FAMILY INCOME, THE HOME ENVIRONMENT, SUSTAINED ATTENTION, GENETIC SUSCEPTIBILITY, AND CHILDREN'S READING OUTCOMES: A STRUCTURAL EQUATION MODELING ANALYSIS By June N. Westdal A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of School Psychology – Doctor of Philosophy 2018 FAMILY INCOME, THE HOME ENVIRONMENT, SUSTAINED ATTENTION, GENETIC SUSCEPTIBILITY, AND CHILDREN'S READING OUTCOMES: A STRUCTURAL EQUATION MODELING ANALYSIS ABSTRACT By June N. Westdal Seemingly small reading delays in early childhood have the potential to compound into larger reading difficulties later in childhood (Bast & Reitsma, 1998; Foster & Miller, 2007). Children growing up in at-risk households are especially vulnerable to falling behind in reading. The objective of this study was to explore the successive interactions and indirect effects of environmental and within child variables that influence reading outcomes for at-risk children. The work was informed by Bronfenbrenner’s bioecological model (Bronfenbrenner & Ceci, 1994), prevailing models of children’s reading development (National Reading Council, 2000), and mediational theories on the effects of poverty (Yeung, Linver, & Brooks-Gunn, 2002). This study examined the associations between income, the early home environment (home literacy and maternal depression), sustained attention, genetic susceptibility, and children's reading outcomes in kindergarten and third-grade. Data were drawn from a nationally representative dataset of at-risk families and children, the Fragile Families and Child Wellbeing Study (FFCWS). The primary analysis techniques were latent variable structural equation modeling (SEM) that examined the mediated and moderated pathways between environmental and within child variables. The final study sample consisted of approximately 2,062 children and their primary caregivers, mostly mothers. Several direct associations were significant. Results indicated that households with more income had children with better reading scores in kindergarten, but not in third grade. Children’s early sustained attention predicted their kindergarten and third-grade reading scores. Mothers’ endorsements of depression did not predict their children’s reading in kindergarten or third grade. Homes with more home literacy had children with higher reading scores in kindergarten, but the direct effects of the early home literacy environment did not persist until third grade. Analyses only supported the indirect path through the home literacy environment. More specifically, homes with more income had more enriched home literacy environments, and children exposed to better home literacy environments had better reading outcomes in kindergarten and third grade. Moderation analyses did not support the hypothesis that DRD4 long allele would differentiate the associations between income, home literacy environment, and children’s third-grade reading outcomes. Post- hoc analyses were conducted using two group SEM comparison testing. A unique and novel significant moderation effect was identified, where the DRD4 long allele moderated the direct association between the early home literacy environment and children’s kindergarten letter-word reading. The findings provide support for the importance of the home environment during the early developmental period and the genetic susceptibility of children with the DRD4 long allele during kindergarten. Copyright by JUNE N. WESTDAL 2018 ACKNOWLEDGEMENTS Thank you to my advisor and dissertation chair, Dr. Jodene Fine for advising me through research, coursework, and clinical work with wisdom, kindness, patience, and sincerity. Thank you for teaching me to think with brains and with heart. Thank you to my guidance committee, Drs. Kristin Rispoli, Anne Bogat, and Lori Skibbe, for your time, suggestions, questions, and feedback that pushed my research and critical thinking. Thank you to the Fragile Families Child and Wellbeing Study for granting me access to a compelling and meaningful dataset. Thank you to the College of Education at Michigan State University for generously supporting this dissertation project. Thank you to my graduate school friends, especially Marianne Clinton, Heather Schmitt, Danielle Balaghi, Allie Siroky, and “The Hateful Eight.” I could not imagine my life without your friendships, endless jokes, and unwaivering support. Thank you to my dad, for teaching me to do my best and to own my mistakes and learn from them. Thank you to Jason Fernandez, I could write another 150 page dissertation on how lucky I am to have you in my life. Thank you, everyone, I do not know how to express my gratitude enough. I’d like to dedicate this dissertation to my mom and my grandparents. Thank you for being my inspiration. v TABLE OF CONTENTS LIST OF TABLES .................................................................................................................. viii LIST OF FIGURES .................................................................................................................... x CHAPTER 1: INTRODUCTION ............................................................................................... 1 CHAPTER 2: LITERATURE REVIEW ..................................................................................... 8 Biology by Environment Frameworks ................................................................................... 8 The Importance of Reading ................................................................................................... 9 Reading Development ....................................................................................................... 11 Income ................................................................................................................................ 16 Poverty risks ............................................................................................................... 17 Mediation and poverty ................................................................................................. 21 Early Home Literacy Environment ..................................................................................... 24 Poverty and home literacy ........................................................................................... 27 Maternal Depression ........................................................................................................ 29 Maternal depression and literacy .................................................................................. 32 Sustained Attention ............................................................................................................ 34 Sustained attention and literacy ................................................................................... 36 Sustained Attention, Home Literacy, and Maternal Depression ......................................... 40 DRD4, Attention, and Literacy .......................................................................................... 47 Purpose of the Present Study .............................................................................................. 51 Research Questions, Hypotheses, Analyses, and Rationale ................................................. 55 CHAPTER 3: METHOD .......................................................................................................... 66 Study Model ....................................................................................................................... 66 Data Source ....................................................................................................................... 66 FFCWS design .............................................................................................................. 67 FFCWS sample demographics ...................................................................................... 69 Final study sample ....................................................................................................... 74 Sampling weights ......................................................................................................... 74 Variables ............................................................................................................................ 79 Reading achievement ................................................................................................... 80 Early home literacy environment ................................................................................. 81 Maternal depression ..................................................................................................... 84 DRD4 ......................................................................................................................... 86 Sustained attention ........................................................................................................ 87 Income ......................................................................................................................... 88 Covariates ..................................................................................................................... 89 Statistical Analyses .............................................................................................................. 92 Preliminary analyses ..................................................................................................... 92 Structural equation modeling ......................................................................................... 96 vi SEM estimation and model fit ............................................................................... 98 Path coefficients ................................................................................................... 99 Equivalent and near equivalent models ................................................................. 101 CHAPTER 4: RESULTS ........................................................................................................ 103 Preliminary Analyses ........................................................................................................ 103 Analyzing the Structural Model ........................................................................................ 104 Research Question One ..................................................................................................... 106 Testing the cross-sectional structural model. .............................................................. 107 Direct 1a. ................................................................................................................... 107 Indirect 1b .................................................................................................................. 108 Influence of covariates ............................................................................................... 108 Research Question Two ................................................................................................... 111 Testing the longitudinal structural model. .................................................................. 112 Direct 2a. ................................................................................................................... 112 Indirect 2b .................................................................................................................. 113 Influence of covariates ............................................................................................... 113 Research Question Three .................................................................................................. 116 Analysis plan ............................................................................................................. 117 Testing the structural model ...................................................................................... 119 Moderation................................................................................................................. 120 Post Hoc Analyses .......................................................................................................... 121 Alternative Models .......................................................................................................... 126 HLE .......................................................................................................................... 126 CHAPTER 5: DISCUSSION .................................................................................................. 130 Early Income Influences Reading in Kindergarten ........................................................... 131 Early Home Literacy Environment Matters for Kindergarten Reading ............................. 133 Effects of the Early Home Literacy Environment Do Not Directly Persist into Third Grade .............................................................................................................................. 135 Early Maternal Depression is not Associated with Reading ............................................. 137 Early Sustained Attention Influences Reading Directly ................................................... 138 DRD4 Moderation ........................................................................................................... 140 Differential susceptibility ......................................................................................... 143 RD and ADHD .......................................................................................................... 145 Conclusions and Clinical Implications ............................................................................. 147 Limitations and Future Directions ................................................................................... 150 APPENDIX ............................................................................................................................ 153 REFERENCES ....................................................................................................................... 155 vii LIST OF TABLES Table 1. Original FFCWS Data Sample ............................................................... 68 Table 2. Table 3. Table 4. Table 5. Table 6. Table 7. Table 8. Table 9. Table 10. Table 11. Table 12. Table 13. Table 14. Table 15. Table 16. Table 17. Table 18. Table 19. Table 20. Table 21. Comparison of Included and Excluded Samples (Categorical) ................ 72 Comparison of Included and Excluded Samples (Continuous) ................ 73 Comparison of HLE Indicators for Included and Excluded Samples ....... 73 Child Age in Final Study Sample ........................................................... 74 Comparison of Unweighted and Weighted Samples (Categorical) .......... 76 Comparison of Unweighted and Weighted Samples (Continuous) .......... 77 Unweighted and Weighted HLE Indicators (Categorical) ...................... 78 Overview of Study Variables and Sources .............................................. 79 Confirmatory Factor Analysis for HLE .................................................. 83 Frequency of CICI-SF Distribution for All Cases ................................... 85 Stability of Maternal Depression Wave 1- Wave 9 ................................. 86 DRD4 Allele Frequency ......................................................................... 87 Comparison by Income Status (Categorical) .......................................... 93 Comparison by Income Status (Continuous) .......................................... 94 Comparison of HLE Indicators by Income (Categorical) ....................... 94 Correlation Matrix for Final Sample ...................................................... 95 Final CFA for a Two-Factor Model ..................................................... 105 RQ1 Parameter Estimates for Direct and Indirect Effects .................... 110 RQ2 Parameter Estimates for Direct and Indirect Effects ..................... 115 AIC and BIC Comparisons for Moderation ......................................... 120 viii Table 22. Table 23. Parameter Estimates for Two Group Model ......................................... 125 CFA for Materials and Parenting Latent Variables ............................... 128 ix Figure 1. Figure 2. Figure 3. Figure 4. Figure 5. Figure 6. Figure 7. Figure 8. Figure 9. Figure 10. Figure 11. Figure 12. Figure 13. LIST OF FIGURES Cross-sectional Conceptual Model ......................................................... 58 Longitudinal Conceptual Model ............................................................. 63 Exploratory Conceptual Model............................................................... 65 Cross-sectional Conceptual Model ......................................................... 97 Longitudinal Conceptual Model. ............................................................ 97 Exploratory Conceptual Model............................................................... 97 Moderated-Mediation Conceptual Model ............................................ 101 Moderated-Mediation Statistical Model with Paths .............................. 101 Near Equivalent Conceptual Model ..................................................... 102 Cross-sectional Model with Unstandardized Estimates ......................... 111 Longitudinal Model with Unstandardized Estimates ............................ 116 Moderation Model with Unstandardized Estimates .............................. 121 Two Group Model with Unstandardized Estimates .............................. 126 x CHAPTER 1 INTRODUCTION Children’s reading skill are influenced by within child and environmental variables. A within child variable of growing interest is sustained attention. Sustained attention is known to contribute to reading outcomes in children, and is a well-established predictor of reading ability (Lam & Beale, 1991; Kibby, Lee, & Dyer, 2014; Stern & Shalev, 2013). Early home variables, such as the literacy environment and maternal depression, are becoming recognized as increasingly important to children’s early reading too. Though the extant literature has studied these variables individually, how environmental and within child variables interact to produce well-developed reading skills in children is not yet well understood. The goal of the present study is to better understand the associations between income, the home environment (home literacy and maternal depression), sustained attention, and children's reading outcomes in kindergarten (cross-sectional) and third grade (longitudinal). The study examined these variables using moderated-mediation models in a large, at-risk, low-income sample with structural equation modeling analyses. As an exploratory question, the study also investigated the role of a gene related to sustained attention (DRD4 VNTR) in these relations. The research was guided by the bioecological model (Bronfenbrenner & Ceci, 1994), which states that children's developmental outcomes are a result of the successive interactions between environmental and within child (cognitive and genetic) factors. Children’s reading development aligns with this perspective, as the National Reading Council (2000) states that the process by which children learn to read is multifaceted and involves multiple factors working in synchrony. Like the bioecological model, the prevailing theory of reading development postulates children’s reading ability is influenced by the interactions between environmental and 1 within child factors (Pennington, 2005). Accordingly, some young readers lag behind in reading because of genetic influences, others lag behind due to environmental circumstances, and many more experience delays resulting from interactions between genetics and environments. One environmental circumstance that significantly affects reading development is family income. Children from poor and near-poor households begin school with weaker emergent literacy and attention skills than children from adequate income homes, which puts them at a disadvantage for experiencing reading success in kindergarten and early elementary school (Duncan & Brooks-Gunn, 2000; Payne, Whitehurst, & Angell, 1994). The negative early effects of poverty have long reaching repercussions, as research demonstrates delayed reading in kindergarten contributes to the high incidence of poor academic achievement in elementary, middle, and high school among children from poor households (Duncan & Brooks-Gunn, 2000; Yoshikawa, Aber, & Beardsless, 2012). The influence of income on children’s reading outcomes has been widely examined; however, the mediational pathways (i.e., investment and family stress models) that justify how income indirectly influences children’s reading outcomes have inconsistently been applied in existing studies. Presently, few studies include both investment-related variables and family stress- related variables when investigating how income influences children’s reading outcomes. The current study addressed these gaps in the literature by using variables and methodology informed by both the investment and family stress models. The home literacy environment and maternal depression were identified as two variables that are consistent with the investment and family stress models, respectively. Caregivers and parents in the home environment are the first to contribute to children’s early literacy development by engaging in shared book reading and providing children access to literacy 2 materials, like books and toys. Showing literacy concepts to children before age five is linked to better performance in letter-word knowledge, decoding, vocabulary, reading fluency, and reading comprehension (Cunningham & Stanovich, 1998; Davidse et al., 2011; Senechal & LeFevre, 2002; Whitehurst & Lonigan, 1998). Additionally, existing theory identifies mothers’ depression as one of the pathways that influence children’s reading outcomes (Lovejoy, Graczyk, O’Hare, & Neuman, 2000; Wilson & Durbin, 2010). Mothers with depression have been shown to read to their children less, read for shorter periods of time, and ask fewer questions about books (Bigatti, Cronan, & Anaya, 2001; McLennan & Kotelchuck, 2000; Reissland, Sherpherd, & Herrera, 2003). These suboptimal reading practices are associated with poorer reading outcomes in children. Although both the early home literacy environment (investment path) and maternal depression (family stress path) have been shown to influence early reading outcomes separately, it is uncertain how the early home literacy environment and mother’s depression mediate the relations between income and children’s reading outcomes when considered in the same model. Along with these environmental factors, within child factors should be recognized. Studies investigating the comorbidity of reading disability (RD) and attention- deficit/hyperactivity disorder (ADHD) were the first to highlight the importance of attention (e.g., sustained attention, focused attention, shifting, etc.) to the reading process (Germanò, Gagliano, & Curatolo, 2010; Wilcutt et al., 2010). Since then, sustained attention has emerged as an important construct in reading (Fan, McCandliss, Sommer, Raz, & Posner, 2002; Prochonow, Tunmer, &Chapman, 2013; Rowe & Rowe, 1992; Stern & Shalev, 2013). Reading requires sustained attention and effort, especially for beginning readers. Therefore, it is not surprising that sustained attention predicts reading achievement consistently, such that children with better sustained attention have better reading scores, while children with poor sustained attention have 3 lower early reading scores (Rabiner, & Coie, 2000; Spira & Fischel, 2005; Walcott, Scheemaker, & Bielski, 2010). Despite numerous studies investigating the separate effects of sustained attention, home factors (home literacy environments and maternal depression), and low income on children’s reading outcomes, few studies have considered these constructs together. Davidse, de Jong, Bus, Huijiuts, and Swaab (2010), Haak, Downer, and Reeve (2012), the NICHD Early Child Care Research Network (2003), and Razza, Martin, and Brooks-Gunn (2012) have found preliminary evidence for the associations between sustained attention, the early home environment (e.g., variables related to early reading environments and maternal mental health), and reading outcomes. For instance, the NICHD, (2003) found that children’s sustained attention mediated the relation between the early home literacy environment and reading outcomes. Martin, Razza, & Brooks-Gunn (2012) also found significant mediating effects for sustained attention in a similarly designed study, but only for a near-poor group of children – not a poor group of children. Despite encouraging findings, these results are far from conclusive. Studies have found mediating effects (NICHD, 2003; Martin, Razza, & Brooks- Gunn, 2012), moderating effects (Haak, Downer, & Reeve, 2012), and null effects (Davidse, de Jong, Bus, Huijiuts, & Swaab, 2010) when examining the role of sustained attention in the relation between early home environments and language/literacy outcomes. The mechanisms by which sustained attention and the home environment relate to children’s reading outcomes are ambiguous, possibly because of a lack of longitudinal analyses, flaws with variable construction, or exclusion of investment- or family stress- related variables. The present study addressed these limitations and found results that provide more evidence to contribute to the growing literature on the associations between sustained attention, the home environment, and reading. 4 Finally, it is important to note that reading is culturally created. Reading is not “innate,” it is learned (Pennington, 2008). This means that to some degree, learning to read is not associated with specific genes that map onto reading ability (Peterson & McGrath, 2009). In other words, genes influence the development of the cognitive systems that support reading, such as attention, and asynchrony between cognitive systems can result in children’s difficulties in reading development. Therefore, the current study aims to incorporate a genetic variable into its analyses. Recently, there has been considerable interest in several dopamine genes related to attention, specifically the dopamine D4 receptor gene (DRD4). DRD4 affects dopamine production in the prefrontal cortex, an area of the brain that regulates attentional behaviors (Posner & Rothbart, 2007). The long variation (7-repeat allele) of DRD4 has been linked to lower dopamine receptor efficiency, which has been shown to result in lower levels of behavioral attention (Lasky-Su et al., 2008; Schmidt, Fox, Perez-Edgar, Hu, & Hamer, 2001; Tripp & Wickens, 2008). Since children must sustain attention to learn to read, carriers of long DRD4 may be at risk for developing delays in reading due to lower abilities to sustain attention; however, a precise picture has not emerged regarding whether the DRD4 gene represents a risk allele because some studies have found contradictory evidence (e.g., Gizer et al., 2009; Hsiung, Kaplan, Petryshen, Lu, & Field, 2004; Kegel & Bus, 2015). Thus, the current study aims to include a novel research question about the role of DRD4 in relation to income, sustained attention, the home environment, and reading outcomes. In summary, it is evident the early home environment and sustained attention matter in the development of reading. What is less evident is how much sustained attention relates to aspects of the early home environment (e.g., home literacy and maternal depression) to influence 5 reading outcomes for children in poverty. Previous studies have found varying mechanisms (mediation and moderation), have used incomplete variables, have not recognized mediational models of poverty, and have not examined longitudinal processes. The present study addressed these limitations by incorporating mediational poverty theory (the investment and family stress models) into the selection and construction of its variables, investigating cross sectional and longitudinal processes, and including an exploratory genetic variable. To examine how income, the early home literacy environment, maternal depression, and sustained attention influence children’s reading outcomes, the current study employed an existing dataset called the Fragile Families and Child Wellbeing Study (FFCWS; Reichman, Teitler, Garfinkel, & McLanahan, 2001), which draws from a population of at-risk families. Families were labeled “fragile” because of the social and economic risks associated with raising children in single-parent households (3/4 unmarried mothers, 1/4 married mothers). The Fragile Families and Child Wellbeing data set collected genetic data from children, which allowed the current study to explore a genetic component of attention. The use of this data set had the unique benefit of allowing the present study to test an exploratory hypothesis about the influences of DRD4 in the relations between income, sustained attention, early home literacy environment, maternal depression, and reading outcomes. It also allowed for the use of structural equation modeling (SEM) analyses due to the large sample size. Specifically, SEM was used to answer these three questions: 1. What are the associations between early income, early home literacy environments, early maternal depression, early sustained attention, and letter-word identification outcomes in kindergarten (cross-sectional model)? 6 2. What are the associations between early income, early home literacy environments, early maternal depression, early sustained attention, and passage comprehension outcomes in third grade (longitudinal model)? 3. Does DRD4 influence the associations between early income, early home literacy environments, early maternal depression, early sustained attention, and reading outcomes (exploratory model)? 7 CHAPTER 2 LITERATURE REVIEW The goal of the present study is to better understand the associations between income, the early home literacy environment, maternal depression, sustained attention, and children's reading outcomes in kindergarten and third grade. This chapter begins by providing an outline of the bioecological framework and a brief review of reading development. Next, the influences of income on children’s reading outcomes are examined, and the mediational pathways (i.e., investment and family stress models) that explain how income influences children’s reading outcomes are presented. Home literacy environments are reviewed in the context of the investment meditational model and maternal depression is reviewed in the context of the family stress mediational model. Next, the influence of children’s sustained attention on reading outcomes is discussed, and specific findings from studies examining sustained attention and early home literacy environments are reviewed. An exploratory use of a genetic variable linked to sustained attention is briefly examined. This chapter concludes with the purpose of the present study. Biology by Environment Frameworks Children’s development is influenced by interactions between their environments and their individual characteristics. Many researchers have proposed models that attempt to explain the distinct relations (Bronfenbrenner & Ceci, 1995; Bronfenbrenner, 1977; Engel, 1977; Hinde, 1992; Pennington, 2005; Sameroff, 2010; Rutter, 1987; Scarr, 1992; Zubin & Spring, 1977) and although there are differences in the hypothesized mechanisms between individual and environmental factors, the general notion that individual-by-environmental interactions influence human development is widely accepted. Given reading development is the accumulation of 8 environmental, cognitive, and biological factors, it is appropriate to use one of these models to investigate how environmental factors and children’s individual characteristics interact to support the development of reading. The present study is guided by Bronfenbrenner’s bio-ecological model (Bronfenbrenner & Ceci, 1994), which posits that children’s developmental outcomes are a result of the successive interactions between their environments and their personal characteristics. Specifically, Bronfenbrenner’s bio-ecological theory states that biology-by-environment interactions, called proximal processes, reflect the intimate interactions between the developing child and his/her immediate environments. Proximal processes vary systematically by intensity, form, and direction due to the characteristics of the developing child and the characteristics of the environment—both immediate and more remote—in which proximal processes take place. Many studies have investigated how children’s environments and individual characteristics interact to influence their reading outcomes (e.g., Afflerbach, Cho, Kim, Crassas, & Doyle, 2013; Dilworth-Bart, 2012; Hindman, Connor, Jewkes, & Morrison, 2008; Razza, Martin, & Brooks-Gunn, 2012). While various environmental and individual characteristics have been examined in relation to reading, there are four constructs of interest to the present study: income, early home literacy environments, maternal depression, and sustained attention. These constructs were investigated to uncover how they relate to cross-sectional and longitudinal reading outcomes. Since a developmental approach to reading is taken, a review of the importance of reading and the developmental progress of reading is necessary. The Importance of Reading Learning to read is one of the most critical milestones for children. Reading is a crucial skill for children, especially during the beginning of elementary school because it builds the 9 foundation for learning and academic success in middle and high school (Foulin, 2005; Sénéchal, LeFevre, Smith-Chant, & Colton, 2001; Whitehurst & Lonigan, 1998). Children who read early and well develop more efficient reading strategies, understand school content better, and show continued growth across other academic domains (Cunningham & Stanovich, 1997). Conversely, children who do not have well developed early reading skills miss opportunities to acquire reading strategies, misunderstand school content, and show less growth over academic domains. If reading difficulties are not addressed early, reading problems intensify and become more challenging to remediate (Vaughn et al., 2009; Torgesen 2004; Torgesen, 2002). Studies have repeatedly shown that students who are not at grade level upon completion of first grade have dramatically lower chances of being at or above grade level in reading later in school (Spira, Bracken, & Fischel, 2005; Wyner, Bridgeland, & Diiulio, 2007). Over time, poor readers fall further behind while average and skilled readers move further ahead, a phenomenon called the Mathew Effect (Stanovich, 1986). As students with reading difficulties progress through school, they are more likely to experience academic problems, dissatisfaction in school, poor approaches to learning, and low motivation in school (Cunningham & Stanovich, 1991). These difficulties extend beyond the academic domain; children who struggle with reading are also at a higher risk for behavioral and social-emotional problems, such as disruptive and defiant behavior, depression, anxiety, and poor social skills (Ackerman et al., 2008; Bierman et al., 2013; Morgan et al., 2008; Maughan, Rowe, Loeber, & Stouthamer-Loeber, 2003; Trzesniewski, Moffitt, Caspi, Taylor, & Maughan, 2006). Put together, reading is an important skill for children, and is associated with academic, social-emotional, and behavioral well-being. To better understand reading, it is important to review the process of reading acquisition. 10 Reading Development The goal of reading is to extract meaning from printed text. Reading allows children to learn academic content, refine and extend their verbal reasoning abilities, and share information with teachers and peers. It is a complex and multifaceted process, requiring the integration of multiple subskills (Pennington, 2009). Children begin developing the subskills needed to read before entering formal schooling. Beginning in infancy, children learn that language comprises patterns of sounds. Young children's abilities to perceive and produce speech are some of the earliest predictors of children's later reading abilities (Dickinson, Golinkoff, & Hirsh-Pasek, 2010; Scarborough, Neuman, & Dickinson, 2009; Tsao, Liu, & Kuhl, 2004). Once children have acquired language, emergent literacy skills become influential in the reading process (Lonigan, Burgess, & Antony, 2000; Pinto, Bigozzi, & Tarchi, 2016). Emergent literacy skills are children’s developing competencies in literacy before entering formal school and before being exposed to formal reading instruction (Whitehurst & Lonigan, 2001). Emergent literacy skills have been defined in many ways; most definitions include: children’s knowledge of letter sounds, language skills, phonological sensitivity, alphabetic naming, vocabulary, print awareness, and phonological awareness (Justice & Ezell, 2001; Sénéchal, LeFevre, Smith-Chant, & Colton, 2001). Researchers have categorized these skills in several manners. Justice & Ezell (2001) used print knowledge (e.g., print awareness, word awareness, and graphic awareness) and phonological awareness (i.e., the awareness of letter sounds) domains to categorize emergent literacy, while Whitehurst and Lonigan (1998) used the terms “inside out” to describe phonological awareness and letter knowledge skills and “outside- in” to describe language, conceptual knowledge, and story comprehension skills. 11 According to Whitehurst and Lonigan (1998) reading skills can be categorized as "outside-in" skills or "inside-out." Outside-in skills include: the language, vocabulary, content, and auditory comprehension skills that children need to aurally comprehend text read by parents. Outside-in skills ultimately lead to reading comprehension skills that allow children to draw meaning from text. Inside-out skills include: letter name and letter sound knowledge, phonological awareness, and decoding skills. Inside-out skills contribute to children’s direct ability to read. Combined, these skills support children’s ability to decode words fluently (inside- out) and comprehend (outside-in) the meaning of text. Numerous studies have supported and replicated the outside-in and inside-out framework (Lonigan, Schatschneider, Westberg, & the National Early Literacy Panel, 2009; Sénéchal & LeFevre, 2002; Storch & Whitehurst, 2002). Children with more emergent literacy skills at the beginning of kindergarten learn to read sooner, read better, and understand more reading instruction than children with less emergent literacy skills (Lonigan, Burgess, & Anthony, 2000; Storch & Whitehurst, 2002; Whitehurst & Lonigan, 2001). The strongest predictor of children’s end of kindergarten reading is children’s beginning of kindergarten emergent literacy skills (Burgess, 2001; Puranik, Lonigan, & Kim, 2011; Wapole, Chow, & Justice, 2004). In part, this is because reading performance is greatly influenced by the emergent literacy skills children learn at home before the beginning of formal schooling (Storch & Whitehurst, 2002). A common measure of emergent literacy is letter-word identification. Letter-word identification is the ability to identify letters of the alphabet and to read sight words and is typically measured by requiring children to name letters and read words aloud from a list. Letter- word reading is typically assessed in preschool and kindergarten, as it represents foundational reading skills necessary for later reading skills (e.g., decoding, fluency, comprehension) 12 (National Reading Panel, 2000). Letter-word reading requires the following emergent literacy skills: knowledge of letter sounds and names, alphabetic skills, and simple consonant-vowel- consonant (cvc) word decoding. In the last decade, researchers have consistently found letter knowledge to be a significant preschool predictor of learning to read, sometimes even the most influential predictor (Catts, Fey, Zhang, & Tomblin, 2001; Connet et al., 2000; Foulin, 2005; Piasta & Wagner, 2010). Emergent literacy prepares children for learning how to decode. Decoding is the next subskill in the development of reading. Decoding is the process of translating print into the phonetic code of language. Readers learn how to split words into letters or letter groups and match the letters to their respective sounds according to language rules; then they put the letters or letter groups back together to read the word as a whole (National Reading Panel, 2000). Skilled readers decode when they encounter less familiar words, difficult words, or new words (Otaiba, Kosanovich, Torgesen, Kamhi & Catts, 2012). Decoding measures examine readers’ abilities to connect visual text with associated letter sounds. This is accomplished through asking readers to read unfamiliar words, such as “frip,” aloud. Skilled decoding is the first step towards fluency. The National Reading Panel (2000) describes fluency as the ability to read quickly, accurately, and with proper expression. Fluency is achieved when readers no longer have to decode each word and instead can dedicate their cognitive resources to reading with speed and ease (National Reading Panel, 2000). Fluency measures examine readers’ speed and accuracy while reading. This is accomplished through asking readers to read a passage quickly and accurately. Though important, skilled fluency does not indicate children’s competence because fluency can occur without understanding what is read (Pikulski & Chard, 2005; Farstrup & Samuels, 2002). The first three years of formal education are dedicated to teaching young 13 children how to read (i.e., decoding and fluency). After third grade, inferring information from text is one of the primary avenues of classroom learning (Chall & Jacobs, 2003; Fountas, & Pinnell, 1996). Comprehension is the goal of reading. It is the ability to understand, maintain, and think about the content of the text (National Reading Panel, 2000). To successfully comprehend text, the National Early Literacy Panel (2009) identified pre-requisite skills that would support the development of comprehension, these are: alphabet knowledge, phonological awareness, rapid automatic naming of letters or digits, rapid automatic naming of objects or colors, writing or writing name, and phonological memory. Along with these skills, young readers must also be proficient in several cognitive skills (e.g., language, attention, memory) before this can be achieved (Peterson & McGrath, 2009). Comprehension is measured by asking readers to read a passage, look at pictures, and respond to literal or inferential questions about the passage. Reading comprehension is the primary mode of learning in late elementary, middle, and high school. Typically reading comprehension is thought to be achieved in third grade, as this is when school begin to move away from “learning to read” and towards “reading to learn,” (Chall, 1983). At the secondary level, under current Common Core standards, students are required to read increasingly difficult texts to build domain-specific knowledge and become content area literate (National Governors Association Center for Best Practices & Council of Chief State School Officers, 2010). Taken together, emergent literacy skills are an essential foundation for successful decoding, fluency, and comprehension to develop in a hierarchical progression (Pikulski & Chard, 2005; National Reading Panel, 2000). Children generally must have emergent literacy skills to have the “building blocks” to be competent with decoding. Skilled decoding leads to 14 fluent reading, (Share & Stanovich, 1995) and fluent reading allows children to comprehend text (Chall, 1983; National Reading Panel, 2000). Due to the hierarchical relations between reading skills, emergent literacy, decoding, fluency, and comprehension are commonly assessed in order determine a child’s level of proficiency with reading (Fuchs, Fuchs, Hosp, & Jenkins, 2001; Hudson, Lane, & Pullen, 2005; Pikulski & Chard, 2005). As such, the present study examined children’s reading at kindergarten and third grade. Kindergarten is approximately the time children become proficient in letter-word identification. To be successful readers, children should be able to name letters and their corresponding sounds and decode simple cvc words in kindergarten; thus, examining these skills with a letter-word identification measure. By third grade, children should have mastered decoding and should be reading fluently. They should use their mental resources towards comprehending the meaning of text; therefore, third grade is an appropriate time point to examine children’s proficiency in reading comprehension. If these reading skills are not mastered at the appropriate grades, delays in reading compound into larger problems in later elementary school and secondary school (Bast & Reitsma, 1998; Foster & Miller, 2007). While much is still unknown about the exact mechanisms by which children become proficient in these four reading skills, it is clear these skills do not develop in isolation. The environment supports, or hinders, the development of these skills. Learning to read successfully is the accumulation and integration of children’s environmental and individual factors (Olson et al., 2011; Pennington, 2005; Rayner, Pollatsek, Ashby, & Clifton, 2012; Veroeven, Reitsma, & Siegal, 2011). The following section discusses an environmental factor that greatly influences children’s reading, income. 