UFDC Home  Search all Groups  UF Institutional Repository  UF Institutional Repository  UF Theses & Dissertations   Help 
Material Information
Thesis/Dissertation Information
Subjects
Notes
Record Information

Full Text 
PAGE 1 1 THE LONGITUDINAL EFFECT OF TRAUMATIC STRESS AND ATTACHMENT DIFFICULTIES ON ACADEMIC ACHIEVEMENT FOR YOUNG CHILDREN By JEFF DRAYTON WOLFGANG A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PA RTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2013 PAGE 2 2 2013 Jeff Drayton Wolfgang PAGE 3 3 To my best friend PAGE 4 4 ACKNOWLEDGMENTS Thank you to the committee members Drs. Cirecie West Olatunji, Edil Torres Rivera, Walter Lana Leite, Ana Puig, Kristen Mary Kemple, and Pamela Pallas for their time and support throughout my graduate studies. Dr. James Algina provided guidance, discussion, and reinforcement on the methodol ogy. Dr. Linda Behar Horenstien extended useful lessons with planning and disseminating research. Amy Rigby, University of Florida alumni, administered assistance during the formatting process. Deepest gratitude to the Milch family, their constant source of encouragement and warmth provided potential and possibility for my education. Honor of grace to Mrs. Wank and Mr. Herstam for sharing their wisdom and kindness. Esteem to Mr. and Mrs. Case for their friendship. Admiration for the genuine mentoring, ad vising, patience, and teaching received from my chair, Dr. Cirecie West Olatunji. Having an exceptional supervisor whose enthusiasm is a testament to the aspiring advisory role. This paramount knowledge has provided a well rounded experience during my grad uate studies journey. PAGE 5 5 TABLE OF CONTENTS page A CKNOWLEDGMENTS ................................ ................................ ................................ ............... 4 LIST OF TABLES ................................ ................................ ................................ ........................... 8 LIST OF FIGURES ................................ ................................ ................................ ......................... 9 ABSTRACT ................................ ................................ ................................ ................................ ... 11 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ .................. 13 Research Questions ................................ ................................ ................................ ................. 17 Hypotheses ................................ ................................ ................................ .............................. 18 Significance of the Study ................................ ................................ ................................ ........ 18 Limitations ................................ ................................ ................................ .............................. 20 D ef i n i t i o n o f Te r m s ................................ ................................ ................................ ............... 21 2 LITERATURE REVIEW ................................ ................................ ................................ ....... 23 Persistent Academic Underachievement and Sociodemographic Factors .............................. 24 Race and Academic Achievement ................................ ................................ ................... 25 Gender and Academic Achievement ................................ ................................ ............... 25 Class and Academic Achievement ................................ ................................ .................. 26 Attachment and Academic Achievement ................................ ................................ ............... 28 Secure Attachments ................................ ................................ ................................ ......... 29 Insecure Attachments ................................ ................................ ................................ ...... 30 Attachment and Academic Underachievement ................................ ............................... 32 Traumatic Stress ................................ ................................ ................................ ..................... 33 Traumatic Stress Theory ................................ ................................ ................................ .. 35 Ecosystemic perspective ................................ ................................ .......................... 36 Individual neurobiological factors ................................ ................................ ........... 37 Developmental level at time of event ................................ ................................ ....... 38 Severity of impact of the event ................................ ................................ ................ 38 S ocio political context of the person (pre and post trauma) ................................ .... 39 Life events that have occurred before and after the traumatic event ........................ 40 Transgenerational Traumatic Stress ................................ ................................ ................ 40 Systemic Oppression and Hegemony ................................ ................................ .............. 43 Traumatic Stress and Attachment in Early Childhoo d ................................ ........................... 46 Effects of Traumatic Stress ................................ ................................ ............................. 47 Hyper/hypo Arousal ................................ ................................ ................................ ........ 48 Avoida nce ................................ ................................ ................................ ........................ 48 Re experiencing ................................ ................................ ................................ ............... 49 Traumatic Stress and Underachievement ................................ ................................ ............... 49 PAGE 6 6 Interventions for Traumatic Stress With Early Childhood ................................ ..................... 51 3 METHODOLOGY ................................ ................................ ................................ ................. 55 Research Constructs and Theoretical Assumptions ................................ ................................ 56 Setting ................................ ................................ ................................ .............................. 57 Participants ................................ ................................ ................................ ...................... 57 Researcher Subjectivity ................................ ................................ ................................ .......... 59 Operational Definition of Variables ................................ ................................ ....................... 60 Instrumentation ................................ ................................ ................................ ....................... 62 Data Analysis ................................ ................................ ................................ .......................... 66 Specification of Measurement Model ................................ ................................ ............. 67 Confirmatory Factor Analyses ................................ ................................ ........................ 71 Development of the Proxy Latent Variable ................................ ................................ ..... 71 Measurement Confirmatory Factor Analyses ................................ ................................ .. 72 Structural Model of LGCM ................................ ................................ ................................ .... 72 The Interaction ................................ ................................ ................................ ................. 73 Final Structural Model. ................................ ................................ ................................ .... 74 4 RESULTS ................................ ................................ ................................ ............................... 82 Participant Multivariate Descriptive Information ................................ ................................ ... 83 Specification of Measurement Model ................................ ................................ ..................... 85 The Latent Growth Curve Model ................................ ................................ .................... 85 The linear model ................................ ................................ ................................ ....... 85 The free loading model ................................ ................................ ............................ 86 The quadratic free loading model ................................ ................................ ............. 87 The quadratic orthogonal polynomial model ................................ ........................... 88 The HMLM quadr atic orthogonal polynomial model ................................ .............. 88 The HMLM cubic orthogonal polynomial model ................................ .................... 89 Confirmatory Factor Analysis ................................ ................................ ................................ 90 Difficult Attachment Symptoms EFA and CFA Development ................................ ....... 91 Traumatic Stress Symptoms CFA Development ................................ ............................. 92 The Structural LGCM ................................ ................................ ................................ ............. 94 Mean Values for Structural LGCM ................................ ................................ ................. 94 The Conditional LGCM ................................ ................................ ................................ .. 95 Reading and Mathematics Growth Trajectory and TSS ................................ ......................... 96 Reading Achievement LGCM With One TIC ................................ ................................ 98 Mathematics Achievement LGCM With One TIC ................................ ......................... 99 Reading and Mathematics Growth Trajectory and DAS ................................ ........................ 99 Reading Achievement LG CM With One TIC ................................ ............................... 101 Mathematics Achievement LGCM With One TIC ................................ ....................... 101 Reading and Mathematics Growth Trajectory and TSS and DAS ................................ ....... 102 Reading Achievement LGCM With Two TICs ................................ ............................. 103 Mathematics Achievement LGCM With Two TICs ................................ ..................... 104 Reading and Mathematics Growth Trajectory and TSS, DAS, Gender, SES, and Culture .. 106 Reading Achievement LGCM With Seven TICs ................................ .......................... 107 PAGE 7 7 Mathematics Achievement LGCM With Seven TICs ................................ ................... 111 Reading and Mathematics Growth Trajectory and TSS, DAS, Gender, SES, Culture, and the Interaction Between TSS and D AS ................................ ................................ ...... 115 Reading Achievement LGCM With Eight TICs ................................ ........................... 117 Mathematics Achievement LGCM With Eight TICs ................................ .................... 121 Grand Summary of Results ................................ ................................ ................................ ... 126 5 DISCUSSION ................................ ................................ ................................ ....................... 152 Findings From the Study ................................ ................................ ................................ ...... 152 Traumatic Stress Symptoms ................................ ................................ .......................... 152 Difficult Attachment Symptoms ................................ ................................ .................... 153 Interaction Betwe en TSS and DAS ................................ ................................ ............... 154 Academic Achievement and Sociodemographic Factors: Gender, SES, and Culture .. 155 Limitations ................................ ................................ ................................ ............................ 156 Implications for Practice ................................ ................................ ................................ ....... 157 Specialized Early Childhood Assessments ................................ ................................ .... 157 Att achment and Traumatic Stress Assessments ................................ ............................ 158 Pediatric Counseling ................................ ................................ ................................ ...... 159 Implications for Policy ................................ ................................ ................................ ......... 159 Recommendations for Future Research ................................ ................................ ................ 162 Summary of Study ................................ ................................ ................................ ................ 163 REFERENCES ................................ ................................ ................................ ............................ 167 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ....... 190 PAGE 8 8 LIST OF TABLES Table page 3 1 Orthogonal polynomial coefficients for unequal time points 0, 4, 16, 40, 64 months. ..... 75 4 1 Covariance coverage of the proportion of data present in the sample reading achievement data for this study. ................................ ................................ ....................... 129 4 2 Covariance coverage of the proportion of data present in the sample mathematics achievement data for this study. ................................ ................................ ....................... 130 4 3 The latent intercept, linear, quad ratic, and cubic factor coefficients, their standard errors, and p values for the cultural variable other people of color for the research .......... 131 4 4 The latent intercept, linear, quadratic, and cubic factor coefficients, their standard errors, and p values for the variable for gender of responding parent/guardian for the research questions 4 and 5 as applied to reading a nd mathematics achievement ................................ ................................ ................................ .......................... 131 4 5 Standardized correlations, observed means, and model estimated means for reading achievement using the HMLM cubic polynomial model. ................................ ............... 132 4 6 Standardized correlations, observed means, and model estimated means for mathematics achievement using the HMLM cubic polynomial model. .......................... 132 4 7 Comparison of the model estimated trajectory means for reading achievement. ............ 132 4 8 Comparison of the model estimated trajectory means for mathematics achievement. .... 133 4 9 EFA factor structure used to select the indicators for difficult attachment symptoms. ... 133 4 10 Descriptive statistics for difficult att achment symptoms. ................................ ................ 13 3 4 11 Descriptive statistics for traumatic stress symptoms. ................................ ...................... 134 4 12 Standardized factor loadings, their standard errors, factor correlations, and estimates for latent variables traumatic stress symptoms and difficult attachment symptoms. ................................ ................................ ................................ ........................ 135 5 1 Pediatric Counseling Competencies. ................................ ................................ ............... 166 PAGE 9 9 LIST OF FIGURES Figure page 3 1 Linear latent growth curve measurement model of reading achievement. ........................ 75 3 2 Linear latent growth cu rve measurement model of mathematics achievement. ................ 76 3 3 Freed loading latent growth curve measurement model for reading achievement. ........... 76 3 4 Freed loading latent growth curve measurement model for mathematics achievement. ................................ ................................ ................................ ...................... 77 3 5 HMLM cubic polynomial latent growth curve structural model with nine time invariant cova riates for reading achievement. ................................ ................................ ... 78 3 6 HMLM cubic polynomial latent growth curve structural model with nine time invariant covariates for mathematics achievement. ................................ ........................... 79 3 7 HMLM cubic polynomial latent growth curve structural model with nine time invariant covariates and the interaction between traumatic stress symptoms and difficult attachment symptoms for reading achievement. ................................ .................. 80 3 8 HMLM cubic polynomial latent growth curve structural model with nine time invariant covariates and the interaction between traumatic stress symptoms and difficult attachment symptoms for m athematics achievement. ................................ .......... 81 4 1 The density plots (on top) and histogram plots (on bottom) of the five measurement periods for reading achievement. ................................ ................................ ..................... 136 4 2 The density plots (on top) and histogram plots (on bottom) of the five measurement periods for mathematics achievement. ................................ ................................ ............. 137 4 3 The randomly selected plots of fif teen individuals observed longitudinal trajectories for reading achievement. ................................ ................................ ................................ .. 138 4 4 The randomly selected plots of fifteen individuals observed longitudinal trajectories for mathematics achi evement. ................................ ................................ ......................... 139 4 5 The total information curve of all of the indicators as a function of the latent variable difficult attachment symptoms. ................................ ................................ ........................ 140 4 6 The characteristic curves of all of the indicators as a function of the latent variable difficult attachment symptoms. ................................ ................................ ........................ 140 4 7 The total information curve of all of the in dicators as a function of the latent variable traumatic stress symptoms. ................................ ................................ .............................. 141 PAGE 10 10 4 8 The characteristic curve of the indicator P1ATTENI as a function of the latent variable traumatic stress sympt oms. ................................ ................................ ................ 141 4 9 The plot of the reading achievement linear trend. ................................ ........................... 142 4 10 The plot of the reading achievement quadratic trend. ................................ ...................... 142 4 11 The plot of the reading achievement cubic trend. ................................ ............................ 143 4 12 The plot of the mathematics achievement linear trend. ................................ ................... 143 4 13 The plot of the mathematics achievement quadratic trend. ................................ ............. 144 4 14 The plot of the mathematics achievement cubic trend. ................................ .................... 144 4 15 The path model analyzing the LGCM of reading achievement with the time invariant covariate traumatic stress symptoms. The red line is an insignificant estimate all others were significan t. ................................ ................................ ................................ .... 145 4 16 The path model analyzing the LGCM of mathematics achievement with the time invariant covariate traumatic stress symptoms. All paths were significant. .................... 146 4 17 The path model analyzing the LGCM of reading achievement with the time invariant covariate difficult attachment symptoms. The red line is an insignificant estimate all others were significant. ................................ ................................ ................................ .... 147 4 18 The path model analyzing the LGCM of mathematics achievement with the time invariant covariate difficult attachment symptoms. The red line is an insignificant estimate all others were significant. ................................ ................................ ................. 148 4 19 The path model analyzing the LGCM of reading achievement with the time invariant covariates traumatic stress symptoms and difficult attachment symptoms. The red line is an insignificant estimate all others were significant. ................................ ............ 149 4 20 The path model analyzing the LGCM of mathematics achievement with the time invariant covariates traumatic stress symptoms and difficult attachm ent symptoms. The red line is an insignificant estimate all others were significant. ............................... 150 4 21 This describes the overall results across all five models with the cubic orthogonal polynomial la tent growth curve model with seven time invariant covariates (TICs) and one time invariant covariate interaction. ................................ ................................ ... 151 PAGE 11 11 Abstract of Dissertation Presented to the Graduate School of the University of Florida i n Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE LONGITUDINAL EFFECT OF TRAUMATIC STRESS AND ATTACHMENT DIFFICULTIES ON ACADEMIC ACHIEVEMENT FOR YOUNG CHILDREN By Jeff Drayton Wolfgang August 2013 Chair: C irec i e West Olatunji Major: Marriage and Family Counseling National educational achievement statistics show that academic underachievement is a significant problem for all students in the United States and for culturally diverse students in particular. The relations hip of attachment and its interaction with traumatic stress has been proposed as an alternative explanation for the persistent underachievement found among marginalized children. This study explored the relationship between early childhood mental health an longitudinally using an ecosystemic view. Traumatic stress theory was used to explore the relationship between traumatic stress symptoms and difficult attachment symptoms affect the growth trajectory of reading and mathematics achievement for children in kindergarten through fifth grade (n=8,368) while controlling for gender, culture, and socio economic status. This quantitative study used the national database Early Childhood Longitudinal Study Kindergarten 1998 cohort (ECLS K) collected by the National Center for Education Statistics (NCES) in order to design a latent growth curve model that in corporated traumatic stress symptoms and difficult attachment symptoms and their effects on the growth trajectory. The results of this study indicated that both traumatic stress symptoms and difficult attachment symptoms predicted PAGE 12 12 significantly lower initi al scores in reading and mathematics as well as lower growth trends over time. The results of this study are significant because they suggest that interactive environmental factors, such as traumatic experiences and attachments issues impact academic achie vement for young children over time. Recommendations for future research include intervention studies that investigate ways to mediate the effects of traumatic stress and attachment issues during the early childhood years. PAGE 13 13 CHAPTER 1 INTRODUCTION Early el ementary school is a critical period for young children and an important time to identify any difficulties with behavior and academic concerns, as these can be associated with later negative outcomes for students; including possible aggressive behavior, gr ade retention, suspension, academic failure, and school dropout (Darney, Reinke, Herman, Stormont, & Ialongo, 2013). Disproportionate poor academic performance tends to be highest among socially marginalized students and these academic underperformance iss ues are often assumed to be due the prevention of academic underachievement and instead aggravates underperformance for culturally diverse students (Ford, Harr is, Tyson & Trotman, 2002; Goodman &West Olatunji, 2010; Sue & Sue, 2013; Wycoff, 1996) Deficit perspectives often keep educators from accurately assessing the risks and barriers to academic achievement by marginalized student ez, 2006). As researchers and educators continue to explore underachievement among marginalized students, relatively few researchers have examined how attachment experiences and psychological trauma may be alternative explanations to this problem (Bergin & Bergin, 2009; Goodman, Miller, & West Olatunji, 2012; Moss, Bureau, St Laurent, & Tarabulsy, 2011; Moss & St Laurent, 2001 ). Researchers have suggested that there are connections between attachment and subsequent academic achievement (Bergin & Bergin, 200 9; Cutrona, Cole, Colangelo, behaviors through differential attachment styles that includes secure and insecure attachments. along a wellness spectrum, that is broken down into 4 styles: (a) secure (most well), (b) insecure avoidant, (c) insecure resistant, and (d) insecure PAGE 14 14 disorganized disoriented (most risk to wellness) (Bergin & Bergin, 2009; Delius, Bovenschen, & Spangler, 2008). The less supportive attachment relationships, such as avoidant, resistant, and disorganized disoriented attachment behaviors often lead to less academic success (Bergin & Bergin, 2009; Pianta & Harbers, 1996) A n area research has shown to compromis e healthy attachment is traumatic stress experiences (Moss et al., 2011). experiences in schools (Goodman et al., 2012; Goodman & West Olatunji, 2008b). Traumatic stress theo ry utilizes an ecosystemic perspective to identify the systemic effects that may be out of control (Carlson, 1997); such as hegemony (the dominance of one social group over another) (Carter, 2007; Goodman & West Olatunji, 2010) and transgenerational traumas (Danieli, 2007). T he effects of systemic oppression are important in reconceptualizing childhood mental health and overall functioning for culturally diverse a nd socially marginalized young children (West Olatunji, 2010). Symptoms of traumatic stress in early childhood include interupted attachment displays of distress such as inconsolable crying, disorientation, diminished interest, aggression, withdrawing from peers, and thoughts or feelings that disrupt normal activities (De Young, Kenardy, & Cobham, 2011a, 2011b; Goodman & West Olatunji, 2010; Johnson, 1997; Lieberman & Van Horn, 2008; Modrowski, Miller, Howell, & Graham Bermann, 201 3 ). Psychological and som atic health problems affect substantial groups of children leading to underachievement and often when students fall behind early they never catch up (Lau, 2002). h PAGE 15 15 deficits, conduct problems, or autism (Goodman et al., 2012; Levine & Kline, 2007). By using traumatic stress theory and attachment health to examine the experien ces of children in their early years, it may improve the ways in which educators and counselors intervene with children. attachment relationship (Metzger, Erdman, & Ng, 2010; Zeanah & Smyke, 2009). provide direct services for children (Bagnato, 2006; Blanco & Ray, 2011). Professional counselors, more often than others, focus on a n ecosystemic perspective, to look at persons, their problems, and resiliencies while considering them in a relational and circumstantial context (Dollarhide, 2012; Lewis, 2012; Lim, Lim, Michael, Cai, & Schock, 2010; Waller, 2001). By using this approach counselors are able to conceptualize children, families, and communities in culturally sensitive and reciprocal fashions (West Olatunji, Frazier, & Kelly, 2011). Unfortunately, ocio emotional development ages zero to eight years. When young children experience developmental obstacles, such as events that are negative, sudden, and/or out of control (Goodman & West Olatunji, 2008b), they can experience developmental challenges. The se experiences are often complex, multidimensional, and difficult to assess in those affected (Follette, Palm, & Pearson, 2006; Follette & Vijay, 2009). When these difficulties are due to chronic stress, children often exhibit acting out behaviors and dist ress (Johnson, 1997), resulting in symptoms that appear to be behavioral and not related to crisis. During extreme stresses children can often have lowered abilities to respond to social conditions ( Nuru Jeter, Sarsour, Jutte, & Boyce, 2010 ). This can affe such as planning, working memory, and mental flexibility (Johanson, 2006; Nuru Jeter et al., PAGE 16 16 2010; Wilkerson, Johnson & Johnson, 2008), leading to academic underac hievement (Goodman et al., 2012). Conceptualization is hindered by the often disparaging nature of descriptions, discussions about social class still being framed within mainstream values, and differences in parenting styles often resulting in families in poverty becoming victims to classism or racism when our communities promote strategies of parenting and caring for children that do not align with their cultural logic of child rearing (Biles, Mphande Finn, & Stroud, 2012; Foss, Generali, & Kress, 2011; S ue & Sue, 2013; West Olatuji & Gibson, 2012). Unfortunately, the field of early childhood mental health does not have consensus on any standards of conceptualization and treatment or collected evidence for the validity or utility of the different approache s (Egger & Angold, 2009). Despite increased attention towards psychosocial development, many children with emotional or behavioral difficulties still go untreated (Gleason, Zeanah, & Dickstein, 2010; Horwitz et al., 2007). There is a minimal use of cultura l perspectives and they do not adequately consider developmental issues (Belfer, 2008). Using a post hoc data set, the purpose of the study is to examine the relationship between the longitudinal trajectory of academic achievement from kindergarten throug h fifth grade with The results of this study could indicate the impact of early chil dhood traumatic stress and attachment difficulties on developmentally competent counseling interventions in early childhood and elementary schools. This study cou ld also support the use of alternative perspectives to the deficit oriented view of low achievement among marginalized students (Goodman et al., 2012). This could aid PAGE 17 17 abilities and needs. Early childhood intervention research is interested in what is most effective and under which circumstance. This study extends prior research by exploring how issues with attachment and traumatic stress moderate academic achievement. The results of this study may indicate a need for counseling interventions in schools and alternative perspectives to the deficit oriented view of low achievement among marginalized students (Goodman et al., 2012). This could help counselors and educators while reducing mislabeled academic performance problems as low ability or problem behavior. Research Questions In order to develop an understanding of the relationship between children stress symptoms and difficult attachment symptoms with the outcome variables for academic achievement, I will examine a longitudinal multi level model beginning at school entry into kindergarten and continuing to fifth grade. I am also interes ted in looking at the relationship with Therefore, this study will examine five research questions: (1) What is the relationship between the growth trajectory of readi ng and mathematics achievement and traumatic stress symptoms for children in kindergarten through fifth grade?, (2) What is the relationship between the growth trajectory of reading and mathematics achievement and difficult attachment symptoms for children in kindergarten through fifth grade?, (3) What is the relationship between the growth trajectory of reading and mathematics achievement and traumatic stress symptoms and difficult attachment symptoms for children in kindergarten through fifth grade?, (4) What is the relationship between the growth trajectory of reading and mathematics achievement and traumatic stress symptoms and difficult attachment symptoms for children in kindergarten PAGE 18 18 through fifth grade, while controlling for gender, culture, and socio economic status? and finally, (5) What is the relationship of an interaction effect between traumatic stress symptoms and difficult attachment symptoms on the growth trajectory of reading and mathematics achievement for children in kindergarten through fi fth grade, while controlling for gender, culture, and socioeconomic status? Hypotheses The null hypotheses for each of the five research questions were as follows: ( Ho1) There will be a negative relationship between the growth trajectory of reading and ma thematics achievement and traumatic stress symptoms for children in kindergarten through fifth grade ( Ho2) There will be a negative relationship between the growth trajectory of reading and mathematics achievement and difficult attachment symptoms for chi ldren in kindergarten through fifth grade ( Ho3) There will be a negative relationship between the growth trajectory of reading and mathematics achievement and traumatic stress symptoms and difficult attachment symptoms for children in kindergarten through fifth grade ( Ho4) There will be a negative relationship between the growth trajectory of reading and mathematics achievement and traumatic stress symptoms and difficult attachment symptoms for children in kindergarten through fifth grade, while controlli ng for gender, culture, and socioeconomic status. ( Ho5) There will be a negative relationship of an interaction effect between traumatic stress symptoms and difficult attachment symptoms on the growth trajectory of reading and mathematics achievement for c hildren in kindergarten through fifth grade, while controlling for gender, culture, and socioeconomic status. Significance of the Study Identifying the relationship between reading and mathematics achievement growth trajectories with traumatic stress sympt oms and difficult attachment symptoms, as well as their interaction effects, could allow counselors and educators to better understand the relationship these variables have with reading and mathematics achievement. Thereby, allowing us to formulate more ac curate and effective direct intervention with young children It could further PAGE 19 19 attachment symptoms interact with academic performance, thus enabling counselors and educ could aid in the development of more culturally sensitive and developmentally appropriate interventions that will potentially better address the effects of traumat ic stress and attachment difficulties, thereby, moving us closer to improving academic performance for marginalized children. By continuing the conversation about how traumatic stress theory may aid policy makers in conceptualizing the needs of students, t hey may better meet equity and accountability. It could help policy makers develop standards that better serve culturally diverse students. Using the ecosystemic perspective may shed light on the contexts that are enabling disproportionality. This would be a step towards social justice and advocacy, by aligning public interests with student needs (Bemak & Chung, 2011). This study will extend previous research on academic underachievement and traumatic stress by using a longitudinal nationally representativ e sample of primary school students and including the attachment symptoms of these students. This study will attempt to investigate issues of culture on the prevalence of traumatic stress and difficult attachment symptoms with their relationship to reading and mathematics achievement growth trajectories among kindergarten to fifth grade students. Therefore, this study will serve to provide counselors, educators, and policy makers ways to improve academic outcomes for early childhood children. These results might be useful in informing training, professional development, and further development of school wide initiatives. Teachers may be able to receive training on recognizing traumatic stress and attachment difficulty symptomology, thereby, allowing them to make appropriate referrals to mental health professionals. As well as allow teachers to be trained in PAGE 20 20 early childhood educational interventions that will remedy the effects of traumatic stress and attachment difficulties within the classroom environment. T o conclude this research will be advocating for school counselors and other mental health professionals working in educational settings to provide trainings to teachers, further develop programs, and directly intervene with early childhood children individ ually and in group counseling settings. Limitations This study will use a nationally norm referenced data set collected by the National Center for Education Statistics (NCES). There are limitations to generalizability and construct validity because this s tudy will use pre collected data for post hoc analyses. Concerns surround construct underrepresentation and construct irrelevant variance. These validity concerns and the items that will be selected will be done in relation to a literature review and discu ssion with experts in ord er to ensure the ir appropriat eness (American Educational Research Association, 1999). PAGE 21 21 D ef i n i t i o n o f Te r m s F o l l o w i n g a r e d e f i n i t i on s o f t e r m s u s e d i n t h e s t ud y. A C A D E M I C A C H I E V E M E N T y the benchmarks within their school system (Robinson & Biran, 2006), including standardized test scores, grades, and level of school completion. A C A D E M I C U ND E R A C H I E V E M E N T Educational performance that is not corresponding to & West Olatunji, 2010) that includes lower grades, behavior problems, placement in special education courses, retention, high retention, and failing to graduate are indicators. A F R I C AN A M E R I C A N An in dividual who self identifies as belonging to a cultural or ethnic group from the U.S. who has African ancestry (Paniagua, 2001). A R O U S A L An indicator of traumatic stress characterized by any of the following: (1) disturbance in sleeping patterns, (2) difficulty with concentration, (3) hypervigilance or augmen ted startle response, and (4) irritability and anger ( 4th ed., text rev.; DSM IV TR; American Psychiatric Association, 2000 ). A VO I D A N CE An indicator of