TRAUMA AND RESILIENCE AMONG MALTREATED YOUTH: THE ROLE OF SOCIAL
SUPPORT IN PROMOTING ACADEMIC AND SOCIAL-EMOTIONAL WELL-BEING
By
Madison L. Chapman

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
School Psychology - Doctor of Philosophy
2017

ABSTRACT
TRAUMA AND RESILIENCE AMONG MALTREATED YOUTH: THE ROLE OF SOCIAL
SUPPORT IN PROMOTING ACADEMIC AND SOCIAL -EMOTIONAL WELL-BEING
By
Madison L. Chapman
This study utilized data from the National Survey of Child and Adolescent Well-Being II
(NSCAW II; NSCAW, 2007) to examine the role of social support in fostering resilience among
victims of child abuse and neglect. Broadly, resilience refers to achieving or maintaining typical
development in one or more areas of functioning following adverse life experiences (Masten,
2001). Numerous extant studies have established that social support is a protective factor that
promotes resilience (e.g., Afifi & MacMillan, 2011). This previous research, however, has not
adequately analyzed social support as a protective factor among youth who have experienced
abuse or neglect. Social support provided at the time of maltreatment and support from multiple
sources may be critical for resilient outcomes. Prior research has also primarily focused on
mental health and behavioral outcomes and has neglected analysis of academic outcomes. To
address these gaps in literature, this study assessed how social support from caregivers and peers
was related to both academic and social-emotional well-being within a sample of maltreated
youth. Through Structural Equation Modeling (SEM), distinct patterns were observed. Perceived
emotional support from caregivers was related to greater academic engagement and fewer
externalizing behavior problems. Perceived emotional support from peers was related to fewer
internalizing behavior problems. Inclusion of both caregiver and peer social support measures
provided the most accurate conceptual model. These results indicate that social support from
different sources meet different needs among victims of maltreatment. Therefore, multiple
sources of social support should be promoted to foster resilience among maltreated youth.

Copyright by
MADISON L. CHAPMAN
2017

To my husband, Bryan Chapman.
Thank you for your unconditional love and support.

iv

ACKNOWLEDGEMENTS

I would like to acknowledge the individuals who had a significant role in helping me
reach this accomplishment. I am incredibly grateful to my advisor, Dr. Evelyn Oka, for
encouraging me to pursue my interest in trauma and helping me create opportunities to do so.
With all that she has taught me, I have most appreciated her ability to teach me the importance of
having both personal motivation and attainable ambition. I would also like to express my
appreciation to my dissertation committee members, Dr. Jodene Fine, Dr. Cary Roseth, and Dr.
Laura Apol, whose guidance and feedback challenged me to become a more critical and better
researcher. I am especially thankful to Dr. Laura Apol for her time and attention while helping
me refine my interests and identify my true passion.
My husband, Bryan Chapman, has encouraged and supported me through every step of
this journey. He celebrated my wins and helped me get back up after my defeats. I appreciate the
many times he reminded me to work on my dissertation and the many times he reminded me to
take a break and have fun. I appreciate the sacrifices he made for me to follow this dream. I
could not have done this without him.
I am grateful to my parents, Gail and Preston Kloss, and to my brother Alex Kloss. They
showed me the importance of serving the community, gave me a passion for higher education,
and taught me how to persevere through challenges. They have provided incredible support to
help me attain this dream. I am so grateful that I knew a home consisting of love and joy and
understood that not all children had the same advantage.

v

Most importantly, I am eternally grateful for my relationship with Jesus. It is because of
His sacrifice that I have hope for eternity. My faith in Him is what led me to earn a PhD and use
it to bring that same hope and love to children in significant need.
Finally, I would like to acknowledge the individuals who contributed to the design and
data collection of the National Survey of Child and Adolescent Well-Being studies. Through this
rich data set, researchers including myself can learn more about the needs of youth in the child
welfare system and identify ways to better serve those needs. Financial support for this
dissertation research was provided through the Michigan State University Dissertation
Completion Fellowship.

vi

TABLE OF CONTENTS

LIST OF TABLES .................................................................................................................x
LIST OF FIGURES ...............................................................................................................xii
CHAPTER 1 ..........................................................................................................................1
INTRODUCTION .................................................................................................................1
CHAPTER 2 ..........................................................................................................................9
LITERATURE REVIEW ......................................................................................................9
Child Welfare .............................................................................................................9
Trauma .......................................................................................................................10
Acts of Commission (Child Abuse) ...............................................................11
Acts of Omission (Child Neglect)..................................................................12
Academic Well-Being ................................................................................................18
Academic Achievement .................................................................................18
Academic Engagement ..................................................................................19
Social-Emotional Effects of Trauma .........................................................................21
Internalizing ...................................................................................................21
Externalizing ..................................................................................................21
Resilience ...................................................................................................................22
Social Support ............................................................................................................25
Caregiver Social Support ...............................................................................29
Peer Social Support ........................................................................................29
Purpose of the Present Study .....................................................................................30
Research Questions and Hypotheses .........................................................................31
Research Question 1 ......................................................................................31
Hypothesis 1...................................................................................................31
Research Question 2 ......................................................................................31
Hypothesis 2...................................................................................................31
Research Question 3 ......................................................................................32
3a. .......................................................................................................32
Hypothesis 3a. ....................................................................................32
3b........................................................................................................32
Hypothesis 3b.....................................................................................32
3c ........................................................................................................32
Hypothesis 3c .....................................................................................32
3d........................................................................................................32
Hypothesis 3d.....................................................................................32
3e ........................................................................................................32
Hypothesis 3e .....................................................................................32
Research Question 4 ......................................................................................32
4a ........................................................................................................33

vii

Hypothesis 4a .....................................................................................33
4b........................................................................................................33
Hypothesis 4b.....................................................................................33
4c ........................................................................................................33
Hypothesis 4c .....................................................................................33
4d........................................................................................................33
Hypothesis 4d.....................................................................................33
4e ........................................................................................................33
Hypothesis 4e .....................................................................................34
CHAPTER 3 ..........................................................................................................................35
METHODS ............................................................................................................................35
Study Model ...............................................................................................................35
Study Design ..............................................................................................................37
Data Collection ..............................................................................................38
Original NSCAW Sample ..........................................................................................39
Final Sample for Current Study .................................................................................41
Child Welfare Characteristics ........................................................................42
Sample Criteria ..............................................................................................45
Missing Data ..................................................................................................46
Sample Weights .............................................................................................47
Variables and Measures .............................................................................................47
Demographics ................................................................................................48
Trauma ...........................................................................................................50
Caregiver Social Support ...............................................................................52
Peer Social Support ........................................................................................54
Academic Achievement .................................................................................56
Academic Engagement ..................................................................................57
Internalizing Behaviors ..................................................................................59
Externalizing Behaviors .................................................................................60
CHAPTER 4 ..........................................................................................................................61
RESULTS ..............................................................................................................................61
Preliminary Analyses .................................................................................................61
Descriptive Statistics ..................................................................................................63
Structural Model ........................................................................................................68
Research Question 1 ......................................................................................68
Caregiver Support ..............................................................................68
Peer Support .......................................................................................69
Final Model ........................................................................................70
Research Question 2 ......................................................................................71
Research Question 3 ......................................................................................74
Type ...................................................................................................74
Severity ..............................................................................................74
Substantiation.....................................................................................74
Age .....................................................................................................75

viii

Gender ................................................................................................75
Research Question 4 ......................................................................................75
Type ...................................................................................................76
Severity ..............................................................................................76
Substantiation.....................................................................................77
Age .....................................................................................................77
Gender ................................................................................................77
CHAPTER 5 ..........................................................................................................................79
DISCUSSION ........................................................................................................................79
Resilience ...................................................................................................................79
Differential Effects.....................................................................................................84
Significance of Peer Social Support ..........................................................................91
Trauma Symptoms .....................................................................................................92
Limitations .................................................................................................................95
Implications for Practice ............................................................................................98
Future Research .........................................................................................................100
APPENDIX ............................................................................................................................103
REFERENCES ......................................................................................................................106

ix

LIST OF TABLES

Table 1

Demographic Characteristics of NSCAW II..................................................41

Table 2

Demographic Characteristics of Sample ........................................................42

Table 3

Child Welfare Characteristics of Sample .......................................................44

Table 4

Summary of NSCAW II Measures Corresponding to Constructs .................48

Table 5

Confirmatory Factor Analysis for One-Factor Model
of Trauma Symptoms .....................................................................................52

Table 6

Confirmatory Factor Analysis for One-Factor Model of
Caregiver Social Support ...............................................................................54

Table 7

Confirmatory Factor Analysis for One-Factor Model of
Peer Social Support ........................................................................................56

Table 8

Confirmatory Factor Analysis for One-Factor Model of
Academic Engagement ..................................................................................59

Table 9

Estimated Correlation Matrix for Observed Variables ..................................62

Table 10

Estimated Correlation Matrix for Latent Variables. ......................................62

Table 11

Descriptive Statistics for Sample ...................................................................66

Table 12

Unstandardized and Standardized Factor Loadings and Significance
Levels for Model Assessing Caregiver Social Support .................................69

Table 13

Unstandardized and Standardized Factor Loadings and Significance
Levels for Model Assessing Peer Social Support ..........................................70

Table 14

Unstandardized and Standardized Factor Loadings and Significance
Levels for Final Model...................................................................................71

Table 15

Model Fit Indices Comparing the Relative Impact of Caregiver
Support and Peer Support ..............................................................................72

Table 16

Results of Chi-Square Test for Difference Testing........................................73

Table 17

Predictors of Trauma Symptoms ...................................................................75

x

Table 18

Predictors of Social Support ..........................................................................78

xi

LIST OF FIGURES

Figure 1

Conceptual Model ..........................................................................................37

Figure 2

Standardized Factor Loadings of Model Assessing
Caregiver Social Support ...............................................................................81

Figure 3

Standardized Factor Loadings of Model Assessing
Peer Social Support ........................................................................................81

Figure 4

Standardized Factor Loadings of Final Model...............................................85

Figure 5

Structural Equation Model Assessing Caregiver Social Support ..................104

Figure 6

Structural Equation Model Assessing Peer Social Support ...........................105

xii

CHAPTER 1
INTRODUCTION
Child maltreatment is a persistent and pervasive problem affecting millions of children
within the United States. In 2013, the U.S. Department of Health and Human Services (DHHS)
reported that approximately 3.9 million children were the subjects of child welfare investigations
(U.S. DHHS, 2015). These investigations were initiated following reports of suspected abuse or
neglect toward the child. Of these 3.9 million children, an estimated 670,000 children were found
to be victims of substantiated maltreatment (U.S. DHHS, 2015). Studies examining self-reported
abuse and neglect during childhood, however, suggest that this prevalence rate may be
significantly greater. In a national sample of adolescents, 28.4% of respondents reported physical
abuse during childhood, 11.8% reported physical neglect, and 4.5% reported sexual abuse
(Hussey, Chang, & Kotch, 2006). These data suggest that many children are exposed to abuse
and neglect that may not have been reported to authorities.
Maltreatment can result in harm to children’s social-emotional and academic functioning.
Research shows that child maltreatment severely interferes with development in multiple
domains including cognitive, social-emotional, and behavioral functioning (Hildyard & Wolfe,
2002). Children and adolescents who experience maltreatment, especially physical neglect, also
perform more poorly on standardized tests of intellectual functioning than non-maltreated
children (Hildyard & Wolfe, 2002). Maltreated children are more likely to have anxious or
disorganized attachments to their caregivers than nonmaltreated children (Hildyard & Wolfe,
2002). These negative effects of maltreatment often continue to affect individuals throughout
childhood and even through adulthood.

1

Maltreatment can also lead to children experiencing symptoms of posttraumatic stress.
Among a sample of children (ages 8-14) referred to child welfare for suspected abuse or neglect,
11.7% reported clinically significant symptoms of posttraumatic stress using the Post-Traumatic
Stress subscale of the Trauma Symptom Checklist for Children, a standardized survey assessing
common symptoms of trauma (Kolko, Hurlburt, Zhang, Barth, Leslie, & Burns, 2010).
Although maltreatment is generally associated with these negative outcomes, some
children and adolescents exhibit typical social-emotional and academic developmental outcomes.
This positive adaptation in the presence of risk is referred to as resilience (Masten, 2001). The
study of resilience related to child maltreatment has led to significant advances in the
understanding of adaptation within these adverse experiences. Static variables and adaptation
systems at the individual and contextual level simultaneously influence development.
Research shows that certain factors or characteristics are associated with more resilient
outcomes. These factors, referred to as protective factors, have been shown to reduce or
moderate the negative effects of maltreatment. These factors include static variables, such as the
age of the child at the time of maltreatment (Ogle, Rubin, & Siegler, 2013), the type of
maltreatment (Wecsler-Zimring, & Kearney, 2011), and severity of maltreatment (Hyman &
Williams, 2001). Later onset of maltreatment, while controlling for gender and race, is associated
with fewer behavioral and mental health problems in adulthood (Kaplow & Widom, 2007).
Children who experience forms of neglect tend to have fewer symptoms of trauma compared to
those that experience forms of abuse (Wecsler-Zimring & Kearney, 2011). Children who
experience less severe maltreatment experience fewer symptoms of traumatic stress compared to
those that experience severe maltreatment (Hyman & Williams, 2011).