15 Although all children should begin school with the expectation to learn to read successfully, many children from poor and near-poor households are at an increased risk for developing reading problems (Lee & Burkam, 2002). Children from poor and near-poor families disproportionately experience reading difficulties and low educational attainment, which can result in negative outcomes that extend past formal education and affect job attainment, overall well-being, and job salary (Arnbak, 2004; Brooks-Gunn & Duncan, 1997; Gibb, Fergusson, & Horwood, 2012; Duncan, Yeung, & Brooks-Gunn, 1998; McLoyd, 1998; Welsh, Nix, Blair, Bierman, 2010). Furthermore, poverty and low-income are associated with physical, cognitive, socio-emotional, and behavioral risks that can negatively influence or interfere with reading achievement. Income The official definition the U.S. federal government uses to determine eligibility for federal assistance programs is, “living in a household with a gross income under the official poverty line.” In the context of the current study in 2001, the poverty line was $8,590 for a family of one, $11,610 for a family of two, $14,630 for a family of three, and $17, 650 for a family of four (U.S. Department of Health and Human Services). In the United States, approximately 25% of children come from households that are below the poverty level and are classified as in “poverty,” or in “poor” households, and approximately 20% of children are living in “low-income,” or “near poor,” households that have family income that is at or slightly above the poverty level (Cauthen & Fass, 2008; Duncan & Brooks-Gunn, 2000; National Center for Children in Poverty, 2010)1. Currently, children from poor or near poor households experience 1 It is important to note that the definitions for “low-income” and “poor” and “poverty” are subject to debate. In this study, the terms are used per the definitions provided; however, some researchers distinguish that “poverty” represents the extent to which families do without 16 many more barriers to social, emotional, physical, and academic well-being than their peers from more affluent households (Aber, Bennett, Conley, Li, 1997; Duncan & Magnuson, 2013). Poverty risks. Physically, there are numerous health issues associated with poverty. Children who are financially disadvantaged are more likely to be born underweight (below 2,500 grams), which is associated with a variety of physical, cognitive, learning, and social-emotional problems (Aarnoudse-Moens, Weisglas-Kuperus, van Goudoever, & Oosterlaan, 2009; Bhutta, Cleves, Casey, Cradock, & Anard, 2002; Healy et al., 2013). Nutritional deficits are more frequent in poor households. Lower caloric intake, protein-energy insufficiency, and vitamin and mineral deficiencies are observed more frequently among poor children and negatively affect children’s physical and cognitive growth and development (Babu, Gajanan, &Sanyal, 2014; Black et al., 2013; Paciorek, Stevens, Finucane, Ezzati, & Nutrition Impact Model Study Group, 2013). Child Indicators Research Group (1997) described that short height for weight (i.e., stunting), a measure of nutrition, is more prevalent among children from poor and near-poor households than children from households with adequate income, especially when children’s families are poor across multiple years. Residence in poor neighborhoods increases the likelihood of exposure to dangerous toxins, like lead. Lead exposure is linked to kidney toxicity, metabolism damage, and cognitive impairments (American Academic of Pediatrics Committee on Environmental Health, 2005; Lanphear et al., 2005; Koller, Brown, Spurgeon, & Levy, 2005; Schwartz, 1994). Altogether, these negative health outcomes associated with poverty impair resources (e.g., financial, emotional, mental, spiritual, support systems, relationships, role models, and knowledge of hidden rules), while “low-income” represents finances alone (Lacour & Tissington, 2011). While the terminology differs, it is agreed up that poverty affects children’s reading achievement (and other outcomes) through other mediating factors. 17 children’s physical and cognitive development (Lipina & Colombo, 2009; Duncan & Brooks- Gunn, 2000). Children in less advantaged households also have a higher risk of developing social and emotional problems. This includes internalizing problems like depression, low self-esteem, and anxiety (Slopen et al., 2010; Dearing, McCartnery, & Taylor, 2006; Lupien, King, Meaney, & McEwen, 2001), and externalizing problems, such as aggression, opposition, and conduct problems (Ackerman, Brown, & Izard, 2004; McLeod & Nonnemaker, 2000; Pachter, Auinger, Palmer, & Weitzman, 2006). Studies have also found that children from low-income households demonstrate poorer peer relationships, lower popularity, and more disruptive classroom behaviors than their economically advantaged counterparts (Bolger, Patterson, Thompson, & Kupersmidt, 1995; Connell & Prinz, 2002; Fantuzzo, Bulosky-Shearer, McDermott, & McWayne, 2007; Raver, Blackburn, Bancroft, & Torp, 1999; Ridge, 2004). In part, this may be because of poverty-related stressors and fewer opportunities to engage with peers (Duncan & Brooks-Gunn, 1997; Fantuzzo et al., 2007). For example, income influences parents’ selection of early childcare for their children. Children whose parents can afford better quality early childcare (e.g., more structured activities, number of hours, more space, better teacher-child interactions) have better social and emotional outcomes than children whose parents could not afford quality early childcare (Dearling, McCartney, & Taylor, 2009; Mistry, Biesanz, Taylor, Burchinal, & Cox, 2004; NICHD Early Child Care Research, 2002). This is thought to occur because better early child care teaches children foundational social, emotional, and academic skills that promote success (Bierman, Torres, Domitrovich, Welsh, & Gest, 2009; Peisner-Feinberg et al., 2001). Lastly, poor children disproportionately experience learning difficulties and low educational attainment (Fergusson, Horwood, & Boden, 2008; Duncan & Brooks-Gunn, 1997; 18 Kierman & Mensah, 2011; Welsh, Nix, Blair, Bierman, 2010). Upon school entry, there is already a difference in achievement between children from poor and near-poor households and children in financially stable homes (National Assessment of Educational Progress, 2004; Burkam, 2013). Rowan and colleagues (2004) found that low-income kindergarten students scored at the 30th percentile, middle-income students scored at the 45th percentile, and high- income students scored at the 70th percentile on reading assessments. Achievement differences at school entry widen throughout elementary, middle, and high school (Fergusson, Horwood, & Boden, 2008; Harkins & Rouse, 2005; Lee & Burkam, 2002). Compared to children in more affluent families, children in poverty have lower graduation rates because they are more likely to drop out due to poor educational performance, pregnancy, or financial reasons (Bailey, Jenkins, & Leinbach, 2005; Jimerson, Egeland, Sroufe, & Carlson, 2000; Murnane, 2013; Thayer, 2000). The differences in achievement may be exacerbated by less developed cognitive and language skills (Bowey, 1995; Farah et al., 2006; Raz & Bryant, 1990; Whitehurst & Lonigan, 1998) likely due to less cognitive stimulation, less literacy exposure, and less language in the home environment (Brooks-Gunn, Klebanov, & Liaw, 1995; Sarsour et al., 2011). Achievement gaps might also be explained by less parental awareness about the importance of early reading exposure in the home. Kuo, Franke, Regalado, and Halfon (2004) found that 37% of low-income parents of young children stated their child’s pediatric health care provider had not discussed the importance of reading. Forty-seven percent of these parents indicated that they would have found information about the importance and benefits of early reading valuable. Negative academic outcomes are more severe if households are experiencing poverty while children are in early childhood (i.e., 0 – 6 years of age) (Dickerson & Popli, 2016; Kiernan & Mensah, 2009; Raver, Blair, & Willoughby, 2012). Studies examining the effect of chronic 19 poverty (i.e., poverty lasting more than two years) and incremental poverty (i.e., poverty during certain stages of children’s development) found that poverty was more detrimental during early childhood (Duncan & Brooks-Gunn, 1997; Hulme, Najman, Hayatbakhsh, Heron, Bor, O’Callaghan, Williams, 2009). Children who experienced the effects of poverty in early childhood had lower rates of high school completion than did children who experienced the effects of poverty during middle childhood (Hulme, Moore, & Shepherd, 2001). Interestingly, neuroimaging studies have found physiological evidence for the adverse cognitive and academic outcomes described (Hackman, Farah, & Meaney, 2010; Noble et al., 2015; Noble, Houston, Sroufe, & Carlson, 2005). Generally speaking, research utilizing Magnetic Resonance Imaging (MRI) has shown that income is correlated with children’s brains’ cortical thickness and surface area activation (Noble et al., 2015). Imaging results evidenced that income is associated with the size of brain structures that manage memory, cognitive control, attention, and language (Farah et al., 2006; Mezzacappa, 2004; Noble, McCandliss, & Farah, 2007; Noble et al., 2015), but not with areas of the brain that control visual and spatial processing (Farah et al., 2006; Noble, McCandliss, & Farah, 2007). Since these studies were correlational, the causal forces behind these associations are not clear, especially since there was high variability between brain structure among children in all income groups. These findings do not suggest that income directly leads to an unchallengeable trajectory, rather, the findings support the idea that income is associated with mediators that, in turn, negatively affect the brains of the most financially disadvantaged children (Noble, McCandliss, & Farah, 2007; Noble et al., 2015). Overall, poverty is associated with negative physical, social-emotional, and academic outcomes for children. Differences between children from high and low income families have 20 even been observed in brain structures. In relation to reading outcomes specifically, children who experience the effects of poverty in early childhood tend to have worse reading scores than children who experience the effects of poverty later in development. Duncan, Yeung, Brooks- Gunn, and Smith (1998) hypothesized that low income early in childhood matters more for achievement outcomes because of the importance early literacy skills (e.g., letter naming and letter sounds) in determining the course of schooling for children. Since poverty has a strong association with low emergent literacy skills, and preschool ability is highly predictive of children’s later academic skills, poor and near-poor children are at a disadvantage. The sum of this evidence suggests that income is associated with factors that negatively influence children. Two models attempt to explain the pathways by which income affects early childhood outcomes: the investment and family stress models. Although the investment and family stress models have been used to explain academic, cognitive, and behavioral outcomes, the following section explains the theories in relation to reading outcomes specifically. Mediation and poverty. Income does not directly affect children’s reading outcomes, rather, income influences children's development and outcomes through meditational pathways. The investment model and family stress models are two meditational paths that explain how poverty influences children’s outcomes. First, the parent investment model suggests that the effect of family income on children’s reading outcomes is apparent in parents’ decisions about how to allocate their money, time, energy, and support (Conger & Conger, 2008; Haveman & Wolfe, 1994; Yueng, Linver, & Brooks-Gunn, 2002). Monetary and time investments are considered together according to the investment model. The amount of money parents spend on materials to help children learn to read (e.g., purchasing books, quality early child care), and the time parents spend with children in joint literacy activities (e.g., reading together or visiting the 21 library) are considered “investments” that have the potential to enhance children's reading and language skills. Despite valuing books and reading programs and valuing time spent reading with their children, poor and near-poor families may not have the disposable income to afford book or the time to read with their children (DeBaryshe & Binder, 1994). Second, the family stress model suggests that the quality and amount of stimulating activities that children experience are influenced by parental mental health. Since there is a paucity of evidence on paternal mental health, most of the theory is based on maternal mental health. Poverty has been shown to cause material hardship, which, in turn, triggers maternal stress, relational conflict, and maternal depression (Du Rocher Schudlich & Cummings, 2003; McLeod & Kessler, 1990; Wadsworth, Raviv, Compas, & Connor-Smith, 2005). Mothers experiencing relational conflict or who are depressed are more likely to withdraw from their children or to become hostile toward them (Buehler & Gerard, 2002; Krishnakumar & Buehler, 2000; Pinderhughes, Dodge, Bates, Pettit, & Zelli, 2000). Additionally, depressed mothers in poverty have sporadic parent-child learning experiences (e.g., shared reading time), harsh discipline style, and low warmth (Gou & Harris, 2000; Bigatti, Cronan, & Anaya, 2001). These poor parenting practices are associated with poor reading outcomes in children (Downey & Coyne, 1990). Integrating these two lines of research, the investment model suggests that poverty restricts parents' abilities to invest money and time into stimulating materials and experiences for their children, thus, negatively affecting their children’s reading development (Conger, 2005; Guo & Harris, 2000; Linver, Brooks-Gunn, & Kohen, 2002; Yeung, Linver, & Brooks-Gunn, 2002). The family stress model (Conger & Elder, 1994) hypothesizes that material hardship puts strain on parents’ relationships with partners and decreases their mental health, which negatively 22 impacts parenting behaviors. Children have fewer occasions to develop emergent literacy skills when parents provide them with insufficient opportunities to be exposed to books, listen to stories, and practice reading (Bus, Van Ijzendoorn, & Pellegrini, 1995; Senechal & LeFevre, 2002; Senechal, 2009). The investment and family stress models have been tested using large-scale studies such as the National Longitudinal Survey of Youth (NLSY) (n=12,686) (Guo & Harris, 2000), the Infant Health and Development Program (IHDP) (n=493) (Linver, Brooks-Gunn, & Kohen, 2002), the Panel Study of Income Dynamics Study (n=868) (Yeung, Linver, & Brooks-Gunn, 2002), and the Early Childhood Longitudinal Study, Kindergarten Class of 1998– 1999 (n= 21,255) (Gershoff, Aber, Raver, & Lennon, 2007) with success. These studies show that investment and family stress paths have been repeatedly and successfully identified through structural equation modeling (SEM) with similar results (Gershoff, Aber, Raver, & Lennon, 2007; Guo & Harris, 2000; Linver, Brooks-Gunn, & Kohen, 2002; Yeung, Linver, & Brooks- Gunn, 2002). Despite these consistent findings, many of the studies use broad definitions of investment and family stress. Furthermore, no study has examined the family stress and investment path solely in relation to children’s reading outcomes. Informed by the investment and family stress pathways, the current study used an early home literacy environment variable and a maternal depression variable, respectively. Along with supporting with these theoretical models, the early home literacy environment and maternal depression have each been shown to be influential factors in children’s development. A home literacy environment measure is appropriate given that many of the questions used to measure home literacy environments align with questions used to measure the investment path (e.g., how many books in the home, how frequently does the parent read with the child). A maternal 23 depression measure is appropriate since the family stress pathway posits that material hardship negatively impacts mental health, particularly depressive symptoms, and, in turn, influences parenting behavior (Conger & Elder, 1994; Gershoff, Aber, Raver, & Lennon, 2007). Although including a parenting behavior variable along with the maternal depression variable would be ideal, the FFCWS did not include a comprehensive measure of parenting. Thus, even though maternal depression is not precisely aligned with family stress pathway, maternal depression is an important component of the family stress mediational pathway. Moreover, these is no scarcity of research on how maternal depression influences children’s reading outcomes, and there is substantial evidence for the influence of maternal depression on children’s early literacy skills (e.g., Baker & Iruka, 2013; Barbarin et al., 2006; Bigatti, Cronan, & Anaya, 2001; Foster, Lambert, Abbot-Shim, McCarty & Franze, 2005; Greenberg, et al., 1999; Reissland, Shepherd, & Herrera, 2003). Thus, both measures are instrumental components of the early environment that support children’s emergent literacy development. The following sections will review the influence of the early home literacy environment and maternal depression on children’s reading outcomes. Early Home Literacy Environment The home literacy environment is broadly defined as the quality and quantity of literacy- related stimulation, supports, and interactions available to children in the home setting. “Home literacy environment” has also been referred to as, “early literacy environment,” “early literacy exposure,” and “home literacy practices,” in the extant literature. The home is where children first encounter language and literacy, participate in literacy activities with parents, and observe parents engage in literacy-related activities before formal education in preschool or kindergarten. Several decades of research have found that exposure to literacy in the home environment is vital 24 to children’s development of emergent literacy skills, including: language, auditory comprehension, letter naming, letter sounds, phonological awareness, and decoding (Anglum, Bell, & Roubinek, 1990; Burgess, 2011; Christian, Morrison, & Bryant, 1998; Foy & Mann, 2003; Frijters, Barron, & Brunello, 2000; Foster, Lambert, Abbot-Shim, McCarty, & Franze, 2005; Griffin & Morrison, 1997; Hood, Conlon, & Andrews, 2008; Payne, Whitehurst, & Angell, 1994; Senechal & LeFevre, 2002; Senechal, LeFevre, Thomas, & Daley, 1998). Collectively, these findings demonstrate when children have more access to literacy materials and reading practices (e.g., shared book reading, variety of books, print exposure, parental modeling) they have better emergent reading skills, more interest in books and literacy, and more language skills. The extant data also shows that emergent literacy begins developing in infancy and continues developing throughout early childhood. The importance of the home literacy environment is widely accepted. Exposing young children to more literacy concepts in the home has been linked to better performance in their letter-word knowledge, decoding, vocabulary, fluency, and comprehension in elementary school (Cunningham & Stanovich, 1998; Davidse et al., 2011; Hood, Conlon, & Andrews, 2008; Niklas & Schneider, 2013; Senechal & LeFevre, 2002; Whitehurst & Lonigan, 1998). Moreover, the benefits of early home literacy environments extend beyond reading benefits in early elementary school (Burgess, Hecht, & Lonigan, 2002; Hood, Conlon, & Andrews, 2008; Lonigan, Burgess, & Anthony, 2000; Piasta, Justice, McGinty, & Kaderavek, 2012; Senechal & LeFevre, 2002); good home literacy environments have been associated with better reading achievement in late elementary and middle school (Cunningham & Stanovich, 1993; Hart et al., 2009; Froiland, Peterson & Davidson, 2013). Despite robust findings supporting the importance of the home literacy environment, there is no widely accepted definition of the home literacy environment. 25 There are various methods of measuring the home literacy environment (Phillips & Lonigan, 2009). Most research on the home literacy environment has focused on determining the frequency of shared parent-child reading using parent questionnaires (de Jong & Leseman, 2001; Foy & Mann, 2003; Frijters, Barron, & Brunello, 2000; Griffin & Morrison, 1997). This approach has consistently shown shared parent-child reading encounters are correlated with growth in children's oral language. Studies that use wider definitions of the home literacy environment, such as frequency reading in the home, the number of books in the home, the types of books in the home, parental encouragement of reading, parental instruction, and the age- appropriateness of the books in the home, find that the home literacy environment is associated with children’s phonological skills, print knowledge, and letter-naming and letter-sound knowledge (Evans, Shaw, & Bell, 2000; Levy, Gong, Hessels, Evans, & Jared, 2006; Senechal & LeFevre, 2002). The Home Literacy Environment Scale (Griffin & Morrison, 1997), Stony Brook Family Reading Survey (Whitehurst, 1992), the Parent as Reader Scale (PARS; DeBaryshe & Binder, 1994), and the Parent Reading Belief Inventory (PRBI; DeBaryshe & Binder, 1994) are commonly used short questionnaires that attempt to measure the home literacy environment with a more diverse scope. The existing data set, the FFCWS, did not ask questions that measure literacy teaching practices explicitly. The current early home literacy environment measure includes questions that match traditional measures of the home literacy environment, and tap the “resources” expended by parents. In other words, the measure of the home literacy environments used in the current study measures the frequency of literacy interactions and the types and number of books in the home. 26 Poverty and home literacy. Children from low-income households have lower emergent literacy skills than their peers in adequate income households (Whitehurst & Lonigan, 2001; Duncan & Brooks-Gunn, 1997; Zill & Resnick, 2006). Evidence suggests that this is partly due to poor home literacy environments. Studies that have investigated the association between income and home literacy environments found significant relations, even when controlling for parental educational attainment and careers. Bracken & Fischel (2008) found that home literacy environments (specifically parent- child reading) predicted children’s emergent literacy outcomes among preschool children from low-income backgrounds (mean income $23,132) attending Head Start. The home literacy environment across three dimensions was measured (child reading, parent reading, and parent- child reading) using the Stony Brook Family Reading Survey (Whitehurst, 1993). Activities such as the frequency with which the parent read to the child, the duration of reading sessions with the child, the frequency of visits to the library, and the number of books in the home appropriate for the child’s age predicted children’s reading outcomes more than children’s reading characteristics (e.g., frequency children asked to be read to) or parent characteristics (e.g., how often parents read alone or how much parents enjoyed reading). These findings have been replicated by other studies investigating the home literacy environment among low-income families (e.g., Raikes et al., 2006; Griffin & Morrison, 1997; Storch & Whitehurst, 2001; Britto & Brooks-Gunn, 2001). Rodriguez and colleagues (2009) found that the frequency of children's participation in literacy activities, the quality of mothers' engagements while reading, and the provision of age- appropriate learning materials each uniquely contributed to low-income children’s language and reading skills, beyond individual child characteristics (e.g., temperament, cognitive 27 ability). Britto and Brooks-Gunn (2001) found that mothers' expressive language while reading to their children was associated with their children’s expressive language. Results from the study also demonstrated that mothers’ warmth best predicted children's reading achievement scores. Of the home literacy predictors, academic stimulation (the number of books in the home) was least predictive of children's reading achievement scores among low-income African American preschoolers. Britto and Brooks-Gunn (2001) and Rodriguez and colleagues (2009) found preliminary support for including parental attributes to the conceptualization of home literacy environments when studying reading in low-income samples. These studies have contributed immensely to understanding how the home literacy environment influences children’s emergent reading skills in low-income households; however, the studies have not fully accounted for the important variables influenced by poverty. For example, Storch and Whitehurst (2001), Bracken and Fischel (2008), and Rodriguez and colleagues (2009) accounted for investment-related variables by including questions about the frequency of time spent book reading and the number of books in the home; however, the studies did not explicitly measure or control for the family stress factors (e.g., maternal depression or stress). This is important because theories of poverty indicate that parental stress and depression should be accounted for when studying low-income families. Thus, an influential factor that contributes to the home literacy environment for low-income samples is missing because a parental stress/depression component was not included. The current study aims to address this gap in the literature by recognizing maternal mental health in its model. This component is vital to understanding how poverty influences children’s reading outcomes. Despite mothers’ efforts to provide their children with literacy experiences in the home environment, individual differences in mothers’ mental health may 28 account for discrepancies in the literacy opportunities (e.g., shared book reading) that children are provided. Mothers’ mental health factors, such as depression, may influence children’s emergent literacy skills if it hinders positive literacy-related interactions between parent and child in the home environment. Accordingly, it is important to consider the effects of maternal depression on children’s reading outcomes. Maternal Depression Depression is a growing concern for parents of young children, especially for mothers (Davé, Petersen, Sherr, & Nazareth, 2010; Hasin, Goodwin, Stinson, & Grant, 2005). The DSM- 5 defines depression in adults as episodic or persistent symptoms of depressed mood designated by feelings or observations of sadness, emptiness, and hopelessness; loss of interest or pleasure in daily activities; adverse changes in sleeping and eating patterns; fatigue; feelings of worthlessness or excessive guilt; difficulty thinking or concentrating; and/or suicidal ideation or attempts (APA, 2013). Experiencing these symptoms impedes a person’s ability to function in daily life. About 6-17% of women experience at least one instance of depression either during pregnancy, the postpartum period, or the childbearing years (Evans et al., 2001; O’Hara & Swain, 1996; Burke, Burke, Rae, & Reiger, 1991; Somerset et al., 2006). Many factors have been implicated as putting women at risk for depression, including complicated medical conditions, previous mental health diagnoses, stressful life events, marital discord, lack of social support (APA, 2000; Dobson & Dozois, 2008; Witt et al., 2011). There has been a growing discussion regarding the disproportionate impact of maternal depression on women in poverty, as poverty is a significant risk factor for depression. Rates for maternal depression are twice as high for women in poverty than for women with adequate income (25% vs. 12%, respectively) (Boury, Larkin, & Krummel, 2004; Lennon, Blome, & 29 English, 2001; U.S. Department of Health and Human Services, 2006). The Center on the Developing Child (Harvard University, 2009) states one in four women living at or below the poverty line endorses moderate to severe depressive symptoms. The striking occurrence of maternal depression in women living in poverty is attributed to several factors, such as inadequate housing, lower levels of education, and decreased career opportunities, partner/marital discord, neighborhood violence, lack of childcare, and lack of social support (O’Hara, 1997; Witt et al., 2011). Furthermore, there are few resources available to help mothers in poverty cope with these problems. Numerous studies have investigated the effects of maternal depression on children’s outcomes. Existing theory identifies mothers’ parenting as one of the primary pathways through which mother’s depressive symptoms influence children’s outcomes (Lovejoy, Graczyk, O’Hare, & Neuman, 2000; Wilson & Durbin, 2010). This extant literature suggests that depressive symptoms are associated with increased occurrences of negative parenting practices (e.g., hostility, intrusiveness, and negative interactions) and decreased occurrences of positive parenting practices (e.g., warmth, sensitivity, and responsiveness) (Carter, Garrity-Rokous, Chazan-Cohen, Little, & Briggs-Gowan, 2001; Crockenberg & Leerkes, 2003; Lovejoy, Graczyk, O'Hare, & Neuman, 2000; McLearn, Minkovitz, Strobino, Marks, & Hou, 2006). Maternal depression has been widely studied in relation to children’s academic, cognitive, social, and behavioral outcomes (Goodman, Rouse, Connell, Broth, Hall, & Heyward, 2011; Peterson & Albers, 2001; National Research Council and Institute of Medicine, 2009). High levels of maternal depression are associated with increased behavioral problems (Dawson et al., 2003; Gartstein & Fagot, 2003; Gross, Shaw, Burwell, & Nagin, 2009), impaired language development (Paulson, Keefe, Leiferman, 2009), reading problems (Sohr-Preseton & 30 Scarmamella, 2006; Zaslow, Ahir, Dion, Ahluwalia, & Sargent, 2001), and social problems (Alperns & Lyons-Ruth, 1993; Feldman et al., 2009; Goodman, Rouse, Connell, Broth, Hall, & Heyward, 2011) in toddlers and early elementary-aged children. These studies suggest that maternal depression represents an added layer of risk for young children, especially those who are in poverty (Stein, Malmberg, Sylva, Barnes, & Leach, 2008). The effect of maternal depression on children’s outcomes are differentiated by the length of maternal depressive symptoms, the severity of maternal symptoms, and the time of exposure (Brennan, Hammen, Andersen, Bor, Najman, & Williams, 2000; Kurstjens, & Wolke, 2001; Sohr-Preston & Scaramella, 2006). Research has distinguished between children’s exposure to maternal depression during pregnancy, during infancy, during toddlerhood, and/or chronically throughout her child’s development. Infancy and toddlerhood are sensitive periods in which the early environment has the potential to influence children's cognitive and language development significantly (Murray & Cooper, 1997). Children's exposure to maternal depression early in development appears to be associated with prominent risk for children’s outcomes (Sohr-Preston & Scaramella, 2006; Peterson & Albers, 2001). Mothers with early depressive symptoms are more likely to use poor parenting practices (e.g., harsh parenting, punitive punishment style) and are more likely to continue to use poor parenting practices as their children develop (McLearn, Minkovitz, Stronbino, Marks, & Hou; 2006). Although maternal depression has been shown to be associated with poor outcomes in children at any time, maternal depression during the infancy and toddlerhood is thought to be especially detrimental to children’s outcomes (Luoma et al., 2001). Findings from longitudinal studies support this notion, as children with mothers with depression in the early developmental years have lower school readiness and verbal 31 comprehension at 36 months (NICHD Study of Early Childcare 1999) and more impaired cognitive functioning at 5 years of age (Brennan et al. 2000; Goodman et al., 2011). Maternal depression and literacy. Research indicates that mothers’ depression is negatively related to children’s early literacy outcomes (Baker & Iruka, 2013; Barbarin et al., 2006; Bigatti, Cronan, & Anaya, 2001; Foster, Lambert, Abbot-Shim, McCarty & Franze, 2005; Greenberg, et al., 1999; Reissland, Shepherd, & Herrera, 2003). As mentioned, maternal depression influences a variety of parenting practices related to children’s reading development (Kiernan & Huerta, 2008). Several studies have identified the specific reading-related practices affected by maternal depression. Using the 1988 National Maternal and Infant Health Survey and the 1991 Longitudinal Follow-Up Survey, McLennan & Kotelchuck (2000) found that maternal depression was related to a decrease in the frequency of daily reading to children, as mothers who rated more symptoms of depression read less to their three-year-old children. Interestingly, this finding was true only for women with a male partner. Bigatti, Cronan, and Anaya (2001) found that depressed mothers of one to three-year-old children read significantly less frequently to their children than non- depressed mothers. When depressed mothers did read to their children, they read for shorter periods of time, and asked fewer leading and exploratory questions about the story to their children. In a similar line of research, Reissland, Shepherd, and Herrera (2003) found that depressed mothers differed from non-depressed mothers in the quality of their book reading to their six-month to one-year-old children. Depressed mothers read faster and did not pause to ask their children questions about the story. Non-depressed mothers paused more often to ask their children questions about the story and adjusted their speech based on the ages of their children. 32 The authors concluded that depressed mothers were less attuned to their children’s reading needs and engaged in a less child-centered approach towards reading. Put together, depressed mothers engaged in less shared reading with their children than non-depressed mothers. The quality of the shared reading conducted by depressed mothers was less engaging and informative than non-depressed mothers. Given the increased risk of maternal depression due to poverty and the implications of maternal depression on parenting practices, future research should incorporate a measure of maternal depression into studies of children’s reading outcomes in poor samples. Including maternal depression might provide information about the disparities children from low-income experience since it represents the family stress path of poverty. Notably, children do not develop early literacy skills within a vacuum; instead, emergent literacy skills develop due to interactions children have with their literacy environments. Child-level factors have been shown to account for slightly more variance in initial literacy competence than home factors (Hindman, Skibbe, Miller, & Zimmerman, 2010), which suggests the importance of accounting for both environmental and child-level factors. For instance, research has shown that learning emergent literacy skills in the home depends on within child factors, such as attention, to sustain learning and practice (Cartwright, 2012; Shaywitz & Shaywitz, 2001). Duncan and colleagues’ (2006) seminal study showed that across six longitudinal data sets, the strongest predictors of children’s later reading achievement were school entry math, reading, and attention abilities. Further support for the importance of attention comes from an increasing number of studies that show children’s sustained attention as a significant predictor of reading (Blair & Razza, 2007; Martin, Razza, & Brooks-Gunn, 2012; McClelland, Cameron, Conner, Farris, Jewkes, & Morrison, 2007). Therefore, along with the 33 environmental factors that influence reading, it is important to consider the contribution of children’s attention skills to their reading outcomes (Rowe & Rowe, 1992). Sustained Attention The early acquisition of reading skills draws heavily upon children’s attentional systems. Attention is a broad cognitive construct that includes maintaining alertness, orienting to perceived events, and detecting signals for cognitive processing (Posner & Peterson, 1990). Preliminary support for a relation between attention and reading outcomes comes from extensive research demonstrating high comorbidity between attention deficit hyperactivity disorder (ADHD) and reading disability (RD) (Willcutt & Pennington, 2000; Shaywitz et al., 1995; Willcutt, Pennington, Olson, & DeFries, 2007; Willcutt et al., 2010); 15-35% of individuals with RD have comorbid ADHD (Germano, Gagliano, & Curatolo, 2010; Shaywitz et al., 1995). Many researchers have attempted to understand the epidemiology of this comorbidity. A significant body of evidence now indicates that attention and reading problems have a shared genetic etiology (e.g., Wilcutt et al., 2007; Ebejer et al., 2010; Zumberge, Baker, & Manis, 2007). Twin and behavioral genetics studies have identified shared pleiotropy (genes) that contributes to how attention and reading performance are related biologically and genetically. Thus, multiple lines of research have linked attention to reading; however, examining specific attention mechanisms as they relate to reading may help scholars illuminate the precise roles of attention in the reading process. One aspect of attention that is gaining interest in reading development is sustained attention (Fan, McCandliss, Sommer, Raz, & Posner, 2002; Prochonow, Tunmer, &Chapman, 2013; Rowe & Rowe, 1992; Stern & Shalev, 2013). Although there are a variety of specific attentional mechanisms that influence reading (e.g., orienting attention, joint attention, and 34 selective attention), sustained attention is thought to be a basic attentional function that determines “higher” levels of attention, and is necessary for the development of other cognitive skills (Sarter, Givens, & Bruno, 2001). It is not surprising that the literature on attention and reading indicates that sustained attention is an important skill for reading. Sustained attention is the ability to maintain focus on and readiness to respond to stimuli while inhibiting distractions from stimuli for a period of time (Anderson, 2008; Pardo, Fox, & Raichle, 1991). It requires children to maintain focus on relevant stimuli and inhibit distractions from irrelevant stimuli over a period of time (Petersen & Posner, 2012). It is referred to as vigilance in early publications (e.g., Kupietz, Samuel, & Richardson, 1978), and the absence of sustained attention can be referred to as inattention. Barkley (1997) described inattention as a multidimensional construct that resulted in behavioral and cognitive difficulties with sustaining attention and participating in goal-oriented actions/ thoughts. Since sustained attention is a necessary component of reading successfully, considering how sustained attention is measured is imperative. Sustained attention is measured using parent or teacher rating scales, continuous performance tasks, or clinical tests (e.g., subtests). Rating scales are indirect measures of sustained attention that require a rater to assess attention behaviors. Continuous performance tests (CPTs) are direct measures of sustained attention. There are many CPTs in existence (e.g., TOVA, Conners, Gordon); however, all have a similar framework for assessing sustained attention. Children are typically told to observe the presentation of stimuli (e.g., letters, numbers, pictures, tones); some are the target stimuli and others are the distractor stimuli. Children are asked to indicate the presence of the target stimuli by pushing a button and to ignore the distractor stimuli by refraining from pushing a button. 35 CPTs vary regarding stimuli modality (e.g., auditory or visual) and type of stimuli used (e.g., letters, shapes, tones, colors). Clinical subtests are standardized and normed clinical batteries that allow for the measurement of diverse attentional capabilities (e.g., sustained attention, divided attention, focused attention). Examples of standardized tests of sustained attention are the Test of Everyday Attention for Children (TEA-CH; Manly, Robertson, Anderson, & Nimmo-Smith, 2008) or the Leiter International Performance Scale –revised (Leiter-R; Roid & Miller, 1997). Cancellation tests are common tests of visual sustained attention that require children to find as many ‘target’ stimuli as possible on a piece of paper filled with very similar distractor stimuli. Examples of cancellation tests are the “Focused Attention” subtest on the Leiter-R or the “Sky Search” subtest on the TEA-CH. Sustained attention and literacy. Reynolds and Besner (2006) suggest that sustained attention is critical in each step of the hierarchical reading process (i.e., emergent literacy, decoding, fluency, and comprehension). Research using indirect and direct measures of attention support this statement. The ability to sustain attention has been linked to better emergent literacy skills, decoding, fluency, and comprehension (Fan, McCandliss, Sommer, Raz, & Posner, 2002; Kupietz, 1990; Rowe & Rowe, 1992; Stern & Shalev, 2013). Studies of teacher and parent-rated (indirect) sustained attention and reading provide a foundation for the importance of sustained attention to reading development. Velting and Whitehurst (1997) used SEM to examine reading skills in a sample of low SES Head Start Students. The researchers did not find significant relations between inattention- hyperactivity in kindergarten and decoding and letter-word naming in kindergarten; however, they found a significant path between children’s first grade inattention-hyperactivity and 36 decoding and letter-word reading skills in first grade. The researchers commented that the inconsistent path could be influenced by their parent-rated measure of attention, which combined inattention with hyperactivity and focused more on the dimension of hyperactivity (rather than inattention). Rabiner, Coie, and the Conduct Problems Research Group (2000) used path analysis to examine how a teacher-rated inattention dimension (as opposed to a hyperactivity dimension) related to children’s kindergarten, first-grade, and fifth-grade reading outcomes (WJ letter-word identification and passage comprehension, respectively). The researchers found that teacher- rated inattention problems in kindergarten put children at risk for low letter-word reading scores in first grade and low passage comprehension scores in fifth grade. Prochnow, Tunmer, & Chapman (2013) examined how teacher-rated inattention related to reading outcomes at school entry and seven years later. The researchers used study-created measures of decoding, spelling, reading comprehension and the Teacher Report Form of the Child Behavior Checklist (CBCL; Achenbach, 1991). In support of their longitudinal hypothesis, the authors found relatively robust associations between inattention and early reading, and between later reading comprehension and inattention. These studies showed that lower teacher or parent ratings of children’s inattention are associated with greater skills in emergent literacy skills, decoding, fluency, and comprehension. These robust findings suggest that sustained attention is integral to the reading process; attention in preschool and early elementary school is related to the development of early literacy skills, and has implications for reading achievement in later elementary school. These studies set the basis for investigating the relation between children’s sustained attention and reading skills. Nevertheless, there are some inherent problems with indirect measures of sustained attention. Indirect measures can be biased since they are dependent on a respondent and can be influenced 37 by the construction of the measure (e.g., Velting and Whitehurst’s inattention-hyperactivity measure) (Abikoff, Courtney, Pelham Jr., & Koplewicz, 1993; Servera, Lorenzo-Seva, Cardo, Rodriguez-Fornells, & Burns, 2009). Rating scale measures may result in biased information regarding children’s attention. Research suggests that responders, especially teachers, complete rating scales based on general academic performance or overall perception rather than the attentional behaviors of children, also called a “negative halo” effect (Abikoff, Courtney, Pelham Jr., & Koplewicz, 1993; Nisbitt & Wilson, 1977; Stevens & Quittner, 1998). Meaning, teacher and parent rating scales provide a global perspective of children’s attention rather than specific aspects of attention (Riccio, Reynolds & Lowe, 2001). Moreover, males from ethnic minority groups and low socioeconomic status households tend to be rated more poorly than their Caucasian, middle-class peers (DuPaul, Power, Anastopoulos, Reid, McGoey, & Ikeda,1997; Reid et al., 1998). Since the current study is investigating sustained attention in a diverse, at-risk sample, a direct measure of attention may provide a less biased, more precise perspective on children’s sustained attention and reading. Studies using direct measures of sustained attention, such as CPTs and standardized tests, have found significant relations to reading. Sims and Lonigan (2013) used teacher rating scales (Conners’ Teacher-ratings Scale Re-standardized; Conners, 1990) and a direct measure of sustained attention (CPT; Rosvold et al., 1956) to predict print knowledge vocabulary and phonological awareness among typically developing preschool children. The researchers found that more omission errors (inattention) on the CPT predicted lower print knowledge, vocabulary, and phonological awareness scores. The teacher ratings of sustained attention showed an inconsistent link between inattention and pre-literacy. The researchers concluded that the direct measure of sustained attention was a better predictor of children’s emergent literacy outcomes. 