traumatic stress characterized by numbing as demonstrated by any of the following: (a) avoidance of tho ughts, feeling, behaviors, or activities that are related to a traumatic event, (b) memory loss to a traumatic event, (c) detachment behaviors and attitudes toward others, (d) sense of a foreshortened future, and (e) flat affect or reduced interest in prio r interests ( 4th ed., text rev.; DSM IV TR; American Psychiatric Association, 2000 ). C U L T U R A L L Y D I V E R S E An individual who self identifies as a member of a cultural group other than European American, such as African American, Latina/o, Asian American/Paci fic Islander, Native American (Sue & Sue, 2013) C U L T U RE Group identity founded on shared worldviews and values (West Olatunji, 2012). E XT E R N A L I Z I N G B E H A V I O R S An indicator of traumatic stress characterized by problematic behaviors as demonstrated by agg ression and delinquency (Graham Bermann & Levendosky, 1998). H E G E M O N Y Domination or privilege of one individual or group over another (Goodman & West Olatunji, 2010). I N T E R N A L I Z I N G B E H A V I O R S An indicator of traumatic stress characterized by problematic b ehaviors as demonstrated by anxiety/depression, withdrawal, and somatic complaints (Graham Bermann & Levendosky, 1998) L AT I N A / O An individual who self identifies as a member of a cultural group from the U.S. with ancestry from a Latin country (Paniagua, 2005). PAGE 22 22 M U L T I C U L T U R A L C OUNSELING C OM P E T E N C I E S (MCC) Benchmarks for clinical competence that include (1) self awareness of biases and worldview, (2) knowledge regarding the worldviews of others, and (3) skills appropriate to engaging with culturally differe nt individuals (Sue & Sue, 2013). N AT I V E A M E R I C A N An i n d i v i d u a l w h o s e l f i d e n t i f i e s a s b e l on g i n g t o a c u lt u r a l gro u p f r o m t h e U S w i t h i n d i g e no u s a n c e s t ry ( Paniagua, 2005 ) P EDIATRIC C OUNSELING An understanding and application of the intersection of var ious developmental issues when conceptualizing assessing, and intervening with young children from birth to five years of age (Wolfgang, Frazier, West Olatunji, & Barrett, 2012). R E E X P ER I E N C I N G An indicator of traumatic stress characterized by any of the following (a) frequent and troublesome thoughts or nightmares about a traumatic event, (b ) belief that a traumatic event is reoccurring, and (c) phsycological reactivity or psychological distress triggered by cues (internal and external) that resemble a tr aumatic event ( 4th ed., text rev.; DSM IV TR; American Psychiatric Association, 2000 ) S O C I A L L Y M A R G I N A L I Z E D I n d i v i d u a l s w ho e xp e r i e n c e o pp r e s s i o n or social bias in society due to demographic factors, such as culture, ethnicity, language, religion, sexual orientation, gender, or socio economic status (Staton, Evans, & Lucey, 2012) S O C I O E C O N O M I C S T AT U S ( S E S ) A measure of social standing based on hou sehold & Nord, 2005). S Y S T E M I C O PP R E SS I O N Differential treatment and systematic marginalization of groups and individuals based on characteristics including culture, gend er, class, sexual identity, religion, and more (Burstow, 2003). T R A N S G E N E R A T I O N AL T R A U M A Trauma that is passed down from one generation to another (Dass Brailsford, 2007). T R A U M A / T R A U M A T I C E V E N T / E X P E R I E N C E An event t h a t i s p e r c e i v e d b y a n i n d i v i d u a l t o be n e g a t i v e un c o n t r o l l a b l e a n d s udd e n t h a t c a n i n v o l v e a c t u a l or t h r e a t e n e d p h y s i c a l p a i n, i n j u r y or d e a t h or a c t u a l or t hr e a t e n e d e m o t i o n a l p a i n (C a r l s o n, 1997 ) i n w h i c h t h e i n d i v i d u a l f e e l s n o s e n s e o f c o n t r ol (C a r t e r, 2007) T R A U M AT I C S T R E SS p s y c h o l o g i c a l s t r e s s t h a t r e s u l t s f r o m a n e g a t i ve e v e nt l e a d i n g t o a ro u s a l a n d a v o i d a n c e o f t h e a ss o c i a t e d s t i m u l i t h a t t r i gg e r s e x p e r i e n c i ng t h a t o c c u rr e n ce i n w h i c h a n i n d i v i d u a l f e e l s n o s e n se o f c o n t r ol ( 4th ed., text rev.; DSM IV TR; American Psychiatric Asso ciation, 2000 ; Ro t h s c h il d 2000 ; S c a e r 2001) PAGE 23 23 CHAPTER 2 LITERATURE REVIEW This study explores the relationship between early childhood mental health and development (Brooks Gunn, Berlin, & Fuligni, 2000; Zeanah & Zeanah, 2009) and explores how this mig diversity of cultures and linguistic groups continues to grow and it is unfortunate that academic underachievement and high dropout rates persists among certain groups of sc hool children (Chu, 2011; Darney et al., 2013). National legislation has forced schools to be accountable for differences in sociodemographic factors in order to create equity in achievement outcomes. These sociodemographics include (a) race or ethnicity, (b) gender, as well as (c) socioeconomic status (Chatterji, 2006; Crumpton & Gregory, 2011; Moore, Henfield, & Owens, 2008) Failure to sufficiently address this concern has been, in part, due to researchers' lack of consideration for macro systemic influe nces or lack of focus on early intervention. Most efforts of dealing with underachievement have been from a deficit orientation which selectively contributes to academic underachievement for students who are culturally diverse and/or from lower socioeconom ic Relatively recently professional counselors and other mental health caregivers have begun to explore the effects of attachment experiences, as well as the prevalence o f traumatic stress on early childhood normal development. This has stimulated a need for all service providers working with young children to be able to address these complex issues. Mental health a o r attachment history may exacerb ate behavioral disorders (Goodman et al., 2012; Levine & Kline, 2007). This chapter discusses the literature relevant to the study of th e following topics: (a) persistent academic underachievement PAGE 24 24 and sociodemographic factors (b) attachment and academic achievement, (c) traumatic stress, (d) traumatic stress and attachment in early childhood, and (e) traumatic stress and underachievement. Persistent Academic Underachievement and Sociodemographic Factors In the U.S. according to the National Assessment of Educational Progress (NAEP), 34% of children w ere found to be at or above proficient in reading and 40% of children were found to be at or above proficient in mathematics (National Center for Education Statistics, 2011a; 2011b). Looking at differences in race/ethnicity show that the percentage of chil dren of color below basic proficiency in reading (Black 51%, Latino/a 49%, and Native American 53%) are still below their White peers (21%) and the percentage of children of color below basic proficiency in mathematics (Black 48%, Latino/a 48%, and Native 44%) are still below their White peers (39%) ; however the gap is closer than in the reading proficiencies. Gender differences persist in reading where females are performing higher than males, however, in mathematics males were only 1 point higher on ave rage than females. Finally, in both reading and mathematics proficiencies low income children are performing significantly lower than higher income children. Educational research has portrayed a deficit orientation toward socioeconomic class and cultural ly diverse persons within the constraints of Western ideologies (Chilisa, 2012; King, 2005; Smith, 1999; West Olatunji & Gibson, 2012). It has become an important discussion where closing the achievement gap is an ongoing goal of educators, counselors, par ents, and politicians (Biles et al., 2012). PAGE 25 25 Race and Academic A chievement Studies of educational inequality have revealed the existence of a persistent gap among African Americans, Latino/as, and Native American students (Crumpton & Gregory, 2011; Morris, 2008). Researchers continue to show that many of our teachers still have low intelligence (Chu, 2011). Children seem to get classified as culturally impoverishe d because their families and communities do not instill values and experiences typical of White, middle class families and communities (Foster, Lewis, & Onafowora, 2003). When comparing students of color to White students, students of color show disproport ionate poor performance on high stakes tests, lower grades, higher school dropout rates, and do not show achievement levels that match with their academic ability (Moore, Ford, & Milner, 2005) Educational researchers discuss how the culture and organizati on of schools situates children of color as academically and behaviorally deficient and thus places them at risk for special education placement African American children, who are male and poor (Moore et al., 2008). Gender and Academic A chievement During the past decade, there has been an increasing gender achievement gap with male students lagging behind their female counterparts on a number of important indicat ors of school success (Clark, Flower, Walton, & Oakley, 2008). Gender and SES differences have been found mathematics achievement in the first and fifth grades (Chatterji, 2006; Dearing, McCartney, & Taylor, 200 9). These studies focus on parenting practices as the way for remediation, this pejorative assumption that parenting style is to blame often results in families in poverty becoming victims to classism or racism when our communities promote strategies of pa renting and caring for children that do not align with their PAGE 26 26 cultural logic of child rearing (Biles et al., 2012; Foss et al., 2011; Sue & Sue, 2013; West Olatuji & Gibson, 2012 ). Several studies suggest that socio political factors such as academic releva nce (view that education will lead to better jobs), lack of male role models in schools, and cultural attitudes about gender expectations might contribute to these differences (Clark et al., 2008; Crumpton & Gregory, 2011). However, Morris (2008) provides support that there might be intersectionality with gender, race, and socioeconomic class. He suggests that boys may be more inclined to interests that involve physical and experiential activities. Class and Academic A chievement The intense lack of convers ation around social class and educational achievement can be experience marginalization, disenfranchisement, and lack of access to tools that make educational achievement available (Staton, Evans, & Lucey, 2012; Sturm & Slaughter, 2012).The belief that poor people do not value education is far less accurate than the idea that poor people are often not valued in the educational process (Pressley & Sifford, 2012). In fact, researchers have found that it is nearly impossible to identify potentially high achieving students who are living in poverty based on the current system of i dentifying students for gifted programs, due to limited access to enrichment programs, after school activities, and being misidentified for special education (Biles et al., 2012). Despite parents of color encouraging their children to develop educational goals, racism and poverty continue to create disparities, especially among African American, Latina/o, and Native American children; as well as some of the less known Asian Pacific Islanders, such as the Hmongs, Tongas, Cambodians, and Laotians (Sue, 2010) Native Americans perform well during the first four years of school, but by the end of fourth grade they begin significantly decreasing in PAGE 27 27 academic performance (Juntunen et al., 2001; Juntunen & Cline, 2010). A main theme in the results was that initial poverty levels than with ethnicity or gender, although African American children showed significant gaps compared with White children even when other background characteristics were controlled (Chatterji, 2006). By fifth grade, poor children are as much as two times more likely to lack proficiency in mathematics and reading skills than children who are not poor (Dearing et al., 2009). Therefore, children living in poverty present a pr ofound challenge to today's educators and counseling professionals as they are significantly more likely to report increased levels of anxiety, depression, greater incidences of behavioral difficulties, and a lower level of positive engagement in school (A matea & West Olatunji, 2007; Brennan, Shaw, Dishion, & Wilson, 2012). Without addressing these risks our low achieving students face the proximal risks of disengagement, greater incidence of school failure, developmental difficulties and delays, low achiev ement, and the more distal and debilitating risks of grade retention, dropout, restricted job options, and low wages (Amatea & West Olatunji, 2007; Crumpton & Gregory, 2011). Most efforts of dealing with the above sociodemographic factors and underachievem ent have been from a deficit orientation. This selectively contributes to academic underachievement for students who are culturally diverse and/or from lower socioeconomic backgrounds (Ford et al., 2002). A better way to address these concerns may be throu gh a focus on early intervention. Early intervention focuses on ecological factors that influence young children's growth and development. In particular, healthy attachment has been linked to wellness. With the expansion of applications of attachment theor y being made beyond just infant caregiver relationships, 2009; Metzger et al., 2010; Moss et al., 2011). PAGE 28 28 Attachment and Academic Achievement The predictive vali academic achievement has been found to account for between 17% and 24% of the variance in the Iowa Test of Basic Skills (ITBS) scores in second through fourth grades (Pianta & Harbers, 1996 ). The connections between attachment, healthy development, and subsequent academic achievement (Bergin & Bergin, 2009; Cutrona et al., 1994; Pianta & Harbers, 1996), provide support that attachment behaviors are strongly related to school outcomes in pres chool and elementary (Moss & St Laurent, 2001; Pianta & Harbers, 1996). Often, t hese developmental years involve rapid growth, expansion of skills, and a progression of social emotional attachment experiences (Squires, 2012). Therefore, it is important to understand that in early childhood the course of development is non linear and dynamic (Walker & Archibald, 2006), and requires counselors and educators to conceptualize children of diverse cultures across time rather than looking at fixed ages and stages (Peluso, Miranda, Firpo Jimenez, & Pham, 2010; Siegel, 2003), as well as understanding that culture plays a particularly important role in the formation of attachment quality and how that psychological development (Metzger et al., 2010). Attachment quality describes the mental health of young children based on the nature of the attachment relationship. For many young children attachments indicate healthy dependencies that encourage secure development involving interdependence as well as autonomy, and are related to survival in the form of a safe haven or secure base from which to explore both inner and outer worlds (Metzger et al., 2010; Zeanah & Smyke, 2009). Attachment relationships for early childhood capture the balance of attachment and exploratory behavior as well as distinguish communication and defensive behaviors that underlie attachment (Solomon & George, 2008). Attachment research describes this development through PAGE 29 29 differential a ttachment behaviors, along a wellness spectrum, that includes secure and insecure attachment. This is further broken down into 4 styles, 1 secure and 3 insecure: (a) secure, (b) insecure avoidant, (c) insecure resistant, and (d) insecure disorganized disor iented (Bergin & Bergin, 2009; Delius et al., 2008). Secure at tachments are considered the more well and insecure disorganized disoriented styles are the more risky to wellness (Solomon & George, 2011a). Secure A ttachments The secure style is distinguished the child is able to engage in culturally appropriate reciprocal interaction or conversation involving some form of direct expression of feelings, desires, and the ability to negotiate conflict or disa greement (Solomon & George, 2008). This style is fostered by caregiver(s) encouragement, sensitivity, consistency, security, and responsive social emotional interactions (Andreassen & West, 2007; Lounds, Borkowski, Whitman, Maxwell, & Weed, 2005). Attachm ent theory provides a four phase process to organize the rapid changes that take place during secure early childhood development (Sheperis, Renfro Michel, & Doggett, 2003; Tomlinson, Murray, & Cooper, 2010; Van Ijzendoorn & Sagi Schwartz, 2008) The four p hases infants reflexively attach, seek out human connection, and respond to familiar smells, sights, and sounds (Lieberman & Van Horn, 2008); using these relati onships to help regulate sleeping, feeding, discomfort, and activity (Zeanah & Zeanah, 2009) Phase II (3 to 6 months) infants begin orienting to familiar people, are emotionally expressive, and are attuned and responsive to erman, 1993; Marvin & Britner, 2008; Rosenblum, Dayton, & Muzik, 2009; Zeanah & Smyke, 2009). Phase III (6 months on), infants become weary of strangers and actively seek proximity to familiar caregivers using them to learn socialization PAGE 30 30 behaviors that inv olve verbal and nonverbal signals (Gauvian & Parke, 2010; Rosenblum et al., 2009), such as displays of happiness, sadness, anger, or fear Phase IV (from the second to third year on) begins the formation of complex verbal communications, progressively grea ter walking ability, increased abilities to negotiate with caregivers (Delius et al., 2008; Marvin & Britner, 2008), and a greater ability to coordinate behaviors in terms of the goals of others while responding to them empathically (Rosenblum et al., 2009 ). Insecure A ttachments What distinguishes the insecure attachments is that to varying degrees the child loses confidence in the caregiving system to be responsive and available during times of distress and/or trauma (Kobak & Madsen, 2008; Perlman & Doyle, 2012). Avoidant attachment styles are distinguished by caregiver(s) who are unresponsive or diminish the importance of being responsive, are emotionally unavailable and have difficulty with closeness, and utilize parenting strategies that oversee children from afar or assign care to someone else (Bergin & Bergin, 2009; George & Solomon, 2008). This rejection or discouragement of the child's attachment behaviors creates a lack of confidence in caregiver(s) availability that can result in children seeming in different to the presence or absence of their caregiver, readily exploring while displaying detached and neutral responses to others, however these children do not avoid interaction altogether, rather the child does subtly minimize opportunities for intera ction (Kobak & Madsen, 2008; Solomon & George, 2008). Resistant attachment styles are characterized by caregiving practices that lead to close upsetness has escalated to high levels before soothing (Solomon & George, 2008). These parents endorse emphasizing positive aspects of the attachment relationship, while having difficulty with unhappiness or disappointment from the child. This resistance to distressing experienc es results PAGE 31 31 in children lacking confidence in caregiver responsiveness causing them to maximize expressions of distress in order to elicit responses from caregivers (Kobak & Madsen, 2008). This often takes the form of exaggerated responses to problems and c onflicts, resulting in ambivalent passive aggressive behavior where the child resists exploration and retaliates with temper tantrum like behaviors or becomes coy and tends to regress to more infantile behavior (Solomon & George, 2008) This can make the c hild difficult to soothe, and results in children who underregulate their emotions while easily feeling threatened, frustrated, anxious, and helpless (Bergin & Bergin, 2009). Disorganized disoriented attachment styles are distinguished by caregiving practi ces that can be hostile or confrontational, submissive or helpless, frightened or out of control, or disengaged with respect to the child, the caregiver(s) own emotions, and with the relationship (Hazen, Jacobvitz, Higgins, Allen, & Jin, 2011; Solomon & Ge orge, 2011a). These caregiver(s) have difficulty providing reassurance, and often passively place children at risk, while portraying themselves as helpless to manage or protect the child and can be flooded by powerful, negative feelings (Solomon & George, 2011b). This attachment style is complex and has a wide variety of possible behaviors and responses that can take various forms of externalizing as well as internalizing behaviors in order to maintain a connection with the caregiver(s) and are conceptualiz adapt to the caregivers emotional needs (Hazen et al., 2011; Solomon & George, 2011a). These attempts at gaining caregivers attention are most distinguished by displays of confusion/disorientation, hostility, freezing respon ses that are followed by rejection or fear of the caregiver, or the child displays caregiving responses such as reassuring, pleasing or cheering up, PAGE 32 32 or being overly attentive (with hints of apprehension) (Hazen et al., 2011; Moss et al., 2011; Solomon & Ge orge, 2008; 2011a; 2011b). Studies have documented links between attachment support and performance in academic achievement, thus caregiver(s) support can be anticipated to have a significant effect on ). In particular the insecure attachment styles, tend to show difficulties self regulating and processing social emotional information in the classroom, have the poorest performance in schools, and display the greatest academic risk (Moss et al., 2011; Sol omon & George, 2011a). Attachment and Academic U nderachievement As they enter into schools, children who progress through the four phases of secure attachment behavior are more likely to develop better attitudes and skills with greater curiosity toward aca demic achievement (Bergin & Bergin, 2009). However, the less supportive attachment relationships, such as avoidant, resistant, and disorganized disoriented attachment behaviors often lead to less academic success (Bergin & Bergin, 2009; Pianta & Harbers, 1 996). In preschool and early school age, insecure attachment behaviors have been associated with disruptive behavior patterns that cause anxieties and fears around school performance (Moss et al., 2011). There have been observed gender differences in how a ttachm ent stressors affect academic achievement as well. Boys and girls can develop different strategies for coping with attachment stressors, a trend that has been observed is that males are more likely to respond to attachment stressors with flight, figh t, or freeze responses, and females are more likely to respond with tending (caretaking) and befriending behaviors (Hazen et al., 2011). This is impactful because the children whom teachers identify as being more disruptive and showing PAGE 33 33 greater underachieve ment academically, are those viewed as being externalizing (aggressive, whiny, rule breaking), and are more often boys (Hazen et al., 2011). Insecure attachments, such as, avoidant, resistant, and disorganized disoriented attachment behaviors tend to lead students to exhibit ADHD like symptoms, show less curiosity, be more dependent, exhibit impaired ability to control attention and memory, while having shorter attention spans, fear of failure, seek less help from teachers, and thus have less academic succe ss (Bergin & Bergin, 2009; Moss & St Laurent, 2001). One of the areas that have been shown to compromise children's healthy attachment has been traumatic stress. There are multiple factors that relate to the severity and impact of traumatic stress on young children. Additionally, some scholars have suggested that there are multiple layers of traumatic stress within families, due to societal factors (Moss et al., 2011). Traumatic Stress Mental health practices have been guided and greatly influenced by the American rev.; DSM IV TR; American Psychiatric Association, 2000; Eriksen & Kress, 2006). According to the DSM IV TR, a traumatic event is defined as one in whi witnessed, or was confronted with an event or events that involved actual or threatened death or IV TR; American Psychiatric Associa tion, 2000, p. 467). Exposure to a traumatic event is a necessary condition that precedes the later development of posttraumatic stress disorder (PTSD) (Friedman, Resick, Bryant, & Brewin, 2011). Additionally, in order to qualify for a diagnosis of PTSD, t he DSM IV TR requires that individuals report experiencing dominant emotions of fear, helplessness, and horror during the trauma (Brunet, Akerib, & Birmes, 2007; Hathaway, Boals, & Banks, 2010). Followed by symptoms of reexperiencing, avoidance/numbing, a nd PAGE 34 34 hyperarousal that cause the person to exhibit clinically significant distress or functional impairment, but only after a one month period of time has elapsed (Friedman et al., 2011). PTSD in young children is diagnosed using the same three criteria as a dults: reexperiencing, avoidance, and hyperarousal (Perlman & Doyle, 2012; Zero to Three, 2005). Although PTSD has been widely reported in children and adolescents, the criteria was developed and tested from adult samples, where no one was younger than 15 years of age (Scheeringa, Zeanah, & Cohen, 2011), which raises concerns regarding its suitability and validity for children (De Young et al., 2011b). Research has indicated that the DSM IV TR PTSD criteria is not developmentally sensitive enough to adequat ely capture the symptoms experienced by early childhood children, particularly infants and preschoolers, and underestimates the number of children experiencing considerable posttraumatic distress and impairment (De Young et al., 2011a; Scheeringa et al., 2011). The mental health profession has been advocating for an expanded conceptualization of trauma in order to addresses the limitations deep Davis & Ocampo, 2005; Burstow, 2003; 2005) Empirical outcomes of indiv iduals experiencing discrimination due to culture have shown that this oppression has harmful effects on both physical and mental health (Harrell, Hall & Taliaferro, 2003; Paradies, 2006; Williams, Mohammed, Leavell, & Collins, 2010; Williams, Neighbors & Jackson, 2003). Research has found that ecological factors, such as racism, harassment, and oppression, may result in such as it does not take into consideration th at for young children, traumatic effects can have immediate negative influences, however these effects are not always apparent immediately following the incidents that caused them and often can remain dormant, accumulating over years PAGE 35 35 or even decades (Levin e, 1997). Some scholars who have evaluated the DSM suggest that it limits our understanding of trauma and minimizes the importance of individual perception (Rothschild, 2000) as well as ignores macro systemic influences (Lopez et al., 2006; Ivey & Ivey, 19 98). This creates important issues for children because the PTSD definition in the DSM criteria are not developmentally sensitive, they require that a traumatic event be identified and physically dangerous, and often excludes macrosystemic influences such as transgenerational influences, as well as experiences of oppression and discrimination (Burstow, 2003; Danieli, 1998; De Young et al., 2011a; Scheeringa et al., 2011; West Olatunji & Conwill, 2011). Therefore, I propose to utilize traumatic stress theory as it takes into account ecological factors, does not require a known traumatic event, and does not put the older than one month time constraint for the assessment process. By using traumatic stress theory, we can more clearly conceptualize children's nee ds and formulate more expedient interventions that are grounded in Traumatic Stress T heory Traumatic stress theory identifies the effects of events on people that meet the following criteria: (a) perception that a n event is negative, (b) the event is perceived to be sudden, (c) and there is a sense of lack of control over the event (Goodman & West Olatunji, 2008b, 2010). When people perceive of an event as being negative, sudden, and uncontrollable, Carlson (1997) theorized that these type of events are tempered by five factors: (a) individual neurobiological factors, (b) developmental level at time of event, (c) severity of impact of the event(s), (d) the socio political context of the person (pre and post trauma), and (e) the life events that have occurred before and after the traumatic event. A way that has been proposed to explore the five factor areas has been to use an ecosystemic perspective. PAGE 36 36 Ecosystemic perspective An ecosystemic framework views people as i ntegrated beings and feels that exploring child development requires a framework that observes and explores the norms, rules, and systems of our society (Bronfenbrenner,1986; Conyne & Cook, 2004). Counselors and educators xperiences can be found within several systems such as, (a) the immediate family, preschool, and elementary school experiences, (b) that there are linkages exte nded family have influences, (d) that these experiences are situated in the customs, values, and laws of the surrounding culture as well as the family culture, and (e) the dimension of time and its effects on human development are significant ( Cook, Conyne Savaeau, & Tang, 2004; Dass Brailsford, 2007). Professional counselors and educators who adopt an ecosystemic perspective assume that mental health issues are influenced by multiple intersecting factors, which include individual characteristics, family, school, community, and other contexts (Dass Brailsford). Since all of these systems are influenced by the passage of time, we attempt to understand the ever development and lookin g to see if their experience of change is altered with time (Roysircar & Pignatiello, 2011). Therefore, the ecosystemic model assumes that we are active participants in our development, that there is continual interaction, mutual influence, and constant c hange, as well Chronister, McWhirter, & Kerewsky, 2004). This allows counselors to view young children, their challenges, and resiliencies in a relational and conditional frame of reference (Dollarhide, 2012; Lewis, 2012; Lim et al., 2010; Waller, 2001). The ecosystemic perspective allows for accounting for symptoms and difficulties in historical contexts, as well as accounting for current cultural, PAGE 37 37 political, and institutional i ssues that are impacting young children (Goldenberg & Goldenberg, 2008). By addressing these sociopolitical frameworks, counselors are able to conceptualize children, families, and communities in sociocultural context s when working with culturally diverse families, thereby extending the system of care to incorporate more systemic influences, such as community support systems (Constantine, Myers, Kindaichi, & Moore, 2004; Foss et al., 2011; West Olatunji et al., 2011) in developmentally and responsive ways. Individu al neurobiological factors Two biological factors may influence responses to trauma in early childhood (Carlson, 1997). The first is a biological or genetic predisposition to vulnerability or resilience to trauma. In the first three years of lif language signals, such as tone of voice, gestures, facial expressions of affect, perception of emotion, and regulation of these states are often dependent on brain responsiveness in the followin g areas: (a) relative internal and external safety, (b) support for exploration and acceptance of sensation, and (c) pendulation of expressing the internal and external emotions (Siegel, 2003; Levine, 2010). The second is biological change(s) that may occu r in response to traumatic experiences (Carlson). This is often found in stressful events during early development, such as, neglect, physical and sexual abuse, failure of the attachment bond, and traumatic incidents (hospitalization, death of parents, car accidents, etc.) (Rothschild, 2000; Levine). These experiences can cause brain development to be reduced or its activity to become especially happens in the hippocamp us, the area of the brain that is involved in emotion and memory organization (Goodman et al., 2012). PAGE 38 38 Developmental level at time of event The effects of the developmental level of the child at the time of trauma are complex and are often equated with age ; suggesting that age at the time of trauma is expected to be negatively related to severity of symptoms (Carlson, 1997). Studies are now showing that prenatal infants, newborns, and very young children are the most at risk to stress and trauma due to thei r undeveloped nervous, motor, and perceptual systems (Levine & Kline, 2007). More importantly, current ecosystemic models of child development provide evidence suggesting that neurobiological vulnerability factors may actually be indicators of sensitivity in response to experience, which could be bad in an adverse environment but good in a favorable one; raising the possibility that children who respond poorly to traumatic experiences may also respond well to positive changes in context provided by interven tions (Levine & Kline; Masten & Narayan, 2012). Severity of impact of the event The intensity, nature, and duration of the trauma can affect symptoms, but it is often the an event has (Blaustein & Kinniburgh, 2012; Carlson, 1997). Traumas that are more intense are thought to be more likely to provoke overwhelming fear and helplessness which often enhances the perception of negative impact due to the discomfort children hav e with the sensations and emotions evoked by fearful and out of control situations (Levine & Kline, 2007; Levine, 2010). The severity of symptoms can be made worse if adults do not consider nor move towards attunement with the ns (reactions); where attunement is the capacity to accurately read physiologically) (Blaustein & Kinniburgh; Johnson, 1997). PAGE 39 39 Socio political context of the person (pre and post trauma) Children's social context directly shapes their neural architecture, as interpersonal relationships shape the way genes, personality, and behavior are expressed (Siegel, 2012). The family can be a common source of warmth and safety, but i t can also be a source of direct threats to provide basic needs (neglect) ( Seccombe, 2002; Turner et al., 2012; Werner, 2000 ). The socioeconomic environment t hat a child is situated in can influence the rate of traumatic stress. It has been shown that trauma is significantly higher among lower socioeconomic status (SES) children, further suggesting that low SES families experience trauma at a higher rate or hav e fewer resources with which to relieve the effects of trauma (Goodman et al., 2012). Further, studies show that traumatic stress theory should include the complex ways in which gender, race, ethnicity, and socioeconomic status (SES) combines to affect pa tterns of disproportionate distribution of mental health issues and disease (Burstow, 2003; 2005; Davis et al., 2011; Cholewa & West Olatunji, 2008; Howell & McFeeters, 2008; Knitzer & Perry, 2009; Liu & Estrada Hernndez, 2010; West Olatuji & Gibson, 2012 ; Williams et al., 2010). In p articular, African Americans, Latina/os, Asian Americans, and Native Americans as marginalized people of color experience systemic oppression as a form of traumatic stress (Carter & Forsyth, 2009; Harrell et al., 2003; Parad ies, 2006; Williams et al., 2003). As well, important dimension in trauma assessment, particularly to determine if there are long term intergenerational experienc es that affect the family and succeeding generations (Danieli, 1998; 2007). PAGE 40 40 Life events that have occurred before and after the traumatic event Prior events that make a trauma seem more controllable and less negative would have a moderating effect, while prior events that make a trauma seem less controllable and more negative would have a sensitizing effect (Carlson, 1997). Likewise, having to cope with negative life experiences after traumatic events, such as living in poverty, single parent households, family instability, stressful work experiences, and living in dangerous community environments would be expected to impair children's recovery from trauma ( Amatea & West Olatunji, 2007; Carlson; Werner & Smith, 1982; West Olatunji, Sanders, Mehta, & Behar Horenstein, 2010) It is also important in traumatic stress theory to explore what the individual, family, and effects of traumatic stress (West Olatunji et al ., 2010). Findings support a need for an ecosystemic framework when looking at survivors of trauma, especially in areas of vulnerability and resilience, in that it becomes necessary to trace a history of multiple traumata along the time dimension at differ ent stages of development and assessing the transgenerational implications (Danieli, 1998; 2007). Transgenerational Traumatic S tress Transgenerational trauma takes an ecosystemic perspective as it explores the impact of trauma, its transmissions, and repe ated patterns of behavior, symptoms, roles, and values as adopted by children, their family members, and communities through the generations (Danieli, 1998; Solomon & George, 2011b; West Olatunji & Conwill, 2011). This transgenerational traumatic stress re fers to thoughts, feelings, behaviors, and disrupted conceptualizations or traumatic beliefs and attachment patterns that are passed down from one generation to another (Danieli, 2007; Dass Brailsford, 2007; Solomon & George). In many cases these parents w ould normally be sensitive and responsive, however, when in the presence of traumatic experiences PAGE 41 41 these parents can alarm infants and young children with repeated exhibitions of frightened, dissociative, or overprotective behavior (Hesse, Main, Abrams, & Rifkin, 2003; Schwerdtfeger & Goff, 2007) When looking at internalizing and externalizing behaviors in early childhood, such as unusual levels of aggression, frightened, dissociative, or threatening play in in the presence of otherwise sensitive parenting it becomes necessary to explore transgenerational histories for traumatic stress (Blaustein & Kinniburgh, 2010; Hesse et al.). Transgenerational trauma are complex and often have at least four components: (a) a parent(s)/caregiver(s) expressions of traum posttrauma adaptational styles (Danieli, 2007). The conspiracy of silence has been described by children and survivors of trauma as the lack of consideration and experience of neglect, misjudgment, or avoidance of their experience by others, which prevents them from testifying these experiences, thus leading to a fragmented sense of self amplified feelings, loneliness, isolation, and mistrust (Danieli, 1984, 1998; Fivush, 2010; Henderson & Thompson, 2011; McKinney, 2007). Mental health professionals have internment during World War II, African American holocaust, and collective trauma and systemic