2

Protective factors also include individual characteristics such as gender (DuMont,
Widom, & Czaja 2007), intellectual functioning (Masten, Garmezy, Tellegen, Pellegrini, Larkin,
& Larsen, 1988), temperament (Martinez-Torteya, Anne Bogat, Vvon Eye, & Levendosky,
2009), self-esteem (Cicchetti, Rogosch, Lynch, & Holt, 1993), and coping skills (Banyard &
Williams, 2007; Rosenthal, Feiring, & Taska, 2003). After controlling for differential
probabilities of experiencing trauma and maltreatment, males are less likely to develop
symptoms of posttraumatic stress following a traumatic event compared to females (Tolin & Foa,
2006). Children with higher intellectual functioning (Masten et al., 1988), easy temperament
(e.g., positive mood, low reactivity; Martinez-Toreya et al., 2009), positive self-esteem (Cicchetti
et al., 1993), and effective and flexible coping skills (Banyard & Williams, 2007; Rosenthal,
Feiring, & Taska, 2003) are more likely to experience resilient outcomes following exposure to a
potentially traumatic event.
External or contextual protective factors include having stable and supportive caregivers
(Herrenkohl, Herrenhokl, & Egolf, 1994), socioeconomic advantages (Masten et al., 1988),
connections to extended family networks (Masten et al. 1988), and neighborhood advantages
(DuMont et al., 2007). A contextual protective factor with substantial empirical evidence is
social support (Afifi & MacMillan, 2011). Children who have supportive relationships with
caregivers or others are more likely to exhibit positive and resilient outcomes in the presence of
risk due to maltreatment. It is hypothesized that this perceived social support fosters the
development of resilience among maltreated children.
This study addressed several gaps in existing literature. First, social support was
examined through relationships with peers in addition to relationships with caregivers. Many
previous studies have demonstrated that social support from caregivers is associated with

3

resilience among maltreated children (e.g., Dingelder, Jaffee, & Mandell, 2010; Muller, GoebelFabri, Diamond, & Dinklage, 2000). Fewer studies, however, have examined social support from
peers. In one study examining social support and adaptive outcomes among African American
children exposed to maltreatment, investigators assessed children’s ratings of support from both
caregiver and peer relationships (Lamis, Wilson, King, & Kaslow, 2014). Within this analysis,
support from both caregivers and peers was related to positive social functioning among children
with a history of physical or emotional abuse. The magnitude of the relation between social
support and social functioning was equal between caregiver support and peer support. Among
children with a history of sexual abuse, only caregiver support was related to positive social
functioning. This study, however, assessed only one outcome measure: social and adaptive
functioning. The study also included only African American children between the ages of 8 and
12 living in an urban setting. Due to these limitations, the findings may not generalize to other
areas of functioning or other populations of maltreated children. Another study examined social
support from multiple sources, coping, perfectionism, and depression among adolescents with a
history of maltreatment (Flett, Drucman, Hewitt, & Wekerle, 2012). This study included a small
sample of 58 participants and did not include a direct measure of maltreatment or trauma history.
Neither of these studies utilized a framework of risk and resilience to examine the role of social
support in promoting resilient or adaptive outcomes. Due to these factors, existing literature in
this area has not adequately examined the role of peer social support in the development of
resilience following maltreatment.
Examining the effects of peer relationships is important for understanding how social
support may promote adaptive functioning among victims of child maltreatment. A significant
portion of children’s social interactions occur at school with peers. These relationships become

4

increasingly important during adolescence. In childhood and early adolescence, relationships
with parents or caregivers are generally the central source of social support (Helsen, Vollebergh,
& Meeus 2000). In adolescence, however, this central source of support gradually shifts to peers
(Helsen et al., 2000). As peers become more important to adolescents, these relationships may
also have a greater effect on adolescents’ functioning. Furthermore, caregiver relationships may
become even less significant among victims of child maltreatment. Many of these children are
victimized by a parent or other closely related caregiver. This violation of trust and security
would likely negatively affect the relationship with the offending caregiver, but may also affect
the child’s ability to trust other caregivers in their lives. Children with a history of maltreatment
may rely more on peer relationships for support. By examining perceived social support from
peer relationships, this study investigated how peer relationships may uniquely contribute to
resilience following maltreatment.
Another way in which this study addressed gaps in existing literature was by examining
academic outcomes indicating resilience. Extant studies have primarily conceptualized resilience
as the absence of harmful outcomes and have therefore focused on assessing emotional and
behavioral outcomes. Many studies examining resilience among maltreated youth have assessed
the incidence of psychopathology (e.g., Kolko et al., 2010), such as internalizing behavior
problems (e.g., Thompson & Tabone, 2010) and externalizing behavior problems (e.g.,
Thompson & Tabone, 2010). Although it is important to examine these outcomes, they do not
provide a complete picture of children’s functioning. Examining academic outcomes within this
study provided a more complete understanding of resilience among victims of maltreatment.
Academic achievement and academic engagement significantly predict important outcomes later
in life such as educational attainment, employment, socio-economic status, and overall life

5

satisfaction. Examining academic outcomes is important not only to understand children’s
current functioning and resilience, but also to predict future outcomes and resilience.
A third way in which this study addressed gaps in existing literature was by examining
the perspectives of children. Studies examining the effects of maltreatment generally sample
adult populations with a history of maltreatment during childhood. Few studies in this area have
included child participants. The existing literature therefore has not adequately examined
resilience at the time of maltreatment. Furthermore, previous studies have primarily assessed
adult outcomes, rather than functioning during childhood. Many studies have examined
perceived social support from family and friends in adulthood as a protective factor among
individuals with a history of child maltreatment (e.g., Pepin & Banyard, 2005; McLewin &
Muller, 2006; Vranceanu, Hobfoll, & Johnson, 2007; Seeds, Harkness, & Quilty, 2010; Evans,
Steel, & DiLillo, 2013; Folger & Wright, 2013; Wilson & Scarpa, 2014). Generally, these studies
ask adult participants to report previous experiences of maltreatment during childhood, current
perceptions of social support, and current outcomes such as mental health and social functioning.
It is presumed that these extant studies have primarily used adult populations due to ethical and
practical issues involved in conducting research with the vulnerable population of child victims
of maltreatment. While these studies provide estimates regarding the role of social support in
developing resilient outcomes, retrospective methods have poorer reliability and validity than
concurrent studies. These previous studies may not provide a complete understanding of how
children develop resilient outcomes and how social support may play a role in this process.
Assessing a sample of children at that time of their involvement with the child welfare
system/protective services allowed for the examination of children’s perceptions of their social
support and resilience. It also helped answer the question as to whether caregiver and peer social

6

support matters immediately following the report of maltreatment. Currently, research provides
evidence for social support as a protective factor leading to long-term positive outcomes but not
for how social support relates to short-term outcomes. This information is necessary to develop
effective short-term and long-term services and supports for children following maltreatment.
The current study addressed these gaps in previous research by examining maltreatment
type, perceived trauma, and perceived social support from caregivers and peers immediately
following the report of maltreatment and both social-emotional and academic outcomes
approximately 18 months after the report of maltreatment. The purpose of this study was to
examine the role of social support as a moderator, to determine whether it affects the relation
between trauma and outcomes among a sample of children referred to child welfare following
suspected abuse or neglect. Perceived social support from multiple sources including caregivers
and peers were evaluated. This addressed the gap within current literature by examining nonfamilial relationships in addition to familial relationships. Examining perceived social support
from these sources provided a more comprehensive indicator of children’s overall social support
and can provide educators with a better understanding of how to effectively support the needs of
this population within the school setting.
Resilience was assessed by measuring functioning after maltreatment within the socialemotional (externalizing behaviors and internalizing behaviors) and academic domains
(academic achievement and academic engagement). Existing studies examining resilience within
this population often conceptualize resilience as a lack of negative outcomes, particularly socialemotional outcomes. Resilience research, however, has shown that both the lack of negative
outcomes and the presence of positive functioning are important in assessing resilience (Masten,

7

2001). Therefore, the absence of negative outcomes and the presence of positive outcomes were
assessed in this study.
This study examined the role of social support from caregivers and peers and identified
the unique role of each type of social support. Results from this study may help professionals,
especially educators, understand the importance of providing this kind of support in the social
context following maltreatment. Implications include informing the services and supports for
victims of maltreatment within the home and school settings.

8

CHAPTER 2
LITERATURE REVIEW
Child Welfare
Child Welfare refers to the system of government and non-government run services to
protect children and promote family stability. The primary federal legislation addressing child
welfare is the Child Abuse Prevention and Treatment Act (CAPTA; Child Welfare Information
Gateway, 2011). This legislation provides funding for child welfare services to states and sets
minimum definitions of child maltreatment. CAPTA was enacted in 1974 and was most recently
amended and reauthorized in December 2010 (Child Welfare Information Gateway, 2011).
Child welfare reports are made when an individual knows or has reasonable cause to
suspect that a child has been subjected to abuse or neglect. These reports may be made by
caregivers, teachers, physicians, or other adults who know the child. Following a report of
suspected maltreatment, Child Protective Services (CPS) staff may initiate an investigation to
determine whether the child has been or is at risk of being harmed. CPS staff then collect
information to determine whether there is evidence of abuse or neglect (i.e., substantiated) or
there is insufficient evidence to conclude the child has been or is being maltreated. If the
investigation results in low risk for future abuse and ongoing safety concerns, the family may be
referred for voluntary CPS services (e.g., parenting classes, family intervention). If there is a
continued risk of harm, the child may remain with the family or may be placed in out-of-home
care. CPS services are provided to the child and family throughout this process. Potential
outcomes of a CPS investigation are that the child may be reunified with the family, may remain
in custody of a relative, may be adopted or live with a permanent legal guardian after parental
rights are terminated, or may live independently with permanent family connection.

9

Trauma
Broadly, trauma refers to distress following a significantly stressful or adverse event or
experience. The American Psychiatric Association (2013) defines a traumatic event as exposure
to actual or threatened death, serious injury, or sexual violation. This exposure may involve the
individual directly experiencing the event, witnessing the event in person, learning that the event
occurred to a close family member or close friend, or experiencing repeated direct exposure to
adverse detail of the event through media, pictures, television, or movies (APA, 2013).
Traumatic experiences may occur as an isolated event (e.g., a single incident of sexual assault),
multiple discrete events (e.g., repeated physical assault), or as an ongoing condition (e.g.,
proximity to war conditions). When these experiences cause significant distress or impairment in
one or more area of functioning, this is referred to as trauma.
Children and adolescents may be exposed to different types of traumatic events. Many
common life experiences can be traumatic, as described by the criteria above. These events
include situations such as the loss of a parent or other close family member or friend, severe
acute or chronic medical conditions, or experiencing a house fire. In addition to these potentially
traumatic events, many children and adolescents are exposed to maltreatment.
Childhood maltreatment refers to several types of abuse including direct physical abuse,
sexual abuse, and emotional abuse, or neglect of a child’s physical or emotional needs (Child
Welfare Information Gateway, 2014). CAPTA provides a broad definition of maltreatment and
requires states to define maltreatment minimally as “any recent act or failure to act on the part of
a parent or caretaker, which results in death, serious physical or emotional harm, sexual abuse, or
exploitation, or an act or failure to act which presents an imminent risk of serious harm” (Child
Welfare Information Gateway, 2014).

