skills. Moreover, language development is important for children’s overall well-being. Beyond the established cognitive and academic benefits, oral language skills are strongly associated with children’s social and emotional development (Cooper et al., 2002). Children are much more likely to develop relationships and friendships if they are able to communicate with other people using social and pragmatic skills (e.g., describe what they are doing, ask questions, Cooper et al., 2002). This lack of communication skills may result in challenges in creating and maintaining friendships, because it will be more difficult for children to join their peers in play and other activities, as well as in creating and maintaining positive interactions with peers and adults (Artman et al., n.d.). A child’s ability to communicate their needs and emotions, and an adult’s response to this communication, is what establishes the learning pathways necessary to grow in all other areas of development (e.g., physical, cognitive, and emotional, Artman et al., n.d.). One rationale for this strong relationship between language and socioemotional development is that when children have higher language skills, their frustration with situations is lower because they are better able to communicate what they need and how they feel (Roben et al., 2013). When children are able to communicate these needs and feelings, in particular using their expressive language (including vocabulary, syntax, pragmatics), they are also better able to regulate their behavior depending on the social situation (Vallotton & Ayoub, 2011). This allows children to get the most out of each situation and continue learning and experiencing new things (Vallotton & Ayoub, 2011). When a child does not feel understood 4 due to poor oral language skills, they may feel less secure and will typically be less likely to communicate, diminishing their social and emotional communication skills. If these skills are compromised, this may lead to challenges for the child, including lower academic achievement and increased negative feelings toward themselves (Artman et al., n.d.). Recognizing speech and language difficulties as risk factors for child well-being, children with disabilities, including speech and language disabilities, are more at risk for abuse or maltreatment in comparison to their peers without disabilities (Sullivan & Knutson, 2000). In a study examining the maltreatment in over 50,000 children with specific disabilities, researchers found that children with speech and language impairments are 4.7 times more likely to experience neglect and physical abuse, 6.6 times more likely to experience emotional abuse, and 2.9 times more likely to experience sexual abuse than typically developing children (Sullivan & Knutson, 2000). This abuse may stem from a lack of understanding on the abuser’s end in regards to the child’s disability, the frustration of a parent not being able to understand their child’s needs/wants, or the increase in family stress related to the child’s struggles (Westby, 2007). It has been found that adults are likely to expect more than is developmentally appropriate from children with non-visible disabilities (e.g., developmental language disorders, Westby, 2007). Conversely, children who have been abused or neglected are more likely to present with language disorders, including having difficulties with expressive language (e.g., expressing their needs and feelings), and maintaining consistent and logical narrative dialogue (Westby, 2007). Clearly language development is important for success in school and life as even a slight delay in language skills may impact children’s development and well-being dramatically. Thus, understanding typical development is essential to ensuring that success. 5 The Development of Language For typically developing children, language development starts in infancy and continues to develop throughout life. In the first year of life, children are developing receptive language as they are beginning to comprehend what others are saying to them prior to the production of their first words (Harris et al., 2011). At birth, infants have the ability to learn any language (Kuhl et al., 1992). However, this capacity begins to change within the first year of life (Kuhl et al., 1992). Infants begin learning language from observation of caregivers and people they hear around them, rather than explicit instruction (Harris et al., 2011). In Vygotsky’s sociocultural theory, this is conceptualized in the idea that human development results from a dynamic interaction between individuals and society, with children continuously learning from the most prominent people in their lives including family members and educators (Vygotsky, 1978). They first learn the words that they hear most frequently from their environments (Harris et al., 2011). At about 2 months of age, infants are able to differentiate among phonemes, as well as produce speech-like sounds (e.g., cooing) by imitating the people surrounding them, even in languages not of their primary home language (Klee & Stokes, 2011). At about 6-7 months of age, infants begin to babble (e.g., “ma, ma, ma” Klee & Stokes, 2011). Also at this age, children can recognize sound patterns in language, frequent words and their meanings, as well as their own name (Harris et al., 2011). At around 10-12 months of age, infants begin to recognize the stress patterns of their home language. In addition, infants start to reduce their ability to discriminate across all phonemes and instead begin to specialize in the sounds of their home language or the language(s) they hear most often (Kuhl et al., 2008). 6 At about 12 months of age, infants begin to say single words (Klee & Stokes, 2011), including commonly heard nouns (i.e., food, animals) and verbs (e.g., eat, drink; Hoff, 2008). By 18-20 months, infants begin to combine words into their first phrases (e.g., “more juice”, “daddy play” Klee & Stokes, 2011). Between 18-24 months, infants’ expressive vocabulary rapidly increases, with a mean of 6,000 words by 2 years of age (Neuman, 2011). Beyond 24 months, it becomes increasingly difficult to estimate a child’s vocabulary size due to the number of new words they are rapidly learning (Klee & Stokes, 2011). With the increase in expressive vocabulary at 24 months comes usage of pronouns, grammatical markers (e.g., - ing), plurals (e.g., dogs), and past tense markers (i.e., -ed, Klee & Stokes, 2011). Also around 24 months, children begin to combine words into simple sentences (e.g., “what X doing?” Garrard, 1991). There are many skills and processes necessary for children to learn before they can produce a full, successful narrative. This includes understanding aspects of phonological awareness (i.e., the ability to understand the sound structure of words), metalinguistic ability (i.e., the ability to understand that language can be manipulated), and producing complex sentences using appropriate syntax. Some children start showing the beginning phases of phonological awareness as early as 2 years of age (Schuele et al., 2007). By 3 years of age, children start to develop some aspects of metalinguistic ability, by producing complex sentences (e.g., “He’s crying ‘cuz he got hurt” Klee & Stokes, 2011). With increases in vocabulary, syntactical structure, and phonological awareness, at around 2-3 years children begin producing simple narratives (Miller & Sperry, 1988). Narratives produced can include creating short stories on their own (McCabe & Rollins, 1994), or retelling The Development of Narratives 7 stories that they were either told or read to from people in their learning environments (Harris & Schroeder, 2012). By 3 to 4 years of age, children develop an awareness of rhyme, which is one of the first emerging skills of phonological awareness (Schuele et al., 2007). As children get older, they gain more phonological awareness skills, including initial sound recognition and noticing when words share beginning sounds (Schuele et al., 2007). By 4 years of age, there is a greater grasp on metalinguistic awareness, producing even more complex sentences that contain embedded clauses (e.g., “Don’t run or you’ll fall down and get hurt” Klee & Stokes, 2011). By 5 years of age, children have a strong command of oral language and use a range of words across the parts of speech (e.g., nouns, verbs, adverbs, adjectives), are able to hold conversations with peers and adults on one topic, and are able to understand and produce language about abstract events (e.g., Describing what happened at a birthday party in the past, Klee & Stokes, 2011). Children at this older age also are able to produce more creative, individualized narratives (Schick & Melzi, 2010). These more developed narratives include more complex semantics and syntax, including microstructural and macrostructural language components (described in detail below). Of course, this long process of early language development varies significantly across children. Sex Differences in Language Development While children do not differ by sex on many early developmental skills including motor skills like crawling, standing and walking (Onis, 2007; Smith et al., 2017), sex differences in language do appear early on. Research has found that starting from infancy, girls are more likely to use as well as understand more types of communicative items (e.g., gestures, eye contact, Eriksson et al., 2012). It has also been found that as children get older 8 into their toddler years (i.e., up until about 2 years of age), girls are found to show earlier word production skills and combine words earlier (Eriksson et al., 2012). Beyond what is known about early development, social influences may exacerbate these sex-based differences. Children’s language develops within a social context (Gee, 2003; Vygotsky, 1978). More specifically, children learn language from the people and environments surrounding them. Due to the social nature of language and language development, social and cultural factors influence this development, particularly for children of different sexes (Vygotsky, 1978). There are many ways in which social factors may bring about differences in language development