oppression experienced by indigenous Native peoples (Danieli, 2007; Dekel & Goldblatt, 2008; Frazier, Goodman, St. Juste, & West Olatunji, 2009; Goodman & West Olatunji, 2008a; West Olatunji & Conwill, 2011). Violent historical events, discrimination, and intersecting oppressions experienced by prior generations can potential ly affect the lives of future generations in the form of unresolved PAGE 42 42 grief and ongoing trauma (Dass Brailsford, 2007; Goodman & West Olatunji, 2008b). Historical racism and modern day oppression contribute to the pervasive mistrust many peoples of color fee l toward the criminal justice system, educational institutions, and other government agencies, which are often viewed with suspicion and cynicism (Dass Brailsford, 2007; McCubbin & McCubbin, 2005).Within this context, the conspiracy of silence can be used as a defense for trying to prevent total collapse and breakout of intrusive traumatic memories and emotions (Danieli, 2007). However, it often inhibits meaningful dialogue that allows children, their ate trauma into a resilient experience (Danieli, 1998; Goodman & West Olatunji, 2008a). Many infants, young children, and school age children develop disorganized/disoriented attachments as a result of unresolved losses, traumas, and discord within the f amily that can lead to increased vulnerability to traumatic stress (Hesse et al., 2003; Moss et al., 2011; Solomon & George, 2011b). Parents posttrauma adaptational styles that become inflexible survival strategies tend to render young children, families, and communities vulnerable to further trauma (Danieli, lifeli ne: birth, developmental stages, transitions, changes, and death (Danieli, 2007). Overtime, these effects of silence and inflexibility may become transgenerational thus affecting succeeding generations (Danieli, 2009). The resulting emotional and behavior al effects of transgenerational t raumatic stress include high rates of depression, anger, anxiety, hypervigilance, low self esteem, suicidal ideation and behavior, substance abuse, violence, loss of language, cultural identity, and traditions (Dass Brailsf ord, 2007; Duran, Duran, Yellow Horse Brave Heart, & Yellow Horse Davis, 1998; Felsen, 1998; Raphael, Swan, & Martinek, PAGE 43 43 1998; Simons & Johnson, 1998). It is important to remember that these infants, toddlers, and preschoolers operate in the present even wh en they have become carriers of transgenerational trauma (Lieberman &Van Horn, 2008), therefore their temporal perspectives encompass past, present, and future; making historical traumas hold present meaning (Verbos, Gladstone, & Kennedy, 2011). To be heali ng and even self actualizing, the articulation and integration of traumatic experiences must be examined from an ecosystemic perspective (Danieli, 2007; McKinney, 2007). Systemic Oppression and H egemony Race, in the United States historically captures no t cultural practices and beliefs but societally imposed stigmatizations and marginalization ( Williams et al., 2010) These stigmatizations and marginalizations adversely affect health in complex ways in which race, ethnicity, and socioeconomic status (SES) combine to reflect systemic oppression, exploitation, and social inequality ( Williams et al.) This translates into race being defined as a social construction in which people are grouped and ranked based on their skin color, language, and physical featur es (Carter et al., 2011). Whether purposeful or unintentional these dehumanizing acts have led to people of color being reduced to things (Freire, 1970/2000). These limiting acts of systemic oppression and educational hegemony have been conceptualized as sources of traumatic stress symptoms and potential sources of transgenerational trauma, thereby providing greater understanding of the mechanisms of trauma (Bryant Davis & Ocampo 2005; Buhin & Vera, 2009; Carter, 2007; Carter & Forsyth, 2009; Carter et al. 2011; Lowe, Okubo, & Reilly, 2012; Turner & Pope, 2009; Sue et al., 2007; Sue & Sue, 2013). In the US, African Americans (West Olatunji & Conwell, 2011), Latina/os (Guzmn & Carrasco, 2011), Asian Americans (Kim, 2011), Native Americans (Garrett & Portma n, 2011), and Arab Americans (Nassar McMillan, 2011) have been shown to experience disproportionate PAGE 44 44 amounts of harmful mental and physical health effects due to oppressive experiences (Harrell et al., 2003; Paradies, 2006; Williams et al., 2003).These syst emic oppressions are dehumanizing and degrading, however, they are most often experienced as racial microaggressions (Choudhuri, Santiago Rivera, & Garrett, 2012; Freire, 1970/2000). Racial Microaggressions have been defined as brief and commonplace daily verbal, behavioral, or environmental indignities, whether intentional or unintentional, that communicate hostile, derogatory, or negative racial slights and insults toward people of color (Sue et al., 2007; Sue, 2010). These microaggressions are often del ivered environmentally through the physical surroundings of target groups, where they are made to feel unwelcome, isolated, unsafe, and alienated (Sue & Sue, 2013). traumatic stress theory criteria needing to be perceived as negative, sudden, and uncontrollable. A qualitative study by Lowe et al. (2012) supported this by asking participants to reflect on racist events, which triggered many to experience strong negati ve emotions and racism based traumatic stress. The responses from these participants resulted in symptoms that can be correlated with traumatic stress responses such as hypervigilance, reexperinecing of racist events, physiological arousal, avoidance, memo ry loss, detachment, estrangement from others, restricted affect, depression, increased aggression, and flashbacks. These authors also found that the participants were further traumatized when attempting to confide in someone about the racist experience wh ere this resulted in their experiences being met by resistance, denial, blame, and minimization; leading the victim to enhance the behaviors of social withdrawal, mistrust of others, lowered self efficacy, and feelings of powerlessness (Lowe et al.). PAGE 45 45 There fore, research shows that one's ability to trust, develop and maintain close relationships, achieve in school, perform at work, and function as a parent can all be impacted by racism related stressors (Choudhuri et al., 2012). Advocates of traumatic stress theory articulate an association between traumatic stress and systemic oppression, as evidenced in socially marginalized students' disengagement during the education process, acting out behaviors (internalizing and externalizing) that result in disproport ionate special education referrals, depression, and their low levels of self mastery within schools (West Olatunji & Conwill, 2011; Carter, 2007). These traumatic stress experiences place children on a trajectory of low education, low job prospects, and lo w income that often leads to the aggregation of other stressors ( Williams et al., 2010) In the presence of high levels of traumatic stress during early childhood, k of disease, and increased emotional and behavioral disturbances ( Williams et al.) Many children from these socially marginalized groups realize that they will likely encounter racially/ethnic based educational discrimination in the future (Buhin & Vera, 2009). Therefore, it is likely that when combined with a lack of economic and social advantages this may lead them to believe that their efforts in school are futile and that academic pursuits will have relatively little payoff (Buhin & Vera). Scholars b elieve that prevention of racism is a social justice issue considering the pervasively detrimental effects of racism on all members of our society and especially traumatizing effects for people of color (Buhin & Vera, 2009; Sue & Sue, 2013). Without taking some form of action for our early childhood children of color, studies show this will likely lead to continued wide experiences of transgenerational trauma (Turner & Pope, 2009; West Olatunji & Conwill, 2011). Alternately, professionals can utilize soluti ons to presenting problems PAGE 46 46 grounded in developmentally and culturally centered approaches in order to aid people of color in enhancing adaptive and positive outcomes from traumatic experiences due to systemic oppression (racism, classism, homophobia, and s exism), while reducing the symptoms of traumatic stress such as anxiety, fear, and depression (Carter et al., 2011; Goodman & West Olatunji, 2008a; West Olatunji & Conwill, 2011; Wolfgang, West Olatunji, Frazier, & Barrett, 2012; Wynn & West Olatunji, 2008 ). Traumatic Stress and Attachment in Early Childhood Each year more than four million children in the United States are exposed to a traumatic event (Levine & Kline, 2007). This is particularly alarming, because prenatal infants, newborns, and very young children are the most severely affected by stress and trauma due to their undeveloped nervous, motor, and perceptual systems (McCubbin & McCubbin, 2005; Levine & Kline). It is paramount to dispel the figment where many adults presuppose that infants, todd lers, and young children (ages 0 to 8) are the most resilient, rather this stage of development is when they are most vulnerable to the effects of trauma (Blaustein & Kinniburgh, 2010). It becomes important than to assess children looking at both traumatic stress and attachment disturbances (Lieberman, 2004). An issue that causes young children to be overlooked in assessing for traumatic stress and attachment difficulty is that most mental health professionals still approach children as if they were miniat ure adults, using scaled down versions of adult assessment techniques or adult models of normal/abnormal behavior and attempting to extrapolate these upon young children ( Straus, 1999; Vernon & Clemente, 2005). T here is also a greater need for emphasis on cultural differences in order to avoid clinician, teacher, or day care provider biases; secondarily this will also assist in culturally sensitive prevention practices, conceptualization, interpretation, and implementation of treatment (Paniagua, 2001; Sue & Sue, 2013). PAGE 47 47 Effects of Traumatic S tress The neurological responses that accompany traumatic stress responses in young children include increased levels of adrenaline (activation of sympathetic nervous system), decreased levels of cortisol and serotonin (reduced ability to moderate sympathetic nervous system), and increased levels of endogenous opioids (resulting in pain reduction, emotional blunting, and memory impairment) (Johnson, 1997). This can lead to an entangled experience of acting out and expres sing distress in the form not being able to organize material into beginning, middle, and end (Johnson). This often leads to children remembering things out of context from the ir experiences having difficulty with emotional acclimation (limbic brain) due to the activation of fight, flight, or freezing responses; and children can have difficulty with transitions and reengaging in relationships (Johnson; Johanson, 2006; Wilkerson et al., 2008). According to traumatic stress theory, experiences that can over whelm the four phases of secure attachment described above are separated into five categories: (a) accidents and falls, (b) medical and surgical procedures, (c) violent acts and attacks, (d) grief and loss, and (e) environmental stressors, such as transgen erational trauma (Carlson, 1997; Levine & Kline, 2007). The attachment system is highly responsive to indications of danger ; however, when infants, toddlers, and young children experience events that are perceived as negative, sudden, and out of control (C arlson, 1997; Goodman & West Olatunji, 2008b; Siegel, 2012), the attachment system can be interrupted (Lieberman, 2004) When infants, toddlers, and young children experience traumatic stress, they look to their familiar caregiver(s) for referencing ; howev er, when severe stresses depletes the caregiver(s) resources they can become neglectful or difficulty understanding the cues for safety versus risk ( Lieberman; Lieberman & Van Horn, 2008). When these type s of interruptions occur infants, toddlers, and young children will show PAGE 48 48 various signs of hyper/hypo arousal, avoidance, or re experiencing through verbal and nonverbal behaviors ( Bailey, Moran, & Pederson, 2007; De Young et al., 2011a, 2011b; Mongillo, Briggs Gowan, Ford, & Carter, 2009). Hyper/hypo A rousal When infants, toddlers, and young children become overwhelmed, they may exhibit fluctuations of hypo to hyperarousal; displaying frantic distress reactions s uch as wailing, gasping, and flailing about or expressing emotional numbing such as excessive sleeping, dazed expression, or averting their eye gaze (Frazier e al., 2009; Levine & Kline, 2007; Lieberman &Van Horn, 2008). They can also display inconsolable crying, shallow breathing, vomiting or spitting, arching the back, biting, and pushing away (Levine & Kline; Montirosso, Borgattie, & Tronick, 2010), particularly during feeding and sleeping times (Casanueva et al., 2010). Toddlers and young children can a lso show arousal by being overly silly, having trouble adjusting to changes, having trouble calming down when upset, being easily bothered by loud noises or bright lights, and have trouble falling asleep or staying asleep, experience recurring nightmares, loss of appetite, vomiting, diarrhea, experience recurrent and intrusive memories, a decline in pre school and kindergarten performance, learning problems, and difficulties with time and sequencing (Dass Brailsford, 2007; De Young et al., 2011a, 2011b; Jo hnson, 1997; Levine, 1997; Modrowski et al., 20 13 ). Avoidance Avoidance in infants can take the form of sleeping more than usual, withdrawal from family members and friends by displaying less affection, consistently looking away, or avoiding facial conta ct (De Young et al., 2011a, 2011b; Klein, Devoe, Miranda Julian, & Linas, 2009). In older toddlers and young children it can become fear of being separated from caregiver(s), crying PAGE 49 49 and refusal to follow directions, disorientation, and extreme sadness (Bog at, Levendosky, von Eye, & Davidson, 2011; De Young et al.). Re e xperiencing Re experiencing or recollection symptoms in infants do not always appear distressing, thus are harder to detect, and often coincide with experiences, behaviors, and sounds that oc curred during the distressing event (Graf, Schiestl, & Landolt, 2011) It can often coincide with avoidance and arousal where distress is shown during sleep, play, or feeding times (Klein et al., 2009). In toddlers and young children it becomes clearer tha t play or drawings are rigid, repetitive, and reenacting anxious qualities of the traumatic experiences, young children verbalize repeated questions around the event, can be seen to space out, and are able to clarify that they are having a nightmare (De Yo ung et al., 2011a, 2011b; Modrowski et al., 201 3 ; Mongillo et al., 2009). According to extant research, children who experience traumatic stress inducing events often exhibit the expected symptoms of re experiencing, avoidance, and arousal (De Young et al. 2011a, 2011b; Klein et al., 2009; Modrowski et al., 201 3 ). However, internalizing behaviors and externalizing behaviors can be common outcomes of traumatic stress (Briggs Gowan et al., 2010) and are important to include in our assessments when identifyin g traumatic stress. Traumatic Stress and Underachievement This section discusses traumatic stress and its effects on academic achievement, it attempts to summarize the above literature and connects the traumatic experiences children can have with insecure attachments, transgenerational trauma, systemic oppression, as well as the many traumatic events that can cause young children to perceive things as negative, sudden, and out of control, thus affecting their academic achievement. Bergin and Bergin (2009) and Moss and St Laurent (2001) have shown that insecure attachments can be disruptive as children PAGE 50 50 express their externalizing behaviors such as, exhibiting ADHD like symptoms, displaying less curiosity, being more dependent on teachers for direction, displ aying more aggression, and exhibiting impaired ability to control attention and memory. They also described the internalizing behaviors as causing issues when these children show fears of failure, seek less help from teachers, and display frightened and di ssociative behavior during class. Thus, studies show that considerable cognitive and attentional resources are necessary to cope with caregiver(s) who induce fear or feelings of abandonment or threat in children which results in the above behaviors and con siderably restricts the mental resources available for learning and exploration as a result of the trauma (Moss et al., 2011). When looking at the resulting emotional and behavioral effects of transgenerational traumatic stress, these have been shown to in clude high rates of depression, anger, anxiety, hypervigilance, low self esteem, suicidal ideation and behavior, substance abuse, violence, loss of language, loss of cultural identity, and loss of cultural traditions (Dass Brailsford, 2007; Duran et al., 1 998; Felsen, 1998; Raphael et al., 1998; Simons & Johnson, 1998). This trauma is complex due to possible intersectionality with other ecosystemic stressors (Morris, 2008). The affects include the above internalizing and externalizing behaviors with added d ifficulties around time orientation often leading to disorganized experiences (Danieli, 2007), this is especially detrimental in the classroom (Johnson, 1997). The associations between traumatic stress and systemic oppression, have been shown to have very significant negative outcomes as socially marginalized students disengage during the education process, due to difficulties trusting and maintaining close relationships, and displaying acting out behaviors (internalizing and externalizing), which lead to d isproportionate poor performance, low levels of self mastery, high dropout rates, and difficulties showing PAGE 51 51 achievement levels that match with their academic ability (Carter, 2007; Choudhuri et al., 2012; Moore et al., 2005; West Olatunji & Conwill, 2011). The importance of these trends is that m any times teachers are the first to notice trauma symptoms as they manifest in the classroom and on the playground. However, these often get interpreted through a deficit lens leading to misdiagnoses of anxiety, ADHD depression, and/or conduct disorders (Levine & Kline, 2007). Interventions for T raumat ic Stress W ith Early C hildhood It is important for counselors to consider developmental capabilities, growth fostering relationships, transformative acts of love, and cultural empathy when assessing, conceptualizing, and designing age appropriate interventions for infants, toddlers, and young children (Ahmed, Wilson, Henriksen, & Jones, 2011; Comstock et al., 2008; Duran, Firehammer, & Gonzalez, 2008; Seto, Young, Becke r, & Kiselica, 2006; Vernon & Clemente, 2005). By conceptualizing the relevant cultural variables, counselors can make more accurate and comprehensive assessments and provide more effective treatments (Crethar, Torres Rivera, & Nash, 2008). This comprehens ive approach to cultural and sociopolitical variables is a part of the social justice dialogue of professional counselors. By incorporating these culturally and developmentally sensitive ecosystemic perspectives the healing of trauma can become a natural p experiencing their bodies, emotions, and thoughts; thereby engaging and balancing the instinctual inner awareness of their body and its health (Levine, 1997; Sapolsky, 2004). A ddress ing the possible underres ponding or overresponding behaviors (Johnson, 1997) requires counselors, teachers, and parents to be aware that traumatic effects are not always apparent immediately following the incidents that caused them and often can remain dormant, accumulating over y ears or even decades (Levine). PAGE 52 52 Counselors, educators, and care providers can adjust their basic skills to better work with young children by focusing on three specific areas: (a) setting up a safe and warm environment, (b) establish trust by displaying tr ust through culturally sensitive gestures, tone of voice, facial expressions, and (c) nurture a non judgmental understanding of young clients while focusing on exploration, empowerment, and acceptance (Erdman & Lampe, 1996) By engaging in these three prac tices the care giving professional should be able to help young children (ages 0 to 8) work through a variety of social, emotional, behavioral, and learning challenges (Henderson & Thompson, 2011). Counselors, educators, and care providers can bring war mth and vitality back to (McDevitt & Ormrod, 2010) that create mediated learning so young children can develop the ability to self define, contextualize, and transform t heir reality into healthy developmental journeys (Wolfgang et al., 2012). This gentle, non threating, re balancing of the energy that was interrupted when the trauma occurred can create restorative opportunities (Levine, 1997). In summary, underachievemen t is still an issue of concern for all children in reading and mathematics but particularly for children of color, male children in reading, and for lower income children in general (National Center for Education Statistics, 2011a; 2011b). Educational res earch has portrayed a deficit orientation toward socioeconomic class and culturally diverse persons (Chilisa, 2012; King, 2005; Smith, 1999; West Olatunji & Gibson, 2012) and informs our continued discussions around closing the achievement gap (Biles et al ., 2012). Studies show that more support needs to be provided to young children in the area of early intervention that is supportive of culturally sensitive socio emotional development (Rubin, 1998). PAGE 53 53 In working with early childhood school aged children th e literature suggests that attachment theory might provide enhanced insight into how children engage their teachers during subsequent academic achievement (Bergin & Bergin, 2009; Cutrona et al., 1994; Pianta & Harbers, 1996), provides support that attachment behaviors are strongly related to school outcomes in preschool and elementary (Moss & St Laurent, 2001; Pianta & Harbers). Insecure attachments, such as avoidant, resi stant, and disorganized disoriented attachment tend to lead students to have less academic success (Bergin & Bergin; Moss & St Laurent). Academic underachievement might be influenced by a compromise in children's healthy attachment. This literature suggest s that there might be a relationship between traumatic stress and insecure attachment (Moss et al., 2011). While there have been some significant studies that inform our understanding of the role of traumatic stress in the persistent underachievement of so me groups of school children, attachment theory helps to better explain this phenomenon and may prove useful in developing solutions at an early stage in children's schooling experiences. Early childhood development requires conceptualization to be from an ecosystemic frame 2010), and the extended timeframes can be understood (Peluso et al., 2010; Siegel, 2003). Traumatic stress theory identifies the effects of traumatic events on people as being perceived as negative, sudden, and uncontrollable (Carlson, 1997; Goodman & West Olatunji, 2008b, 2010). These traumatic events can be unknown due to transgenerational influences (Blaustein & Kinniburgh, 2010; Hesse et al., 2003), the impact of racism related stressors (Choudhuri et al., 2012), and the extent to which the child is exhibiting reexperiencing, arousal, avoidance, PAGE 54 54 internalizing behaviors, and externalizing behaviors (Bailey et al., 2007; De Young et al., 20 11a, 2011b; Mongillo et al., 2009). By incorporating these culturally and developmentally sensitive ecosystemic perspectives the healing of trauma can become a natural process (Levine, 1997; Sapolsky, 2004). Counselors, educators, and care providers can ad just their basic skills to better work with young children and the care giving professional should be able to help young children (ages 0 to 8) work through a variety of social, emotional, behavioral, and learning challenges (Erdman & Lampe, 1996; Henderso n & Thompson, 2011). Professional counselors therefore can assist educators and care providers in applying these skills and knowledge within early childhood centers, elementary schools, pre K classrooms, and community mental health agencies. PAGE 55 55 CHAPTER 3 METH ODOLOGY This study explore d the issues of academic underachievement and their relationship with early childhood mental health and development (Brooks Gunn, Berlin, & Fuligni, 2000; Zeanah & Zeanah, 2009). T he relationship between traumatic stress symptoms and difficult attachment symptoms in respect to academic achievement among early childhood school aged children was explored. This study propose d to examine the following five research questions: (1) What is the relationship between the growth trajectory o f reading and math achievement and traumatic stress symptoms for children in kindergarten through fifth grade?, (2) What is the relationship between the growth trajectory of reading and math achievement and difficult attachment symptoms for children in kin dergarten through fifth grade?, (3) What is the relationship between the growth trajectory of reading and math achievement and traumatic stress symptoms and difficult attachment symptoms for children in kindergarten through fifth grade?, (4) What is the re lationship between the growth trajectory of reading and math achievement and traumatic stress symptoms and difficult attachment symptoms for children in kindergarten through fifth grade, while controlling for gender, culture, and socioeconomic status? and finally, (5) What is the relationship of an interaction effect between traumatic stress symptoms and difficult attachment symptoms on the growth trajectory of reading and math achievement for children in kindergarten through fifth grade, while controlling for gender, culture, and socioeconomic status? The purpose of this chapter is to provide an overview of how the methodology was chosen for examining th e above relationship s and the proposed ways of using them to examine ly Childhood Longitudinal Study, Kindergarten (ECLS K) post hoc data set. This chapter will cover the development of the model, the setting of the post PAGE 56 56 hoc data set, the participants, variables, instrumentation, data collection, and proposed data analysis. Research Constructs and Theoretical Assumptions This study and its research questions were shaped by my experiences working with young children as a professional counseling intern and through the collaboration and discussions in the Pediatric Counseling R esearch Group formed at the University of Florida. This team of researchers gathers to discuss, assist, and develop new areas within counseling research. The construct of traumatic stress evolved from the theory development, research, and discussions from Dr. West Olatunji and Dr. Goodman. Their assumptions and theoretical constructs were has published many articles in the area of traumatic stress and continues to publish on the constructs surrounding traumatic stress and academic achievement (Frazier et al., 2009; Goodman, 2009; Goodman et al., 2012; Goodman & West Olatunji, 2008a, 2008b, 2010). Her work took items designed to measures the levels of stress and dis tress in children and created the latent factor for traumatic stress. This construc t was based on common symptoms of traumatic stress among children, as well as the premise that some children with indicators high for traumatic stress may have lower academi c achievements (4th ed., text rev.; DSM IV TR; American Psychiatric Association, 2000; R. D. Goodman, personal communication, March 7, 2013; Zero to Three, 2005). My area s of stud y have centered on early childhood development and attachment. This has been attachment relationships affects traumatic stress and academic achievement. As was informed through the literature review, a longitudinal design was important in order to un derstand the effects on early childhood development and academic achievement. This literature review PAGE 57 57 suggest ed hypothesized that attachment behaviors might be an important aspect of the relati onship between traumatic stress and academic achievement. Therefore, I include d indicators that the literature suggest ed might make up the latent variable difficult attachment symptoms, in order to explore this relationship. Setting This study use d data extracted from the ECLS K, which was characterized by a dual frame multistage complex stratified cluster sampling design that included a series of design weights in order to provide acceptable estimates about minority and poor children which were oversampl ed (Dawson & Williams, 2008; DiPerna, Lei, & Reid, 2007; Jyoti, Frongillo, & Jones, 2005; Ready, LoGerfo, Burkam, & Lee, 2005; Tourangeau, L, & Nord, 2005) In the baseline year, this nationally representative and longitudinal study of 21,409 children who entered kindergarten are situated within a primary sampling unit made up of four identified geographic areas, the second stage units were randomly sampled from 934 public schools and 346 private schools within these geographic counties, and the third stag e units were the students, 24 kindergarteners within each public school and 12 within each private school (Carlson et al., 2008; Dawson & Williams; Tourangeau et al.; Tourangeau, Nord, L, Pollack, & Atkins Burnett, 2006) The ECLS K base data was collecte d during the 1998 1999 academic school year and continued until intervals: fall and the spring of kindergarten (1998 99), fall and spring of 1st grade (1999 2000), spring of 3rd grade (2002), spring of 5th grade (2004), and spring of 8th grade (2007) (National Center for Education Statistics [NCES], 2001). Participants The target sample of this longitudinal study include d the children, their parent(s)/guardian(s), and teachers from th e ECLS K data collection base year respondents PAGE 58 58 (kindergarten, Fall 1998), all of whom were followed to their fifth grade year (Spring 2004). Therefore, the repeated measures will include the direct cognitive assessments for reading and mathematics data fo r children in the kindergarten year through 5th grade. The complex stratified cluster sampling design was addressed using the software program M plus v.6.12 (Muthn & Muthn, 2010) T he estimation method utilized for all models was the robust maximum likel ihood (MLR) method of estimation, which provide d Satorra Bentler corrected standard errors and test statistics (Jackman, Leite, & Cochrane, 2011). The MLR method also has the ability to obtain unbiased parameter estimates using the stratification (C16FPSTR =regional strata), cluster (C16FPPSU=the county primary sampling unit), and weights (C1_6FP0=sampling weight) information made available in the ECLS K data set This method use d Taylor series design based methods in order to adjust the data allowing for co rrected standard errors (DiPerna et al. 2007). Because some of the files in this data set includ e d cases that have exclusions all the way through, they w ould not have K 5 longitudinal weights. For example, exclusions will include base year respondents who did not participate in spring of first grade, spring of third grade, and/or spring of fifth grade. Other exclusions from the total sample are based on the following circumstances: (a) if the children died or moved out of the country, (b) if the children m oved during the study, and (c) children for whom there are neither first grade, third grade, nor fifth grade data. These exclusions cause d participants to not have any K 5 longitudinal weights and therefore w ere not included in the final data that was extr acted (Tourangeau et al., 2006). The children selected in this study were chosen for various criteria, such as (a) they participated in various measures on their abilities and skills around academic achievement in reading and mathematics, (b) exhibited beh aviors as reported by parent(s)/guardian(s) and/or PAGE 59 59 teachers based on evaluations from a number of social emotional dimensions that might suggest they potentially could be experiencing symp toms of traumatic stress at entry into the kindergarten year (Fall & Spring, 1998 1999), and (c) the parent(s)/guardian(s) provided perceptions of their own behaviors and/or reported perceptions about the child based on evaluations of the social emotional relationship of the child and the parent(s)/guardian(s) which may p rovide symptoms of potential difficult attachment relationships at entry into the kindergarten year (Spring, 1999). Therefore, this was a academic progress during the kindergarten and early elementary school years and its relationship of difficult attachment relationships at school entry into kindergarten ( Tourangeau et al., 2006 ). Researcher Sub jectivity The purpose of this researcher subjectivity statement in a quantitative study is based on an emancipatory methodological position. In an emancipatory research framework the researcher is explicit in describing their own subjectivity and in operat ionalizing the values and beliefs that underpin how the researcher conceptualizes and interprets the data in the analyses ( West Olatunji et al., 2012, May). This is done to make the researchers cultural, historical, and contemporary experiences known so th at their complexities, power relations, and life experiences can be contextualized within the findings (Ponterotto, 2005; Rose & Glass, 2008; Tillman, 2002, 2006). This is important to make explicit as it informs how decisions were made in sampling, theory and methodology. As a European American male who has lived and worked with multicultural populations from around the world, I see our theories and methodologies as being culturally situated; the meanings and assumptions rely heavily on our values and bel iefs. Having had lived experiences PAGE 60 60 with marginalizing early childhood events of complex traumatic stress, both these lived experiences and my professional training shape my choice in theory, methodology, how I conceptualize, and the questions that I bring to this study. As a professional counselor I work from an ecosystemic perspective, addressing client needs from a developmental, culturally sensitive, strength based, and preventive fashion. I make a conscious effort to be aware of my biases and attempt t o ground my personal perspective and professional practice in empirically supported models of practice so that my contributions to clinical work and the research conversation are supported by best practices within the profession of multicultural counseling with early childhood populations. Operational Definition of Variables The variables I used for this study include d : (a) the latent variable traumatic stress symptoms, (b) the latent variable difficult attachment symptoms, (c) the direct cognitive assessme nts for reading and mathematics academic achievement, (d) gender, (e) gender of parent/guardian present during in person at home assessments, (f) culture/ethnicity, and (g) socioeconomic status (Marsh et al., 2009). The latent variable traumatic stress in clude d a confirmatory factor analysis model that constitutes the observed composite variables: (a) a ttention qualities of children (P1ATTENI) from the Fall parent questionnaire and the following from the Spring of kindergarten Teacher Social Rating Scale (SRS) data: (b) s elf control (T2CONTRO), (c) i nterpersonal s kills (T2INTERP), (d) e xternalizing behaviors (T2EXTERN), and (e) internalizing behaviors (T2INTERN). The latent variable traumatic stress symptoms, was determined to be present through identifica tion of the following symptoms of traumatic stress: (a) reexperiencing, (b) avoidance, (c) arousal, (d) externalizing, and (e) internalizing behaviors. Re experiencing was indicated by the variable Self ntrol behavior is PAGE 61 61 observed sometimes to never. The avoidance measure was indicated by the variable Interpersonal Skills, which looks at whether friendships, comforting other children, expressing feelings and op inions in positive ways, and showing sensitivity is observed sometimes to never. Arousal was indicated by attention qualities of children (P1ATTENI) perceived to be less well than other chil dren. Externalizing behaviors involved observing very often to often acting out behaviors such as arguing, fighting, getting angry, acting impulsive, and disturbing activities. Internalizing behaviors involved observing very often to often anxiety, lonelin ess, low self esteem, and sadness. The latent variable attachment difficulties, includes a confirmatory factor analysis model that includes the following ordinal categorical observed variables from the Spring kindergarten parent data: (a) by the end of a long day, hard to be warm and loving toward child (P2HRDWRM), (b) child does things that bother me (P2CHDOES), (b) feel trapped as a parent (P2FLTRAP), (c) often feel angry with child (P2FEELAN), and (d) child is harder to care for (P2CHHARD). These variab les are used to approximate the attachment relationship between parent and child. The repeated measures reading and mathematics academic achievement include d five a cademic achievement time variables include d assessments from: (1) Fall 1998 kindergarten (C1R4RSCL), (2) Spring 1999 kindergarten (C2R4RSCL), (3) Spring 2001 first grade (C4R4RSCL), (5) Spring 2002 third grade (C5R4RSCL), and (6) Spring 2004 fifth grade ( C6R4RSCL). Mathematics academic achievement time variables include d assessments from: (1) Fall 1998 kindergarten (C1R4MSCL), (2) Spring 1999 kindergarten (C2R4MSCL), (3) PAGE 62 62 Spring 2001 first grade (C4R4MSCL), (5) Spring 2002 third grade (C5R4MSCL), and (6) S pring 2004 fifth grade (C6R4MSCL). The variable gender was dummy coded indicating male (0) or female (1). The variable responding parent/gu a was dummy coded indicating male (0) or female (1). The variable ethnicity was dummy coded into four variables so that White will be the reference group (0), while Asian, Black, Latino/a, and Other People of Color will make up the four occupation, education level, and household income. This study theorize d that the time variant dependent variables reading and mathematics academic achievement might have be en impacted by the time invariant covariates (TICs) traumatic stress symptoms and difficult attachment symptoms. The TICs gender, parent/gu a were included to measure reading and mathematics academic achievements relationship with the sociodemographic factors as well as controlling for unknown variables that impact the possible relatio nship between academic achievement and traumatic stress symptoms and difficult attachment symptoms. Instrumentation Th is study use d direct cognitive assessments, interviews with parents, and teacher