10

The Centers for Disease Control developed recommended definitions of terms related to
child maltreatment to promote and improve consistency of research for public health practices
(Leeb, Paulozzi, Melanson, Simon, & Arias, 2008). Like the federal definition provided by the
U.S. DHHS, the CDC defines child maltreatment as “Any act or series of acts of commission or
omission by a parent or other caregiver that results in harm, potential harm, or threat of harm to a
child” (Leeb et al., p. 11). Acts of commission involve words or overt actions that cause harm,
potential harm, or threat of harm and are generally referred to as abuse. Acts of omission involve
failure to provide for needs or protect from harm or potential harm and are generally referred to
as neglect. Utilizing this differentiation recommended by the CDC allows results from this study
to be comparable to previous and future literature related to child maltreatment. Within study
analyses, types of maltreatment were categorized as abuse (physical, sexual, or emotional abuse)
or neglect (physical neglect and parental substance abuse/domestic violence).
Acts of Commission (Child Abuse). Child abuse includes acts of physical abuse, sexual
abuse, and emotional abuse. Physical abuse is defined by federal and state laws as nonaccidental
physical injury to the child (Welfare Information Gateway, 2014). This physical injury may
include striking, kicking, burning the child, or any other action that results in physical injury
toward the child. Within 38 states, the definition of physical abuse also includes action that
threatens the child with physical harm or causes risk of harm to the child (Welfare Information
Gateway, 2014).
Sexual abuse is defined by CAPTA as: “The employment, use, persuasion, inducement,
enticement, or coercion of any child to engage in, or assist any other person to engage in, any
sexually explicit conduct or simulation of such conduct for the purpose of producing a visual
depiction of such conduct; or the rape, and in cases of caretaker or interfamilial relationships,

11

statutory rape, molestation, prostitution, or other form of sexual exploitation of children, or
incest with children.” (U.S. Department of Health and Human Services, 2010, p. 31). While the
federal definition of sexual abuse details specific acts of sexual abuse, some state legislation
includes general definitions of sexual abuse (Welfare Information Gateway, 2014).
Psychological abuse, or emotional abuse, is specified in most state definitions of
maltreatment. These definitions generally include language such as “injury to the psychological
capacity or emotional stability of the child as evidenced by an observable or substantial change
in behavior, emotional response, or cognition” such as “anxiety, depression, withdrawal, or
aggressive behavior” (Welfare Information Gateway, 2014, p. 3). This injury may result from
verbal aggression towards children including shouting and intimidation or manipulation
including emotional blackmail (Types of Child Abuse, 2015). Overall, psychological abuse
involves causing psychological or emotional harm to a child.
Acts of Omission (Child Neglect). Definitions of neglect typically involve language
such as “the failure of a parent or other person with responsibility for the child to provide needed
food, clothing, shelter, medical care, or supervision to the degree that the child’s health, safety,
and well-being are threatened with harm” (Welfare Information Gateway, 2014, p. 2). Some
states include additional provisions within definitions of neglect to include failure to educate the
child as required by law and failing to provide any special medical treatment or mental health
care needed by the child (Welfare Information Gateway, 2014). Neglect involves some form of
caregivers neglecting to provide for the physical needs of children.
Parental domestic violence and parental substance abuse are also common causes of child
welfare investigations. Parental domestic violence includes exposure to the physical assault
toward one or more relatives by another relative. This typically involves a male relative, such as

12

a father or husband/partner, using physical violence toward women and children (Types of Child
Abuse, 2015). This violence may also involve physical abuse or emotional abuse toward the
child or other family members. Parental substance abuse may include actions such as: prenatal
exposure of a child to harm due to mother’s substance use, use of a controlled substance by a
caregiver that impairs the caregiver’s ability to adequately care for the child, or manufacture of a
controlled substance in the presence of a child or on the premises occupied by a child (Welfare
Information Gateway, 2014). These actions cause direct harm to children and are therefore
included in definitions of child maltreatment.
Childhood maltreatment is a common problem within the United States (Wildeman,
Emanuel, Leventhal, Putnam-Hornstein, & Waldfogel 2014). Each year, approximately 0.9% of
children are confirmed victims of maltreatment according to the DHHS Child Maltreatment 2013
report (US DHHS, 2015). This rate has been relatively stable between 2009 and 2013 (US
DHHS, 2015). Estimates of the lifetime prevalence of childhood maltreatment, however, are
much greater. A study utilizing data from CPS investigations between 2004 and 2011 estimates
that 12.5% of children will experience persistent or severe maltreatment leading to a confirmed
CPS investigation by the age of 18 (Wilderman et al., 2014). Furthermore, self-report studies
suggest even greater prevalence rates. A study conducted in 2011 using telephone interviews
found that over 40% of children were maltreated during childhood (Finkelhor, Turner, Shattuck,
& Hamby, 2013). These lifetime estimates suggest that many children and adolescents will
experience a form of maltreatment during childhood and this maltreatment may not result in a
confirmed CPS investigation.
Furthermore, the prevalence of child maltreatment varies based on type of maltreatment.
Prevalence studies generally include the four primary types of maltreatment (physical abuse,

13

emotional abuse, sexual abuse, and neglect), but do not often include domestic violence, parental
substance abuse, and/or abandonment. Within the confirmed cases of maltreatment in 2013,
nearly 80% were victims of neglect, 18% were physically abused, 9% were emotionally abused,
and 9% were sexually abused (US DHHS, 2015). Estimates of the lifetime prevalence of
maltreatment suggest somewhat different patterns. Among adolescents reporting lifetime
victimization, emotional abuse was the most commonly reported form of maltreatment (25%),
followed by neglect (22%), physical abuse (18%), and sexual abuse (0.5%; Finkelhor et al.
2014). These data indicate that neglect is the most common form of reported and substantiated
maltreatment. Data from self-reports, however, indicate that neglect, physical abuse, and
emotional abuse may be experienced at similar rates during childhood.
Children’s trauma symptoms following maltreatment appears to be affected by several
factors including the type of trauma, severity of trauma, substantiation, age, and gender. Whereas
children exposed to any type of maltreatment generally exhibit symptoms of trauma, children
who experience physical or sexual maltreatment appear to exhibit greater symptoms of trauma
than children who experience neglect (Wechsler-Zimring & Kearney, 2011).
Severity of trauma has been conceptualized in a variety of ways. Studies have examined
chronicity of maltreatment as the number of different forms of trauma, number of reported
maltreatment incidents, and length of time maltreatment occurred. Children who experience
multiple forms of trauma (i.e., polyvictimization), a greater number of maltreatment incidents,
and greater duration of maltreatment are more likely to have symptoms of posttraumatic stress
and depression and externalizing behavior problems than children that experience less chronic
maltreatment (Ford, Wasser, & Connor, 2011; English, Graham, Litrownik, Everson, &

14

Bangdiwala, 2005). Severity of trauma significantly predicts poorer outcomes following
maltreatment.
Substantiation of trauma may also predict symptoms of trauma, but current research in
this area has resulted in mixed findings. A determination of substantiation is made after a child
protective services investigation. Substantiated cases indicate that the investigation resulted in
sufficient evidence to conclude the alleged maltreatment occurred. Unsubstantiated cases
indicate that the investigation did not find sufficient evidence of maltreatment. While this
determination aims to provide an indicator regarding the occurrence of maltreatment, this
indicator may have poor reliability and validity due to differences across caseworkers and CPS
agencies, the identification and report of maltreatment cases, a lack of credible evidence, or other
factors (Hussey et al., 2005). Therefore, the determination of substantiation may under-identify
actual occurrences of maltreatment. Several studies have investigated the utility of assessing
substantiation in maltreatment research and have resulted in mixed findings.
Some studies have found that youth with substantiated investigations of maltreatment are
more likely to engage in delinquent behaviors (Snyder & Smith, 2014). Other investigations,
however, have found no significant differences between unsubstantiated and substantiated
investigations on behavior, adaptive skills, and trauma symptom outcomes (Hussey et al., 2005).
Substantiated reports of maltreatment do predict a higher risk of an additional report of alleged
maltreatment within 12 months, regardless of gender, race, age, or type of maltreatment (Fuller
& Nieto, 2009). These previous findings suggest that the indication of substantiation may or may
not predict negative outcomes including symptoms of trauma. This study aims to further
investigate these questions by assessing the predictive validity of substantiation on perceived
trauma symptoms following child maltreatment.

15

Age and developmental stage have been shown to significantly predict development of
trauma symptoms. Childhood trauma exposure is related to more severe symptoms of PTSD
compared to trauma exposure in adulthood, even after controlling for differences in cumulative
trauma exposure or subjective differences in traumatic experiences (Ogle et al., 2013). This
indicates that understanding trauma exposure and development of posttraumatic stress symptoms
is particularly important among child populations. Individuals experience many developmental
changes during childhood and adolescence. These differences may cause youth at different ages
to develop qualitatively different responses to traumatic events. For example, adolescence is a
time of identity formation and increasing independence. Experiencing a traumatic event during
this developmental stage may have greater effect on overall well-being than experiencing a
traumatic event at a younger age. Differences in trauma symptoms based on developmental
stages have primarily examined large ranges in age such as childhood, adolescence, young
adulthood, middle-age, and older adulthood (Ogle et al., 2013). Within these developmental
stages, however, there may be further variance in how individuals respond to trauma. Therefore,
examining age differences at narrower intervals is necessary to understand how children and
adolescents may experience trauma differently.
The development of trauma symptoms following a traumatic experience also varies by
gender. A meta-analysis of studies examining sex differences in the relation between
experiencing potentially traumatic events and developing posttraumatic stress symptoms found
that males are more likely to report experiencing a traumatic event (Tolin & Foa, 2006). Sex
differences were also found based on the type of traumatic experiences reported. Females were
more likely to experience sexual abuse during adulthood or childhood compared to males (Tolin
& Foa, 2006). Males and females were equally likely, however, to experience nonsexual abuse

16

during childhood (Tolin & Foa, 2006). After experiencing a potentially traumatic event, females
are twice as likely as males to develop symptoms of posttraumatic stress (Tolin & Foa, 2006).
This research examining gender differences in trauma symptoms has not examined child
populations. However, it is likely that similar gender differences would emerge within a sample
of maltreated children.
Overall, trauma results when these events and experiences cause symptoms of
posttraumatic stress that significantly interfere with functioning. Symptoms of posttraumatic
stress include re-experiencing, avoidance, negative cognitions and mood, and arousal (APA,
2013). Re-experiencing a traumatic event includes any type of undesired and repetitive memory
of the event. After experiencing a form of maltreatment, a child may have spontaneous memories
of the event (e.g., flashbacks), recurrent dreams related to the event, or prolonged mental
repetition of the events. Avoidance includes purposeful avoidance of any external reminder of
the event. A child with a history of maltreatment may try to avoid the physical location abuse
occurred (e.g., the family home) or may experience significant psychological distress in
situations similar to the maltreatment (e.g., close physical proximity to an adult male). Negative
cognitions and mood includes significant myriad feelings caused by the traumatic event.
Maltreatment may cause a child to have a distorted sense of blame of self or others, diminished
interest in previously enjoyable activities, or a persistent negative mood. Arousal includes
aggressive or reckless behavior, sleep disturbances, and hypervigilance. A child with a history of
maltreatment may engage in aggressive behavior toward others, especially in a defensive
manner, as a result of experiencing trauma. Additionally, children may have difficulty sleeping
or relaxing after experiencing maltreatment.

17

While many of these symptoms represent typical responses to potentially traumatic
experiences, they become problematic when they interfere with functioning. This distress may
affect social functioning, work or academic functioning, or other important areas. Significant
trauma may be expressed by exhibiting many different types of symptoms, greater severity of
symptoms, or greater impairment in one or more area of functioning. Child maltreatment and
resulting trauma significantly impairs children’s functioning in multiple domains, particularly
academic well-being and social-emotional well-being.
Academic Well-Being
While trauma appears to affect both social-emotional and academic well-being, less
research has been conducted examining the academic effects of trauma and child maltreatment
compared to behavioral effects. Studies have shown that child maltreatment can affect multiple
areas of academic performance across schooling years (Kendall-Tackett & Eckenrode, 1996). In
early childhood, children with a history of maltreatment show greater problems with receptive
and expressive language skills, important early indicators of academic performance (Allen &
Oliver, 1982). This places maltreated children at significant disadvantage as they enter
Kindergarten. Academic deficits continue through postsecondary education. Maltreated children
are less likely to complete secondary school qualifications or gain a university degree than
nonmaltreated children (Boden, Horwood, & Fergusson, 2007). Two primary domains affected
by trauma and maltreatment include academic achievement and academic engagement.
Academic Achievement. During school years, maltreated children tend to have poorer
academic achievement compared to nonmaltreated children. Academic achievement may be
measured through letter grades, performance on standardized tests of achievement, or other
outcome measures. In a study comparing abused and neglected children to nonmaltreated

18

counterparts, maltreated children had significantly lower grades and were more likely to repeat a
grade level (Kendall-Tackett & Eckenrode, 1996).
A study examining predictors of standardized academic achievement as measured by the
Wide Range Achievement Test (WRAT) assessed sociodemographic variables, preschool
experience, maternal factors, and abuse status among abused and matched non-abused
counterparts. Among the sample of 165 abused children and 169 non-abused children,
sociodemographic variables accounted for 15.8% of variance in reading achievement, 15.6% of
variance in spelling scores, and 15.2% in mathematics scores (Kinard, 1999). Adding preschool
experience and maternal factors did not have a significant effect on outcome measures. Abuse
status significantly improved the model and accounted for an additional 1.3% change in variance
for reading scores, an additional 1.6% for spelling, and an additional 1.2% for mathematics
(Kinard, 1999). Abused children had significantly lower academic achievement scores than nonabused children within all three academic domains (Kinard, 1999).
Academic Engagement. Academic engagement refers to self-directed behaviors related
to pursuing academic tasks. This includes behaviors such as pursuing optimally challenging
tasks, paying attention, and completing assignments. Academic engagement is an important
aspect of children’s functioning as it predicts academic achievement (Connell & Wellborn,
1991). Among all children, academic engagement tends to decline during middle school and high
school (Juvonen, Espinosa, & Knifsend, 2012). Children with a history of maltreatment,
however, tend to have even lower academic engagement compared to nonmaltreated children.
Preschool and school-age children with a history of maltreatment generally show lower levels of
academic engagement in school as measured by parent and teacher report (Aber, Allen, Carlson
& Cicchetti, 1989; Toth & Cicchetti, 1996). A study of 229 children including 146 children with

19

a history of maltreatment examined the relation between maltreatment and academic and
behavioral adjustment. Children with a history of maltreatment exhibited lower academic
engagement than nonmaltreated children as assessed by teacher ratings and school records
(Shonk & Cicchetti, 2001).
Furthermore, academic outcomes are significantly affected by the cognitive and socialemotional effects of trauma. Trauma following maltreatment is related to many areas of
cognitive functioning including attention, memory, and planning. These cognitive abilities are
necessary for students to attend to, benefit from, and remember academic content. Students with
a history of maltreatment therefore may likely have significant difficulty engaging in academic
tasks.
Trauma can also result in internalizing and externalizing behavior problems which affect
academic performance. Externalizing behavior problems may cause students to engage in
aggressive behaviors in school that may lead to disciplinary action such as suspensions.
Internalizing behavior problems may cause students to become withdrawn which may lead to
less engagement with teachers or peers. These social-emotional symptoms of trauma may make
it difficult for students to attend and benefit from instruction.
Trauma and child maltreatment, thus, are predictive of poorer academic achievement and
engagement, but current research has not examined this relation within a large sample of
maltreated youth. These academic outcomes are important not only for learning and development
during childhood, but also for opportunities such as postsecondary education and employment
into adulthood. Understanding how trauma affects these outcomes can aid efforts to promote
resilient outcomes following maltreatment.