38 Razza, Martin, and Brooks-Gunn (2012) examined the effect of sustained attention (Leiter-R; Roid & Miller, 1997) and lack of impulsivity (Leiter-R) on academic outcomes (reading, math, and behavior) among low-income, at-risk students (Fragile Families and Child Wellbeing Study). The researchers were interested in how sustained attention at age five predicted passage comprehension, mathematic problem solving, and behavior problems at age nine. After controlling for language, child temperament, and maternal depression, regressions showed that the direct measure of sustained attention significantly predicted passage comprehension and applied math problem solving, while impulsivity was predictive of behavior problems. Interestingly, income did not moderate the relations between sustained attention and reading as the authors predicted. This may be because the authors created two separate groups (i.e., below the poverty line and above the poverty line) to investigate whether income moderated the association between attention and reading. Because the authors created dichotomous income group, they likely lost variability in their analyses. Taken together, the extant studies using behavioral ratings and direct measures of sustained attention have effectively linked sustained attention to reading. Inattention was shown to be predictive of reading achievement; however, the strength of the relation was dependent on the type of measure of attention (e.g., behavioral rating, direct measure) and the relation was not consistent in a low-income sample (Razza, Martin, & Brooks-Gunn, 2012). More information is needed to understand how sustained attention is associated with reading for children in low- income households. Examining the early home literacy environment and maternal depression with sustained attention may provide valuable information about how sustained attention influences reading for children in low-income households. 39 Sustained Attention, Home Literacy, and Maternal Depression The early home literacy environment, maternal depression, and children’s sustained attention play a critical role in the development of reading according to Bronfenbrenner’s bioecological framework. Specifically, reading abilities are thought to emerge in the social context of the early home literacy environment and are believed to be shaped by continuous interactions between children and their environments (proximal processes). Parents play an important role in the development of children’s early reading skills by sharing in the exploration of books or directing their sustained attention towards reading materials (Charland, Perron, Boulard, Chamberland, & Hoffman, 2015; Evans, Williamson, & Pursoo, 2008). Research supports this idea. When studied together, more early literacy exposure, better maternal mental health, and better sustained attention were associated with better emergent literacy outcomes in children, whereas low literacy exposure and poor sustained attention are associated with poorer emergent literacy outcomes (Downer & Pianta, 2006; NICHD ECCRN, 2003; Martin, Razza, & Brooks-Gunn, 2010). It may also be that the home literacy environment acts as a protective factor for children at risk for reading difficulty such that the early home literacy environment prepares children to take advantage of more formal literacy instruction (Haak, Downder, & Reeve, 2012). The following studies have examined the associations between the home literacy environment, maternal depression, sustained attention, and reading/language outcomes. Haak, Downer, and Reeve (2012) examined whether the home literacy environment and parent-rated attention influenced letter-word reading and expressive vocabulary for four-year-old children. The home literacy environment was measured using a composite of the Learning Materials, Language Stimulation, and Academic Stimulation Home Observation Measurement of the Environment subscales (HOME; Caldwell, & Bradley, 1984). Gender, ethnicity, income, 40 maternal education, maternal sensitivity, and cognitive ability were entered as covariates. Regressions showed that early home literacy environments and sustained attention predicted letter-word reading. Moderation showed no effect of sustained attention on reading scores. However, significant moderation effects for sustained attention on language scores were found, as children with typical attention scores who were exposed to high early literacy had significantly higher expressive vocabulary when compared to children with atypical attention scores and the same amount of high early literacy exposure. The authors concluded that sustained attention and home literacy are important factors in the development of children’s language outcomes. The NICHD Early Child Care Research Network (2003) investigated whether a direct measure of sustained attention and the home literacy environment was related to children’s letter-word reading outcomes. Sustained attention was assessed via continuous performance task (CPT; Rosvold, Mirsky, Sarason, Bransome, & Beck, 1956), and environment was measured by the maternal sensitivity and cognitive stimulation composites of the HOME (Caldwell, & Bradley, 1984) and the total scores from videotaped ratings of maternal sensitivity and cognitive stimulation averaged across four age points (6, 15, 36, and 54 months) (NICHD Early Child Care Research Network, 2000). School readiness was assessed with a composite of the Woodcock Johnson-Revised (WJ-R; Woodcock & Johnson, 1990) letter-word reading and applied problem- solving subtests. Using Baron and Kenny's (1986) model for mediation. The researchers found a significant mediation effect; the home environment predicted children's sustained attention, and sustained attention, in turn, predicted reading and math achievement. In other words, results showed that poor sustained attention reduced the effect of the home literacy environment on children’s achievement scores, while better sustained attention increased the effect of the home 41 literacy environment on children’s achievement scores. The NICHD study suggests that children with better sustained attention benefit from the home environment more than children with poor sustained attention. It is important to note that although the findings were statistically significant, the researchers combined the WJ-R letter word and math problem solving subtests into a single measure; thus, is it not clear whether sustained attention and the home literacy environment relate to reading achievement specifically. Davidse, de Jong, Bus, Huijbregts, and Swaab (2010) studied the relations between children’s short-term memory, sustained attention, and home literacy environment on their language and letter-word recognition using a sample of pre-school-aged Dutch children from middle-income families. The home literacy environment was measured with one question assessing the frequency that parents read to their child per week (0 = not so often, 1 = every other day, 2 = on a daily basis) and a Dutch checklist of book title recognition (Stichting Collective Propaganda van het Nederlandse Boek; CPNB, 2006). Children’s reading outcomes were assessed with a study-created letter-knowledge test (Alpha reliability equaled .90) and children’s language outcomes were assessed with the Dutch version of the Peabody Picture Vocabulary Test (PPVT-III; Schlichting, 2005). Sustained attention was assessed with a direct measure called the Amsterdam Neuropsychological Tests (De Sonneville, 2005). Hierarchical regression showed that the home literacy environment was predictive of children’s letter-word and language scores. The authors hypothesized that sustained attention would moderate the effects of the home literacy environment and literacy outcomes; however, tests for moderation were not significant. Consequently, the researchers concluded there was no support for the hypothesis that sustained attention moderated learning from the home literacy environment. The 42 authors suggested that no significance was found because parent-child book reading sessions are highly structured and children are not required to sustain attention to the book. Lastly, Razza, Martin, and Brooks-Gunn (2010) examined the relations between the early “family” environment, sustained attention, poverty status, and school readiness using the Fragile Family and Child Wellbeing data set. Children were grouped into two groups by family income and labeled “poor” and “near poor.” The researchers recognized the investment and family stress mediators by including measures that address parental investments and maternal stress/depression in their conceptualization of the “family” environment. Environment was measured using a composite of the maternal warmth, cognitive stimulation, and learning material HOME subscales (Caldwell, & Bradley, 1984); a composite of parenting stress (Parenting Stress Index; Abidin, 1995); and a depression screener (Composite International Diagnostic Interview Short Form; Kessler, Andrews, Mroczek, Ustun, & Wittchen, 1998). School readiness was measured with a scale of receptive language (Peabody Picture Vocabulary Test-III; Dunn & Dunn 1997) and externalizing behaviors (Child Behavior Checklist; Achenbach, 1991). Sustained attention was measured with the sustained attention subtest of the Leiter International Performance Scales, Revised (Leiter-R; Roid & Miller, 1997). For the near-poor group, results from multiple regression analyses showed that sustained attention mediated 25.25% of the total effect of maternal warmth on children’s receptive vocabulary. No mediation effects were found for the poor group. For both groups, children’s sustained attention significantly predicted receptive vocabulary. Notably, this study did not include a measure of reading and, instead, used a measure of language. This study was the only study to include maternal depression as a variable or as a control. 43 Altogether, four conclusions can be made from these studies. First, the well-documented, positive associations between the early home literacy environment and early reading achievement were supported by all studies’ findings (Davidse, de Jong, Bus, Huijbregts, and Swaab; 2010; Haak, Downer, and Reeve, 2012; NICHD Early Child Care Research Network; 2003; Razza, Martin, and Brooks-Gunn, 2010). Since all studies took a cross sectional approach to the research, longitudinal relations between the variables of interest cannot be made. Second, sustained attention was inconsistently shown as a mechanism underlying the associations between home literacy environments and children’s reading/language outcomes. This could be because significant results were found depending on whether sustained attention was conceptualized as a mediator or a moderator variable. Razza, Martin, and Brooks-Gunn (2010) and the NICHD Early Child Care Research Network (2003), found significant results for sustained attention as a partial mediator. Both studies’ results showed that children with higher sustained attention also better reading outcomes; however, Razza, Martin, and Brooks-Gunn’s (2010) finding was only significant for the near-poor group. On the other hand, sustained attention was not significant when it was considered as a moderator of the relation between early home literacy and reading outcomes (Davidse, de Jong, Bus, Huijbregts, & Swaab, 2010; Haak, Downer, & Reeve, 2012). Haak, Downer, and Reeve (2012) found significant moderation effects of sustained attention on home literacy environments and children’s expressive vocabulary outcomes but not reading outcomes, while Davidse, de Jong, Bus, Huijbregts, & Swaab (2010) did not find significant moderator effects of sustained attention on the relation between home literacy, language, and reading. Third, the inconsistent results could also be a function of the varying measures of sustained attention and reading outcomes used by the researchers in these studies. Although all 44 studies used directed measures of attention, all but one study (Martin, Razza, & Brooks-Gunn, 2012) used CPTs. The studies also used differing measures of reading; some studies used letter- word identification (Haak, Downer, & Reeve, 2012; Davidse, de Jong, Bus, Huijbregts, & Swaab, 2010), another combined letter-word identification with applied math problem solving into a composite (NICDH, 2010), and another used only a language measure (Martin, Razza, & Brooks-Gunn, 2012). The present study aims to address these inconsistencies to determine whether sustained attention functions as a mediator in the hypothesized associations. Fourth, only one study (Razza, Martin, and Brooks-Gunn, 2012) included income as a variable in the examination of sustained attention, maternal depression and home literacy; however, the researchers did not include a distinct reading outcome variable. Razza, Martin, and Brooks-Gunn’s results showed sustained attention as a partial mediator of the association between income, the home environment and children’s language, but only for near-poor children. This finding might be explained by a methodological problem with the study’s design and measure selection. Razza, Martin, and Brooks-Gunn (2012) created two groups, a “poor” group and “near-poor” group. By creating two binary groups and restricting the sample size based on income, the study lost variability in its design. This might account for why the mediation between sustained attention, home environment, and children's outcomes were only significant for near-poor children. As such, an income variable with more variability is necessary as small differences in household income could influence the associations between home literacy, sustained attention, maternal depression, and reading outcomes. An alternative explanation for the differences between income groups may be how the researchers applied the meditational framework that describes how income influences children's outcomes by the investment and family stress model. Razza, Martin, and Brooks-Gunn used 45 Baron and Kenny’s (1986) model for determining criteria for mediation. Although this is a well- validated method, newer structural equation modeling techniques, such as Sobel tests and bootstrapping, allow the associations between variables to be examined simultaneously and might provide a more comprehensive picture of how income, sustained attention, maternal depression, and the home environment influence children’s reading development. In summary, there is a need for further research on the differential roles of early sustained attention, early home literacy environments, and maternal depression in the development of reading in children from poor and near-poor households. Despite promising findings, there are some holes in the research to address. While it is clear that sustained attention is an essential element of the reading process, more research is needed to determine the role it plays when considering income, home literacy environments, maternal depression, and children’s reading outcomes together. The current study aims to address some of the concerns in the existing literature by:1) investigating cross-sectional and longitudinal paths; 2) test sustained attention as a mediator; 3) including variables that are appropriate for low-income samples (investment and family stress informed paths); and 4) using an income variable with more variability. It is important to consider that factors beyond sustained attention, the home literacy environment, maternal depression, and income may be responsible for the mixed findings in the literature. Therefore, it may be beneficial to include a genetic variable in the present study to conduct an exploratory analysis. A benefit of using the Fragile Families data set is the inclusion of genetic variables. During the in-home component of the year nine wave, saliva was collected from focal children to extract genetic information to allow researchers to test hypotheses about the associations among genes, family environments, and child outcomes (Waldfogel, Craigie, & Brooks-Gunn, 2010). Reading ability appears to be at least partially influenced by genetic 46 factors. Untangling the genetic components that influence reading may provide valuable information about how the home environment interacts with cognitive/biological factors to influence reading. One of the genes measured by the Fragile Family and Child Wellbeing Study was the DRD4 VNTR, a gene associated with both attention and reading. DRD4, Attention, and Literacy The present study used a bioecological approach to understand how environmental and individual factors influence children's development. An examination within the bioecological framework would not be complete if it did not acknowledge the influence of genetics. Studies examining the roles of gene by environment interplay, or G X E (e.g., Moffitt, Caspi, & Rutter, 2006) have been guided by the diathesis-stress framework and the differential-susceptibility framework. The diathesis-stress perspective posits that certain genetic characteristics create psychological vulnerability to stressful environments and that individuals with vulnerable genes in stressful environments develop negative outcomes (Zuckerman, 1999). The differential susceptibility framework posits that certain individuals may be susceptible to negative and positive environments (Belsky, 2005; Belsky & Pluess, 2009; Belsky, Bakermans-Kranenburg, & Van IJzendoorn, 2007). Meaning, certain individuals with susceptibility genes flourish in positive environments and decline in negative environments. These frameworks are not contradictory per se. It may be that some genes and environments interact in a diathesis-stress manner and some genes and environments interact in a differential susceptibility manner (Belsky, 2005). New methodologies in behavioral genetics have allowed researchers to link individual genes to behavioral and psychological symptomology. Knowledge about a child’s genotype may 47 not have immediate practical implications (Mullineaux, DiLalla, & Fisher, 2015); however, it can contribute to understanding how certain children may be more responsive or vulnerable to certain aspects of their environment, such as the home literacy environment. In turn, this can lead to recommendations that modify the environment, such as promoting early reading for all children. Also, knowledge about genotypes can provide information as to why some children are more responsive to specific treatments (Moffit, 2005). Furthermore, it can inform the selection of measures examining specific endophenotypes (behavioral expressions) of DRD4, such as sustained attention, to screen children at risk for developing poor reading outcomes (Kegel & Bus, 2012). The gene of interest is the dopamine receptor DRD4. DRD4 is one of the most studied genes in child psychology due to its relation to several behavioral, socio-emotional, and cognitive children's outcomes (Bakermans-Kranenburg, Van IJzendoorn, Pijlman, Mesman, & Juffer, 2008; Buil, Koot, Olthof, Nelson, van Lier, 2015; Kluger, Siegfried, & Ebstein, 2002; Smith et al., 2012). DRD4 has been identified as relevant to the behavioral expression of both RD and ADHD (Barkley et al., 2010; Gizer, Ficks, & Waldman, 2009). Additionally, DRD4 is associated with attention problems, like inattention, hyperactivity, and impulsivity, (Auerbach, Benjamin, Faroy, Geller & Ebstein, 2001; Gizer & Waldman, 2012; Kieling, Roman, Doyle, Hutz, & Rohde, 2006; McCracken et al., 2000; Schmidt, Fox, Perez-EdgarHu, & Hamer, 2001). The literature has found that DRD4 affects dopamine production in the prefrontal cortex, an area of the brain involved in regulating attention (Posner & Rothbart, 2007). The long variant (7- repeat allele) of DRD4 has been linked to lower dopamine reception efficiency in the brain, which is associated with lower levels of attention (Tripp & Wickens, 2008). 48 When DRD4 was studied directly in relation to reading, the direct significance of DRD4 was negligible. Hsiung and colleagues (2004) showed a marginally significant link (p = .06) between the 7-repeat allele and reading disability, but Marino and colleagues (2003) did not. This may be because DRD4 does not directly influence reading, instead, DRD4 directly influences attention. Since attention has been shown to influence reading, DRD4 may influence the relation between attention and reading, such that DRD4 directly influences attention and attention influences reading. Currently, there are only four studies that examine the indirect relation between DRD4 and children’s reading outcomes. Kegel and Bus (2012) studied the indirect relation between DRD4 to reading by examining the role of executive attention in a kindergarten-aged sample. The researchers hypothesized that executive attention would mediate the relations between DRD4 and “alphabetic skills.” Executive attention was defined as children’s ability to manage working memory, focus attention on target stimuli, and ignore non-target stimuli. It was measured by creating a composite using principle component analysis from direct measures of digit span forward, digit span backward, and a Stroop task. Reading was measured by creating a composite from direct measures of spelling, rapid automatic reading, letter knowledge, and word reading. Results showed that children with the 7-repeat DRD4 allele had poorer executive attention and poorer alphabetic skills. A Sobel test revealed that executive attention was a mediator variable, such that children’s DRD4 status (7-repeat or no 7-repeat) predicted performance on executive attention, which then predicted alphabetic scores. The researchers concluded that children with the DRD4 7-repeat allele were more at risk for delays in alphabetic skills because of attention problems. 49 Kegel, Bus, and van Ijzendoorn (2011) believed that the DRD4 7-repeat allele may increase risk for inattention and differentiate response to feedback during reading instruction. They investigated the association of children’s genotype (DRD4 7-repeat allele or no 7-repeat allele) to responses to a reading feedback intervention. Children were assigned a literacy intervention that had either: 1) positive feedback for a correct response, or 2) no feedback for a correct response. After a 15-week intervention, children with the 7-repeat DRD4 allele in the positive feedback version of the reading intervention showed significant improvement on post- intervention reading tests. Children with the 7-repeat DRD4 allele in the no-feedback version of the reading intervention showed little growth on the post intervention reading test (although they did not differ significantly from the control group). Children without the 7-repeat DRD4 alleles were not significantly influenced by the feedback or no-feedback intervention conditions. The researchers concluded that the intervention showed that DRD4 7-repeat reflected differential susceptibility. Later in 2015, Belsky, van Ijzendoorn, Plak, Kegel, & Bus tested a reading intervention for kindergarten students to determine whether DRD4 7-repeat alleles would predict students’ performance. The researchers split children into groups by reading ability (delayed or not delayed) and provided a reading intervention. Results at the end of the intervention showed that that DRD4 moderated the effects of the intervention. In the delayed group, children with the DRD4 7-repeat showed the most growth in reading scores (d = 0.56), while children without DRD4 7-repeat showed less reading growth (d = -0.09). In the not delayed group, children with the DRD4 7-repeat showed the same growth as children without DRD4 7-repeat. The researchers concluded that DRD4 influenced children’s motivation, attention, and reinforcement mechanism (which were addressed by the reading intervention), and the pattern of reading group showed 50 differential susceptibility pattern of functioning (Belsky et al., 2007; Belsky & Pluess, 2009). A replication of this study was published recently (October 2016) by the same authors (Plak, Merkenbach, Kegel, van Ijzendoor, & Bus, 2016). In summary, although there is little support for the direct relation between DRD4 and reading, research by Kegel and Bus (2011) found evidence for the mediating effects of attention on DRD4 and reading. Kegel, Bus, and van Ijzendoorn (2012), and Belsky, van Ijzendoorn, Plak, Kegel, and Bus (2015) found evidence for the moderating effect of DRD4 on predicting children’s response to reading interventions. These studies contribute to the large number of existing studies have found that genes moderate and mediate and the impact of social environments (Kim-Cohen et al., 2006; Simons, Beach, & Barr, 2012). Results showed that children with the 7-repeat allele were less able to regulate their attention during learning; thus, it may be that their reading development was negatively affected, in part, by their difficulties sustaining attention during reading practices at home and in school. Since there are only four studies that examine the relation between DRD4 and reading, the present study takes an exploratory approach to investigate the relations between DRD4, income, the home literacy environment, maternal depression, sustained attention, and reading outcomes. Purpose of the Present Study The existing literature provides ample evidence for the importance of early home literacy environments, maternal depression, and sustained attention when examining reading outcomes. There is also abundant support for the deleterious effects of poverty on children’s reading outcomes, such that reading problems occur early, persist throughout children’s schooling, are difficult to remediate, and increase risk for behavioral problems, social-emotional problems, and 51 high school incompletion (Crooks, 1995; Duncan & Brooks-Gunn, 1997). Despite this evidence, there are several gaps and methodological issues in the existing literature. First, the existing studies examining home literacy environments and maternal depression using at-risk samples do not completely account for the mediational theories that explain how poverty influences children’s reading outcomes. Many studies only account for the investment path and do not account for the family stress path (e.g., Britto & Brooks-Gunn, 2001; Storch & Whitehurst, 2001; Bracken & Fischel, 2008). By not including or controlling for either path, these studies do not acknowledge an important component of poverty that influences children’s reading. The present study addressed this limitation by including measures of maternal depression and early home literacy environments. Although one study included variables informed by the investment and family stress path, the Razza, Martin, & Brooks-Gunn (2012) study is not without limitations. Contrary their hypotheses, Razza, Martin, and & Brooks-Gunn found that income was only influential for the near-poor group but not for the poor group. This might be because the researchers divided their sample into dichotomous groups, the poor (0- 100% of the poverty line) and the near poor (100-300% of the poverty line). Binary grouping often causes studies to lose variability. Since the Fragile Family data set is populated with low- income families in the first place, un-grouping the sample may provide more interpretable results and allow for more meaningful conclusions about how monetary differences in household income influence parental investments and maternal mental health. The present study used a categorical measure of income with five categories. This strategy allows more variability into the analyses, and allowed conclusions to be made about how differences in household income associated with increases or decreases in investments in the home literacy environment and maternal mental health. 52 Non-significant findings in the Razza, Martin, and Brooks-Gunn study might have also resulted from the use of a binary measure of parental investments in reading (i.e., the HOME). The HOME measures the environment by recording whether target objects are present “1” or not present “0” (i.e., the presence of a book about letters). This is problematic because homes with one book and homes with ten books were coded the same, although there is considerable difference between owning one and owning ten books; thus, quantifiable differences in the home literacy environment were not recognized. Thus, along with not accounting for differences in income, the Razza, Martin, & Brooks-Gunn study did not adequately account for the variation in material investments either. The present study used measures of household income and parental reading investments with more variability that may allow for more interpretable results, described further in the methods section. Second, this study examined the mediating mechanism underlying the associations between sustained attention, home literacy environments, maternal depression, and children’s reading outcomes. Differing results were found depending on whether sustained attention was conceptualized as a mediator or a moderator variable. Razza, Martin, and Brooks-Gunn (2010) and the NICHD Early Child Care Research Network (2003), found significant results for sustained attention as a partial mediator. Alternatively, sustained attention was not significant when it was considered as a moderator of the relation between early home literacy and reading outcomes (Davidse, de Jong, Bus, Huijbregts, & Swaab, 2010; Haak, Downer, & Reeve, 2012). Therefore, the hypothesis that children with better sustained attention would benefit more from the home literacy was not supported (Davidse, de Jong, Bus, Huijbregts, & Swaab, 2010; Haak, Downer, & Reeve, 2012). It may be that home literacy is beneficial to children regardless of their attention abilities (Davidse, de Jong, Bus, Huijbregts, & Swaab, 2012). It also might be that 53 sustained attention is not a moderator during the early developmental period since caregivers structure and lead many of the home literacy activities. Sustained attention might become more influential to children’s reading outcomes at a later time, perhaps when reading depends more on children’s self-regulating capacities, such as during early elementary school (Altemeier, Abbott, & Berninger, 2008). Since this study focused on reading outcomes and attention during the early developmental period and there is existing evidence supporting sustained attention as a mediator, the current study examined sustained attention as a mediator rather than a moderator. Third, this study examined the longitudinal effects of early home literacy environments and sustained attention on children’s reading outcomes. Presently, no research has investigated whether the early effects of home literacy environments and sustained attention persist beyond children’s early elementary school reading outcomes. Fourth, this study took advantage of genetic data collected by the Fragile Family and Child Wellbeing study to understand whether a gene (DRD4) affects the relations between income, early home literacy environments, maternal depression, sustained attention, and children's reading outcomes. Since no study has attempted to examine whether DRD4 influences these relations in a longitudinal study design, this aspect of the study was exploratory in nature. In summary, few studies have considered the combined roles of the early literacy environment, maternal depression, and sustained attention on reading outcomes. This study employed statistical methods (SEM) that accounted for income, sustained attention, home literacy environments, maternal depression, and reading outcomes within a complex conceptual framework. SEM allowed the present study to use a moderated-mediation model that could estimate complex associations between the multiple predictor variables. The following research questions and hypothesizes were tested. 54 Research Questions, Hypotheses, Analyses, and Rationale Research Question One. What are the direct and indirect associations between income, early home literacy environments, early maternal depression, early sustained attention, and letter-word reading in kindergarten? Hypothesis 1a. Lower income categories, lower home literacy environments, lower sustained attention, and higher maternal depression were hypothesized to be associated with lower letter-word reading scores. Conversely, higher income categories, higher home literacy environments, higher sustained attention, and lower maternal depression were hypothesized to be associated with higher letter-word reading scores. Hypotheses 1b. It was hypothesized that families with lower income would have homes with lower literacy environments, which would be associated with lower reading scores. Families with lower income were expected to have children with lower sustained attention development and, therefore, lower reading scores. It was also expected that families with lower income would have mothers with more endorsements of depression, and mothers with more depression would have children with lower reading scores. Conversely, families with higher income were hypothesized to have home environments with more literacy and, therefore, children with better developed reading. Families with higher income were expected to have children with better developed sustained attention, which would be associated with better developed reading. Families with higher income were expected to have mothers with less endorsements of depression, and less depressed mothers would have children with higher reading scores. Rationale. Early home literacy environment. Numerous studies have found that exposure to literacy in the home environment is vital to children’s development of early reading skills (Senechal & LeFevre, 2002; Senechal, LeFevre, Hudson, & Lawson, 1996; Senechal, LeFevre, 55 Thomas, & Daley, 1998; Whitehurst & Lonigan, 1998). Exposing young children to literacy concepts before age five has been linked directly to better performance in letter-word reading (Cunningham & Stanovich, 1998; Davidse et al., 2011; Hood, Conlon, & Andrews, 2008; Senechal & LeFevre, 2002; Whitehurst & Lonigan, 1998). Consistent with the existing literature, the early home literacy environment was predicted to be directly associated with children’s letter-word identification scores, such that children from households with more home literacy would have higher letter-word reading than children from households with less home literacy. Mediator. The meditational theories of poverty (i.e., the investment and family stress models) posit that income influences children’s outcomes by affecting family processes and investments (Yeung, Linver, & Brooks-Gunn, 2002; Duncan & Brooks-Gunn, 1997). The home literacy environment is a good representation of the investment path because the home literacy environment measure is comprised of questions that reflect families’ literacy resources, like the numbers of books in the home and the frequency of parent-child shared reading. The early home literacy environment in kindergarten is expected to mediate the association between income and the letter-word identification outcome, such that lower income would be associated with lower early home literacy environment which would lead to lower letter-word identification scores. Higher income would be associated with higher early home literacy environment, which would be linked with higher letter-word identification scores. Rationale. Maternal Depression. Studies have shown that maternal depression negatively impacts children’s reading outcomes. Depressed mothers engaged in less shared reading with their children, and the quality of the shared reading conducted by depressed mothers was less engaging and informative than non-depressed mothers. Both practices led to poor performance in their children’s emergent literacy (Bigatti, Cronan, & Anaya, 2001; McLennan & Kotelchuck, 56 2000). Consistent with the existing literature, maternal depression was predicted to be directly associated with children’s letter-word identification scores, such that mothers who endorsed more symptoms of depression would have children with less developed letter-word reading and mothers who endorse less symptoms of depression would have children with more developed letter-word reading. Mediator. The meditational theories of poverty (i.e., the investment and family stress models) posit that income influences children’s outcomes by affecting family processes and investments (Yeung, Linver, & Brooks-Gunn, 2002; Duncan & Brooks-Gunn, 1997). Maternal depression is a component of the family process meditational path because it represents a mental health factor related to income (Razza, Martin, & Brooks-Gunn, 2010). Although a parenting variable is missing, maternal depression is predicted to mediate the relation between income and letter-word identification scores, such that families with higher income were expected to have mothers with less endorsements of depression, and less depressed mothers would have children with higher letter-word reading scores. Conversely, families with low income were expected to have mothers with more endorsements of depression, and more depressed mothers would have children with lower letter-word reading scores. Rationale. Sustained Attention. Sustained attention is critical in each step of the hierarchical reading process (Reynolds & Besner, 2006); children with better sustained attention abilities have better emergent literacy skills, like decoding and letter-word knowledge (Fan, McCandliss, Sommer, Raz, & Posner, 2002; Stern and Shalev, 2012). Previous research has shown support for the predicted direct association between sustained attention and letter-word identification scores. Consistent with the existing literature, sustained attention was hypothesized to be directly related to letter-word identification scores. Children with better developed 57 sustained attention were hypothesized to have better developed letter-word reading. Children with less developed sustained attention were hypothesized to have less developed letter-word reading. Mediator. When sustained attention and reading were examined in a poor and near poor sample, Razza, Martin, and Brooks-Gunn (2010) found that sustained attention partially mediated the association between income, the family environment, and children’s language, but only for near-poor children. The NICHD Early Child Care Research Network (2003), also found similar mediating effects. When moderating effects were examined, results were not significant. Davidse, de Jong, Bus, Huijbregts, & Swaab, 2010; Haak, Downer, & Reeve, 2012) did not find significant moderator effects of sustained attention on the relation between home literacy, language, and reading. As such, there is evidence for the mediating effect of sustained attention. Families with higher income were expected to have children with better developed sustained attention, which would be associated with better developed letter word reading. In contrast, families with lower income were expected to have children with less developed sustained attention, which would be associated with less developed letter word reading. Figure 1. Cross-sectional Conceptual Model Research question 2 (longitudinal model). What are the direct and indirect associations between early income, early home literacy environments, early maternal depression, early sustained attention, and passage comprehension in third grade? 58 Hypothesis 2a. Lower income categories, lower home literacy environments, lower sustained attention, and higher maternal depression were hypothesized to be associated with lower passage comprehension reading scores. Conversely, higher income categories, higher home literacy environments, higher sustained attention, and lower maternal depression were hypothesized to be associated with higher passage comprehension reading scores. Hypothesis 2b. Families with lower income were expected to have homes with lower literacy environments, which would be associated with lower passage comprehension scores. Families with lower income would have children with lower sustained attention development and, therefore, lower passage comprehension scores. Families with lower income would have mothers with more endorsements of depression, and mothers with more depression would have children with lower passage comprehension scores. Conversely, families with higher income would have home environments with more literacy and, therefore, children with better developed passage comprehension. Families with higher income would have children with better developed sustained attention, which would be associated with better passage comprehension. Families with higher income would have mothers with less endorsements of depression, and less depressed mothers would have children with higher passage comprehension scores. Rationale. Early home literacy environment. Early exposure to early literacy in the home environment is vital to children’s development of early reading skills and is predictive of children’s later reading skills (Burgess, Hectch, & Lonigan, 2002; Hart et al., 2009; Van Steensel, 2006. Good early home literacy environments have been linked to better performance on reading comprehension measures in later elementary and middle school (Keith et al., 1993; Peterson & Davidson, 2013). Thus, there is evidence for the direct association of the early home literacy environment with children’s later reading comprehension outcomes. Consistent with the 59 existing literature, the early home literacy environment is predicted to be directly associated with passage comprehension scores. Specifically, children from homes with better early home literacy environments would have better passage comprehension in third grade. In contrast, children from homes with poor early home literacy environments would have worse passage comprehension in third grade. Mediator. As mentioned, the investment and family stress models posit that income influences children’s outcomes by affecting family processes and investments (Yeung, Linver, & Brooks-Gunn, 2002; Duncan & Brooks-Gunn, 1997). Research shows that academic problems are much more prevalent for children who experience the effects poverty in early childhood because of the importance early literacy skills (e.g., letter naming and letter sounds) in determining the course of schooling for children (Duncan, Yeung, Brooks-Gunn, & Smith, 1998). It may be that the lack of exposure to materials that promote reading at an early age negatively influences children’s later reading outcomes. The early home literacy environment in kindergarten is expected to mediate the relation between income and the passage comprehension outcome, such that lower income would be associated with lower early home literacy environment which would lead to lower letter-word identification scores in kindergarten. Lower letter-word scores in kindergarten would then be associated with lower passage comprehension scores in third grade. Alternatively, higher income was hypothesized to be associated with higher early home literacy environment, which would be linked with higher letter-word identification scores in kindergarten. Higher letter-word scores in kindergarten would then be associated with better passage comprehension scores in third grade. Rationale: Maternal Depression. Studies have shown that maternal depression negatively impacts children’s later reading outcomes, especially maternal depression during children’s early 60 developmental stages (Brennan, Hammen, Andersen, Bor, Najman & Willians, 2000; Dahlen, 2016). Consistent with the existing literature, maternal depression was predicted to be directly associated with passage comprehension scores, such that children with mothers with more symptoms of early depression would have low passage comprehension scores, and children with mothers with less symptoms of early depression would have higher passage comprehension scores. Mediator. The meditational theories of poverty (i.e., the investment and family stress models) posit that income influences children’s outcomes by affecting family processes and investments (Yeung, Linver, & Brooks-Gunn, 2002; Duncan & Brooks-Gunn, 1997). Maternal depression is a component of the family process meditational path because it represents a mental health factor related to income (Razza, Martin, & Brooks-Gunn, 2010). Although there was no parenting measure, maternal depression was still predicted to mediate the relation between income and passage comprehension scores, such that families with higher income would have mothers with less endorsements of depression, and less depressed mothers would have children with better developed letter-word reading in kindergarten. Better letter-word reading in kindergarten would then be associated with better passage comprehension in third grade. Conversely, families with lower income would have mothers with more endorsements of depression, and more depressed mothers would have children with poorly developed letter-word reading in kindergarten. Poor letter-word reading in kindergarten would then be associated with poor passage comprehension in third grade. Rationale. Sustained attention. Longitudinal studies show support for the enduring relation between sustained attention and later reading achievement (Rabiner, & Coie, 2000; Velting & Whitehurst, 1997), such that children who display poor attention in preschool or 61 kindergarten exhibit lower levels of reading achievement in later elementary and middle school. These findings show support for the predicted direct relation between sustained attention and passage comprehension. Specifically, children with better early sustained attention would also have better passage comprehension in third grade, while children with worse early sustained attention would have worse passage comprehension in third grade. Mediator. Despite past research indicating that sustained attention may be a key ingredient for emergent literacy, it is unclear whether early sustained attention mediates the link between income and later reading outcomes. Previous research found that early sustained attention predicted later reading (Rabiner & Coie, 2000; Velting & Whitehurst, 1997) and indicated that children from low-income families are more likely to develop problems with attention and other cognitive factors (e.g., memory, executive functions) (Lengua, 2002; McLoyd, 1998; NICHD ECCRN, 2004; Schmitz, 2003). As such, there is evidence for the mediating effect of sustained attention longitudinally. The study hypothesizes that sustained attention in kindergarten is expected to mediate the relation between income and the passage comprehension outcome, such that families with higher income would have children with better developed sustained attention, which would be associated with better developed letter word reading in kindergarten. Better letter-word reading in kindergarten would then be associated with better passage comprehension in third grade. Conversely, families with lower income would have children with poorly developed sustained attention, which would be associated with poorly developed letter word reading in kindergarten. Poor letter-word reading in kindergarten would then be associated with poor passage comprehension in third grade. 