questionnaires. Socioeconomic status (SES) was calculated by the NCES through a composite of and household income ( Tourangeau et al. 2006).The direct cognitive assessment battery covered three content areas, but only reading and mathematics was constantly measured longitudinally from kindergarten to fifth grade. The questions for the cognitive assessments were developed by consu lting with experts in the field of education to ensure that they developmentally reflected PAGE 63 63 content areas and content validity (Pollack, Atkins Burnett, Najarian & Rock, 2005; Tourangeau et al., 2006). The questions that passed screening for construct valid ity evidence and sensitivity, were then field tested, allowing the creation of the cognitive assessments provided in either number right scores or raw score count (Item Response Theory (IRT) scale scores or standardized scores [T scores]). In this study, I use d the IRT scale scores, which have reliabilities for the reading and mathematics of .93 and .94, respectively (Tourangeau et al., culture. The re were four co mposite continuous variables from the Spring of kindergarten Teacher Social Rating Scale (SRS) data that were used for the traumatic stress symptoms variable They included : (a) T2CONTRO, (b) T2INTERP, (c) T2EXTERN, and (d) T2INTERN. rated the fr equency with the child exhibited the social skill or behavior on the following scale: (1) never, (2) sometimes, (3) often, or (4) very often The scale for T2CONTRO and T2INTERP were reverse scored so that when regressed on reading and mathematics achievem ent their factor loadings and traumatic stress symptoms factor coefficients would be more inline and interpretable with scores of children weak in these areas. The parent interview consisted of approximately 330 questions and covered topics such as child health, parent and family characteristics, child care, and school experiences. For the purpose of this study, I use d P1ATTENI from the Child Health and Wellbeing questionnaire for the traumatic stress latent variable. This question asks parents to rate ho w the child pays attention in comparison to other children his or her age. The parent had the following response choices: The last two categories were parcel ed into PAGE 64 64 This resulted in three aggregated ordinal categories for P1ATTENI: (1) better than other children, (2) as well as other chi ldren, and (3) slightly less well/much less well than other children. I initially selected eleven variables from the Parent Interview Questionnaire that were used as indicators for the difficult attachment symptoms latent variable during the exploratory f actor analysis (EFA) Five of these come from the Discipline, Warmth, and Emotional Supportiveness questionnaire (DWQ). These were chosen based on the literature and from a personal communication by email with Dr. Gail Mulligan, the project officer for the Early Childhood Longitudinal Studies National Center for Education Statistics (Mulligan, personal communication, October, 15 2012). She stated that there were no direct measures of attachment included in the ECLS K, however there are questions located in the Discipline, Warmth, and Emotional Supportiveness questionnaire (DWQ) section of the Spring kindergarten parent interview questionnaire that are related to the parent/child relationship and that these have been of interest to attachment researchers. The five included the following: P2HRDWRM, P2CHDOES, P2FLTRAP, P2FEELAN, and P2CHHARD). These five questions asked the parent to rate the following: (a) at the end of a long day, did parent/guardian find it hard to be warm and loving toward child (P2HRDWRM), (b) if the child does things that really bothers the parent/guardian (P2CHDOES), (c) parent/guardian feels trapped by their responsibilities as a parent (P2FLTRAP), (d) parent/guardian often feels angry with the child (P2FEELAN), and (e) child seems harde r to care for than most (P2CHHARD). Completely true and mostly true were parceled into one category, thus creating three ordinal categories. PAGE 65 65 Four indicators were taken from the CFQ : Critical Family Processes questionaire (P2ARGCHD, P2ARGAFF, P2ARGHEA, P2HITCHR) and two indicators were taken from the PPQ : Parent Psychological Well being and Health questionnaire (P2DEPRES a nd P2FEARFL). The four indicators from the CFQ asked the parent to rate the following: (a) do you and your partner have arguments about the child (P2 ARGCHD ), (b) do you and your partner ague about showing love and affection (P2 ARGAFF ), (c) when you have a serious disagreement with your partner do you argue heatedly or shout at each other (P2 ARGHEA ), (d) when you have a serious disagreement with your partner how often do you end up hitting or throwing things at each other (P2 HITCHR ) The parent had the follo often sometimes hardly ever never us, (P2ARGCHD, P2ARGAFF, P2ARGHEA, P2HITCHR) had 4 ordinal categories each The two indicators from the PPQ asked the parent to rate the following: (a) how often duri ng the past week have you felt depressed ( P2DEPRES ), and ( b ) how often in past week have you felt fearful (P 2 FEARFL ) The parent had the following choices for these two questions: Thus, (P2DEPRES and P2FEARFL) had 4 ordinal categories each. called for a follow up phone interview lasting approximately 5 minutes. Of those re interviewed, 94% r eported the same answers as in the original interview (Tourangeau et al., 2006). Providing a relative sense that parents were answering the questions with consistency. The teacher assessments were taken from the Teacher Social Rating Scale (SRS) which w as an adaptation, used with permission, of the Social Skills Rating Scale: Elementary Scale A (SSRS) by Gresham and Elliott ( Tourangeau et al., 2006 ) From the SRS, the Self Control scale, PAGE 66 66 Interpersonal Skills scale, Externalizing Problem Behaviors, and In ternalizing Problem Behaviors scale had split half reliabilities of .80, 0.89, .90, and .78 respectively (U. S. Department of Education (NCES), 2006). Data Analysis Informed by K data set around these proposed variables and personal communications with R. D. Goodman (Goodman, 2009, personal communication, March 7, 2013; Goodman et al., 2012), and the literature review from chapter two I used a latent growth curve model (LGCM) with five rep eated measures and nine TICs (Duncan, Dunca n, & Strycker, 2009). I utiliz ed M plus when analyzing latent variable multi level longitudinal data where subjects are followed over time with repeated measures of each variable of interest. This data analyses us ed a methodological substantive process that explore d relationships using exploratory factor analyses (EFA ; Marsh et al., 2009 ), declar ed these relationships with confirmatory factor analyses (CFA ; Asparouhov & Muthn, 2009 ), and analyz ed the causal nature of these relationships using structural equation modeling (SEM) so that it c ould be applied to the LGCM (Bollen & Curren, 2006; Marsh et al., 2009). This analysi s was conducted systematically, looking at the LGCM from simple structures to more complex structures, allowing the trends in the analysis to be seen in a step wise manner so that misspecifications can be determined from the simplest model. In this study model estimation used full information maximum likelihood (FIML) as this approach handles both missing completely at random (MCAR) and missing at random (MAR) data, and is utilized by default in M plus ( Muthn & Muthn, 2010) FIML uses all the informatio n of the observed data and maximizes the likelihood of the model given the observed data (Wang & Wang, 2012). PAGE 67 67 There were s everal robust estimators a vailable in M plus for dealing with nonnormality I use d the MLR estimator (Muthn & Muthn, 2010). According to Muthn and Muthn (2009 March 23 ), MLR provides maximum likelihood parameter estimates using a sandwich estimator that allows for robust test statistic and standard errors. Since this study is using a complex survey data set, I ran the TYPE=COM PLEX estimator, which allows MLR to be robust to non normality and non independence of observations. The MLR statistic is asymptotically equivalent to the Yuan Bentler test statistic, which does not make a specific distributional assumption (Muthn & Muthn, 2009 March 23 ; Yuan & Bentler, 2000). Specification of Measurement Mode l The first step beg a n with the specification of the two linear LGCM measurement models, one for reading achievement and the second for mathematics achievement. The goal was to determine if these two models fit adequately ( Figure 3 1 and Figure 3 2 ) (Hoyle, 1995; Kaplan, 2009; Kline, 2011). Adequacy of fit has predominantly focused on assessments of goodness o f fit (GOF) using the GOF test statistic and the likelihood ratio difference test (LRT) (Leite & Stapleton, 2011). However, because this study used the MLR estimator the LRT was calculated by using procedures suggested by Bryant and Satorra (2012). Adequacy of fit. The adequacy of fit was determined using the following GOF indices: (a) the comparative fit index (CFI), (b) the Tucker Lewis Index (TLI), (c) the standardized root mean residual (SRMR), and (d) the root mean squared error of approximation (RMSEA) ( Grimm & Ram, 2009; Kline, 2011 ; Leite & Stapleton, 2011 ; Marsh et al., 2009 ; Wu & West, 2010). The recommended level of acceptance for CFI & TLI was anything greater than .95 SRMR was anything less than 0.09, while RMSEA was anything less than 0 .06 ( Jackson, Gillaspy, & Purc Stephenson, 2009 ) D ue to some models needing numerical integration the AIC and BIC will be PAGE 68 68 used to determine adequacy of fit because the and re g ular fit indices ( CFI, TLI, SRMR, and RMSEA ) are not able to be estimate d ( Grimm & Ram, 2009; 2011; Grimm, Ram, & Hamagami, 2011) However, Leite and Stapleton (2011) found that CFI, TLI, and SRMR are affected by growth shape, while SRMR was shown to be affected by sample size and performed worse as the sample size increased RMSEA was found to be the best of the above at distinguishing misspecifications of growth shape in LGCM, and so the RMSEA is of most importance in the measurement model for the linear growth curve model. T he traditional structural equation modeling (SEM) p rograms have been shown to use the wrong baseline model for t he CFI and TLI therefore these were corrected using procedures described by Wu and West (2010) If either/both models were found to have inadequate fit to the growth trends of the data, as was a nticipated, due to missing data or possible nonlinear trends in the data that might be potential causes to convergence issues the model was modified using the following methods: (1) modification indices, (2) nonlinear trajectories using nonlinear freed lo ading LGCM (Bollen & Curren, 2006), (3) nonlinear modeling in the form of orthogonal polynomial functions estimating quadratic and cubic trends (Kelley & Maxwell, 2008; Kim, Cicchetti, Rogosch, & Manly, 2009; Mayer, Steyer, & Mueller, 2012; Shin, Davison, Long, Chan, & Heistad, 2013; Wang & Wang, 2012), and finally (4) moving to a hierarchical multivariate linear model (HMLM ; Raudenbush, Brennan, & Barnett, 1995; Raudenbush & Byrk, 2002; Thum, 1997; Usami, 2009). Modification indexes. The modification index es used estimate d the amount the overall model statistic that w ould decrease if a particular fixed to zero parameter were to be freely estimated and the greater the value of a modification index the better the predicted improvement in overall fit ( Kline, 2011). The modification indexes that make theoretical sense and were PAGE 69 69 predicted to make a substantive change to the models were added one at a time. The additional parameters that were added were chosen based on the literature review and attempt ed th rough respecifications to improve the model adequacy of fit. In this study the linear LGCM prove d inadequate, I attempt ed a non linear freed loading model. The freed loading model consist ed of fixing the first loading on the slope factor to 0, the secon d loading to 1, and freely estimat ing the remaining three loadings from the data (DiPerna et al. 2007) The freed loading model g i ve s flexibility in fitting non linear forms and is less demanding than the more involved use of poly nomial trajectories ( Figure 3 3 and Figure 3 4 ) (Bollen & Curren, 2006). In this way the model w as allo w ed to let the data inform the growth curve trajectory. These two LGCM s for reading and mathematics achievement require d the utilization of general orthogonal polynomial functions for unequal time intervals ; this was in order to fix the loadings to specific a priori values that are not correlated (Hedeker & Gibbons, 2006) Orthogonal polynomials were used in order to allow the model to estimate possible nonlinear quadratic and cubic trends in the trajectories without the typical multicollinearity issues (Kelley & Maxwell, 2008; Kim et al. 2009; Mayer et al. 2012; Shin et al. 2013; Wang & Wang, 2012). Table 3 1 provides the unequal time intervals and the subsequent orthogonal polynomials that were used for these models ; these orthogonal polynomials were created with the R statistical software v. 3.0.1 using the contr.poly command The advantage of the nonlinear quadratic (change in the rate of change over time in the linear growth) and cubic growth (change in the change in the rate of change over time in the linear growth) is it offers more complex explanations within the linear trend (Bollen & C urren, 2006; Muthn & Muthn, 2010). The other advantage to orthogonal polynomial curves are as PAGE 70 70 follows: (a) they allow for more accurate estimate s of developmental curves, (b) can be estimated for unequal time intervals, and (c) allow for a possible satu rated model to be estim ated if neede d (Hox, 2010; Snijders & Bosker, 2012). Final measurement model. The final model used in this study combine d the use of the general orthogonal polynomial model with the use of the hierarchical multivariate linear model ( HMLM ), which was developed for multilevel hierarchical linear modeling (HLM) by Raudenbush and Bryk (2002) Multilevel and SEM approaches have been seen as practical implementations of the same growth curve approach, differing on data structure and model a pproach, however, identical models can be constructed using both approaches (Grimm & Ram, 2012). M plus is able to estimate an HMLM model by (a) setting the variances of all trend latent variables (intercept, slope, and higher level polynomials) equal to ze ro, and (b) allowing the residuals for all observed reading and mathematics achievement variables to correlate ( Raudenbush & Bryk) The HMLM had an unrestricted covariance matrix (Raudenbush et al. 1995; Raudenbush & Byrk ; Thum, 1997; Usami, 2009), in the sense that the variances and covariances were no longer a within person covariance; rather, it captured all variation and covariation among the five repeated observations regardless of missing data. This addressed the convergence issues and biases in the estimates due to the missing data (Raudenbush, Bryk, Cheong, Congdon, & du Toit, 2011). HMLM model s ha ve the ability to be tailored to handle a variety of covariance structures (Hox, 2010; Raudenbush et al. 1995; Raudenbush & Byrk, 2002; Thum, 1997; Usami 2009) and they w ere developed as a way of estimating multivariate normal models from incomplete data (Raudenbush & Byrk). By capturing all variation and covariation among repeated PAGE 71 71 observations regardless of missing data, the HMLM address ed the convergenc e issues and biases in the estimates due to missing data ( Figure 3 5 and Figure 3 6 ) (Raudenbush et al. 2011). Confirmatory Factor Analyses The process of developing the confirmatory factor analy sis for the two TIC latent variables was different. For traumatic stress symptoms, it follow ed the traditional latent variable measurement specification that is based on prior developed theory and previous research that used logit and multiple regression a nalyses (Asparouhov & Muthn, 2009; Goodman, 2009; Goodman et al., 2012; R. D. Goodman, personal communication, March 7, 2013). The indicators for traumatic stress symptoms w ere adapted for a structural equation model analysis utilizing measurement model c onfirmatory practices (CFA; Appleyard, Yang, & Runyan, 2010). The latent variable difficult attachment symptoms, however, ha d not been developed in previous work. Therefore, this proxy variable w as developed through exploratory techniques that utilize d us ual SEM based standard errors and overall tests of mode l fit (Asparouhov & Muthn, 2009). This data analyses use d a methodological substantive process that explores relationships using exploratory factor analyses (EFA), declaring these relationships with c onfirmatory factor analyses (CFA), and analyzing the causal nature of these relationships using structural equation modeling (SEM) so that it c ould be applied to the LGCM (Bollen & Curren 2006; Marsh et al., 2009). Development of the Proxy Latent V ariable The latent variable difficult attachment symptoms was estimated using M plus v.6.12 This involve d that us ed an oblique geomin rotation and constrain ed the EFA models to extract 1 to 4 factors from the eleven selected indicators that were covered in the instrumentation section (Muthn & Muthn, 2010). This provide d insight into which items were : (a) most adequately loading, (b) provid ing conceptual clarity, (c) interpretab le and (d) PAGE 72 72 theoretical salience with the hypothesized factors (Klin e, 2011). The goal was to have an adequate number of indicators and for these to load on only the difficult attachment symptoms latent factor. The items of importance load ed greater than .40 on the relevant factor and less than .40 on the other possible fa ctors (Ang, 2005). I discard ed those that were loading less than .40 and the model was reconstructed using the high loading indicators. Measurement C on firmatory Factor A nalyses The next step utilize d measurement model confirmatory factor analyses (CFA; Ap pleyard et al. 2010) to provide evidence that these two latent factors ha d adequacy of fit (Hoyle, 1995; Kaplan, 2009; Kline, 2011). Due to the presence of categorical indicators, the CFA models were run with the indicators specified as continuous in orde r to get a reference AIC & BIC, as well as the chi square ( ) test statistic, loglikelihood ratio test (LRT), CFI, TLI, SRMR, and RMSEA to provide an idea of the models fit (Kline). I then compare d th ose AIC & BIC to an estimation run with the indicators specified as ordinal categorical, so that a logit model (which does not provide a fit test or fit indices) would run A gain the modification indexes were c onsult ed as well as looking at the reliability of the indicators (Leite & Zuo, 2011). Structural Model of LGCM The next stage of the analysis involve d develo ping the structural model for th e latent growth curve This include d a repeated measure of reading and mathematics achievement nested within individuals as the level one dependent variable. The level two variables were the following time invariant covariat es ( TIC s) or co varying independent variables that do not change over time (Wang & Wang, 2012): (a) the latent variable traumatic stress symptoms, (b) the latent variable difficult attachment symptoms, (c) gender, (d) gender of parent/guardian present duri ng assessments (interviewee gender) (e) culture/ethnicity, and (f) socioeconomic status. PAGE 73 73 Th e structural model w as a HMLM cubic orthogonal polynomial LGCM which was used to answer the first four research questions. Therefore, four different structural mode ls were implemented and each of these f our models include d an estimate for reading achievement and an estimate for mathematics achievement ( Figure 3 5 and Figure 3 6 ). The I nteraction In order to l ook at the interactions between the latent variables traumatic stress and difficult attachment symptoms, I utiliz ed M plus latent moderated structural equation ( LMS ) approach for the interaction modeling (Dimitruk, Schermelleh Engel, Kelava, & Moosbrugger 2007; Kelava & Brandt, 2009; Klein & Schermelleh Engel 2010). The LMS method was especially developed for the analysis of nonlinear structural equation models, such as when exploring interaction effects of latent variables, which explicitly take heteros kedasticity into account and do not require the formation of product variables (Kelava et al., 2011; Klein & Moosbrugger, 2000; Marsh et al., 2004; Dimitruk et al., 2007). Simulation studies have shown that LMS provides consistent, asymptotically unbiased, asymptotically efficient, and asymptotically normally distributed estimates which help the analysis to not be affected by problems resulting from low reliability of latent interactions (Dimitruk et al.). LMS was established by Schermelleh Engel, Klein, a nd Moosbrugger (1998) and Klein and Moosbrugger (2000) to better solve the non normality problems experienced by latent interaction effects, as these effects are not distributed normally even when indicators of first order factors are normally distributed (Marsh, Wen, & Hau, 2006). LMS was derived under the assumption of multivariate normality based on the concepts of mixture and conditional distributions, allowing the observed non normal distributions to be partitioned into aggregates of normal distributio ns while holding one or more other variables constant, each at a particular value (Cham, West, Ma, & Aiken, 2012; Jackman et al., 2011; Kelava et al., 2011). This allows PAGE 74 74 for asymptotically correct standard errors for nonlinear effects by applying the expec tation maximization (EM) algorithm which results in maximum likelihood (ML) parameter estimates (Jackman et al., 2011; Lyhagen, 2007). This is important for applied researchers using inferential statistics, as having biased estimated standard errors can cr itically affect modeling performance around significance testing (Dimitruk et al, 2007). SEM often incorporates linear relationships among latent variables and uses the model chi square and GOF indices in order to compare the fitted covariance matrix of a target model to an unstructured covariance matrix to assess global and adequacy of fit, however, when nonlinear terms are included (polynomial functions and interaction terms) this becomes problematic (Klein & Schermelleh Engel, 2010). The problem is that these models are not nested within the saturated model that is represented by the unstructured covariance matrix, thus not allowing the use of the GOF test nor any of the GOF indices as these tests require specification of a saturated model (Kelava et al., 2011; Little, Bovaird, & Widaman, 2006). The restrictions on the mean and covariance structures therefore, only allow for comparisons of non nested models using the Akaike Information Criteria (AIC) and Bayesian Information Criteria (BIC); when using the AIC and BIC indexes smaller values indicate better models both in terms of model fit and model parsimony (Kelava et al.; Little et al.). F inal S tructural M odel. The final structural model for the fifth research question used a HMLM cubic orthogonal latent growth curve model with nine TICs and an interaction between traumatic stress symptoms and difficult attachment symptoms using the LMS method This included an estimate for reading achievement and an estimate for mathematics achievement ( Figure 3 7 and Figure 3 8 ). PAGE 75 75 Table 3 1. Orthogonal polynomial coefficients for unequal time points 0, 4, 16, 40, 64 months Trend Time 1 Time 2 Time 3 Time 4 Time 5 Intercept 0.4472136 0.4472136 0.4472136 0.4472136 0.4472136 Linear 0.4610972 0.3867267 0.1636151 0.2826080 0.7288310 Quadratic 0.4149989 0.1496717 0.4305151 0.6198129 0.4856573 Cubic 0.4128476 0.1300984 0.6735047 0.5713284 0.1805729 Figure 3 1 Linear l atent g rowth c urve measurement m odel of reading achievement. PAGE 76 76 Figure 3 2 Linear l atent g rowth c urve measurement m odel of mathematics achievement. Figure 3 3. Freed loa ding l atent g rowth c urve measurement m odel for reading achievement. PAGE 77 77 Figure 3 4. Freed loading latent growth curve measurement model for m athematics achievement. PAGE 78 78 F igure 3 5. HMLM c ubic polynomial l atent g rowth c urve structural m odel with nine tim e invariant covariates for reading achievement. PAGE 79 79 Figure 3 6. HMLM cubic polynomial latent growth curve structural model with nine time invariant covariates for mathematics achievement. PAGE 80 80 Figure 3 7. HMLM cubic polynomial latent growth curve structural model with nine time invariant covariates and the interaction between traumatic stress symptoms and difficult attachment symptoms for reading achievement. PAGE 81 81 Figure 3 8. HMLM cubic polynomial latent growth curve structural model with nine time invariant covariates and the interaction between traumatic stress symptoms and difficult attachment symptoms for mathematics achievement. PAGE 82 82 CHAPTER 4 RESULTS The issue of persistent underachievement for certain segments of the student population has been of interest to educational researchers ; however environmental factors have not been traumatic stress symptoms and negative attachment symptoms with their effect on the outco me trajectory of reading and mathematics academic achievement from kindergarten to fifth grade. The study controlled for the effects of gender, culture, and socioeconomic status (SES) on these relationships. My research questions were: (1) What is the rela tionship between the growth trajectory of reading and mathematics achievement and traumatic stress symptoms for children in kindergarten through fifth grade?, (2) What is the relationship between the growth trajectory of reading and mathematics achievement and difficult attachment symptoms for children in kindergarten through fifth grade?, (3) What is the relationship between the growth trajectory of reading and mathematics achievement and traumatic stress symptoms and difficult attachment symptoms for chil dren in kindergarten through fifth grade?, (4) What is the relationship between the growth trajectory of reading and mathematics achievement and traumatic stress symptoms and difficult attachment symptoms for children in kindergarten through fifth grade, w hile controlling for gender, culture, and socioeconomic status? and finally, (5) What is the relationship of an interaction effect between traumatic stress symptoms and difficult attachment symptoms on the growth trajectory of reading and mathematics achi evement for children in kindergarten through fifth grade, while controlling for gender, culture, and socioeconomic status? This quantitative study used the national database Early Childhood Longitudinal Study Kindergarten 1998 cohort (ECLS K) collected by the National Center for Education Statistics (NCES) in order to design a latent growth curve model that incorporated traumatic stress PAGE 83 83 symptoms and difficult attachment symptoms and their effects on the growth trajectory. The purpose of this chapter is to p resent the results from analyzing the research questions using the methods described previously. Participant Multivariate Descriptive Information using full information m aximum likelihood ( FIML ) estimation The covariance coverage of the proportion of data present in the reading achievement model ranged from 88.2% to 100% present ( Table 4 1 ). The covariance coverage of the proportion of data pre sent in the mathematics achievement model ranged from 91.6% to 100% present ( Table 4 2 ). This weighted longitudinal sample consisted of 51.5% males ( ) and 48.5% females ( ) The distribution of the responding parent/guardian was 6 2 % male ( ) and 9 3 8 % female ( ) The distribution of the cultural groups in this sample was 57. 7 % European American ( ) 18. 8 % Latina/o ( ) 16% African American ( ) 2.7% Asian ( ) and 4.8% other people of color (Native American, Alaska Native, Native Hawaiian, Other Pacific Islander, a nd Multi Racial) ( ) D ue to low percentage of the sam ple number the cultural group other people of color means were for the most part found to be not significantly different from the reference group, European American, when answering research question s four and five for the reading and mathematics LGCM. Th erefore it was not included in the discussion of the results, but was allowed to remain in the model as a controlled for dummy coded variable. However, just to note the cultural group other people of color did significantly differ from the mean quadratic s lope factor when answering research questions four and five for the mathematics LGCM, suggest ing PAGE 84 84 that children from th is cultural group show lower negative quadratic trends in their mathematics achievement trajectory than European American children ( Table 4 3 ) Likewise, children whose responding parent/guardian w ere male w ere found to be not significantly different from the reference sample of children whose responding parent/guardian was female for research questions four and five for the reading and mathematics LGCM Therefore it was also not included in the discussion of the results but was allowed to remain in the model as a controlled for dummy coded variable ( Table 4 4 ) In review of the aca demic achievement measures for the students in the sample the observed and estimated means for reading achievement, their standard deviations, and correlation matrix are shown in Table 4 5 The observed and estimated means for mathematics achievement, their standard deviations, and correlation matrix are shown in Table 4 6 These two correlation matrix tables show that these items are significantly correlated and therefore measure similar constructs o ver time. By l ooking at the histogram/densit y plots of the reading ( Figure 4 1 ) and mathematics ( Figure 4 2 ) achievement variables and the plot of individuals observed longitudinal tr ajectories for reading ( Figure 4 3 ) and mathematics ( Figure 4 4 ) data (Grimm & Ram, 201 2 ) these figures provide support for the data to be following nonlinear trends (Grimm, Steele, Mashburn, Burchinal, & Pianta, 2010). This can be seen as the distribution started out positively skewed in time periods one and two. The trend in the data from third grade to fifth grade showed possible platykurtic or flattening of the distribution and some potential multimodal aspect s (Huck, 2008). This distribution reflected the need for using MLR estimation methods in M plus as it has been shown to deal well with nonnormal distribution of the data (Muthn & Muthn, 2009 March 23 ; 2010; Yuan & Bentler, 2000). PAGE 85 85 Specification of Measure ment Model In order to ensure that estimates of the latent growth curve model (LGCM) w ere accurate it was necessary to determine which measurement model was the best fit. I began with the linear and its quadratic LGCM, then the freed loading and its quad ratic LGCM, next lo oked at the quadratic orthogonal polynomial model, and followed by the quadratic and cubic orthogonal polynomial of the hierarchical multivariate linear model ( HM LM ) model. The cubic orthogonal polynomial HMLM model was finally chosen as providing the most accurate estimates for the reading and mathematics achievement data. The Latent Growth Curve Model I began this process by analyzing the fit of the unconditional LGCM. Time was centered at initial assessment and the following E quation 4 1 was used for the linear model: ( 4 1) The linear model I began with the linear model to the reading achievement and mathematics data. The results for the reading data showed that it did not fit the data well ( ), ). The mathematics data showed similarly poor fit to the linear model ( ), PAGE 86 86 The quadratic trend to this model did not converge due to a not positive definite residual covariance matrix for both the reading and mathematics data. Given the poor fit of the linear model, I then examined the utility of the free loading model. The free loading model The free loading LGCM was examined next for the reading an d mathematics data, with the following E quation 4 2 : ( 4 2) The free loading for reading ( ), ), , and mathematics ( showed a large improvement in fit, but still showed problems ade quately estimating the data The largely significant LRT showed the free loading model was significantly different and its lower AIC and BIC values suggested better fit, however, the fit indices suggested a need for model improvement. The poor fit of the f reed loading model required that the utility of the quadratic free loading model be looked at next PAGE 87 87 The quadratic free loading model The estimated factor scores from the non polynomial model were specified and the quadratic factor was added T he followin g specifications are shown for the reading free loading model with quadratic factor in E quation 4 3 : ( 4 3) which did not converge because it did not have a positive definite residual covariance matrix. The mathematics free l oading model with quadratic factor had the following specifications in its E quation 4 4 : ( 4 4) and this did converge, showing large improvement on the model that estimates an adequate fit ( ), ), , however when the TICs were added it wa s not able to converge and so was not used for further analysis. Given the troubles converging the quadratic freed loading model was rejected, I then examined the utility of the quadratic orthogonal polynomial model. PAGE 88 88 The quadratic orthogonal polynomial mod el A quadratic orthogonal polynomial model for the reading and mathematics data was analyzed next using the following E quation 4 5 : ( 4 5) This did not converge because it had a not positive definite residual covariance matrix. This led to the conclusion that the issues of convergence were due to the amount of missing data and thus the HMLM model by Raudenbush and Byrk (2002) was examined as it was designed to deal with missing data issues specifically. The HMLM quadratic orthog onal polynomial model This model is accomplished in M plus by (a) setting the variances of all trend latent factors equal to zero, and (b) allowing the residuals for all observed reading and mathematics variables to correlate ( Raudenbush & Bryk, 2002) The following E quation 4 6 was used for this model: ( 4 6) In the HMLM quadratic orthogonal polynomial the trend changed as this model was able to converge, suggesting that this model is able to address the missing data issue. The HMLM PAGE 89 89 quadratic orthogonal polynomial model fit the data better than previous models for reading data ) when compared to the free loading model, ), The mathematics HMLM quadratic orthogonal polynomial model approximated the parameters better than previous models for the mathematics data when compared to the free loading model, However, RMSEA was slightly above the .06 cutoff when the confidence interval is accounted for and it being th e more important fit index, the HMLM cubic orthogonal polynomial model was explored. The HMLM cubic orthogonal polynomial model The E quation 4 7 for the HMLM cubic orthogonal polynomial model is as follows: ( 4 7) The HMLM cubic orthogonal polynomial model had the best fit to the data for both the reading LGCM ( ), ), , and the mathematics LGCM PAGE 90 90 This was determined by looking at the fit indices, the L RT, and the AIC and BIC values of all models tried. The LRT showed that the HMLM cubic orthogonal polynomial model contributed additional information that causes a significant difference with the HMLM quadratic polynomial model [reading: and mathematics : The AIC and BIC values were lowest suggesting that the HMLM cubic orthogonal polynomial had the best fit of all models for reading and mathematics data. Thus, the HMLM cubic orthogonal polynomial model was selected to test the research questions and hypothesis. The correlation matrix for the reading achievement endogenous variables, their estimated means by the model, and the observed means with standard deviations are presented in Table 4 5 The correlation matrix for the mathematics achievement endogenous variables, their estimated means by the model, and the observed means with standard deviations are presented in Table 4 6 A comparison of the d ifferent model estimated means for reading ( Table 4 7 ) and mathematics ( Table 4 8 ) achievement were compared and it was verified that the HMLM cubic polynomial model had the more stable estimates. C onfirmatory Factor Analysis The process of developing the measurement specifications necessary to proceed with structural analysis was developed in two phases. The first phase consisted of applying EFA procedures with SEM based tests of model fit to the difficult attachment symptoms indicators No EFA procedures were conducted with traumatic stress symptoms because the factor structure had already been theoretically developed and prior research supported its structure. For difficult attachment symptoms a selection of the appropriate items w as made and th en they were specified to the single factor without any cross loadings known as confirmatory factor analysis (CFA) The second phase consisted of analyzing the measurement model for the latent variables, PAGE 91 91 t raumatic stress symptoms and difficult attachment symptoms fitness to the cubic orthogonal polynomial LGCM. The following are the results of these phases. Difficult A ttachment S ymptoms EFA and CFA D evelopment I included 11 items based on the literatur e and the personal communication with Dr. Gail Mulligan (Mulligan, personal communication, October, 15 2012) for an exploratory factor analysis (EFA). The oblique geomin rotation was used and the estimations were constrained to 1 to 4 factors. A 3 factor m odel was chosen as the best fitting result. Its showed that this was not an exact fitting model, however the fit indices did show that it was adequate enough not to reject Of the 11 items, I dropped 6 items from subsequent analyses due to low loading (< 0.40) on the factor of interest and high loading on 2 other factors ( Table 4 9 ) The resulting indicators (P2HRDWRM, P2 CHDOES, P2FLTRAP, P2FELAN, and P2CHHARD) were fit to a CFA analysis. The model was run without specifying the indicators as categorical, causing M plus to assume them as continuous, thus providing a baseline comparison, since logit regression does not provi de indices of fit. This resulted in the following fit test and indices, ). The model was then rerun using logit regression by specifying its indicators as ordinal categorical, providing the following resu lts ( ). As can be seen by the non significant likelihood chi square test for ordered categorical outcomes and the improved lower AIC and BIC values these indicators appear to fit the laten t factor adequately. The descriptive statistics for parents' difficult attachment symptoms indicate that the majority of caregivers reported "not at all true" for the five symptoms: P2HARDWRM ("hard to be warm"), P2CHDOES ( PAGE 92 92 things that really bo ), P2FLTRAP ( P2FEELAN ( ), and P2CHHARD ( ). See Table 4 10 for the descriptive statistics for difficult attachment