20

Social-Emotional Effects of Trauma
Trauma appears to have significant negative effects on children’s social and behavioral
functioning. Children with a history of maltreatment, both abuse and neglect, often experience
internalizing problems and externalizing problems.
Internalizing. Children with a history of maltreatment often exhibit more internalizing
behavior problems compared to nonmaltreated children (Hildyard & Wolfe, 2002). Those with a
history of neglect tend to demonstrate even more internalizing problems than both abused
children and nonmaltreated children (Hildyard & Wolfe, 2002). Neglected children are often
socially withdrawn and do not engage with peers during preschool years (Crittenden, 1992).
During early school years, neglected children tend to continue to have social difficulties and
avoid peer interactions (Kaufman & Cicchetti, 1989). Furthermore, children who experience
chronic maltreatment exhibit greater internalizing problems than children who experience
situational maltreatment (Jaffee & Maikovich-Fong, 2011).
Externalizing. Maltreatment also predicts externalizing behaviors. Children with a history of
maltreatment (abuse or neglect) often exhibit more externalizing behavior problems compared to
nonmaltreated children (Hildyard & Wolfe, 2002). This is especially true of physically abused
children. Children with a history of abuse tend to demonstrate more externalizing problems than
both neglected children and nonmaltreated children (Hildyard & Wolfe, 2002). Physically
abused children are often physically aggressive toward siblings, parents, peers, and teachers,
even more so than children exposed to different types of maltreatment (Crittenden, 1992). Within
a longitudinal study of externalizing behavior problems, more than half (61%) of preschool
children who were the subject of a child welfare investigation demonstrated low or normal
problem behavior trajectories over a six-year time period, while 12% demonstrated persistently

21

high or clinical behavior problems, 23% demonstrated improving externalizing behaviors, and
4% demonstrated worsening problem behaviors (Woodruff & Lee, 2011). Another longitudinal
analysis found that repeated maltreatment also predicted trajectories of externalizing behavior
problems (Li & Godinet, 2013). Children with repeated maltreatment exhibited more significant
behavior problems over time, compared to children with only one child maltreatment
investigation (Li & Godinet, 2013). Overall, children with a history of maltreatment exhibit more
significant externalizing behavior problems, especially those with a history of physical
maltreatment or repeated maltreatment.
Resilience
Although children often experience negative academic and social-emotional outcomes
following trauma such as child maltreatment, some children experience better adjustment than
others (Luthar, 2006; Masten & Obradovic, 2006). These differences can be explained by the
theoretical construct of resilience. Resilience refers to a dynamic process in which individuals
exposed to significant adversity are able to overcome these risk factors to maintain positive
adjustment (Masten, 2007, 2011; Luthar, Cicchetti & Becker, 2013). This definition of resilience
requires two components: risk (the presence of a serious threat to adaptation or development) and
resilience (achievement of positive adaptation and good outcomes). For example, children
exposed to maltreatment who are able to maintain positive adaptation would be considered
resilient.
Previously, researchers thought that resilience was achieved through “special”
characteristics of certain children. More recent empirical and theoretical work in this area,
however, has demonstrated that resilience is an “ordinary” process that results from normal
human adaptation systems (Masten, 2001). These human adaptation systems include both

22

individual factors and social-ecological systems that facilitate development with or without the
presence of risk. These factors may include an individual’s coping strategies, a strong social
network, and the availability and accessibility of culturally relevant services in the community.
In the presence of risk, these adaptation systems mitigate the potential harm caused by adversity
and stress.
It is important to emphasize that resilience is a process, rather than a characteristic or
event. These processes may operate before exposure to adversity by promoting the individual’s
capacity to withstand and cope with future adverse experiences. Examples of these types of
factors may include opportunities to develop autonomy or responsibility and successful coping
with previous stressors. Conversely, these processes may operate after exposure to adversity to
promote positive adaptation. Examples of these types of factors may include external changes
(e.g., removal from the abusive environment) or personal actions (e.g., seeking support through
friendships or professional counseling). Therefore, these influences on resilience may operate
immediately after experiencing adversity such as maltreatment or may operate over a
considerable time span (Rutter, 2013). As individuals experience resilience differently, it is
expected that resilient outcomes may be observed after different lengths of time following
exposure to adversity. Generally, however, research studies assessing resilience examine
functioning outcomes approximately one year after exposure to adversity. This study assessed
indicators of resilience 18 months after the report of child maltreatment.
Among victims of child maltreatment, resilience has been defined as achieving a normal
range of competencies across multiple domains of functioning (Walsh, Dawson & Mattingly,
2010). This study did not measure resilience directly, but rather assessed two important areas of
children’s functioning, academic and social-emotional well-being, as indicators of resilience.

23

Academic engagement and academic achievement were measured to assess children’s
academic functioning. Academic engagement refers to students’ active participation and effort
directed toward academic tasks and goals. Academic engagement is generally assessed as a selfreport of behaviors such as completing homework, paying attention in class, and desiring to try
to do well in school. Academic achievement refers to students’ actual performance on academic
tasks and their academic knowledge and skills. Academic achievement is generally assessed
using class grades or standardized measures of academic achievement. These measures predict
later academic outcomes including postsecondary attendance (Scott & Ingels, 2007), freshmen
and cumulative college grades (Geiser & Santelices, 2007), and highest degree attained (Scott &
Ingels, 2007). Assessing these indicators of academic and social-emotional functioning provided
a measure of current functioning, and an indicator of future educational and mental health
outcomes.
Externalizing behaviors and internalizing behaviors were measured to assess children’s
social-emotional functioning. These behaviors are often assessed through rating scales completed
by caregivers, teachers, or children. Informants reflect on the child’s behavior from a set period
of time (i.e., last 2 weeks or last 2 months) and rate the frequency of specific behaviors.
Externalizing behaviors include engaging in outward aggression toward others and internalizing
behaviors include outward indicators of withdrawal or depression. Longitudinal analyses have
consistently found that behavior problems in childhood and adolescence predict psychopathology
in adulthood (Hofstra, Vand Der Ende, & Verhulst, 2002; Reef, 2010; Loth, Drabick, Leibenluft,
& Hulvershorn, 2014).

24

Within resilience literature, the factors that are found to be related to more resilient
outcomes are referred to as protective factors. Significant research since the early 2000s has
examined the role of many different characteristics and factors that appear to promote resilience.
These protective factors include characteristics at the individual, family, and community level.
Individual-level protective factors include characteristics such as personality traits, intellect, selfefficacy, coping skills, and having an internal locus of control (Ungar, Ghazinour, & Richter,
2013). Family-level protective factors include having sufficient resources such as time and
money, having supportive caregivers and caregiving environments, and families having adaptive
coping strategies (Ungar, Ghazinour, & Richter, 2013). Community-level protective factors
include positive peer relationships, nonfamily member relationships and social support, and
religion (Ungar, Ghazinour, & Richter, 2013). The presence of these positive protective factors
has been shown to predict more resilient outcomes among victims of maltreatment
Social Support
Social support has been found to be a significant protective factor that promotes resilient
outcomes in the presence of risk. Social support has been defined and conceptualized in many
ways. Social support includes the existence of social relationships, the structure of social
networks, and the functions of social support (House, Kahn, McLeod, & Williams, 1985).
Gottlieb (1983) defines social support broadly as “verbal and non-verbal information or advice,
tangible aid, or action that is offered by social intimates or inferred by their presence and has
beneficial emotional or behavioral effects on the recipients” (p. 28).
Types of social support include emotional, instrumental, informational, and appraisal
support (Barrera, 1986). Emotional support involves a subjective feeling that one is cared for,
loved, and valued (Langford, Bowsher, Maloney, & Lillis, 1997). It results from positive social

25

interactions that cause positive affect and a sense of belonging and acceptance. Students
receiving emotional support would feel that they are valuable members of their school
community and are cared for and loved by their guardians. Instrumental support involves the
provision of tangible support in the form of goods and services (Langford et al., 1997). This form
of support is concrete and can be measured more objectively. Students receiving instrumental
support may obtain counseling services from a school employee or other professional.
Information support involves the provision of information during a time of stress to support
problem-solving (Langford et al., 1997). Students receiving informational support may receive
instruction in deep breathing techniques to manage stress or may receive advice regarding how to
manage or alleviate a situation causing stress. Appraisal support involves provision of
affirmations regarding self-evaluation (Langford et al., 1997). Students receiving appraisal
support may receive feedback from a teacher commending the student for appropriately
responding to an interpersonal conflict.
While all types of social support are important and enable individuals to receive helpful
support, emotional support has been regarded as the most significant because it is more salient
than other forms of support. When individuals describe supportive acts they have received, they
primarily describe examples of emotional acts rather than other tangible forms of support
(Gottlieb, 1978). This suggests that individuals regard emotional support as the most important
type of social support. A primary characteristic of social support therefore includes actions that
cause an individual to feel cared for and emotionally supported.
Another important distinction regarding social support is the difference between
perceived and enacted support. Gottieb’s (1983) definition includes these forms of support as
“offered” and “inferred” support. Perceived support involves an individual’s perception

26

regarding the availability of support if needed (Barrera, 1981; 1986). Enacted support involves
the actual occurrence of a socially supportive exchange (Barrera, 1981; 1986). Enacted support
refers to direct actions taken by others with the intent to alleviate others’ distress (Gleason &
Iida, 2015). The course of enacted support has been modeled as a three-stage process (Pearlin &
McCall, 1990). First, the need for support must be communicated directly or indirectly to the
provider. Second, the provider must be able to recognize the need for support and determine
whether he or she can provide it. Third, the actual support is either provided or not provided.
This model hypothesizes that each of these three stages must be enacted appropriately for
successful provision of social support.
Perceived social support does not require each of the stages of enacted social support.
Perceived social support is defined as an individual’s beliefs regarding the availability of
different types of social support from their social ties (Gottlieb & Bergen, 2010). This includes
perceptions of care, respect, acceptance, and relatedness from familial and nonfamilial
relationships. It is the belief or understanding that effective support is available from social
relationships when it is needed. This belief may be consistent or inconsistent with the actual
social support received from social networks. A meta-analysis examining the effect sizes of
studies comparing enacted and perceived social support found that these constructs are weakly
correlated with an average correlation of r = .35, p < .001 (Haber, Cohen, Lucas, Baltes, 2007).
Perceived social support may be influenced by an individual’s perception of relationships,
judgement of the value or potential effectiveness of a supportive relationship, or memory of
previous support needed or received (Laskey & Drew, 1997). This may cause individuals to
perceive that they have more or less social support available than they actually receive.

27

While perceived social support may not accurately measure the degree of support
available and provided to an individual, it is an important construct to assess. Research on
perceived and enacted support has found that individuals’ sense of supportive relationships is a
significant buffer against adversity and stress, possibly even more so than enacted social support
(Gottieb & Bergen, 2010).
This study conceptualizes social support as perceived, rather than enacted, support from
social relationships. Furthermore, emotional forms of social support were emphasized in this
study, rather than tangible forms of support. Social support was measured as the individual’s
perception regarding the availability of emotional support if needed and their beliefs regarding
the supportiveness of their social relationships and environment.
Perceived social support has been one of the most consistently replicated correlates of
resilient outcomes in the presence of trauma (Brewin, Andrews & Valentine, 2000). Studies
examining social support have primarily assessed the role of social support as a main effect;
limited research has studied social support as a buffer or moderator of resilience (Wilson &
Scarpa, 2014). The existing research, however, provides strong evidence that children with
greater perceived social support have more resilient outcomes, including among maltreated
children. Children with greater available social support following maltreatment have fewer
depressive symptoms at follow-up (Dingfelder, Jaffee, & Mandell, 2010) and fewer
posttraumatic stress symptomatology (Muller, Goebel-Fabbri, Diamond, & Dinklage, 2000).
Maltreated children with low social support during childhood, however, are more likely than
nonmaltreated children to become abusers during adulthood (Litty, Kowalski, & Minor, 1996).
This indicates that perceived social support is particularly important in developing positive
outcomes among children with a history of maltreatment.