between girls and boys (Eriksson et al., 2012). For example, parents communicate differently with girls and boys, accentuating politeness with girls and instruction with boys (Gleason & Ely, 2002). This idea is reinforced by observational work identifying that children living in heterosexual, two- parent households model their parents via role-play, especially seeing sex differences in their speech and language that they use (Gleason & Ely, 2002). Research has also found that adults interact differently with children of different sexes, playing with action-oriented toys more often with boys and caring-oriented toys more often with girls (Caldera et al., 1989). This is important because more language is generated from playing with caring-oriented toys like dolls (i.e., playing doctor to figure out how to make the doll feel better) than playing with action-oriented toys (i.e., racing cars, Caldera et al., 1989). In addition, mothers tend to use more exploratory play (i.e., playing with toys like blocks that you can manipulate) with boys, and symbolic play (i.e., playing with toys like dolls and cooking sets in a way that represents the real world) with girls (Pierce, 2009). Symbolic play involves more discussion among children about what the toys are doing/saying, whereas 9 exploratory play involves less discussion since children do not need to converse. Thus, it is no surprise that symbolic play is highly correlated with more advanced language skills (Pierce, 2009). These early communication differences may relate to the way children develop language, including narrative language. This is because these differences in communication from adults are creating experiences in which children are using language to express themselves and expectations of language in different ways. For example, research has found that narratives created by girls have been found to include more content centered around relationships and emotions than boys (Schick & Melzi, 2010). This could be due to the fact that mothers are more likely to talk with their daughters about emotions, especially when it comes to relationships (Fivush et al., 2000). Considering these varied language opportunities for girls and boys, it is not surprising that girls demonstrate higher/earlier language achievement than do boys. Previous research has found that girls have an advantage in acquiring language early on, in the first years of life (Bouchard et al., 2009; Eriksson et al., 2012; Maccoby, 1966). Girls are earlier than boys in saying their first words (i.e., mommy), as well as creating their first word combinations (i.e., hi mommy, Eriksson et al., 2012, Maccoby, 1966). Research has also found that in general, girls are more verbal than boys, often producing more words and have higher vocabularies, as well as producing longer sentences by utilizing grammar and syntax in more sophisticated ways (Bouchard et al., 2009, Maccoby, 1966). It is possible that some of this variation is due to the ways in which language has been measured and assessed in young children. 10 While research has found that in the first years of life, girls develop stronger language than boys, research has also found that, after a few years, boys “catch up” and differences are no longer identified. The results of multiple studies indicate that sex differences in language fade in the first few years of school (Bouchard et al., 2009; Coates, 1993). However, language is such an important skill and protective factor in children’s development that any sign of delay in language may put young boys at risk due to the potential loss in learning opportunities when their language skills were lower than that of girls. Thus, it is essential to identify the language trajectories of young boys to identify when they are different from girls and when they “catch up”. Research has found that once children start school and are exposed to shared book reading and vocabulary lessons every day, they learn a lot of new language (Wasik & Bond, 2001). In addition to the microstructural language skills of a narrative that are developed (i.e., new words including multiple parts of speech), this can include macrostructural language skills of a narrative, such as language related to the title, characters, and aspects of plot. Research has also found that when children are exposed to stories, their comprehension increases as observed in their retellings of stories (Isbell et al., 2004). This was specifically observed through an increase in the ability to label macrostructural components of language, like the setting of the story, the characters involved in the story, and the significance of the story (Miller & Pennycuff, 2008). Macrostructural language skills may develop earlier than microstructural skills, demonstrating a broad continuum of language development. Thus, by assessing microstructural language components as well as macrostructural language components, we may be able to identify whether sex-based differences exist across these skills or in just some areas of language. 11 Critique of Methods that Have Been Previously Used In children under 3 years of age, previous research studying the differences in girls’ and boys’ verbal ability has been completed more commonly through parent report instruments (Eriksson et al., 2012). Unfortunately, issues exist with parent reporter bias (Ladd & Burgess, 2001), which impacts the accuracy of the assessment. These concerns for reporter bias include social desirability biases (Fowler, 2002); that is, completing the survey in a way to be sure that whoever is reading their answers does not judge their parenting negatively. Research has found that while sometimes parents can be reliable in reporting physical milestones (i.e., first step), they sometimes are unable to reliably report language milestones (Miller et al., 2017). While parents can be reliable reporters depending on the information they are being asked to report, this is typically found reliable from studies using decontextualized assessments that focus on vocabulary usage (Miller et al., 2017). A direct assessment of language that is contextualized may prevent issues from parent report biases and provide more accurate data on the differences in girls’ and boys’ verbal ability. However, direct assessments of language can also have several limitations. For example, some direct assessments of language only assess some constructs (Bachman, 2007). Only examining certain pieces of language is problematic because that does not describe the entire picture of a child’s language abilities (Bachman, 2007). Direct assessments of language also often have bias towards various groups including sex, ethnicity, and nationality (Campbell et al., 1997). It could be that language assessments in which children of minority groups score lower are in fact only scoring lower because of their differences in background and experience rather than in linguistic deficits (Campbell et al., 1997). For example, oftentimes direct language assessments examine children’s 12 language in ways that can be influenced by life experiences and general knowledge (Campbell et al., 1997). This could be problematic for children who haven’t had as many experiences in life including events outside of their neighborhood (Campbell et al., 1997). By contextualizing language within a framework familiar to a child, assessors will see the child’s true capability of language production. If assessors test children on decontextualized assessments using vocabulary that children have never heard before, then children are more likely to score lower on assessments and assessors will not see the child’s true language ability. Previous research studying differences in girls’ and boys’ verbal ability also has been examined through the use of standardized assessments (Eriksson et al., 2012). Much of this previous work on young children’s language development has focused on vocabulary development specifically (Barbu et al., 2015; Eriksson et al., 2012; Eriksson, 2016; Miller et al., 2017). Assessments include asking children to provide names and/or selecting images to identify actions and objects and asking children to finish sentences with the appropriate vocabulary (i.e., this cat is big and this cat is ___, in which we would expect the child to say ‘small’). Such an approach provides a limited and decontextualized model of language; vocabulary assessments evaluate children’s understanding of words in isolation. Such assessments do not invite children to share what they do know within a familiar context and they do not allow children to relate things to the real world. Rather, these assessments ask children about a narrow set of words which may or may not reflect their personal experiences. Issues exist with these standardized assessments, including the fact that they are oftentimes not ecologically valid since they ask children to sit somewhere quiet, away from their peers, and do activities that are uncommon to normal, everyday life 13 The Benefits of Narrative Assessment (Brassard & Boehm, 2007). A more comprehensive, contextualized evaluation of language, such as that provided by a narrative assessment, might allow children to demonstrate more of their language skills and subsequently depict a different trajectory of development for girls and boys. This is the focus of the present study. A narrative is defined as the production of a series of events that are in a sequential order (Engel, 1995); narrative is the telling of an event or story. We use narratives regularly in our daily dialogue as we converse with others about a range of topics. The use of narratives is very important to many cultures, since they are used in the passing down of stories from generation to generation (Schick & Melzi, 2010). Children use narratives to relate events, establish and maintain friendships, and express their thoughts and feelings about important topics (McCabe & Bliss, 2003). Narrative is a creative and developmentally appropriate way to examine multiple components of children’s language. Narratives can be collected from children, including young children (Justice et al., 2010), in order to assess a variety of skills, including broad language skills such as semantics and