symptoms When looking at the plots of the total information and characteristics curves for the indicators of difficult attachment symptoms four out of the five indicators are stable and one of the indicators P2HRDWRM is not as stable ( Fi gure 4 5 and Figure 4 6 ) Traumatic S tress S ymptoms CFA D evelopment The model of indicators developed from prior research on traumatic stress with this data set led to the following indicators to be specified (T2CONTRO, T2INTE RP, T2EXTERN, T2INTERN, and P1ATTENI) to load on the latent variable traumatic stress symptoms. Again the model was run first without specifying the indicator P1ATTENI as categorical to provide a baseline comparison. The resulting fit test and indices ( ) show that the model is a good fit. The fact that TLI is below the current standard of 0.95 can be attributed t o P1ATTENI not being specified as categorical. This model was rerun specifying P1ATTENI as an ordinal categorical indicator which provided an showing a relatively large increase in model fit. Thus, these indicators were accepted as fitting this factor adequately. The descriptive statistics of traumatic stress symptoms suggested that teachers rated most children as showing average behaviors, and a small proportion showed signs of difficulty ( Ta ble 4 1 1 ) Given the shape of the total information curve of traumatic stress symptoms it is hypothesized that it is a bimodal distribution ( Figure 4 7 ) T he plot of the characteristic curve for the P1ATTENI indicator as a PAGE 93 93 fu nction of traumatic stress symptoms suggests that this indicator is not as stable in it s distribution ( Figure 4 8 ) The final step in the latent variables measurement model involved estimating how they fit together. Again I ran traumatic stress symptoms and difficult attachment symptoms together as continuous variables to estimate a baseline comparison. The fit test and indices were as follows ( ), showing the combined model was a great fit. When specified to include the ordinal categorical indicators it r esulted in these values ( ). The likelihood chi square test for the binary and ordered categorical outcomes was non significant and the AIC and BIC showed a considerable model improvement providing support that this model is sufficient. The factor loadings for the measurement model that includes traumatic stress symptoms and difficult attachment symptoms shows discrepancies in teacher ability to identify internalizing behaviors (T2INTERN) parents had trouble discriminating (P1ATTENI), and parents possibly showed trouble disclosing when they were having trouble being warm with their children (P2HRDWRM) ( Table 4 1 2 for the factor loadings, standard errors, factor correlations, and estimates ). Andreassen and West (2007) and Chapin (2006) have found that both teacher and parent self report data has been skewed toward socially desirable answers. This suggests that some of the t roubles with the indicators could be point ing to social desirability biases or even cultural differences in interpreting the meaning of socially desirable questions. PAGE 94 94 The Structural LGCM Th e unconditional cubic orthogonal polynomial model for reading and math ematics provided the best description for the growth model. This model estimated the values for the mean intercept, mean linear factor, mean quadratic factor, and the mean cubic factor for the trajectory for all children in the sample from kindergarten to fifth grade. However, because the general orthogonal polynomial coefficients are rescaled measures of the time scores, both will be provided. Mean Values for Structural LGCM The non orthogonalized values outside of the parenthesis and the orthogonaliz ed values are presented inside the parentheses. The reading achievement data observed the initial value of mean reading achievement to be with a ( which increased linearly at a mean rate of with a units per four months increase in time ( Figure 4 9 ; T his increase in slope on average was de celerating at a rate of unit per units per four months increase in time with a ( Figure 4 10 ; However, looking at this plot one can see a hug e acceleration from Fall of kindergarten to Spring of kindergarten, which then results in a continuing deceleration till fifth grade. This average quadratic deceleration is also decelerati ng ( attenuat ed ) or reduced over time at an average rate of wit h a ( Figure 4 11 ; Thus, the mean intercept, mean slope, mean quadratic slope, and mean cubic slope differed significantly from zero These trends were graphed to give a v isual perspective of these nonlinear trends. Looking at the mathematics achievement data, the mean initial value of mathematics achievement was equal to with a ( which PAGE 95 95 increased linearly at a mean r ate of units per four months increase in time with a ( Figure 4 12 ; This increase in slope was decelerat ing at a n average rate of unit per units per four months in crease in time with a ( Figure 4 13 ; L ooking at this plot, one can see the huge acceleration from Fall kindergarten to Spring kindergarten which results in a continuing dec eleration until fifth grade. This average quadratic deceleration is also decelerati ng ( attenuated ) or reduced over time at an average rate of with a ( Figure 4 14 ; The mea n intercept, mean slope, mean quadratic slope, and mean cubic slope of mathematics achievement differed significantly from zero These trends in mathematics were graphed to give a visual perspective of these nonlinear trends. Because variance of the intercept, slope, quadratic, and cubic factors were set equal to zero, individual differences were not estimated for this model. This was acceptable, because the purpose of this study was to look at how traumatic stress symptoms and difficult attachmen t mathematics achievement from Fall of kindergarten to Spring of fifth grade. Therefore, it was not important to estimate the individual differences, particularly since the missing data made models that estimat ed these parameters not possible. The Conditional LGCM The values of the intercept slope factor, linear slope factor, quadratic slope factor, and cubic slope factor will be based on the rescaled time scale for the results pertaining to the conditional LGC M and the research questions. In this next part of the analysis the conditional LGCM was examined using Equation 4 8 and Equation 4 9 The amount of time invariant PAGE 96 96 covariates (TICs) included in the models differed for each research question. The full model is described by E quation 4 10 The level 1 equation is: ( 4 8) The level two equation in its full form is: ( 4 9) The combined model is: (4 10) Reading and Mathematics Growt h Trajectory and TSS The first research question asked: What is the relationship between the growth trajectory of reading and mathematics achievement and traumatic stress symptoms for children in kindergarten through fifth grade? This question was examined with this nested level 2 model using Equation 4 11 : PAGE 97 97 ( 4 1 1 ) where represents the time invariant covariate (TIC): traumatic stress symptoms (TSS). The following are the mean estimates for reading achievement for childr en who are low on TSS: mean intercept = mean linear slope = mean quadratic slope = and the mean cubic slope = The following are the mean estimates for mathematics achiev ement for children who are low on TSS: mean intercept slope factor = mean linear slope = mean quadratic slope = and the mean cubic slope = The latent variable TSS has bot h continuous and ordinal categorical data, which requires numerical integration, so the normal test and fit indices are not available. Thus, likelihood based indexes were used to make model comparisons ( Grimm & Ram, 2009).These included the loglikel ihood been estimated, its loglikelihood, AIC, and BIC values were used as a compariso n due to this model being nested within the models being examined for the research questions. The path model for the structural equation of reading achievement with the TIC TSS can be seen in Figure 4 15 and mathematics achie vement in Figure 4 16 The reading data results of the LRT, AIC, and BIC ) suggest that the addition of TSS to the model contribute d additional information t hat causes a significant difference, however this causes an increase in complexity such that the PAGE 98 98 AIC and BIC are quite a bit larger. This is also the case for the mathematics data ) wher e the addition of TSS causes a significant difference, yet increases the complexity with larger AIC and BIC values. The AIC and BIC suggest that the addition of TSS does not fit as well, however, there is not a measure to determine if the fit is still adeq uate. Therefore, the stability of the model was also checked by examining the estimated means, residual variances, and the estimates from standardized output in M plus for both reading and mathematics data The estimated means showed little difference to the cubic orthogonal polynomial model estimated means for reading and mathematics the residual variances were signific ant and positive for both, and the showed no change from Table 4 1 2 that is based on the CFA for TSS and DAS; thus supporting the models fit adequately. Reading Achievement LGCM With One TIC In the reading LGCM, TSS sign ificantly predicted a negative intercept factor ( ), negative linear slope factor ( ), and a positive quadratic slope factor ( ) ( Figure 4 15 ) The estimates for TSS did not show a significant difference from the mean cubic slope ( ). Indicating that, on average, children high in TSS are observed to have lower initial reading achievement scores with a lower rate of change over time suggesting that a one unit increase in TSS is associated with a 7.31 decrease in the slope of the trajectory of reading achievement. TSS also predicted a slower deceleration for child ren high in TSS as c ompared to children low in TSS. PAGE 99 99 Mathematics Achievement LGCM W ith One TIC TSS for the mathematics achievement, significantly predicted a negative intercept factor ( ), a negative linear slope factor ( ), a positive quadratic slope factor ( ), and a negative cubic slope factor ( ) ( Figure 4 16 ) This suggest that, on average, children high in TSS are observed to have lower initial mathematics achievement scores with lower rates of change that indicate a one unit increase in TSS is associated with a 5.58 decrease in the slope of the trajectory of mathematics achievement over time. The positive quadratic slope factor attenuates the mean quadratic cup down trend, thus slowing the plateau effect for children high in TSS. The neg ative cubic slope factor is reducing the mean cubic slopes amplification of the quadratics decelerating effect on the growth rate over time. Based on the results of the first statistical analysis, there is a significant relationship between traumatic stres s symptoms and reading and mathematics achievement. Children who are reported as having higher levels of traumatic stress also have lower levels of initial performance and performance over time. These findings support prior research in this area (Goodman e t al., 2012) and go further to suggest that these lower levels of achievement can persist across grade levels and developmental trajectories Reading and Mathematics Growth Trajectory and DAS The second research question explored: What is the relationship between the growth trajectory of reading and mathematics achievement and difficult attachment symptoms for children in kindergarten through fifth grade? This question was examined with th e nested level 2 model in Equation 4 12 : PAGE 100 100 ( 4 1 2 ) where represents the time invariant covariate (TIC): difficult attachment symptoms (DAS). The following are the mean estimates for reading achievement for children who are low on DAS: mean intercept slope factor = mean linear slope = mean quadratic slope = and the mean cubic slope = The following are the mean estimates for mathematics achievement for children who are low on DAS: mean intercept slope factor = mean linear slope = mean quadratic slope = and the mean cubic slope = The path model for the structural equation of reading achievement with the TIC DAS can be seen in Figure 4 17 and mathematics achievement in Figure 4 18 The reading data results of the LRT, AIC, and BIC ) show again that the addition of a TIC, in this case DAS, to the model contributed additional information that caused a significant difference. More complexity resulted and the AIC and BIC values were larger. The mathematics data ) also showed that the addition of DAS caused a significant difference, with the expected increase in the AIC and BIC values. The estimated means maintained accuracy for both reading and mathematics the residual variances were PAGE 101 101 significant and po sitive for both, and the showed no change from Table 4 1 2 thus supporting the models adequacy of fit. Reading Achievement LGCM With One TIC In the reading LGCM, DAS significantly predicted a negative intercept factor ( ), a negative linear slope factor ( ), and a positive quadratic slope factor ( ) ( Figure 4 17 ) The cubic slope factor was n ot found to be significantly different from the mean cubic slope ( ). Thus, on average, children high in DAS are observed to have lower initial reading achievement scores and lower rates of change over time suggest ing that a one unit increase in DAS is associated with a 4.34 decrease in the slope of the trajectory of reading achievement. The concave upward trend of the quadratic slope factor suggest s a slowing of the decelerating trend of the quadratic slope for chi ldren high in DAS. Mathematics Achievement LGCM With One TIC DAS for the mathematics achievement significantly predicted a negative intercept factor ( ), negative linear slope factor ( ), and a positive quadratic slope factor ( ( Figure 4 18 ). DAS di d not estimate a significantly different cubic slope factor from the mean cubic slope ). Indicating that, on average, children high in DAS are observed to have lower initial mathematics achievement scores and a lower growth trend over time that for a one unit increase in DAS is associated with a 2.64 decrease in the slope of the trajectory of mathematics achievement. The positive quadratic slope factor indicated an attenuation of the decelerating effects of the m ean quadratic slope, thus slowing the plateau e ffect for children high in DAS. PAGE 102 102 Based on the results of the second statistical analysis, there is a significant relationship between difficult attachment symptoms and reading and mathematics achievement initia lly and over time. Children whose parents reported as having higher levels of difficult attachment symptoms also have lower levels of initial performance and lower performance over time. These findings support p rior research in this area (Bergin & Bergin, 2009; Pianta & Harbers, 1996 ) and go es further to suggest that these higher levels of difficult attachment symptoms start before s e con d grade, rather in this study that they at least begin in kindergarten, and lead to lower levels of achievement that can p ersist at least until fifth grade. Reading and Mathematics Growth Trajectory and TSS and DAS The third research question explores: What is the relationship between the growth trajectory of reading and mathematics achievement and traumatic stress symptoms a nd difficult attachment symptoms for children in kindergarten through fifth grade? This question was examined with th e nested level 2 model using Equation 4 13 : ( 4 1 3 ) where represents the time invariant covariate (TIC): traumatic stress symptoms (TSS) and represents the time invariant covariate (TIC): difficult attachment symptoms (DAS). The following are the mean estimates for children who are low on TSS and DAS: mean intercept slope factor = mean linear slope = mean quadratic slope = and the mean cubic slope = The following are the mean estimates for mathematics achievemen t for children who are low on TSS and DAS: PAGE 103 103 mean intercept slope factor = mean linear slope = mean quadratic slope = and the mean cubic slope = The path model for the stru ctural equation of reading achievement with the TICs TSS and DAS can be seen in Figure 4 19 and mathematics achievement in Figure 4 20 For the reading data, results of the LRT, AIC, and BIC when compared to the cubic LGCM ) showed again that the addition of the TICs TSS and DAS together contributed significant additional information, with expected complexity resulting in larger AIC and BIC values. The mat hematics data ) also showed this same trend. The estimated means maintained accuracy for both reading and mathematics the residual variances were significant and positive for both, and t he stayed the same for both TSS and DAS, all together supporting model fit. Reading Achievemen t LGCM W ith Two TICs TIC1: TSS and DAS. In the reading LGCM when controlling for DAS, TSS significantly predicted a negative intercept factor ( ), a negative linear slope factor ( ), and a positive quadratic slope factor ( ( Figure 4 19 ). slope factor was not found to be significantly different from the mean cubic slope ( Indicating that children high in TSS have a lower initial reading score and a lower rate of growth over time that suggested a one unit increase in TSS was associated with a 5.90 decrease in the slope of the trajectory of reading achievement. The positive quadratic slope factor attenuates the mean negative quadratic slope. PAGE 104 104 TIC2: DAS and TSS When controlling for TSS, DAS had a sig nificant negative intercept factor ( ), negative linear slope factor ( ), and positive quadratic slope factor ( ) ( Figure 4 19 ). DAS did n ot significantly predict a cubic slope factor that was different from the mean cubic slope ( ). These results indicate that when controlling for TSS, DAS predicted a lower initial reading achievement score with a l ower rate of growth over time showing that for every one unit increase in DAS resulted in a decrease of 3.62 units in the slope of the trajectory. The quadratic slope factor is also estimated to be reducing the effects of the mean negative quadratic slope, which results in children high in DAS not leveling off in the growth trend as fast as children low in DAS. Mathematics Achievement LGCM With Two TICs TIC1: TSS and DAS. When controlling for DAS for the mathematics data, TSS significantly predicted a nega tive intercept factor ( ), negative linear slope factor ( ), positive quadratic slope factor ( and a negative cubic slope fac tor ( ) ( Figure 4 20 ) Indicating that, on average, children high in TSS are observed to have lower initial mathematics achievement scores and lower rates of change over time such tha t a one unit increase in TSS was associated with a 4.80 decrease in the slope of the trajectory of mathematics achievement. The positive quadratic slope factor showed an attenuation of the mean quadratic slope thus resulting in children high in TSS not hav ing their growth rate slowed down as fast. The negative cubic slope factor is also slowing the mean cubic slopes trend of speeding up of this slowing down of the mathematics achievement growth rate PAGE 105 105 TIC2: DAS and TSS. DAS for the mathematics data significa ntly predicted a negative intercept factor ( ), negative linear slope factor ( ), positive quadratic slope factor ( while cont rolling for TSS ( Figure 4 20 ) DAS did not predict a significant difference from the mean cubic slope ( ). This suggests that while controlling for TSS, children high in DAS have a lo wer initial mathematics score with a downward trajectory over time that for every one unit increase in DAS is associated with a 1.99 decrease in the slope of mathematics achievement trajectory. The positive quadratic slope factor is attenuating the mean qu adratic slope, resulting i n a flatter growth trajectory. Based on the results of the third statistical analysis, there is a significant relationship between traumatic stress symptoms and difficult attachment symptoms with reading and mathematics achievemen t. Children who are reported as having higher levels of traumatic stress and difficult attachment symptoms also have lower levels of initial performance and performance over time. These findings support prior research that has looked at both traumatic stre ss and attachment behaviors (Bailey et al. 2007; De Young et al., 2011a, 2011b; Lieberman, 2004; Lieberman & Van Horn, 2008; Mongillo et al. 2009 ) suggesting difficulties, but not specifically looking at academic achievement. study connected the effects of traumatic stress to reading and mathematics achievement at the cross sectional fifth grade level. The findings of this study go further to suggest that traumatic stress symptoms and difficult attachment symptoms together pre dict lower levels of achievement and that these lower levels can persist across grade levels as well as developmental trajectories. PAGE 106 106 Reading and Mathematics Growth Trajectory and TSS, DAS, Gender, SES, and Culture The fo u rth research question explores: Wha t is the relationship between the growth trajectory of reading and mathematics achievement and traumatic stress symptoms and difficult attachment symptoms for children in kindergarten through fifth grade, while controlling for gender, culture, and socioeco nomic status? This question was examined with th e level 2 model in Equation 4 14 : ( 4 1 4 ) where is the time invariant covariate ( TIC ) : traumatic stress symptoms (TSS), is the TIC : difficult attachment symptoms (DAS), is the TIC gender, represents the TIC gender of interviewed parent/caregiver, represents th e TIC SES, the next four variables represent dummy codes of the variable culture: represents the TIC Asian American culture, represents the TIC African American culture, represents the TIC Latina/o culture represents th e TIC other people of color culture However, as mentioned earlier, TIC and were found to be insignificant in the model and ar e not discussed in the results. The following are the mean estimates for children from average SES families who se interviewed caregiver was fe male, the child is male, European American and the children are low on TSS and DAS: mean intercept slope factor = mean linear slope = mean quadratic slope = a nd the mean cubic slope = The following are the mean estimates for mathematics achievement for children from average SES, whose interviewed caregiver was female, the child is male, European PAGE 107 107 American and the children are low on TSS and DAS: mean intercept slope factor = mean linear slope = mean quadratic slope = and the mean cubic slope = For the reading data, results of the LRT, AIC, and BIC when comp ared to the cubic LGCM with TSS and DAS as TICs ) showed that the addition of the TICs gender, SES, and culture to the reading HMLM cubic orthogonal polynomial model with the TICs TSS and DAS contributed significant additional information. Interestingly the addition of these four TICs resulted in a reduction in the AIC and BIC values. This suggests that gender, SES, and culture have significant information that they add to this model causin g it to fit the data better than TSS and DAS alone. The results for the mathematics data ) showed similar results to the trends displayed by the reading data. The addition of the TICs gender, SES, and culture to the mathematics HMLM cubic orthogonal polynomial model with the TICs TSS and DAS contributed significant additional information. Again there was a reduction in the AIC and BIC values. This suggests that gender, SES, and culture have significant information for the mathematics model as well. The estimated means maintain ed accuracy for both reading and mathematics the residual variances were significant and positive for both, and the stayed similar for both TSS and DAS, all together supporting model fit. Reading Achievement LGCM With S even TICs TIC1: TSS and DAS, Gender, SES, Culture. For the reading achievement trajectory data, TSS was examined first while controlling for DAS, gender, SES, and culture. TSS significantly predicted a negative intercept factor ( ), a PAGE 108 1 08 negative linear slope factor ( ), and a positive quadratic slope factor ( ). The cubic slope factor was found not be significantly different from the mean cubic slope ( ). Indicating that, on average, children high in TSS are observed to have lower in itial reading achievement scores with a downward rate of change that reflects with a one unit increase in TSS this was associated with a 4 01 decrease in the slope of the trajectory of reading achievement. The concave up quadratic slope factor acts against the mean concave down quadratic slope, resulting in a flattenin g of the growth rate over time. TIC2: DAS and TSS, Gender, SES, Culture. DAS was examined next controlling for TSS, gender, SES, and culture. DAS significantly predicted a negative intercept f actor ( ) and a negative linear slope factor ( ). It showed that the quadratic slope factor ( ) and the cubic slope factor ( ) are not estimated to be significantly different from the mean quadratic and cubic slopes. This suggests that when all other TICs are controlled for, on average, children high in DAS have lower initial reading achievem ent scores and show a downward rate of change over time. Thus, with a one unit change in DAS there is an associated 2. 44 decrease in the slope tra jectory of reading achievement. TIC3: Gender and TSS, DAS, SES, Culture. The results of the analysis for gende r controlling for TSS, DAS, SES, and culture were looked at next. Gender showed a significant positive intercept factor ( ), a positive linear slope factor ( ), and a negative quadratic slope factor ( ). However, gender did not show a significant diffe rence from the mean PAGE 109 109 cubic slope ( ). This indicates that females, on average, show higher reading achievement scores at initial assessment and have an increasing rate of change in reading achievement over time. Th e concave downward quadratic slope factor accentuates the already negative quadratic slope, resulting in females have a faster reduction in the rate of growth over time. So even though females start out higher in scores and have a higher rate of growth ove r time, their growth rate is decelerating at a faster rate. Where males start out lower in reading achievement, have lower rates of change, but their growth rate does not decelerate as fast. TIC4: SES and TSS, DAS, Gender, Culture. Examining Socioeconomic status (SES) while controlling for TSS, DAS, gender, and culture significantly predicted a positive intercept factor ( ), a positive linear slope factor ( ), and a negativ e quadratic slope factor ( ). SES did not predict a significantly different cubic slope factor from the mean cubic slope ( ) This estimation predicts that on average, c hildren whose families are high in SES are observed to have the highest initial reading achievement scores with a large positive rate of change over time respectively. This predicts that a one unit increase in SES is associated with a 7.3 3 unit increase in the slope of the trajectory of reading achievement. Those high in SES also show a larger concave down quadratic trend resulting in a greater decelerating growth rate over time. This suggests that reading achievement plateaus for high SES children as they head toward third and fifth grade. TIC5: Asian American Culture and TSS, DAS, Gender, SES, AA, Latina/o culture. The relationship of A sian American culture on reading achievement while controlling for TSS, DAS, gender, SES, AA, and Latina/o culture showed significant predictions for a negative linear PAGE 110 110 slope factor ( ) and cubic slope factor ( ) The intercept factor ( ) and the quadratic slope factor ( ) were found to not be significantly different from their respective mean slopes. These results show that children of A sian American culture, on average, for reading achievement show a lower growth rate over time as compared to children of European Americ an and Latina/o cultures and that their cubic factor is maintaining the quadratics deceleration TIC 6: AA Culture and TSS, DAS, Gender, SES, Asian American, Latina/o culture. The relationship of African American culture on reading achievement while control ling for TSS, DAS, gender, SES, Asian American, and Latina/o culture showed significant predictions for a negative intercept factor ( ) and negative linear slope factor ( ), however, the quadratic slope factor ( ) and cubic slope factor ( ) were found to not be different from their respective mean slopes. Importantly, African American cult ures negative intercept slope factor was almost as large as children who were high on TSS. These results suggest that children of African American culture, on average, begin kindergarten with lower initial mean reading achievement scores while showing a lo wer growth rate over time as compared to children of European American Asian American, and Latina/o cultures. TIC 7 : Latina/o c ulture and TSS, DAS, Gender, SES, A A culture. The relationship of Latina/o culture on reading achievement while controlling for T SS, DAS, gender, SES, Asian American, and African American culture showed a significant negative intercept factor ( ), but was found to not have significant differences in the mean linear slope ( ), mean quadratic slope ( PAGE 111 111 ), or the mean cubic slope ( ). Therefore these estimates suggest that, on average, children from Latina/o cultures only show differences in their initial reading achievement scores, which are lower as compared to European American children. Based on the results of the fourth statistical analysis, there is a significant relationship between traumatic stress symptoms difficu lt attachment symptoms gender, SES, Asian American culture, AA culture, and Latina/o culture with reading achievement. Children who are reported as having higher levels of traumatic stress difficult attachment symptoms are male, low in SES, and Asian Am erican, African American or Latina/o also have lower levels of initial performance and /or performance over time. These findings support prior research that has looked at traumatic stress (Goodman et al., 2012) difficult attachment symptoms ( L ieberman, 20 04; Pianta & Harbers, 1996 ) gender ( Clark et al., 2008; Dearing et al., 2009 ) low in SES ( Chatterji, 2006 ) and are either African American or Latina/o (Crumpton & Gregory, 2011; Morris, 2008) as indicators for lower reading achievement The findings of this study go further to suggest that all seven covariates predict lower levels of reading achievement and that these lower levels can persist across grade levels from kindergarten through at least fifth grade as well as developmental trajectories. Mathema tics Achievement LGCM With Seven TICs TIC1: TSS and DAS, Gender, SES, Culture. For the mathematics achievement trajectory data, TSS was examined while controlling for DAS, gender, SES, and culture. TSS significantly predicted a negative intercept factor ( ), a negative linear slope factor ( ), a positive quadratic slope factor ( ), and a negative cubic slope factor ( PAGE 112 112 ). Indicating that, on average, children high in TSS are observed to have lower initial mathematics achievement scores and a lower rate of change over time that suggested with a one unit increase in TSS there is an associated 4. 13 decrease in the slope of the trajectory of mathematics achievement. The positive quadratic slope factor attenuates the effects of the negative mean quadratic slope thus reducing the deceleration of the growth trajectory over time. While the negative cubic slope factor also is working against the mean positive cubic slopes effects, thus reducing its amplification of the cup down quadratic trend. TIC2: DAS and TSS, Gender, SES, Culture. DAS was examined next controlling for TSS, gender, SES, and culture DAS significantly predicted a negative intercept factor ( ). DAS did not show significantly different results from the mean linear slope ( ), mean quadratic slope ( ), or the mean cubic slope ( ). This suggests that when all other TICs are controlled for, on average, children high in DAS had lower initial mathematics achievement scores, b ut similar rates of change (linear slope and other polynomial slo pes) to children low on DAS. TIC3 : Gender and TSS, DAS, SES, Culture. The results of the analysis for gender controlling for TSS, DAS, SES, and culture were looked at next. Gender showed a si gnificant negative intercept factor ( ), a negative linear slope factor ( ), and a positive quadratic slope factor ( ). However, gender did not show a significant diff erence from the mean cubic slope ( ). This indicates that females, on average, show lower mathematics achievement scores at initial assessment and have a lower rate of change in mathematics achievement over time. T he concave upward quadratic slope factor works PAGE 113 113 against the negative mean quadratic slope, resulting in a slight reduction in the deceleration trend over time. TIC4: SES and TSS, DAS, Gender, Culture. Examining Socioeconomic status (SES) while controlling f or TSS, DAS, gender, and culture significantly predicted a positive intercept factor ( ), a positive linear slope factor ( ), and a negative quadratic slope factor ( ). SES did not predict a significant difference fr om the mean cubic slope ( ). This estimation predicted that on average, children whose families are high in SES are observed to have higher initial mathematics achievement scores with higher rate of change over ti me. This suggests that a one unit increase in SES is associated with a 6 20 increase in the slope of the trajectory of mathematics achievement. Those high in SES also show a concave down quadratic slope factor that accentuates the mean cup down quadratic s lope, causing children high in SES to have a greater decelerating rate of change over time. This suggests that mathematics achievement may have a pla teau or ceiling effect as well. TIC5: Asian American Culture and TSS, DAS, Gender, SES, AA, Latina/o cultur e. The relationship of Asian American culture on mathematics achievement while controlling for TSS, DAS, gender, SES, AA, and Latina/o culture showed significant predictions for a positive linear slope factor ( ) a positive quadratic slope factor ( ) and a negative cubic slope factor ( ). The intercept factor ( ) was found to not be signifi cantly different from its respective mean slopes. These results show that children of Asian American culture, on average, for mathematics achievement show a higher growth rate over time as compared to children of European American African American, and La tina/o PAGE 114 114 cultures T he positive or cup up quadratic slope factor suggests that this higher growth rate is accelerating and the negative cubic slope factor shows this acceleration is reducing over time TIC 6 : AA Culture and TSS, DAS, Gender, SES, Latina/o cul ture. The relationship of African American culture on mathematics achievement while controlling for TSS, DAS, gender, SES, Asian American, and Latina/o culture showed significant predictions for a negative intercept factor ( ), negative linear slope factor ( ), and a positive quadratic slope factor ( ). The cubic slope factor ( ) was foun d not to be significantly different from the mean cubic slope. In mathematics achievement an even larger negative mean intercept slope factor is estimated than children who were high on TSS. These results suggest that children of African American culture, on average, begin kindergarten with lower initial mean mathematics achievement scores while showing a lower rate of change in mathematics achievement over time and this trend is attenuating the mean quadratic trend of cup down or decelerating rate of chan g e over time. TIC 7 : Latina/o culture and TSS, DAS, Gender, SES, AA culture. The relationship of Latina/o culture on mathematics achievement while controlling for TSS, DAS, gender, SES, Asian American, and African American culture showed a significant negati ve intercept factor ( ) and positive quadratic slope factor ( ) The linear slope factor ( ) and cubic slope factor ( ) was found to not be significantly different from the respective mean slope factors Therefore these estimates suggest that, on average, children from Latina/o cultures begin school with a lower initial mathematics achievement score as compared to European American children and d ue to a cup up quadratic PAGE 115 115 slope factor have less of a deceleration in growth over time. Latina/o children do not differ on the other trajectories of mathematics achievement over time. Based on the results of the fourth stati stical analysis, there is a significant relationship between traumatic stress symptoms, difficult attachment symptoms, gender, SES, Asian American, AA culture, and Latina/o culture with mathematics achievement. Children who are reported as having higher le vels of traumatic stress, difficult attachment symptoms, are female, low in SES, and are either African American or Latina/o also have lower levels of initial performance and performance over time. These findings support prior research that has looked at t raumatic stress (Goodman et al., 2012), difficult attachment symptoms (Lieberman, 2004; Pianta & Harbers, 1996), gender (Clark et al., 2008; Dearing et al., 2009), low in SES (Chatterji, 2006), and are either African American or Latina/o (Crumpton & Gregor y, 2011; Morris, 2008) as indicators for lower mathematics achievement. The findings of this study go further to suggest that six of the covariates predict lower levels of mathematics achievement and that these lower levels can persist across grade levels from kindergarten through at least fifth grade as well as through developmental trajectories ; with the exception of Asian American culture which actually shows larger growth rates than their European American peers Reading and Mathematics Growth Trajector y and TSS, DAS, Gender, SES, Culture, and the Interaction B etween TSS and DAS The fifth research question explores: What is the relationship of an interaction effect between traumatic stress symptoms and difficult attachment symptoms on the growth traject ory of reading and mathematics achievement for children in kindergarten through fifth grade, while controlling for gender, culture, and socioeconomic status? This question was examined with th e level 2 model in Equation 4 15 : PAGE 116 116 ( 4 1 5 ) where is the time invariant covariate ( TIC ): traumatic stress symptoms (TSS), is the TIC : difficult attachment symptoms (DAS), is the TIC gender, represents the TIC gender of interviewed parent/caregiver, represents the TIC SES, the next four variables represent dummy codes of the variable culture: represents the TIC Asian American culture, represents the TIC African American culture, represents the TIC Latina/o culture represents the TIC other people of color culture The final TIC represents TSS DAS (interaction between TSS and DAS). However, as mentioned earlier, TIC and were found to be insignificant in the model and ar e not discussed in th e results. The following are the mean estimates for reading achievement for higher SES, male, European American children who are low on TSS and DAS: mean intercept slope factor = mean linear slope = mean quadrati c slope = and the mean cubic slope = The following are the mean estimates for mathematics achievement for higher SES, male, European American children who are low on TSS and DAS: mean intercept slope factor = mean linear slope = mean quadratic slope = and the mean cubic slope = For the reading data, results of the LRT, AIC, and BIC when compared to the cubic LGCM with seven TIC s PAGE 117 117 ) showed that the addition of the TIC interaction between TSS and DAS had significantly change d the model. It also showed that the addition of the interaction resulted in smaller AIC and BIC values showing it provided a better fitting model. The mathematics data ) showed similar results to this trend. The estimated means maintained accuracy for both reading and mathematics the residual variances were significant and positive for both, and the stayed similar for both TSS and DAS, thus, still supporting the model fit. Reading Achievement LGCM With Eight TICs TIC1: TSS and DAS, TSS DAS, Gender, SES, Culture. For the reading achievement trajectory data, TSS was examined first while controlling for DAS, TSS DAS, gender, SES, and culture. TSS significantly predicted a negative intercept factor ( ), a negative linear slope factor ( ), and a positive quadratic slope factor ( ). Indicating that, on average, children high in TSS are observed to have lower initial reading achievement scores with a downward r ate of change that reflects with a one unit increase in TSS that there is an association with a 3.6 1 decrease in the mean slope of the trajectory of reading achievement. The concave up quadratic slope for children high in TSS suggests that they are plateau ing slower than children low in TSS. The cubic slope factor was found to be non significantly different from the mean cubic slope factor ( ). TIC2: DAS and TSS, TSS DAS, Gender, SES, Culture. DAS was examined next controlling for TSS, TSS DAS, gender, SES, and culture. DAS significantly predicted a negative intercept factor ( ) and a negative linear slope factor ( ). It showed no significant differences from the mean PAGE 118 118 quadratic slope ( ) nor from the mean cubic slope ( ). This suggests that when all other TICs are controlled for, on average, children high in DAS show a lower mean initial reading achievement score with a lower linear trend that has a downward rate of change over time ind icating that with a one unit change in DAS there is an associated 2.5 1 decrease in the slope trajec tory of reading achievement. TIC3: TSS DAS and TSS, DAS, Gender, SES, Culture. The interaction next to be examined controlling for TSS alone, DAS alone, gender, SES, and culture. The TSS DAS variable did not show significantly different results from the mean intercept factor ( ), quadratic slope factor ( ) or cubi c slope factor ( ). TSS DAS did show a significantly lower linear slope factor ( ). These results suggest that the interaction between TSS and DAS causes a lower rate of g rowth over time for children high in TSS and in DAS and that one unit of change in is associated with a 2. 