62 Figure 2. Longitudinal Conceptual Model Research Question 3. Research question three was exploratory given the limited evidence to suggest that DRD4 is associated with the variables of interest. Does DRD4 influence the associations between income, early home literacy environment, and reading outcomes? Diathesis-stress (risk model). A diathesis-stress framework proposes that the DRD4 7- repeat allele is a risk allele that places children at risk for poor attention despite positive environmental experiences. The diathesis-stress model posits that children’s outcomes are influenced by the interactions between environmental and individual factors, which result in a disorder or condition (Scarr, 1992; Zuckerman, 1999). Specifically, diathesis states that an individuals’ biological predispositions (vulnerability) to particular disorders can be triggered by stressful life events. Individuals with low predisposition or low vulnerability for a disorder would require exposure to high levels of stress to trigger symptoms of that disorder. Alternatively, individuals with high predisposition or high vulnerability for a disorder would require lower levels of stress for symptoms of a disorder to develop. According to the diathesis-stress model, it was hypothesized that despite high early literacy exposure in the home, children with genetic risk (DRD4 7-repeat) would have poorer letter-word identification and passage comprehension scores than their peers with without genetic risk and high early literacy exposure. In other words, children with genetic risk and high 63 early literacy exposure would have poorer letter-word identification and passage comprehension scores than their peers without genetic risk and high early literacy exposure. Differential susceptibility (plasticity model). Alternatively, the differential susceptibility hypothesis proposes that the DRD4 7-repeat allele is a plasticity allele that responds differentially to children’s environments (e.g., high literacy exposure vs. low literacy exposure). Belsky and colleagues (Belsky et al., 1997; Belsky & Pluess, 2009) reviewed studies that showed evidence for the diathesis-stress model and hypothesized that individual differences (genes) are not just more sensitive to adverse environments. Instead, the “risk gene” was actually a “plasticity gene” where children are more susceptible to adverse environments and enriched environments. Belsky titled this theory the “differential susceptibility” hypothesis. He suggests that genes influence how responsive children are to their environmental contexts; some children are less sensitive to the environment and some children are more sensitive the environment (Belsky et al., 2007). Children with this “sensitive” genetic predisposition are more vulnerable to adverse environments but also flourish more in supportive environments. Dobbs (2009) used the metaphor “orchid” to describe susceptible children and the metaphor “dandelion” to describe non-susceptible children. “Orchids” are strongly dependent on the quality of the environment: they suffer more from bad instruction and profit more from optimal teaching. “Dandelions” are more adaptive to environments and perform adequately no matter the conditions. Although this study samples children from fragile families and there may be fewer environmental factors that would push the children towards thriving, there are still many children who are exposed to optimal early home literacy environments. As such, there is ample opportunity to demonstrate differential susceptibility in “at-risk” populations and better understand how parents and 64 caregivers create positive environments for their children to learn to read despite considerable adversity. The studies that examined DRD4 in relation to reading provide support for the differential susceptibility hypothesis. Kegel and Bus (2012) and Kegel, Bus, and van Ijzendoorn (2015) showed significant moderating and mediating effects of DRD4 between attention, environmental factors, and reading. Specifically, Kegel, Bus, and van Ijzendoorn (2015) found that children with the DRD4 7-repeat allele were more responsive to a reading intervention, such that children with the DRD4 7-repeat who were included in the positive condition reading invention showed the most growth in reading scores and children with the DRD4 7-repeat who were not included in positive condition reading intervention. According to the differential susceptibility model, it was hypothesized that children with the DRD4 7-repeat allele and high home literacy environment would have higher reading scores than their peers with without the DRD4 7-repeat allele and high early home literacy environment. Children with the DRD4 7-repeat allele and low home literacy environment were predicted to have lower reading scores than children without the DRD4 7-repeat allele and low home literacy environment. In other words, children with the DRD4 7-repeat allele would have reading outcomes scores that reflect the quality of their early home literacy environment. Figure 3. Exploratory Conceptual Model 65 CHAPTER 3 METHOD Study Model The study used data from the Fragile Families and Child Wellbeing Study (FFCWS) to examine the mechanisms through which income, the home literacy environment, maternal depression, and sustained attention in the early childhood period associate with children’s reading outcomes in kindergarten and third grade. Specifically, the current study tested three models. First, a cross sectional model investigating the relations between income, the early home literacy environment, maternal depression, sustained attention, and kindergarten letter-word identification outcomes was tested. Second, a longitudinal model examining the relations between income, the early home literacy environment, early maternal depression, early sustained attention, and third grade passage comprehension outcomes was examined. Third, an exploratory model was identified to determine whether the DRD4 gene moderates the relations in the longitudinal model. Structural equation modeling (SEM) analyses were used to determine the associations among the variables in all three models. The latent variable was the early home literacy environment. The observed variables were income, maternal depression, sustained attention, DRD4, kindergarten letter-word identification, and third grade reading comprehension. Covariates included marital status, mothers' age, mothers' education, mothers' race/ethnicity, children's age at assessment, children's language estimate, children's gender. Data Source This study used data from the Fragile Families and Child Wellbeing Study (FFCWS) year five (kindergarten) and year nine (3rd grade) data collection waves. The FFCWS was developed 66 under the efforts of the Center for Research on Child Wellbeing (CRCW), the Center for Health and Wellbeing (CHW), the Columbia Population Research Center (CPRC), and the National Center for Children and Families (NCCF). The study has multiple funding sources, including the Eunice Kennedy Shriver National Institute of Child Health and Human Development (R01-HD- 36916 and 5P30-HD-32030), the National Science Foundation, and the Office of the Assistant Secretary for Planning and Evaluation and Administration for Children and Families (U.S. Department of Health and Human Services). The primary investigator was granted permission to use the restricted-use Fragile Families and Child Wellbeing Study (FFCWS) data for this project under a public data license and restricted data license agreement provided to the dissertation faculty chair (Jodene G. Fine, NCSP, PhD, LP). The Michigan State University Institutional Review Board granted the primary investigator approval to complete this project (IRB #x16-142e). This study takes advantage of multi-informant, longitudinal data that represents an at-risk population to study constructs related to children’s reading achievement. FFCWS design. The FFCWS study is following a longitudinal birth cohort of 4,789 children born in 20 large American cities between 1998 and 2000 (Reichman, Teitler, Garfinkel, & McLanahan, 2001). The FFCWS chose children born in 20 large U.S. cities (i.e., cities with populations greater than 200,000). Sixteen of the 20 cities were selected using a stratified random sample (Reichman, Teitler, Garfinkel, & McLanahan, 2001), the cities were: Indianapolis, IN; Austin, TX; Boston, MA; Santa Ana, CA; Richmond, VA; Corpus Christi, TX; Toledo, OH; New York, NY; Birmingham, AL; Pittsburgh, PA; Nashville, TN; Norfolk, VA; Jacksonville, FL; San Antonio, TX; Philadelphia, PA; and Chicago, IL (Reichman, Teitler, Garfinkel, & McLanahan, 2001). In the recruitment stage mothers were enlisted from maternity 67 ward lists from selected hospitals. Trained FFCWS examiners asked the mothers to complete a screening survey to determine marital status and eligibility for participation in the study. FFCWS sampled children born to unmarried mothers (3/4 unmarried to 1/4 married) to investigate the mothers’ and focal children’s wellbeing. FFCWS self-describes its population as “fragile families” because unmarried parents and their children have a greater risk of family instability, poverty, and socioeconomic disadvantage than traditional families (e.g., nuclear families). To sample “fragile families” the number of unmarried and married births quotas was based on sample cities’ 1996/1997 unmarried birth rates. No new mothers or children were recruited since baseline. The core FFCWS has a longitudinal design. Data has been collected for the past nine years at multiple time points. There have been five data collection time points: after birth of the child (baseline; 1998-2000; n = 4,789), t year one (1999-2000; n = 4,270), at year three (2001- 2003; n = 4,140), at year five (2003-2006; n = 4,055), and at year nine (2007-2010; n = 3,391) (see Table 1). Year 15 data collection is currently underway and is projected to finish in 2018. Table 1. Original FFCWS Data Sample Data Collection Wave Baseline Year One Year Three Year Five Year Nine Sample Size Focal Child Age Year n = 4,789 n = 4,270 n = 4,140 n = 4,055 n = 3,391 Mean 0 1.3 3.6 5.8 9.9 SD 0 .70 .51 .65 .73 1998-2000 1999-2001 2001-2003 2003-2006 2007-2010 Over the waves of data collection, the data was gathered from multiple informants, including primary caregivers (mostly mothers), secondary caregivers, teachers, focal children, and FFCWS examiners. Data was collected using in-person and phone interviews with the 68 primary caregiver and secondary caregiver, mailed surveys filled out by the primary caregivers, phone interviews with focal children’s teacher, mailed surveys filled out by the focal children’s teacher, observations during home visits, and direct focal children assessments during home visits. The data that was collected are public and restricted access. Public access data include information on maternal mental health, child well-being, child grades, child behaviors, and environmental components of the home and neighborhood environment. Restricted access data include information on genetic information on the focal children and biological mothers, medical records, school characteristics, and macroeconomic data. FFCWS sample demographics. The FFCWS collected longitudinal, multi-informant, multi-modal forms of data that are descriptive of a cohort of at-risk children born between 1998 and 2000. Thus, the results from the current study will yield conclusions representative of children and mothers with similar characteristics to this sample. In the baseline sampling year (1998-2000) 4,789 mothers and their children from major city hospitals across the country. Children who were enrolled in the FFCWS were Black/African American (44%), Hispanic (35%), White (17%), and Other (4%). Slightly more children in the study were male (53%). Most the primary caregivers were mothers (96%) followed by grandmothers (2%). Fathers represented less than one percent of the sample. Mothers gave birth primarily between ages 20 through 24 (39%), followed by under age 20 (27%), age 30 and older (18%), and between ages 25 and 29 (16%). The mothers and fathers of children were mostly cohabitating and in a relationship (51%), followed by not cohabitating but in a relationship (31%), not cohabitating and friends (8%), not cohabitating with little to no contact (9%), and father unknown (1%). In terms of income, 27% of mothers made “far below the poverty line,” 18% of mothers were 69 “below the poverty line,” 28% of mothers were “just above the poverty line,” 14% of mothers were “above the poverty line,” and 13% of mothers were “very above the poverty line” in the years 1998-2000. The final sample for the current study included children from the FFCWS study that had primary caregiver reports of early home literacy environments, maternal depression, sustained attention, and genetic data from the primary caregiver reports and in-home data collection at year five (kindergarten), and primary caregiver reports, genetic data, and reading data from in-home data collection at year nine (3rd grade). Since the predictors and outcomes required direct assessment (i.e., sustained attention, reading achievement, genetic sampling), it was important that children have data from the in-home assessments at both wave five and wave nine to be included in this study. Since the variables of interest were collected from the in-home assessment, children who did not participate in both the three and five year wave were excluded. Additionally, given the interests of this study and as to not influence results, children were excluded if they had a parent report of a neurological condition, autism spectrum disorder, visual or hearing impairment, and/or intellectual disability. The final study sample for research questions 1 and 2 was n =2,062. Research question 3 required the inclusion of a genetic variable, DRD4. There were 1,556 participants who consented to providing genetic information. After the application of exclusion criteria, n =1,517 participants were included for research question 3 analyses. To understand the nature of the missing data, a series of comparisons were conducted between the included sample and the excluded sample to determine whether there were notable differences between variables of interest (see Table 2, 3, 4). Although there were several statistically significant differences between the two groups (e.g., language estimate, sustained 70 attention), none were of clinical significance to the current study. Most of the observed differences between the included and excluded groups differed within a few tenths of a point, such as the third-grade language estimate variable (PPVT-III). For example, in third grade children included in the sample scored 92.63 on average on the Peabody Picture Vocabulary Test, Third Edition (PPVT-III) while children excluded scored a 92.84 on average on the same test. The difference between the average scores does not represent a meaningful difference in receptive language. There were some larger differences observed. Income in the included sample had a higher frequency of low income families than the excluded sample. There is also a higher frequency of mothers with higher education in the included sample than the excluded sample. Missing data analyses revealed that only 14.09% of cases were missing and 1.77% of values were missing from the final study sample (n=2,062). Missing data analyses revealed that the missingness of the values and cases did not appear to be missing due to relatedness to another variable. As such, it is assumed that the cases and values are missing at random (MAR) (Schafer & Graham, 2002). Based on the assumption of MAR and of multivariate normality, full information maximum likelihood (FIML) was selected to address missing data. FIML derives model fit information from a summation across all individual observations for a variable. Then, a likelihood function for each missing case is created (based on summation fit). FIML estimates two models, a null (H0) model and an alternative (H1) model and derives a chi-square estimated to determine difference between the two models. The H0 model assumes that all variables are correlated and is unrestricted. The H1 model assumes the hypothesized restrictions. When conducting SEM in Mplus, FIML was applied automatically by the software. 71 Table 2. Comparison of Included and Excluded Samples (Categorical) Characteristics Included Sample n= 2062 1059 (51.4%) 1003 (48.6%) 403 (19.5%) 1117 (54.2%) 483 (23.4%) 59 (2.9%) 592 (28.7%) 289 (14%) 81 (5%) 221 (10.7%) 399 (19.4%) 478 (23.3%) 424 (21.6%) 398 (20.3%) 839 (42.8%) 298 (15.2%) 236 (11.4%) 1826 (88.5) 451 (21.9%) 446 (21.6%) 515 (25%) 288 (14%) 362 (17.6%) n= 1533 901 (58.8%) 632 (41.2%) Total Child Gender Male Female Child Race/Ethnicity White Black Hispanic Other Mother’s Relationship Married Rom Cohab Rom Sep/Wid/Div Friends No Relationship/Unknown Mothers' Education Less than high school High school or equivalent Some college/technical College or above Mother’s Depression Yes No Income (Categorical) 0-49% 50-99% 100-199% 200-299% 300% + DRD4 Status DRD4 short DRD4 long Excluded Sample n= 2836 1509 (53.2%) 1326 (46.8%) 627 (22.2%) 1209 (42.8%) 853 (30.2%) 135 (4.8%) 700 (33.8%) 245 (11.7%) 57 (2.7%) 229 (11%) 361 (17.4%) 483 (23.3%) 342 (22%) 351 (22.6%) 594 (38.2%) 268 (17.2%) 354 (12.5%) 2491 (87.8%) 442 (21.2%) 350 (16.8%) 566 (27.2%) 284 (13.7%) 435 (20.9%) -- -- Chi-Square (df) 1.673 (1), p=0.196 67.40 (3), p=.000** 23.09 (5), p=.002* 8.88 (3), p= 0.31 1.903 (1), p=0.096 674.31 (5), p=.000** N/A 72 Table 3. Comparison of Included and Excluded Samples (Continuous) Variables Included Sample Means (SD) Mothers’ Variable Mother’s Age Children’s Variables Age at assessment (K) Age at assessment (3rd) Language Estimate PPVT-R (K) PPVT-III (3rd) Sustained Attention Focused Attention Lack of Impulsivity 24.92 (5.95) 5.68 (1.23) 9.20 (.31) 93.90 (15.46) 92.63 (14.15) 12.82 (3.25) 10.12 (2.84) Excluded Sample Mean (SD) 25.53 (6.09) 5.7 (1.52) 9.4 (.41) 94.33 (16.54) 92.84 (16.01) 12.04 (3.6) 9.77 (3.01) Table 4. Comparison of HLE Indicators for Included and Excluded Samples Variables Excluded Sample N (%) Days per week read with child 0 1 2 3 4 5 6 7 Number of book w/ colors None 1-2 3-4 5 or more Number of book w/ numbers None 1-2 3-4 5 or more Number of book w/ songs None 1-2 Included sample N (%) 53 (2.6%) 102 (4.9%) 198 (9.6%) 328 (15.9%) 250 (12.1%) 334 (16.2%) 67 (3.2%) 730 (35.4%) 41 (2.0%) 270 (13.2%) 506 (24.7%) 1133 (60.1%) 64 (3.1%) 328 (16%) 476 (23.3%) 1177 (57.6%) 191 (9.3%) 479 (23.4%) 44 (1.6%) 80 (2.8%) 185 (6.5%) 302 (10.6%) 229 (8.1%) 296 (10.4%) 57 (2.0%) 793 (28%) 11 (1.2%) 92 (10%) 203 (22.1%) 614 (66.7%) 24 (2.6%) 116 (12.7%) 201 (22%) 574 (62.7%) 91 (9.9%) 203 (22.1%) Independent Samples t- test (df) 2.251 (4892), p= .134 123.12 (3560), p=.000** 108.41 (3513), p=.000** 10.315 (3560), p=.001* 21.444 (3344), p =.000** 3.455 (2076), p=.063 .887 (2076), p=.346 Chi-Square (df) 758.258 (10), p= .000** 13.845 (3), p=.003** 8.646 (3), p= .034* 3.552 (3), p=.314 73 Table 4 (cont’d) 3-4 5 or more Number of book w/ alphabet None 1-2 3-4 5 or more How many books in home None 1-10 11-20 More than 20 How often encourage read Less than 1x/month About 1x/month Few times/month Few times/week Every day 449 (22%) 926 (45.3%) 64 (3.1%) 364 (17.8%) 461 (22.5%) 1161 (56.5%) 5 (0.2%) 178 (8.7%) 222 (10.9%) 1641 (80.2%) 71 (3.5%) 32 (1.6%) 236 (11.6%) 693 (34%) 1008 (49.4%) 180 (19.6%) 443 (48.3%) 23 (2.5%) 149 (16.2%) 179 (19.5%) 567 (61.8%) 2 (0.2%) 75 (8.1%) 116 (12.6%) 728 (79%) 31 (3.4%) 21 (2.3%) 94 (10.3%) 281 (30.8%) 486 (53.2%) 7.169 (3), p= .067 2.055 (3), p= .561 6.575 (4), p=.160 Final study sample. The final sample for the current study was (n= 2,062). The mean age of children during year five (kindergarten) was 5.68 years and the mean age of children during year nine (3rd grade) was 9.20 years (see Table 5). Children in the final sample were almost evenly male (51.4%) and female (48.6%). Children were approximately 19.5% White, non-Hispanic, 54.2% Black or African-American, 23.4% Hispanic, race specified, and 2.9% Other race/ethnicity. Table 5. Child Age in Final Study Sample Data Collection Wave Age of Study Child (months) Five (kindergarten) Mean 5.68 SD 1.23 Range 3.81-7.10 Nine (3rd grade) 9.20 0.31 8.62-10.88 Years 2003- 2006 2007- 2010 Sampling weights. To make the study sample representative of the national population at the time of data collection, FFCWS created national weights. Weighting the data allows the 74 included data to describe the population it was taken from (i.e., ¾ unmarried mothers and ¼ married mothers from the major cities between the years 1998 and 2000). Unlike other longitudinal datasets, FFCWS only created cross-sectional weights to make each wave representative of the original sampling frame. There were no longitudinal weights created; therefore, for studies using measures from more than one wave, the FFCWS researchers suggested using the weight of the wave from which the most variables were taken (Carlson & Mathematica Policy Research, 2008). The majority of variables in the current study were taken from the year five wave (kindergarten); thus, participants in the present study were assigned the wave five sample weight “m4natwt.” Although not ideal, this weight best fits the analyses based on the available weights and the recommendations from the FFCWS researchers. The “m4natwt” weight allows for the demographic results of the study to be generalizable to children born in large cities from “fragile families” from the birth cohort of 1998-2000. Unweighted and weighted descriptive analyses for the demographic and predictor variables were conducted (see Tables 6 and 7). The unweighted descriptive statistics represent the FFCWS population where each case was counted equally. The weighted descriptive statistics represents the entire population of children born in 1998-2000 (based on the U.S. Census) from the original 20 sampled cities, and weights cases unequally to reflect the entire population born in 1998-2000 from the original 20 sampled cities. Although descriptive analyses were conducted with the weighted and unweighted samples, SEM analyses were conducted with the unweighted sample. While descriptive analyses with the weighted sample produced appropriately scaled statistics, the SEM analyses with the weighted sample did not. Weights were not used with SEM analyses because it was determined that analyses conducted with weights overestimated the Fragile Families population (Carlson, 75 2008). The overestimation was determined to be due to the complex sampling design, unequal selection probabilities, response rates across cities, hospitals, and births, and lack of availability of longitudinal weights2 (Carlson, 2008). After consulting with researchers from Fragile Families Child and Family Wellbeing Study, the present study added control variables to the model as an alternative to using weights in SEM analyses. The variables added to the analyses were informed by the procedure the FFCWS used to create the weights, known as raking. Raking is a method of stratification to ensure the weighted counts of the sample are consistent with population counts informed by US Census between 1998 and 2000 (Carlson, 2008). The raking variables were: mothers’ marital status, mothers' education level, mothers' race/ethnicity, and mothers' age. Consequently, the variables the FFCWS researchers recommended adding to the present models were: mothers’ marital status, mothers' age, mothers' education, and mothers' race/ethnicity. The final covariate list includes marital status, mothers' age, mothers' education, mothers' race/ethnicity, children's age at assessment, children's language estimate, children's gender. All covariates were retained in the analyses regardless of significance in the models per the recommendation of the FFCWS researchers. Table 6. Comparison of Unweighted and Weighted Samples (Categorical) Characteristics Child Gender Male Female Child Race/Ethnicity White Black Hispanic Other Sample N 1059 1003 403 1117 483 59 51.6 48.4 32.9 27.9 31.8 7.4 Sample % Weighted N Weighted % 51.4 48.6 19.5 54.2 23.4 2.9 302904 232413 176173 149306 170071 39767 2 Weights were not longitudinal. The Fragile Families guidebook suggested applying the weight that corresponded to the year with the most variables (Carlson, 2008). 76 Table 6 (cont’d) Mother’s Relationship Married Rom Cohab Rom Sep/Wid/Div Friends No Relationship/Unknown Mothers' Education Less than high school High school or equivalent Some college/technical College or above Mother meets CIDI Cut-Off Yes No Income (Categorical) 0-49% 50-99% 100-199% 200-299% 300% + DRD4 Status n= 1533 missing DRD4 short DRD4 long 592 289 81 221 399 478 424 398 839 298 236 1826 451 446 515 288 362 529 901 632 47.1 8.8 2.7 18.5 9.4 13.5 33 26.9 18.2 21.4 8.8 91.2 17.3 17 23.9 14.3 27.5 22.4 51.4 26.1 252007 46971 14192 98889 50150 72643 176840 144236 97327 115053 47008 488309 92661 90852 127924 76652 147228 120061 275373 139884 28.7 14 5 10.7 19.4 23.3 21.6 20.3 42.8 15.2 11.4 88.5 21.9 21.6 25 14 17.6 25.6 43.7 30.6 Table 7. Comparison of Unweighted and Weighted Samples (Continuous) Variables Weighted Mean Sample S.D. Sample Mean Sample Range Weighted S.D. Weighted Range 24.92 5.68 9.20 Mothers’ Variable Mother’s Age Children’s Variables Age at assessment (K) Age at assessment (3rd) Language Estimate PPVT-R (K) 93.90 5.95 1.23 0.31 20-50 3.81- 7.10 8.62- 10.88 25.95 7.87 9.79 6.48 1.33 2.97 20-50 3.81-7.10 8.62-10.88 15.46 40-139 95.71 16.86 40-139 77 92.63 Table 7 (cont’d) PPVT-III (3rd) Sustained Attention Focused Attention 12.82 Lack of 10.12 Impulsivity 14.15 3.25 2.84 44-139 1-19 0-17 96.37 12.76 10.12 Sample % 2.6 4.9 9.6 15.9 12.1 16.2 3.2 35.4 2.0 13.1 24.5 59.7 3.1 15.9 23.1 57.0 Sample N 53 102 198 328 250 334 67 730 41 270 506 1233 64 328 476 1177 Table 8. Unweighted and Weighted HLE Indicators (Categorical) Variables Days per week read with child 0 1 2 3 4 5 6 7 Number of book w/ colors None 1-2 3-4 5 or more Number of book w/ numbers None 1-2 3-4 5 or more Number of book w/ songs & rhymes None 1-2 3-4 5 or more Number of book w/ alphabet None 1-2 3-4 5 or more How many books in home None 191 479 449 926 64 364 461 1161 5 9.3 23.2 21.8 44.9 3.1 17.6 22.3 56.3 0.2 78 15.59 3.59 2.69 44-139 1-19 0-17 Weighted N Weighted % 12739 25970 53027 68725 55588 96828 17381 205059 5673 69917 125244 332898 20179 84650 132813 295071 2.4 4.9 9.9 12.8 10.4 18.1 3.2 38.3 1.1 13.1 23.4 62.2 3.8 15.8 24.8 55.1 47112 126958 118113 239377 19485 96030 152754 265163 733 8.8 23.7 22.1 44.7 3.6 17.9 28.5 49.5 0.1 Table 8 (cont’d) 1-10 11-20 More than 20 How often encourage read Less than 1x/month About 1x/month Few times/month Few times/week Every day Variables 178 222 1641 71 32 236 693 1008 8.6 10.8 79.5 3.4 1.6 11.4 33.6 48.8 51977 56071 425052 22882 7777 42957 170469 287648 9.7 10.5 79.4 4.3 1.5 8.0 31.8 53.7 The variables included in the model are Likert-scale/ordered categorical (e.g., CBCL) and continuous (e.g., Leiter-R, PPVT-III) measures. A table of constructs and their associated measures used in the models are provided (see Table 8). The covariate list includes marital status, mothers' age, mothers' education, mothers' race/ethnicity, children's age at assessment, children's language estimate, children's gender. Table 9. Overview of Study Variables and Sources Constructs 1. Early home literacy environment 5^ 2. Early sustained attention 3. Maternal depression 3. Genetic variable 4. Kindergarten letter-word ID 5. Third passage comprehension 6. Income variable 7. Mother’s education** 8. Mother’s age** 79 FFCWS Data Source FFCWS Variable a. Parent survey (wave 5) a. Focused attention (wave 5) b. Lack of impulsivity (wave 5) a. CIDI (wave 5) a. DRD4 VNTR (wave 9) a. WJ-R letter-word ID (wave 5) a. WJ-III passage comp (wave 9) a. Parent survey (wave 3) a. Parent survey (wave 5) a. Parent survey (wave 5) a. PCS a. HV b. HV a. PCS a. HV a. HV a. HV a. PCS a. PCS a. PCS Table 9 (cont’d) 9. Mother’s age** 10. Mother’s relationship status** 11. Children’s early language** 12. Children’s later language** 13. Children’s age at assessment** 14. Children’s age at assessment** a. Parent survey (wave 9) a. Parent survey (wave 5) a. PPVT-R (wave 5) a. PPVT-III (wave 9) a. Home visit assessment (wave5) a. Home visit assessment (wave 9) a. Parent survey (wave 5) a. PCS a. PCS a. HV a. HV a. HV a. HV 15. Ethnicity/race** Note: ^ indicates a latent construct. Confirmatory Factor Analysis was conducted on individual items to identify constructs. ** indicates covariate. Note: PCS = Primary caregiver survey, PCSA = Primary caregiver self-administered, MQ = Mother questionnaire, FQ = Father questionnaire, TS = Teacher survey, CS= Child survey, IO = Interviewer observations, HV = Home visit a. PCS Reading achievement. Reading achievement assessments were conducted with the focal children during the wave five and wave nine home visits. Data was collected over multiple years, so the data represent fall and spring outcomes (see Table 1). Letter-word identification begins to emerge in kindergarten and decoding and fluency should be well developed by third grade. Thus, letter word identification and reading comprehension are appropriate measures to determine whether reading is developing on time (Pikulski & Chard, 2005). Early reading achievement. The Woodcock-Johnson Revised Tests of Achievement (WJ- R; Woodcock & Johnson, 1990) letter-word identification subtest was used as an early reading achievement outcome. The letter-word subtest measured children's skills in identifying words and letters names. For the WJ letter–word identification test, internal reliability for preschool- aged children is good (α =.92) (Woodcock & Mather, 1989). The FFCW researchers derived 80 standardized scores (M = 100, SD = 15) from raw scores recorded at test time using the national norms published by WJ-R test manual. Later reading achievement. The Woodcock-Johnson Tests of Achievement, Third Edition (WJ-III; Woodcock, McGrew, & Mather, 2001) passage comprehension subtest was used as a later reading achievement outcome. The passage comprehension subtest measured children's skills in understanding the meaning of the text they read. The questions required children to identify pictures that correspond with words, use context clues to find missing words, and answer questions about the meaning of the text. Internal reliability for children is good (α =.88) (McGrew, & Woodcock, 2001). The FFCW researchers derived standardized scores (M = 100, SD = 15) from raw scores recorded at test time using the national norms published by the WJ-III test manual. Early home literacy environment. The early home literacy environment informed by the investment models (Haveman & Wolfe, 1994; Mayer, 1997). The parent investment model suggests that the effect of family income on children is apparent in parents’ decisions about how to allocate their money, time, energy, and support. Questions about literacy investments were selected from the wave five caregiver survey based on commonly used questions to measure the home literacy environment (e.g., Evans & Shaw, 2008, Scarborough & Dobrich, 1994; Sénéchal, 2006). Parent responses on these items were used as indicators of the early home literacy environment for the focal child. The FFCWS data set did not provide a composite for early literacy environment, so confirmatory factor analysis was conducted with a sample of 2,602 using MPlus student version 7.4 (Muthén & Muthén, 2012). CFA defines sets of variables that reliably measure the constructs of interest and tests whether underlying latent constructs exist (Kline, 2016; Suhr, 2006). The following questions were included in the CFA: 81 - How many days per week do you read to CHILD? (1, 2, 3, 4, 5, 6, 7 days) - About how many toys, books or games does (child) have that are helping him/her know about colors? (None, 1-2, 3-4, 5 or more) - About how many toys, books or games does (child) have that are helping him/her know about numbers? (None, 1-2, 3-4, 5 or more) - About how many toys, books or games does (child) have that are helping him/her know about nursery rhymes or songs? (None, 1-2, 3-4, 5 or more) - About how many toys, books or games does (child) have that are helping him/her know about the alphabet? (None, 1-2, 3-4, 5 or more) - How many books are in the home? (None, 1-10 books, 11-20 books, and 20 or more books) - How often do you encourage CHILD to read? (less than once a month, about once a month, a few times a month, at least a few times per week, every day) To perform the CFA, each item was identified as an ordered categorical indicator, and the variance of the latent factor was fixed at one to allow each indicator’s factor loading to be estimated freely. There were high correlations (covariances) between indicators in the CFA, so covariances were specified in the model between the days per week parents read with their child and the number of times parents encouraged their child to read. Multiple fit indices suggested good model fit. X2= 125.75, p=0.014 and RMSEA= 0.063 (lower CI = 0.053, upper CI= 0.073), and relative fit indices, CFI =0.986, TLI=0.977, WRMR=1.207. The chi-square index is not significant, which allows for the rejection of the null hypothesis. The CFI and TLI are near 1 (Bentler, 1990). Although the RMSEA exceeds the .05 cutoff (Browne & Cudeck 1992), the 90% lower bound CI is within the .05 range, suggesting an adequate fit. Finally, although the 82 WRMR is slightly above 1, this is not uncommon in studies with large samples (Bollen, 1989; Hu & Bentler, 1999; Yu, 2002). Standardized factor loadings ranged from 0.877 to 0.270 (Table 9). Two indicators had a low communality value (<.4) (i.e., days per week read to child and how often read to child), indicating less explained variance within the construct; however, the indicators were retained because the items had significant factor loadings and the items played a role in the interpretations of the early home literacy environment. Days per week parents read to their child and how often parents encourage their children to read, had standardized loadings of 0.307 and 0.270, respectively. As predicted by the existing literature, the indicators selected loaded onto a single latent construct representing the early home literacy environment. Lower scores on this measure correspond to lower early home literacy environment and higher scores on this measure correspond to higher early home literacy environment. The HLE latent variable was tested for skewness and results indicated skewness of -1.56. Although this variable is not normally distributed, SEM with maximum likelihood (ML) can still be used; however, there are caveats as SEM has been shown to be less robust when data are ordinal or non-normal (i.e., very skewed or kurtotic). Nonetheless, some literature suggests that non-normal variables can be used if skewness and/or kurtosis is within +/- 1.5 – 2.0 (Garson, 2007). Table 10. Confirmatory Factor Analysis for HLE HLE Days per week read with CHILD Number of books about color Number of books about numbers Number of books about rhymes/songs Number of books about the alphabet Number of books in the house How often do you encourage CHILD to read Unstandardized 0.982 2.602 3.104 2.309 2.939 1.891 0.877 Standardized 0.307 0.754 0.877 0.678 0.837 0.565 0.270 83 Maternal depression. Questions about maternal depression were selected from the wave five primary caregiver survey. The Composite International Diagnostic Interview-Short Form, Section A (CIDI-SF; Kessler et al., 1998) was administered to mothers. The short form of the CIDI interview takes a portion of the full set of CIDI questions and generates the probability that the respondent would be positively diagnosed as experiencing a Major Depressive Episode if the full CIDI interview were given. The CIDI questions are consistent with the Diagnostic and Statistical Manual of Mental Disorders-Fourth Edition (DSM-IV; APA, 1994). The CIDI is a standardized assessment instrument intended for use in research studies. Mothers were asked whether they have had feelings of dysphoria (depression) or anhedonia (inability to enjoy what is usually pleasurable) in the past year that lasted for two weeks or more. If so, mothers were asked whether the symptoms lasted most of the day and whether the symptoms occurred every day of the two-week period. If so, they were asked more specific questions about: - Losing interest - Feeling tired - Change in weight - Trouble sleeping - Trouble concentrating - Feeling worthless - Thinking about death The FFCWS used the CIDI-SF to classify respondents according to the criteria for a DSM-IV major depressive episode (MD). Likelihood for MD was determined by mothers’ affirmatively answering three or more questions. FFCWS then classified mothers into “1” likely 84 MD case or “2” unlikely MD case. The distinction between mothers with major depressive disorder, major depressive episodes that occur as part of a bipolar disorder, or major depressive episodes that occur in the course of psychotic disorders was not made. Higher scores indicate more endorsements of feeling symptoms of depression, while lower scores indicate less endorsements of feeling symptoms of depression. Scores ranged from 0 – 7 (Table 10). Table 11. Frequency of CICI-SF Distribution for All Cases Probability of MD Caseness Mothers (n) Short Form Depression Score 0.0001 0 0.0568 1 0.2351 2 0.5542 3 0.8125 4 0.8895 5 0.9083 6 7 0.9083 Note: Table adapted from CRCW (2015) 2,405 9 19 38 81 187 1809 61 Although the maternal depression variable is informed by only the wave five primary caregiver survey, research on the stability of maternal depression in the FFCWS suggests maternal depression is stable over time and across income categories (see table 12). Research on the stability of maternal depression by Turney (2012) showed between 40 and 50% of mothers were depressed across two consecutive waves, as 50% of mothers depressed at the wave 1 survey were also depressed at the wave 3 survey, 39% of mothers depressed at the wave 3 survey were also depressed at the wave 5 survey, and 41% of mothers depressed at the wave 5 survey were also depressed at the wave 9 survey. These depression stability statistics are consistent with other low-income, high-risk data sets (McLennan, Kotelchuck, & Cho, 2001; Pascoe, Stoilfi, & Ormond, 2006). Given this consistent approximation of maternal depression across waves, it is appropriate to use a wave 5 as the entry point for examining maternal depression to maintain 85 consistence across mediator variables (i.e., the sustained attention and HLE mediator variables are also from wave 5). 1095 (25%) 811 (18.5%) 1094 (25.0%) 599 (13.7%) 765 (17.5%) Table 12. Stability of Maternal Depression Wave 1- Wave 9 Income (Categorical) year 1 to year 3 Wave 1 0-49% 50-99% 100-199% 200-299% 300% + Income (Categorical) year 3 to year 5 Wave 3 0-49% 50-99% 100-199% 200-299% 300% + Income (Categorical) year 5 to year 9 Wave 5 0-49% 50-99% 100-199% 200-299% 300% + Note: adapted from Turney (2012) 957 (22.6%) 819 (19.4%) 1059 (25.0%) 573 (13.5%) 823 (19.4%) 893 (21.5%) 796 (19.2%) 1081 (26.1%) 572 (13.6%) 797 (19.3%) Wave 3 957 (22.6%) 819 (19.4%) 1059 (25.0%) 573 (13.5%) 823 (19.4%) Wave 5 893 (21.5%) 796 (19.2%) 1081 (26.1%) 572 (13.6%) 797 (19.3%) Wave 9 610 (17.5%) 693 (19.9%) 1008 (28.9%) 480 (13.8%) 691 (19.8%) DRD4. Saliva DNA samples were taken at the wave nine home visit with the Oragene DNA sample collection kit (DNA Genotek). Focal children were instructed to spit into a container until the volume of saliva was 2 ml. The FFCWS examiner capped the container, and a liquid preservative was released. The container was then put into a small plastic biohazard bag with safety precautions and sent to be analyzed by the Bendheim-Thoman Center for Research on Wellbeing or the Columbia Population Research Center. Genotypes were obtained by gel electrophoresis (CRCW, 2015). Genotypes at DRD4 contains a 48bp repeat polymorphism in exon III of chromosome 11 (Dreber et al., 2009). Table 11 shows how the alleles were coded. 1,533 focal children in the current sample (n = 2062) had genetic data available. 