syntactical structure, to specific language components, such as narrative elements that are microstructural and macrostructural (Justice et al., 2010). Microstructural elements are focused on the words the child uses including the child’s use of verbs, adverbs, and adjectives (Justice et al., 2010). Macrostructural elements of narratives can include broad elements of stories including title, characters, and plot. A narrative assessment will permit the examination of both microstructural and macrostructural components of children’s language, which may identify when and for which sex differences exist (Fiorentino & Howe, 2004). It is 14 important to study both microstructural and macrostructural components of children’s language, because the two components together cover what is needed for a successful narrative. Microstructural components of language provide the specific words that put the story together, and macrostructural components of language help guide how to create a narrative for the story. Theoretically, narratives require the use of and thus, allow for the examination of, all areas of language—content, form, and use—included in the theoretical models presented above (i.e., Bloom & Lahey, 1978; Lindfors, 1987). In children’s narratives, language content is represented as the words that are used to tell a story. Language content is how meaning is portrayed through the words (Bloom & Lahey, 1978). Thus, it is reflected in the complexity of children’s language through the richness of the vocabulary they use to tell their stories. Language form is expressed as how the child puts words together to make sentences, how the sentences are combined to create scenes in a story, and how the scenes in a story are combined to create a well thought out, complete story. The ways in which children put words together to combine sentences, scenes, and a complete story are through the use of microstructural language components. Lastly, language use is represented by how the child puts a story together, specifically with macrostructural language components. This includes whether or not the child uses aspects of a narrative, like characters, a title, and a plot (described in detail below). Another important benefit of assessing language through a narrative is that narratives capture language within a context. When obtaining a narrative language sample, professionals are able to collect large amounts of children’s language and use this language to understand how children use language in a contextualized setting (Costanza-Smith, 15 2010), that is, in a real life context as opposed to an abstract setting (i.e., asking a child what they did at school yesterday instead of asking a child a question from a standardized assessment like what would you do in this scenario?). This suggests that obtaining language samples through narrative is less biased than standardized language assessments (Justice et al., 2006). This is because a narrative is a contextualized way of assessing language since it captures language at the story level rather than at the word or phrase level as occurs with decontextualized assessments that target parts of speech in isolation. Research has found that context matters when assessing younger children (Bachman, 2007). In addition, narratives reflect an authentic way to examine language; children are able to apply their knowledge of real-life, familiar situations to their assessment. Since narratives are a contextualized measure of language, they are both socially (i.e., having conversations with family members about how they’re feeling or what they did at school that day) and academically (i.e., having a conversation about a book they read in school that day) meaningful (Pindzola et al., 2016). Assessing narrative skills in children are particularly beneficial to clinicians, including speech-language pathologists. Narratives allow clinicians to assess children’s expressive language skills (Justice et al., 2006). Research has found that assessing children’s narrative skills allows clinicians to differentiate children who are developing typically from children who may have a language impairment, since children with language impairments are more likely to use fewer microstructural language components (Justice et al., 2006). Narratives are particularly useful to clinicians working in settings where they have difficulty setting large time frames aside for assessment, since some collections of 16 children’s narratives can occur briefly (e.g., during conversation, while children are playing, observing the children in the classroom, Southwood & Russell, 2004). Although advantages of using narrative to assess children exist, including allowing professionals to identify a baseline as well as long-term language growth starting in preschool (Justice et al., 2010), professionals find narrative assessment arduous due to the time they take to collect, transcribe, and analyze. Professionals are also less likely to collect narratives unless they have normed data to compare their students to, since often they have only enough time to assess and evaluate using one or two assessments and they won’t want to choose one that is not normed (Justice et al., 2006). Fortunately, the Narrative Assessment Protocol (NAP) was created to ease the burden of professionals who otherwise might not choose to use a narrative in their language assessment battery (Justice et al., 2010). The NAP-2 (Bowles et al., 2020) was created to enhance the NAP by scoring narratives from a video recording of the child saying their narrative, removing transcription from the task list of assessing a narrative sample. The NAP is useful because it allows professionals to analyze language samples for both microstructural and macrostructural components of language, which is rare since narrative assessments often only examine macrostructural components of language (Justice et al., 2006). The NAP is relatively quick to collect and analyze, given the time it takes for a child to produce their narrative is typically less than ten minutes and the time it takes to score from video also takes approximately ten minutes (Bowles et al., 2020; Justice et al., 2010). The assessment is both ecologically valid and contextualized, enhancing the validity of the measure (Bowles et al., 2020; Justice et al., 2010). Due to the multiple testing forms available, the measure allows clinicians and educators to collect narratives as a baseline to examine children’s 17 language skills at the beginning of school and over time to monitor progress (Bowles et al., 2020; Justice et al., 2010). There are multiple ways to capture narratives when assessing children. Researchers can assess children while they play freely (Southwood & Russell, 2004). If assessors are looking for opportunities for children to create complex and longer phrases, children can be assessed by asking them to create stories (Southwood & Russell, 2004). For example, researchers can ask children to tell a past personal experience, tell a fictional story, or explain or describe an object or a process (Elleseff, 2017). In order to create more opportunities for more complex phrases, assessment of children can occur while adults have conversations with them (Southwood & Russell, 2004). A final example of capturing a narrative is through a narrative retelling in which the assessor reads a child a story and asks him/her to retell it (Elleseff, 2017). For young children, assessing narrative skills through a retelling is beneficial (Schneider & Dube, 2005); the retelling allows children to create longer and more complex stories compared to spontaneous productions (Schneider & Dube, 2005). To understand the sex differences in narrative language development identified by previous research (Bouchard et al., 2009; Eriksson et al., 2012), the present study utilized a contextualized assessment, the NAP-2, to examine children’s content, form, and use of language across time. The study identified how preschoolers produce narratives and examined whether or not narrative skills develop differentially for girls vs. boys across time. The following research questions guided this study. Current Study 18 1. What are the frequencies of young children’s use of microstructural and macrostructural components of language in a narrative retelling? 2. Do the frequencies of young children’s use of microstructural and macrostructural components of language in a narrative retelling change across time? 3. Do the frequencies of young children’s use of microstructural and macrostructural components of language in a narrative retelling differ by sex over time? I expected that in the initial time point, children would have low frequencies of both microstructural and macrostructural components of language and that they would increase in each of the subsequent timepoints two through four. I expected that girls would use more microstructural and macrostructural components of language compared to boys. I expected that boys might use fewer microstructural and macrostructural components of language early on and, as they get older, would use more equal amounts of microstructural and macrostructural components of language as compared to girls. 