95 decrease in the slope trajectory of reading achievement over time. A further implication of these results is that TSS is dependent on the e ffects of DAS and DAS is dependent on the effects of TSS towards their overall effect on reading achievement. TIC4: Gender and TSS, DAS, TSS DAS, SES, Culture. The results of the analysis for gender controlling for TSS, DAS, were looked at next. Gender showed a significant positive intercept factor ( ), a positive linear slope factor ( ), and a negative quadratic slope factor ( ). Gender did not show significantly different results than the mean cubic slope factor ( ). This indicated that females, on average, showed higher reading achievement scores at initial asses sment and have a PAGE 119 119 higher rate of change in reading achievement over time. The concave downward nonlinear trend predicted by the quadratic slope factor suggests that females are slowing the rate that their negative (cup down) quadratic growth is decelerating their linear growth ; more so than is happening to males over time. Where males start out lower in reading achievement, have lower rates of positive change, but show a s lower deceleration in the change trajectory over time. TIC5: SES and TSS, DAS, TSS DAS, Gender, Culture. Next socioeconomic status (SES) is examined while controlling for TSS, DAS, TSS DAS, gender, and culture. SES significantly predicted a positive intercept factor ( ), a positive linear sl ope factor ( ), and a negative quadratic slope factor ( ). SES did not predict a significantly difference from the mean cubic slope factor ( ) This estimation predicts that on average, children whose families are higher in SES are observed to have higher initial reading achievement scores with a larger positive rate of change over time that suggests that a one unit increase in SE S is associated with a 7.30 increase in the slope of the trajectory of reading achievement. Those high in SES also show a larger concave down or greater decelerating quadratic trend. This suggests that reading achievement plateaus faster for higher SES chi ldren as they head toward third and fifth grade. TIC 6 : Asian Culture and TSS, DAS, TSS DAS, Gender, SES, AA, Latina/o culture. The relationship of A sian American culture on reading achievement while controlling for TSS, DAS, gender, SES, AA, and Latina/o culture showed significant predictions for a negative linear slope factor ( ) and cubic slope factor ( ) T he intercept factor ( ) and the quadratic slope factor ( ) were found to not PAGE 120 120 be significantly different from their respective mean slopes. These results show that children of A sian American culture, on average, for reading achievement show a lower growth rate over time as compared to children of European Amer ican and Latina/o cultures and that their cubic factor is maintaining the quadratics deceleration TIC 7 : AA Culture and TSS, DAS, TSS DAS, Gender, SES, Asian, Latina/o culture. The relationship of African American culture on reading achievement while controlling for TSS, DAS, Latina/o culture showed significant predictions for a ne gative intercept factor ( ) and negative linear slope factor ( ), however, the quadratic slope factor ( ) and cubic slope facto r ( ) were found to be no t significant ly different from their respective means. Again African American cultures negative mean intercept slope factor was close to being as large as children who were high on TSS. Th ese results suggest that children of African American culture, on average, begin kindergarten with lower initial mean reading achievement scores showing a lower rate of change in reading achievement over time as compared to children of European American A sian American, and Latina/o cultures. TIC 8 : Latina/o culture and TSS, DAS, TSS DAS, Gender, SES, Asian, AA culture. The relationship of Latina/o culture on reading achievement while controlling for TSS, DAS, TSS DAS, gender, SES, and other cultures showed a significant negative intercept factor ( ), but was not found to be significantly different from the mean predicted linear slope factor ( ), quadratic slope factor ( ), and cubic slope factor ( ). Therefore these estimates suggest that, on average, children from Latina/o PAGE 121 121 cultures begin school with a lower initial reading achievement score as compared to European American children and are similar to the mean s for the other trajectory trends. Based on the results of the fifth statistical analysis, there is a significant relationship between traumatic stress symptoms, difficult attachment symptoms, the interaction between TSS DAS gender, SES, Asian American, A frican A merican and Latina/o culture with reading achievement. Children who are reported as having higher levels of traumatic stress, difficult attachment symptoms, have an interaction between TSS DAS, are male, low in SES, and Asian American, African American, or Latina/o also have lower levels of initial performance and/or performance over time. These findings support prior research that has looked at traumatic stress (Goodman et al., 2012), difficult attachment symptoms (Lieberman, 2004; Pianta & Harbers, 199 6), gender (Clark et al., 2008; Dearing et al., 2009), low in SES (Chatterji, 2006), and are either African American or Latina/o (Crumpton & Gregory, 2011; Morris, 2008) as indicators for lower reading achievement. The findings of this study go further to suggest that all seven covariates including the interaction between TSS DAS predict lower levels of reading achievement and that the se predictions can persist across grade levels from kindergarten through at least fifth grade as well as through development al trajectories. Mathematics Achievement LGCM With Eight TICs For the mathematics gender, SES, and culture. TSS signif icantly predicted a negative intercept factor ( ), a negative linear slope factor ( ), a positive quadratic slope factor ( ), a nd negative cubic slope factor ( ). Indicating that when all PAGE 122 122 other TICs are controlled for, on average, children high in TSS are observed to have lower initial mathematics achievement scores with a lower downward rate of change that reflects with a one unit increase in TSS that there is an associated 3. 73 decrease in the mean slope of the trajectory of mathematics achievement over time. The concave up or positive quadratic slope factor for children high in TSS sugg ests that they are decelerating at a slower rate in their growth trajectory as compared to children low in TSS. TIC2: DAS and TSS, TSS DAS, Gender, SES, Culture. DAS was examined next controlling for TSS, TSS DAS, gender, SES, and culture. DAS significantly predicted a negative intercept factor ( ) a negative linear slope factor ( ) and a positive quadratic slope factor ( ). It showed no significant differences for the mean cubic slope factor ( ). This suggests that when all other TICs are controlled for, on average, children high in DAS show a lower initial mathematics achievement score with a lower growth rate over time that reflects with a one unit increase in DAS that there is an associated 1.3 6 decrease in the mean slope of the trajectory of mathematics achievement over time. The concave up or positive quadratic slope factor for children high in DAS suggests that they are decelerating at a slower rate in their growth trajectory as compared to c hildren low in DAS next to be examined controlling for TSS alone, DAS alone, gender, SES, and culture. The mean intercept factor ( ), quadratic slope factor ( ) or cubic slope factor ( significantly lower linear slope factor ( ). These results PAGE 123 123 suggest that the interaction between TSS and DAS causes a lower rate of growth over time for children high in TSS and in DAS and that one h a 2.85 decrease in the slope trajectory of mathematics achievement over time. Another implication of these results is TSS is dependent on the effects of DAS and DAS is dependent on the effects of TSS towards their overall effect on mathematics achievemen t. TIC4: Gender and TSS, DAS, TSS DAS, SES, Culture. The results of the analysis for gender controlling for TSS, DAS, TSS DAS, SES, and culture were looked at next. Gender showed a significant positive intercept factor ( ), a negative linear slope factor ( ), and a positive quadratic slope factor ( ). Gender did not show significantly different results than the mean cubic slope factor ( ). This indicated that females, on average, showed lower mathematics achievement scores at initial assessment and lower rates of change in mathematics achievement over time. The concave up ward quadratic slope fac tor for females is predicting a smaller deceleration of the negative growth trajectory over time than the mean for males Where males start out higher in mathematics achievement, have higher rates of positive change, but show a higher deceleration in the c hange trajectory over time. TIC5: SES and TSS, DAS, TSS DAS, Gender, Culture. Next socioeconomic status (SES) is examined while controlling for TSS, DAS, TSS DAS, gender, and culture. SES significantly predicted a positive intercept factor ( ), a positive linear slope factor ( ), and a negative quadratic slope factor ( ). SES did not predict a significant difference from the mean cubic slope factor ( ). This PAGE 124 124 estimation predicts that on average, children whose families are higher in SES are observed to have higher initial mathematics achievement scores with a larger positive rate of change over time that suggests that a one unit increase in SES is associated with a 6. 15 increase in the slope of the trajectory of mathematics achievement. Those high in SES also show a larger concave down or greater decelerati on in the growth trend over time This suggests that mathematics achievement plateaus faster for higher SES children as they head toward third and fifth grade. TIC 6 : Asian Culture and TSS, DAS, TSS DAS, Gender, SES, AA, Latina/o culture. The relationship of A sian American culture on mathematics achievement while controlling for TSS, DAS, gender, SES, AA, and Latina/o culture showed significant predictions for a positive linear slope fact or ( ) quadratic slope factor ( ) and a negative cubic slope factor ( ) T he intercept factor ( ) was found to not be significantly different from its mean slope. These results show that children of A sian American culture, on average, for mathematics achievement show a higher growth rate over time as compared to children of European Americ an African American, and Latina/o cultures The positive or concave up quadratic slope factor trajectory is decelerating at a slower rate and their cubic slope factor causes the mean cubic slope trajectory to be relatively close to zero Essentially this mean that Asian American children do not show as much of a plateau effect in mathematics achievement as compared to their peers. TIC 7 : AA Cultur e and TSS, DAS, TSS DAS, Gender, SES, Asian, Latina/o culture. The relationship of African American culture on mathematics achievement while controlling for Asian American, and Latina/o culture showed significant predictions for a negative intercept factor ( ), a negative PAGE 125 125 linear slope factor ( ), and a positive quadratic slope factor ( ). African Am erican culture did not show significant difference from the mean cubic slope ( ). African American cultures negative intercept slope factor showed the largest decrease in initial mathematics achievement of any oth er TIC in this study. These results indicate that children of African American culture, on average, begin kindergarten with lower initial mean mathematics achievement scores showing a lower rate of change in mathematics achievement over time as compared to children of European American Asian American, and Latina/o cultures. The positive or concave upward quadratic slope factor suggests that African American children have a slower deceleration of the growth trend over time ; thus plateau ing slower than their peers. TIC 8 Asian, AA culture. The relationship of Latina/o culture on mathematics achievement while controlling for TSS, ercept factor ( ) and positive quadratic slope factor ( ), but was not found to be significantly different from the mean predicted linear slope ( ) nor the mean cubic slope ( ). Therefore these estimates suggest that, on average, children from Latina/o cultures begin school with a lower initial mathematics achievement score and their cup up quadratic slope factor causes the deceleration of the growth rate to be slowed, thus they do not plateau in growth as fast as European American children. Based on the results of the fifth statistical analysis, there is a significant relationship between tr aumatic stress symptoms, difficult attachment symptoms, the interaction between TSS DAS gender, SES, Asian American culture, AA culture, and Latina/o culture with PAGE 126 126 mathematics achievement. Children who are reported as having higher levels of traumatic stress, difficult attachment symptoms, have an interaction between TSS DAS, are fe ma le, low in SES, and are either African American or Latina/o also have lower levels of initial performance and performance over time. These findings support prior research that has looked at traumatic stress (Goodman et al., 2012), difficult attachment symp toms (Lieberman, 2004; Pianta & Harbers, 1996), gender (Clark et al., 2008; Dearing et al., 2009), low in SES (Chatterji, 2006), and are either African American or Latina/o (Crumpton & Gregory, 2011; Morris, 2008) as indicators for lower mathematics achiev ement. The findings of this study go further to suggest that six of the covariates including the interaction between TSS DAS predict lower levels of mathematics achievement and that these predictions can persist across grade levels from kindergarten throug h at least fifth grade as well as through developmental trajectories. Grand Summary of Results The results of this study provide evidence that there is a significant relationship between traumatic stress symptoms, difficult attachment symptoms, and their i nteraction on the longitudinal trajectory of reading and mathematics achievement for young children at kindergarten through fifth grade, even after controlling for the effects of gender, SES, and culture. Thus, children high in traumatic stress symptoms co nsistently had a lower initial reading and math ematics achievement scores while also consistently showing lower growth rates over time The consistent quadratic trends for reading and quadratic and cubic trends in math ematics provide support for the hypoth esis that traumatic stresses effects are not linear, but persist ove r time and developmental level. Children high in difficult attachment symptoms also consistently showed lower initial reading and mathematics achievement scores However, the effects becam e different when gender, SES, and culture were included. At this time in reading, difficult attachment symptoms PAGE 127 127 showed lower growth trends over time but no effects on the mean quadratic or cubic trends. In mathematics, difficult attachment symptoms showed lower initial scores in mathematics, but with the interaction it also showed nonlinear influences, which hint that its effects might be more on development than measured achievement. When looking at gender the results are consistent that females show low er initial scores in math ematics and persiste nt lower growth ov er time while having nonlinear quadratic effects. Likewise, males consistently show lower initial scores in reading and persistent lower growth rates over time while having nonlinear quadratic effects. SES also consistently showed that lower SES children had relatively large lower initial reading and mathematics scores, had lower rates of growth over time, and nonlinear quadratic effects Cultur al effects were more interesting, when the model di d not include the interaction, culture showed significantly lower initial score s in reading and mathematics for African American and Latina/o children as well as lower growth rates over time ; Asian Americans surprisingly were included in the lower growth r ates over time However, when the interaction between traumatic stress and difficult attachment symptoms was includ ed in the model, culture predicted lower initial scores in reading and mathematics for African American and Latina/o children, while suggesti ng lower growth rates over time for Asian and African American children An interesting effect of the interaction was significant cubic slope factor estimation for Asian children. This suggests that there are significant levels of traumatic stress and chal lenges to the attachment behaviors of culturally different children within the educational system that persist over time and across developmental levels. This consistency within gender, SES, and culture suggests a systemic issue within the education system for these patterns to persist across developmental trajectories and grade levels PAGE 128 128 These findings support prior research about young children that has looked at the effects of traumatic stress (Goodman et al., 2012), difficult attachment symptoms (Lieberma n, 2004; Pianta & Harbers, 1996), gender (Clark et al., 2008; Dearing et al., 2009), low SES (Chatterji, 2006), and either African American or Latina/o cultures (Crumpton & Gregory, 2011; Morris, 2008) as indicators for lower reading and mathematic s achiev eme nt. The findings of this study go further to suggest that all s even covariates including the interaction between TSS DAS predict effects earlier than other studies have before T hese effects may be larger than previously expected, and persist across gra de levels from kindergarten through at least fifth grade as well as s ho w ing significant negative trends in developmental trajectories ( Figure 4 21 ) PAGE 129 129 Table 4 1. Covariance coverage of the proportion of data present in the s ample reading achievement data for this study. 1 2 3 4 5 6 7 8 9 10 11 12 1. P1ATTENI 0.997 2. P2HRDWRM 0.985 0.988 3. P2CHDOES 0.984 0.987 0.988 4. P2FLTRAP 0.984 0.986 0.986 0.987 5. P2FEELAN 0.985 0.987 0.987 0.986 0.988 6. P2CHHARD 0.984 0.987 0.987 0.986 0.987 0.988 7. READ1 0.919 0.912 0.912 0.911 0.912 0.912 0.922 8. READ2 0.947 0.939 0.939 0.938 0.940 0.939 0.916 0.950 9. READ4 0.959 0.951 0.951 0.950 0.951 0.951 0.906 0 .935 0.962 10. READ5 0.954 0.946 0.946 0.945 0.947 0.946 0.887 0.914 0.932 0.958 11. READ6 0.949 0.941 0.941 0.940 0.941 0.941 0.882 0.910 0.927 0.935 0.953 12. T2CONTRO 0.958 0.950 0.949 0.949 0.950 0.949 0.893 0.920 0.931 0.923 0.919 0.961 13. T2INTERP 0.954 0.946 0.945 0.944 0.946 0.945 0.889 0.916 0.926 0.919 0.915 0.953 14. T2EXTERN 0.959 0.951 0.951 0.950 0.952 0.951 0.894 0.921 0.932 0.924 0.920 0.958 15. T2INTERN 0.955 0.947 0.947 0.946 0.947 0.947 0.892 0.919 0.929 0.921 0.917 0.954 16 GENDER 0.997 0.988 0.988 0.987 0.988 0.988 0.922 0.950 0.962 0.958 0.953 0.961 17. WKSESL 0.997 0.988 0.988 0.987 0.988 0.988 0.922 0.950 0.962 0.958 0.953 0.961 18. R_A1 0.997 0.988 0.988 0.987 0.988 0.988 0.922 0.950 0.962 0.958 0.953 0.961 19. R_B2 0.997 0.988 0.988 0.987 0.988 0.988 0.922 0.950 0.962 0.958 0.953 0.961 20. R_L3 0.997 0.988 0.988 0.987 0.988 0.988 0.922 0.950 0.962 0.958 0.953 0.961 21. R_U4 0.997 0.988 0.988 0.987 0.988 0.988 0.922 0.950 0.962 0.958 0.953 0.961 22. P1SEX_1 0.997 0.988 0.988 0.987 0.988 0.988 0.922 0.950 0.962 0.958 0.953 0.961 13 14 15 16 17 18 19 20 21 22 13. T2INTERP 0.957 14. T2EXTERN 0.954 0.963 15. T2INTERN 0.950 0.955 0.959 16. GENDER 0.957 0.963 0.959 1.000 17. WKSESL 0.957 0.963 0.959 1.000 1.000 18. R_A1 0.957 0.963 0.959 1.000 1.000 1.000 19. R_B2 0.957 0.963 0.959 1.000 1.000 1.000 1.000 20. R_L3 0.957 0.963 0.959 1.000 1.000 1.000 1.000 1.000 21. R_U4 0.957 0.963 0.959 1.000 1.000 1.000 1.000 1.000 1.000 22. P1SEX_1 0.957 0.963 0.959 1.000 1.000 1.000 1.000 1.000 1.000 1.000 PAGE 130 130 Table 4 2. Covariance coverage of the proportion of data present in the sample mathematics achievement data for this study. 1 2 3 4 5 6 7 8 9 10 1 1 12 1. P1ATTENI 0.997 2. P2HRDWRM 0.985 0.988 3. P2CHDOES 0.984 0.987 0.988 4. P2FLTRAP 0.984 0.986 0.986 0.987 5. P2FEELAN 0.985 0.987 0.987 0.986 0.988 6. P2CHHARD 0.984 0.987 0.987 0.986 0.987 0. 988 7. READ1 0.971 0.964 0.963 0.962 0.964 0.963 0.974 8. READ2 0.980 0.972 0.972 0.971 0.973 0.972 0.967 0.984 9. READ4 0.975 0.967 0.967 0.965 0.967 0.967 0.957 0.968 0.978 10. READ5 0.958 0.951 0.950 0.949 0.951 0.950 0.940 0.95 0 0.951 0.962 11. READ6 0.950 0.942 0.942 0.941 0.942 0.942 0.931 0.941 0.942 0.940 0.954 12. T2CONTRO 0.958 0.950 0.949 0.949 0.950 0.949 0.941 0.951 0.945 0.927 0.920 0.961 13. T2INTERP 0.954 0.946 0.945 0.944 0.946 0.945 0.937 0.947 0.940 0.923 0. 916 0.953 14. T2EXTERN 0.959 0.951 0.951 0.950 0.952 0.951 0.941 0.952 0.946 0.928 0.921 0.958 15. T2INTERN 0.955 0.947 0.947 0.946 0.947 0.947 0.938 0.949 0.942 0.925 0.917 0.954 16. GENDER 0.997 0.988 0.988 0.987 0.988 0.988 0.974 0.984 0.978 0.962 0. 954 0.961 17. WKSESL 0.997 0.988 0.988 0.987 0.988 0.988 0.974 0.984 0.978 0.962 0.954 0.961 18. R_A1 0.997 0.988 0.988 0.987 0.988 0.988 0.974 0.984 0.978 0.962 0.954 0.961 19. R_B2 0.997 0.988 0.988 0.987 0.988 0.988 0.974 0.984 0.978 0.962 0.954 0.96 1 20. R_L3 0.997 0.988 0.988 0.987 0.988 0.988 0.974 0.984 0.978 0.962 0.954 0.961 21. R_U4 0.997 0.988 0.988 0.987 0.988 0.988 0.974 0.984 0.978 0.962 0.954 0.961 22. P1SEX_1 0.997 0.988 0.988 0.987 0.988 0.988 0.974 0.984 0.978 0.962 0.954 0.961 13 14 15 16 17 18 19 20 21 22 13. T2INTERP 0.957 14. T2EXTERN 0.954 0.963 15. T2INTERN 0.950 0.955 0.959 16. GENDER 0.957 0.963 0.959 1.000 17. WKSESL 0.957 0.963 0.959 1.000 1.000 18. R_A1 0.957 0.963 0.9 59 1.000 1.000 1.000 19. R_B2 0.957 0.963 0.959 1.000 1.000 1.000 1.000 20. R_L3 0.957 0.963 0.959 1.000 1.000 1.000 1.000 1.000 21. R_U4 0.957 0.963 0.959 1.000 1.000 1.000 1.000 1.000 1.000 22. P1SEX_1 0.957 0.963 0.959 1.000 1.000 1.000 1.000 1.000 1.000 1.000 PAGE 131 131 Table 4 3. The latent intercept, linear, quadratic, and cubic factor coefficients, their standard errors, and p values for the cultural variable other people of color for the research questions 4 and 5 as applied to rea Variable Reading LGCM Mathematics LGCM Other People of Color Factor Coefficient S.E. value Factor Coefficient S.E. value Research Question 4 : Intercept 3.84 3.80 0.312 5.62 3.15 0.074 : Linear Slope 1.11 3.17 0.726 0.88 2.66 0.740 : Quadratic Slope 1.65 1.29 0 .200 1.62 0.60 0.007 : Cubic Slope 0.72 1.44 0.615 0.27 0.81 0.737 Research Question 5 : Intercept 3.93 3.82 0.303 5.65 3.14 0.072 : Linear Slope 1.25 3.20 0.697 0.98 2. 66 0.712 : Quadratic Slope 1.62 1.29 0.209 1.58 0.59 0.008 : Cubic Slope 0.76 1.44 0.596 0.28 0.81 0.729 Note. Bolded values are significantly different from the mean quadratic slope factor. Table 4 4. Th e latent intercept, linear, quadratic, and cubic factor coefficients, their standard errors, and p values for the variable for gender of responding parent/guardian for the research questions 4 and 5 as applied to reading and mathematics achievement Variable Reading LGCM Mathematics LGCM Gender of Responding Parent/Guardian Factor Coefficient S.E. value Factor Coefficient S.E. value Research Question 4 : Intercept 0.67 2.29 0.7 69 1.16 2.41 0.63 1 : Linear Slope 1.01 1.23 0.41 1 0.10 1.38 0.94 4 : Quadratic Slope 0.02 0.88 0 .9 79 0.03 0.57 0.96 1 : Cubic Slope 1.70 0.91 0.06 0 0.09 0.52 0.86 1 Research Question 5 : Intercept 0. 83 2. 31 0. 7 2 0 1.22 2.41 0.61 4 : Linear Slope 1. 17 1.2 3 0. 343 0.22 1.38 0.8 77 : Quadratic Slope 0. 02 0.87 0.9 83 0.06 0.57 0.91 4 : Cubic Slope 1. 74 0. 90 0.0 54 0.08 0.52 0.87 4 PAGE 132 132 Table 4 5 Standardized c orrelations observed means, and model estimated means for readin g achievement using the HMLM cubic polynomial model. Read Time 1 Read Time 2 Read Time 3 Read Time 4 Read Time 5 Reading Time 1: Fall 1998 1.00 Reading Time 2: Spring 1999 0.83 1.00 Reading Time 3: Spring 2000 0.68 0.76 1.00 Readin g Time 4: Spring 2002 0.56 0.60 0.74 1.00 Reading Time 5: Spring 2004 0.52 0.56 0.70 0.86 1.00 Estimated Mean 34.73 45.79 76.82 125.42 148.35 Observed Mean 34.56 45.76 76.67 125.4 0 148.37 Standard Deviations 10.30 14.08 23.56 27.23 25.49 Not es. is inserted for duplicates to the correlation matrix. Table 4 6 Standardized correlations, observed means, and model estimated means for mathematics achievement using the HMLM cubic polynomial model. Math Time 1 Math Time 2 Math Time 3 Math Time 4 Math Time 5 Math Time 1: Fall 1998 1.00 Math Time 2: Spring 1999 0.83 1.00 Math Time 3: Spring 2000 0.72 0.78 1.00 Math Time 4: Spring 2002 0.69 0.73 0.78 1.00 Math Time 5: Spring 2004 0.63 0.68 0.74 0.87 1.00 Estima ted Mean 25.77 35.64 61.19 97.64 121.70 Observed Mean 25.77 35.98 60.92 97.95 121.88 Standard Deviations 9.40 12.13 18.05 24.18 23.95 Notes. is inserted for duplicates to the correlation matrix. Table 4 7 Comparison of the model estimated traje ctory means for reading achievement Read Time 1 Read Time 2 Read Time 3 Read Time 4 Read Time 5 Observed Means 34.56 45.76 76.67 125.40 148.37 Linear Estimates 36.10 43.50 65.70 110.11 154.53 Free Loading Estimates 34.67 45.92 77.41 125.91 147.93 H MLM Quadratic Orthogonal Polynomial Estimates 34.86 46.63 78.19 124.51 148.43 HMLM Cubic Orthogonal Polynomial Estimates 34.73 45.79 76.82 125.42 148.35 PAGE 133 133 Table 4 8 Comparison of the model estimated trajectory means for mathematics achievement Mat h Time 1 Math Time 2 Math Time 3 Math Time 4 Math Time 5 Observed Means 25.77 35.98 60.92 97.95 121.88 Linear Estimates 27.95 34.50 54.16 93.49 132.81 Free Loading Estimates 25.73 36.06 61.09 98.27 121.39 HMLM Quadratic Orthogonal Polynomial Estimates 25.92 35.16 60.27 98.79 121.71 HMLM Cubic Orthogonal Polynomial Estimates 25.77 35.64 61.19 97.64 121.70 Table 4 9 EFA factor structure used to select the indicators for difficult attachment symptoms. Factor Structure 1 2 3 P2CHDOES 0.66 0.29 0.3 5 HRDWRMB 0.51 0.24 0.30 P2FLTRAP 0.73 0.27 0.47 P2FEELAN 0.83 0.26 0.45 P2CHHARD 0.69 0.28 0.36 P2ARGCHD 0.20 0.63 0.17 P2ARGAFF 0.22 0.59 0.24 P2ARGHEA 0.14 0.70 0.20 P2HITCHR 0.27 0.58 0.33 P2DEPRES 0.37 0.39 0.61 P2FEARFL 0.29 0 .26 0.83 Note. Factor loadings greater than .50 and are low loading on other factors are shown in bold. Table 4 10 Descriptive statistics for difficult attachment symptoms Mean SD Min Max Difficult Attachment Symptoms 0.06 0.55 1.25 2.93 Proportion of Occurrence Category 1 Category 2 Category 3 P2HARDWRM 0.76 0.18 0.06 P2CHDOES 0.89 0.05 0.06 P2FLTRAP 0.96 0.02 0.02 P2FEELAN 0.98 0.01 0.01 P2CHHARD 0.95 0.02 0.03 Category 1: Not at all true Category 2: Somewhat true Category 3: Completely/Mostly true PAGE 134 134 Table 4 1 1 Descriptive statistics for traumatic stress symptoms Mean SD Min Max R esponse Scale Traumatic Stress Symptoms 0.08 0.54 1.02 2.13 Continuous Composite Variables T2CONTRO 1.76 0.60 1. 00 4.00 1. very often 2. often 3. sometimes 4. never T2INTERP 1.82 0.62 1.00 4.00 1. very often 2. often 3. sometimes 4. never T2EXTERN 1.60 0.60 1.00 4.00 1. never 2. sometimes 3. often 4. very often T2INTERN 1.53 0.49 1.00 4.00 1. never 2. sometim es 3. often 4. very often Ordinal Categorical Variable Proportion of Occurrence Category 1 Category 2 Category 3 P1ATTENI 0.25 0.60 0.15 1. Better than others 2. As well as others 3. Slightly less/ much less well than ot hers PAGE 135 135 Table 4 1 2 Standardized factor loadings, their standard errors, factor correlations and estimates for latent variables traumatic stress symptoms and difficult attachment symptoms. Factor Loadings Traumatic Stress Symptoms Difficult Attachment Symptoms T2CONTRO 0.9 4 (0.01) 0.90 T2INTERP 0.84 (0.01) 0.70 T2EXTERN 0.78 (0 .01) 0.61 T2INTERN 0.3 4 (0.03) 0.11 P1ATTENI 0.3 6 (0.02) 0.13 P2HRDWRM 0.47 (0.04) 0.22 P2CHDOES 0.7 1 (0.04) 0.51 P2FLTRAP 0.80 (0.04) 0.64 P2FEELAN 0.77 (0.06) 0.60 P2CHHARD 0.75 (0.05) 0.56 Factor Correlations Traumatic Stre ss Symptoms 0.25 (0.05) Difficult Attachment Symptoms Note. is used for correlations equal to one or are repeated values. PAGE 136 136 Figure 4 1 The density plots (on top) and histogram plot s (on bottom) of the five measurement periods for reading achievement. PAGE 137 137 Figure 4 2 The density plots (on top) and histogram plot s (on bottom) of the five measurement periods for mathematics achievement. PAGE 138 138 Figure 4 3 The randomly selected plot s of fifteen individuals observed longitudinal trajectories f or reading achievement PAGE 139 139 Figure 4 4 The randomly selected plot s of fifteen individuals observed longitudinal trajectories for mathematics achievement. PAGE 140 140 Figure 4 5 The total information curve of all of the indicators as a function of the latent vari able difficult attachment symptoms. Figure 4 6 The characteristic curves of all of the indicators as a function of the latent variable difficult attachment symptoms. PAGE 141 141 Figure 4 7 The total information curve of all of the indicators as a function of the latent variable traumatic stress symptoms. Figure 4 8 The characteristic curve of the indicator P1ATTENI as a function of the latent variable traumatic stress symptoms. PAGE 142 142 Figure 4 9 The plot of the reading achievement linear trend. Figure 4 10 The plot of the reading achievement quadratic trend. PAGE 143 143 Figure 4 11 The plot of the reading achievement cubic trend. Figure 4 12 The plot of the mathematics achievement linear trend. PAGE 144 144 Figure 4 13 The plot of the mathematics achievement quadr atic trend Figure 4 14 The plot of the mathematics achievement cubic trend. PAGE 145 145 Figure 4 15 The path model analyzing the LGCM of reading achievement with the time invariant covariate traumatic stress symptoms. The red line is an insignificant estima te all others were significant. PAGE 146 146 Figure 4 16 The path model analyzing the LGCM of mathematics achievement with the time invariant covariate traumatic stress symptoms. All paths were significant. PAGE 147 147 Figure 4 17 The path model analyzing the LGCM of read ing achievement with the time invariant covariate difficult attachment symptoms. The red line is an insignificant estimate all others were significant. PAGE 148 148 Figure 4 18 The path model analyzing the LGCM of mathematics achievement with the time invariant cov ariate difficult attachment symptoms. The red line is an insignificant estimate all others were significant. PAGE 149 149 Figure 4 19 The path model analyzing the LGCM of reading achievement with the time invariant covariates traumatic stress symptoms and difficult attachment symptoms. The red line is an insignificant estimate all others were significant. PAGE 150 150 Figure 4 20 The path model analyzing the LGCM of mathematics achievement with the time invariant covariates traumatic stress symptoms and difficult attachment symptoms. The red line is an insignificant estimate all others were significant. PAGE 151 151 Figure 4 21. This describes the overall results across all five models with the cubic orthogonal polynomial latent growth curve model with seven time invariant covariate s (TICs) and one time invariant covariate interaction. PAGE 152 152 CHAPTER 5 DISCUSSION The purpose of this chapter is to discuss the findings regarding this study, the implications of these findings, and how these can be applied to practice, policy, and future resea rch. The relationships explored in this study used complex methodologies and covered support the need for further work and developm ent in this area. Findings F rom the Study development is significantly affected by the following intersecting aspects of their lives: traumatic stress symptoms, difficult attachment experiences, gender SES, and culture/ethnicity. The following sections are a beginning to this discussion. Traumatic Stress Symptoms This study found relatively large effects from the latent variable traumatic stress symptoms (TSS) of young children as assessed at entry into kindergarten. The results of children expressing the various forms of hyper/hypo arousal, avoidance, or re experiencing through verbal and nonverbal behaviors ( Bailey et al. 2007; De Young et al., 2011a, 2011b; Mongillo et al. 2009) were related to significant levels of underachievement. While controlling for the other covariates, TSS was associated with a relatively large reduction in achievement scores at entry into kindergarten as well as a persistent lower growth rate over time. There was also ev idence of significant nonlinear trends that often resulted in a flattening or premature plateauing effect on mathematics achievement trajectories. The TSS measurement model adapted from prior research (Goodman, 2009; Goodman et al., 2012; R. D. Goodman, personal communication, March 7, 2013) appeared well suited to PAGE 153 153 in self control, having trouble forming interpersonal relationships, and high in externalized behaviors explained 90%, 70%, and 60% respectively of the variance in the LGCM. However, the measures of internalizing behavior and attention qualities of the children appear to address less of this issue as they only explained 11% and 13% of variance respectively. There could be several possible reasons for this. These may not be developmentally appropriate as young children may not be displaying these behaviors yet. Alternatively, it could be the issue where behaviors are passed off as per sonal or cultural deficits, which often inhibit educators from accurately assessing the risks and barriers for marginalized students (Robinson & Biran, 2006; Or it could be that parents and teachers are not aware of how to prop erly assess young children for internalizing behaviors and attention qualities. This supports the need for better developmentally appropriate direct assessments for children in this area. Difficult Attachment Symptoms The explorat ory nature of this latent variable provided insight into the effects of relationships that get in the way of warmth and support for young children. This study found significant effects for reading and mathematics where children high in DAS showed persistently lower achievement sco res at entry into kindergarten and had reduced growth trends over time, while controlling for the other covariates. DAS also showed effects on the nonlinear growth trends associated with mathematics achievement. The DAS measurement model was chosen based o n literature reviewed (Bergin & Bergin, 2009; George & Solomon, 2008; Kobak & Madsen, 2008; Perlman & Doyle, 2012) and was moderate in its ability to explain the reading and mathematics longitudinal trends. As other covariates were added to the model, DAS showed differing results. However, a strong interaction effect was shown to be present with traumatic stress symptoms which could explain some of this PAGE 154 154 variability could be due to the fact that attachment behaviors were not directly assessed and these indicators explained parent appraisals of the relationships with their children. Regardless, four of the five items were found to explain above 50% of the variance in reading and mathematics longitudinal achievement. I nteraction Between TSS and DAS An important contribution this study had to current literature was exploring the interaction relationship between TSS and DAS, while controlling for the other covariates. The significant interaction effect between traumatic s tress and attachment adds to the conversation of et al. 2007; De Young et al., 2011a, 2011b; Lieberman, 2004; Lieberman & Van Horn, 2008; Mongillo et al. 2009 ). The int eraction effect between TSS and DAS was statistically significant. Therefore the effects of TSS on the linear slope factor ( ) are dependent on the value of DAS and likewise the effect of DAS on the linear slope factor ( ) are dependent on the value of TSS ; m eaning high in TSS and DAS both from direct effects and from their interactions. This informs the need to do further studies exploring the extent to which traumatic stress