28

Children may experience social support from a variety of relationships. Two salient
relationships in children’s’ lives include relationships with caregivers and peers. These different
sources of support may both provide children with perceived emotional support. Relationships
with caregivers and friends can produce close personal connections. Children may disclose
personal information such as their feelings, desires, and concerns with individuals in a caregiver
role or a friend role. Caregivers and peers may also both express positive affect toward the child.
While both of these relationships may provide children with perceived emotional support, they
may do so in different ways. Furthermore, children may rely on different sources of support
differentially based on their developmental stage, life circumstances, or type of social support
needed.
Caregiver Social Support. Relationships with caregivers are typically children’s first
and primary personal relationship. Children develop attachment relationships with caregivers as
their emotional needs are met. Caregivers generally provide a sense of security, love, and
warmth toward children. This provides children with emotional forms of social support they
need.
Peer Social Support. Relationships with peers typically become more salient once
children begin attending school or day care. Children begin to rely more on peer relationships
compared to caregiver relationships as they get older. Among children with a history of
maltreatment, peer relationships may provide a more secure source of support than caregiver
relationships. After allegations of maltreatment, children may be removed from their biological
or primary caregiver’s home and may experience transition in their living environment.
Furthermore, the child’s caregiver may be the subject of a child welfare investigation causing a
lack of trust for their primary caregiver or others in a caregiver role. Without this trust and

29

reliability, children may not seek social support from caregivers. Instead, they may seek and
benefit from greater social support from peers. Peer social support may serve a unique purpose
among children with a history of maltreatment.
Purpose of the Present Study
The current literature provides sufficient evidence that children with a history of
maltreatment tend to experience symptoms of trauma and poorer social-emotional and academic
outcomes, compared to children with no history of maltreatment. The literature also supports the
positive role of social support in promoting resilience and positive outcomes among children
including those with a history of maltreatment. Although these issues have been studied within
previous research, several gaps still exist within current literature.
This study addressed three primary gaps within current trauma and resilience literature.
First, current research primarily measures social support between the child and caregiver. This
study assessed both caregiver and peer social support as potential protective factors. By
examining both perceived caregiver social support and perceived peer social support, this study
can draw conclusions relevant to both the home and school environment.
Second, studies of trauma and resilience have primarily utilized social-emotional
outcome variables such as delinquency, external behavior problems, and Posttraumatic Stress
Disorder symptoms. This study assessed both social-emotional (i.e., internalizing behavior
problems and externalizing behavior problems) and academic outcomes (i.e., academic
achievement and academic engagement). By examining both social-emotional and academic
outcomes, this study examined indicators of resilience among a broader range of domains
relevant to children’s lives. Addressing these gaps aided in providing a more comprehensive
understanding of the role of social support in moderating the relation between trauma and

30

outcomes relevant to children’s lives. Additionally, this provided information relevant to both the
home and school settings.
Third, most extant studies utilize retrospective, self-report interviews with adult samples,
rather than assessing measures with child samples. This study addressed this gap by assessing
children’s current perceived trauma symptoms, perceived social support, and outcomes. By
examining current trauma symptoms, social support, and outcomes, this study aimed to obtain
more reliable and valid estimates of these variables. This provided more appropriate conclusions
relevant to children at the time of maltreatment.
Research Questions and Hypotheses
The following research questions were assessed within this study. The first two questions
address the role of social support from caregivers and peers as a protective factor following child
maltreatment. The third and fourth questions addressed how maltreatment and demographic
characteristics may predict perceived trauma symptoms and perceived social support.
Research Question 1. Does social support serve as a protective factor, moderating the
relation between trauma and academic and social-emotional well-being?
Hypothesis 1. As children’s perceived social support from caregiver and peer
relationships increases, the negative relation between trauma and outcomes will decrease.
Research Question 2. Is caregiver social support or peer social support a more important
moderator between trauma and outcomes within the overall conceptual model?
Hypothesis 2. Peer social support will exhibit a stronger moderating relationship within
the conceptual model compared to caregiver social support, as measured by parameter estimates
and goodness of fit indices.

31

Research Question 3. Do children’s perceived trauma symptoms vary based on type of
maltreatment, severity of maltreatment, substantiation, age, or gender?
3a. Do children’s perceptions of trauma differ between children referred primarily for
abuse (physical abuse, emotional abuse, sexual abuse) versus neglect (physical neglect, parental
domestic violence/substance abuse)?
Hypothesis 3a. Children with investigations primarily due to abuse will exhibit greater
trauma symptoms compared to those with investigations due to neglect.
3b. Do children’s perceptions of trauma differ based on the level of harm they
experienced as reported by caseworkers?
Hypothesis 3b. Children judged to have experienced moderate to severe harm will
exhibit greater trauma symptoms compared to those with mild or low harm as reported by the
caseworker.
3c. Do children’s perceptions of trauma differ based on the determination of
substantiation of maltreatment?
Hypothesis 3c. Children with substantiated investigations of maltreatment will exhibit
greater trauma symptoms compared to those with unsubstantiated investigations of maltreatment.
3d. Do children’s perceptions of trauma differ based on age?
Hypothesis 3d. Younger children will exhibit greater trauma symptoms compared to
older adolescents.
3e. Do children’s perceptions of trauma differ between males and females?
Hypothesis 3e. Females will exhibit greater trauma symptoms compared to males.
Research Question 4. Do children’s perceived social support vary based on type of
maltreatment, severity of maltreatment, substantiation, age, or gender?

32

4a. Do children’s perceptions of social support differ between children referred primarily
for abuse (physical abuse, emotional abuse, sexual abuse) versus neglect (physical neglect,
parental domestic violence/substance abuse)?
Hypothesis 4a. Children with investigations due to neglect will report greater support
from caregivers compared to those with investigations due to abuse. Children with investigations
due to abuse will report greater support from peers compared to those with investigations due to
neglect.
4b. Do children’s perceptions of social support differ based on the level of harm they
experienced as reported by caseworkers?
Hypothesis 4b. Children with moderate to severe harm will perceive less social support
from both caregivers and peers compared to children with mild or no harm as reported by
caseworkers.
4c. Do perceptions of social support differ based on the determination of substantiation of
maltreatment?
Hypothesis 4c. Children with substantiated investigations of maltreatment will perceive
less social support from both caregivers and peers compared to children with unsubstantiated
investigations of maltreatment.
4d. Do perceptions of social support differ based on age?
Hypothesis 4d. Younger children will perceive greater support from caregivers
compared to older children. Older children will perceive greater support from peers compared to
younger children.
4e. Does perceived social support differ between males and females?

33

Hypothesis 4e. Females will report greater perceived support from peers compared to
males. Males will report greater perceived support from caregivers compared to females.

34

CHAPTER 3
METHODS
Study Model
Two conceptual models were proposed (see Appendix A). The conceptual models depict
the hypothesized relation between trauma, social support, and academic and social-emotional
outcomes through caregiver relationships and peer relationships. Model 1 evaluates the latent
variable of caregiver social support as a moderator while Model 2 evaluates the latent variable of
peer social support as a moderator.
Overall, the conceptual model (Figure 1) represents the hypothesis that social support
serves a unique and important role among children with a history of maltreatment. Following
previous literature on resilience and social support, it was hypothesized that social support
moderates the relation between trauma symptoms and outcomes. To assess moderation,
interaction terms between each form of social support and trauma were assessed. This evaluated
whether social support acts as a protective factor, buffering children from the potentially harmful
outcomes associated with trauma and maltreatment. Children with greater support from
caregivers and peers were hypothesized to have greater social resources to cope with the trauma
and would therefore be more likely to overcome adversity.
The proposed study models also depict the hypothesis that caregiver social support and
peer social support serve different roles in the lives of children with a history of maltreatment.
Consequently, caregiver and peer social support were assessed as separate latent variables.
Children engage in relationships with caregivers and peers differently and may meet different
needs through these types of relationships. These differences may be even more prominent
among children with a history of maltreatment. Therefore, it was hypothesized that while

35

caregiver social support and peer social support may both act as protective factors, they may
relate differently to the overall conceptual model. Children with a history of maltreatment,
particularly adolescents, may rely more on relationships with peers compared to relationships
with caregivers. Social support from peer relationships may be more important than caregiver
social support.
Predictors of perceived trauma symptoms were also identified within the study model. As
previously discussed, individuals respond to potentially traumatic events differently. Previous
research has identified factors that predict the risk of developing symptoms of trauma following
a potentially traumatic event. It was hypothesized that perceived trauma symptoms can be
explained by the type of maltreatment, severity of maltreatment, substantiation of maltreatment,
age, and gender. Similarly, the study model includes predictors of social support. These variables
may also influence how the child develops social relationships and the nature of those
relationships. It was hypothesized that perceived social support can be explained by the type of
maltreatment, severity of maltreatment, substantiation of maltreatment, age, and gender.
Overall, symptoms of trauma predict academic and social-emotional well-being. Children
who experience significant trauma following maltreatment generally experience negative
outcomes. Trauma affects many areas of functioning which may negatively influence outcomes
including academic and social-emotional well-being. It was hypothesized that trauma predicts
academic well-being (achievement and engagement) and social-emotional well-being
(internalizing and externalizing behaviors). Children with greater trauma are more likely to
experience poorer academic and social-emotional outcomes.

36

Figure 1
Conceptual Model

Study Design
Data from the second National Survey of Child and Adolescent Well-Being (NSCAW II)
were utilized for this study. The initial NSCAW was a longitudinal study administered and
funded by the Administration for Children and Families of the U.S. Department of Health and
Human Services. The purpose of the NSCAW was to study outcomes of abused and neglected
children and their involvement with the child welfare system. Although other national
longitudinal studies of child welfare had been conducted, the NSCAW was the first to collect
data from children and families and to relate well-being to a wide range of individual, family,
community, and agency factors. The NSCAW study was conducted between October 1999 and

37

December 2007. The NSCAW II study comprised a new cohort of children and included
essentially the same design as the NSCAW and was conducted between February 2008 and
December 2012.
The NSCAW study was designed and implemented through collaboration and
consultation with various research entities and individuals. Collaborators included the Research
Triangle Institute (RTI), University of California at Berkley (UCB), Caliber Associates (Caliber),
University of North Carolina at Chapel Hill, and Child and Adolescent Services Research Center
at San Diego Children’s Hospital. A team of experts in child welfare, longitudinal data
collection, agency services research, developmental psychology, and data analysis met in April
1997 to design the procedures of the NSCAW study. The Technical Workgroup engaged in
ongoing advising and consultation during the study and included experts in child welfare
agencies and systems, social welfare policy, child and youth development, and other areas.
Data from the NSCAW and NSCAW II are available through licensing agreements with
the National Data Archive on Child Abuse and Neglect (NDCAN) at Cornell University. A
General Release data set and a Restricted Release data set are available. The General Release
data set contains variables that were removed or altered to further ensure de-identification of
participants. The Restricted Release data set, while still including de-identified information,
contains a greater number of variables and specific responses from variables. This study utilized
the Restricted Release data set of the NSCAW II to understand the role of social support in
fostering resilience within a large sample of maltreated youth.
Data Collection. The NSCAW II was conducted over multiple waves of data collection
using multiple informants. Three waves of data collection were conducted between February
2008 and December 2012. Wave 1, or baseline data collection, was completed between March

38

2008 and September 2009. Wave 2 was completed between October 2009 and January 2011, 18
months after the close of the child welfare investigation of suspected maltreatment. Wave 3 was
completed between June 2011 and December 2012, 36 months after the close of the child welfare
investigation. NSCAW II includes a cohort of 5,872 children, aged birth to 17.5 years old who
had contact with the child welfare system within a 15-month period. Children, parents, nonparent
caregivers (e.g., foster parents, kin caregivers, group home caregivers), teachers, and
investigative caseworkers completed interviews to provide information about the child during
each wave of data collection.
Original NSCAW Sample
The original NSCAW sample design included two stages of a stratified sample. In the
first stage, the U.S. was divided into nine strata. Four states were excluded due to requiring child
welfare agencies to make initial contact with potential participants, rather than NSCAW
researchers. Eight strata comprised the 8 states with the largest number of child welfare cases.
The ninth strata included the remaining 38 states. Primary Sampling Units (PSUs) were created
by geographic areas that encompass the population of one child protective service agency. A
random selection of PSUs was conducted. This resulted in 92 responding and eligible PSUs
within the original NSCAW study. Within the NSCAW II sampling, the same counties were
approached for participation resulting in 76% retention of the original PSUs.
In the second stage, cases were randomly selected from each PSU. Each PSU compiled
lists of children investigated or assessed for suspected abuse or neglect during a 14-month period
between February 2008 and April 2009. Only children investigated as victims were included in
sampling (rather than perpetrators). Only one child per family could be included in sampling. A

39

random sample of children and families was selected from each of the sampling domains within
each PSU.
There are several issues to consider regarding the representation of the NSCAW
sampling. Sampling procedures oversampled some populations (e.g., infants and cases receiving
ongoing services) to provide a sufficient sample size of these subgroups. Additionally, some
PSUs did not allow inclusion of cases that were unsubstantiated following the investigation.
Sampling also excluded agencies with few children investigated for suspected abuse and neglect
and excluded PSUs that required the child welfare agency to contact potential participants before
the researchers. Additionally, administrative files may have had incomplete lists of cases
investigated for abuse and neglect at the time of sampling. These issues could have created
noncoverage bias. NSCAW researchers examined these issues and compared data to other
representative national longitudinal studies. They found that there were small bias estimations
for differences between the sampled group and the estimated population regarding age, gender,
race/ethnicity, and substantiated or unsubstantiated cases of abuse and neglect.
Demographic characteristics of the original NSCAW II sample are provided in Table 1
(N=5,872). The NSCAW II data set includes children from less than one year old (31.4%) to 17
years old (1.1%). The sample includes a greater percentage of younger children compared to
older adolescents. The NSCAW II sample included approximately half male (51.4%) and half
female (48.6%) participants. The majority of the unweighted sample were White/Non-Hispanic
(41.7%), while significant proportions were Hispanic (27.7%) and Black/Non-Hispanic (22.7%).