19 METHODS Participants The present study used archival data from a larger study designed to examine the reliability and validity of the narrative assessment tool, the Narrative Assessment Protocol, second edition (NAP-2; Bowles et al., 2020). Participants of the NAP-2 study included 243 children ages 3 to 5.5 years (M = 4.28 years, SD = .65). Children who met eligibility criteria (i.e., children proficient in English, and children who did not have any significant language or developmental delay according to parent report in the family questionnaire) joined the study (see Appendix A for family questionnaire). Forty-four percent of the children were girls. Per parent report, children reflected diverse ethnicities including 60% White/Caucasian, 18% Black/African American, 17% Multiracial, and 3% Asian (2% unreported). In addition, 7% of children were identified as Latinx. When looking at caregivers’ highest degree earned, 6% did not have a high school diploma, 25% earned a high school diploma, 7% earned an Associate’s degree, 25% earned a Bachelor’s degree, 26% earned a Master’s degree, and 9% earned a Doctoral degree (2% other or did not report). To be included in this present study, participants had to participate in three of the four time points. This inclusion criteria resulted in a sample size of 219 children ages 3 to 5.5 years. Of the children, 45.7% were female. Children reflected diverse ethnicities including 62.2% White/Caucasian, 18.0% Black/African American, 17.5% Multiracial, and 2.3% Asian (1.4% unreported). In addition, 7.6% identified as Latinx. While all children were assessed in English, another language was identified as the primary language for 4.1% of children. When looking at caregivers’ highest degree earned, 5.5% earned less than 20 a high school diploma, 3.2% earned a high school diploma or an equivalent degree, 3.7% earned a high school diploma or an equivalent degree as well as a technical training certificate, 16.5% completed some college with no degree, 5.5% earned an Associate’s degree, 21.1% earned a Bachelor’s degree, 5.5% earned at least one year of coursework beyond a Bachelor’s degree, 22.5% earned a Master’s degree, 3.2% earned a professional diploma based on at least one year of coursework beyond a Master’s degree, and 9.6% earned a Doctoral degree (3.7% other). Measures Children’s oral language production was assessed using the Narrative Assessment Protocol (NAP-2, Bowles et al., 2020). The NAP-2 is administered using one of four wordless picture books that are 16 pages long. Titles of the picture books are: Wolf Cleans His Bedroom, Tiger Gets Ready for Bed, Raccoon Makes Lemonade, and Rabbit Goes for a Bike Ride. Children are told that the assessor will tell them a story using the pictures in the book and then it will be their turn to tell a story using the pictures. The assessor reads the child the book using a script, and then prompts the child by telling them, “Now it is your turn to tell me a make-believe story using the pictures in this book”. The goal of the assessment is for the child to retell the story that was read to them. However, it is acceptable if they add or change elements of the story that was read to them by the assessor. Scripts are between 364 and 375 words long. Each story follows the same format containing a conventional opening, the setting of the stage, a goal for the main character, multiple sub-goals, a problem to solve, a solution to that problem, and a conventional ending. Children were randomly assigned to wordless picture books, one at each timepoint of the study and never repeated. To allow for ease and reliability of coding later on at the 21 lab, the assessments of children were video recorded. Ten percent of narratives were double coded, with an inter-coder agreement of .85. The internal consistency was α = .81 as computed by Rasch analysis using Winsteps (Bowles et al., 2020). The measure has strong evidence of content and construct validity, which was computed by Rasch measurement techniques (Bowles et al., 2020). Trained and reliable research assistants scored narratives in the lab using the video recordings. Prior to training, three master coders coded 18 videos to be used in training and reliability sessions. Prior to scoring, research assistants had to become reliable. Research assistants had to read sections of the Syntax Handbook (Justice & Ezell, 2002) to become familiar with key terms relevant to coding (e.g., verb, noun, clause, phrase). They then watched two videos to become familiar with the scoring process. They then practiced scoring five videos and compared their scores to the master codes. Lastly, they scored three videos and were required to reach 85% agreement with the master codes in order to demonstrate reliability in coding. If they did not become reliable, they were allowed to try again with three new videos. They were able to repeat this reliability training up to a total of four times. The coding protocol includes 19 items which are title, conventional opening, character reference, temporal ordering, emotion reference, elongations, similes and metaphors, time reference, tier two verbs, tier two adverbs, tier two adjectives, interrogative “wh” questions, conjoined adverbial phrase, elaborated noun phrase, sub- goal, sub-problem, sub-solution, sub-resolution, and conventional ending. For definitions and examples of each of these codes see Table 1. Microstructural codes include tier two verbs, tier two adverbs, tier two adjectives, interrogative “wh” questions, conjoined adverbial phrase, and elaborated noun phrase. Macrostructural codes include title, 22 conventional opening, character reference, temporal ordering, emotion reference, elongations, similes and metaphors, time reference, sub-goal, sub-problem, sub-solution, sub-resolution, and conventional ending. Procedures To recruit the participants, research staff distributed informational fliers about the study at childcare centers, schools, and community centers in two mid-western states. Parents called or emailed the research lab to enroll their child into the study. Upon entry into the study, caregivers completed demographic questionnaires and signed consent forms. Children gave verbal assent to participate. The children and their caregivers were compensated for their participation with a gift card. The procedures of this study were approved by the university Institutional Review Board. Prior to data collection, research assistants were trained on the narrative assessment protocol by reading the assessment carefully and becoming familiar with the story and scoring. They were given time to practice administering the assessment with peers. Once an experienced assessor observed the research assistants administering the assessment and provided them with feedback and it was determined that they were administering correctly, they were approved to assess children in the study. Trained research assistants assessed children in a quiet location convenient to the parent, either at the child’s school or home, or at a community building (i.e., local library). Data was collected every six months for a total of four time points. Data was collected via video recording so that the recorded videos could be returned to the university laboratory for coding by trained coders. 23 Data Analysis Descriptive statistics including means, standard deviations, and frequencies were used to examine the frequencies of young children’s use of microstructural and macrostructural components of language in a narrative retelling. Microstructural components of language that were examined included tier two verb, tier two adverb, tier two adjective, conjoined adverbial phrase, and elaborated noun phrase. Macrostructural components of language that were examined included title, conventional opening and ending, character reference, temporal ordering, emotion reference, elongations, similes/metaphors, gratuitous terms, time reference, interrogative ‘wh’ questions, sub- goal, sub-problem, sub-solution, and sub-resolution. Growth curve analysis was used to examine whether the frequencies of young children’s use of microstructural and macrostructural components of language in a narrative retelling change across time. A two- level model in which time reflects level-one and child sex is included at level-two was fit to the data to examine the hypothesized differential trajectories for girls and boys. 24 RESULTS Description of Children’s Narrative Skills To answer research question 1, What are the frequencies of young children’s use of microstructural and macrostructural components of language in a narrative retelling, descriptive statistics were used. At timepoint 1, children used an average of 1.26 microstructural language components (see Table 2). However, variation existed (SD = 1.79; Range = 0-9.00). While most children used no microstructural language components, when they did, often this score reflected the use of tier two verbs and elaborated noun phrases. Children’s use of these language components increased over time and by timepoint 4, children used an average of 2.40 microstructural language components (SD = 2.19; Range = 0-9.00). These microstructural language components used by children primarily were made up of tier two verbs, tier two adverbs, and elaborated noun phrases. The majority of children (52.40%) in timepoint 1 did not use any microstructural language components in their narratives. Over the second, third and final timepoint, however, the number of children who used zero microstructural language components decreased dramatically (T2 = 41.80%; T3 = 28.80%; T4 = 21.40%). At timepoint 1, children used an average of 8.88 macrostructural language components (see Table 3). However, variation existed (SD = 5.18; Range = 0-26.00). In general, this reflects their use of sub-goals, sub-problems, sub-solutions, sub-resolutions and conventional endings in their narrative. Children’s use of these language components grew over time and by timepoint 4, children used an average of 14.54 macrostructural language components (SD = 4.03; Range = 0-23.00). The same five macrostructural language components were shown in timepoint 4, with the addition of children referencing 25 microstructural and macrostructural components of language in a narrative retelling change characters. Unlike microstructural language components, macrostructural language component usage was much higher, demonstrating that these skills were developed earlier. For example, in timepoint 1, 88.9% of children used 1 or more macrostructural language component. Over the second, third, and final timepoint, this use of 1 or more macrostructural language components increased dramatically (T2 = 92.00%; T3 = 96.2%; T4 = 98.8%). To answer part of research question 2, Do the frequencies of young children’s use of across time, microstructural language component data was modeled first, with an intercept only model with three parameters (intercept mean, intercept variance, and residual variance), and second, with a linear model with six parameters (intercept and slope means, intercept and slope variances and their covariance, and residual variance). For the intercept only model, model fit was poor: X2 (11) = 86.168; RMSEA = .143; CFI = .00. For the linear