looked at the interaction between mother child and their effect on academic success (Pianta & Harbers, 1996), however, the effects of ex periences that are negative, sudden, and out of control are not all explained by mother child interactions. Therefore, direct assessments that explore the relationship between attachment experiences and traumatic stress experiences will be important for fu ture work in the area of early childhood mental health. PAGE 155 155 Academic Achievement and Sociodemographic Factors: Gender, SES, and Culture The effects of gender on the academic achievement were significant in predicting gender estimated a strong relationship with the reading and mathematical achievement trajectory. Gender explained linear effects as well as nonlinear quadratic growth in children When controlling for other covariates, females showed lower mathematics scor es and higher reading scores. Likewise, male children performed higher in mathematics and lower in reading. This trend has been supported by many prior studies (Chatterji, 2006; Dearing et al. 2009; Morris, 2008). In this sample, SES significant ly predict achievement scores, their linear trends, and nonlinear quadratic and cubic slopes for both reading and mathematics These results indicated agreement with the literature review that even after controlling for other covariates, lower initial reading and mathematics scores as well as reduction levels (Chatterji, 2006). This study found that children high in SES have lar ge gains at initial levels, large gains in growth over time, and interestingly have larger quadratic trends that level off growth ar ound third through fifth grade. mathematics scores for both African American and Latina/o children when compared to their European American peers. A surprising finding was that Asian American children predicted a significant lower growth trend in reading. T he trend of African American ch ildren showing large significant gaps compared with European American children, even when other background characteristics were controlled was shown in this study as large negative intercepts were seen and lower growth rates over time were prevalent (Chatt erji, 2006). A positive occurrence since th is data was collected is that even though European American s continue to score PAGE 156 156 higher on average than African American and Latina/o students the gap is narrowing as both African American and Latina/o st udents have made greater gains in both reading and mathematics achievement as shown in the National Center for Education Statistics ( 2012 ) The Of particular interest was the similarity between the large negative intercept factor for b oth children of African American culture and children high in TSS. This hints that evidence of environmental factors may better explain the experiences of children of color. For example, traumatic stress symptoms for African American children across SES ma y be related to contextual factors, such as racism and other forms of bias embedded institutionally in schools and other systems as previously discussed by Sue and Sue (2013) and Staton et al., (2012) Limitations This study, as mentioned before, used a n ationally norm referenced data set collected by the National Center for Education Statistics (NCES). There were several limitations to this study needing to be mentioned. The most obvious concern surrounds generalizability and construct validity, due to th e pre collected nature of the data for post hoc analyses. Concerns surround construct underrepresentation with the latent variables traumatic stress symptoms and difficult attachment symptoms (American Educational Research Association, 1999). The low values found in the internalizing and attention constructs of TSS showed that there is a more complex process going on that may require more sophisticated questions or methods for colle cting items for this construct. As for DAS, the issue is more involve d, as construct underrepresentation, construct irrelevant variance, and construct validity are potential issues. This latent variable did not have previous studies that created a factor structure to draw from and there were not direct measures of attachmen t from the data set used. This required explorative methods to construct and assess PAGE 157 157 difficult attachment symptoms in this post hoc analysis. This was done with an extensive literature review and consultation with experts. Regardless of these issues, the re sults still suggest that there were significant effects from the theoretically driven latent variable diffi cult attachment symptoms (DAS). Implications for Practice Health in early childhood is dependent on the quality of attachment relationships as well as responsiveness to adverse situations that can create traumatic stress responses. In order to address the growing needs of young children, counselors and educators need to be skilled at providing developmentally appropriate and culturally sensitive ecosy stemically informed interventions. From a counseling developmental perspective young children are seen as being capable of defining and contextualizing their experiences. Thus, through their attachment relationships and cultural perspectives they are in a continuous process of transforming their reality to meet their needs. This requires counselors and educators to use specialized assessment and cultural perspect ives such as non verbal communication and behaviors that allow counselors and educators to provide direct services. Specialized Early Childhood Assessments persistent reading and mathematics underachievement is suggested to be the result. It therefore becomes important for counselors and educators who are going to work with young children to practice s suggested that the best approaches comprise of multimodal perspectives that adjusts basic interviewing skills used require (a) establishment of safety, (b) nee d for warmth and care, (c) developing trust by PAGE 158 158 showing trust through culturally appropriate gestures, vocal tones, and facial expressions, (c) using explorative, empowering, ways grounded in theory in order to better make contact with young children at the ir own level (Erdman & Lampe, 1996 ) This becomes particularly important when they may be experiencing traumatic stress and/or difficulties in their attachment relationships or educational experiences By incorporating differentiated cultural orientations of independence and interdependence, counselors and educators can apply a more ecosystemically accurate model of practices to work with young children (Suizzo et al., 2008). Attachment and Traumatic Stress Assessments eriences for quality of attachment or level of distress that may be traumatic due to sudden, negative events, that are perceived as out of control (Goodman & West Olatunji, 2008b), a direct ecosystemic observation process is required. This means observing the child and their caregivers (parents, guardians teachers, daycare, etc.) in many different environments that they are normally situated in, such as at home, in the classroom, daycare, or any community activities. These observations allow for a contextu alized assessment so that any discrepancies can be used to fo rmulate a plan of intervention. relationships or in their experience of traumatic stress requires counsel ors and educators to have specialized training. This training should familiarize counselors and educators in ways of observing indicators of somatic, physiological, and behavioral symptoms in young children (Levine & Kline, 2007). This means understanding the context, such as lower SES, culturally marginalized, or in the case of attachment looking at displays of avoidance, resistance, or disorganized behaviors. When looking at traumatic stress it is important to look at non verbal behaviors that cause the c hild to avoid, be repetitive, or highly aroused thus making it difficult to work with others. These neurological, behavioral, and relational systems require specialized PAGE 159 159 attention, which has led to a new area in development for counselors working with young children ages 0 to 5 called pediatric counseling (Wolfgang et al., 2012). Pediatric Counseling Pediatric counseling is a method developed for working with young children that incorporates the basic pedagogies and assumptions of the counseling profession s uch as preventative and strength based approaches to wellness (Mellin, Hunt, & Nichola, 2011). family and community, comprising of more than play therapy, and promotes professionalism within and outside of the counseling profession (Wolfgang et al., 2012). Pediatric counseling goes further than allowing for young children ( ages 0 to 5 ) to express themselves; rather these techniques are designed to provide opport unities for movement and transformation. Pediatric counselors provide young children the opportunity to make sense of their experiences in culturally sensitive ways through sociodramatic play (McDevitt & Ormrod, 2010), creating mediated learning experience s where young children can develop the ability to self define, contextualize, and change their reality into healthy developmental journeys. By making experiences and atta chment difficulties can be addressed in the moment so that the condition does not become worse, thus ending in academic underachievement. We therefore, can move from ameliorative practices with children and underachievement, to ones of prevention and great er success. Implications for Policy Underachievement for young children with traumatic stress and/or attachment difficulties creates disadvantages in schools and it is important to intervene through education and mental health policy. One of the ways t o en hance the educational performance of young children PAGE 160 160 r equires the assistance of educational policy makers. What this study and previous studies have been finding is that there is need for greater attention on environmental factors that affect the socio emot ional development in children ( Koller & Bertel, 2006; Lockhart & Keys, 1998 ; Paylo, 2011 ) With support for these environmental factors children will be more likely to perform to their full abilities in school. Therefore, it becomes necessary for education al policy maker s to provide direct funding for school counselors within elementary schools who are specially trained to assist in th e socioemotional needs of children. It also requires that school counseling be included in the educational reform agenda (P russe, Goodnough, Donegan, & Jones, 2004). By funding school counselors in this area it will allow for more accurate assessment conceptualization, and intervention leading to better outcomes for all children thus allowing the school counselor to help th e school in educational equity and accountability Moving school practices from remediation to preventative methods. Young children experienc ing traumatic stress and/or attachment difficulties a re in need of cultural ly sensitivity and developmental ly appro priate intervention. School counselors who are trained as mental health professionals (Koller & Bertel, 2006; Lockhart & Keys, 1998; Paylo, 2011), are capable of providing i nterventions which allow for support rather than defici t based practices ; t hereby, avoiding further o ppressi ve experiences, cultural insensitivities, and other experiences that allow for disproportionate levels of traumatic stress and/or attachment difficulties and development trajectories (Ford et a l. 2002; Goodman &West Olatunji, 2010; Sue & Sue, 2013; Wycoff, 1996). This can be most efficiently achieved by providing direct funding for school counseling positions in elementary schools as well as developing an agenda that allocates job descriptions so that trained professional counselors can provide the mental health services need ed for young PAGE 161 161 children (Carey, Harrington, Martiin, & Stevenson, 2012 ; Koller & Bertel, 2006 ). The importance of having school counselors in elementary schools can impact mo re than just direct services delivery, as school counselors can serve as consulta nts for teachers and administrators around the socio emotional needs of young children This is of great need since educator's are challenge d by not only needing to teach cont ent driven academic pedagogy but also being required to create classroom environment s where all learners' needs are addressed (Koller & Bertel, 2006) Since counselors focus on developmental models, particularly school counselors trained in pediatric cou nseling, can aid teachers in developing more accurate assessments and develop more effective interventions for young children. Fundamentally by utilizing these more effective interventions the development of child ren will improve b ecause of t his important interplay between emotional health and school success (Koller & Bertel, 2006). A llow ing counselors, teachers, and administrator s to prepare and prevent exacerbated effects from child ren would thus creat e improved outcomes not only in physical health, mental health, but also educational growth. Mental Health policy makers can support the needs of young children by including them in service delivery rather than continuing the predominant focus on adult mental health services that has been the trend for more than ten years ( Lockhart & Keys, 1998). By allocating more attention and support for the use of developmentally and culturally sensitive practices for young children, it would lessen the burden now affecting adult mental health. Early interventions would lessen the negative effects of distre ss so that they do not continue into academic underachievement persistent mental health problems, or the growing juvenile ju stice difficulties for our youth ( Koller & Bertel, 2006 ) Policy makers can make it a priority for mental health PAGE 162 162 counselors who work with young children to address issues of distress earlier instead of waiting to improve a stronger collaborative e ffort (Paylo, 2011) The earlier the traumatic stress and/or attachment difficulties are assessed; it will provide opportunity for a prevention focus This would be in place of current practices that often wait until young children achieve developmental co mpetence where they can articulate their concerns. This allows counselors and educators to implement preventative interventions that emphasi ze resilience and wellness. Recommendations for Future Research Given that pediatric counseling (Wolfgang et al., 20 12) is a developing paradigm, it is important that professional counselors reach agreement regarding the definitions, procedures, and culturally sensitive interventions that make up this new area in the field. To advance our understanding of pediatric coun seling a Delphi study is recommended (Clayton, 1997; Fish & Busby, 2005; Vzquez Ramos, Leahy, & Estrada Hernndez, 2007). The Delphi methods would allow counseling researchers proficient in the area of early childhood work to create a consciousness about cross cultural issues in development in counseling young children that in working with young children in early childhood would be able to refine these experts k nowledge and opinions through subsequent reviews; with the eventual outcome being a converging consensus about pediatric counseling (Solmonson, Roaten, & Cheryl, 2010; Vzquez Ramos et al., 2007). This would provide a foundation for further developing our understanding of how traumatic stress attachment difficulties and their interactions impact young children. Delphi studies provide a flexible and time efficient way to come to consensus (Fish & Busby, 2005) and would help the shaping of the future of ped iatric counseling. This would be an PAGE 163 163 ideal way to move forward with pediatric counseling as its own entity within the professional counseling field. The results could provide valid, reliable feedback from current experts on how to address the issues of pedi atric counseling and illustrate these against the other practices utilized today. This practice could aid in the refining of the pediatric counseling competencies presented in Table 5 1 Summary of Study In sum, academic underac hievement is a significant national problem that has disproportionately affected culturally diverse students. Prior r esearch has suggested that sociodemographic factors are significant contributors towards this disproportionality, these include: (a) hegemo n y within the educational system positions culturally diverse students as (b) academic relevance or the view that education will lead to better jobs and the intersectionality of race and culture with gender has been a contributing component to the gender differences found in both reading and mathematics achievement ( Clark et al., 2008; Crumpton & Gregory, 2011; Morris, 2008), and (c) that children from poor families are marginalized, di senfranchised, and are not provided access to tools that make educational achievement possible (Biles et al., 2012; Chatterji, 2006; Pressley & Sifford, 2012 ). Additionally, researchers have found that attachment quality is significantly correlated with ac hievement outcomes particularly in preschool and elementary (Moss & St Laurent, 2001; Moss et al., 2011; Solomon & George, 2011a; Pianta & Harbers, 1996 ) It has also been suggested that difficult attachment relationships are significantly connected with t raumatic stress experiences and that both should be assessed ( Lieberman, 2004; L ieberman & Van Horn, 2008). Traumatic stress experiences have been shown to be connected with insecure attachments, transgenerational trauma, systemic oppression, as well as th e many events that can cause young children to perceive things as negative, sudden, and out of control ; PAGE 164 164 thus affecting academic achievement (Bergin & Bergin, 2009; Dass Brailsford, 2007; Danieli, 2007 ; Johnson, 1997; Moss & St Laurent, 2001; Moss et al., 2 011 ; West Olatunji & Conwill, 2011). The gap that this study fills is that these conditions were assessed together and their effects were tracked over time from kindergarten to fifth grade. No other studies reviewed had started at the kindergarten years an d looked at the relationships longitudinally towards reading and mathematics achievement The research design of this study was a longitudinal nonlinear latent growth curve model with time invariant covariates This study explored the relationships of the covariates with the mean trajectory of a national cohort of kindergarteners till fifth grade in reading and mathematics academic achievement. The significant findings of this study are that there is a significant interaction between traumatic stress sympto ms and difficult attachment symptoms that predicts individuals high in traumatic stress and difficulty with attachments show lower growth over time. Traumatic stress symptoms and difficult attachment symptoms significantly estimate lower initial values in achievement assessments and that low growth persist over time. It was also shown that difficult attachment symptoms ha ve nonlinear e ffects on the growth rate when the other covariates and interaction are controlled for. The effects of gender, SES, and cult ure show significant lower initial values and lower growth over time for males, who are poor, and are culturally diverse in the area of reading achievement. Also, gender, SES, and culture show significant lower initial values and lower growth over time for females, who are poor, and culturally diverse in mathematics achievement The significance of these findings to this research area is that they emphasize the extreme importance of considering the impact that environmental characteristics have on young chi and development PAGE 165 165 The recommendations for practice and policy involv e a strong suggestion that profession al counselors who become skilled at working with young children and in work ing within the school s ystem such as school counselors a nd other counselors trained in pediatric counseling be employed to assist young children in navigating difficult environmental circumstances. This means that educational policy makers need to make specific budget allocations to include counselors in schoo ls, that the role of these counselors need to be specifically for counseling children, and support for supervision and continuing education need to be implemented. The recommendations for future research involve the further development of the pediatric cou nseling paradigm, through the use of a Delphi study, allowing for valid, reliable, feedback from experts on how to proceed forward with the development of pediatri c counseling at its own entity. PAGE 166 166 Table 5 1. Pediatric Counseling Competencies. 1. Counselor education training incorporating a paradigm of specific training geared to children age 0 to 5 populations. 2. Specialized skills, training, and dialogue to work with children so counselors are equipped with the tools that acknowledge characteristics and cul tural nuances that are specific to child populations. a. Foundational elements of communication, with a focus on the use of audible and tactile messages, eye contact and facial recognition, and kinetic or experiential means of expression. b. Training in first a id and CPR. c. Cross culturally effective interventions. d. Empowerment of child client skills. 3. Knowledge and skill of development and the variety of developmental models a child could have within context of culture. 4. Be comfortable with differences in terms o f culture, beliefs, individual, and social experiences of where a child is in their development. 5. Recognize the impact of human development with communication and gestures that are 6. Knowledge abou t own developmental cultural beliefs and influences while helping to identify strengths and resources of young children ages 0 to 5. 7. Consultation with colleagues, research, and people beyond family/caregiver within the peers, etc.) on the developmental process, cultural trends, school and federal perspectives of development, implications of how child and 8. Focus on interventions geared toward young child ren (ages 0 to 5) that incorporate direct service delivery and secondary family/caregiver support services. 9. define identity, contextualization experiences, and transform their reality. 10. Advocate for the development, exploration of new paths to development, and culture centered treatment plans with children; thus, creating further applications of multimodal therapies within pediatric counseling PAGE 167 167 REFERENCES Ahmed, S., Wilson, K. B., Henriksen, R. C., & Jones, J. W. (2011). What does it mean to be a culturally competent counselor?. Journal for Social Action in Counseling & Psychology, 3 (1), 17 28. Amatea, E. S., & West Olatunji, C. A. (2007). Joining the conversation about educating ou r poorest children: Emerging leadership roles for school counselors in high poverty schools. Professional School Counseling, 11 (2), 81 89. American Educational Research Association., American Psychological Association., National Council on Measurement in E ducation., & Joint Committee on Standards for Educational and Psychological Testing (U.S.). (1999). Standards for educational and psychological testing Washington, DC: American Educational Research Association. American Psychiatric Association. (2000). Di agnostic and statistical manual of mental disorders (4th ed., text rev.). Washington, DC: Author. Andreassen, C., & West, J. (2007). Measuring socioemotional functioning in a national birth cohort study. Infant Mental Health Journal, 28 (6), 627 646. DOI: 1 0.1002/imhj.20157 Ang, R. P. (2005) Development and validation of the teacher student relationship inventory using exploratory and confirmatory factor analysis. The Journal of Experimental Education, 74 (1), 55 74. DOI: 10.3200/JEXE.74.1.55 74 Appleyard, K. Yang, C., & Runuam, D. K. (2010). Delineating the maladaptive pathways of child maltreatment: A mediated moderation analysis of the roles of self perception and social support. Developmental & Psychopathology, 22 337 352. DOI: 10.1017/S095457941000009X Asparouhov, T. & Muthn, B. (2009). Exploratory structural equation modeling. Structural Equation Modeling: A Multidisciplinary Journal, 16 (3), 397 438. DOI: 10.1080/1070551090300 8204 Bagnato, S. J. (2006). Of helping and measuring for early childhood intervention: Reflections on issues and school psychology's role. School Psychology Review, 35 (4), 615 620. Bailey, H. N., Moran, G., & Pederson, D. R. (2007). Childhood maltreatment complex trauma symptoms, and unresolved attachment in an at risk sample of adolescent mothers. Attachment & Human Development, 9 (2), 139 161. DOI: 10.1080/14616730701349721 Belfer, M. L. (2008), Child and adolescent mental disorders: The magnitude of the problem across the globe. Journal of Child Psychology & Psychiatry, 49 (3), 226 236. DOI: 10.1111/j.1469 7610.2007.01855.x Bemak, F. & Chung, R. C. Y. (2011). Applications in social justice counselor training: Classroom without walls. Jo urnal of Humanistic Counseling, 50 (2), 204 219. DOI: 10.1002/j.2161 1939.2011.tb00119.x PAGE 168 168 Bergin, C., & Bergin, D. (2009). Attachment in the classroom. Educational Psychology Review, 21 (2), 141 170. DOI: 10.1007/s10648 009 9104 0 Biles, K., M phande Finn, J., & Stroud, D. (2012). Social class and the schools: Beyond Ruby Payne. In D. C. Sturm & D. M. Gibson (Eds.), Social class and the helping professions: (pp. 151 163). NY: Ro utledge Talyor & Francis Group. Blanco, P. J., & Ray, D. C. (2011). Play Therapy in Elementary Schools: A best practice for improving academic achievement. Journal of Counseling & Development, 89 (2), 235 243. Blaustein, M., & Kinniburgh, K. M. (2010). Trea ting traumatic stress in children and adolescents: How to foster resilience through attachment, self regulation, and competency NY: Guilford Press. Bogat, G. A., Levendosky, A. A., von Eye, A., & Davidson, W. S., II (2011). Effects of intimate partner vio lence on the attachment relationship between mother and child: Data from a longitudinal study beginning during pregnancy. In S. A. Graham Bermann & A. A. Levendosky (Eds.), How intimate partner violence affects children: Developmental research, case studie s, and evidence based intervention (pp. 19 46). Washington, DC: American Psychological Association. Bollen, K. A., & Curran, P. J. (2006). Latent curve models: A structural equation perspective Hoboken, NJ: John Wiley & Sons, Inc. Brennan, L. M., Shaw, D. S., Dishion, T. J., & Wilson, M. (2012). Longitudinal predictors of school age academic achievement: Unique contributions of toddler age aggression, oppositionality, inattention, and hyperactivity. Journal of Abnormal Child Psychology, 40 (8), 1289 1300. DOI: 10.1007/s10802 012 9639 2 Briggs Gowan, M. J., Carter, A. S., Clark, R., Augustyn, M., McCarthy, K. J., & Ford, J. D. (2010). Exposure to potentially traumatic events in early childhood: Differential links to emergent psychopathology. Journal of Child Psychology & Psychiatry, 51 (10), 1132 1140. DOI: 10.1111/j.1469 7610.2010.02256.x Bronfenbrenner, U. (1986). Ecology of the family as a context for human development: Research perspectives. Developmental Psychology, 22 (6), 723 742. DOI: 10.1037/0012 1649.22.6.723 Brooks Gunn, J., Berlin, L. J., & Fuligni, A. S. (2000). Early childhood intervention programs: What about the family?. In J. P. Shonkoff & S. J. Meisels (Eds.), Handbook of Early Childhood Interven tions (pp. 549 577). NY: Cambridge University Press. Brunet, A., Akerib, V., & Birmes, P. (2007). Don't throw out the baby with the bathwater (PTSD is not overdiagnosed). Canadian journal of psychiatry, 52 (8), 501 502. PAGE 169 169 Bryant, F. B., & Satorra, A. (2012). Principles and practices of scaled difference chi square testing. Structural Equation Modeling: A Multidisciplinary Journal, 19 (3), 327 398. DOI: 10.1080/10705511.2012.687671 Bryant Davis, T., & Ocampo, C. (2005). The Trauma of Raci sm: Implications for Counseling, Research, and Education. Counseling Psychologist 33 (4), 574 578. DOI: 10.1177/0011000005276581 Buhin, L., & Vera, E. M. (2009). Preventing racism and promoting social justice: Person centere d and environment centered interventions. The Journal of Primary Prevention, 30 (1), 43 59. DOI: 10.1007/s10935 008 0161 9 Burstow, B. (2003). Toward a radical understanding of trauma and trauma work. Violence Against Women 9 (11), 1293 1317. DOI: 10.1177/10 77801203255555 Burstow, B. (2005). A critique of posttraumatic stress disorder and the DSM. Journal of Humanistic Psychology, 45 (4), 429 445. DOI: 10.1177/0022167805280265 Carey, J., Harrington, K., Martin, I., & Stevenson, D. (2012). A statewide evalua tion of the outcomes of the implementation of ASCA national model school counseling programs in Utah high schools. Professional School Counseling 16 (2), 89 99. Carlson, E. B. (1997). NY: The Guildford Press. Carls on, S. A., Fulton, J. E., Lee, S. M., Maynard, L. M., Brown, D. R., Kohl III, H. W., & Dietz, W. H. (2008). Physical education and academic achievement in elementary school: Data from the early childhood longitudinal study. American Journal of Public Healt h, 98 (4), 721 727. DOI: 10.2105/AJPH.2007.117176 Carter, R. T. (2007). Racism and psychological and emotional injury recognizing and assessing race based traumatic stress. The Counseling Psychologist 35 (1), 13 105. DOI: 10.1177/0011 000006292033 Carter, R. T., & Forsyth, J. M. (2009). A guide to the forensic assessment of race based traumatic stress reactions. Journal of the American Academy of Psychiatry & the Law, 37 (1), 28 40. Carter, R. T., Mazzula, S., Victoria, R., Vazquez, R., Initial development of the race based traumatic stress symptom scale: Assessing the emotional impact of racism. Psychological Trauma: Theory, Research, Practice, and Policy, 5 (1), 1 9. DOI: 0.1037/a0025911 Casan ueva, C., Goldman Fraser, J., Ringeisen, H., Lederman, C., Katz, L., & Osofsky, J. (2010). Maternal Perceptions of temperament among infants and toddlers investigated for maltreatment: Implications for services need and referral. Journal of Family Violence 25 (6), 557 574. DOI: 10.1007/s10896 010 9316 6 PAGE 170 170 Cham, H., West, S. G., Ma, Y., & Aiken, L. S. (2012). Estimating latent variable interactions with nonnormal observed data: A comparison of four approaches. Multivariate Behavioral Research 47 (6), 840 876. DOI: 10.1080/00273171.2012.732901 Chapin, J. R. (2006). The achievement gap in social studies and science starts early: Evidence from the Early Childhood Longitudinal Study. The Social Studies, 97 (6), 231 238. DOI: 10.3200/TSSS.97.6.231 238 Chat terji, M. (2006). Reading achievement gaps, correlates, and moderators of early reading achievement: Evidence from the Early Childhood Longitudinal Study (ECLS) kindergarten to first grade sample. Journal of Educational Psychology 98 (3), 489 507. DOI: 10.1037/0022 0663.98.3.489 Chilisa, B. (2012 ). Indigenous research methodologies CA: Sage Publications, Inc. Cholewa, B. & West Olatunji, C. (2008). Exploring the relationship among cultural discontinuity, psychological distress, and a cademic outcomes with low income, culturally diverse students. Professional School Counseling. 12 (1), 54 61. Choudhuri, D. D., Santiago Rivera, A., & Garrett, M. T. (2012). Counseling & diversity CA: Brooks/Cole. Chronister, K. M., McWhirter, B. T., & Ker ewsky, S. D. (2004). Counseling and ecological prevention practice. In R. K. Conyne & E. P. Cook (Eds.), Ecological counseling: An innovative approach to conceptualizing person environment interaction (pp. 315 338) Alexandria, VA: American Counseling Asso ciation. Chu, S. (2011). Perspectives in understanding the schooling and achievement of students from culturally and linguistically diverse backgrounds. Journal of Instructional Psychology, 38 (3/4), 201 209. Clark, M. A., Flower, K., Walton, J., & Oakley, E. (2008). Tackling male underachievement: Enhancing a strengths based learning environment for middle school boys. Professional School Counseling, 12 (2), 127 132. Clayton, M. J. (1997). Delphi: A technique to harness expert opinion for critical decision making tasks in education. Education Psychology: An International Journal of Experimental Educational Psychology, 17 (4), 373 386. DOI: 10.1080/0144341970170401 Comstock, D. L., Hammer, T. R., Strentzsch, J., Cannon, K., Parsons, J., & Salazar, G. (2008). R elational cultural theory: A framework for bridging relational, multicultural, and social justice competencies. Journal of Counseling & Development, 86 (3), 279 287. DOI: 10.1002/j.1556 6678.2008.tb00510.x Constantine, M. G., Myers, L., Ki ndaichi, M., & Moore III, J. L. (2004). Exploring indigenous mental health practices: The roles of healers and helpers in promoting well being in people of color. Counseling & Values, 48 (2), 110 125. PAGE 171 171 Conyne, R. K., & Cook, E. P. (2004). Understanding perso ns within environments: An introduction to ecological counseling. In R. K. Conyne & E. P. Cook (Eds.), Ecological counseling: An innovative approach to conceptualizing person environment interaction (pp. 3 35) Alexandria, VA: American Counseling Associati on. Cook, E. P., Conyne, R. K., Savaeau, C., & Tang, M. (2004). The process of ecological counseling. In R. K. Conyne & E. P. Cook (Eds.), Ecological counseling: An innovative approach to conceptualizing person environment interaction (pp. 109 140) Alexa ndria, VA: American Counseling Association. Crethar, H. C., Torres Rivera, E., & Nash, S. (2008). In search of common threads: Linking multicultural, feminist, and social justice counseling paradigms. Journal of Counseling & Development, 86 (3), 269 278. D OI: 10.1002/j.1556 6678.2008.tb00509.x Crumpton, H., & Gregory, A. (2011). I'm not learning: The role of academic relevancy for low achieving students. The Journal of Educational Research 104 (1), 42 53. DOI: 10.1080/002206709035673 98 Cutrona, C. E., Cole, V., Colangelo, N., Assouline, S. G., & Russell, D. W. (1994). Perceived parental social support and academic achievement: An attachment theory perspective. Journal of Personality and Social Psychology, 66 (2), 369 378. Danieli, Y. ( 1984). Psychotherapist's participation in the conspiracy of silence about the holocaust. Psychoanalytic Psychology, 1 (1), 23 42. DOI: 10.1037/0736 9735.1.1.23 Danieli, Y. (1998). Introduction: History and conceptual foundations. In Y. Danieli (Ed.), Inter national handbook of multigenerational legacies of trauma (pp. 1 17). NY: Kluwer Academic/Plenum Publishing Corporation. Danieli, Y. (2007). Assessing trauma across cultures from a multigenerational perspective. In J. P. Wilson & C. S K. Tang (Eds.), Cross cultural assessment of psychological trauma and PTSD (pp. 65 89). NY: Springer. Danieli, Y. (2009). Massive trauma and the healing role of reparative justice. Journal of Traumatic Stress, 22 (5), 351 357. DOI: 10.1002/jts.20441 Darney, D., Reinke, W. M., H erman, K. C., Stormont, M., & Ialongo, N. S. (2013) Children with co occurring academic and behavior problems in first grade: Distal outcomes in twelfth grade. Journal of School Psychology, 51 (1), 117 128 DOI: 10.1016/j.jsp.2012.09.005 Dass Brailsford, P. (2007). A practical approach to trauma: Empowering interventions CA: Sage Publications, Inc. (2011). Building the capacity of family day care educators to promote and emotional wellbeing: an exploratory cluster randomised controlled trial. BMC Public Health, 11 (1), 842 849. DOI: 10.1186/1471 2458 11 842 PAGE 172 172 Dawson, A. B., & Williams, S. A. (2008). The impact of language status as an acculturative str essor on internalizing and externalizing behaviors among Latino/a children: A longitudinal analysis from school entry through third grade. Journal of Youth & Adolescence, 37 (4), 399 411. DOI: 10.1007/s10964 007 9233 z De Young, A. C., Kenardy, J. A., & Cob ham, V. E. (2011a). Diagnosis of posttraumatic stress disorder in preschool children. Journal of Clinical Child & Adolescent Psychology, 40 (3), 375 384. DOI: 10.1080/15374416.2011.563474 De Young, A. C., Kenardy, J. A., & Cobham, V. E. (2011b). Trauma in e arly childhood: A neglected population. Clinical Child & Family Psychology Review, 14 (3), 231 250. DOI: 10.1007/s10567 011 0094 3 Dearing, E., McCartney, K., & Taylor, B. (2009). Does higher quality early child care promote low and reading achievement in middle childhood?. Child Development, 80 (5), 1329 1349. DOI: 10.1111/j.1467 8624.2009.01336.x Dekel, R., & Goldblatt, H. (2008). Is there intergenerational transmission of trauma? The case of combat veterans' children. American Journal of Orthopsychiatry 78 (3), 281 289. DOI: 10.1037/a0013955 Attachment & Human Development, 10 ( 4), 395 414. DOI: 10.1080/14616730802461425 Dimitruk, P., Schermelleh Engel, K., Kelava, A., & Moosbrugger, H. (2007). Challenges in nonlinear structural equation modeling. Methodology: European Journal of Research Methods for the Behavioral & Social Scien ces, 3(3) 100 114. DiPerna, J. C., Lei, P W., & Reid, E. E. (2007). Kindergarten predictors of mathematical growth in the primary grades: An investigation using the early childhood longitudinal study kindergarten cohort. Journal of Educational Psychology, 99 (2), 369 379. DOI: 10.1037/0022 0663.99.2.369 Dollarhide, C. T. (2012), Humanism is alive and well: A review of humanistic perspectives on contemporary counseling issues. Journal of Humanistic Counseling, 51 (1): 2 5. DOI:10.1002/j.21 61 1939.2012.00001.x Duncan, T. E., Duncan, S. C., & Strycker, L. A. (2009). An introduction to latent variable growth curve modeling: Concepts, issues, and applications (2nd ed.). Mahwah, NJ: Lawrence Erlbaum Associates, Inc. Duran, E., Duran, B., Yellow Horse Brave Heart, M., & Yellow Horse Davis, S. (1998). Healing the American Indian soul wound. In Y. Danieli (Ed.), International handbook of multigenerational legacies of trauma (pp. 341 354). NY: Plenum. PAGE 173 173 Duran, E., Firehammer, J., & Gonzalez, J. (2008). Liberation psychology as the path toward healing cultural soul wounds. Journal of Counseling & Development, 86 (3), 288 295. DOI: 10.1002/j.1556 6678.2008.tb00511.x Egger, H. L., & Angold, A. (2009). Classification of psychopathology in early childhood. In C. H. Zeanah, Jr., (Ed), Handbook of infant mental health (3rd ed., pp. 285 300). NY: The Guilford Press. Erdman, P., & Lampe, R. (1996). Adapting basic skills to counsel children. Journal of Counseling and Development, 74 374 377. Eriksen, K., & Kr ess, V. E. (2006). The DSM and the professional counseling identity: Bridging the gap. Journal of Mental Health Counseling, 28 (3), 202 217. Felsen, I. (1998). Transgenerational transmission of effects of the Holocaust. In Y. Danieli (Ed.), International ha ndbook of multigenerational legacies of trauma (pp. 43 68). NY: Plenum. Fish, L. S., & Busby, D. M. (2005). The Delphi method. In D. H. Sprenkle & F. P. Piercy (Eds.), Research methods in family therapy (pp. 238 253). NY: The Guilford Press. Fivush, R. (20 10). Speaking silence: The social construction of silence in autobiographical and cultural narratives. Memory, 18 (2), 88 98. DOI: 10.1080/09658210903029404 