40

Table 1
Demographic Characteristics of NSCAW II
Characteristic
Sample N Sample %
Child Age
0
1,845
31.4%
1
763
13%
2
329
5.6%
3
290
4.9%
4
277
4.7%
5
261
4.4%
6
220
3.7%
7
235
4.0%
8
218
3.7%
9
181
3.1%
10
199
3.4%
11
170
2.9%
12
156
2.7%
13
165
2.8%
14
166
2.8%
15
167
2.8%
16
164
2.8%
17
66
1.1%
Child Gender
Male
3,017
51.4%
Female
2,855
48.6%
Child Race/Ethnicity
Black/Non-Hispanic
1827
31.2%
White/Non-Hispanic
2003
34.2%
Hispanic
1614
27.6%
Other
407
7.0%

Weighted N

Weighted %

170,448
162,454
177,959
225,884
167,705
165,368
146,127
149,135
144,667
117,450
125,118
120,642
116,206
124,225
108,283
110,125
93,990
49,060

6.9%
6.6%
7.2%
9.1%
6.8%
6.7%
5.9%
6.0%
5.8%
4.7%
5.1%
4.9%
4.7%
5.0%
4.4%
4.4%
3.8%
2.0%

1,258,849
1,215,997

50.9%
49.1%

561,361
1,032,610
685,894
186,274

22.7%
41.7%
27.7%
7.5%

Final Sample for Current Study
The current study will utilize a sample of children ages 11 to 17 years old children at the
first wave of the NSCAW II (2008-2012) data set who had valid measures of trauma symptoms,
caregiver social support, peer social support, and internalizing and externalizing behaviors, direct
assessment of reading and mathematics achievement, parent report of internalizing and
externalizing behaviors, and a sample weight variable to estimate population statistics. The final

41

sample meeting these criteria consisted of 697 children. Trauma and social support were assessed
at Wave 1 of data collection. Academic and social-emotional well-being were assessed at Wave
2 of data collection. Of the sample of 697 children, 497 completed measures at the time of Wave
2 sampling. Unweighted and weighted demographic characteristics of the sample are presented
in Table 2 (N=697). Children’s ages within the estimated population ranged from 11 years
(20.7%) to 17 years (3.8%) following sampling criteria. The majority of the estimated population
was female (62.5%). White/Non-Hispanic children accounted for an estimated 41.7% of the
population, Black/Non-Hispanic children accounted for 20.3% of the population, Hispanic
children accounted for 27.8% of the population, and 10.2% of the estimated population included
children of other race/ethnicities.
Table 2
Demographic Characteristics of Sample
Characteristic
Sample N
Sample % Weighted N Weighted %
Child Age
11
133
19.1%
101,608
20.5%
12
120
17.2%
99,139
20.0%
13
109
15.6%
85,297
17.2%
14
116
16.6%
75,655
15.3%
15
107
15.4%
70,604
14.2%
16
80
11.5%
46,062
9.3%
17
32
4.6%
17,671
3.6%
Child Gender
Male
301
43.2%
193,952
39.1%
Female
396
56.8%
302,087
60.9%
Child Race/Ethnicity
Black/Non-Hispanic
182
26.1%
101,123
20.3%
White/Non-Hispanic
274
39.3%
207,323
41.8%
Hispanic
165
23.6%
136,906
27.6%
Other
76
10.9%
50,685
10.2%
Child Welfare Characteristics. In addition to demographic characteristics of this
sample, it is also important to understand participants’ experience with the child welfare system.
Table 3 presents unweighted and weighted child welfare characteristics of the study sample. The

42

majority of the estimated population included children that were the subject of a child welfare
investigation primarily due to physical abuse (33.1%) or neglect (29.4%), while a smaller
proportion of the population included children exposed to emotional abuse (8.2%), sexual abuse
(12.8%), or parental substance abuse/domestic violence (16.4%). These categories refer to the
primary type of maltreatment that was investigated by child protective services report used to
select the case in the NSCAW II sampling. Children may have experienced other forms of
maltreatment within or prior to the child welfare investigation. After recoding these variables to
dichotomous categories, 54.1% of the population experienced a form of abuse (physical, sexual,
or emotional abuse) while 45.9% experienced a form of neglect (physical neglect or parental
substance abuse or domestic violence). Child welfare caseworkers provided a subjective rating of
the severity of harm the child experienced by responding to the following question: “Regardless
of the outcome of the investigation, how would you describe the level of harm to the child?
Would you say…none, mild, moderate, or severe?” The majority of the estimated population
experienced low levels of harm (none or mild) as reported by the caseworker (73.7%), while a
still significant percentage experienced high levels of harm (moderate or severe) due to
maltreatment (19.9%). Most cases had prior reports of maltreatment (68.3%), were handled
through an investigation (80.9%), and resulted in substantiated charges (73.7%). Following the
report of alleged maltreatment, most cases did not receive child welfare services (67.1%) and
most children remained in the home with a biological parent (88.6%). Other settings included a
home with an adoptive parent, formal or informal kin-care, foster care, a group home, or another
type of out of home setting.

43

Table 3
Child Welfare Characteristics of Sample
Characteristic
Sample N Sample % Weighted N
Weighted %
Type of Maltreatment
Physical Abuse
218
31.3%
164,415
33.1%
Emotional Abuse
52
7.5%
40,569
8.2%
Sexual Abuse
110
15.8%
63,578
12.8%
Neglect
199
28.6%
146,032
29.4%
Parental Substance
Abuse/Domestic Violence
118
16.9%
81,442
16.4%
Type of Maltreatment - Dichotomized
Abuse
380
54.5%
268,562
54.1%
Neglect
317
45.5%
227,474
45.9%
Level of Harm
None
220
31.6%
250,209
50.3%
Mild
210
30.1%
146,354
23.4%
Moderate to Severe
260
37.3%
98,767
19.9%
Prior Reports of Maltreatment
Yes
477
69.4%
339,001
68.3%
No
206
29.6%
139,861
28.1%
Case Procedures
Investigation
580
83.2%
401,744
80.9%
Assessment
83
12.0%
71,779
14.4%
Assessment that later
resulted in an investigation
26
3.7%
18,166
3.7%
Other
5
0.7%
4,066
0.8%
Case Substantiated
Yes
339
48.6%
124,331
73.7%
No
337
48.4%
347,528
26.3%
Services Received
Yes
408
58.5%
163,579
32.9%
No
289
41.5%
332,460
67.1%
Child Setting
In-Home: Bio Parent
466
66.8%
415,662
88.6%
In-Home: Adoptive Parent
25
3.6%
16,388
3.5%
Formal Kin Care
48
6.9%
11,202
2.4%
Informal Kin Care
47
6.9%
30,122
6.4%
Foster Care
77
11.0%
15,029
3.2%
Group Home/Res Program
31
4.4%
6,414
1.4%
Other OOH Arrangement
7
1.0%
1,219
0.2%

44

Sample Criteria. Data utilized for this study were limited to children between the ages
of 11 and 17.5. This study aimed to investigate the role of social support among adolescents.
This developmental period is characterized by greater dependence on peer relationships and less
dependence on family relationships for social support. Therefore, it was hypothesized that peer
social support may promote resilience among adolescents affected by abuse and neglect.
Additionally, several of the variables of interest were only administered to children age 11 years
and older (e.g., Rochester Assessment Package for Schools). Excluding children younger than 11
years old allowed for the use of these measures. The maximum age of children included in Wave
1 of the NSCAW II was 17.5. This was used as the maximum age within this sample to allow for
a greater sample size and include participants in the developmental periods of early adolescence
through late adolescence.
The current sample was limited based on the type of maltreatment. Case workers reported
the primary type of maltreatment of the investigation selected from NSCAW II sampling from 16
categories (e.g., physical maltreatment, neglect, abandonment, domestic violence, voluntary
relinquishment). Child welfare investigations in which abuse or neglect were not the primary
cause of maltreatment (e.g., abandonment, educational maltreatment, child in need of services)
were excluded from the sample as to not confound results (N=326). Within the NSCAW II data
set, physical neglect and lack of supervision neglect were categorized separately. For the purpose
of this study, these types of maltreatment were combined to form one category of neglect.
Similarly, domestic violence and substance abuse by a parent were combined to form one
category of parental domestic violence/substance abuse. This allowed for consistent categories in
these areas with a sufficient sample size to include in analyses. The current sample was restricted

45

to the following five types of maltreatment: physical abuse, emotional abuse, sexual abuse,
neglect, and parental domestic violence/substance abuse.
Given the interest in examining peer social support within the school setting and
academic outcomes, children who were not attending school at the time of Wave 1 interviews
were excluded from the final sample (N=28). Finally, cases in which a sample weight variable
was not provided within the NSCAW II Wave 2 data set were excluded (N=3). No other
sampling criteria were applied.
Missing Data. Within each measure, data for individual items or composites is missing
for some participants. During child interviews, participants had the option to respond with
“Don’t know” or “Refuse to answer” for each questionnaire item. Individual items used to create
latent variables within this study had up to 24 missing responses within the 697 cases. The
majority of participants, however, had complete or mostly complete responses. Within the
trauma measure, one case was missing all items. Within the caregiver social support measure,
seven cases were missing all items. Within the peer social support measure, one case was
missing all items. Within the academic achievement measures, ten cases were missing scores for
the reading subtest and eleven cases were missing scores for the mathematics subtest. The
academic engagement measure did not have any cases missing all items. Similarly, the child
report of internalizing and externalizing behaviors did not have any missing cases. The parent
report of internalizing and externalizing behaviors was missing 12 cases. Overall, up to 2% of
cases were missing for each variable utilized in this study. To account for this missing data,
maximum likelihood estimation was used to estimate missing observations. This procedure
involves estimating values through the mean and variance of collected responses and identifying
the most probable value given the model. Using maximum likelihood estimation provided more

46

reliable parameter estimates, standard errors, and test statistics compared to other analyses to
address missing data (i.e., multiple imputation).
Sample Weights. Descriptive analyses on the final sample include unweighted and
weighted means. Unweighted descriptive statistics represent the NSCAW II population in which
each case was counted equally. Weighted descriptive statistics estimate the entire population of
children at the time of sampling, and accounts for children who were not sampled. As previously
discussed, the NSCAW II utilized complex sampling procedures, including oversampling of
selected populations. To adjust for differential selection probabilities and non-response bias,
sample weights were designed within the NSCAW II to account for the unequal probabilities of
selection within the national sample. Therefore, structural equation analyses were computed
using sample weights, which allows for results that are generalizable to the population.
Variables and Measures
Variables were constructed using selected measures from the NSCAW II data set. A
description of the measures for each predictor and outcome variable is described below in
relation to the conceptual model. These variables include demographic variables, trauma,
caregiver social support, peer social support, academic achievement, academic engagement,
internalizing behaviors, and externalizing behaviors. See Table 4 for a complete list of latent
constructs and variables.

47

Table 4
Summary of NSCAW II Measures Corresponding to Constructs
Construct
NSCAW II Variables
NSCAW II Data Source
Type of Maltreatment
Type of Abuse
Caseworker
Severity of Maltreatment
Level of Harm to Child
Caseworker
Substantiation
Substantiated Indicator
Caseworker
Age
Child age - Years
Child/Caregiver/Caseworker
Gender
Child Gender
Child/Caregiver/Caseworker
Trauma Symptoms*
Posttraumatic Stress subscale of
Child
Trauma Symptom Checklist for
Children
Caregiver Social Support* Rochester Assessment Package for Child
Schools
Peer Social Support*
Loneliness and Social
Child
Dissatisfaction Questionnaire
Academic Achievement*
Woodcock-Johnson III Tests of
Direct Child Assessment
Achievement Letter-Word
Identification and Applied
Problems subtests
Academic Engagement*
School Engagement Questionnaire Child
Internalizing Behaviors*
Youth Self Report Internalizing
Child/Caregiver
composite
Child Behavior Checklist
Internalizing composite
Externalizing Behaviors* Youth Self Report Externalizing
Child/Caregiver
composite
Child Behavior Checklist
Externalizing composite
Note: * indicates a latent construct. Confirmatory Factor Analysis was conducted on
individual items to identify constructs.
Demographics. Demographic information regarding the child and circumstances of
maltreatment were analyzed to evaluate whether children’s trauma symptoms and social support
differed based on certain characteristics. These characteristics have been shown in previous
studies to explain differences in trauma symptoms. Type of maltreatment, severity of
maltreatment, substantiation of maltreatment, age, and gender were included as demographic
variables.
Type of maltreatment was assessed with the same question used to select the sample. The
measure was a project developed question assessing the primary type of maltreatment within the