model, model fit was good: X2 (8) = 8.753; RMSEA = .017; CFI = .966. Microstructural language component data at the beginning was estimated to be 1.099 (SE = .083, p < .01) and the estimated average linear change was .367 (SE = .050, p < .01). The estimated variation in the intercepts was .059 (SE = .214, p = .784) and the estimated variation in the slope was .078 (SE = .071, p = .276). The linear growth curve model was selected as the final unconditional model for microstructural language components given that it fit the data better than the intercept only model. Since the data fits this linear growth curve model, this provides evidence that children’s microstructural language grows linearly across time. Exploring Linear Models of Microstructural Language Components 26 microstructural and macrostructural components of language in a narrative retelling change Exploring Linear Models of Macrostructural Language Components To answer part of research question 2, Do the frequencies of young children’s use of across time, macrostructural language component data was modeled first, with an intercept only model with three parameters (intercept mean, intercept variance, and residual variance), then with a linear model with six parameters (intercept and slope means, intercept and slope variances and their covariance, and residual variance), and finally with a latent basis model with eight parameters (intercept and slope means, intercept and slope variances and their covariances, residual variances, and basis coefficient at timepoints 2 and 3). For the intercept only model, model fit was poor: X2 (11) = 293.678; RMSEA = .277; CFI = .00. For the linear model, model fit was good: X2 (8) = 17.154; RMSEA = .058; CFI = .944. A latent basis model was run and did not improve model fit (X2 (6) = 14.399; RMSEA = .065; CFI = .948), therefore, the linear model was chosen. Macrostructural language component data at the beginning was estimated to be 8.421 (SE = .280, p < .01) and the estimated average linear change was 1.949 (SE = .110, p < .01). The estimated variation in the intercepts was 14.238 (SE = 2.132, p < .01) and the estimated variation in the slope was .215 (SE = .357, p = .546). Since the data fits this linear growth curve model, this provides evidence that children’s macrostructural language grows linearly across time. To answer the first part of research question 3, Do the frequencies of young children’s differ by sex over time, a multigroup analysis was conducted in order to assess if there were differences based on sex in the development of microstructural language components. 27 Multigroup Analysis use of microstructural and macrostructural components of language in a narrative retelling First, a model was run where all parameters were constrained to be equal across both groups. For this model, there were 20 degrees of freedom with a chi-square of 25.612. Next, a model was run allowing the means of level and slope to vary based on sex. For this model, there were 17 degrees of freedom with a chi-square of 17.032. The chi-square difference test showed a significant difference where p < .05, therefore this model with unconstrained means was determined to be a better representation of the data. For girls, the unstandardized microstructural language component data at timepoint 1 was estimated to be 1.216 (SE = .127, p < .01) and the estimated average linear change was .419 (SE = .082, p < .01). For boys, the unstandardized microstructural language component data at timepoint 1 was estimated to be 1.002 (SE = .109, p < .01) and the estimated average linear change was .322 (SE = .061, p < .01). The standard errors and p-values listed above show there were significant mean level and slope differences on microstructural language components based on sex. In addition, these differential trajectories provide evidence that girls start out at timepoint 1 using more microstructural language components, and by timepoint 4 girls continued to use more microstructural language components (see Figure 1). To answer the second part of research question 3, Do the frequencies of young retelling differ by sex over time, a multigroup analysis was also conducted in order to assess if there were differences based on sex in the development of macrostructural language components. First, a model was run where all parameters were constrained to be equal across both groups. For this model, there were 20 degrees of freedom with a chi-square of 44.082. Next, a model was run allowing the means of level and slope to vary based on sex. For this model, there were 17 degrees of freedom with a chi-square of 35.234. The chi- children’s use of microstructural and macrostructural components of language in a narrative 28 square difference test showed a significant difference where p < .05, therefore this model with unconstrained means was determined to be a better representation of the data. For girls, the unstandardized macrostructural language component data at timepoint 1 was estimated to be 9.309 (SE = .935, p < .01) and the estimated average linear change was 1.791 (SE = .152, p < .01). For boys, the unstandardized macrostructural language component data at timepoint 1 was estimated to be 7.694 (SE = .389, p < .01) and the estimated average linear change was 2.067 (SE = .158, p < .01). The different standard errors and significant p-values listed above show there were mean level and slope differences on macrostructural language components based on sex. In addition, these differential trajectories provide evidence that while girls start out at timepoint 1 using more macrostructural language components, by timepoint 4 boys nearly approached the same score as girls in their use of macrostructural language components (see Figure 2). 29 DISCUSSION The findings of this study contribute new knowledge about the early trajectories of girls’ and boys’ language development by describing early narrative growth for microstructural and macrostructural skills and pinpointing when and where differences exist. Differences in girls’ and boys’ narrative skills begin early and persist for microstructural skills. For macrostructural skills, however, this work identifies that while boys begin preschool with lower skills, they narrow the advantage of the girls within a years’ time. If the study continued on for another timepoint of 6 months, it is expected that boys would catch up to and surpass girls in their use of macrostructural language components. Importantly, narrative assessment is one valuable strategy for identifying the differential trajectories of young children’s language development. Identifying when and on which skills girls and boys differ in their language development provides critical information for researchers, educators, and therapists regarding when and on which skills to intervene. Girls and Boys Demonstrate Varying Growth Trajectories for Language Sex is one variable that explains some of the variation in children’s initial skills and growth trajectories for language. Girls and boys begin preschool with differing skills and grow in their language skills at differing rates from the ages of 3-5.5 years. I found that over 18 months, at each timepoint of 6 months, children of both sexes increased in their use of microstructural and macrostructural language components, both growing linearly over time. In addition, for microstructural components of language, girls start ahead early on in preschool using significantly more advanced language, and boys do not catch up within 18 months’ time. With macrostructural components of language, however, girls start ahead 30 early in age using more components within their narratives and boys eventually narrow the gap using nearly as many components 18 months later. Based on the trajectories identified, I estimate that if the study continued on for another timepoint of 6 months, boys would catch up to and surpass girls in their use of macrostructural language components. For microstructural components of language, girls retain an advantage. While previous foundational work has identified an early advantage for girls (Bouchard et al., 2009; Eriksson et al., 2012; Maccoby, 1966), the results of this study expand the previous trajectory identified for infants and toddlers to preschoolers age 3-5.5 years old noting a continuation of this female advantage. This could potentially be due to the fact that parents talk to and interact with their daughters and sons differently. Differences in communication early on, both in interactions and in play, between parents and children may relate to how children develop language, more specifically narrative language. For example, mothers talk more and use more supportive speech with daughters compared to sons in children age 3-4 years (Fivush et al., 2000). Even in their play, parents use approaches that elicit more language from girls, including caring-oriented and symbolic play, than with boys, with whom they tend to engage in exploratory and action-oriented play (Caldera et al., 1989; Pierce, 2009). This identified lower skill in microstructural components of language for boys during this critical time of rapid language development is a problem. Due to their less developed expressive language, boys may miss out on opportunities to use language to express their wants, needs, and ideas. These children may struggle to develop positive relationships with their peers, because they are unable to use language to explain events in their lives, as well as their feelings, wants, and needs (Cooper et al., 2002). Research has 31 found that children who have limitations in their language ability are less likely to be identified as preferred playmates by their peers (Gertner et al., 1994); higher level of language ability was a better predictor of peer status than age or intelligence (Gertner et al., 1994). Similarly, research has found that dual language learners with lower English language struggle to develop peer relations and are rated lower in popularity by their peers in preschool (Gertner et al., 1994). Lower language skills may influence boys’ literacy skills as previous work has shown language to be important for literacy development (Harris et al., 2011; Whorrall & Cabell, 2015). Understanding