86 Previous research has shown that the distribution of DRD4 allele variants differs across ethnic groups (Chang et al., 1996). Across populations, the 2-, 4-, and 7-repeat alleles are most common (Wang et al., 2004). Between Caucasian, Asian, and African American/Black populations, the 4-repeat allele is most common and is considered the non-susceptibility/non-risk variant. The DRD4 susceptibility variant differs across populations. Further evidence suggests that 7+ repeat alleles are associated with decreased efficiency during dopamine bindings compared to 4-repeat alleles (Asghari et al., 1996; Wang et al., 2004). Thus, 7-repeat and higher alleles (i.e., 7-, 8-, 9-, 10- repeats) were coded as the risk/susceptibility variant, DRD4 long (coded as 1), and all other alleles (i.e., 3-, 4-, 5-, 6- repeats) were coded as non-risk/non- susceptibility variants, DRD4 short (coded as 0). Non-risk/susceptibility variant, DRD4 short, was recorded for 901 children (58.8%) and risk/susceptibility, DRD4 long, was recorded for 632 children (41.2%). The distribution of DRD4 alleles is commensurate with other studies investigating DRD4. Table 13. DRD4 Allele Frequency DRD4 Status DRD4 short DRD4 long n= 1533 901 (58.8%) 632 (41.2%) Sustained attention. Children’s sustained attention was measured using indicators from the Leiter International Performance Scale-Revised (Leiter-R) Attention Sustained subtest. The Leiter International Performance Scale-Revised (Leiter-R) (Roid & Miller, 1997) Attention Sustained subtest was used to measure sustained attention during the wave five home visit. The Leiter-R sustained attention test has been used as a measure of sustained attention for pre-school and kindergarten students in several studies (e.g., Domitrovich, Cortes, & Greenberg, 2007; Jensen et al., 2010). The Attention Sustained measure is a cancellation test. FFCWS examiners 87 showed children a page with a target picture on the top and pictures of non-target and target pictures on the bottom. Children were instructed to cross out as many target pictures as possible without crossing out non-target pictures. There were four trials to the test. The first three trials lasted 30 seconds and the last trial lasted 60 seconds. Two scores were created to represent sustained attention: focused attention and lack of impulsivity. The number of cross-outs of objects matching the target reflected focused attention and the number of cross-outs of non-target pictures was reverse-scored to reflect lack of impulsivity. Scores were standardized against a national norming sample (M = 10, SD = 3). The task has acceptable internal reliability (α = .83) and good test-retest reliability (r = .85) for children between four and five (Roid & Miller, 1997). The FFCW researchers derived scores (M = 10, SD = 3) from raw scores recorded at test time using the Leiter-R 1997 norms. Income. To better understand how income level affects children’s reading achievement, a variable indicating whether the family was in poverty at wave three was created by the Fragile Families study. Wave 3 income data allowed the study to make time-ordered mediational and longitudinal inferences from the SEM analyses (Tate, 2015). It allowed this study to measure family income during children’s early development years, as poverty during the early years have been shown to be more influential to children’s development than poverty during later development (Dickerson & Popli, 2016; Kiernan & Mensah, 2009; Raver, Blair, & Willoughby, 2012). A categorical variable for income also accounts for family size. The variable was coded categorically, and a categorical variable was used in the analyses. Categorical data was coded according to five categorical values representing household income established by the U.S. Department of Health and Human Services as:1= 0–49%, 2 = 50–99%, 3= 100–199%, 4 =200–299%, and 5 = 300%. The FFCWS measured income by dividing total 88 household income in the prior 12 months by the official poverty threshold for the year in which the interview was conducted. Wave 3 data was collected between 2001 and 2003. Thus, in 2001, this was $8,590 for a family of one, $11,610 for a family of two, $14,630 for a family of three, and $17, 650 for a family of four. In 2002, this was $8,860 for a family of one, $11,940 for a family of two, $15,020 for a family of three, and $18,100 for a family of four. In 2003, this was $8,980 for a family of one, $12,120 for a family of two, $15,260 for a family of three, and $18,400 for a family of four (U.S. Department of Health and Human Services). Covariates. The following covariates were used to test the proposed models: mothers’ marital status, mothers' age, mothers' education, mothers' race/ethnicity, children's age at assessment, children's language estimate, children's gender. These variables were included in the model to control for their differential influence on the early home literacy environment, sustained attention, DRD4, reading outcomes, and income. Mothers’ marital status. A variable indicating mothers’ marital status at wave five was created by FFCWS. The variable was coded as married, cohabitating in a relationship, not cohabitating in a relationship, separate / widowed / divorced, and friends. Mothers’ marital status at wave five was included as a control variable in the analyses because of its use as a raking variable in the creation of the FFCWS weights (see section on weights for full explanation). Mothers’ age. The FFCWS examiner recorded mothers’ ages at wave five data collection. In the model that tests kindergarten letter-word reading outcomes (cross sectional model), the age of mothers at the wave five home visit was used. In the model that tests third grade passage comprehension reading outcomes (longitudinal model), the age of mothers at the wave nine home visit was used. Mothers’ age at the time of data collection was included as a 89 control variable in the analyses because of its use as a raking variable in the creation of the FFCWS weights (see section on weights for full explanation). Mothers’ education. A variable indicating the mothers' level of education at wave five was created by FFCWS. The variable was coded as: less than high school, high school or GED, some college, and bachelor’s degree or higher. Mother’s education has been shown to be related to the variables of interest (Burchinal, Peisner-Feinberg, Pianta, & Howes, 2002; Carneiro, Meghir, & Parey, 2012; Curenton & Justice, 2008). Ethnicity/race. A variable indicating the focal children’s ethnicity from the primary caregiver survey wave five was created by the FFCWS. This composite included the following categories: White, non-Hispanic; Black or African-American; Hispanic, race specified; or Other. Ethnicity/race was included because some studies have found racial differences in academic achievement (Chatterji, 2006; Kainz & Vernon-Feagans, 2007). Gender. Focal children’s gender was categorized by primary caregiver report from the primary caregiver survey from wave five. Sustained attention skills have not been shown to vary according to gender as measured by direct assessments (Levy, 2006); however, disorders that may involve impaired sustained attention have been shown to differ according to gender, as boys are more frequently diagnosed with attention disorders than girls (American Psychiatric Association, 2013). Early reading differences between genders have also been observed, as girls frequently read better than boys (Below, Skinner, Fearrington, & Sorrell, 2010; Logan & Johnston, 2010) Children’s age at reading outcome assessment. Children’s ages during the reading achievement assessments were recorded by the FFCWS. In the model that tests kindergarten letter-word reading outcomes (cross sectional model), the age of the focal child at wave five 90 home visit was used. In the model that tests third grade passage comprehension reading outcomes (longitudinal model), the age of the focal child at the wave nine home visit was used. Children’s age at the time of reading testing was included as a control variable in the analyses because it is important to control for age in order to remove variance related to age-related maturity (i.e., sustained attention skills [DeLuca et al., 2003] and reading skills [Guttentag & Haith, 1978] increase with age). Kindergarten language estimate. A measure of receptive vocabulary from the wave five home visit represents focal children’s language development in kindergarten. The Peabody Picture Vocabulary Test, Revised (PPVT-R; Dunn & Dunn, 1981) has been shown to be a reliable estimate of language development. For children ages 3 to 6, the PPVT-R demonstrates high internal reliability (α = .87) (Dunn & Dunn, 1981). The PPVT requires children to point to one of four pictures that correspond with an orally presented target word. The FFCW researchers derived standardized scores (M = 100, SD = 15) from raw scores recorded at test time using the national norms provided by the PPVT-R test manual. Since early language ability is a large predictor of reading achievement (Scarborough, 2009; Puolakanaho et al., 2009; Storch & Whitehurst, 2002), it is important to control for this potential influence in the study. Third grade language estimate. A measure of receptive vocabulary from the wave nine home visit represents the children’s language development in third grade. The Peabody Picture Vocabulary Test, Third Edition (PPVT-III; Dunn & Dunn, 1997) has been shown to be a reliable estimate of language development. For children ages 3 to 6, the PPVT demonstrates high internal reliability (α = .94) and validity (Williams & Wang, 1997). The PPVT requires children to point to one of four pictures that correspond with an orally presented target word. The FFCW researchers derived standardized scores (M = 100, SD = 15) from raw scores recorded at test 91 time using the national norms provided by the PPVT-III test manual. Since language ability is a large predictor of reading achievement (Scarborough, Neuman, & Dickinson, 2009; Puolakanaho et al., 2009; Storch & Whitehurst, 2002), it is important to control for this potential influence in the study. Statistical Analyses Two statistical programs were used to conduct analyses. Statistical Package for the Social Sciences, version 23 (SPSS; IBM Corp, 2015) was used to clean the data, run descriptives, and conduct preliminary analyses. Data was downloaded from a secure zip file and the final sample was created based on inclusion and exclusion criteria. MPlus student version 7.4 (Muthén & Muthén, 2012) was used to create the latent variable and to model the data in the main analyses. Preliminary analyses. Preliminary data analyses were conducted to better understand the included and excluded data. Descriptive statistics were run on all variables of interest. Graphs were examined to address potential problems with outliers and skewed distributions. Outliers were included in the study given the large sample size, cross sectional and longitudinal study design, and wide range of potential responses on measures of attention, reading achievement, and language (Tabachnick & Fidell, 2007). Correlation matrices were created to screen for multicollinearity (see Table 18). A few predictor variables were highly correlated; the income variable and the descriptive variable “mother’s education” were highly correlated. Given that these two measures are historically group together as a measure of socioeconomic status and important to the current research questions, these variables were retained. Means (continuous), standard deviations (continuous), frequencies (categorical), and percentages (continuous and categorical) were examined. Full descriptive statistics for the weighted and unweighted final sample are on Tables 6 and 7. 92 For the purposes of comparing descriptive statistics only, children and mothers in poor and near-poor households were compared to children and mothers in adequate income households (Table 14). When SEM analyses were conducted, the income variable was categorical to allow variability in the models; however, comparisons were conducted to better understand the sample. To test for significant differences independent samples t-tests (continuous) and chi-square test of independence (categorical) were conducted on the variables. Table 14. Comparison by Income Status (Categorical) Characteristics Below Poverty Above Poverty Total Child Gender Male Female Child Race/Ethnicity White Black Hispanic Other Mother’s Relationship Married Rom Cohab Rom Sep/Wid/Div Friends No Relationship/Unknown Mothers' Education Less than high school High school or equivalent Some college/technical College or above Mother meets CIDI Cut-Off Yes No DRD4 Status Missing DRD4 short DRD4 long N (%) 462 (51.5%) 435 (48.5%) 81 (9%) 579 (64.5%) 221 (24.6%) 16 (1.8%) 102 (11.4%) 183 (20.4%) 73 (8.2%) 14 (1.6%) 220 (24.5%) 282 (31.5%) 507 (56.6%) 353 (28.3%) 126 (14.1%) 9 (1%) 125 (13.9%) 772 (86.1%) 246 (27.4%) 365 (40.7%) 286 (31.9%) 93 N (%) 597 (51.2%) 568 (48.8%) 322 (27.6%) 538 (46.2%) 262 (22.5%) 43 (3.7%) 490 (42.1%) 187 (16.1%) 65 (5.6%) 39 (3.3%) 179 (15.4%) 191 (16.4%) 277 (23.8%) 308 (26.4%) 385 (33%) 195 (16.7%) 111 (9.5%) 1054 (90.5%) 283 (24.3%) 901 (43.7%) 632 (30.6%) Chi-Square (df) .014 (1), p=.907 128.806 (3), p= .000** 258.774 (5), p= .000** 344.254 (3), p =.000** 9.713 (1), p =.002* 6.011 (2), p=.060 Table 15. Comparison by Income Status (Continuous) Variables Below Poverty Means (SD) Above Poverty Mean (SD) Independent Samples t- test (df) Mothers’ Variable Mother’s Age Children’s Variables Age at assessment (K) Age at assessment (3rd) Language Estimate PPVT-R (K) PPVT-III (3rd) Sustained Attention Focused Attention Lack of Impulsivity 25.75 (5.52) 5.29 (2.93) 9.33 (4.20) 88.42 (15.12) 87.48 (12.59) 12.37 (3.30) 9.83 (2.96) 28.92 (6.12) 5.36 (3.05) 9.27 (3.64) 98.10 (14.37) 96.55 (14.01) 13.16 (3.17) 10.34 (2.72) 19.91 (2059), p= .000** .447 (2051), p= .022* 21.74 (1958), p= .000** 2.27 (2052), p= .000** 10.36 (1954), p =.000** 3.891 (1849), p=.000** 3.137 (1849), p=.000** Table 16. Comparison of HLE Indicators by Income (Categorical) Variables Days per week read with child 0 1 2 3 4 5 6 7 Number of book w/ colors None 1-2 3-4 5 or more Number of book w/ numbers None 1-2 3-4 5 or more Number of book w/ songs None 1-2 3-4 Below Poverty Above Poverty Chi-Square (df) N (%) 29 (3.2%) 54 (6%) 86 (9.6%) 162 (18.1%) 108 (12%) 134 (14.9%) 12 (2.1%) 305 (34%) 19 (2.1%) 163 (18.3%) 244 (27.4%) 466 (52.2%) 39 (4.4%) 196 (22.1%) 225 (25.4%) 427 (48.1%) 112 (12.6%) 250 (28.1%) 219 (24.6%) 19.733 (7), p=.006* 52.321 (3), p=.000** 70.944 (3), p=.000** 76.278 (3), p=.000** N (%) 24 (2.1%) 48 (4.1%) 112 (9.6%) 166 (14.2%) 142 (12.2%) 200 (17.2%) 48 (4.1%) 425 (36.5%) 22 (1.9%) 107 (9.2%) 262 (22.6%) 767 (66.2%) 25 (2.2%) 132 (11.4%) 251 (21.7%) 750 (64.8%) 79 (6.8%) 229 (19.8%) 230 (19.9%) 94 Table 16 (cont’d) 5 or more Number of book w/ alphabet None 1-2 3-4 5 or more How many books in home None 1-10 11-20 More than 20 How often encourage read Less than 1x/month About 1x/month Few times/month Few times/week Every day Table 17. Correlation Matrix for Final Sample 309 (34.7%) 40 (4.5%) 220 (24.7%) 213 (23.9%) 419 (47%) 4 (0.5%) 123 (13.9%) 121 (13.6%) 640 (72.1%) 38 (4.3%) 19 (2.1%) 120 (13.6%) 313 (35.4%) 395 (44.6%) 617 (53.4%) 24 (2.1%) 144 (12.4%) 248 (21.4%) 742 (64.1) 1 (0.1%) 178 (8.7%) 222 (10.9%) 1641 (80.2%) 33 (2.9%) 13 (1.1%) 116 (10%) 380 (32.9%) 613 (53.1%) 79.205 (3), p=.000** 74.665 (3), p=.000** 19.781 (4), p=.001* 95 Regression analyses served as preliminary analyses for the primary SEM analysis to ensure the conditions for mediation and moderation were met. Regressions between each pathway were performed to test the direct contribution of each predictor variable to the reading outcome variable while controlling for mothers’ marital status, mothers' age, mothers' education, mothers' race/ethnicity, children's age at assessment, children's language estimate, children's gender. Regression beta coefficients, slope, and adjusted rsquare were examined to determine the impact of the predictor on the reading outcome variables. Interaction variables for HLE (home literacy environment) X DRD4 were created to test for moderation (hypothesis 3). Structural equation modeling. SEM allows for simultaneous estimation of measurement and structural relations. It also allows testing the relations between observed dependent variables and latent independent variables to determine if a model is consistent with the data from a confirmatory approach (Byrne, 1994). This is appropriate to use to test hypotheses and models developed from the FFCWS non-experimental data. The conceptual models in Figure 4 and 5 display the proposed relationships between exogenous variables (independent variables in SEM), and endogenous variables (dependent variables in SEM) in the cross-sectional model and in the longitudinal model, respectively. Figure 6 shows the full structural model. Latent variables are shown in circles and observed variables are shown in squares. 96 Figure 4. Cross-sectional Conceptual Model Covariates: mothers’ marital status, mothers' age, mothers' education, mothers' race/ethnicity, children's age at assessment, children's language estimate, children's gender. Figure 5. Longitudinal Conceptual Model Covariates: mothers’ marital status, mothers' age, mothers' education, mothers' race/ethnicity, children's age at assessment, children's language estimate, children's gender. Figure 6. Exploratory Conceptual Model Covariates: mothers’ marital status, mothers' age, mothers' education, mothers' race/ethnicity, children's age at assessment, children's language estimate, children's gender. 97 SEM estimation and model fit. Model estimation refers to describing the parameter estimates of the model, evaluating the hypothesized model’s fit to the observed data, and considering equivalent or near equivalent models (Kline, 2016). Model estimation was determined with full information maximum likelihood (FIML). FIML estimation assumes that the model is correctly specified based on theory, the data observations are independent, the endogenous variables are normally distributed, and the exogenous variables are normally distributed. Model fit evaluation and equivalent models are described in the following sections. The fit between the hypothesized model and the observed data were determined by evaluating absolute fit. Absolute fit indices measure the hypothesized model's covariance matrix and the sample covariance matrix and compare the two covariance matrices (Lei & Wu, 2007). Absolute fit indices produce statistics that represent the comparison between hypothesized model's covariance matrix and the sample data’s covariance matrix. These statistics can be evaluated to determine whether the hypothesized model actually fits the sample data. To assess the goodness-of-fit in SEM, four fit indices were used: chi-square, Root Mean Square Error of Approximation (RMSEA), comparative fit index (CFI), and Tucker-Lewis index (TLI). For SEM, chi-square significance testing calculates the chi-square difference statistic between the constrained and unconstrained models (Hooper, Couglan, & Mullen, 2008). If significant, this index suggests that the model parameters are unequal across groups, if not significant, this index suggests that the model parameters are equal across groups. If there is good model fit, the chi-square statistic is not significant (p > .05) (Hooper et al., 2008). The comparative fit index (CFI) and Tucker-Lewis Index (TLI) compares the hypothesized model with a null model, and its values fall between 0 and 1. A CFI or TLI value greater than .95 indicate good fit, it is suggested that the statistic is at least .91 to ensure that incorrect models are 98 not accepted (Hu & Bentler, 1999). The Standardized Root Mean Squared Residual (RMSEA) statistic indicates the difference between the observed correlation matrix and the predicted correlation matrix. A RMSEA statistic of 0 indicates perfect fit, an RMSEA of .05 or less is considered good fit, and a RMSEA between .05 and .08 suggests a reasonable fit (Browne & Cudeck, 1992). Path coefficients. After model fit was established, the paths between variables were examined for significance using standardized coefficients, unstandardized coefficients, slope, and standard error. Direct paths. Direct paths between predictor variables and outcome variables were used to answer hypothesis 1a and 2a, “Income, the early home literacy environment, maternal depression, and sustained attention would be directly associated with reading outcomes (letter- word identification and passage comprehension).” Mediation. Mediation is an effect that one variable causes changes in another variable, which in turn leads to changes in the outcome variable (Little, 2013; Kline, 2016). The intervening variable is the mediator. The change effects are conducted from one predictor variable, through the mediator, to the final outcome variable. Paths were examined to answer research questions 1b and 2b, “The early home literacy environment would mediate the association between income and reading outcomes. Maternal depression would mediate the association between income and reading outcomes. Sustained attention would mediate the association between income and reading outcomes.” The significance levels of these hypothesized indirect effects were calculated with bootstrapping in Mplus and Sobel tests in MPlus. Bootstrapping is a method for deriving sampling distributions empirically by repeated sampling with replacement from the data (Kline, 2016). Repeated sampling is conducted at least 99 1,000 times by the software to correct for bias. Next, confidence intervals are checked at the .05 level. If 0 falls between the confidence intervals, the mediation path is not significant. Sobel’s (1982) significance test tests for the direct and indict associations of the independent variable (income) with the dependent variable (reading outcomes) via the mediators (home literacy, maternal depression, and sustained attention). Moderation. Moderation is when effects are conditional on another variable. For example, the effects of a predictor variable on an outcome variable are conditional on the level (high score, medium score, or low score) of the moderator variable. Moderation is typically evaluated by created an interaction term with the predictor and moderator variables and examining the slopes for significance (Kline, 2016; Preacher, Rucker, & Hayes, 2007). Paths were examined to answer research questions 3, “DRD4 would moderate the direct relations between home literacy environments and passage comprehension reading outcomes.” Moderated mediation. Mediation and moderation are commonly examined using SEM. Models that examine mediation and moderation together are referred to as mediated moderation or moderated mediation (e.g., Baron & Kenny, 1986; Preacher, Rucker, & Hayes, 2007; Muller, Judd, & Yzerbyt, 2005). The general term “conditional indirect effects,” encompasses both moderated mediation and mediated moderation. Preacher, Rucker, & Hayes (2007) define conditional indirect effects as, “the magnitude of an indirect effect at a particular value of a moderator (or at particular values of more than one moderator) (pp. 186).” The following conditional indirect effects model from Hayes PROCESS (2012) was used to hypothesize the relations between income, home literacy environments, maternal depression, sustained attention, DRD4, and reading outcomes (Figure 6). Moderated mediation involves a moderator variable 100 (DRD4), which moderates the relationship between the mediator (home literacy environment) and the dependent variable (passage comprehension reading outcomes) (research question 3). Figure 7. Moderated-Mediation Conceptual Model Figure 8. Moderated-Mediation Statistical Model with Paths X is the predictor, M1, M2, M3, and M4 are the mediators, V is the moderator, M2V is the interaction term, and Y is the outcome (Model 73; Hayes PROCESS, 2012). Equivalent and near equivalent models. Equivalent models explain the same data as well as the current study’s model but have a different pattern of causal effects among the same variables (Kline, 2016). Near equivalent models fit similarly to the current study’s model but have different paths. Both equivalent and near-equivalent models fit similarly to the hypothesized models because they (ideally) have identical or near-identical fit indices, 101 correlation, and covariance matrices. The following near -equivalent model was tested against the final model (Figure 6). This near-equivalent model hypothesizes that the home literacy environment would better be explained by combining investments and family stress variables into a single latent variable. Figure 9. Near Equivalent Conceptual Model 102 CHAPTER 4 RESULTS Preliminary Analyses Means and standard deviations of predictors, covariates, and outcomes variables for the study sample are displayed on Table 6, 7, 8. Unweighted descriptive statistics represent the FFCWS population where each case was counted equally. Weighted descriptive statistics represent the entire population of children born in large cities between 1998 and 2000 to primarily single-parent households. Descriptive analyses were conducted with the weighted and unweighted samples. SEM analyses were conducted with the unweighted sample. Although descriptive analyses with the weighted sample produced appropriately scaled statistics, the SEM analyses with the weighted sample did not. Weights were not used with SEM analyses because it was determined that analyses conducted with weights overestimated the Fragile Families population (Carlson, 2008). The overestimation was determined to be due to the complex sampling design, unequal selection probabilities, response rates across cities, hospitals, and births, and lack of availability of longitudinal weights3 (Carlson, 2008). After consulting with researchers from Fragile Families Child and Family Wellbeing Study, the present study added control variables to the model as an alternative to using weights in SEM analyses. The variables added to the analyses were informed by the procedure the FFCWS used to create the weights, known as raking. Raking is a method of stratification to ensure the weighted counts of the sample are consistent with population counts informed by the 3. Weights were not longitudinal. The Fragile Families guidebook suggested applying the weight that corresponded to the year with the most variables (Carlson, 2008). 103 United States Census between 1998 and 2000 (Carlson, 2008). The raking variables were: marital status, mothers' education level, mothers' race/ethnicity, and mothers' age. Consequently, the variables the FFCWS researchers recommended adding to the present models were: marital status, mothers' age, mothers' education, and mothers' race/ethnicity. The final covariate list includes marital status, mothers' age, mothers' education, mothers' race/ethnicity, children's age at assessment, children's language estimate, children's gender. All covariates were retained in the analyses regardless of significance in the models per the recommendation of the FFCWS researchers. Analyzing the Structural Model Before running the structural equation model analyses, confirmatory factor analysis (CFA) was conducted to test the measurement model. The model had two latent factors, home literacy environment, informed by seven indicators, and maternal depression, indicated by nine observed variables. The two-factor CFA fit the data well. Unweighted estimates included, X2 = 91.74, p=0.0105, df = 63, RMSEA = 0.021, lower bound of 90% CI = 0.010, CFI = 0.999, TLI = 0.999, and WRMR= 0.810. Hu and Bentler (1999) suggest that RMSEA < 0.06, TLI > 0.95, CFI > 0.95 describe well fitting models. The factor loadings, modification indices, fit indices, and theoretical consideration were examined (Schreiber, Stage, King, Nora, & Barlow, 2006), and no modifications were deemed necessary. The indicators' standardized factor loadings were above 0.6 with the exception of two indicators in the HLE latent variable, days per week parents read to their child and how often parents encourage their children to read. These indicators were retained because of their theoretical contribution to the latent factor. The standardized factor estimate, R- square value, and p-value are reported in Table 18. 104 Table 18. Final CFA for a Two-Factor Model Unstandardized Standard Error Standardized R2 Factor Loadings Maternal Depression Losing interest 1.000 Feeling tired 1.048 Change in weight 1.006 Trouble sleeping 1.029 Trouble concentrating 1.024 Feeling down/worthless 1.026 0.937 Thinking about death HLE 1.000 Colors 2.453 Numbers Rhymes/songs 2.984 2.515 Alphabet 3.012 # books in the home Often Read 1.970 0.832 Encourage 0.000 0.009 0.013 0.011 0.013 0.011 0.017 0.000 0.288 0.351 0.301 0.353 0.246 0.138 0.963 1.009 1.011 0.969 0.990 0.986 0.909 0.291 0.714 0.869 0.732 0.877 0.573 0.242 0.935 0.898 0.927 0.939 0.982 0.973 0.826 0.510 0.754 0.536 0.769 0.329 0.085 0.059 P value 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Two methods for analyzing structural equation models using unweighted, complex sample data were used in the study. Research questions 1 and 2 used a bootstrapping method with a maximum likelihood estimator (ML), which allowed for bootstrapping procedures that offer non-symmetric confidence intervals. This was useful for parameter estimates that have non- normal sampling distributions, such as for mediation (indirect effects). Unfortunately, this method did not allow for analysis of random effects that were necessary for testing moderation hypotheses. Therefore, research question 3 used a second method that took random effects into account, and could test for moderation using an interaction term method with a maximum likelihood estimator. This approach allowed for the creation of an interaction term essential for testing research question 3 to determine differential susceptibility or diathesis stress. It also allowed the interaction term to be created with a latent variable (HLE) and a binary variable 105 (DRD4). Despite these advantages, there are several limitations to the method. Bootstrapping was unavailable, and indirect effects were estimated individually (Muthen & Muthen, 2015). Additionally, traditional absolute and relative fit statistics (e.g., chi-square, RMSEA, CFI, and TLI) were unavailable. Instead, Akaike Information Criterion (AIC) and Bayesian Information Criterion (BIC) predictive fit indices were provided. Research Question One What are the direct and indirect associations between income, early home literacy environments, early maternal depression, early sustained attention, and letter-word identification in kindergarten? Research question one sought to test the direct and indirect associations between income, early home literacy environments, early maternal depression, early sustained attention, and letter-word identification in kindergarten using children in a fragile sample (n=2,062) using SEM. Lower income categories, lower home literacy environments, lower sustained attention, and higher maternal depression were hypothesized to be associated with lower reading scores. Conversely, higher income categories, higher home literacy environments, higher sustained attention, and lower maternal depression were hypothesized to be associated with higher reading scores (1a). It was hypothesized that families with lower income would have homes with lower literacy environments, which would be associated with lower reading scores. Families with lower income were expected to have children with lower sustained attention development and, therefore, lower reading scores. It was also expected that families with lower income would have mothers with more endorsements of depression, and mothers with more depression would have children with lower reading scores. Conversely, families with higher income were hypothesized to have home environments with more literacy and, therefore, children with better developed reading. Families with higher income were expected to have 106 children with better developed sustained attention, which would be associated with better developed reading. Families with higher income were expected to have mothers with less endorsements of depression, and less depressed mothers would have children with higher reading scores (1b). Testing the cross-sectional structural model. SEM was conducted to test the overall model and the data were unweighted. All covariates were retained in the analyses regardless of significance to account for the removal of weights per the recommendations of the FFCWS researchers. Table 19 displays the results of the structural model. The hypothesized cross- sectional model fit the data well. Unweighted estimates included, X2 = 207.707, p=0.003, df = 155, RMSEA = 0.018, lower bound of 90% CI = 0.011, CFI = 0.999, TLI = .999, and WRMR= 0.848 Direct 1a. As hypothesized, categories of income were positively associated with children's letter-word reading scores, (B= 0.617, p<0.05), suggesting that children from families with more income had better developed letter-word reading. Sustained attention was positively associated with children's letter word identification scores, (B= 0.650, p<0.00), suggesting that children with better developed sustained attention also had better developed letter-word reading. Higher endorsements on the home literacy environment measure were positively associated with letter-word identification scores, (B= 2.750, p<0.05), suggesting that children who had more exposure to literacy in the home environment had higher letter-word reading scores. Contrary to the study's hypothesis, mothers' ratings of depressive symptoms were not linked to their children's letter-word identification scores, (B= -0.870, p=0.116). The estimated model predicted 16% of the variance in sustained attention, 20% of the variance in the home literacy 107 environment, 3% of the variance in maternal depression, 26% of the variance in letter word reading outcome. Indirect 1b. Sobel test statistics and bootstrapping (confidence intervals) are reported for indirect effects. Although direct effects for the association between income category and letter- word identification scores were significant (B= 0.752, p<0.05, lower CI= 0.116, upper CI = 1.410), total indirect effects were not (B= 0.135, p=0.226, lower CI= -0.082, upper CI= 0.359). Specifically, the mediating association between income category, sustained attention, and letter- word identification was not significant (B= 0.096, p=0.104, lower CI = -0.016, upper CI= 0.225). The mediating association between income category, maternal depression, and letter-word identification score was not significant (B= -0.062, p=0.269, upper CI= -0.204, upper CI= 0.022). The only significant mediating association was between income category, home literacy environment, and letter-word identification scores (B= 0.101, p<0.05, lower CI= 0.012, upper CI= 0.274). This finding suggests families with more income had homes with higher literacy environments, and higher home literacy environments were linked with children's letter-word identification scores in kindergarten. Specifically, functioning through the home literacy environment, increases in income representing 50-100% of the poverty line were associated with less than one point increase on the letter-word measure. Together, the significance of income and the early home environment through direct and indirect analyses illustrates the importance of family income and literacy materials in the home for children’s development of early reading skills. Influence of covariates. Covariates were included in the model to control for their effect on letter-word identification scores, home literacy environment, maternal depression, income, and sustained attention. Additionally, covariates were included to accommodate the FFCWS 108 sampling procedure (see section on weights). Covariates used to rake weights (mothers’ marital status, education, race/ethnicity, and mothers’ age) were retained regardless of path significance per the FFCWS researchers’ suggestion. The final covariate list was: mothers' marital status, maternal depression, mothers' education, mothers' age, race/ethnicity, children's age at assessment, and children's language estimate. Mothers’ marital status was not associated with the amount of literacy in the homes (B= - 0.002, p=0.562), nor was it linked with children’s letter-word reading (B= 0.095, p=0.467) nor with sustained attention (B= -0.003, p=0.936). Mothers' marital status was not associated with their ratings of depressive symptoms (B= 0.001, p=0.953). Mothers with more education had homes with better literacy environments (B= 0.046, p<0.05) and children whose letter-word reading (B= 2.478, p<0.05) was better developed. Mothers education level was not related to their children’s sustained attention (B= 0.121, p=0.351). Mothers who were older had homes with better literacy environments (B= -0.005, p<0.05) and had children whose sustained attention was better developed (B= -0.046, p<0.05); however, mothers’ ages were not associated with their own depressive symptoms (B= -0.003, p= 0.680) or their children’s letter-word identification scores (B= -0.022, p=0.752). Race/ethnicity was associated with mothers’ endorsements of depressive symptoms (B= -0.097, p<0.05), the amount of literacy in the home (B= -0.220, p<0.05), children’s scores on measures of letter-word reading (B= 0.734, p<0.05) and sustained attention (B= 0.453, p<0.005). Children’s ages at the time of assessment did not influence the amount literacy in their homes (B= 0.001, p=0.859) or performance on a sustained attention (B= 0.064, p=0.066) or their mothers’ symptoms of depression (B= 0.028, p=0.65); however, children who were older were more likely to have higher scores on measures letter-word reading (B= 0.293, p<0.05). Children 109 with better language development also had better developed letter-word reading (B= 0.259, p<0.000) and sustained attention (B= 0.069, p<0.000), and were more likely to come from homes with better literacy environments (B= 0.004, p<0.000). However, children's language estimates were not associated with their mother's depressive symptoms (B= -0.001, p=0.633). Children’s gender significantly predicted several outcomes, as girls were more likely to have higher exposure to literacy in their homes (B= 0.036, p<0.05), and have more developed letter-word reading (B= 2.478, p<0.05) and sustained attention (B= 0.909, p<0.000) than boys. Children's gender was not associated with their mothers’ depressive symptoms (B= 0.026, p=0.781). Table 19. RQ1 Parameter Estimates for Direct and Indirect Effects Unstandardized S.E. Factor Loadings Standardized R2 P-value 0.000 0.460 0.576 0.469 0.577 0.325 0.174 0.000 0.013 0.013 0.015 0.014 0.011 0.012 0.276 0.731 0.873 0.735 0.881 0.541 0.213 0.930 0.978 1.001 0.988 0.989 0.982 0.967 0.201 0.076 0.00 0.535 0.00 0.763 0.00 0.540 0.00 0.777 0.00 0.293 0.00 0.045 0.00 0.021 0.865 0.00 0.956 0.00 0.912 0.00 0.976 0.00 0.978 0.00 0.976 0.00 0.958 0.00 1.000 2.785 3.415 2.799 3.451 2.006 0.768 1.000 1.053 1.078 1.063 1.067 1.078 1.060 HLE Often read Colors Numbers Rhymes/songs Alphabet # books in the house Encourage Maternal Depression Losing interest Feeling tired Change in weight Trouble sleeping Trouble concentrating Feeling down/worthless Thinking about death Incomeà LWID S. Attention à LWID M. Depression à LWID HLE à LWID Direct Effects Unstandardized S.E. 0.617 0.323 0.127 0.650 0.554 -0.870 2.750 1.884 Indirect Effects Unstandardized S.E. 110 Standardized R2 0.064 -- -- 0.160 -- 0.062 0.058 -- P-value 0.050 0.000 0.116 0.044 Standardized R2 P-value Table 19 (cont’d) Total Total Indirect Income à sustained attentionà LWID Income à maternal depression à LWID Income à HLE à LWID 0.752 0.135 0.096 -0.062 0.101 0.322 0.112 0.059 0.056 0.070 0.078 0.014 0.010 -0.006 0.011 -- -- -- -- -- 0.019 0.226 0.104 0.269 0.046 Figure 10. Cross-sectional Model with Unstandardized Estimates Research Question Two What are the direct and indirect associations between early income, early home literacy environments, early maternal depression, early sustained attention, and passage comprehension in third grade? The second research question tested the longitudinal direct and indirect associations between income, early home literacy environments, early maternal depression, early sustained attention, and third-grade passage comprehension using the FFCWS (n=2,062). Lower income categories, lower home literacy environments, lower sustained attention, and higher maternal depression were hypothesized to be associated with lower passage comprehension reading scores. Conversely, higher income categories, higher home literacy environments, higher sustained attention, and lower maternal depression were hypothesized to be associated with 111 higher passage comprehension reading scores (2a). Families with lower income were expected to have homes with lower literacy environments, which would be associated with lower reading scores. Families with lower income would have children with lower sustained attention development and, therefore, lower reading scores. Families with lower income would have mothers with more endorsements of depression, and mothers with more depression would have children with lower reading scores. Conversely, families with higher income would have home environments with more literacy and, therefore, children with better developed reading. Families with higher income would have children with better developed sustained attention, which would be associated with better developed reading. Families with higher income would have mothers with less endorsements of depression, and less depressed mothers would have children with higher reading scores (2b). Testing the longitudinal structural model. SEM was conducted to test the overall model and the data were unweighted. All covariates were retained in the analyses regardless of significance to account for the removal of weights per the recommendation of the FFCWS researchers. Table 20 displays the results of the structural model. The hypothesized longitudinal model fit the data well. Unweighted estimates included, X2 = 226.774, df = 165, p = 0.001, CFI = 0.998, TLI= 0.998, RMSEA = 0.020, CI 90% lower = 0.013, and WRMR = 0.851. The estimated model predicted 10% of the variance in sustained attention, 19% of the variance in the home literacy environment, 3% of the variance in maternal depression, 21% of the variance in letter word reading, and 44% of the variance in passage comprehension reading outcomes. Direct 2a. Contrary to the hypotheses, categories of income were not associated with children’s passage comprehension scores, (B= 0.019, p=0.943). Nor was the home literacy (B= 0.863, p=0.500), or maternal depression (B= -0.627, p=0.369) associated with children’s passage 112 comprehension scores. Sustained attention was the only variable significantly associated with children’s passage comprehension scores, (B= 0.238, p<0.05), such that children with better sustained attention in kindergarten had higher passage comprehension scores in third grade. Indirect 2b. Sobel test statistics and bootstrapping (confidence intervals) are reported for indirect effects. Although the total direct effects between income and passage comprehension were not significant, the total indirect effects between income category and passage comprehension scores were significant (B= 0.219, p<0.05, lower CI= 0.052, upper CI= 0.428 ). Specifically, the mediating association between income category, sustained attention, letter-word identification, and passage comprehension was not significant (B= 0.024, p=0.065, lower CI = 0.000, upper CI= 0.052). The mediating association between income category, maternal depression, letter-word identification, and passage comprehension was not significant (B= - 0.008, p=0.463, lower CI= -0.039, upper CI= 0.006). The only significant mediating association was between income category, home literacy environment, letter-word identification, and passage comprehension, (B= 0.030, p<0.05, lower CI= 0.008, upper CI= 0.058). This finding indicates families with higher income had homes with better early literacy environments, and higher early literacy environments improved children's letter-word reading in kindergarten. Better letter-word reading in kindergarten led to better passage comprehension in third grade. Influence of covariates. Covariates were included in the model to control for their effect on letter-word identification and passage comprehension scores, home literacy environment, maternal depression, income, and sustained attention. Additionally, covariates were included to accommodate the FFCWS sampling procedure (see section on weights). Covariates were freely allowed to predict all variables. Again, covariates that were used to rake weights (mothers' marital status, education, race/ethnicity, and age) were retained regardless of path significance 113 per the FFCWS researchers’ suggestion. The final covariate list was: mother's marital status, maternal depression, mothers' education, mothers' age, race/ethnicity, children's age at assessment, and children's language estimate. Mothers' marital status was not associated with the amount of literacy in their homes (B= 0.001, p=0.679) or their endorsements of depressive symptoms (B= 0.015, p=0.430), nor was mothers' marital status associated with their children's passage comprehension (B= -0.095, p=0.319) or early sustained attention (B= 0.019, p=0.660). Better educated mothers had better early home literacy environments (B= 0.053, p<0.05) and children with better early sustained attention (B= 0.240, p<0.05). Children's reading comprehension was not influenced by their mothers' education (B= 0.386, p=0.284). Maternal education was unrelated to maternal depression (B= 0.072, p= 0.245). Older mothers had homes with higher literacy environments (B= 0.005, p<0.05) and children with better early sustained attention (B= 0.05, p<0.05); however, mothers’ ages did not predict their children’s passage comprehension (B= 0.016, p=0.764). Mothers' ages did not influence their ratings of depressive symptoms (B= 0.001, p=0.924). Race/ethnicity influenced mothers’ ratings of depression (B= -0.027, p<0.05), the quality of the early literacy environment in the home (B= -0.038, p<0.05), as well as children’s development of sustained attention (B= 0.232, p<0.05); however, race/ethnicity was unrelated to children’s passage comprehension (B= -0.101, p=0.792). Children’s age at assessment was related to their passage comprehension scores, as older children scored higher than younger children on the passage comprehension measure (B= 0.133, p<0.05). Children’s age at assessment was unrelated to the amount of literacy in their early home environment (B= 0.001, p=0.847), their mother’s early ratings of depression (B=0.006, p=0.595), or their sustained attention abilities (B= 0.009, p=0.753). 114 Children with better language development at nine years of age had more literacy in their homes when they were young (B= 0.004, p<0.000). Children with better language skills also had better passage comprehension (B= 0.420, p<0.000) and early sustained attention skills (B= 0.044, p<0.000); however, children’s language was unrelated to their mothers’ early ratings of depression (B= 0.001, p=0.711). Children's gender was also significant, as girls had more early literacy in their homes (B= 0.032, p<0.05), had better developed passage comprehension (B= 1.853, p<0.05), and early sustained attention skills (B= 1.120, p<0.000) than boys; however, children’s gender was unrelated to their mothers’ early endorsements of depressive symptoms (B= -0.041, p=0.656). Table 20. RQ2 Parameter Estimates for Direct and Indirect Effects Unstandardized S.E. Standardized R2 0.290 0.739 0.889 0.716 0.859 0.577 0.228 0.010 0.008 0.008 0.006 0.006 0.016 0.009 0.190 0.084 0.546 0.790 0.513 0.738 0.333 0.052 0.023 0.925 0.951 0.987 0.838 0.987 0.982 0.875 Standardized R2 -- 0.005 0.067 -- -- -0.053 0.022 -- P-value 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 P-value 0.943 0.031 0.369 0.500 HLE Often read Colors Numbers Rhymes/songs Alphabet # books in the house Encourage Maternal Depression Losing interest Feeling tired Change in weight Trouble sleeping Trouble concentrating Feeling down/worthless Thinking about death Incomeà PC S. Attention à PC M. Depression à PC HLE à PC Factor Loadings 0.000 0.363 0.448 0.359 0.428 0.292 0.784 0.000 0.013 0.013 0.012 0.018 0.017 0.016 1.000 2.668 3.295 2.575 3.167 2.039 0.784 1.000 1.014 1.034 1.027 0.951 0.971 0.949 Direct Effects Unstandardized S.E. 0.268 0.019 0.238 0.110 0.414 -0.627 0.863 1.281 115 Table 20 (cont’d) Total Total Indirect Income à sustained attentionà LWID à PC Income à maternal depression à LWIDà PC Income à HLE à LWIDà PC Indirect Effects Unstandardized S.E. 0.259 0.219 0.094 0.200 0.024 0.013 Standardized R2 -- 0.026 -- 0.024 0.003 -- -0.008 0.030 0.011 0.012 -0.001 0.003 -- -- P-value 0.398 0.034 0.065 0.463 0.034 Figure 11. Longitudinal Model with Unstandardized Estimates Research Question Three Research question three is exploratory given the limited evidence to suggest that DRD4 is associated with the variables of interest. Does DRD4 influence the associations between income, early home literacy environment, and reading outcomes? The third research question aimed to determine whether the mediation model would differ by the presence (n= 625) or the absence (n= 892) of the DRD4 long allele (moderation) (total n =1,517). The hypotheses for this research question differentiated between children carrying a short ( 6- repeat) or long (7+ repeat) DRD4 allele. Specifically, this question examined whether the association between income, early home literacy environment, sustained attention, maternal depression, and passage comprehension 116 differs due to DRD4. This is known as mediated-moderation, which is when mediated effects are altered at different values of the moderator. Two theories were hypothesized. One, the diathesis-stress model postulated the DRD4 long allele is a risk allele. This hypothesis suggests that despite high early literacy exposure in the home, children with genetic risk (DRD4 long) would have poorer reading scores than their counterparts without genetic risk and high early literacy exposure. In other words, children with genetic risk would have poorer reading outcomes regardless of their home literacy environment. Two, the differential susceptibility hypothesis postulated that the DRD4 long allele is a sensitive allele (Kegel, Bus, & van Ijzendoom, 2011). Children who have the DRD4 long allele and who have experienced high home literacy environment would have higher reading scores than children with DRD4 long allele and low early home literacy environments. In other words, children with the DRD4 long allele would have reading outcomes scores that reflect the quality of their early home literacy environment. For children with the DRD4 short allele, their letter- word reading scores would not significantly benefit or suffer due to the effects of the home literacy environment. Analysis plan. There are two methods for estimating mediated-moderation effects in SEM. The multiple group method compares one model that constrains factor loadings and intercepts of observed indicators to be equal across groups against a second model that allows the factor loadings and intercepts observed indicators to vary across groups (Kline, 2015). The estimates between the two models are compared to determine if a freely estimated model fits better/worse than a constrained model. If the results of the freely estimated model are better, differences between each groups’ freely estimated model can be compared. Differences between the parameter estimates and factors loadings between the groups would indicate significant 117 differences between the group with the DRD4 long allele and the group with the DRD4 short allele. The second method creates an interaction term between the DRD4 variable (binary present or not present) and the home literacy environment variable. The estimate of the interaction term is interpreted as a variable in the structural equation model. If significant (p<0.05), the slopes of the interaction term and the reading outcome are plotted and compared to determine differential susceptibility or diathesis stress. Belsky, Bakermans-Kranenburg, & van IJzendoorn (2007) and Ellis et al., (2011) outline several requirements to consider when determining diathesis-stress or differential susceptibility. First, there must be a crossover interaction; second, the slope of interaction term must significantly different than zero; third, the DRD4 gene should not be significantly related to the HLE (which would indicate bi-directional effect) or to the reading outcome (indicating dual risk). As described previously, the second SEM method was selected to test the diathesis-stress vs. differential susceptibility hypotheses. However, there are two main limitations to the interaction method for estimating moderation. One, bootstrapping was unavailable and, two, traditional fit statistics such as chi-square, RMSEA, CFI, and TLI, were unavailable. Instead, AIC and BIC predictive fit indices were provided to determine the best fitting structural equation model. AIC is an estimate of a constant and the distance between the model’s unknown true likelihood function and the model’s fitted likelihood function (Dziak, Coffman, Lanza, & Li, 2012). BIC is a similar estimate under a Bayesian assumption (Dziak, Coffman, Lanza, & Li, 2012; Raftery, 1995). Both are useful goodness of fit measures when ML estimation is used (Burnham & Anderson, 1998). Lower AIC and BIC estimates mean that a model is more likely to be the true model (Dziak, Coffman, Lanza, & Li, 2012). Multiple models were generated, and 118 models with smaller AIC and BIC values were interpreted as more likely to be replicated in the population where the sample was drawn. Testing the structural model. The structural model identified in research question 2 (longitudinal model) was tested (n=1,517). When analyses were conducted with the full structural model from research question 2, there were problems with conversion due to the letter- word identification variable because it was conceptualized as a secondary mediator. When the letter-word identification variable was removed as the secondary mediator and, instead, conceptualized as a control variable, the model converged. Like the previous SEM models, the data were unweighted. An interaction variable, HLExDRD4, was created to test moderation and was centered before analyses (Cohen, West, & Aiken, 2003). To compare AIC and BIC for best fit, the model without covariates was analyzed first. Next, covariates were added to the model and all covariates were allowed to freely predict the reading comprehension outcome variable. Thereafter, covariates were allowed to freely predict exogenous variables. All covariates were retained in the analyses regardless of significance to account for the removal of weights. To interpret AIC and BIC, Kass and Raftery (1995) and Muthen (2010) suggest that a difference of more than 10 is substantial evidence against the model with the higher value. Accordingly, the model without covariates (AIC= 23590.101, BIC= 23902.101) fit the data best (see Table 21). The estimated model predicted 11% of the variance in sustained attention, 1% of the variance in the interaction term, 11% of the variance in the home literacy environment, 3% of the variance in maternal depression, 20% of the variance in passage comprehension reading outcomes. Figure 12 displays the results of the best fitting structural equation model. 