48

child protective services report used to select the case in the NSCAW II sampling. Response
options included in the sample were: physical abuse, emotional abuse, sexual abuse, neglect, and
parental domestic violence/substance abuse. These responses were dichotomized into the
categories of abuse (physical, emotional, and sexual abuse) and neglect (physical neglect and
parental domestic violence/substance abuse).
Severity of maltreatment was assessed using the caseworker’s report of the level of harm
the child experienced as a result of the maltreatment under investigation. Child welfare
caseworkers provided a subjective rating of the severity of harm the child experienced by
responding to the following question: “Regardless of the outcome of the investigation, how
would you describe the level of harm to the child?” Response options were: none, mild,
moderate, or severe. These responses were recoded into three categories of “none,” (N=220)
“mild,” (N=210) and “moderate to severe” (N=260). These categories provided comparable
sample sizes across groups and allowed for assessing differences based on the level of harm the
child experienced as judged by the caseworker. This variable was treated as an ordinal variable
as it was assessed within the original NSCAW II as an ordinal variable. Response options
included ordered categories, but the difference between response options were not equivalent.
The difference between the categories of “none” and “mild” is not equal to the difference
between the categories of “mild" and moderate to severe.”
Substantiation was assessed using the caseworker’s report of the outcome of the child
protective services report used to select the case in the NSCAW II sampling. Following a child
welfare investigation of a report of maltreatment, child welfare services determined whether the
allegation of abuse or neglect is substantiated or unsubstantiated. A determination of
substantiation indicated that credible evidence of the abuse or neglect exists. This determination

49

was made by the child welfare caseworker evaluating whether there was enough credible
evidence to suggest that the probability of child maltreatment having occurred is over 50%
(Action, 2008). A child welfare case may not be deemed “substantiated” if there is not enough
evidence to support the allegation of abuse or neglect. Therefore, this measure is not a direct
assessment of whether or not maltreatment occurred. Substantiation determinations are used to
classify individuals as abusers and neglectors within an automated tracking system and to
provide accountability to caregivers and other stakeholders regarding the outcome of a child
welfare investigation (Action, 2008). Substantiation determinations are not necessary for
children to receive child protective services, to determine the safety of children, or to initiate
court proceedings or make an adjudication decision (Action, 2008). Within the NSCAW II
dataset, caseworkers indicated whether the investigation was substantiated with a rating of “Yes”
(N=339) or “No” (N=337).
Age and gender were also measured using project developed questions completed by the
child, caregiver and caseworker. Each of these variables were created within the NSCAW II data
set by combining the corresponding child interview, caregiver interview, and caseworker
interview questions. Age represents the child’s age at the time of sampling in years with a range
of 11 years to 17 years. Gender was assessed with binary male and female response options. If
child, caregiver, or caseworker responses differed, the majority response was assigned.
Trauma. Perceived trauma symptoms were measured using the child interview responses
to a standardized measure of posttraumatic stress symptoms. Trauma symptoms were measured
at each wave of data collection in the original NSCAW II study. For the purpose of this study,
trauma symptoms were assessed at the initial wave of data collection to provide a measure of
initial trauma. The posttraumatic stress subscale of the Trauma Symptom Checklist for Children

50

(TSCC; Briere, 1996) is a frequently used measure of self-reported indicators of posttraumatic
stress symptoms. Reliability of the TSCC was r=.87 within the standardization sample.
Reliability within the current study was r=.85. This is the only measure available within the
NSCAW II data to directly assess symptoms of trauma and was only completed by children.
Children responded “1=Never,” “2=Sometimes,” “3=Lots of times,” or “4=Almost all of the
time” to ten items. These items assessed posttraumatic stress symptoms including intrusive
recollections of traumatic events, sensory re-experiencing and nightmares, dissociative
avoidance, and fears related to the traumatic experience. Specific item statements or questions
are not available within the NSCAW II dataset.
Children’s responses to the TSCC items were used as an indicator of overall symptoms of
trauma. Since individual item responses were provided within the NSCAW II data set, factor
analyses were conducted to verify the latent variable structure. Confirmatory Factor Analysis
(CFA) using the maximum likelihood parameter estimate (estimator MLR) specifying all ten
items representing the construct of trauma using Mplus 7.4 (Muthén & Muthén, 2012) was
conducted with 696 participants within the sample that completed this measure. Highly
correlated items were identified and integrated within the CFA model (T2 with T1; T7 with T6).
See Table 5 for final factor loadings and standardized and unstandardized coefficients. All items
loaded significantly on the hypothesized construct of perceived trauma symptoms. Fit indices
included RMSEA = 0.037, with 90% CI = [0.024, 0.051], and CFI/TLI fit indices close to 1,
indicating the model fits the data reasonably well. This variable will indicate the extent to which
children report experiencing symptoms of posttraumatic stress following alleged maltreatment
and is represented by a normed T-Score. Higher scores on this measure correspond to greater
symptoms of trauma.

51

Table 5
Confirmatory Factor Analysis for One-Factor Model of Trauma Symptoms
Items Unstandardized Standard Error (SE) Standardized
T1
1.000
0.000
0.620
T2
1.230
0.083
0.763
T3
1.196
0.092
0.742
T4
0.093
0.087
0.577
T5
1.319
0.069
0.818
T6
1.105
0.087
0.685
T7
0.785
0.099
0.487
T8
1.309
0.089
0.812
T9
1.417
0.083
0.879
T10
1.066
0.075
0.661
Note: All factor loadings were significant (p<.001).
Caregiver Social Support. Perceived caregiver social support was assessed by a
measure completed by the child at the initial wave of data collection. This provided an
assessment of the child’s perceived support from the primary caregiver following the report of
maltreatment. The Rochester Assessment Package for Schools (RAPS; Connell, 1990; Lynch &
Cicchetti, 1991) assessed children’s feelings about their relationship with their caregivers.
Domains assessed included autonomy, involvement, emotional support, and structure. Reliability
for this measure within the original NSCAW study was high (r=.88). Reliability within the
current study was r=.91. Children responded “1=Not at all true,” “2=Not very true,” “3=Sort of
true,” or “4=Very true” to the following 12 items about their current caregiver:
•

When I’m with my [caregiver], I feel good. How true is this?

•

When I’m with my [caregiver], I feel mad. How true is this?

•

When I’m with my [caregiver], I feel unhappy. How true is this?

•

My [caregiver] enjoys spending time with me. How true is this?

•

My [caregiver] does a lot to help me. How true is this?

•

My [caregiver] doesn’t seem to have enough time for me. How true is this?

52

•

My [caregiver] doesn’t seem to know how I feel about things. How true is this?

•

My [caregiver] trusts me. How true is this?

•

My [caregiver] doesn’t let me make any of my own decisions. How true is this?

•

My [caregiver] is fair with me. How true is this?

•

My [caregiver] doesn’t think I can do very much. How true is this?

•

I don’t know what my [caregiver] wants. How true is this?

Children’s responses to these items were used as an indicator of overall perceived social
support from their relationship with the current caregiver. Since individual items were provided
within the NSCAW II data set and a total composite was not created, factor analyses were
conducted to combine items to represent the construct of interest. CFA was conducted with
responses from the 688 participants within the sample that completed this measure. Responses
for negatively phrased items were reverse coded (C2, C3, C6, C7, C9, C11, and C12). Strongly
correlated items were identified and integrated within the CFA model (C3 with C2; C10 with C8;
C12 with C7). These items assessed similar aspects of caregiver social support or opposite
aspects (i.e., feel mad and feel unhappy) and therefore were expected to be correlated. See Table
6 for final factor loadings and standardized and unstandardized coefficients. All items loaded
significantly on the hypothesized construct of perceived caregiver social support. Fit indices
included RMSEA = 0.049, with 90% CI = [0.039, 0.059], and CFI/TLI fit indices close to 1,
indicating the model fits the data reasonably well. Higher scores on this measure correspond to
higher perceived social support from the caregiver. This measure provided a reliable indicator of
children’s perceptions of social support provided by the child’s caregiver.

53

Table 6
Confirmatory Factor Analysis for One-Factor Model of Caregiver Social Support
Standard Error
Items
Unstandardized
Standardized
(SE)
C1: Feel good
1.000
0.000
0.787
C2: Feel mad
0.917
0.058
0.722
C3: Feel unhappy
0.988
0.041
0.777
C4: Enjoys spending time with me
1.032
0.045
0.812
C5: Does a lot to help me
1.117
0.065
0.879
C6: Doesn't have time for me
0.821
0.058
0.646
C7: Doesn't know how I feel
0.706
0.070
0.556
C8: Trusts me
0.725
0.072
0.570
C9: Doesn't let me make decisions
0.572
0.061
0.450
C10: Is fair with me
0.874
0.063
0.688
C11: Doesn't think I can do much
0.808
0.068
0.635
C12: I don't know what CG wants
0.700
0.078
0.551
Note: All factor loadings were significant (p<.001).
Peer Social Support. Peer social support was assessed by another measure completed by
the child at the initial wave of data collection. The Loneliness and Social Dissatisfaction
Questionnaire for Young Children (Asher & Wheeler, 1985) assessed children’s feelings of
loneliness and whether important relationship provisions were being met within the school
setting. Reliability for this measure within the original NSCAW study was high (r=.89).
Reliability within the current study is r=.90. Children responded “1=Never,” “2=Hardly ever,”
“3=Sometimes,” “4=Most of the time,” or “5=Always” to the following 16 items:
•

It’s easy for me to make new friends at school. How often is this true about you?

•

I have nobody to talk to at school. How often is this true about you?

•

I’m good at working with other kids at school. How often is this true about you?

•

It’s hard for me to make friends at school. How often is this true about you?

•

I have lots of friends at school. How often is this true about you?

•

I feel alone at school. How often is this true about you?

54

•

I can find a friend when I need one. How often is this true about you?

•

It’s hard to get kids in school to like me. How often is this true about you?

•

I don’t have anyone to play with at school. How often is this true about you?

•

I get along with other kids at school. How often is this true about you?

•

I feel left out of things at school. How often is this true about you?

•

There are no kids at school that I can go to when I need help. How often is this
true about you?

•

I don’t get along with other kids at school. How often is this true about you?

•

I’m lonely at school. How often is this true about you?

•

I am well liked by the kids at school. How often is this true about you?

•

I don’t have any friends at school. How often is this true about you?

Children’s responses to these items were used as an indicator of overall perceived social
support from peer relationships at school. Since individual items were provided within the
NSCAW II data set and a total composite was not created, factor analyses were conducted to
combine items to represent the construct of interest. CFA was conducted with 696 participants
within the sample that completed this measure. Negatively phrased items were reverse coded
(P2, P4, P6, P8, P9, P11, P12, P13, P14, P16). See Table 7 for final factor loadings and
standardized and unstandardized coefficients. All items loaded significantly on the hypothesized
construct of perceived peer social support. Fit indices included RMSEA = 0.048, with 90% CI =
[0.041, 0.055], and CFI/TLI fit indices close to 1, indicating the model fits the data reasonably
well. This measure provided a reliable indicator of children’s perceptions of social support
provided by peers in the school setting. Higher scores on this measure correspond to greater
perceived support.

55

Table 7
Confirmatory Factor Analysis for One-Factor Model of Peer Social Support
Unstandardized Standard Error (SE) Standardized
Items
1.000
0.000
0.665
P1: Easy to make friends
0.949
0.086
0.631
P2: Nobody to talk to
0.941
0.064
0.626
P3: Good at working with other kids
1.027
0.072
0.683
P4: Hard to make new friends
1.171
0.073
0.779
P5: Have lots of friends
1.241
0.078
0.826
P6: Feel alone
1.034
0.069
0.688
P7: Can find a friend
1.119
0.056
0.744
P8: Hard to get kids to like me
1.120
0.087
0.745
P9: Don't have anyone to play with
0.969
0.074
0.644
P10: Get along with others
1.124
0.059
0.748
P11: Feel left out
0.883
0.073
0.587
P12: No kids to go to
0.992
0.072
0.660
P13: Don't get along with kids
1.210
0.069
0.805
P14: Lonely at school
1.035
0.057
0.688
P15: Well-liked by kids
1.124
0.080
0.747
P16: Don't have any friends
Note: All factor loadings were significant (p<.001).
Academic Achievement. Academic achievement was assessed through a standardized
measure of achievement completed approximately 18 months after the initial wave of data
collection. This measure provided an assessment of academic outcomes following the experience
of maltreatment.
Selected subtests of the Woodcock-Johnson III Tests of Achievement (WJ-III;
Woodcock, McGrew, & Mather, 2004) were administered to measure children’s academic
achievement in the areas of reading and mathematics. The subtests administered included letterword identification and applied problems. Letter-Word Identification assesses children’s ability
to identify letters and words, without needing to understand the meaning of any word. Applied
Problems assesses children’s ability to analyze and solve mathematics problems, recognizing
necessary procedures and performing simple calculations. Standard scores were obtained for
each of these subtests. These scores were used to estimate the latent construct of Academic

56

Achievement. The WJ-III Letter-Word Identification and WJ-III Applied Problems subtest
scores were standardized and utilized as observed variables within analyses to provide a valid
and reliable estimate of the latent construct of academic achievement.
Academic Engagement. Academic engagement was assessed through a self-report
questionnaire completed by the child approximately 18 months after the initial wave of data
collection. This measure provided an assessment of academic outcomes following the experience
of maltreatment.
The School Engagement Questionnaire was part of the Drug Free Schools Outcome
Study Questions (DFSCA; U.S. Department of Education). Within the NSCAW study, this
measure had strong reliability (r = .84). Reliability within the current study was r=.74. Children
rated their disposition toward learning and school responding “1=Never,” “2=Sometimes,”
“3=Often,” or “4=Almost Always” to the following 11 items:
•

How often do you enjoy being in school?