when and where boys are behind in language affords the opportunity to alleviate potential negative effects of low/limited language, which can influence literacy (Harris et al., 2011; Whorrall & Cabell, 2015) and overall well-being (Westby, 2007). For macrostructural components of language, boys narrow the gap. One important contribution of this study, which responds to an important limitation of studies that only examine microstructural language components, is that the deficit in early language skills in terms of macrostructural language components identified for boys does not persist. While previous work identifying language deficits for boys have primarily examined microstructural components of language (e.g., Bouchard et al., 2009; Eriksson, 2012), the present study indicates that for macrostructural components of language, boys do narrow the original gap identified in skills across preschool. This narrowing of the gap could be due to increased exposure to stories, story features, and story structure through read-aloud experiences traditional of preschool classrooms. For example, early on girls are more likely than boys to sit at home with an adult and read a book or listen to stories (Barrs, 1993); boys are less likely to be read to prior to school entry (Barrs, 1993). Once boys enter 32 preschool, they typically receive regular exposure to stories, which may contribute to the narrowing of the gap identified for these skills prior to school entry. This is likely because book reading is a common practice in preschool classrooms, with one study finding that early childhood educators spent 20% of each day providing children with language and literacy-learning opportunities including book reading (Pelatti et al., 2014). Further, children tend to be familiar with narrative; stories with genres of fantasy and contemporary realistic fiction, that is, narratives, as opposed to other genre like non-fiction, poetry, etc., are the most common type read in preschool classrooms (Crisp et al., 2016; Pentimonti et al., 2011). As children are exposed to more read-alouds of storybooks, they are exposed to more macrostructural language components including learning about books’ titles, use of characters, conventional openings and endings, use of time and temporal terms, and aspects including character goals, problems, solutions and resolution. These are important components of language captured by the NAP-2, the measure of language used in this study. While the present study identified the number of components that children cover, other features of children’s narrative including the quality of language could be examined in future work. Teachers and clinicians can capitalize on these macrostructural language skills in order to promote stronger microstructural language components, particularly for boys. They can provide opportunities for children to tell stories as well as to use language to talk about characters, goals, problems, and solutions during read-alouds, which can prompt language growth in microstructural skills as well (Hargrave & Sénéchal, 2000; Miller & Pennycuff, 2008; Wasik & Bond, 2001; Zucker et al., 2013). When teachers and clinicians participate in repetitive interactive book reading, preschool and kindergarten aged 33 Narrative Assessments Capture Important Variation in Language Development children’s comprehension and vocabulary development increases (McGee & Schickedanz, 2007). Studies have found that an increase of shared reading in early childhood classrooms results in an increase in children’s expressive vocabulary (Grolig et al., 2019; Zucker et al., 2013), as well as an increase in grammar, comprehension monitoring, and narrative comprehension skills (Grolig et al., 2019). Engaging children in shared book reading by asking them to label or point to images can promote vocabulary growth (Sénéchal et al., 1995). The use of narrative assessment in this study provides further evidence that narrative approaches can be used successfully to assess young children’s language (Bowles et al., 2020; Justice et al., 2010). As in previous work (Bowles et al., 2020; Justice et al., 2010), my study demonstrates that narrative assessment is capable of identifying variation in very young children’s language. My study is an important expansion of this previous work describing children’s narrative skills using 1-2 timepoints (e.g., Justice et al., 2010) by articulating children’s longitudinal trajectories of narrative development across four timepoints. Practically, my study confirms that due to the multiple equivalent books, the NAP-2 can be used to assess children over multiple timepoints using the same measure to identify growth in language for children across this age range (i.e., 3-5.5 years). Moreover, by collecting data using a narrative, I could examine multiple components of children’s language better reflecting the multiple features of language – content, form, and use – included in theories of language development (i.e., Bloom & Lahey, 1978; Lindfors, 1987). This measure allowed me to collect data to assess a variety of skills, including broad language skills such as semantics and syntactical structure, to specific 34 language components, such as narrative elements that are microstructural and macrostructural (Bowles et al., 2020; Justice et al., 2010). In children’s narratives, language content is the words that come together to create a story. In my study, this includes some of the microstructural codes of the NAP-2 including tier two verbs, tier two adverbs, and tier two adjectives. Language form is the process in which a child puts words together to create sentences, and then puts sentences together to create scenes, and finally puts scenes together to create a complete story. In the NAP-2, this includes some of the microstructural codes like interrogative “wh” questions, conjoined adverbial phrases, and elaborated noun phrases. Finally, language use is the process in which a child puts together a story using the macrostructural language components. In the NAP-2, these codes include providing a story title, a conventional opening, character references, temporal ordering, emotion references, elongations, similes and metaphors, time references, sub-goal, sub-problem, sub-solution, sub-resolution, and providing a conventional ending. Using a narrative assessment, I was able to discover how language develops across time and how this differs between girls and boys. This approach is divergent from studies using only vocabulary or decontextualized assessments of language. One study found that girls used more gestures, produced more vocabulary, and were earlier in combining words, than boys and that this sex difference increased with age (Eriksson, 2012); while my study found that boys catch up to girls in certain language skills. Another study found that girls have more advanced language skills compared to boys at a young age (Bouchard et al., 2009). That study, however, used decontextualized assessments to study language, while my study used a more comprehensive assessment, which gave a fuller picture of children’s language production. Narrative assessments capture details about both the microstructural 35 and macrostructural components of children’s language skills and development. Obtaining language samples through narrative is less biased than many standardized language assessments (Justice et al., 2006). This is because narrative is a contextualized way of assessing language because it captures language at the story level rather than at the word or phrase level; decontextualized assessments target parts of speech in isolation (e.g., Barbu et al., 2015; Eriksson et al., 2012; Eriksson, 2016; Miller et al., 2017). The use of contextualized assessments may be one reason why we can see the early differences in the use of macrostructural components of language between girls and boys. In addition to standardized assessments of language, narratives have a role to play in examining children’s language development. Narrative assessments have multiple benefits for use by practitioners including therapists and educators. Narratives can be gathered in many ways through natural settings including during typical classroom activities, while children are at play, and during therapy sessions (Elleseff, 2017; Southwood & Russell, 2004). This variety of settings is beneficial because it permits testing at times when the child is utilizing language the most, allowing for the most realistic version of their use of language. The data gathered from narratives can be used for multiple purposes, including collecting for microstructural and macrostructural language components, as demonstrated in this study, but also for examining number of words and parts of speech (Justice et al., 2006). This allows professionals to see many aspects of children’s language, rather than just one specific feature. This multicomponent approach allows professionals to identify strengths and areas where they might work with children to improve language. Narratives reflect an authentic way to examine language; children are able to apply their knowledge of real-life, familiar situations to their assessment (Pindzola 36 Clinical Implications et al., 2016). By using narratives, professionals are providing children with the best opportunity to portray their language skills, rather than zooming in on one specific piece of language and not seeing the whole picture. Narratives are socially and academically meaningful (Pindzola et al., 2016), and represent a contextualized assessment of language. More specifically looking at the NAP-2, it is beneficial that clinicians can code the data collected from this assessment without transcribing (Bowles et al., 2020). Although becoming reliable in the coding system is required, alleviating the transcription of every narrative decreases time spent on completing and processing the assessment data. Training of the NAP-2 coding system can be found at http://www.narrativeassessment.com. The findings of this study reiterate a female advantage for language, specifically microstructural components, as well as a novel finding that boys do narrow the original gap in macrostructural components across the 18 months