119 Table 21. AIC and BIC Comparisons for Moderation Model Without covariates Covariates constrained Covariates free Log -11728.050 -11381.730 -23820.102 AIC 23590.101 22897.459 44981.987 BIC 23902.101 23207.803 40394.271 Moderation. Research question 3 sought to explore whether the DRD4 moderated the association between home literacy environment and passage comprehension reading outcomes in third grade. Overall, results suggest that DRD4 did not moderate associations between the home literacy environment and reading comprehension outcomes in third grade. Contrary to the hypotheses, the interaction variable HLExDRD4 did not significantly predict reading comprehension in third grade (B= 0.152, p=0.843). As such, the slopes of the interaction term and the reading outcome could not be plotted to determine differential susceptibility or diathesis stress. The direct paths were examined. Homes with higher income had children with better passage comprehension scores (B= 0.772, p<0.05) and better sustained attention scores(B= 0.162, p<0.05). More income in the home was also linked with better home literacy environments (B= 0.247, p<0.05), and mothers with fewer symptoms of maternal depression (B= -0.055, p<0.05). Children with better developed sustained attention also had better passage comprehension (B= 0.390, p<0.05). Children from homes with more early literacy had better developed passage comprehension in third grade (B= 1.161, p<0.05). It is important to highlight that this model did not include covariates; therefore, significant findings must be interpreted cautiously. 120 Figure 12. Moderation Model with Unstandardized Estimates Post Hoc Analyses Post-hoc analyses were conducted to determine whether DRD4 moderated other associations between variables in the structural model specified in research question 2 and to attempt to identify the full longitudinal model identified in research question 2 (with the double mediation). Since post-hoc analyses aimed to examine the differences between children with the DRD4 long allele and children with the DRD4 short allele, the one-factor latent measurement model was tested for group invariance. Group invariance was tested by examining configural, metric, and scalar invariance to compare the models. A configural invariance model with single-factor models was estimated within each group. The factor mean was fixed to 0 and the factor variance was fixed to 1 for identification within the DRD4 long group and the DRD4 short group. The configural model had a good fit, thus a series of model constraints were applied to determine decreases or increases in model fit resulting from measurement invariance. Next, metric invariance model examined the unstandardized item factor loadings across groups. The factor means were fixed to 0 in both groups, but the factor variance was fixed to 1 in the DRD4 long group but was freely estimated 121 in the DRD4 short group. The metric invariance model fit well and did not result in a significant decrease in fit relative to the configural model, − 2 ∆, p>0.05. This suggests that there is "weak invariance," which means the items were related to the latent HLE factor equivalently across groups. Finally, scalar invariance model examined the unstandardized item intercepts across groups. The factor mean was fixed to 0 and the variance was fixed to 1 for DRD4 long, but the factor mean and variance were estimated for DRD4 short. The factor loadings were constrained to be equal across groups and residual variances were allowed to differ across groups. The scalar invariance model fit well and did not result in a significant decrease in fit relative to the metric invariance model, − 2 ∆, p>0.05. Overall, results suggest that the measurement model met requirements for invariance, thereby indicating that HLE can be interpreted equally across the DRD4 long and DRD4 short groups. The structural equation model from research question 2 was tested separately for each group. Results demonstrated that the model fit the data adequately for children with DRD4 long, as unweighted estimates were: X2 = 182.077, df = 270, p = 0.0115, CFI = 0.999, TLI= 0.999, RMSEA= 0.017, CI 90% lower = 0.008 , and WRMR = 1.057. The estimated model predicted 10% of the variance in sustained attention, 16% of the variance in the home literacy environment, 3% of the variance in maternal depression, 23% of the variance in letter word reading, and 37% of the variance in passage comprehension reading outcomes. Results demonstrated that the model fit the data adequately for children with DRD4 short, as unweighted estimates were: X2 = 143.476, df = 270, p = 0.0115, CFI = 0.999, TLI= 0.999, RMSEA= 0.017, CI 90% lower = 0.008 , and WRMR = 1.057. The estimated model predicted 10% of the variance in sustained attention, 16% of the variance in the home literacy environment, 5% of the variance 122 in maternal depression, 21% of the variance in letter word reading, and 47% of the variance in passage comprehension reading outcomes. The same model tested above was used to specify a structural equation model to test for equivalence across samples of children with DRD4 long allele (n= 625) and children without DRD4 short allele (n= 892). Two models were identified. In one model, structural paths and factor loadings were freely estimated. This model fit the data adequately, X2 = 226.774, df = 165, p = 0.001, CFI = 0.998, TLI= 0.998, RMSEA = 0.020, CI 90% lower = 0.013, and WRMR = 0.851. The results of the freely estimated model were compared to a second model. The second model constrained the factor loadings and paths to be equal across groups. The fit of the model worsened when the factor loadings, covariate estimates, and structural paths were constrained to be equal across groups, however, the model fit remained acceptable X2 = 325.553, df = 270, p = 0.0115, CFI = 0.999, TLI= 0.999, RMSEA= 0.017, CI 90% lower = 0.008 , and WRMR = 1.057. Although both models are acceptable, these statistics indicate that recognizing the two groups, DRD4 long and DRD4 short, is more parsimonious than grouping children with DRD4 long and DRD4 short into the same group. Thus, analyzing children with DRD4 long and DRD4 short separately is justified. In the freely estimated models, direct effects were not significant for either group. No differences in significance were observed when comparing the original paths between groups, suggesting no moderation effects. Categories of income were not directly associated with children's passage comprehension scores for the DRD4 long (B= 0.400, p=0.176) or DRD4 short (B= 0.044, p=0.866) groups. Home literacy environment was not associated with passage comprehension for DRD4 long (B= 0.530, p=0.192) or DRD4 short (B= 0.214, p=0.537) groups, nor was maternal depression associated with passage comprehension for DRD4 long (B= - 123 0.417, p=0.423) or DRD4 short (B= -0.460, p=0.313). Sustained attention was not associated with passage comprehension for either group, DRD4 long (B= 0.201, p=0.111) or DRD4 short (B= 0.185, p=0.065.) The direct associations remained consistent between the DRD4 long and DRD4 short groups, suggesting that DRD4 does not moderate the direct associations measured in the original model (see Table 22 and Figure 13). An interesting finding was observed when paths not originally identified in the longitudinal model were examined. The association between early home literacy environment and children's letter-word reading was significant for the DRD4 long group (B= 1.260, p<0.05), but not for the DRD4 short group (B= 0.606, p=0.171). This finding suggests a moderating effect of D4D4 on the association between the early home literacy environment and children’s letter- word reading scores in kindergarten. Children with DRD4 long from homes with higher early home literacy had better letter-word reading than children with DRD4 long from homes with low early home literacy environments. Alternatively, the amount of literacy in the home was not significantly associated with children's letter-word reading for children with DRD4 short. Sobel test statistics and bootstrapping (confidence intervals) are reported for indirect effects. The total indirect effects between income category and passage comprehension scores were not significant for DRD4 long (B= 0.560, p= 0.471, lower CI= 0.026, upper CI= 1.181) or DRD4 short (B= 0.181, p=0.078, lower CI= -0.335, upper CI= 0.609) groups. Specifically, the mediating association between income category, sustained attention, letter-word identification, and passage comprehension was not significant for DRD4 long (B= 0.012, p= 0.339, lower CI = -0.006, upper CI= 0.049) nor DRD4 short (B= 0.020, p=0.389, lower CI = -0.013, upper CI= 0.046). The mediating association between income category, maternal depression, letter-word identification, and passage comprehension was not significant for DRD4 long (B= 0.000, 124 p=0.953, lower CI= -0.044, upper CI= 0.006). It was not significant for the DRD4 short group either (B= 0.006, p= 0.624, lower CI= -0.011, upper CI= 0.035). The only significant mediating association was between income category, home literacy environment, letter-word identification, and passage comprehension for DRD4 long (B= 0.019, p<.05, lower CI= 0.005, upper CI= 0.069) and DRD4 short (B= 0.014, p<.05, lower CI= 0.000, upper CI= 0.053). The indirect associations remained consistent between the DRD4 long and DRD4 short groups, signifying that DRD4 does not moderate the indirect associations (see Table 22). Overall, although a significant moderation effect by DRD4 was observed, a significant moderated-mediation effect by DRD4 was not observed Table 22. Parameter Estimates for Two Group Model Unstandardized Standardized Unstandardized Standardized HLE Often read Colors Numbers Rhymes/songs Alphabet # books in the house Encourage Maternal Depression Losing interest Feeling tired Change in weight Trouble sleeping Trouble concentrating Feeling down Thinking about death DRD4 long 0.239 0.738 0.840 0.662 0.828 0.522 0.368 0.972 0.971 0.990 0.981 0.896 0.872 0.889 0.259 0.753 0.847 0.727 0.888 0.553 0.401 0.963 0.975 1.007 0.997 0.919 0.895 0.901 125 DRD4 short 0.239 0.738 0.840 0.662 0.828 0.522 0.368 0.972 0.971 0.990 0.981 0.896 0.872 0.889 0.340 0.766 0.860 0.693 0.554 0.566 0.393 0.974 0.973 0.999 0.982 0.900 0.887 0.894 Table 22 (cont’d) Income à PC S. Attention à PC Depression à PC HLE à PC HLEà LWID 0.400 (0.176) 0.201 (0.111) -0.417 (0.423) 0.530 (0.192) 1.260 (0.023) 0.560 (0.06) 0.160 (0.06) 0.012 (0.339) Total Total indirect Income à attentionà LWIDà PC Income à depression à LWID à PC Income à HLE à LWIDà PC Direct Effects 0.049 0.059 Indirect Effects -0.038 0.054 0.115 0.069 0.020 0.001 0.044 (0.866) 0.185 (0.065) -0.460 (0.313) 0.214 (0.537) 0.606 (0.171) 0.181 (0.462) 0.136 (0.150) 0.012 (0.389) 0.005 0.052 -0.041 0.020 0.052 0.022 0.017 0.001 0.000 (0.953) 0.000 0.006 (0.624) 0.001 0.019 (0.165) 0.002 0.014 (0.200) 0.002 Figure 13. Two Group Model with Unstandardized Estimates Red= DRD4 long; Black = DRD4 short Alternative Models HLE. To examine whether the HLE variable would be better explained by two separate latent factors a confirmatory factor analysis (CFA) was conducted. Indicators from the HLE variable were separated to represent a materials variable and a parenting variable. Like the 126 original HLE latent variable, questions about literacy investments were selected from the wave five caregiver survey based on commonly used questions to measure the home literacy environment (e.g., Evans & Shaw, 2008, Scarborough & Dobrich, 1994; Sénéchal, 2006). Parent responses on these items were used as indicators of the early home literacy environment for the focal child. The latent home literacy material factor included five indicators: - About how many toys, books or games does (child) have that are helping him/her know about colors? (None, 1-2, 3-4, 5 or more) - About how many toys, books or games does (child) have that are helping him/her know about numbers? (None, 1-2, 3-4, 5 or more) - About how many toys, books or games does (child) have that are helping him/her know about nursery rhymes or songs? (None, 1-2, 3-4, 5 or more) - About how many toys, books or games does (child) have that are helping him/her know about the alphabet? (None, 1-2, 3-4, 5 or more) - How many books are in the home? (None, 1-10 books, 11-20 books, and 20 or more books) The latent home literacy parenting factor included two indicators: - How many days per week do you read to CHILD? (1, 2, 3, 4, 5, 6, 7 days) - How often do you encourage CHILD to read? (less than once a month, about once a month, a few times a month, at least a few times per week, every day) The measurement model fit the data well. Absolute fit indices included : X2= 125.75, p=0.014 and RMSEA= 0.063 (lower CI = 0.053, upper CI= 0.073), and relative fit indices, CFI =0.986, TLI=0.977, WRMR=1.207. Hu and Bentler (1999) suggest that RMSEA < 0.06, TLI > 0.95, CFI > 0.95 describe well fitting models. 127 Table 23. CFA for Materials and Parenting Latent Variables Unstandardized Standard Error HLE Materials Colors Numbers Rhymes/songs Alphabet # books in the home HLE Parent Often Read Encourage 1.000 1.170 0.957 1.186 0.689 1.000 1.735 Factor Loadings 0.000 0.011 0.014 0.010 0.024 0.000 0.023 Standardized R2 0.731 0.847 0.727 0.865 0.584 0.621 0.567 0.208 -- -- -- -- -- 0.102 -- -- P value 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Since the measurement model fit the data well, an alternative structural model was examined. An alternative model with the latent material and parenting factor was examined to determine whether two separate latent variables would result in a more parsimonious cross sectional model. SEM was conducted to test the overall cross sectional model and the data were unweighted. All covariates were retained in the analyses regardless of significance to account for the removal of weights per the recommendations of the FFCWS researchers. The alternative cross-sectional model fit adequately. Unweighted estimates included, X2 = 680.922, p=0.00, df = 74, RMSEA = 0.062, lower bound of 90% CI = 0.058, CFI = 0.920, TLI = 0.866, and WRMR= 2.005. The estimates from the alternative model fit worse than the original model that conceptualized HLE as a single latent construct (X2 = 263.791, p=0.00, df = 81, RMSEA = 0.033, lower bound of 90% CI = 0.029, CFI = 0.975, TLI = 0.962, and WRMR= 1.254). Within the study’s context, this suggests that HLE is better explained with a single variable that combines indicators about material and parenting factors. 128 Similarly, to examine whether the HLE variable would be better explained by two separate latent factors SEM was conducted. The data were unweighted and all covariates were retained in the analyses regardless of significance to account for the removal of weights per the recommendations of the FFCWS researchers. The alternative longitudinal model fit poorly. Unweighted estimates included, X2 = 1717.430, p=0.00, df = 79, RMSEA = 0.101, lower bound of 90% CI = 0.097, CFI = 0.740, TLI = 0.556, and WRMR= 3.118. The estimates from the alternative model fit far worse than the original model that conceptualized HLE as a single latent construct (X2 = 256.156, df = 89, p = 0.00, CFI = 0.978, TLI= 0.964, RMSEA = 0.031, CI 90% lower = 0.027, and WRMR = 1.102). Again, within the study’s context, this suggests that HLE is better explained with a combination of material and parenting factors. Altogether, these results indicate the original latent HLE variable provides a better representation of the home literacy environment than two separate latent variables (materials and parenting) in the cross-sectional and longitudinal models. This suggests that literacy books and materials and parenting behaviors, such as joint reading, should be considered in simultaneously during the early developmental period when promoting children’s literacy. 129 CHAPTER 5 DISCUSSION The objective of this study was to explore the successive interactions and indirect effects of environmental and within child variables that contribute to reading outcomes for at-risk children. The work was informed by Bronfenbrenner’s bioecological model (Bronfenbrenner & Ceci, 1994), prevailing models of children’s reading development (National Reading Council, 2000), and mediational theories on the effects of poverty (Yeung, Linver, & Brooks-Gunn, 2002). Children from poor and near-poor households begin school with at a disadvantage, as they often have less developed early literacy skills than their peers from adequate income households (Duncan & Brooks-Gunn, 2000; Payne, Whitehurst, & Angell, 1994). Although income has been widely established as an influential environmental variable in studies examining children’s reading outcomes, less is known about the mediational pathways by which income influences children’s reading outcomes. By focusing on the indirect and interactive (i.e., moderated) associations among at-risk children and their environments, this study extended previous research and provided insight into the various pathways by which income influences children’s reading outcomes. The present study examined the associations between income, the early home environment (home literacy and maternal depression), sustained attention, and children's reading outcomes in kindergarten and third-grade using a large, low-income sample. Using latent variable structural equation modeling techniques and a nationally representative data set of at- risk families and children, this study examined the complex mediated and moderated pathways between environmental and within child variables. First, the direct and indirect associations between income, early home literacy environment, early maternal depression, and early sustained 130 attention on children's letter-word identification in kindergarten were examined. Next, the direct and indirect longitudinal effects of income, early home literacy environment, early maternal depression, and early sustained attention on children's passage comprehension in third-grade were examined. Finally, a genetic variable associated with attention, DRD4, was examined as a moderator of the longitudinal associations. Latent variable structural equation modeling analyses tested hypothesized pathways among early income, early home literacy environment, early maternal depression, early sustained attention, and DRD4 on children's kindergarten and third- grade reading outcomes. All measurement and structural models were acceptable, suggesting the paths were interpretable. The hypotheses were examined and partially supported. Early Income Influences Reading in Kindergarten It was hypothesized that households with higher income would have children with higher reading scores in kindergarten and third grade. Higher income during children’s early development was related to better reading in kindergarten, but not in third grade. Despite the somewhat limited range in the income variable since the FFCWS study purposefully oversampled at-risk families (e.g., single-parent household, low-income), results showed households with more income had children with higher letter-word reading scores in kindergarten. Given that the poverty line between 2001 and 2003 was defined as approximately $8,000-9,000 for a family of one, $11,000-12,000 for a family of two, $14,000-15,500 for a family of three, and $17,000-18,500 for a family of four (U.S. Department of Health and Human Services), this finding suggests that even small increases or decreases in household earnings can influence reading outcomes for children living in poor families. As established by the existing literature, family income in early childhood is a powerful and significant predictor of children’s reading, likely due to the association of income and factors relevant to children's development, 131 such as neighborhood, early childhood education, healthcare, and nutrition (Brooks-Gunn, Klebanov, & Liaw, 1995; Sarsour et al., 2011). Even though early household income was significantly associated with children’s letter- word reading scores in kindergarten, the longitudinal influence of early income substantially weakened over time, such that early income was no longer associated with children’s passage comprehension in third grade. Although the null finding was contrary to the study’s hypothesis, it was not surprising. Research has shown that the influence of household income is more prominent in early childhood and the direct effects of early poverty weaken over time (Dickerson & Popli, 2016; Duncan & Brooks-Gunn, 1997; Kiernan & Mensah, 2009; Raver, Blair, & Willoughby, 2012). It may be that the effects of early poverty make their effects apparent directly in early childhood, while longer lasting effects are observed through mediated pathways later in development (Yeung, Liner, & Brooks-Gunn, 2002). Mediational findings of the home literacy environment discussed later in this section support this claim. It is not disputed that early poverty is a risk factor for child development; nonetheless, many children from poor households excel in reading despite the effects of poverty (Kiernan & Mensah, 2011; Raver & Knitzer, 2002). It is possible that factors not measured in this study, such as quality child care and parenting practices, may have minimized the adverse effects of early financial hardship and promoted resilience in children. Research on the resilience of specific subgroups of children suggests that certain factors might protect children from the adverse effects of early poverty (Rutter, 1990). Quality early child care, for example, has been shown to improve children’s academic outcomes, despite growing up in low-income households (Burchinal et al., 2011). Research has identified parental values, beliefs, and behaviors regarding reading and positive parent-child relationships as resilience factors that protect children from the 132 harmful effects of poverty (Davis-Kean, 2005; Masten et al., 1990). Given these mixed findings, more research is certainly necessary to understand how exposure to resilience promoting practices can improve children’s reading. Understanding factors that promote resilience in children can inform intervention efforts for children who are most at-risk. For instance, teaching parents about the value of reading or specific practices on how to read to their children may promote good reading habits that having long-lasting, positive implications. Early Home Literacy Environment Matters for Kindergarten Reading Results supported the hypothesis that homes with better literacy environments would have children with higher reading scores in kindergarten. Despite a sizeable positive skew on the home literacy environment measure, children from homes with higher literacy environments scored three more points on the letter-word reading measure than children from homes with lower literacy environments. Although this finding might initially appear negligible, as three points only represents 1/5 of a standard deviation on the letter-word measure, early differences in reading achievement are very resistant to interventions and tend to widen as children age (Bast & Reitsma, 1998; Foster & Miller, 2007). Early and seemingly small differences have the potential to compound into more considerable differences since research suggests learning to read is a developmental progression whereby later reading skills, such as comprehension and fluency, build on earlier reading skills, like letter-word reading and decoding (Pikulski & Chard, 2005; National Reading Panel, 2000). Therefore, three point reading differences should be monitored carefully; if certain children continue to demonstrate delayed reading skills, evidence-based reading interventions should be introduced. This finding aligns with the existing literature documenting the importance of the early home literacy environment to children's early reading skills (see Burgess, Hecht, & Lonigan, 133 2002; Weigal, Martin, & Benett, 2007, Whitehurst & Lonigan, 2000). Interestingly, the confirmatory factor analysis indicated “books about the alphabet and letters” contributed most to the home literacy environment factor. The present study’s results are commensurate with findings from other studies examining materials in the home literacy environment. For instance, Son and Morrison (2010) examined how components of the home environment related to children's reading achievement. Interestingly, Son and Morrison (2010) also found the material component of the home literacy environment was the most accessible component for parents to increase over time. Acquiring more books was more likely to occur in the home environment than increasing joint parent-child reading experiences, promoting academic stimulation, or providing additional language stimulation. One conclusion may be that parents who purchased more age-appropriate books for their children were engaging more in literacy-related behaviors with their children (e.g., reading together, promoting academic and language stimulation). Together, these findings suggest that the material aspect of the home literacy environment may have some influence on children from low income families’ early letter-word reading. Introducing books into the homes of low-income families and teaching parents how to interact with their children appears to be an accessible point of early intervention. It was also hypothesized the home literacy environment might indirectly link income and children’s reading outcomes as hypothesized by the investment theory of income (Yeung, Linver, & Brooks-Gunn, 2002). Mediation hypotheses were partially supported. Mediation analyses were not significant for overall indirect effects; however, the path for early income, home literacy environment, and letter-word reading was significant, though this finding should be interpreted cautiously since the overall indirect effects was not significant. Households with more income had higher home literacy environments, and children from homes with higher 134 literacy environments had better developed letter-word reading. This finding suggests some of the effect of income on children’s reading outcomes was detected in the FFCWS parents’ decisions to allocate their money towards books about the alphabet, numbers, colors, and songs, as well as the time parents spent reading with their children. In the context of the mediational theory, these were shown as “investments” that enhanced their children's letter word reading development. Functioning through the home literacy environment and as defined by the poverty thresholds, every 50 to 100% increase in family income (approximately $5,000-15,000 increase depending on family size) was associated with less than one point increase on the letter-word measure. An increase of household income in exchange for a less than one point change in letter- word reading appears negligible and may be due to the large sample size of the data set (Kline, 2016). Nonetheless it is important to point out that income represents money allocated towards shelter, food, healthcare, and other necessities that influence children’s development. After considering necessary expenses, it is easy to see how materials for the home literacy environment might be overlooked (Yeung, Linver, & Brooks-Gunn, 2002). Since the study did not include a “value” measure, it is unclear whether parents and caregivers valued purchasing literacy materials but could not afford to. It may be that despite valuing books, reading programs, and time spent reading with their children, poor and near-poor families did not have the disposable income or time to afford books or time to read with their children (DeBaryshe & Binder, 1994). Effects of the Early Home Literacy Environment Do Not Directly Persist into Third Grade Differing from the study’s hypothesis, the direct effects of the early home literacy environment did not persist longitudinally into the third grade. Homes with higher early literacy 135 environments were not associated with higher passage comprehension scores in third grade. Likewise, the indirect effects of the early home literacy environment were similarly minimal. Differing with the study’s hypotheses, the total indirect effects were significant for only one path of the longitudinal model. While the mediation path between income, depression, and reading as well as the mediation path between income, sustained attention, and reading were not significant, the mediation path between income, home literacy environment, letter-word reading, and passage comprehension was significant. Specifically, households with more income had higher home literacy, children exposed to better home literacy environments had better developed letter-word reading, and children with better letter word reading in kindergarten had higher passage comprehension in third grade. The statistical significance of this finding aligns with previous research demonstrating the longitudinal indirect effect of income on reading for children from at- risk families through the investment model (Yeung, Linver, & Brooks-Gunn, 2002). However, despite the statistical significance, it is important to recognize this mediation effect explained less than a one point difference in reading outcomes, suggesting a marginal clinically significant change in reading scores. Clinically, the mediational effects are considered small, as income, home literacy, and kindergarten letter-word reading only accounted for a 0.025 gain in third-grade passage comprehension scores. Similar to the mediation effect in kindergarten, the significant statistical effect should be tempered by the practical implication of this finding; a less than one point change in reading scores does not translate into a meaningful representation of children’s reading abilities. It is likely that other factors in the environment beyond the early home literacy environment are responsible for low reading achievement in the third grade. The present findings 136 may be an artifact of the limited perspective of the home literacy environment given the home literacy environment variable represented investments. Early Maternal Depression is not Associated with Reading Contrary to the study’s hypothesis, mothers' depressive symptoms were not significantly associated with their children's reading outcomes in kindergarten or third grade. Mothers' depressive symptoms were not directly related to children’s reading outcomes in kindergarten or third grade, nor did mothers’ depressive symptoms mediate the association between income and children's reading outcomes in kindergarten or third grade. Preliminary correlations indicated income was associated with more instances of maternal depression, which aligns with the extant literature since robust findings demonstrate that mothers living in lower household income reported more symptoms of depression than mothers in higher income households (Carlson & Corcoran, 2001; Johnson & Flake, 2007; McLoyd, 1990; Pachter et al., 2006). However, it appears mothers' depressive symptoms do not influence their children's reading scores in the FFCWS. This finding is unexpected given the evidence from the extant literature that mothers who endorse more depressive symptoms read less frequently and for shorter duration to their children (Bigatti, Cronan, & Anaya, 2001) and evidence that depressed mothers are more likely to have children with less well-developed language and reading (Sohr-Preseton & Scarmamella, 2006; Zaslow, Ahir, Dion, Ahluwalia, & Sargent, 2001). Although this finding is contrary to the results from many studies (Baker & Iruka, 2013; Barbarin et al., 2006; Bigatti, Cronan, & Anaya, 2001; Reissland, Shepherd, & Herrera, 2003), one explanation might be the use of maternal depression variable without a complementary maternal parenting behavior variable as a representation of the “family stress path” resulted in the null findings. It is likely that maternal depression does not always translate into parenting 137 behaviors, and the omission of a parenting behaviors variable resulted in failure to capture how the family stress path functions in the FFCWS sample. One possible conclusion is that although maternal risk factors (e.g., depression, stress) are predictive of children's outcomes, mothers experiencing depression can still parent efficiently to promote their children's literacy competence (Lovejoy, Graczyk, O'Hare, & Neuman, 2000; Mason, Briggs, & Silver, 2011). Another explanation comes from a study conducted by Foster, Lambert, Abbott-Shim, McCarty, and Franze, 2005). Examining low-income students and families attending Head Start, the authors found that home literacy (e.g., books, time reading) significantly predicted children’s reading, but social risk (e.g., maternal depression, parenting) did not. Interestingly, the home literacy environment variable held the most clinical significance, as homes with more early literacy exposure were associated with three more points on a letter-word reading measure. In this context, the current study’s results underscore the importance of stimulating and diverse literacy materials to children’s literacy in the home environment above and beyond maternal mental health characteristics. This finding also supports the importance of examining both family stress variables and investment variables simultaneously when studying at-risk families. Early Sustained Attention Directly Influences Reading Along with environmental variables, the bioecological model postulates considering within child factors that contribute to reading, such as attention. Results supported the hypothesis that children with higher scores on a sustained attention measure had higher letter-word reading scores in kindergarten. Children with better sustained attention scored approximately half a point higher on the letter-word reading measure than children with less sustained attention. This finding is consistent with the extant literature on the importance of attentional systems for the early acquisition of reading skills (Fan, McCandliss, Sommer, Raz, & Posner, 2002; Kupietz, 138 1990; Rowe & Rowe, 1992; Stern & Shalev, 2013) and particularly of sustained attention to reading (Grigorenko, Kornev, Rakhlin, & Krivulskaya, 2011; Pham, Fine, & Semrud-Clikeman, 2011; Pham, 2016). It is likely that sustained attention aids children’s reading development by maintaining alertness to text, orienting to essential text features, and detecting changes in processing (Posner & Peterson, 1990). Sustained attention is thought to be a fundamental aspect of attention that determines "higher" levels of attention (e.g., executive attention), and is necessary for the development of other cognitive skills, including reading (Sarter, Givens, & Bruno, 2001). Although attention was observed as directly related to kindergarten reading, the hypothesis that children’s early sustained attention would mediate the association between income and children’s letter-word reading was not substantiated. This finding differed from previous studies, as Razza, Martin, and Brooks-Gunn’s (2010) and the NICHD Early Child Care Research Network (2003) found significant results for sustained attention as a partial mediator. Differences in methodology may account for the divergent findings between the previous studies and the present study. For instance, Razza, Martin, and Brooks-Gunn’s (2010) and the NICHD Early Child Care Research Network (2003) used regression while the current study used structural equation modeling with bootstrapping. Alternatively, Razza, Martin, and Brooks-Gunn (2010) used two income groups, while the present study used four. The use of SEM in the present study allowed for simultaneous consideration of other environmental factors that may mediate the association between income and children’s reading. Furthermore, the use of an income variable with more categories allowed for more conclusions about the effect of income beyond below or above the poverty line. As such, in context of home literacy environments and maternal depression, it can be concluded that children’s sustained attention was not a significant 139 mediator in kindergarten. Similar findings of direct and mediational results extended to longitudinal analyses. Results supported the hypothesis that children with better early sustained attention measure also had better passage comprehension in third grade. Children with better sustained attention scored higher on the passage comprehension measure than children with lower early sustained attention. This association aligns with the extant literature on the predictive power of early attention. Duncan and colleagues' (2006) seminal study showed school entry attention abilities as one of the strongest predictors of children's later reading achievement across six longitudinal data sets. The ability to sustain attention early in development has been linked to better literacy skills, such as decoding, fluency, and comprehension in other studies too (Fan, McCandliss, Sommer, Raz, & Posner, 2002; Kupietz, 1990; Rowe & Rowe, 1992; Stern & Shalev, 2013). Comparable to the finding from the cross-sectional model, results showed no evidence for the hypothesis that children's early sustained attention would mediate the association between income and children's letter-word and passage comprehension. Since there was a null finding for mediation in the cross-sectional model, this finding was not surprising. Factors not examined by this study may better account for the unsubstantiated mediation effect, as other within child cognitive or biological factors might account for children's sustained attention. DRD4 Moderation Differing from the study’s hypothesis, the DRD4 allele did not moderate the associations between home literacy environment and children’s passage comprehension in third grade, and a moderated-mediation effect was not observed. However, many direct associations were significant. In a model without covariates, children from households with more income had 140 higher reading comprehension scores than children from households with less income. Children with better early sustained attention had better developed reading comprehension than children with poorer early sustained attention. Children exposed to more enriched early home literacy environments had higher reading comprehension scores than children with less enriched home literacy environments. Children with less depressed mothers had better third-grade reading comprehension scores than children with more depressed mothers. Importantly, these findings were significant when covariates were not included in the analyses (since the BIC and AIC were lowest for the model without covariates). There are several explanations for these findings. First, the significant findings in the absence of covariates indicate the strong influence of marital status, mothers' age, mothers' education, mothers' race/ethnicity, children's age at assessment, children's language estimate, and children's gender on third-grade reading comprehension scores. Descriptive statistics indicate particularly strong correlations between children’s gender, maternal education, and language estimate on the third-grade passage comprehension outcome, suggesting that the omission of these covariates account for much of the variance in the passage comprehension. Second, the statistical analysis used to examine the moderation hypothesis can only examine one moderated path at a time (e.g., the effect of DRD4 on the association between HLE and passage comprehension). Although multiple group SEM could have been applied to determine moderation, it would not have allowed for differential susceptibility or diathesis stress hypotheses to be tested, as differential susceptibility and diathesis stress testing requires the slopes from an interaction variable to be plotted. Third, the structural equation model from research question two did not converge when the letter-word reading variable was included a second mediator. Thus, the letter-word reading variable was removed from the original model and used as a control variable. Given the predictive power of 141 early reading and significant mediation paths from income through HLE and letter-word reading to passage comprehension, it is likely that the exclusion of the letter-word reading variable negatively influenced this statistical analysis. To address the weaknesses described, post-hoc analyses were conducted using multiple group SEM. Although multiple group SEM does not allow for formal differential susceptibility and diathesis stress testing, it simultaneously estimates models and allows for moderation to be estimated on all paths. Multiple group SEM analyses revealed a novel result. A path not initially identified in the longitudinal model (research question 2) between home literacy environment and early letter-word reading was significant for moderation. The association between home literacy and letter-word reading was significant for the DRD4 long group but not for the DRD4 short group. This finding suggests a moderating effect of D4D4 on the association between home literacy environment and letter-word reading scores in kindergarten. Specifically, children with the DRD4 long allele who had better home literacy environments had higher letter-word reading scores. An increase in the early home literacy environment (e.g., approximately two more books, parent-child reading more than one time per month) was associated with half a point increase in letter-word reading for children with DRD4 long. Conversely, the association between home literacy environment and letter-word reading was not significant for children with the DRD4 short allele. This difference between groups suggests that the home literacy environment was not associated with a significant increase in letter-word reading for children with the short DRD4 allele, but was associated with a significant increase in letter-word reading for children with the long DRD4 allele. The strength of this finding is statistically robust; however, the clinical implications should be interpreted cautiously. More specifically, the HLE was associated with a 1.4 point 142 increase in letter word reading for the DRD4 long group and the HLE was associated with a 0.8 point increase in letter word reading for the DRD4 short group, suggesting an approximate 0.6 point difference in letter word reading between groups. Much like the mediational findings, the clinical implications of a 0.6 point difference is minimal. Furthermore, since the indicators of the HLE variable were highly skewed, it is difficult to determine an exact number of books or reading encounters associated with the 0.6 increase in letter-word reading scores for children with DRD4 long who were exposed to high early home literacy environments. Nonetheless, the finding that DRD4 influences the effects of the home literacy environment on reading skills has not been shown in the literature before and contributes theoretical considerations on how biological and environmental factors interact to influence children’s development. The findings suggest children’s genetics influence how they respond to the home literacy environment and might explain why some children who experience stressors associated with poverty are resilient despite poverty’s adverse effects (Rutter, 2012). Due to the small clinical significance of the findings, it is likely that a 0.6 difference in letter word reading does not represent a meaningful difference in children’s reading abilities between DRD4 groups. It would not be prudent to differentiate assessment (e.g., provide genetic testing) or intervention (e.g., provide more early reading to some groups of children) based on these results. Instead, given that early reading experiences have been shown to be beneficial for all children, early home reading experiences should continue to be promoted at an universal level. Differential susceptibility. Although differential susceptibility could not be formally tested, the theoretical underpinnings of differential susceptibility align with the finding that DRD4 moderates the relation between the home literacy environment and children's letter-word 143 reading in kindergarten. Briefly stated, the differential susceptibility model postulates that specific groups of children (e.g., children with DRD4 long) may be more susceptible to the environment than their peers (e.g., children with DRD4 short). While children with genetic susceptibility may lag behind in suboptimal environments, they excel in optimal environments (Belsky, Bakermans-Kranenburg, & Van IJzendoorn, 2007; Belsky & Pluess, 2009). This phenomenon was observed partly in the present study. Results revealed that the DRD4 long “susceptible” group did better than the low susceptible group (DRD4 short) when provided with “optimal” high home literacy environments, but the reverse pattern was not observed. In other words, the "for-better” pattern was observed, but the for-worse pattern (e.g., how children with DRD4 long perform in “suboptimal” environments) was not established. Conclusions from the results showed that children with DRD4 long and high home literacy environments excelled, but children without DRD4 long did not necessarily benefit or suffer from the environment. Similar results were found in a lab study of reading and DRD4. When studying reading in a group of four-year-old preschool children, Plak, Kegel, & Bus (2015) examined whether DRD4 moderated children’s responses to a phonemic reading intervention with specific feedback (e.g., praise). Results showed significantly more reading improvements for children with the DRD4 long allele who were exposed to the reading intervention. Children with the DRD4 short allele marginally improved. Although differential susceptibility was not formally established in this study either, the results are promising because a differential susceptibility finding could inform targeted intervention for children. Children with DRD4 long responded particularly well to praise in the phonemic intervention, suggesting that adding additional praise could significantly increase children's reading achievement. 