Follette, V., Palm, K., & Pearson, A. (2006). Mindfulness and trauma: Implications for treatment. J ournal of Rational Emotive & Cognitive Behavior Therapy, 24 (1), 45 61. DOI: 10.1007/s10942 006 0025 2 Follette, V. M., & Vijay, A. (2009). Mindfulness for Trauma and Posttraumatic Stress Disorder. In F. Didonna (Ed.), Clinical Handbook of Mindf ulness (pp. 299 317). NY: Springer. DOI: 10.1007/978 0 387 09593 6 Ford, D. Y., Harris III, J. J., Tyson, C. A., & Trotman, M. F. (2002). Beyond deficit thinking: Providing access for gifted African American students. Roeper Review 24 (2), 52 58. Fos s, L. L., Generali, M. M. & Kress, V. E. (2011). Counseling people living in poverty: The CARE model. Journal of Humanistic Counseling, 50 (2), 161 171. DOI: 10.1002/j.2161 1939.2011.tb00115.x Foster, M. L., L. Lewis, J ., & Onafowora, L. (2003). Anthropology, culture, and research on teaching and learning: Applying what we have learned to improve practice. The Teachers College Record 105 (2), 261 277. Frazier, K. N., Goodman, R. D., St. Juste, S., & West Olatunji, C. (20 09). Transgenerational trauma and child sexual abuse: Reconceptualizing cases involving young survivors of CSA. Journal of Mental Health Counseling, 31 (1), 22 33. PAGE 174 174 Freire, P. (2000). Pedagogy of the oppressed NY: The Continuum International Publishing Gro up Inc. (Original work published 1970) Friedman, M. J., Resick, P. A., Bryant, R. A., & Brewin, C. R. (2011). Considering PTSD for DSM 5. Depression and anxiety 28 (9), 750 769. Garrett, M. T., & Portman, T. A. A. (2011). Counseling & diversity: Counseling Native Americans CA: Brooks/Cole. Gauvian, M., & Parke, R. D. (2010). Socialization. In M. H. Bornstein (Ed.), Handbook of cultural developmental science (pp. 239 258). NY: Pschology Press Taylor & Francis Group. George, C., & Solomon, J. (2008). The caregiving system: A behavioral systems approach to parenting. In J. Cassidy & P. R. Shaver (Eds.), Handbook of attachment: Theory, resea rch, and clinical applications (2nd ed., pp. 833 856). NY: The Guilford Press. Gleason, M. M., Zeanah, C. H., Jr., & Dickstein, S. (2010). Recognizing young children in need of mental health assessment: Development and preliminary validity of the early c hildhood screening assessment. Infant Mental Health Journal, 31 (3), 335 357. DOI: 10.1002/imhj.20259 Goldenberg, I., & Goldenberg, H. (2008). Family therapy: An overview Belmont, CA: Thomson Brooks/Cole. Goodman, R. (2009). Academic achievement and traumatic stress among primary school students. (Ph.D. dissertation). University of Florida, Florida. Retrieved January 27, 2013, from Dissertations & Theses. (Publication No. AAT 3467699). Goodman, R. D., Miller, M. D., & West Ola tunji, C. A. (2012). Traumatic stress, socioeconomic status, and academic achievement among primary school students. Psychological Trauma: Theory, Research, Practice, and Policy 4 (3), 252. DOI: 10.1037/a0024912 Goodman, R. D., & West Olatunji, C. A. (2008 a). Transgenerational trauma and resilience: Improving mental health counseling for survivors on Hurricane Katrina. Journal of Mental Health Counseling, 30 (2), 121 136. Goodman, R. D., & West Olatunji, C. A. (2008b). Traumatic stress, systemic oppression, and resilience in post Katrina New Orleans. Spaces for Difference: An Interdisciplinary Journal 1 (2). Retrieved from: http://escholarship.org/uc/item/4x8082z1 Goodman, R. D., & West Olatunji, C. A. (2010). Educational hegemony, traumatic stress, and Afric an American and Latino American students. Journal of Multicultural Counseling & Development, 38 (3), 176 186. Graf, A., Schiestl, C., & Landolt, M. A. (2011). Posttraumatic stress and behavior problems in infants and toddlers with burns. Journal of Pediatr ic Psychology, 36 (8), 923 931. DOI: 10.1093/jpepsy/jsr021 PAGE 175 175 Graham Bermann, S.A., & Levendosky, A. A. (1998). Traumatic stress symptoms in children of battered women. Journal of Interpersonal Violence, 13 111 129. Grimm, K. J., & Ram, N. (2009). No nlinear growth models in Mplus and SAS Structural Equation Modeling: A Multidisciplinary Journal, 16 (4), 676 701. DOI: 10.1080/10705510903206055 Grimm, K. J., & Ram, N. (2012). Growth curve modeling from an structural equation modeling perspective. In T. Little, B. Laursen, & N. Card (Eds.), Handbook of developmental research methods (pp. 411 431). New York: Guilford Publications. Grimm, K. J., Ram, N., & Hamagami, F. (2011). Nonlinear growth curves in developmental research. Child Development, 82(5), 1357 1371. DOI: 10.1111/j.1467 8624.2011.01630.x Grimm, K. J., Steele, J. S., Mashburn, A. J., Burchinal, M., & Pianta, R. C. (2010). Early behavioral associations of achievement trajectories. Developmental Psychology, 46 (5), 976 983. DOI: 10.1037/a0018878 Guzmn, M. R., & Carrasco, N. (2011). Counseling & diversity: Counseling Latino/a Americans CA: Brooks/Cole. Harrell, J. P., Hall, S., & Taliaferro, J. (2003). Physiological responses to raci sm and discrimination: An assessment of the evidence. American Journal of Public Health, 93 (2), 243 248. Hathaway, L. M., Boals, A., & Banks, J. B. (2010). PTSD symptoms and dominant emotional response to a traumatic event: an examination of DSM IV Criteri on A2. Anxiety, Stress & Coping 23 (1), 119 126. DOI: 10.1080/10615800902818771 Hazen, N. L., Jacobvitz, D., Higgins, K. N., Allen, S., & Jin, M. Y. (2011). Pathways from disorganized attachment to later social emotional problems: The role of gender and pa rent child interaction patterns. In J. Solomon & C. George (Eds.), Disorganized attachment and caregiving (pp. 167 206). NY: The Guilford Press. Hedeker, D., & Gibbons, R. D. (2006). Longitudinal data analysis Hoboken, NJ: John Wiley & Sons, Inc. Henderso n, D. A., & Thompson, C. L. (2011). Counseling children (8th ed.). CA: Brooks/Cole. Hesse, M., Main, M., Abrams, K. Y., & Rifkin, A. (2003). Unresolved states regarding loss or d frightening ideation in the offspring of traumatized, non maltreating parents. In M. F. Solomon & D. J. Siegel (Eds.), Healing trauma: Attachment, mind, body, and brain (pp 57 106). NY: W.W. Norton. Hooks, B. (2000). Where we stand: Class matters NY: R outledge. PAGE 176 176 Horwitz, S. M., Kelleher, K. J., Stein, R. E. K., Storfer Isser, A., Youngs trom, E. A., Park, E. R., & Hoagwood, E. K. (2007). Barriers to the identification and management of psychosocial issues in children and maternal depression. Pediatrics, 1 19 (1), 208 218. DOI: 10.1542/peds.2005 1997 in urban/rural areas. Journal of Health Care for the Poor & Underserved,19 (1), 237 247. DOI: 10.1353/hpu.2008.0 008 Hox, J. J. (2010). Multilevel analysis: Techniques and applications (2nd ed.). NY: Routledge. Hoyle, R. H. (1995). The structural equation modeling approach: Basic concepts and fundamental issues. In R. H. Hoyle (Ed.), Structural equation modeling: Con cepts, issues, and applications (pp. 1 15). CA: Sage Publications, Inc. Huck, S. W. (2008). Reading statistics and research (5th ed.). Boston, MA: Pearson Education, Inc. Ivey, A. E., & Ivey, M. B. (1998). Reframing DSM IV: Positive strategies from developmental counseling and therapy. Journal of Counseling & Development 76 (3), 334 350. Jackman, M. G., Leite, W. L., & Cochrane, D. J. (2011). Estimating latent variable interactions with the unconstrained approach: A com parison of methods to form product indicators for large, unequal numbers of items. Structural Equation Modeling, 18 (2), 274 288. DOI: 10.1080/10705511.2011.557342 Jackson, D. L., Gillaspy, Jr, J. A., & Purc Stephenson, R. (2009). Reporting practices in con firmatory factor analysis: An overview and some recommendations. Psychological Methods, 14 (1), 6 23. DOI: 10.1037/a0014694 Johanson, G. J. (2006). The use of mindfulness in psychotherapy. Annals of the American Psychotherapy Association, 9 15 24. Johnson, K. (1997). Trauma in the lives of children: Crisis and stress management techniques for counselors, teachers, and other professionals Alameda, CA: Hunter House, Inc. Juntunen, C. L., Barraclough, D. J., Broneck, C. L., Seibel, G. A., Winrow, S. A., & Mor in, P. M. (2001). American Indian perspectives on the career journey. Journal of Counseling Psychology, 48 (3), 274 285. DOI: 10.1037//OO22 O167.48.3.274 Juntunen, C. L., & Cline, K. (2010). Culture and self in career development: Working with Americna Indi ans. Journal of Career Development, 37 (1), 391 410. DOI: 10.1177/0894845309345846 academic performance, weight gain, and socia l skills. The Journal of Nutrition, 135 (12), 2831 2839. PAGE 177 177 Kaplan, D. (2009). Structural equation modeling: Foundations and extensions (2nd ed.). CA: SAGE Publications, Inc. Kelava, A., & Brandt, H. (2009). The extended unconstrained approach, Review of Psyc hology, 16 (2), 123 131. Kelava, A., Werner, C. S., Schermelleh Engel, K., Moosbrugger, H., Zapf, D., Ma, Y., ... & West, S. G. (2011). Advanced nonlinear latent variable modeling: Distribution analytic LMS and QML estimators of interaction and quadratic e ffects. Structural Equation Modeling: A Multidisciplinary Journal 18 (3), 465 491. DOI: 10.1080/10705511.2011.582408 Kelley, K., & Maxwell, S. E. (2008). Delineating the average rate of change in longitudinal models. Journal of Educational & Behavioral Sta tistics, 33 (3), 307 332. DOI: 10.3102/1076998607306074 Kim, B. S. K. (2011). Counseling & diversity: Counseling Asian Americans CA: Brooks/Cole. Kim, J., Cicchetti, D., Rogosch, F. A., & Manly, J. (2009). Child maltreatment and tra jectories of personality and behavioral functioning: Implications for the development of personality disorder. Development & Psychopathology, 21 (3), 889 912. DOI: 10.1017/S0954579409000480 King, K. R. (2005). Wh y is discrimination stressful? The mediating role of cognitive appraisal. Cultural Diversity & Ethnic Minority Psychology, 11 (3), 202 212. DOI: 10.1037/1099 9809.11.3.202 Klein, A., & Moosbrugger, H. (2000). Maximum likeli hood estimation of latent interaction effects with the LMS method. Psychometrika 65 (4), 457 474. Klein, A. G., & Schermelleh Engel, K. (2010). Introduction of a new measure for detecting poor fit due to omitted nonlinear terms in SEM. AStA Advances in Sta tistical Analysis 94 (2), 157 166. DOI: 10.1007/s10182 010 0130 5 Klein, T. P., Devoe, E. R., Miranda Julian, C., & Linas, K. (2009). Young children's responses to September 11th: The New York City experience. Infant Mental Health Journal 30 (1), 1 22. DOI : 10.1002/imhj.20200 Kline, R. B. (2011). Principles and practice of structural equation modeling (3rd ed.). NY: The Guilford Press. Knitzer, J., & Perry, D. (2009). Poverty and infant and toddler development: Facing the complex challenges. In C. H. Zean ah, Jr. (Ed.), Handbook of infant mental health (3rd ed., pp.135 152). NY: The Guilford Press. PAGE 178 178 Kobak, R., & Madsen, S. (2008). Disruptions in attachment bonds: Implications for theory, research, and clinical intervention. In J. Cassidy & P. R. Shaver (Ed s.), Handbook of attachment: Theory, research, and clinical applications (2nd ed., pp. 23 47). NY: Guilford Press. families, and schools: Revisiting the preservic e training and preparation of school based personnel. Education & Treatment of Children, 29 (2), 197 217. Lau, B. W. K. (2002). Does the stress in childhood and adolescence matter? A psychological perspective. The Journal of the Royal Society for the Promot ion of Health, 122 (4), 238 244. DOI: 10.1177/146642400212200411 Leite, W. L., & Stapleton, L. M. (2011). Detecting growth shape misspecifications in latent growth models: An evaluation of fit indexes. The Journal of Experimental Education, 79 (4), 361 381. DOI: 10.1080/00220973.2010.509369 Leite, W. L., & Zuo, Y. (2011). Modeling latent interactions at level 2 in multilevel structural equation models: An evaluation of mean centered and residual centered unconstrained approaches, Structural Equation Modelin g: A Multidisciplinary Journal, 18 (3), 449 464. DOI: 10.1080/10705511.2011.582400 Levine, P. A. (1997). Waking the tiger: Healing trauma Berkeley, CA: North Atlantic Books. Levine, P. A. (2010). In an unspoken voice: How the body releases trauma and rest ores goodness Berkeley, CA: North Atlantic Books. Levine, P. A., & Kline, M. (2007). miracle of healing. Berkeley, CA: North Atlantic Books. Lewis, R. E. (2012). Ecohumanism: Integrating humanism and r esilience theory. In M. B. Scholl, A. S. McGowan, & J. T. Hansen (Eds.), Humanistic perspectives on contemporary counseling issues (pp. 191 214). NY: Routledge Taylor & Francis Group. Lieberman, A. F. (1993). The emotional life of the toddler NY: Free Pr ess. Lieberman, A. F. (2004). Traumatic stress and quality of attachment: Reality and internalization in disorders of infant mental health. Infant Mental Health Journal, 25 (4), 336 351. DOI: 10.1002/imhj.20009 Lieberman, A. F., & Van Horn, P. (2008). Psychotherapy with infants and young children: Repairing the effects of stress and trauma on early attachment. NY: The Guilford Press. Lim, S., Lim, B. K. H., Michael, R., Cai, R., & Schock, C. K. (2010). The trajectory of counse ling in China: Past, present, and future trends. Journal of Counseling & Development, 88 (1), 4 8. DOI: 10.1002/j.1556 6678.2010.tb00141.x PAGE 179 179 Little, T. D., Bovaird, J. A., & Widaman, K. F. (2006). On the merits of orthogonalizing powered and product terms: Im plications for modeling interactions among latent variables. Structural Equation Modeling: A Multidisciplinary Journal, 13 (4), 497 519. DOI: 10.1207/s15328007sem1304_1 Liu, W. M., & Estrada Hernndez, N. (2010). Counselin g and advocacy for individuals living in poverty. In M. J. Ratts, R. L. Toporek, & J. A. Lewis (Eds.), ACA advocacy competencies: A social justice framework for counselors (pp. 43 53). VA: American Counseling Association. Lockhart, E. J., & Keys, S. G. (19 98). The mental health counseling role of school counselors. Professional School Counseling, 1 (4), 3 6. Lopez, S. J., Edwards, L. M., Pedrotti, J. T., Prosser, E. C., LaRue, S., Spalitto, S. V., & Ulven, J. C. (2006). Beyond the DSM IV: Assumptions, altern atives, and alterations. Journal of Counseling & Development 84 (3), 259 267. DOI: 10.1002/j.1556 6678.2006.tb00404.x Lounds, J. J., Borkowski, J. G., Whitman, T. L., Maxwell, S. E., & Weed, K. (2005). Adolescent parenting and attachment during infancy and early childhood. Parenting: Science & Practice, 5 (1) 91 118. DOI: 10.1207/s15327922par0501_4 Lowe, S. M., Okubo, Y., & Reilly, M. F. (2012). A qualitative inquiry into racism, trauma, and coping: Implications for supporting victims of racism. Professiona l Psychology: Research & Practice, 43 (3), 190 198. DOI: 10.1037/a0026501 Lyhagen, J. (2007). Estimating nonlinear structural models: EMM and the Kenny Judd model. Structural Equation Modeling: A Multidisciplinary Journal, 14(3), 391 403. DOI: 10.1080/10705510701301487 Marsh, H. W., Muthn, B., Asparouhov, T., Ldtke, O., Robitzsch, A., Morin, A. J. S., & Trautwein, U. (2009). Exploratory structural equation modeling, integrating CFA and university teaching. Structural Equation Modeling: A Multidisciplinary Journal, 16 (3), 439 476. DOI: 10.1080/10705510903008220 Marsh, H. W., Wen, Z., & Hau, K T. (2004). Structural equation models o f latent interactions: Evaluation of alternative estimation strategies and indicator construction. Psychological Methods 9 (3), 275 300. DOI: 10.1037/1082 989X.9.3.275 Marsh, H. W., Wen, Z., & Hau, K T. (2006). Structural equation models of latent interac tion and quadratic effects. In G. R. Hancock & R. O. Mueller (Eds.), Structural equation modeling: A second course (pp. 225 265). Greenwich, CN: Information Age Publishing. Marvin, R. S., & Britner, P. A. (2008). Normative development: The ontogeny of atta chment. In J. Cassidy & P. R. Shaver (Eds.), Handbook of attachment: Theory, research, and clinical applications (2nd ed., pp. 269 294). NY: Guilford Press. PAGE 180 180 Masten, A. S., & Narayan, A. J. (2012). Child development in the context of disaster, war, and terr orism: Pathways of risk and resilience. Annual Review of Psychology, 63 (1), 227 257. DOI: 10.1146/annurev psych 120710 100356 Mayer, A., Steyer, R., & Mueller, H. (2012). A general approach to defining latent growth components. Structural Equation Modelin g: A Multidisciplinary Journal, 19 (4), 513 533. DOI: 10.1080/10705511.2012.713242 McCubbin, L. D., & McCubbin, H. I. (2005). Culture and ethnic identity in family resilience: Dynamic processes in trauma and transformation of indigenous people. In M. Ungar (Ed.), Handbook for working with children and youth pathways to resilience across cultures and contexts (pp. 27 44) CA: Sage Publications, Inc. McDevitt, T. M., & Ormrod, J. E. (2010). Child development and education. Upper Saddle River, NJ: Pe arson Education, Inc. McKinney, K. (2007), "Breaking the conspiracy of silence": Testimony, traumatic memory, and psychotherapy with survivors of political violence. Ethos, 35 (3): 265 299. DOI: 10.1525/eth.2007.35.3.265 Mellin, E. A., Hunt, B., & Nichols, L. M. (2011). Counselor professional identity: Findings and implications for counseling and interprofessional collaboration. Journal of Counseling & Development, 89 (2), 140 147. Metzger, S., Erdman, P., & Ng, K. M. (2010). Attachm ent in cultural contexts. In P. Erdman & K. M. Ng (Eds.), Attachment: Expanding the cultural connections (pp. 3 12). NY: Routledge Taylor & Francis Group. Modrowski, C. A., Miller, L. E., Howell, K. H., & Graham Bermann, S. A. (2013). Consistency of trauma symptoms at home and in therapy for preschool children exposed to intimate partner violence. Psychological Trauma: Theory, Research, Practice, and Policy, 5 (3), 251 258. DOI: 10.1037/a0027167 Mongillo, E. A., Briggs Gowan, M., Ford, J. D., & Carter, A. S. (2009). Impact of traumatic life events in a community sample of toddlers. Journal of Abnormal Child Psychology, 37 (4), 455 468. DOI: 10.1007/s10802 008 9283 z communicative gestures. In R. A. Lanius, E. Vermetten, & C. Pain (Eds.), The impact of early life trauma on health and disease: The hidden epidemic (pp.103 111). NY: Cambridge University Press. Moore, I. L., Henfield, M. S., & Owens, D. (2008). African Ame rican males in special education: Their attitudes and perceptions toward high school counselors and school counseling services. American Behavioral Scientist 51 (7), 907 927. DOI: 10.1177/0002764207311997 PAGE 181 181 Moore, J. L., Ford, D. Y., & Milner, H. R. (2005). Underachievement among gifted students of color: Implications for educators. Theory Into Practice, 44 (2), 167 177. DOI: 10.1207/s15430421tip4402_11 ithmetic: Social class, masculinity, and schooling in a rural context. Gender & Society, 22 (6), 728 751. DOI: 10.1177/0891243208325163 Moss, E., Bureau, J F., St Laurent, D., & Tarabulsy, G. M. (2011). Underst anding disorganized attachment at preschool and school age: Examining divergent pathways of disorganized and controlling children. In J. Solomon & C. George (Eds.), Disorganized attachment and caregiving (pp. 52 79). NY: The Guilford Press. Moss, E., & St Laurent, D. (2001). Attachment at school age and academic performance. Developmental Psychology, 37 (6), 863 874. DOI: 10.1037//0012 I649.37.6.863 Muthn, L. K., & Muthn, B. O. (2009, March 23). MLR acronym? [Online Mplus Structural Equation Modeling discu ssion forum comment]. Retrieved from http://www.statmodel.com/discussion/messages/11/2156.html?1279899364 Muthn, L. K. Muthn, B. O. ( 2010 ). Mplus Users' Guide (6th ed.) CA : Muthen & Muthen. Nassar McMillan, S. C. (2011). Counseling & diversity: Coun seling Arab Americans CA: Brooks/Cole. National Center for Education Statistics. (2001). K base year public use data files and electronic codebook [NCES 2001 029 (revised)]. Washington, DC: US Department of Education. National Center for Education Statistics (2011a). 2011 (NCES 2012 458). Institute of Education Sciences, U.S. Department of Education, Washington, D.C. National Center for Education Statistics (2011b). Reading 2011 (NCES 2012 457). Institute of Education Sciences, U.S. Department of Education, Washington, D.C. N ational Center for Education Statistics (2012). Progress 2012 (NCES 2013 456 ). Institute of Educatio n Sciences, U.S. Department of Education, Washington, D.C. Nuru Jeter, A. M., Sarsour, K., Jutte, D. P., & Boyce, W. T. (2010). Socioeconomic predictors of health and development in middle childhood: Variations by socioeconomic status measure and race. Iss ues in Comprehensive Pediatric Nursing, 33(2) 59 81. DOI: 10.3109/01460861003663953 PAGE 182 182 relationship between poverty and special edu cation placement. Educational Researcher 35 (6), 6 11. DOI: 10.3102/0013189X035006006 Paniagua, F. A. (2001). Diagnosis in a multicultural context: A casebook for mental health professionals CA: Sage Publications. Paradies, Y. (2006). A systematic review of empirical research on self reported racism and health. International Journal of Epidemiology, 35 (4), 888 901. DOI: 10.1093/ije/dyl056 Paylo, M. J. (2011). Preparing school counseling students to aid families: Integrating a family systems perspective. Th e Family Journal, 19 (2), 140 146. DOI: 10.1177/1066480710397130 Peluso, P. R., Miranda, A. O., Firpo Jimenez, M., & Pham, M. T. (2010). Attachment dynamics and Latin cultures: Areas of convergence and divergence In P. Erdman & K. Ng (Eds.), Attachment: Expanding the cultural connections (pp. 281 295), NY: Routledge Taylor & Francis Group. Perlman, S., & Doyle, A. (2012). Trauma experienced in early childhood. In L. L. Levers (Ed.), Trauma counseling: Theories an d interventions (pp. 132 145). NY: Springer Publishing Company, LLC. Prusse, R., Goodnough, G. E., Donegan, J., & Jones, C. (2004). Perceptions of school counselors and school principals about the national standards for school counseling programs and the transforming school counseling initiative. Professional School Counseling, 7 (3), 152 161. Pianta, R. C., & Harbers, K. L. (1996). Observing mother and child behavior in a problem solving situation at school entry: Relations with academic achievement. Jour nal of School Psychology, 34 (3), 307 322. Pollack, J., Atkins Burnett, S., Najarian, M., and Rock, D. (2005). Early Childhood Longitudinal Study, Kindergarten Class of 1998 99 (ECLS K), Psychometric Report for the Fifth Grade (NCES 2006 036rev). Washingto n, DC: National Center for Education Statistics. Ponterotto, J. G. (2005). Qualitative research in counseling psychology: A primer on research paradigms and philosophy of science. Journal of Counseling Psychology, 52 (2), 126 136. DOI: 10.1037/0022 0167.52. 2.126 Pressley, C., & Sifford, A. (2012). Poverty: Urban and rural. In D. C. Sturm & D. M. Gibson (Eds.), landscape of class in America (pp. 35 50). NY: Routledge Talyor & Fran cis Group. Raphael, B., Swan, P., & Martinek, N. (1998). Intergenerational aspects of trauma for Australian Aboriginal people. In Y. Danieli (Ed.), International handbook of multigenerational legacies of trauma (pp. 327 339). NY: Plenum. PAGE 183 183 Raudenbush, S. W., Brennan, R. T., & Barnett, R. C. (1995). A multivariate hierarchical model for studying psychological change within married couples. Journal of Family Psychology, 9 (2), 161. Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models: Applications and data analysis methods (2nd ed.). CA: Sage Publications, Inc. Raudenbush, S. W., Bryk, A. S., Cheong, Y. F., Condon, Jr. R. T., & du Toit, M. (2011). HLM 7: Hierarchical linear and nonlinear modeling Lincolnwood, IL: Scientific Software International, Inc. Raykov, T., & Marcoulides, G. A. (2011). Introduction to psychometric theory NY: Routledge Taylor & Francis Group in kindergarten literacy learning: Do cl assroom behaviors make a difference?. The Elementary School Journal, 106 (1), 21 38. DOI: 10.1086/496905 Robinson, J., & Biran, M. (2006). Discovering self: Relationships between African identity and academic achievement. Journal of Black Studies, 37 (1), 4 6 68. DOI: 10.1177/0021934704273149 Rose, J. & Glass, N. (2008). The importance of emancipatory research to contemporary nursing practice. Contemporary Nurse, 29 (1), 8 22. Rosenblum, K. L., Dayton, C. J., & Muzik, M. (2009). Infant social and emotional development: Emerging competence in relational context. In C. H. Zeanah, Jr. (Ed.), Handbook of infant mental health (3rd ed., pp.80 103). NY: The Guilford Press. Rothschild, B. (2000). The body remembers: The psycho physiology of trauma and trauma treatment NY: Norton. Roysircar, G. & Pignatiello, V. (2011). A multicultural ecological assessment tool: Conceptualization and practice with an Asian Indian immigrant woman. Journal of Multicultural Counseling & Developmen t, 39 (3), 167 179. DOI: 10.1002/j.2161 1912.2011.tb00149.x Rubin, K. H. (1998). Social and emotional development from a cultural perspective. Developmental Psychology, 34 (4), 611 615. Sapolsky, R. M. (20 04). (3rd ed.). New York: Henry Holt and Co. Scaer, R. C. (2001). The body bears the burden: Trauma, dissociation, and disease Binghamton, NY: The Haworth Press, Inc. Scheeringa, M. S., Zeanah, C. H., & Cohen, J. A. (2011). PTS D in children and adolescents: Toward an empirically based algorithma. Depression & Anxiety 28 (9), 770 782. DOI: 10.1002/da.20736 PAGE 184 184 Schermelleh Engel, K., Klein, A., & Moosbrugger, H. (1998). Estimating nonlinear effects using a latent moderate d structural equations approach. In R. E. Schumacker & G. A. Marcoulides (Eds.), Interaction and nonlinear effects in structural equation modeling (pp. 203 238). Mahwah, NJ: Lawrence Erlbaum Associates, Inc. Schwerdtfeger, K. L., & Goff, B. S. N. (2007) Intergenerational transmission of trauma: Exploring mother infant prenatal attachment. Journal of Traumatic Stress, 20 (1), 39 51. DOI: 10.1002/jts.20179 Seccombe, K. (2002). "Beating the odds" versus "changing the odds": Poverty, resilience, and family policy. Journal of Marriage & Family, 64 (2), pp. 384 394. Seto, A., Young, S., Becker, K. W., & Kiselica, M. S. (2006). Application of the triad training model in a multicultural counseling course. Counselor Education & Supervision, 45 (4), 304 318. DOI: 1 0.1002/j.1556 6978.2006.tb00006.x Sheperis, C. J., Renfro Michel, E. L., & Doggett, R. A. (2003). In home treatment of reactive attachment disorder in a therapeutic foster care system: A case example. Journal of Mental Health Counseling, 25 (1), 76 88. Shin T., Davison, M. L., Long, J. D., Chan, C. K., & Heistad, D. (2013). Exploring gains in reading and mathematics achievement among regular and exceptional students using growth curve modeling. Learning & Individual Differences, 23 DOI: 10.1016/j.lindif.20 12.10.002 Siegel, D. J. (2003). An interpersonal neurobiology of psychotherapy: The developing mind and the resolution of trauma. In M. F. Solomon & D. J. Siegel (Eds.), Healing trauma: Attachment, mind, body, and brain (pp. 1 56), NY: W. W. Norton & Compa ny, Inc. Siegel, D. J. (2012). The developing mind: How relationships and the brain interact to shape who we are (2nd ed.). NY: London: Guilford Press. Simons, R. L., & Johnson, C. (1998). An examination of competing explanations for the intergenerational transmission of domestic violence. In Y. Danieli (Ed.), International handbook of multigenerational legacies of trauma (pp. 553 570). NY: Plenum. Smith, L. T. (1999). Decolonizing methodologies: Research and indigenous peoples NY: Zed Books Ltd. Snijders T. A. B., & Bosker, R. J. (2012). Multilevel analysis: An introduction to basic and advanced multilevel modeling (2nd ed.). CA: Sage Publications, Inc. Solmonson, L., L., Roaten, G., & Cheryl, S. (2010). Emerging concerns regarding the hiring practices o f school counselors: A delphi study. Journal of Professional Counseling: Practice, Theory & Research, 38 (2), 37 51. PAGE 185 185 Solomon, J., & George, C. (2008). The measurement of attachment security and related constructs in infancy and early childhood. In J. Cass idy & P. R. Shaver (Eds.), Handbook of attachment: Theory, research, and clinical applications (2nd ed., pp. 383 416). NY: Guilford Press. Solomon, J., & George, C. (2011a). The disorganized attachment caregiving system: dysregulation of adaptive processes at multiple levels. In J. Solomon & C. George (Eds.), Disorganized attachment and caregiving (pp. 3 24). NY: The Guilford Press. Solomon, J., & George, C. (2011b). Disorganization of maternal caregiving across two generations: The origins of caregiving he lplessness. In J. Solomon & C. George (Eds.), Disorganized attachment and caregiving (pp. 25 51). NY: The Guilford Press. Summers & R. Chazan Cohen (Eds.), Understand ing early childhood mental health (pp.99 123). Baltimore, MD: Paul H. Brooks Publishing Co. Staton, A. R., Evans, W., & Lucey, C. (2012). Understanding social class in the United States. In D. C. Sturm & D. M. Gibson (Eds.), Social class and the helping pr ofessions: A (pp. 17 34). NY: Routledge Talyor & Francis Group. Straus, M. B. (1999). No talk therapy for children and adolescents NY: W.W. Norton. Sturm, D. C., & Slaughter, E. (2012). Mid dle class America. In D. C. Sturm & D. M. Gibson (Eds.), landscape of class in America (pp. 51 68). NY: Routledge Talyor & Francis Group. Sue, D. W. (2010). Microaggressions in everyday life: Race, gender, and sexual orientation Hoboken, N.J: John Wiley & Sons, Inc. Sue, D. W., Capodilupo, C. M., Torino, G. C., Bucceri, J. M., Holder, A., Nadal, K. L., & Esquilin, M. (2007). Racial microaggressions in everyday life: implication s for clinical practice. American Psychologist 62 (4), 271 286. DOI: 10.1037/0003 066X.62.4.271 Sue, D. W., & Sue, D. (2013). Counseling the culturally diverse: Theory and practice (6 th ed.). Hoboken, N.J: John Wiley & Sons, Inc. Suizzo, M., Chen, W C., Ch eng, C C., Liang, A. S., Contreras, H., Zanger, D., & Robinson, C. (2008). Parental beliefs about young children's socialization across US ethnic groups: Coexistence of independence and interdependence. Early Child Development & Care, 178 (5), 467 486. DOI: 10.1080/03004430600823917 Thum, Y. M. (1997). Hierarchical linear models for multivariate outcomes. Journal of Educational & Behavioral Statistics, 22 (1), 77 108. Tillman, L. C. (2002). Culturally sensitive research approaches: An African American perspec tive. Educational Researcher, 31 (9), 3 12. DOI: 10.3102/0013189X031009003 PAGE 186 186 Tillman, L. C. (2006). Researching and writing from an African American perspective: Reflective notes on three research studies. International Journal of Qualitative Studies in Educa tion, 19 (3), 265 287. DOI: 10.1080/09518390600696513 Tomlinson, M., Murray, L., & Cooper, P. (2010). Attachment theory, culture, and Africa: Past, present, and future. In P. Erdman & K. Ng (Eds.), Attachment: Expanding the cultural connections (pp181 194). NY: Routledge. Tourangeau, K., L, T., and Nord, C. (2005 ). Early Childhood Longitudinal Study, Kindergarten Class of 1998 99(ECLS K), Fifth Grade Methodology Report (NCES 2006 037). U.S. Department of Education. Washington, DC: National Center for Educat ion Statistics. Tourangeau, K., Nord, C., L T., Pollack, J.M., and Atkins Burnett, S. (2006). Early Childhood Longitudinal Study, Kindergarten Class of 1998 99 (ECLS for the ECLS K Fifth Grade Data Files and Electronic Codebooks (NCES 2006 032). U.S. Department of Education. Washington, DC: National Center for Education Statistics. Turner, H. A., Finkelhor, D., Ormrod, R., Hamby, S., Leeb, R. T., Mercy, J. A. & Holt, M. (2012). Family context, victimization, and child trauma symp toms: Variations in safe, stable, and nurturing relationships during early and middle childhood. American Journal of Orthopsychiatry, 82 (2), 209 219. DOI: 10.1111/j.1939 0025.2012.01147.x Turner, S. L., & Pope, M. (2009). North America's Native peoples: A social justice and trauma counseling approach. Journal of Multicultural Counseling & Development, 37 (4), 194 205. U.S. Department of Education, National Center for Education Statistics (2006). ECLS K Base Year Public Use Electronic Codebook Errata. (NCES 2 001 029e). Washington, DC: Author. Usami, S. (2009). Applying Gibbs sampler for multivariate hierarchical linear model. World Academy of Science, Engineering & Technology 54 340 342. Van Ijzendoorn, M. H., & Sagi Schwartz, A. (2008). Cross cultural patt erns of attachment: Universal and contextual dimensions. In J. Cassidy & P. R. Shaver (Eds.), Handbook of attachment: Theory, research, and clinical applications (2nd ed., pp. 880 905). NY: Guilford Press. Vzquez Ramos, R., Leahy, M., & Estrada Hernndez, N. (2007). The Delphi method in rehabilitation counseling research. Rehabilitation Counseling Bulletin, 502 (2), 111 118. DOI: 10.1177/00343552070500020101 Verbos, A. K., Gladstone, J. S., & Kennedy, D. M. (2011). Native American values and man agement education: Envisioning an inclusive virtuous circle. Journal of Management Education, 35 (1), 10 26. DOI: 10.1177/1052562910384364 PAGE 187 187 Vernon, A., & Clemente, R. (2005). Assessment and intervention with children and adolescents: Development and multicul tural approaches (2nd ed.). Alexandria, VA: American Counseling Association. Walker, J. F., & Archibald, L. M. D. (2006). Articulation rate in preschool children: A 3 year longitudinal study. International Journal of Language & Communication Disorders, 41 ( 5) 541 565. DOI: 10.1080/10428190500343043 Waller, M. A. (2001). Resilience in ecosystemic context: Evolution of the concept. American Journal of Orthopsychiatry, 71 (3), 290. Wang, J. & Wang, X. (2012). Structural equation modeling: Applications using Mpl us. United Kingdom: Higher Education Press. Werner, E. E. (2000). Protective factors and individual resilience. In J. P. Shonkoff & S. J. Meisels (Eds.), Handbook of early childhood interventions (2nd edition) (pp. 115 132). NY: Cambridge University Press. Werner, E. E., & Smith, R. S. (1982). Vulnerable, but invincible: A longitudinal study of resilient children and youth NY: McGraw Hill. West Olatunji, C. A. (2010). ACA advocacy competencies with culturally diverse clients. In M. J. Ratts, R. L. Toporek & J. A. Lewis (Eds.), ACA advocacy competencies: A social justice framework for counselors (pp. 55 63). VA: American Counseling Association. West Olatunji, C. (2012). Equal access, unequal resources: Appreciating cultural, social and economic diversity i n families. In E. Amatea (Ed.), Building culturally responsive family school partnerships (2nd ed., pp. 143 166). Upper Saddle River, NJ: Pearson. West Olatunji, C., Behar Horenstein, L., Algina, J., Oliver, B., Russell Brown, K., Pemberton, D., Bowie, M., Archer Banks, D., & Shure, L. (2012, May). Emancipatory educational research. In C. West Olatunji & L. Behar Horenstein (Chairs), Emancipatory educational research conference. Conference conducted at the meeting of The Emancipatory Educational Research Gr oup, University of Florida College of Education, Gainesville, FL. West Olatunji, C. A., & Conwill, W. (2011). Counseling & diversity: Counseling African Americans CA: Brooks/Cole. West Olatunji, C., Frazier, K. N., & Kelley, E. (2011), Wraparound counseli ng: An ecosystemic approach to working with economically disadvantaged students in urban school settings. Journal of Humanistic Counseling, 50 (2), 222 237. DOI: 10.1002/j.2161 1939.2011.tb00121.x West Ol atunji, C. & Gibson, D. M. (2012). A global and historical introduction to social class. In D. C. Sturm & D. M. Gibson (Eds.), Social class and the helping professions: A (pp. 3 16). NY: Rou tledge Talyor & Francis Group. PAGE 188 188 West Olatunji, C., Sanders, T., Mehta, S., & Behar Horenstein, L. (2010). Parenting practices among low income Parents/Guardians of academically successful fifth grade African American children. Multicultural Perspectives, 12 (3), 138 144. DOI: 10.1080/15210960.2010.504475 Wilkerson, D., Johnson, G., & Johnson, R. (2008). Children of neglect with attachment and time perception deficits: Strategies and interventions. Education, 129 (2), 343 352. Williams, D. R., Mohammed, S. A., Leavell, J. and Collins, C. (2010), Race, socioeconomic status, and health: Complexities, ongoing challenges, and research opportunities. Annals of the New York Academy of Sciences, 1186 69 101. DOI: 10.1111/j.1749 6632.2009.05339.x Williams, D. R., Neighbors, H. W., & Jackson, J. S. (2003). Racial/ethnic discrimination and health: Findings from community studies. American Journal of Public Health, 93 (2), 200 208. Wolfgang, J. D., Frazier, K., West Olatunji, C., & Barrett, J. (2012). Developing cross cultural competence: Applying development and prevention Ideals to counseling young children. Educational Resources Information Clearinghouse ERIC No: ED528898, pp. 1 16. Wolfgang, J. D., West Olatunji, C., Frazier, K., & Barrett, J. (2012). Culture centered pediatric counseling interventions for young children. In G. R. Walz, J. C. Bleuer, & R. K. Yep (Eds.), Ideas and research you can use: VISTAS 2012 Retrieve d from http://www.counseling.org/Resources Wu, W., & West, S. G. (2010). Sensitivity of fit indices to misspecification in growth curve models. Multivariate Behavioral Research, 45 (3), 420 452. DOI: 10. 1080/00273171.2010.483378 Wycoff, S. E. M. (1996). Academic performance of Mexican American women: Sources of support that serve as motivating variables. Journal of Multicultural Counseling & Development 24 (3), 146 155. Wynn, R., & West Olatunji, C. (2008 ). Culture centered case conceptualization using NTU psychotherapy with an African American gay male client. Journal of LGBT Issues in Counseling, 2 (4), 308 325. DOI: 10.1080/15538600802501995 Yuan, K H., & Bentler, P. M. (2000). Three likelihood based met hods for mean and covariance structure analysis with nonnormal missing data. Sociological Methodology, 30 (), 165 200. Zeanah, C. H., Jr., & Smyke, A. T. (2009). Attachment disorders. In C. H. Zeanah, Jr. (Ed.), Handbook of infant mental health (3rd ed., pp .421 434). NY: The Guilford Press. Zeanah, C. H., Jr., & Zeanah, P. D. (2009). The scope of infant mental health. In C. H. Zeanah, Jr. (Ed.), Handbook of infant mental health (3rd ed., pp.5 21). NY: The Guilford Press. PAGE 189 189 Zero to Three (2005). DC:0 3R: Diagno stic classification of mental health and developmental disorders of infancy and early childhood (revised edition) Washington, D.C: Zero to Three Press. PAGE 190 190 BIOGRAPHICAL SKETCH Jeff Drayton Wolfgang received a Bachelor of Science, in fall 2003, graduating wit h c um Laude status in his zoology major. In fall of 2008, he received a Master of Education and Specia list in Education, majoring in marriage and family counseling. In summer 2013, he received his Doctor of Philosophy, majoring in marriage and family counseling. He received his bachelor s, master s, and doctora te degree s from the University of Florida. Hi s research has focused on traumatic stress, attachment, marginalized and oppressed populations, multicultural counseling, and pediatric counseling. A theme of his research, publications, presentations, and practice are that theories and methodologies need to embody emancipatory frameworks thus leading to mutually empowering and transformative relationships. 