of the present study. These findings have important implications for clinical practice. Future intervention work should identify teaching strategies for supporting young boys to engage with and use language earlier; low language skills are related to a number of social and academic difficulties including struggling to develop and maintain positive relationships with peers (Cooper et al., 2002; Gertner et al., 1994) and later literacy/reading difficulties (Harris et al., 2011; Whorrall & Cabell, 2015). Clinicians should provide resources for children to develop these microstructural and macrostructural skills earlier on. For example, clinicians can incorporate the use of narrative in therapy as well as engage families in incorporating 37 narrative usage at home; exposing boys to more books is an approachable therapy for both clinicians and families. Clinicians should also recognize the needs of typically developing boys, so that they can provide assistance and guidance to increase these language skills. Since this study found that the boys catch up to girls on macrostructural language components but are still behind in microstructural components, clinicians can leverage these strengths to increase boys’ early weaknesses. Clinicians should engage boys in narrative strategies including promoting boys to talk about macrostructural components of language like characters, plot, etc., which will help them to use more language and prompt more use of microstructural language components. Clinicians should assess multiple components of language including microstructural and macrostructural components. Note that the findings of this study identify different growth for different components of language; thus, it is recommended that clinicians assess multiple components of language when working with young children. Narrative assessments, like the NAP-2 (Bowles et al., 2020), assess both microstructural and macrostructural components of language within the same assessment. This study helps recognize that both microstructural and macrostructural language components develop early in development. Using a narrative assessment, the NAP-2 (Bowles et al., 2020), this study found that examining both microstructural and macrostructural components of language offers a more complete understanding of young children’s language development than assessing one or the other alone or assessing language using decontextualized assessments. Conclusions 38 Specifically, girls’ and boys’ developmental trajectories differ for microstructural as compared to macrostructural components of language. Although boys demonstrate lower language skills initially in both areas, findings identified areas of strength and need for boys. Both researchers and practitioners can benefit from assessing children using a narrative task because it provides a multicomponent assessment of language examining multiple skills necessary for later reading success. Intervention work should include creating teaching strategies focused on supporting young boys to engage with and use language earlier, perhaps by leveraging their macrostructural skills to support microstructural language development. 39 APPENDICES 40 APPENDIX A. Family Questionnaire 41 Family Questionnaire (cont’d) 42 APPENDIX B. Tables Table 1. Narrative Assessment Protocol Codes NAP Codes Definition Code Does the child use a title before Title starting their story? Conventional Opening Does the child use an introduction to open their story? Ex. Once upon a time… Character Reference Does the child include specific character names in their story besides just using pronouns (i.e., he or she)? Ex. Mama Raccoon Sentence using temporal language Temporal Ordering (first, before, after, then, now) to indicate the ordering of events not including the use word ‘then’ Emotion Reference Does the child include any emotions? Ex. She was happy. Does the child exaggerate any Elongations words to add emphasis? Ex. She was sooooo happy. Similes and Metaphors Comparing one thing with another Ex. The color was as bright as the sun. Does the child make references to Time Reference time? Ex. One day… Ex. It was bedtime. Tier Two Verbs Uncommon verb that is longer than 7 letters Ex. The horse galloped. Tier Two Adverbs Uncommon adverb that is longer than 7 letters Ex. She carefully got on her bike. Tier Two Adjectives Uncommon adjective that is longer than 7 letters Ex. The beautiful day 43 Language Component Category Macrostructural Macrostructural Macrostructural Macrostructural Macrostructural Macrostructural Macrostructural Macrostructural Microstructural Microstructural Microstructural Microstructural Microstructural Microstructural Macrostructural Macrostructural Macrostructural Macrostructural Macrostructural Table 1. (cont’d) Sentences that start with a ‘wh’ Interrogative ‘wh’ word (who, what, when, where, Questions why, which, how) Sentence that contains more than Conjoined Adverbial one adverbial phrase Phrases Ex. He climbed slowly and carefully. Noun phrase containing two or Elaborated Noun more adjectives preceding a Phrases singular or plural noun Ex. The furry brown puppy Does the child state the goal they Sub-goal are trying to accomplish? Ex. She wants to share lemonade with her friends. Does the child state a problem? Sub-problem Ex. She can’t find any cups. Does the child state how to solve Sub-solution the stated problem? Ex. She goes inside to get cups. Does the child state that the goal Sub-resolution was accomplished? Ex. She came back with the cups. Conventional Ending Does the child include an ending to their story? Ex. The end 44 Table 2. Children’s Use of Microstructural Language Components Full Sample Timepoint Mean T1 1.26 1.35 T2 1.87 T3 T4 2.39 Table 3. SD 1.79 1.69 1.78 2.19 Range 0-9 0-8 0-8 0-9 Mean 1.34 1.45 2.03 2.68 Girls SD 1.73 1.63 1.86 2.28 Boys Range Mean 0-9 0-8 0-8 0-9 1.21 1.85 1.22 1.65 1.74 1.71 2.15 2.10 SD Range 0-7 0-7 0-7 0-9 Children’s Use of Macrostructural Language Components Full Sample SD Timepoint Mean 8.88 T1 5.18 10.54 5.17 T2 12.55 4.75 T3 T4 14.53 4.03 Range 0-26 0-20 0-21 0-23 Girls SD Range Mean Boys SD Range 0-23 0-19 0-20 0-23 8.12 5.41 0-26 9.71 5.55 0-20 0-21 12.10 5.02 5-23 14.20 4.33 Mean 9.79 4.78 11.48 4.51 13.02 4.36 14.92 3.64 45 Figure 1. APPENDIX C. Figures Microstructural Language Components Use By Sex Across Time T1 T2 T3 T4 Timepoint Re[lecting Six-Month Intervals 0 0.5 1 1.5 2 2.5 3 0 2 4 6 8 10 12 14 16 s t n e n o p m o C l a r u t c u r t s o r c i M f o # s t n e n o p m o C l a r u t c u r t s o r c a M f o # Figure 2. Macrostructural Language Components Use By Sex Across Time T1 T2 T3 T4 Timepoint Re[lecting Six-Month Intervals 46 Girls Boys Girls Boys REFERENCES 47 REFERENCES Artman, K., Meiler, C., Quesenberry, A., & Hemmeter, M. L. (n.d.) Recognizing and Supporting the Social and Emotional Health of Young Children Birth to Age Five [Online Tutorial]. Retrieved from https://www.ecmhc.org/tutorials/social- emotional/index.html Bachman, L. F. (2007). What is the construct? The dialectic of abilities and contexts in defining constructs in language assessment. In J. Fox, M. Wesche, D. Bayliss, L. Cheng, C. E. Turner, & C. Doe (Eds.), Language testing reconsidered (pp. 41-71). The University of Ottawa Press. Barbu, S., Nardy, A., Chevrot, J. P., Guellaï, B., Glas, L., Juhel, J., & Lemasson, A. (2015). Sex differences in language across early childhood: Family socioeconomic status does not impact boys and girls equally. Frontiers in Psychology, 6, 1874. doi: 10.3389/fpsyg.2015.01874 Barrs, M. (1993). Introduction: Reading the difference. In M. Barrs & S. Pidgeon (Eds.). Reading the difference. CLPE. Bloom, L., & Lahey, M. (1978). Language development and language disorders. Wiley. Bouchard, C., Trudeau, N., Sutton, A., Boudreault, M., & Deneault, J. (2009). Gender differences in language development in French Canadian children between 8 and 30 months of age. Applied Psycholinguistics, 30(4), 685-707. doi:http://dx.doi.org.proxy2.cl.msu.edu/10.1017/S0142716409990075 Bowles, R. P., Justice, L. M., Khan, K. S., Piasta, S. B., Skibbe, L. E., & Foster, T. D. (2020). Development of the Narrative Assessment Protocol-2: A tool for examining young children’s narrative skill. Language, Speech, and Hearing Services in Schools. Advance online publication. https://doi.org/10.1044/2019_LSHSS-19-00038 Brassard, M. R., & Boehm, A. E. (2007). Preschool assessment: Principles and practices. Guilford Press. Caldera, Y., Huston, A., & O‘Brian, M. (1989). Social interactions and play patterns of parents and toddlers with feminine, masculine and neutral toys. Child Development, 60, 70–76. doi: 10.2307/1131072 Campbell, T., Dollaghan, C., Needleman, H., & Janosky, J. (1997). Reducing bias in language assessment: Processing-Dependent measures. Journal of Speech, Language, and Hearing Research, 40, 519-525. 48 Women, men and language: A sociolinguistic account of gender differences in Coates, J. (1993). The acquisition of gender-differentiated language. In J. Coates (Ed.), language (2nd ed., pp. 143–167). Longman. Cooper, D. H., Roth, F. P., Speece, D. L., & Schatschneider, C. (2002). The contribution of oral language skills to the development of phonological awareness. Applied Psycholinguistics, 23, 399–416. Costanza-Smith, A. (2010). The clinical utility of language samples. Perspectives on Language Learning and Education, 17, 9-15. 10.1044/lle17.1.9. Crisp, T., Knezek, S. M., Quinn, M., Bingham, G. E., Girardeau, K., & Starks, F. (2016). What’s on our bookshelves? The diversity of children’s literature in early childhood classroom libraries. Journal of Children’s Literature, 42(2), 29-42. Elleseff, T. (2017, July 20). The importance of narrative assessments in speech language pathology. Smart Speech Therapy. https://www.smartspeechtherapy.com/revised- narrative-assessments-in-speech-language-pathology/ Engel, S. (1995). The stories children tell: Making sense of the narratives of childhood. WH Freeman. Eriksson, M. (2016). The Swedish Communicative Development Inventory III: Parent reports on language in preschool children. International Journal of Behavioral Development, 41(5), 647-654. Eriksson, M., Marschik, P., Tulviste, T., Almgren, M., Pérez-Pereira, M., Wehberg, S., Umek, L., Gayraud, F., Kovačević, M., & Gallego, C. (2012). Differences between girls and boys in emerging language skills: Evidence from 10 language communities. The British Journal of Developmental Psychology, 30, 326-43. 