144 Altogether, these findings suggest the importance of the environment for children with the DRD4 long allele. The longitudinal component of the present study further contributes to the existing literature by providing further evidence for a long-held belief – the early developmental period is particularly sensitive for children with genetic susceptibility. It may be that reading acquisition requires a time-dependent hierarchical or sequential development of sub-skills (Pennington, 1991), of which children with certain biological dispositions are very sensitive. The lack of environmental supports and stimuli in the home literacy environment may disrupt the development of the reading system for children with DRD4 long (Ellis, Boyce, Belsky, Bakermans-Kraneburg, & Van Ijzendoorn, 2011), and may place them on a trajectory towards reading problems. Implications of this finding further support the need, importance, and value of early reading interventions for young, at-risk children (Lonigan, Farver, Phillips, & Clancy- Menchetti, 2011). RD and ADHD. The possible sensitivity to reading acquisition for children with DRD4 long may, in part, be explained by research into ADHD and reading disability. As previously mentioned, research has consistently demonstrated the high comorbidity between Attention- Deficit Hyperactivity Disorder (ADHD) and reading disability (RD), since 15-35% of individuals with RD have comorbid ADHD (Germano, Gagliano, & Curatolo, 2010; Willcutt & Pennington, 2000; Shaywitz et al., 1995; Willcutt, Pennington, Olson, & DeFries, 2007; Willcutt et al., 2010). Evidence points to a shared genetic etiology between attention and reading problems (e.g., Wilcutt et al., 2007; Ebejer et al., 2010; Zumberge, Baker, & Manis, 2007); however, not all children with ADHD have comorbid RD. It may be that DRD4, or other dopaminergic genes, is implicated in the comorbid ADHD and RD population. For instance, DRD4 has been variably linked to ADHD symptoms in several studies; however, there are 145 considerable differences in how researchers classify subtypes of ADHD. The differences in classification could be influencing whether DRD4 was found to be significantly associated with ADHD. Some studies have found significant associations between DRD4 long and ADHD (LaHoste et al., 1996; Rowe et al., 1998; Smalley et al., 1998), other studies have only found an association between DRD4 and severe ADHD with combined presentations (Todd et al., 2005), and many recent studies have found evidence refuting the association altogether (Kustanovich et al., 2003; Todd et al., 2005). Certainly, more research is needed to determine whether dopaminergic genes, such as DRD4, are associated with the subpopulation of children with comorbid RD and ADHD. Research on the cognitive manifestations of ADHD and RD are beginning to uncover these complex associations. The multiple-deficit model (Pennington, 2006) postulates disorders are unlikely to be caused by a single gene acting in isolation (Pennington, 2006; Plomin, DeFries, McClearn, & Rutter, 1997; Willcutt et al., 2001). In support of the multiple-deficit model, neuropsychological studies examining genetic pleiotropy of ADHD and RD have found that no single gene accounts for behavioral symptoms of either ADHD and RD. Additionally, there is substantial evidence for the implication of specific environmental factors in the etiology of ADHD and RD, such as low birth weight, adverse childhood experiences, and exposure to specific environments (Jimenez, Wade, Schwartz-Soicher, Lin, & Reichman, 2017; Olson, Wise, Conners, Rack, & Fulker, 1989; Swanson et al., 2007). In this context with the present findings, it may be that early home literacy environments and DRD4 are additional environmental and genetic factors, respectively, implicated in impaired reading outcomes. No other paths were significant for the multiple group testing. Unlike a study conducted by Kegel and Bus (2012), the present study did not find a significant moderation effect of DRD4 146 on the association between attention and reading outcomes. The null effects of DRD4 moderation on sustained attention could be due to the study’s failure to measure more than one aspect of attention (e.g., executive attention, selective attention, orienting attention) and more than one aspect of reading (e.g., alphabetic awareness, blending, decoding). Despite, these shortcomings, these findings provide further support for a perspective of reading and attention disability that incorporates additive and interactive systems of multiple genetic and environmental factors. Conclusions and Clinical Implications Learning to read builds the foundation for academic success in middle and high school (Foulin, 2005; Sénéchal, LeFevre, Smith-Chant, & Colton, 2001; Whitehurst & Lonigan, 1998). Children who read early and well develop more efficient reading strategies, understand school content better, and show continued growth across academic domains (Cunningham & Stanovich, 1997). Often, for these children, emergent literacy is acquired in the home environment preceding formal literacy instruction in school. Unfortunately, a large number of children lack emergent literacy and cognitive skills fundamental to school success at school entry because they were not exposed to enriching literacy environments. These children are not as well prepared to engage in intensive, formal reading instruction as their peers who were exposed to more enriching literacy environments (Plak, 2016). It is well established that children who are not at grade level upon completion of first grade have dramatically lower chances of being at or above grade level in reading later in school (Spira, Bracken, & Fischel, 2005; Wyner, Bridgeland, & Diiulio, 2007). The findings from this study further support this finding. The well-documented, positive associations between the early home literacy environment and children's reading achievement were supported by the statistically significant findings 147 (Payne, Whitehurst, & Angell, 1994; Snow, Burns, & Griffin, 1998; Whitehurst & Lonigan, 1998). The statistically significant positive effects should be tempered by the small clinical significance in children’s reading scores since the less-than-one-point difference between groups suggests that the HLE variable may have been too limited to adequately capture the association between the home environment and reading outcomes. Novertheless, when compared to sustained attention and maternal depression, the home literacy environment was a better predictor of early reading scores and significant efforts should be made to sustain programs that support early home literacy, such as Head Start and the Early Literacy Initiative (along with other state and federally funding pre-school literacy programs). The theory behind Head Start recognizes the family as key to countering the cyclical and pervasive impacts of poverty on children's academic, social, and emotional development (Zigler & Muenchow, 1992). Head Start includes a parent investment component that encourages and provides home learning materials, activities, and experiences. Head Start also supports the critical role that parents play in children's development. The Early Literacy Initiative is a major initiative in the Michigan Department of Education (MDE) due to the 2015 Third-Grade Reading workup. The Michigan Association of Intermediate School Administrators (MAISA) General Education Leadership Network (GELN) created a document that outlines considerations for Pre- K and K-3 Essential Instructional Practices in Early Literacy. The Pre-kindergarten instructional practices in early literacy includes recommendations to use literacy artifacts throughout the classroom, read aloud with print, interactive read aloud with vocabulary and comprehension focus, play with sounds, instruction on letter names, writing, extended conversation, provision of reading material in the classroom, observation and assessment of children’s reading and language, and collaboration with families in promoting literacy. 148 In both of these examples, home-based literacy activities are strongly encouraged and promoted; however, it is unclear how policy is supporting children’s literacy before these children are identified by Head Start or before pre-school and kindergarten entry. As such, it may be prudent to explore policy that incorporates literacy into primary health care provision. Introducing home-based literacy activities earlier in children’s development could occur during well-child visits or routine appointments with pediatricians. Kuo, Franke, Regalado, & Halfon (2004) surveyed families and their pediatric physicians, results indicated 62% percent of parents reported discussing reading with their pediatric physician and, of these parents, 55% reported reading daily to their child. Of the remaining 38% of parents who did not discuss reading with their pediatrician, approximately half reported they would have found a discussion on reading helpful. Reaching out to pediatric providers may be a beneficial avenue to promote even earlier early home literacy for the most at-risk children. Regarding early intervention, the home literacy environment was shown as important for all children. Despite a theoretically important finding that the HLE was more influential for children with a known genetic risk factor (e.g., children with DRD4 long), more research is required before clinical practices change in response to this finding. Presently, it seems prudent to address school entry reading for all children who evidence delayed reading skills in kindergarten; then, as children develop, reading supports should change to meet children’s shifting needs. In summary, although the immediate clinical implications are marginal, the broad clinical implications support a critical period of early development whereby children would benefit from exposure to early literacy experiences. Children should live in home environments with a variety of age-appropriate books about letters, colors, numbers, and songs and children should be read to 149 by caregivers and parents more frequently. Efforts should be made to inform parents of the importance of these home literacy activities. Perhaps an act as simple as providing parents books would support children’s reading outcomes for years to come. Thus, child care professional, psychologists, social workers, educators, and healthcare providers should provide families with information about local libraries, literacy programs, and community programs that provide children with cognitively stimulating materials and encourage engagement in reading activities and stimulating outings (Brooks-Gunn, Berlin, & Fuligni, 2000; Fuligni & Brooks-Gunn, 2000). Limitations and Future Directions There are several limitations to the proposed study. First, since the FFCWS had a non- experimental design, no causal inferences can be made about income, sustained attention, early home literacy environments, maternal depression, genetics, and reading outcomes. An experimental design is necessary to establish the how altering levels of income, the early home literacy environment, maternal depression, and sustained attention correspond to reading outcomes. Thus, prudence is necessary when making conclusions concerning causal relationships about the home literacy environment, maternal depression, sustained attention, and children’s development in reading. Second, because weights were only applied to descriptive statistical analyses, it is unclear how representative the final data set is to the national population between 1998 and 2000 because of the sampling design of the study. Since the FFCWS did not create a longitudinal weight, a cross-sectional weight was applied; therefore, longitudinal results from this study should be interpreted with caution. Third, mediators not assessed in the study could play a role in how income affects children’s reading outcomes, such as father’s mental health, neighborhood effects, or type of childcare (Duncan & Brooks-Gunn, 1997; Guo & Harris, 2000). One missing variable of 150 importance is parenting behaviors. Given the study’s use of the family stress model, a parenting variable would represent a more proximal process than maternal depression. Fourth, formal differential susceptibility testing could not be conducted and although results align with the differential susceptibility theory, the results are not conclusive of differential susceptibility. Fifth, the limitations associated with self-report are present in this study (e.g., rater bias) and there are limitations related to using a single method of construct measurement. Finally, because only one gene was examined in this study, the significant moderation of DRD4 on the home literacy environment and letter-word reading does not constitute a complete GxE interaction and does not represent best practices in GxE methodology. Currently, prominent researchers in the gene by environment field suggest the use genome-wide associations over candidate genes (Dick et al., 2015; Duncan & Keller, 2011; Lee et al., 2012). Thus, future iterations of this research should examine multiple genetic factors and use genome-wide association methodology to better represent true genetic risk. The study highlights important considerations and areas for future research. Given the long-term effects of low income on children’s academic development, longitudinal SEM modeling, such as latent growth modeling, would provide information on the influence of each variable over time. There were significant direct, mediation, and moderation effects in this study, and all analyses revealed that the home literacy environment was the most significant variable when considering children’s reading in the current study’s context. More research should examine the home literacy environment’s effects on children’s reading development, especially with various definitions of home literacy. Development is influenced by multiple genetic factors, and single gene approaches do not explain the intricacy of genetics; therefor, further research should consider using additional genetic variables to examine reading. Genetic variables may 151 provide new explanations and perspectives for why some children learn to read well despite risk factors. Lastly, it is important to point out there is no single pathway through which family income operates on children’s reading and future research should examine other aspects the environment and within child factors. Research on enrollment in preschool, academic interventions, and parenting style are promising variables of interest. 152 APPENDIX 153 Initial IRB Application Determination *Exempt* March 10, 2016 To: Re: Jodene Fine 440 Erickson Hall IRB# x16-142e Category: Exempt 4 Approval Date: March 8, 2016 Title: Parenting style and DAT1 in Relation to Attention Behaviors The Institutional Review Board has completed their review of your project. I am pleased to advise you that your project has been deemed as exempt in accordance with federal regulations. The IRB has found that your research project meets the criteria for exempt status and the criteria for the protection of human subjects in exempt research. Under our exempt policy the Principal Investigator assumes the responsibilities for the protection of human subjects in this project as outlined in the assurance letter and exempt educational material. The IRB office has received your signed assurance for exempt research. A copy of this signed agreement is appended for your information and records. Renewals: Exempt protocols do not need to be renewed. If the project is completed, please submit an Application for Permanent Closure. Revisions: Exempt protocols do not require revisions. However, if changes are made to a protocol that may no longer meet the exempt criteria, a new initial application will be required. Problems: If issues should arise during the conduct of the research, such as unanticipated problems, adverse events, or any problem that may increase the risk to the human subjects and change the category of review, notify the IRB office promptly. Any complaints from participants regarding the risk and benefits of the project must be reported to the IRB. Follow-up: If your exempt project is not completed and closed after three years, the IRB office will contact you regarding the status of the project and to verify that no changes have occurred that may affect exempt status. Please use the IRB number listed above on any forms submitted which relate to this project, or on any correspondence with the IRB office. Good luck in your research. If we can be of further assistance, please contact us at 517-355-2180 or via email at IRB@msu.edu. Thank you for your cooperation. Sincerely, Harry McGee, MPH SIRB Chair c: June Westdal 154 Office of Regulatory Affairs Human Research Protection Programs Biomedical & Health Institutional Review Board (BIRB) Community Research Institutional Review Board (CRIRB) Social Science Behavioral/Education Institutional Review Board (SIRB) Olds Hall 408 West Circle Drive, #207 East Lansing, MI 48824 (517) 355-2180 Fax: (517) 432-4503 Email: irb@msu.edu www.hrpp.msu.edu MSU is an affirmative-action, equal-opportunity employer. REFERENCES 155 REFERENCES Aarnoudse-Moens, C., Weisglas-Kuperus, N., Duivenvoorden, H., Goudoever, H., & Oosterlaan, J. (2013). Executive function and IQ predict mathematical and attention problems in very preterm children. Plos One, 8(2), e55994. doi:10.1371/journal.pone.0055994 Aber, J., Bennett, N., Conley, D., & Li, J. (1997). The effects of poverty on child health and development. Annual Review of Public Health, 18(1), 463-483. doi:10.1146/annurev.publhealth.18.1.463 Abidin, R. R. (1990). Parenting stress index-short form. Charlottesville, VA: Pediatric Psychology Press. Abikoff, H., Courtney, M., Pelham, J., W E, & Koplewicz, H. S. (1993). Teachers' ratings of disruptive behaviors: The influence of halo effects. Journal of Abnormal Child Psychology, 21(5), 519-533. doi:10.1007/BF00916317 Achenbach, T. M. (1991). Manual for the Child Behavior Checklist/4-18 and 1991 profile (p. 288). Burlington, VT: Department of Psychiatry, University of Vermont. Ackerman, B. P., Brown, E. D., & Izard, C. E. (2004). The relations between persistent poverty and contextual risk and children's behavior in elementary school. Developmental Psychology, 40(3), 367-377. doi:10.1037/0012-1649.40.3.367 Afflerbach, P., Cho, B. Y., Kim, J. Y., Crassas, M. E., & Doyle, B. (2013). Reading: What else matters besides strategies and skills?. Reading Teacher, 66(6), 440-448. doi:10.1002/TRTR.1146 Aikens, N. L., & Barbarin, O. (2008). Socioeconomic differences in reading trajectories: The contribution of family, neighborhood, and school contexts. Journal of Educational Psychology, 100(2), 235-251. doi:10.1037/0022-0663.100.2.235 Alexander, K. L., Entwisle, D. R., & Dauber, S. L. (1993). First-grade classroom behavior: Its short-and long-term consequences for school performance. Child development, 64(3), 801-814. doi:10.1111/j.1467-8624.1993.tb02944.x Al Otaiba, S., Kosanovich, M. L., Torgesen, J. K., Kamhi, A. G., & Catts, H. W. (2012). Assessment and instruction for phonemic awareness and word recognition skills. Language and Reading Disabilities, 3, 112-140. 156 Alpern, L., & Lyons-Ruth, K. (1993). Preschool children at social risk: Chronicity and timing of maternal depressive symptoms and child behavior problems at school and at home. Development and Psychopathology, 5(3), 371-387. doi:10.1017/S0954579400004478 American Academy of Pediatrics Committee on Environmental Health. (2005). Lead exposure in children: prevention, detection, and management. Pediatrics, 116(4), 1036. doi: 10.1542/peds.2005-1947 Anderson, P. (2008). Towards a developmental model of executive function. In V. Anderson, R. Jacobs, & P. Anderson (Eds.), Executive functions and the frontal lobes (pp. 3-22). New York, NY: Psychology Press. Anderson, D. R., Burnham, K. P., & White, G. C. (1998). Comparison of Akaike information criterion and consistent Akaike information criterion for model selection and statistical inference from capture-recapture studies. Journal of Applied Statistics, 25(2), 263-282. Anglum, B. S., Bell, M. L., & Roubinek, D. L. (1990). Prediction of elementary student reading achievement from specific home environment variables. Reading Improvement, 27(3), 173. Arnbak, E. (2004). When are poor reading skills a threat to educational achievement?. Reading and Writing, 17(5), 459-482. doi:10.1023/B:READ.0000044595.76174.cc Arnold, E. M., Goldston, D. B., Walsh, A. K., Reboussin, B. A., Daniel, S. S., Hickman, E., & Wood, F. B. (2005). Severity of emotional and behavioral problems among poor and typical readers. Journal of Abnormal Child Psychology, 33(2), 205-217. doi:10.1007/s10802-005-1828-9 Atkins, M. S., Pelham, W. E., & Licht, M. H. (1985). A comparison of objective classroom measures and teacher ratings of attention deficit disorder. Journal of abnormal child psychology, 13(1), 155-167. doi:10.1007/BF00918379 Auerbach, J. G., Benjamin, J., Faroy, M., Geller, V., & Ebstein, R. (2001). DRD4 related to infant attention and information processing: A developmental link to ADHD? Psychiatric Genetics, 11(1), 31-35. doi:10.1097/00041444-200103000-00006 Auinger, P., Lanphear, B. P., Kalkwarf, H. J., & Mansour, M. E. (2003). Trends in otitis media among children in the united states. Pediatrics, 112(3), 514-520. doi:10.1542/peds.112.3.514 Babu, S. C., Gajanan, S. N., & Sanyal, P. (2014). Food security, poverty, and nutrition policy analysis: Statistical methods and applications (Second ed.). Boston;Amsterdam;: Academic Press. 157 Bailey, T., Jenkins, D., Leinbach, T., & Columbia University. Teachers College. (2005). What we know about community college low-income and minority student outcomes: Descriptive statistics from national surveys. Distributed by ERIC Clearinghouse. Baker, L., Fernandez-Fein, S., Scher, D., & Williams, H. (1998). Home experiences related to the development of word recognition. In Metsala, J. L. & Ehri, L. C. (Eds). Word Recognition in Beginning Literacy, (263-287). Mahwah, NJ: Lawrence Erlbaum Associates Publishers. Baker, C. E. (2013). Fathers’ and mothers’ home literacy involvement and children’s cognitive and social emotional development: Implications for family literacy programs. Applied Developmental Science, 17(4), 184–197. doi:10.1080/10888691.2013.836034 Bakermans-Kranenburg, M. J., IJzendoorn, M. H. V., Pijlman, F. T. A., Mesman, J., & Juffer, F. (2008). Experimental evidence for differential susceptibility: Dopamine D4 receptor polymorphism (DRD4 VNTR) moderates intervention effects on toddlers' externalizing behavior in a randomized controlled trial. Developmental Psychology, 44(1), 293-300. doi:10.1037/0012-1649.44.1.293 Barkley, R. A. (1997). Behavioral inhibition, sustained attention, and executive functions: Constructing a unifying theory of ADHD. Psychological Bulletin, 121(1), 65-94. doi:10.1037//0033-2909.121.1.65 Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51(6), 1173-1182. doi:10.1037/0022-3514.51.6.1173 Bast, J., & Reitsma, P. (1998). Analyzing the development of individual differences in terms of Matthew effects in reading: results from a Dutch Longitudinal study. Developmental Psychology, 34(6), 1373. Bierman, K.L., Torres, M.M., Domitrovich, C.E., Welsh, J.A., & Gest, S.D. (2009.) Behavioral and cognitive readiness for school: Cross-domain associations for children attending Head Start. Social Development, 18, 305-323. doi:10.1111/j.1467-9507.2008.00490.x Bigatti, S. M., Cronan, T. A., & Anaya, A. (2001). The effects of maternal depression on the efficacy of a literacy intervention program. Child Psychiatry and Human Development, 32(2), 147-162. doi:10.1023/A:1012250824091 Bhutta, A.T., Cleves, M.A., Casey, P.H., Cradock, M.M., & Anand, K.J. (2002). Cognitive and behavioral outcomes of school-aged children who were born preterm: A meta-analysis. JAMA : The Journal of the American Medical Association, 288(6), 728-737. doi:10.1001/jama.288.6.728 158 Black, R., Victora, C., Walker, S., Bhutta, Z., Christian, P., de Onis, M., . . . Maternal and Child Nutrition Study Group. (2013). Maternal and child undernutrition and overweight in low- income and middle-income countries. Lancet, 382(9890), 427-451. doi:10.1016/S0140- 6736(13)60937-X Belsky, J. (2005). Differential susceptibility to rearing influence. In Ellis, B. J., & Bjorklund, D. F. (Eds.), Origins of the social mind: Evolutionary psychology and child development. (139-163). New York: Guilford Press. Belsky, J., Bakermans-Kranenburg, M. J., & Marinus H. van IJzendoorn. (2007). For better and for worse: Differential susceptibility to environmental influences. Current Directions in Psychological Science, 16(6), 300-304. doi:10.1111/j.1467-8721.2007.00525.x Belsky, J., & Pluess, M. (2009). Beyond diathesis stress: Differential susceptibility to environmental influences. Psychological Bulletin, 135(6), 885-908. doi:10.1037/a0017376 Bracken, S. S., & Fischel, J. E. (2008). Family reading behavior and early literacy skills in preschool children from low-income backgrounds. Early Education & Development, 19(1), 45-67. doi:10.1080/10409280701838835 Blair, C., & Razza, R. P. (2007). Relating effortful control, executive function, and false belief understanding to emerging math and literacy ability in kindergarten. Child Development, 78(2), 647-663. doi:10.1111/j.1467-8624.2007.01019.x Brennan, P. A., Hammen, C., Andersen, M. J., Bor, W., Najman, J. M., & Williams, G. M. (2000). Chronicity, severity, and timing of maternal depressive symptoms: Relationships with child outcomes at age 5. Developmental Psychology, 36(6), 759-766. doi:10.1037//0012-1649.36.6.759 Bolger, K. E., Patterson, C. J., Thompson, W. W., & Kupersmidt, J. B. (1995). Psychosocial adjustment among children experiencing persistent and intermittent family economic hardship. Child Development, 66(4), 1107-1129. doi: 10.1111/j.1467- 8624.1995.tb00926.x Both-de Vries, A. C., & Bus, A. G. (2010). The proper name as starting point for basic reading skills. Reading and Writing, 23(2), 173-187. doi:10.1007/s11145-008-9158-2 Boury, J. M., Larkin, K. T., & Krummel, D. A. (2004). Factors related to postpartum depressive symptoms in low-income women. Women & Health, 39(3), 19-34. doi: 10.1300/J013v39n03_02 Bowey, J. A. (1995). Socioeconomic status differences in preschool phonological sensitivity and first-grade reading achievement. Journal of Educational Psychology, 87(3), 476. doi:10.1037/0022-0663.87.3.476 159 Britto, P. R. and Brooks-Gunn, J. (2001), Beyond Shared Book Reading: Dimensions of Home Literacy and Low-Income African American Preschoolers' Skills. New Directions for Child and Adolescent Development, (2001)92, 73–90. doi:10.1002/cd.16 Bronfenbrenner, U., & Ceci, S. J. (1994). Nature-nuture reconceptualized in developmental perspective: A bioecological model. Psychological Review, 101(4), 568-586. doi:10.1037/0033-295X.101.4.568 Bronfenbrenner, U. (1977). Toward an experimental ecology of human development. American Psychologist, 32(7), 513-531. doi:10.1037/0003-066X.32.7.513 Brooks-Gunn, J., & Duncan, G. J. (1997). The effects of poverty on children. The Future of Children, 55-71. doi: 10.2307/1602387 Brooks-Gunn, J., & Markman, L. (2005). The Contribution of Parenting to Ethnic and Racial Gaps in School Readiness. The Future of Children, 15(1), 139-168. Retrieved from http://www.jstor.org/stable/1602666 Brooks-Gunn, J., Klebanov, P. K., & Liaw, F. (1995). The learning, physical, and emotional environment of the home in the context of poverty: The infant health and development program. Children and Youth Services Review, 17(1), 251-276. doi:10.1016/0190- 7409(95)00011-Z Buehler, C., & Gerard, J. M. (2002). Marital conflict, ineffective parenting, and children's and adolescents' maladjustment. Journal of Marriage and Family, 64(1), 78-92. doi:10.1111/j.1741-3737.2002.00078.x Buil, J. M., Koot, H. M., Olthof, T., Nelson, K. A., & van Lier, Pol A. C. (2015). DRD4 genotype and the developmental link of peer social preference with conduct problems and prosocial behavior across ages 9–12 Years. Journal of Youth and Adolescence, 44(7), 1360-1378. doi:10.1007/s10964-015-0289-x Burkam, D. T. (2013). Educational Inequality and Children: The preschool and early school years. In Rycroft, R. S. (Ed). The Economics of Inequality, Poverty, and Discrimination in the 21st Century, (381). Santa Barbara, CA: ABC-CLIO, LLC. Burke, K. C., Burke, J. D., Rae, D. S., & Regier, D. A. (1991). Comparing age at onset of major depression and other psychiatric disorders by birth cohorts in five US community populations. Archives of General Psychiatry, 48(9), 789-795. doi:10.1001/archpsyc.1991.01810330013002 Burgess, S. R. (2011). Home literacy environments (HLEs) provided to very young children. Early Child Development and Care, 181(4), 445-462. doi:10.1080/03004430903450384 160 Burgess, S. R., Hecht, S. A., & Lonigan, C. J. (2002). Relations of the home literacy environment (HLE) to the development of reading-related abilities: A one-year longitudinal study. Reading Research Quarterly, 37(4), 408-426. doi:10.1598/RRQ.37.4.4 Burchinal, M., McCartney, K., Steinberg, L., Crosnoe, R., Friedman, S. L., McLoyd, V., & Pianta, R. (2011). Examining the Black–White achievement gap among low-income children using the NICHD study of early child care and youth development. Child development, 82(5), 1404-1420. Bus, A. G. (2001). Joint caregiver-child storybook reading: A route to literacy development. In Neuman, S., & Dickson, D. (Eds). Handbook of Early Literacy Research,179-191. New York: Guilford Press. Bus, A., Van Ijzendoorn, M., & Pellegrini, A. (1995). Joint Book Reading Makes for Success in Learning to Read: A Meta-Analysis on Intergenerational Transmission of Literacy. Review of Educational Research, 65(1), 1-21. Retrieved from http://www.jstor.org/stable/1170476 Caldwell, B. M., & Bradley, R. H. (1984). Home observation for measurement of the environment. Little Rock: University of Arkansas at little Rock. Carlson, B. L. & Mathematica Policy Research (2008). Fragile Families and Child Wellbeing Study: Methodology for constructing mother, father, and couple weights for core Telephone public survey data waves 1-4. Retrieved from: http://www.fragilefamilies.princeton.edu/sites/fragilefamilies/files/ff_const_wgts.pdf Carter, A. S., Garrity-Rokous, F. E., Chazan-Cohen, R., Little, C., & Briggs-Gowan, M. J. (2001). Maternal depression and comorbidity: predicting early parenting, attachment security, and toddler social-emotional problems and competencies. Journal of the American Academy of Child & Adolescent Psychiatry, 40(1), 18-26. doi: 10.1097/00004583-200101000-00012 Cartwright, K. B. (2012). Insights from cognitive neuroscience: The importance of executive function for early reading development and education. Early Education & Development, 23(1), 24. doi:10.1080/10409289.2011.615025 Catts, H. W., Fey, M. E., Zhang, X., & Tomblin, J. B. (2001). Estimating the risk of future reading difficulties in kindergarten children: A research-based model and its clinical implementation. Language, Speech, and Hearing Services in Schools, 32(1), 38-50. doi:10.1044/0161-1461(2001/004) Chall, J. S., & Jacobs, V. A. (2003). Poor children's fourth-grade slump. American Educator,27(1), 14. Retrieved November 23, 2016, from http://www.aft.org/periodical/american-educator/spring-2003/classic-study-poor- childrens-fourth-grade-slump 161 Christian, K., Morrison, F. J., & Bryant, F. B. (1998). Predicting kindergarten academic skills: Interactions among child care, maternal education, and family literacy environments. Early Childhood Research Quarterly, 13(3), 501-521. doi:10.1016/S0885- 2006(99)80054-4 Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2013). Applied Multiple Regression/correlation Analysis for the Behavioral Sciences. Routledge. Common Core State Standards Initiative. (2010). Common Core State Standards for Mathematics. Washington, DC: National Governors Association Center for Best Practices and the Council of Chief State School Officers. Conger, R. D., & Conger, K. J. (2008). Understanding the processes through which economic hardship influences families and children. In Crane, D. R., & Heaton, T. B. (Eds). Handbook of Families and Poverty, 64-81. Thousand Oaks, CA: Sage Publications. Connell, C. M., & Prinz, R. J. (2002). The impact of childcare and parent–child interactions on school readiness and social skills development for low-income African American children. Journal of School Psychology, 40(2), 177-193. doi: 10.1016/S0022- 4405(02)00090-0 Crockenberg, S. C., & Leerkes, E. M. (2003). Parental acceptance, postpartum depression, and maternal sensitivity: Mediating and moderating processes. Journal of Family Psychology, 17(1), 80-93. doi:10.1037/0893-3200.17.1.80 Crooks, D. L. (1995). American children at risk: Poverty and its consequences for children's health, growth, and school achievement. American Journal of Physical Anthropology, 38(S21), 57-86. doi: 10.1002/ajpa.1330380605 Cunningham, A., & Stanovich, K. (1997). Early reading acquisition and its relation to reading experience and ability 10 years later. Developmental Psychology, 33(6), 934-945. doi:10.1037//0012-1649.33.6.934 Cunningham, A., & Stanovich, K. (1991). tracking the unique effects of print exposure in children - associations with vocabulary, general knowledge, and spelling. Journal of Educational Psychology, 83(2), 264-274. doi:10.1037//0022-0663.83.2.264 Davé, S., Petersen, I., Sherr, L., & Nazareth, I. (2010). Incidence of maternal and paternal depression in primary care: A cohort study using a primary care database. Archives of Pediatrics & Adolescent Medicine, 164(11), 1038-1044. doi:10.1001/archpediatrics.2010.184 Davidse, N. J., de Jong, M. T., Bus, A. G., Huijbregts, S. C. J., & Swaab, H. (2011). Cognitive and environmental predictors of early literacy skills. Reading and Writing, 24(4), 395- 412. doi:10.1007/s11145-010-9233-3 162 de Jong, P. F., & Leseman, P. P. M. (2001). Lasting effects of home literacy on reading achievement in school. Journal of School Psychology, 39(5), 389-414. doi:10.1016/S0022-4405(01)00080-2 DeBaryshe, B. D., & Binder, J. C. (1994). Development of an instrument for measuring parental beliefs about reading aloud to young children. Perceptual and Motor Skills, 78(3), 1303- 1311. Dearing, E., McCartney, K., & Taylor, B. A. (2009). Does higher quality early child care promote low-income children’s math and reading achievement in middle childhood? Child Development, 80(5), 1329-1349. doi:10.1111/j.1467- 8624.2009.01336.x Dearing, E., McCartney, K., & Taylor, B. A. (2006). Within-child associations between family income and externalizing and internalizing problems. Developmental Psychology, 42(2), 237. doi: 10.1037/0012-1649.42.2.237 De Sonneville, L. M. J. (2005). Amsterdam Neuropsychological Tasks: scientific and clinical applications. Tijdschrift voor Neuropsychologie, 1, 27-41. Dilworth-Bart, J. E. (2012). Does executive function mediate SES and home quality associations with academic readiness?. Early Childhood Research Quarterly, 27(3), 416-425. doi: 10.1016/j.ecresq.2012.02.002 Dickinson, D. K., Golinkoff, R. M., & Hirsh-Pasek, K. (2010). Speaking out for language: Why language is central to reading development. Educational Researcher, 39(4), 305-310. doi:10.3102/0013189X10370204 Dickinson, D. K., McCabe, A., Anastasopoulos, L., Peisner-Feinberg, E. S., & Poe, M. D. (2003). The comprehensive language approach to early literacy: The interrelationships among vocabulary, phonological sensitivity, and print knowledge among preschool-aged children. Journal of Educational Psychology, 95(3), 465-481. doi:10.1037/0022- 0663.95.3.465 Dickerson, A., & Popli, G. K. (2016). Persistent poverty and children's cognitive development: Evidence from the UK millennium cohort study. Journal of the Royal Statistical Society: Series A (Statistics in Society), 179(2), 535-558. doi:10.1111/rssa.12128 Downer, J. T., & Pianta, R. C. (2006). Academic and cognitive functioning in first grade: Associations with earlier home and child care predictors and with concurrent home and classroom experiences. School Psychology Review, 35(1), 11. Retrieved from http://www.jstor.org/stable/3696295 163 Downey, G., & Coyne, J. C. (1990). Children of depressed parents: An integrative review. Psychological Bulletin, 108(1), 50-76. doi:10.1037/0033-2909.108.1.50 Drummond, K. V., & Stipek, D. (2004). Low-income parents' beliefs about their role in children's academic learning. The Elementary School Journal, 197-213. Retrieved from http://www.jstor.org/stable/3202949 Duncan, G. J., & Brooks-Gunn, J. (1997). Consequences of growing up poor. New York, NY: Russell Sage Foundation. Duncan, G. J., Yeung, W. J., Brooks-Gunn, J., & Smith, J. R. (1998). How much does childhood poverty affect the life chances of children? American Sociological Review, 63(3), 406- 423. Retrieved from http://www.jstor.org/stable/2657556 Duncan, G. J., & Brooks-Gunn, J. (2000). Family poverty, welfare reform, and child development. Child Development, 71(1), 188-196. doi: 10.1111/1467-8624.00133 Duncan, G. J., & Magnuson, K. (2013). The long reach of early childhood poverty. In Duncan, G. J., & Magnuson, K. (Eds). Economic Stress, Human Capital, and Families in Asia. Economic stress, human capital, and families in Asia (57-70). Netherlands: Springer. Duncan, G. J., Dowsett, C. J., Claessens, A., Magnuson, K., Huston, A. C., Klebanov, P., et al. (2007). School readiness and later achievement. Developmental Psychology, 43(6), 1428–1446. doi:10.1037/0012-1649.43.6.1428 Dunn, L. M., & Dunn, L. M. (1997). Examiner's manual for the PPVT-III peabody picture vocabulary test: Form IIIA and Form IIIB. AGS. DuPaul, G. J., Reid, R., Anastopoulos, A. D., Lambert, M. C., Watkins, M. W., & Power, T. J. (2016). Parent and teacher ratings of attention-deficit/hyperactivity disorder symptoms: Factor structure and normative data. Psychological Assessment, 28(2), 214-225. doi:10.1037/pas0000166 Du Rocher Schudlich, Tina D, & Cummings, E. M. (2007). Parental dysphoria and children's adjustment: Marital conflict styles, children's emotional security, and parenting as mediators of risk. Journal of Abnormal Child Psychology, 35(4), 627. doi:10.1007/s10802-007-9118-3 Dziak, J. J., Coffman, D. L., Lanza, S. T., & Li, R. (2012). Sensitivity and specificity of information criteria (technical report# 12e119). State College, PA: The Methodology Center, Pennsylvania State University. Ebejer, J. L., Coventry, W. L., Byrne, B., Willcutt, E. G., Olson, R. K., Corley, R., . . . U tbildningsvetenskap. (2010). Genetic and environmental influences on inattention, hyperactivity-impulsivity, and reading: Kindergarten to grade 2. Scientific Studies of Reading, 14(4), 293-316. doi:10.1080/10888430903150642 164 Engel, G. L. (1979). The biopsychosocial model and the education of health professionals. General Hospital Psychiatry, 1(2), 156-165. doi: 10.1016/0163- 8343(79)90062-8 Evans, M. A., Shaw, D., & Bell, M. (2000). Home literacy activities and their influence on early literacy skills. Canadian Journal of Experimental Psychology/Revue canadienne de psychologie expérimentale, 54(2), 65. doi: 10.1037/h0087330 Evans, J., Heron, J., Francomb, H., Oke, S., & Golding, J. (2001). Cohort study of depressed mood during pregnancy and after childbirth. BMJ: British Medical Journal, 323(7307), 257-260. doi:10.1136/bmj.323.7307.257 Fan, J., McCandliss, B. D., Sommer, T., Raz, A., & Posner, M. I. (2002). Testing the efficiency and independence of attentional networks. Journal of Cognitive Neuroscience, 14(3), 340-347. doi:10.1162/089892902317361886 Fantuzzo, J., Bulotsky-Shearer, R., McDermott, P. A., McWayne, C., Frye, D., & Perlman, S. (2007, March). Investigation of dimensions of social-emotional classroom behavior and school readiness for low-incom urban preschool children. School Psychology Review, 36(1), 44+. Retrieved from http://ezproxy.msu.edu.proxy1.cl.msu.edu.proxy2.cl.msu.edu.proxy1.cl.msu.edu.proxy2.c l.msu.edu/login?url=http://go.galegroup.com.proxy1.cl.msu.edu.proxy2.cl.msu.edu/ps/i.d o?p=ITOF&sw=w&u=msu_main&v=2.1&it=r&id=GALE%7CA161846311&sid=sum on&asid=aadf81e46c53b564b964ccf8c358d28e Farah, M. J., Shera, D. M., Savage, J. H., Betancourt, L., Giannetta, J. M., Brodsky, N. L., ... & Hurt, H. (2006). Childhood poverty: Specific associations with neurocognitive development. Brain research, 1110(1), 166-174. doi:10.1016/j.brainres.2006.06.072 Farstrup, A. E., Samuels, S. J., & International Reading Association. (2002). What research has to say about reading instruction, third edition Distributed by ERIC Clearinghouse. Feldman, R., Granat, A., Pariente, C., Kanety, H., Kuint, J., & Gilboa-Schechtman, E. (2009). Maternal depression and anxiety across the postpartum year and infant social engagement, fear regulation, and stress reactivity. Journal of the American Academy of Child & Adolescent Psychiatry, 48(9), 919-927. doi: 10.1097/CHI.0b013e3181b21651 Fergusson, D. M., Horwood, L. J., & Boden, J. M. (2008). The transmission of social inequality: Examination of the linkages between family socioeconomic status in childhood and educational achievement in young adulthood. Research in Social Stratification and Mobility, 26(3), 277-295. doi:10.1016/j.rssm.2008.05.001 Froiland, J. M., Peterson, A., & Davison, M. L. (2013).The long-term effects of early parent involvement and parent expectation in the USA. School Psychology International, 34(1), 33-50. doi: 10.1177/0143034312454361 165 Frijters, J. C., Barron, R. W., & Brunello, M. (2000). Direct and mediated influences of home literacy and literacy interest on prereaders' oral vocabulary and early written language skill. Journal of Educational Psychology, 92(3), 466. doi: 10.1037/0022-0663.92.3.466 Foster, M. A., Lambert, R., Abbott-Shim, M., McCarty, F., & Franze, S. (2005). A model of home learning environment and social risk factors in relation to children's emergent l iteracy and social outcomes. Early Childhood Research Quarterly, 20(1), 13-36. doi:10.1016/j.ecresq.2005.01.006 Foster, W. A., & Miller, M. (2007). Development of the literacy achievement gap: A longitudinal study of kindergarten through third grade. Language, Speech, and Hearing Services in Schools, 38(3), 173-181. Foulin, J. N. (2005). Why is letter-name knowledge such a good predictor of learning to read?. Reading and Writing, 18(2), 129-155. doi:10.1007/s11145-004-5892-2 Fountas, I. C., & Pinnell, G. S. (1996). Guided reading: Good first teaching for all children. Portsmouth, NH: Heinemann. Fox, P. T., Pardo, J. V., & Raichle, M. E. (1991). Localization of a human system for sustained attention by positron emission tomography. Nature, 349(6304), 61-64. doi:10.1038/349061a0 Foy, J. G., & Mann, V. (2003). Home literacy environment and phonological awareness in preschool children: Differential effects for rhyme and phoneme awareness. Applied Psycholinguistics, 24(1), 59-88. doi:10.1017/S0142716403000043 Fuchs, L. S., Fuchs, D., Hosp, M. K., & Jenkins, J. R. (2001). Oral reading fluency as an indicator of reading competence: A theoretical, empirical, and historical analysis. Scientific Studies of Reading, 5(3), 239-256. doi:10.1207/S1532799XSSR0503_3 Gartstein, M. A., & Fagot, B. I. (2003). Parental depression, parenting and family adjustment, and child effortful control: Explaining externalizing behaviors for preschool children. Journal of Applied Developmental Psychology, 24(2), 143-177. doi:10.1016/S0193-3973(03)00043-1 Germanò, E., Gagliano, A., & Curatolo, P. (2010). Comorbidity of ADHD and dyslexia. Developmental Neuropsychology, 35(5), 475-493. doi:10.1080/87565641.2010.494748 Gershoff, E. T., Aber, J. L., Raver, C. C., & Lennon, M. C. (2007). Income is not enough: Incorporating material hardship into models of income associations with parenting and child development. Child Development, 78(1), 70-95. doi:10.1111/j.1467- 8624.2007.00986.x 166 Gibb, S., Fergusson, D., & Horwood, L. (2012). Childhood family income and life outcomes in adulthood: Findings from a 30-year longitudinal study in new zealand. Social Science & Medicine, 74(12), 1979-1986. doi:10.1016/j.socscimed.2012.02.028 Gizer, I. R., Ficks, C., & Waldman, I. D. (2009). Candidate gene studies of ADHD: A meta- analytic review. Human Genetics, 126(1), 51-90. doi:10.1007/s00439-009-0694-x Gizer, I. R., & Waldman, I. D. (2012). Double dissociation between lab measures of inattention and impulsivity and the dopamine transporter gene (DAT1) and dopamine D4 receptor gene (DRD4). Journal of Abnormal Psychology, 121(4), 1011. doi:10.1037/a0028225 Goodman, S. H., Rouse, M. H., Connell, A. M., Broth, M. R., Hall, C. M., & Heyward, D. (2011). Maternal depression and child psychopathology: A meta-analytic review. Clinical Child and Family Psychology Review, 14(1), 1-27. doi:10.1007/s10567-010-0080-1 Griffin, E. A., & Morrison, F. J. (1997). The unique contribution of home literacy environment to differences in early literacy skills. Early Child Development and Care, 127(1), 233. doi:10.1080/0300443971270119 Grigorenko, E. L., Kornev, A. N., Rakhlin, N., & Krivulskaya, S. (2011). Reading-related skills, reading achievement, and inattention: A correlational study. Journal of Cognitive Education and Psychology, 10(2), 140. Gross, H. E., Shaw, D. S., Burwell, R. A., & Nagin, D. S. (2009). Transactional processes in child disruptive behavior and maternal depression: A longitudinal study from early childhood to adolescence. Development and psychopathology, 21(01), 139-156. doi:10.1017/S0954579409000091 Guo, G., & Harris, K. M. (2000). The mechanisms mediating the effects of poverty on children's intellectual development. Demography, 37(4), 431-447. doi:10.1353/dem.2000.0005 Haak, J., Downer, J., & Reeve, R. (2012). Home literacy exposure and early language and literacy skills in children who struggle with behavior and attention problems. Early Education & Development, 23(5), 728. doi:10.1080/10409289.2011.565721 Haskins, R., & Rouse, C. (2005). Closing Achievement Gaps. The Future of Children, Policy Brief, Spring 2005. Princeton, N.J.: Princeton-Brookings; Heckman, James J. and Dimitriy V. Masterov. 