10.1111/j.2044- 835X.2011.02042.x. Fiorentino, L., & Howe, N. (2004). Language competence, narrative ability, and school readiness in low-income preschool children. Canadian Journal of Behavioural Science/Revue Canadienne des Sciences du Comportement, 36, 280-294. 10.1037/h0087237. Fivush, R. A., Brotman, M. P., Buckner, J., & Goodman, S. (2000). Gender differences in parent–child emotion narratives. Sex Roles, 42, 233-253. 10.1023/A: 1007091207068. Fowler, F. J. (2009). Applied social research methods: Survey research methods (4th ed.). SAGE Publications, Inc. 49 Garrard, K. R. (1991). A guide for assessing young children’s expressive language skills through language sampling. National Student Speech Language Hearing Association Journal, 18, 87-95. Gee, J. P. (2003). A sociocultural perspective on early literacy development. In S. B. Neuman & D. Dickinson (Eds.), Handbook of early literacy research (pp. 30-42). Guilford Press. Gertner, B. L., Rice, M. L., & Hadley, P. A. (1994). Influence of communicative competence on peer preferences in a preschool classroom. Journal of Speech and Hearing Research, 37, 913-923. Gleason, J. B., & Ely, R. (2002). Gender differences in language development. In A. McGillicuddy-De Lisi & R. De Lisi (Eds.), Biology, society, and behavior: The development of sex differences in cognition (pp. 127–154). Ablex Publishing. Gomez, R. L., & Gerken, L. (2000). Infant artificial language learning and language acquisition. Trends in Cognitive Science, 4(5), 178-186. Grolig, L., Cohrdes, C., Tiffin-Richards, S. P., & Schroeder, S. (2019). Effects of preschoolers’ storybook exposure and literacy environments on lower level and higher level language skills. Reading and Writing, 32, 1061-1084. https://doi.org/10.1007/s11145-018-9901-2 Hargrave, A., & Sénéchal, M. (2000). A book reading intervention with preschool children who have limited vocabularies: The benefits of regular reading and dialogic reading. Early Childhood Research Quarterly, 15, 75-90. 10.1016/S0885-2006(99)00038-1. Harris, J., Golinkoff, R. M., & Hirsh-Pasek, K. (2011). Lessons from the Crib for the Classroom: How Children Really Learn Vocabulary. In S. B. Neuman & D. K. Dickinson (Eds.), Handbook of early literacy research (Vol. 3, pp. 49-65). The Guilford Press. Harris, Y. R., & Schroeder, V. M. (2012). What the Berenstain Bears can tell us about school readiness: Maternal story grammar style and preschool narrative recall. Journal of Early Childhood Research, 10, 176–195. Hoff, E. (2008). Language experience and language milestones during early childhood. In K. McCartney, & D. Phillips (Eds.), Blackwell handbook of early childhood development (pp. 232-251). Wiley. Isbell, R., Sobol, J., Lindauer, L., & Lowrance, A. (2004). The effects of storytelling and story reading on the oral language complexity and story comprehension of young children. Early Childhood Education Journal, 32, 157-163. 10.1023/B: ECEJ.0000048967.94189.a3. 50 Justice, L. M., Bowles, R., Kaderavek, J., Ukrainetz, T., Eisenberg, S., & Gillam, R. (2006). The Index of Narrative Microstructure: A clinical tool for analyzing school-age children’s narrative performances. American Journal of Speech-Language Pathology / American Speech-Language-Hearing Association, 15, 177-91. 10.1044/1058-0360(2006/017). Justice, L. M., Bowles, R., Pence, K., & Gosse, C. (2010). A scalable tool for assessing children's language abilities within a narrative context: The NAP (Narrative Assessment Protocol). Early Childhood Research Quarterly, 25, 218-234. 10.1016/j.ecresq.2009.11.002. Justice, L.M., & Ezell, H.K. (2002). The syntax handbook: Everything you learned about syntax… but forgot. PRO-ED, Inc. Klee, T. & Stokes, S. F., (2011). Language development. In D. Skuse, H. Bruce, L. Dowdney, & D. Mrazek (Eds.), Child Psychology and Psychiatry (pp. 45-50). Wiley- Blackwell. Kuhl, P., Conboy B., Coffey-Corina S., Padden, D., Rivera-Gaxiola, M., & Nelson, T. (2008). Phonetic learning as a pathway to language: New data and native language magnet theory expanded (NLM-e). Philosophical Transactions of the Royal Society B, 363, 979–1000. Kuhl, P., Williams, K. A., Lacerda, F., Stevens, K. N., & Lindblom, B. (1992). Linguistic experience alters phonetic perception in infants by 6 months of age. Science, 255, 606-608. 10.1126/science.1736364. Ladd, G., & Burgess, K. B. (2001). Do relational risks and protective factors moderate the linkages between childhood aggression and early psychological and school adjustment? Child Development, 72, 1579-1601. Lindfors, J. (1987). Children’s Language and Learning. Prentice-Hall. Maccoby, E. E. (Ed.). (1966). The development of sex differences. Stanford University Press. McCabe, A., & Bliss, L. S. (2003). Patterns of narrative discourse. Allyn & Bacon. McCabe, A., & Rollins, P. R. (1994). Assessment of preschool narrative skills. American Journal of Speech-Language Pathology, 3(1), 45-56. 10.1044/1058-0360.0301.45. McGee, L. M., & Schickedanz, J. A. (2007). Repeated interactive read-alouds in preschool and kindergarten. The Reading Teacher, 60(8), 742-751. Miller, L., Perkins, K. A., Dai, Y. G., & Fein, D. (2017). Comparison of parent report and direct assessment of child skills in toddlers. Research in Autism Spectrum Disorders, 41-42, 57-65. 10.1016/j.rasd.2017.08.002. 51 Miller, S., & Pennycuff, L. (2008). The power of story: Using storytelling to improve literacy learning. Journal of Cross-Disciplinary Perspectives in Education, 1, 36-43. Miller, P., & Sperry, L. L. (1988). Early talk about the past: The origins of conversational stories of personal experience. Journal of Child Language, 15, 293- 315. 10.1017/S0305000900012381. Morgan, P. L., Farkas, G., Hillemeier, M. M., Scheffner Hammer, C., & Maczuga, S. (2015). 24- month-old children with larger oral vocabularies display greater academic and behavioural functioning at kindergarten entry. Child Development, 86, 1351-1370. Neuman, S. B. (2011). Lessons from the crib for the classroom: How children really learn vocabulary. In S. B., Neuman, & D. K. Dickinson (Eds.), Handbook of early literacy research (Vol. 3, pp. 358-372). The Guilford Press. Onis, M. (2007). Assessment of sex differences and heterogeneity in motor milestone attainment among populations in the WHO Multicentre Growth Reference Study. Acta Paediatrica, 95, 66-75. 10.1111/j.1651-2227.2006.tb02377.x. Pelatti, C. Y., Piasta, S. B., Justice, L. M., & O'Connell, A. (2014). Language- and literacy- learning opportunities in early childhood classrooms: Children's typical experiences and within-classroom variability. Early Childhood Research Quarterly, 29(4), 445– 456. https://doi.org/10.1016/j.ecresq.2014.05.004 Pentimonti, J. M., Zucker, T. A., & Justice, L. M. (2011). What are preschool teachers reading in their classrooms? Reading Psychology, 32(3), 197-236. DOI: 10.1080/02702711003604484 Pierce, H. K. (2009). Exploratory, functional, and symbolic play behaviors of toddlers with Autism Spectrum Disorders (FSU_migr_etd-0813)[Doctoral dissertation, Florida State University]. Proquest Dissertations Publishing. Pindzola, R. H., Plexico, L. W., Haynes, W. O., & Haynes, W. O. (2016). Diagnosis and evaluation in speech pathology (9th edition). Pearson. Roben, C. K., Cole, P. M., & Armstrong, L. M. (2013). Longitudinal relations among language skills, anger expression, and regulatory strategies in early childhood. Child Development, 84, 891-905. doi:10.1111/cdev.12027 Schick, A. & Melzi, G. (2010). The development of children's oral narratives across contexts. Early Education and Development, 21, 293-317. 10.1080/10409281003680578. Schneider, P., & Dubé, R. V. (2005). Story presentation effects on children's retell content. American Journal of Speech-Language Pathology, 14(1), 52-60. 52 Schuele, C. M., Skibbe, L., & Rao, P. (2007). Assessing phonological awareness. In K. Pence (Ed.), Assessment in emergent literacy (pp. 275-325). Plural Publishing. Sénéchal, M., Thomas, E., & Monker, J. A. (1995). Individual differences in 4-year-old children's acquisition of vocabulary during storybook reading. Journal of Educational Psychology, 87(2), 218–229. https://doi.org/10.1037/0022- 0663.87.2.218 Smith, L., Van Jaarsveld, C., Llewellyn, C., Fildes, A., Sanchez, G., Wardle, J., & Fisher, A. (2017). Genetic and environmental influences on developmental milestones and movement: Results from the Gemini cohort study. Research Quarterly for Exercise and Sport, 88, 401-407. 10.1080/02701367.2017.1373268 Southwood, F., & Russell, A. F. (2004). Comparison of conversation, freeplay, and story generation as methods of language sample elicitation. Journal of Speech, Language, and Hearing Research, 47, 366-376. 10.1044/1092-4388(2004/030). Sullivan, P. M., & Knutson, J. F. (2000). Maltreatment and disabilities: A population-based epidemiological study. Child Abuse & Neglect, 24, 1257-1273. https://doi.org/10.1016/S0145-2134(00)00190-3 Vallotton, C., & Ayoub, C. (2011). Use your words: The role of language in the development of toddlers' self-regulation. Early Childhood Research Quarterly, 26, 169-181. 10.1016/j.ecresq.2010.09.002. Vygotsky, L. S. (1978). Mind in Society. Harvard University Press. Wasik, B., & Bond, M. A. (2001). Beyond the pages of a book: Interactive book reading and language development in preschool classrooms. Journal of Educational Psychology, 93, 243-250. 10.1037/0022-0663.93.2.243. Westby, C. E. (2007). Child maltreatment: A global issue. Language, Speech, and Hearing Services in Schools, 38(2), 140-148. doi: 10.1044/0161-1461(2007/014). Whorrall, J. & Cabell, S. (2015). Supporting children’s oral language development in the preschool classroom. Early Childhood Education Journal, 44, 335-341. 10.1007/s10643-015-0719-0. Zucker, T. A., Cabell, S. Q., Justice, L. M., Pentimonti, J. M., & Kaderavek, J. N. (2013). The role of frequent, interactive prekindergarten shared reading in the longitudinal development of language and literacy skills. Developmental Psychology, 49(8), 1425- 1439. doi: 10.1037/a0030347 53