2007. “The Productivity Argument for Investing in Young Children.” Retrieved January 23, 2009, from http://jenni.uchicago.edu/human- inequality/papers/Heckman_final_all_wp_2007-03-22c_jsb.pdf; Lynch, Robert G. 2004. Exceptional Returns: Economic, Fiscal and Social Benefits of Investment in Early Childhood Development, Washington, D.C.: Economic Policy Institute. 167 Hart, S. A., Petrill, S. A., DeThorne, L. S., Deater-Deckard, K., Thompson, L. A., Schatschneider, C., & Cutting, L. E. (2009). Environmental influences on the longitudinal covariance of expressive vocabulary: measuring the home literacy environment in a genetically sensitive design. Journal of Child Psychology and Psychiatry, 50(8), 911-919. doi:10.1111/j.1469-7610.2009.02074.x Hasin, D. S., Goodwin, R. D., Stinson, F. S., & Grant, B. F. (2005). Epidemiology of major depressive disorder: results from the National Epidemiologic Survey on Alcoholism and Related Conditions. Archives of General Psychiatry, 62(10), 1097-1106. doi:10.1001/archpsyc.62.10.1097 Haveman, R., & Wolfe, B. (1994). Succeeding generations: On the effects of investments in children. New York, NY: Russell Sage Foundation. Healy, E., Reichenberg, A., Nam, K., Allin, M., Walshe, M., Rifkin, L., . . . Nosarti, C. (2013). Preterm birth and adolescent social functioning-alterations in emotion-processing brain areas. Journal of Pediatrics, 163(6), 1596-1604. doi:10.1016/j.jpeds.2013.08.011 Hernandez, D. J. (2011). Double Jeopardy: How Third-Grade Reading Skills and Poverty Influence High School Graduation. Annie E. Casey Foundation. Retrieved from: http://files.eric.ed.gov/fulltext/ED518818.pdf Holmes, J., Payton, A., Barrett, J., Harrington, R., McGuffin, P., Owen, M., . . . Thapar, A. (2002). Association of DRD4 in children with ADHD and comorbid conduct problems. American Journal of Medical Genetics, 114(2), 150-153. doi:10.1002/ajmg.10149 Hood, M., Conlon, E., & Andrews, G. (2008). Preschool home literacy practices and children's literacy development: A longitudinal analysis. Journal of Educational Psychology,100(2), 252-271. doi:10.1037/0022-0663.100.2.252 Hudson, R. F., Lane, H. B., & Pullen, P. C. (2005). Reading fluency assessment and instruction: What, why, and how? The Reading Teacher, 58(8), 702-714. doi:10.1598/RT.58.8.1 Hsiung, G. Y. R., Kaplan, B. J., Petryshen, T. L., Lu, S., & Field, L. L. (2004). A dyslexia susceptibility locus (DYX7) linked to dopamine D4 receptor (DRD4) region on chromosome 11p15. 5. American Journal of Medical Genetics Part B: Neuropsychiatric Genetics, 125(1), 112-119. doi: 10.1002/ajmg.b.20082 Hulme, D., & Shepherd, A. (2003). Conceptualizing chronic poverty. World Development,31(3), 403-423. doi:10.1016/S0305-750X(02)00222-X IBM Corp. Released 2015. IBM SPSS Statistics for Windows, Version 23.0. Armonk, NY: IBM Corp. 168 Jimerson, S., Egeland, B., Sroufe, L. A., & Carlson, B. (2000). A prospective longitudinal study of high school dropouts examining multiple predictors across development. Journal of school psychology, 38(6), 525-549.doi: 10.1016/S0022-4405(00)00051-0 Jordan, N. C., Kaplan, D., Ramineni, C., & Locuniak, M. N. (2009). Early math matters: kindergarten number competence and later mathematics outcomes. Developmental Psychology, 45(3), 850. Kamhi, A. G., & Catts, H. W. (2013). Language and Reading Disabilities: Pearson New International Edition. Pearson Higher Ed. Kass, R. E., & Raftery, A. E. (1995). Bayes factors. Journal of the American Statistical Association, 90(430), 773-795. Kegel, C. A. T., & Bus, A. G. (2013). Links between DRD4, executive attention, and alphabetic skills in a nonclinical sample. Journal of Child Psychology and Psychiatry, 54(3), 305- 312. doi:10.1111/j.1469-7610.2012.02604.x Kessler, R. C., Andrews, G., Mroczek, D., Ustun, B., & Wittchen, H. (1998). The world health organization composite international diagnostic interview short-form (CIDI- SF). International Journal of Methods in Psychiatric Research, 7(4), 171-185. doi:10.1002/mpr.47 Kieling, C., Roman, T., Doyle, A. E., Hutz, M. H., & Rohde, L. A. (2006). Association between DRD4 gene and performance of children with ADHD in a test of sustained attention. Biological Psychiatry, 60(10), 1163-1165. doi:10.1016/j.biopsych.2006.04.027 Kiernan, K. E., & Mensah, F. K. (2009). Poverty, maternal depression, family status and children's cognitive and behavioural development in early childhood: A longitudinal study. Journal of Social Policy, 38(4), 569-588. doi:10.1017/S0047279409003250 Kiernan, K. E., & Mensah, F. K. (2011). Poverty, family resources and children's early educational attainment: The mediating role of parenting. British Educational Research Journal, 37(2), 317. doi:10.1080/01411921003596911 Kim-Cohen, J., Caspi, A., Taylor, A., Williams, B., Newcombe, R., Craig, I. W., & Moffitt, T. E. (2006). MAOA, maltreatment, and gene–environment interaction predicting children's mental health: new evidence and a meta-analysis. Molecular Psychiatry, 11(10), 903-913. doi:10.1038/sj.mp.4001851 Kluger, A. N., Siegfried, Z., & Ebstein, R. P. (2002). A meta-analysis of the association between DRD4 polymorphism and novelty seeking. Molecular Psychiatry, 7(7), 712-717. doi:10.1038/sj.mp.4001082 169 Kohl, G. O., Lengua, L. J., & McMahon, R. J. (2000). Parent involvement in school conceptualizing multiple dimensions and their relations with family and demographic risk factors. Journal of School Psychology, 38(6), 501-523. doi:10.1016/S0022- 4405(00)00050-9 Koller, K., Brown, T., Spurgeon, A., & Levy, L. (2004). Recent developments in low-level lead exposure and intellectual impairment in children. Environmental health perspectives, 987-994. Retrieved from http://www.jstor.org/stable/3838099 Krishnakumar, A., & Buehler, C. (2000). Interparental conflict and parenting behaviors: A meta- analytic review. Family Relations, 49(1), 25-44. doi:10.1111/j.1741-3729.2000.00025.x Kurstjens, S., & Wolke, D. (2001). Effects of maternal depression on cognitive development of c children over the first 7 years of life. Journal of Child Psychology and Psychiatry, 42(05), 623-636. doi: 10.1017/S0021963001007296 Lanphear, B. P., Byrd, R. S., Auinger, P., & Hall, C. B. (1997). Increasing prevalence of recurrent otitis media among children in the united states. Pediatrics, 99(3), e1-e1. doi:10.1542/peds.99.3.e1 Lanphear, B. P., Hornung, R., Khoury, J., Yolton, K., Baghurst, P., Bellinger, D. C., ... & Rothenberg, S. J. (2005). Low-level environmental lead exposure and children's intellectual function: an international pooled analysis. Environmental Health Perspectives, 894-899. Retrieved from http://www.jstor.org/stable/3436211 Lee, V. E., & Burkam, D. T. (2002). Inequality at the starting gate: Social background differences in achievement as children begin school. Economic Policy Institute: Washington, DC. Lennon, M. C., Blome, J., & English, K. (2001, April). Depression and low-income women: Challenges for TANF and welfare-to-work policies and programs. Research Forum on Children, Families, and the New Federalism, National Center for Children in Poverty. Mailman School of Public Health, Columbia University. Levy, B., Gong, Z., Hessels, S., Evans, M., & Jared, D. (2006). Understanding print: Early reading development and the contributions of home literacy experiences. Journal of Experimental Child Psychology, 93(1), 63-93. doi:10.1016/j.jeep.2005.07.003 Liu, C. H., & Tronick, E. (2013). Rates and predictors of postpartum depression by race and ethnicity: Results from the 2004 to 2007 new york city PRAMS survey (pregnancy risk assessment monitoring system). Maternal and Child Health Journal, 17(9), 1599-1610. doi:10.1007/s10995-012-1171-z Lyytinen, P., Laakso, M. L., & Poikkeus, A. M. (1998). Parental contribution to child’s early language and interest in books. European Journal of Psychology of Education, 13(3), 297-308. doi:10.1007/BF03172946 170 Lipina, S. J., & Colombo, J. A. (2009). Poverty and brain development during childhood: An approach from cognitive psychology and neuroscience (1st ed.). Washington, DC: American Psychological Association. Linver, M. R., Brooks-Gunn, J., & Kohen, D. E. (2002). Family processes as pathways from income to young children's development. Developmental Psychology, 38(5), 719-734. Retrieved from http://ezproxy.msu.edu.proxy1.cl.msu.edu/login?url=http://search.proquest.com.proxy1.c l.msu.edu/docview/224543173?accountid=12598 Lonigan, C., Burgess, S., & Anthony, J. (2000). Development of emergent literacy and early reading skills in preschool children: Evidence from a latent-variable longitudinal study. Developmental Psychology, 36(5), 596-613. doi:10.1037//0012-1649.36.5.596 Lonigan, C. J., & Shanahan, T. (2009). Developing Early Literacy: Report of the National Early Literacy Panel. Executive Summary. A Scientific Synthesis of Early Literacy Development and Implications for Intervention. National Institute for Literacy. Lovejoy, M. C., Graczyk, P. A., O'Hare, E., & Neuman, G. (2000). Maternal depression and parenting behavior: A meta-analytic review. Clinical Psychology Review, 20(5), 561. doi.org.proxy2.cl.msu.edu/10.1016/S0272-7358(98)00100-7 Lupien, S. J., King, S., Meaney, M. J., & McEwen, B. S. (2001). Can poverty get under your skin? basal cortisol levels and cognitive function in children from low and high socioeconomic status. Development and Psychopathology, 13(3), 653-676. doi:10.1017/S0954579401003133 National Institute of Child Health and Human Development (NICHD) Early Child Care Research Network. (2003). Do children’s attention processes mediate the link between family predictors and school readiness? Developmental Psychology, 39(3), 581-593. doi:10.1037/0012-1649.39.3.581 National Reading Panel (US), National Institute of Child Health, & Human Development (US). (2000). Report of the national reading panel: Teaching children to read: An evidence- based assessment of the scientific research literature on reading and its implications for reading instruction: Reports of the subgroups. National Institute of Child Health and Human Development, National Institutes of Health. Mammarella, I. C., Ghisi, M., Bomba, M., Bottesi, G., Caviola, S., Broggi, F., & Nacinovich, R. (2016). Anxiety and depression in children with nonverbal learning disabilities, reading disabilities, or typical development. Journal of Learning Disabilities, 49(2), 130-139. doi:10.1177/0022219414529336 171 Manly, T., Anderson, V., Nimmo-Smith, I., Turner, A., Watson, P., & Robertson, I. H. (2001). The differential assessment of children's attention: The test of everyday attention for children (TEA-ch), normative sample and ADHD performance. The Journal of Child Psychology and Psychiatry and Allied Disciplines, 42(8), 1065-1081. doi:10.1017/S0021963001007909 Marino, C., Giorda, R., Vanzin, L., Molteni, M., Lorusso, M. L., Nobile, M., ... & Battaglia, M. (2003). No evidence for association and linkage disequilibrium between dyslexia and markers of four dopamine-related genes. European child & adolescent psychiatry, 12(4), 198-202. doi:10.1007/s00787-003-0332-4 Martin, A., Razza, R. A., & Brooks-Gunn, J. (2012). Sustained attention at age 5 predicts attention-related problems at age 9. International Journal of Behavioral Development, 36(6), 413-419. doi:10.1177/0165025412450527 Mason, Z. S., Briggs, R. D., & Silver, E. J. (2011). Maternal attachment feelings mediate between maternal reports of depression, infant social–emotional development, and parenting stress. Journal of Reproductive and Infant Psychology, 29(4), 382- 394.doi:10.1080/02646838.2011.629994 Maughan, B., Rowe, R., Loeber, R., & Stouthamer-Loeber, M. (2003). Reading problems and depressed mood. Journal of Abnormal Child Psychology, 31(2), 219-229. doi:10.1023/A:1022534527021 McCardle, P., Scarborough, H. S., & Catts, H. W. (2001). Predicting, explaining, and preventing children's reading difficulties. Learning Disabilities Research & Practice, 16(4), 230- 239. doi:10.1111/0938-8982.00023 McClelland, M. M., Cameron, C. E., Connor, C. M., Farris, C. L., Jewkes, A. M., & Morrison, F. J. (2007). Links between behavioral regulation and preschoolers' literacy, vocabulary, and math skills. Developmental Psychology, 43(4), 947-959. doi:10.1037/0012- 1649.43.4.947 McCracken, J. T., Smalley, S. L., McGough, J. J., Crawford, L., Del'Homme, M., Cantor, R. M., . . . Nelson, S. F. (2000). Evidence for linkage of a tandem duplication polymorphism upstream of the dopamine D4 receptor gene (DRD4) with attention deficit hyperactivity disorder (ADHD). Molecular Psychiatry, 5(5), 531-536. doi:10.1038/sj.mp.4000770 McLearn, K. T., Minkovitz, C. S., Strobino, D. M., Marks, E., & Hou, W. (2006). Maternal depressive symptoms at 2 to 4 months post partum and early parenting practices. Archives of Pediatrics & Adolescent Medicine, 160(3), 279-284. doi:10.1001/archpedi.160.3.279 McLennan, J. D., & Kotelchuck, M. (2000). Parental prevention practices for young children in the context of maternal depression. Pediatrics, 105(5), 1090-1095. doi:10.1542/peds.105.5.1090 172 McLeod, J., & Nonnemaker, J. (2000). Poverty and Child Emotional and Behavioral Problems: Racial/Ethnic Differences in Processes and Effects. Journal of Health and Social Behavior, 41(2), 137-161. Retrieved from http://www.jstor.org/stable/2676302 McLeod, J. D., & Kessler, R. C. (1990). Socioeconomic status differences in vulnerability to undesirable life events. Journal of Health and Social Behavior, 31(2), 162-172. McLoyd, V. C., Jayaratne, T. E., Ceballo, R., & Borquez, J. (1994). Unemployment and work interruption among african american single mothers: Effects on parenting and adolescent socioemotional functioning. Child Development, 65(2), 562-589. doi:10.1111/j.1467- 8624.1994.tb00769.x McLoyd, V. C. (1998). Socioeconomic disadvantage and child development. American Psychologist, 53(2), 185-204. doi:10.1037/0003-066X.53.2.185 Mensah, F. K., & Kiernan, K. E. (2010). Parents’ mental health and children’s cognitive and social development. Social Psychiatry and Psychiatric Epidemiology, 45(11), 1023-1035. Mistry, R. S., Biesanz, J. C., Chien, N., Howes, C., & Benner, A. D. (2008). Socioeconomic status, parental investments, and the cognitive and behavioral outcomes of low-income children from immigrant and native households. Early Childhood Research Quarterly, 23(2), 193-212. doi:10.1016/j.ecresq.2008.01.002 Mullineaux, P. Y., & DiLalla, L. F. (2015). Genetic influences on peer and family relationships across adolescent development: Introduction to the special issue. Journal of Youth and Adolescence, 44(7), 1347-1359. doi:10.1007/s10964-015-0306-0 Murnane, R. J. (2013). U.S. high school graduation rates: Patterns and explanations. Journal of Economic Literature, 51(2), 370-422. doi:10.1257/jel.51.2.370 Muthén, L. K., & Muthén, B. O. (2012). Mplus statistical modeling software: Release 7.0. Los Angeles, CA: Muthén & Muthén. Najman, J. M., Hayatbakhsh, M. R., Heron, M. A., Bor, M., & Williams, G. M. (2009). The impact of episodic and chronic poverty on child cognitive development. J Pediatr, 154(2), 284-289.e1. doi:10.1016/j.jpeds.2008.08.052 National Early Literacy Panel (NELP). (2008). Developing early literacy: Report of the National Early Literacy Panel. Washington, DC: National Institute for Literacy. NICHD Early Child Care Research Network. (2002). Early Child Care and Children's Development Prior to School Entry: Results from the NICHD Study of Early Child Care. American Educational Research Journal, 39(1), 133-164. Retrieved from http://www.jstor.org/stable/3202474 173 Niklas, F., & Schneider, W. (2013). Home literacy environment and the beginning of reading and spelling. Contemporary Educational Psychology, 38(1), 40-50. doi:10.1016/j.cedpsych.2012.10.001 Nisbett, R. E., & Wilson, T. D. (1977). The halo effect: Evidence for unconscious alteration of judgments. Journal of Personality and Social Psychology, 35(4), 250-256. doi:10.1037//0022-3514.35.4.250 Noble, K. G., Houston, S. M., Brito, N. H., Bartsch, H., Kan, E., Kuperman, J. M., . . . Sowell, E. R. (2015). Family income, parental education and brain structure in children and adolescents. Nature Neuroscience, 18(5), 773-778. doi:10.1038/nn.3983 O'hara, M. W., & Swain, A. M. (1996). Rates and risk of postpartum depression—a meta- analysis. International Review of Psychiatry, 8(1), 37-54. Olson, R. K., Keenan, J. M., Byrne, B., Samuelsson, S., Coventry, W. L., Corley, R., . . . Utbildningsvetenskap. (2011). Genetic and environmental influences on vocabulary and reading development. Scientific Studies of Reading, 15(1), 26-46. doi:10.1080/10888438.2011.536128 Paciorek, C., Stevens, G., Finucane, M., Ezzati, M., Nutrition Impact Model Study Group Child, & Nutrition Impact Model Study Group (Child Growth). (2013). Children's height and weight in rural and urban populations in low-income and middle-income countries: A systematic analysis of population-representative data. Lancet Global Health, 1(5), E300- E309. doi:10.1016/S2214-109X(13)70109-8 Pachter, L. M., Auinger, P., Palmer, R., & Weitzman, M. (2006). Do parenting and the home environment, maternal depression, neighborhood, and chronic poverty affect child behavioral problems differently in different racial-ethnic groups? Pediatrics, 117(4), 1329-1338. doi:10.1542/peds.2005-1784 Paulson, J. F., Keefe, H. A., & Leiferman, J. A. (2009). Early parental depression and child language development. Journal of Child Psychology and Psychiatry, 50(3), 254-262. Payne, A. C., Whitehurst, G. J., & Angell, A. L. (1994). The role of home literacy environment in the development of language ability in preschool children from low-income families. Early Childhood Research Quarterly, 9(3), 427-440. doi:10.1016/0885- 2006(94)90018-3 Peterson, S.M, & Albers, B, A. (2001). Effects of poverty and maternal depression on early child development. Child Development, 72(6), 1794-1813. doi:10.1111/1467-8624.00379 Petersen, S. E., & Posner, M. I. (2012;2011;). The attention system of the human brain: 20 years after. Annual Review of Neuroscience, 35(1), 73-89. doi:10.1146/annurev-neuro-062111- 150525 174 Pham, A. V. (2016). Differentiating behavioral ratings of inattention, impulsivity, and hyperactivity in children: Effects on reading achievement. Journal of Attention Disorders, 20(8), 674-683. Pratt, L. A., & Brody, D. J. (2008). Depression in the United States household population, 2005- 2006 (pp. 1-8). US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Health Statistics. Pennington, B. F. (2006). From single to multiple deficit models of developmental disorders. Cognition, 101(2), 385-413. doi:10.1016/j.cognition.2006.04.008 Pérez-Padilla, J., Menéndez, S., & Lozano, O. (2015). Validity of the parenting stress index short form in a sample of at-risk mothers. Evaluation Review, 39(4), 428-446. doi:10.1177/0193841X15600859 Phillips, B. M., & Lonigan, C. J. (2009). Variations in the home literacy environment of preschool children: A cluster analytic approach. Scientific Studies of Reading, 13(2), 146- 174. doi:10.1080/10888430902769533 Piasta, S. B., Justice, L. M., McGinty, A. S., & Kaderavek, J. N. (2012). Increasing young Children’s contact with print during shared reading: Longitudinal effects on literacy achievement. Child Development, 83(3), 810-820. doi:10.1111/j.1467-8624.2012.01754.x Piasta, S. B., & Wagner, R. K. (2010). Developing early literacy skills: A meta-analysis of alphabet learning and instruction. Reading Research Quarterly, 45(1), 8-38. doi:10.1598/RRQ.45.1.2 Pikulski, J. J., & Chard, D. J. (2005). Fluency: Bridge between decoding and reading comprehension. The Reading Teacher, 58(6), 510-519. doi:10.1598/RT.58.6.2 Pinderhughes, E. E., Dodge, K. A., Bates, J. E., Pettit, G. S., & Zelli, A. (2000). Discipline responses: Influences of parents' socioeconomic status, ethnicity, beliefs about parenting, stress, and cognitive-emotional processes. Journal of Family Psychology: Journal of the Division of Family Psychology of the American Psychological Association, 14(3), 380- 400. doi:10.1037//0893-3200.14.3.380 Pinto, G., Bigozzi, L., Vezzani, C., & Tarchi, C. (2016). Emergent literacy and reading acquisition: A longitudinal study from kindergarten to primary school. European Journal of Psychology of Education, doi:10.1007/s10212-016-0314-9 Plak, R. D., Kegel, C. A. T., & Bus, A. G. (2015). Genetic differential susceptibility in literacy- delayed children: A randomized controlled trial on emergent literacy in kindergarten. Development and Psychopathology, 27(1), 69-79. doi:10.1017/S0954579414001308 175 Posner, M. I., & Rothbart, M. K. (2007). Research on attention networks as a model for the integration of psychological science. Annu. Rev. Psychol., 58, 1-23. doi: 10.1146/annurev.psych.58.110405.085516 Preacher, K. J., Rucker, D. D., & Hayes, A. F. (2007). Addressing moderated mediation hypotheses: Theory, methods, and prescriptions. Multivariate Behavioral Research, 42(1), 185-227. doi:10.1080/00273170701341316 Prochnow, J. E., Tunmer, W. E., & Chapman, J. W. (2013). A longitudinal investigation of the influence of literacy-related skills, reading self-perceptions, and inattentive behaviours on the development of literacy learning difficulties. International Journal of Disability, Development and Education, 60(3), 185-207. doi: 10.1080/1034912X.2013.812188 Puranik, C. S., Lonigan, C. J., & Kim, Y. S. (2011). Contributions of emergent literacy skills to name writing, letter writing, and spelling in preschool children. Early Childhood Research Quarterly, 26(4), 465-474. doi:10.1016/j.ecresq.2011.03.002 Rabiner, D., & Coie, J. D. (2000). Early attention problems and children's reading achievement: A longitudinal investigation. Journal of the American Academy of Child & Adolescent Psychiatry, 39(7), 859-867. doi:10.1097/00004583-200007000-00014 Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1(3), 385-401. doi:10.1177/014662167700100306 Raftery, A. E. (1995). Bayesian model selection in social research. Sociological Methodology, 111-163. Raikes, H., Luze, G., Brooks-Gunn, J., Raikes, H. A., Pan, B. A., Tamis-LeMonda, C. S., . . . Rodriguez, E. T. (2006). Mother-child bookreading in low-income families: Correlates and outcomes during the first three years of life. Child Development, 77(4), 924-953. doi:10.1111/j.1467-8624.2006.00911.x Raver, C., Elizabeth T. Gershoff, & Aber, J. (2007). Testing Equivalence of Mediating Models of Income, Parenting, and School Readiness for White, Black, and Hispanic Children in a National Sample. Child Development, 78(1), 96-115. Retrieved from http://www.jstor.org/stable/4139215 Raver, C. C., Blackburn, E. K., Bancroft, M., & Torp, N. (1999). Relations between effective emotional self-regulation, attentional control, and low-income preschoolers' social competence with peers. Early Education & Development, 10(3), 333-350. doi:10.1207/s15566935eed1003_6 176 Raver, C. C., Blair, C., Garrett-Peters, P., & Family Life Project Key Investigators. (2015). Poverty, household chaos, and interparental aggression predict children's ability to recognize and modulate negative emotions. Development and Psychopathology, 27(3), 695-708. doi:10.1017/S0954579414000935 Rayner, K., & Pollatsek, A. (1989). The psychology of reading. Englewood Cliffs, N.J: Prentice Hall. Razza, R. A., Martin, A., & Brooks-Gunn, J. (2012). The implications of early attentional regulation for school success among low-income children. Journal of Applied Developmental Psychology, 33(6), 311-319. doi:10.1016/j.appdev.2012.07.005 Razza, Rachel A., Anne Martin, and Jeanne Brooks-Gunn. "Associations among Family Environment, Sustained Attention, and School Readiness for Low-Income Children." Developmental psychology 46.6 (2010): 1528. ProQuest. Web. 28 Oct. 2016. Raz, I. S., & Bryant, P. (1990). Social background, phonological awareness and children's reading. British Journal of Developmental Psychology, 8(3), 209-225. doi:10.1111/j.2044-835X.1990.tb00837.x Reichman, N. E., Teitler, J. O., Garfinkel, I., & McLanahan, S. S. (2001). Fragile families: Sample and design. Children and Youth Services Review, 23(4), 303-326. doi:10.1016/S0190-7409(01)00141-4 Reid, R., DuPaul, G. J., Power, T. J., Anastopoulos, A. D., Rogers-Adkinson, D., Noll, M., & Riccio, C. (1998). Assessing culturally different students for attention deficit hyperactivity disorder using behavior rating scales. Journal of Abnormal Child Psychology, 26(3), 187-198. doi:10.1023/A:1022620217886 Reissland, N., Shepherd, J., & Herrera, E. (2003). The pitch of maternal voice: A comparison of mothers suffering from depressed mood and non-depressed mothers reading books to their infants. Journal of Child Psychology and Psychiatry, and Allied Disciplines, 44(2), 255-261. doi:10.1111/1469-7610.00118 Riccio, C. A., Waldrop, J. J., Reynolds, C. R., & Lowe, P. (2001). Effects of stimulants on the continuous performance test (CPT): Implications for CPT use and interpretation. The Journal of Neuropsychiatry and Clinical Neurosciences, 13(3), 326-335. doi:10.1176/appi.neuropsych.13.3.326 Ridge, T. (2002). Childhood poverty and social exclusion: From a child's perspective. Bristol: Policy Press. Roberts, J. E., Burchinal, M. R., Jackson, S. C., Hooper, S. R., Roush, J., Mundy, M., . . . Zeisel, S. A. (2000). Otitis media in early childhood in relation to preschool language and school readiness skills among black children. Pediatrics, 106(4), 725-735. doi:10.1542/peds.106.4.725 177 Rodriguez, E. T., Tamis-LeMonda, C. S., Spellmann, M. E., Pan, B. A., Raikes, H., Lugo-Gil, J., & Luze, G. (2009). The formative role of home literacy experiences across the first three years of life in children from low-income families. Journal of Applied Developmental Psychology, 30(6), 677-694. doi:10.1016/j.appdev.2009.01.003 Roid, G. (1997). Miller L. Leiter international performance scale–revised. Wood Dale, IL: Stoelting. Roisman, G. I., Newman, D. A., Fraley, R. C., Haltigan, J. D., Groh, A. M., & Haydon, K. C. (2012). Distinguishing differential susceptibility from diathesis-stress: Recommendations or evaluating interaction effects. Development and Psychopathology, 24(2), 389. doi:10.1017/S0954579412000065 Rosvold, H. E., Mirsky, A. F., Sarason, I., Bransome, E. D., & Beck, L. H. (1956). A continuous performance test of brain damage. Journal of Consulting Psychology, 20(5), 343-350. doi:10.1037/h0043220 Rowe, K. J. & Rowe, K. S. (1992). The relation between inattentiveness in the classroom and reading achievement (Part B): An exploratory study. Child and Adolescent Psychiatry, (31)2, 357-368. doi:10.1097/00004583-199203000-00026 Rutter, M. (1987). Psychosocial resilience and protective mechanisms. The American Journal of Orthopsychiatry, 57(3), 316-331. doi:10.1111/j.1939-0025.1987.tb03541.x Rutter, M., Moffitt, T. E., & Caspi, A. (2006). Gene-environment interplay and psychopathology: Multiple varieties but real effects. Journal of Child Psychology and Psychiatry, and Allied Disciplines, 47(3-4), 226-261. doi:10.1111/j.1469- 7610.2005.01557.x Ryan, R. M., Fauth, R. C., & Brooks-Gunn, J. (2006). Childhood Poverty: Implications for School Readiness and Early Childhood Education. In Spodek, B., Saracho, O. N., (Eds). Handbook of research on the education of young children, 2nd ed., (323-346). Mahwah, NJ: Lawrence Erlbaum Associates Publishers. Sameroff, A. (2010). A unified theory of development: A dialectic integration of nature and nurture. Child Development, 81(1), 6-22. doi:10.1111/j.1467-8624.2009.01378.x Sarsour, K., Sheridan, M., Jutte, D., Nuru-Jeter, A., Hinshaw, S., & Boyce, W. T. (2011). Family socioeconomic status and child executive functions: The roles of language, home environment, and single parenthood. Journal of the International Neuropsychological Society, 17(1), 120-132. doi:10.1017/S1355617710001335 Sarter, M., Givens, B., & Bruno, J. P. (2001). The cognitive neuroscience of sustained attention: Where top-down meets bottom-up Elsevier B.V. doi:10.1016/S0165-0173(01)00044-3 178 Scarr, S. (1992). Developmental theories for the 1990s: Development and individual differences. Child Development, 63(1), 1-19. doi:10.1111/j.1467-8624.1992.tb03591.x Scarborough, H. S. (2009). Connecting early language and literacy to later reading (dis) abilities: Evidence, theory, and practice. In Fletcher-Campbell, F., Soler, J., & Reid, G., (Eds) Approaching Difficulties in Literacy Development: Assessment, Pedagogy and Programmes (23-38). Los Angeles: Sage Publications. Scarborough, H. S., & Dobrich, W. (1994). On the efficacy of reading to preschoolers. Developmental Review, 14(3), 245-302. doi:10.1006/drev.1994.1010 Schlichting, L. (2005). Peabody picture vocabulary test-III-NL. Amsterdam, the Netherlands: Hartcourt Assessment B.V. Schreiber, J.B., Nora, A., Stage, F.K., Barlow, E. A., King, J. (2006). Reporting structural equation modeling and confirmatory factor analysis results: A review. The Journal of Educational Research, 99(6), 323-338. Schmidt, L. A., Fox, N. A., Perez-Edgar, K., Hu, S., & Hamer, D. H. (2001). Association of DRD4 with attention problems in normal childhood development. Psychiatric Genetics, 11(1), 25-29. doi:10.1097/00041444-200103000-00005 Schwartz, J. (1994). Low-level lead exposure and Children′s IQ: A metaanalysis and search for a threshold. Environmental Research, 65(1), 42-55. doi:10.1006/enrs.1994.1020 Sénéchal, M., & LeFevre, J. (2002). Parental involvement in the development of childrens reading skill: A five-year longitudinal study. Child Development, 73(2), 445-460. doi:10.1111/1467-8624.00417 Sénéchal, M., LeFevre, J., Smith-Chant, B. L., & Colton, K. V. (2001). On refining theoretical models of emergent literacy the role of empirical evidence. Journal of School Psychology,39(5), 439-460. doi:10.1016/S0022-4405(01)00081-4 Sénéchal, M., LeFevre, J., Thomas, E. M., & Daley, K. E. (1998). Differential effects of home literacy experiences on the development of oral and written language. Reading Research Quarterly, 33(1), 96-116. doi:10.1598/RRQ.33.1.5 Senechal, M. (2006). Testing the home literacy model: Parent involvement in kindergarten is differentially related to grade 4 reading comprehension, fluency, spelling, and reading for pleasure. Scientific Studies of Reading, 10(1), 59-87. doi:10.1207/s1532799xssr1001_4 Servera, M., Lorenzo-Seva, U., Cardo, E., Rodriguez-Fornells, A., & Burns, G. L. (2009). Understanding trait and sources effects in attention deficit hyperactivity disorder and oppositional defiant disorder rating scales: mothers', fathers', and teachers' ratings of children from the Balearic Islands. Journal of Clinical Child & Adolescent Psychology, 39(1), 1-11. 179 Share, D. L., & Stanovich, K. E. (1995). Cognitive processes in early reading development: Accommodating individual differences into a model of acquisition. Issues in Education, 1(1), 1-57. doi:10.1080/0929704950840022 Shaywitz, S. E., & Shaywitz, B. A. (2008). Paying attention to reading: The neurobiology of reading and dyslexia. Development and Psychopathology, 20(4), 1329-1349. doi:10.1017/S0954579408000631 Shaywitz, B. A., Fletcher, J. M., Holahan, J. M., Shneider, A. E., Marchione, K. E., Stuebing, K. K., ... & Shaywitz, S. E. (1995). Interrelationships between reading disability and attention-deficit/hyperactivity disorder. Child Neuropsychology, 1(3), 170-186. doi:10.1017/S0954579408000631 Sims, D. M., & Lonigan, C. J. (2013). Inattention, hyperactivity, and emergent literacy: Different facets of inattention relate uniquely to preschoolers' reading-related skills. Journal of Clinical Child and Adolescent Psychology, 42(2), 208-219. doi:10.1080/15374416.2012.738453 Simons, R. L., Beach, S. R. H., & Barr, A. B. (2012). Differential susceptibility to context: A promising model of the interplay of genes and the social environment. Biosociology and Neurosociology, 29, 139-163. doi:10.1108/S0882-6145(2012)0000029008 Slopen, N., Fitzmaurice, G., Williams, D. R., & Gilman, S. E. (2010). Poverty, food insecurity, and the behavior for childhood internalizing and externalizing disorders. Journal of the American Academy of Child & Adolescent Psychiatry,49(5), 444-452. Retrieved from http://ezproxy.msu.edu.proxy1.cl.msu.edu/login?url=http://search.proquest.com.proxy1.c l.msu.edu/docview/622320688?accountid=12598 Smith, H. J., Sheikh, H. I., Dyson, M. W., Olino, T. M., Laptook, R. S., Durbin, C. E., . . . Klein, D. N. (2012). Parenting and child DRD4 genotype interact to predict children's early emerging effortful control. Child Development, 83(6), 1932. doi:10.1111/j.1467- 8624.2012.01818.x Snow, C. E., Burns, M. S., Griffin, P., National Academy Press (U.S.), Committee on the Prevention of Reading Difficulties in Young Children, & Educational Resources Information Center (U.S.). (1998). Preventing reading difficulties in young children. Washington, DC: National Academy Press. Sohr-Preston, S. L., & Scaramella, L. V. (2006). Implications of timing of maternal depressive symptoms for early cognitive and language development. Clinical child and family psychology review, 9(1), 65-83. doi:10.1007/s10567-006-0004-2 Somerset, W., Newport, D. J., Ragan, K., & Stowe, Z. (2006). Depressive disorders in women. In Keyes, C., & Goodman, S. (Eds). Women and depression: A handbook for the Social, Behavioral, and Biomedical Sciences, 62-88. Washington DC: Cambridge University Press. 180 Son, S. H., & Morrison, F. J. (2010). The nature and impact of changes in home learning environment on development of language and academic skills in preschool children. Developmental Psychology, 46(5), 1103. Spira, E. G., Bracken, S. S., & Fischel, J. E. (2005). Predicting improvement after first-grade reading difficulties: The effects of oral language, emergent literacy, and behavior skills. Developmental Psychology, 41(1), 225−234. doi:10.1037/0012-1649.41.1.225 Stanovich, K. E. (1986). Matthew effects in reading: Some consequences of individual differences in the acquisition of literacy. Reading Research Quarterly, 21(4), 360-407. doi:10.1598/RRQ.21.4.1 Stanovich, K. E. (2000). Progress in understanding reading: Scientific foundations and new frontiers. New York: Guilford Press. Stein, A., Malmberg, L. E., Sylva, K., Barnes, J., & Leach, P. (2008). The influence of maternal depression, caregiving, and socioeconomic status in the post-natal year on children's language development. Child: Care, Health and Development, 34(5), 603-612. doi:10.1111/j.1365-2214.2008.00837.x Stern, P., & Shalev, L. (2013). The role of sustained attention and display medium in reading comprehension among adolescents with ADHD and without it. Research in Developmental Disabilities, 34(1), 431-439. doi:10.1016/j.ridd.2012.08.021 Stevens, J., & Quittner, A. L. (1998). Factors influencing elementary school teachers' ratings of ADHD and ODD behaviors. Journal of Clinical Child Psychology, 27(4), 406-414. doi:10.1207/s15374424jccp2704_4 Storch, S. A. and Whitehurst, G. J. (2001), The Role of Family and Home in the Literacy Development of Children from Low-Income Backgrounds. New Directions for Child and Adolescent Development, 2001(92), 53–72. doi:10.1002/cd.15 Storch, S. A., & Whitehurst, G. J. (2002). Oral language and code-related precursors to reading: Evidence from a longitudinal structural model. Developmental Psychology, 38(6), 934- 947. doi:10.1037//0012-1649.38.6.934 Thayer, P. B. (2000). Retention of students from first generation and low income backgrounds. The Journal of the Council for Opportunity in Education, (5), 3-8. Retrieved from: http://files.eric.ed.gov/fulltext/ED446633.pdf Torgesen, J. K. (2002). The prevention of reading difficulties. Journal of School Psychology, 40(1), 7-26. doi: 10.1016/S0022-4405(01)00092-9 181 Torgesen, J. K. (2004). Avoiding the devastating downward spiral: The evidence that early intervention prevents reading failure. American Educator, 28(3), 6-19. Retrieved from: http://www.aft.org/periodical/american-educator/fall-2004/avoiding-devastating- downward-spiral Tripp, G., & Wickens, J. R. (2008). Research review: dopamine transfer deficit: a neurobiological theory of altered reinforcement mechanisms in ADHD. Journal of Child Psychology and Psychiatry, 49(7), 691-704. doi: 10.1111/j.1469-7610.2007.01851.x Trzesniewski, K. H., Moffitt, T. E., Caspi, A., Taylor, A., & Maughan, B. (2006). Revisiting the association between reading achievement and antisocial behavior: New evidence of an environmental explanation from a twin study. Child Development, 77(1), 72-88. doi:10.1111/j.1467-8624.2006.00857.x Tsao, F., Liu, H., & Kuhl, P. K. (2004). Speech perception in infancy predicts language development in the second year of life: A longitudinal study. Child Development, 75(4), 1067-1084. doi:10.1111/j.1467-8624.2004.00726.x US Department of Health and Human Services. (2016). The 2016 HHS Poverty Guidelines: One Version of the [US] Federal Poverty Measure. Washington, DC. Retrieved from:https://aspe.hhs.gov/poverty-guidelines Van Steensel, R. (2006). Relations between socio-cultural factors, the home literacy environment and children's literacy development in the first years of primary education. Journal of Research in Reading, 29(4), 367-382. doi:10.1111/j.1467-9817.2006.00301.x Vaughn, S., Wexler, J., Leroux, A., Roberts, G., Denton, C., Barth, A., & Fletcher, J. (2012). Effects of intensive reading intervention for eighth-grade students with persistently inadequate response to intervention. Journal of Learning Disabilities, 45(6), 515-525. doi:10.1177/0022219411402692 Velting, O. N., & Whitehurst, G. J. (1997). Inattention-hyperactivity and reading achievement in children from low-income families: A longitudinal model. Journal of Abnormal Child Psychology, 25(4), 321-331. doi:10.1023/A:1025716520345 Verhoeven, L., Reitsma, P., & Siegel, L. S. (2011). Cognitive and linguistic factors in reading acquisition. Reading and Writing, 24(4), 387-394. doi:10.1007/s11145-010-9232-4 Wadsworth, M. E., Raviv, T., Compas, B. E., & Connor-Smith, J. K. (2005). Parent and adolescent responses to poverty related stress: Tests of mediated and moderated coping models. Journal of Child and Family Studies, 14(2), 283-298. doi:10.1007/s10826-005- 5056-2 Waldfogel, J., Craigie, T., & Brooks-Gunn, J. (2010). Fragile families and child wellbeing. The Future of Children, 20(2), 87-112. doi:10.1353/foc.2010.0002 182 Walpole, S., Chow, S. M., & Justice, L. M. (2004). Literacy achievement during kindergarten: examining key contributors in an at-risk sample. Early Education and Development, 15(3), 245-264. doi: 10.1207/s15566935eed1503_1 Welsh, J. A., Nix, R. L., Blair, C., Bierman, K. L., & Nelson, K. E. (2010). The development of cognitive skills and gains in academic school readiness for children from low-income families. Journal of Educational Psychology, 102(1), 43-53. doi:10.1037/a0016738 Whitehurst, G. J., & Lonigan, C. J. (1998). Child development and emergent literacy. Child Development, 69(3), 848-872. doi:10.1111/j.1467-8624.1998.00848.x Whitehurst, G. J., & Lonigan, C. J. (2003). Emergent literacy: Development from prereaders to readers. In Neuman, S., & Dickson, D. (Eds). Handbook of Early Literacy Research, 1, 11-29. New York: Guilford Press. Whitehurst, G. J. (1992). Stony Brook family reading survey. Stony Brook, NY: Author. First Published in: Payne, A. C., Whitehurst, G. J., & Angell, A. L. (1994). The role of home literacy environment in the development of language ability in preschool children from low-income families. Early Childhood Research Quarterly, 9(3), 427-440. doi:10.1016/0885-2006(94)90018-3 Wight, V. R., Chau, M., Aratani, Y., & The National Center for Children in Poverty. (2010). Who Are America's Poor Children?: The Official Story. New York, NY: National Center for Children in Poverty. Willcutt, E. G., & Pennington, B. F. (2000). Psychiatric comorbidity in children and adolescents with reading disability. The Journal of Child Psychology and Psychiatry and Allied Disciplines, 41(8), 1039-1048. doi:10.1111/1469-7610.00691 Willcutt, E. G., Pennington, B. F., Olson, R. K., & DeFries, J. C. (2007). Understanding comorbidity: A twin study of reading disability and attention-deficit/hyperactivity disorder. American Journal of Medical Genetics Part B: Neuropsychiatric Genetics, 144B(6), 709-714. doi:10.1002/ajmg.b.30310 Willcutt, E. G., Betjemann, R. S., McGrath, L. M., Chhabildas, N. A., Olson, R. K., DeFries, J. C., & Pennington, B. F. (2010). Etiology and neuropsychology of comorbidity between RD and ADHD: The case for multiple-deficit models. Cortex, 46(10), 1345-1361. doi:10.1016/j.cortex.2010.06.009 Wilson, S., & Durbin, C. E. (2010). Effects of paternal depression on fathers' parenting behaviors: A meta-analytic review. Clinical Psychology Review, 30(2), 167-180. doi:10.1016/j.cpr.2009.10.007 Woodcock, R. W., & Johnson, M. B. (1989/1990). Woodcock Johnson Psycho-educational Battery—Revised. Allen, TX: DLM Teaching Resources. 183 Wyner, J. S., Bridgeland, J. M., & Diiulio, J. J. (2007). Achievement trap: How America is failing millions of high achieving students from low- income families. Jack Kent Cooke Foundation Retrieved from: http://files.eric.ed.gov/fulltext/ED503359.pdf Yarosz, D. J., & Barnett, W. S. (2001). Who reads to young children?: Identifying predictors of family reading activities. Reading Psychology, 22(1), 67-81. doi: 10.1080/02702710121153 Yoshikawa, H., Aber, J. L., & Beardslee, W. R. (2012). The effects of poverty on the mental, emotional, and behavioral health of children and youth: Implications for prevention. American Psychologist, 67(4), 272-284. doi:10.1037/a0028015 Zaslow, M. J., Hair, E. C., Dion, M. R., Ahluwalia, S. K., & Sargent, J. (2001). Maternal depressive symptoms and low literacy as potential barriers to employment in a sample of families receiving welfare: are there two-generational implications?. Women & Health, 32(3), 211-251. doi: 10.1300/J013v32n03_03 Zill, N., & Resnick, G. (2006). Emergent literacy of low-income children in Head Start: Relationships with child and family characteristics, program factors, and classroom quality. In Neuman, S., & Dickson, D. K. (Eds). Handbook of Early Literacy Research, 347-371. New York: Guilford Press. Ziol-Guest, K.M., Duncan, G.J., Kalil, A. & Boyce, W.T. Early childhood poverty, immune- mediated disease processes and adult productivity. Proc. Natl. Acad. Sci. USA 109, 17289–17293 (2012). Zubin, J., & Spring, B. (1977). Vulnerability: a new view of schizophrenia. Journal of Abnormal Psychology, 86(2), 103. doi: 10.1037/0021-843X.86.2.103 Zuckerman, M. (1999). Vulnerability to psychopathology: A biosocial model. Washington DC: American Psychological Association. Zumberge, A., Baker, L. A., & Manis, F. R. (2007). Focus on words: A twin study of reading and inattention. Behavior Genetics, 37(2), 433-433. doi:10.1007/s10519-007-9147-2 184