•

How often do you hate being in school?

•

How often do you try to do your best work in school?

•

How often do you find the school work too hard to understand?

•

How often do you find your classes interesting?

•

How often do you fail to complete or turn in your assignments?

•

How often do you get sent to the office, or have to stay after school, because you
misbehaved?

•

How often do you get along with your teachers?

•

How often do you listen carefully or pay attention in school?

•

How often do you get your homework done?

57

•

How often do you get along with other students?
Children’s responses to these items were used as an indicator of overall school

engagement. Since individual items were provided within the NSCAW II data set and a total
composite was not created, factor analyses were conducted to combine items to represent the
construct of interest. CFA was conducted with 507 participants that completed the questionnaire
at the second wave of data collection. Negatively phrased items were reverse coded (SE2, SE4,
SE6, and SE7). Strongly correlated items were identified and integrated within the CFA model
(SE2 with SE1; SE9 with SE3). These items assessed similar or opposite aspects of school
engagement (i.e., Do you enjoy being in school? and Do you hate being in school?) and therefore
are expected to be correlated. See Table 8 for final factor loadings and standardized and
unstandardized coefficients. All items loaded significantly on the hypothesized construct of
perceived peer social support. Fit indices included RMSEA = 0.049, with 90% CI = [0.036,
0.063], and CFI/TLI fit indices close to 1, indicating the model fits the data reasonably well. This
measure provided a reliable indicator of children’s perceived school engagement. Higher scores
on this measure correspond to greater school engagement.

58

Table 8
Confirmatory Factor Analysis for One-Factor Model of Academic Engagement
Items
Unstandardized Standard Error (SE) Standardized
SE1: Enjoy being in school
1.000
0.000
0.616
SE2: Hate being in school
0.824
0.122
0.508
SE3: Try to do best work
1.112
0.119
0.685
SE4: Find school work hard
0.356
0.140
0.220
SE5: Find classes interesting
1.048
0.126
0.645
SE6: Fail to complete assignments
0.687
0.101
0.423
SE7: Sent to the office
0.864
0.118
0.532
SE8: Get along with teachers
0.960
0.115
0.591
SE9: Listen carefully
1.027
0.136
0.633
SE10: Get homework done
0.900
0.126
0.555
SE11: Get along with other students
0.688
0.106
0.424
Note: All factor loadings were significant (p<.01)
Internalizing Behaviors. Social-emotional well-being was assessed using the Child
Behavior Checklist parent questionnaire (CBCL; Achenbach, 1991) and Youth Self Report child
questionnaire (YSR; Achenbach, 1991) completed approximately 18 months after the initial
wave of data collection. This data provided an assessment of behavioral outcomes following the
experience of maltreatment.
The CBCL and YSR are commonly used standardized assessments of children’s behavior
within several domains. The internalizing behaviors subscale of the CBCL and YSR includes
items assessing behaviors indicating withdrawal, somatic problems, anxiety/depression, social
problems, and thought problems. It assesses indicators of behavior problems observed in the
previous two months. The CBCL Internalizing subscale had a reliability coefficient of r=.90 and
the YSR Internalizing subscale had a reliability coefficient of r=.90 within the original NSCAW
study. Using both caregiver and child measures will provide a more reliable indicator of socialemotional well-being than including only one of these measures. Children completed the YSR
and the caregiver with whom children were living at the time of data collection completed the

59

CBCL. Children and caregivers rated a variety of aspects of the child’s behavior during the past
two months on 120 items by responding “0=Not True,” “1=Sometimes True,” and “2=Very
True.”
Scores on the CBCL and YSR are normed providing a measure of behavior problems
compared to same-age peers. The CBCL Internalizing subscale score and the YSR Internalizing
subscale score were standardized and utilized as observed variables within analyses to provide a
valid and reliable estimate the latent construct of internalizing behaviors.
Externalizing Behaviors. The CBCL and YSR were also used to assess externalizing
behaviors. The externalizing behaviors subscale of the CBCL and YSR include items assessing
attention problems, delinquent behavior, and aggressive behaviors. It assesses indicators of
behavior problems observed in the previous two months. The CBCL Externalizing subscale had
a reliability coefficient of r=.92 and the YSR Externalizing subscale had a reliability coefficient
of r=.90 within the original NSCAW study. Similar to the Internalizing Behaviors variable, both
caregiver and child measures were used to provide a more reliable indicator of social-emotional
well-being than using only one indicator. Children and caregivers rated a variety of aspects of the
child’s behavior during the past two months on 120 items by responding “0=Not True,”
“1=Sometimes True,” and “2=Very True.”
Scores on the CBCL and YSR are normed providing a measure of behavior problems
compared to same-age peers. The CBCL Externalizing subscale score and the YSR Externalizing
subscale score were standardized and utilized as observed variables within analyses to provide a
valid and reliable estimate the latent construct of externalizing behaviors.

60

CHAPTER 4
RESULTS
Data were prepared and analyzed on a password protected computer in a locked research
office. The study data set was created using Statistical Package for Social Sciences (SPSS)
version 22 (IBM Corp, 2013). Variables of interest were selected from the NSCAW II data set.
This included variables required for weighting, sampling criteria, demographic characteristics,
and all model variables. A sampling variable was created to indicate inclusion in the sample for
the current study. Negatively phrased items were reverse coded. Reduced categories were created
for type of maltreatment (abuse or neglect) and severity of maltreatment (none, mild, moderate to
severe). Continuous variables (achievement, internalizing behaviors, externalizing behaviors)
were standardized. Descriptive analyses were conducted using complex sample weights and
compared to results without sample weights. Final analyses were conducted using sample
weights.
Preliminary Analyses
Preliminary analyses were conducted to ensure that statistical assumptions were met.
Correlation coefficients were calculated to determine whether there was an observed correlation
between independent and dependent variables. Correlation matrices for observed variables and
latent variables are presented in Table 9 and 10, respectively. Statistically significant correlations
were observed between Trauma, Caregiver Support, Peer Support, Engagement, Internalizing
Behaviors, and Externalizing Behaviors. This suggests there is a significant relation between
these variables that may be investigated further. Achievement was not significantly related to
Trauma, Caregiver Support, Peer Support, or the other outcome measures.

61

This study analyzed outcome variables measured at Wave 2 of the NSCAW II study, 18
months after the start of the maltreatment investigation. As these measures were also collected at
Wave 1, consideration was given to utilizing Wave 1 measures to control for functioning at
Wave 2. This would provide an indication of the potential influence of trauma symptoms on
outcome measures beyond adolescents' prior academic and social-emotional functioning. During
preliminary analyses, the regression of trauma on Wave 2 outcome measures, while controlling
for Wave 1 measures, showed that Wave 1 and Wave 2 measures are highly correlated with each
other. This linear dependence caused estimation issues which made it impossible to control for
Wave 1 measures in further analyses.
Table 9
Estimated Correlation Matrix for Observed Variables
Variables
Type
Severity-2 Severity-3 Substantiation
Type
Severity-2
0.039
Severity-3
-0.069
-0.839***
Substantiation 0.04
0.202***
0.682***
Age
0.006
0.045
0.051
0.095*
Gender
-0.076* -0.024
0.058
0.062
* p < 0.05 ** p < 0.01 *** p < 0.000

Age

0.014

Table 10
Estimated Correlation Matrix for Latent Variables
Variables

Trauma

Trauma
Caregiver
-0.385***
Peer
-0.366***
Achievement -0.082
Engagement -0.189*
Internalizing 0.687***
Externalizing 0.364***
* p < 0.05 ** p < 0.01

Achievement

Caregiver

Peer

0.233***
0.077
0.198**
-0.344**
-0.292**
*** p < 0.000

0.107
0.18*
0.107
-0.602*** -0.144
-0.175*
-0.105

62

Engagement

-0.452***
-0.698***

Internalizing

0.617***

Descriptive Statistics
Cross tabulation analyses were conducted on the demographic and maltreatment
characteristic variables to better understand the population sampled. Gender did not appear to be
significantly related to maltreatment characteristics. Type of maltreatment as a dichotomized
variable (i.e., abuse or neglect) did not differ based on gender with approximately half of
investigations assessing each type of maltreatment for both males (abuse: 52%, neglect: 48%)
and females (abuse: 57%, neglect: 43%). However, gender differences may exist at the more
specific level (e.g., sexual abuse, physical abuse, etc.). Additionally, males and females tended to
have similar caseworker ratings of the level of perceived harm to the child. Among males, the
majority of cases were rated as experiencing little to no harm (none: 50%, mild: 32%,
moderate/severe: 18%). Among females, the majority of cases were also rated as experiencing
little to no harm (none: 49%, mild: 30%, moderate/severe 21%). The determination of
substantiation did not appear to be related to gender either. Of cases involving males, an
estimated 76% were unsubstantiated and 24% were substantiated. Similarly, of cases involving
females, an estimated 72% were unsubstantiated and 28% were substantiated.
The distribution of maltreatment characteristics also appeared to be distributed similarly
across ages. Approximately half of the population were the subject of a child welfare
investigation related to abuse and approximately half of the population were the subject of an
investigation related to neglect across most ages, with the exception of children in the age groups
14-, 16-, and 17-years-old. A greater percentage of cases involved suspected abuse compared to
neglect among 14-year-olds (abuse: 74%, neglect: 26%) and 17-year-olds (abuse: 71%, neglect:
29%). Among 16-year-olds, a greater percentage of cases involved suspected neglect compared
to abuse (abuse: 34%, neglect: 66%). Regarding substantiation, the majority of cases across each

63

age group were found unsubstantiated (49%-78%) compared to substantiated (22%-51%). The
distribution of level of severity was also similar across each age group with the majority of cases
rated by the child’s caseworker as none (37%-55%), followed by moderate (26%-38%), and then
moderate to severe (16%-29%). Overall, age and gender do not appear to be related to the type of
maltreatment investigated, determination of substantiation, or level of harm as perceived by the
caseworker.
While demographic variables do not appear to be related to maltreatment characteristics,
different patterns were observed between some maltreatment characteristics variables. Type of
maltreatment did not appear to be related to determination of substantiation. Among cases
primarily investigated due to abuse, an estimated 75% were unsubstantiated and 25% were
substantiated. Among cases primarily investigated due to neglect, an estimated 72% were
unsubstantiated and 28% were substantiated. Similarly, the perceived level of severity of
maltreatment did not appear to be related to whether a case was investigated primarily due to
abuse (none: 49%, mild: 30%, moderate/severe: 21%) or neglect (none: 50%, mild: 32%,
moderate/severe: 18%). Substantiation did appear to predict the level of harm perceived by the
child’s caseworker. The majority of unsubstantiated cases were reported by caseworkers as
resulting in low levels of harm to the adolescent (none: 63%, mild: 27%, moderate/severe: 10%).
Conversely, the majority of substantiated cases were reported to result in high levels of harm to
the adolescent (none: 10%, mild: 39%, moderate/severe: 51%). This suggests that substantiated
cases of maltreatment were more likely to be rated as causing significant harm to the child
compared to unsubstantiated cases.
Descriptive statistics were computed for the outcome measures prior to standardization.
Descriptive statistics for predictor and outcome variables are presented in Table 11 (N=697). On

64

the academic achievement measures, obtained scores were presented as Standard Scores, which
have a mean of 100 and standard deviation of 15 within the normal population. Scores between
85 and 115 are typically considered in the average range. Within the current sample, academic
achievement scores were generally in the average range. The average reading achievement
standard score (Mean = 91.85, SD = 386.3) was slightly greater than the average mathematics
achievement standard score (Mean = 87.78, SD = 231.4). Observed scores ranged from 6 to 137
for reading and 2 to 125 for mathematics.
On the social-emotional well-being measures, obtained scores were presented as T
Scores, which have a mean of 50 and standard deviation of 10. Scores between 63 and 64 are
considered borderline clinical and scores equal to or greater than 65 are considered clinically
significant. Within the current sample, adolescents’ scores on the Youth Self Report Internalizing
Behaviors subscale ranged from 26 to 81 with a mean score of 48.36. Caregivers’ ratings on the
Child Behavior Checklist Internalizing Behaviors subscale ranged from 33 to 90 with a mean
score of 53.16. Adolescents’ scores on the Youth Self Report Externalizing Behaviors subscale
ranged from 25 to 88 with a mean score of 52.31. Caregivers’ ratings on the Child Behavior
Checklist Externalizing Behaviors subscale ranged from 34 to 91 with a mean score of 55.04.
This suggests that the mean YSR and CBCL scores were slightly greater than the average score
within the population (T=50). Many children exhibited fewer behavior problems than typical
children, but many others exhibited clinically significant behavior problems.

65

Table 11
Descriptive Statistics for Sample
Variable
Trauma Symptoms
T1
T2
T3
T4
T5
T6
T7
T8
T9
T10
Caregiver Social Support
C1-Child feels good with
CG
C2-Child feels mad with
CG
C3-Child feels unhappy
with CG
C4-CG enjoys time with
child
C5-CG does a lot to help
child
C6-CG doesn't have time
for child
C7-CG doesn't know how
child feels
C8-CG trust child
C9-CG doesn't let child
make decisions
C10-CG is fair with child
C11-CG thinks child
can't do much
C12-Child doesn't know
what CG wants
Peer Social Support
P1-Easy to make new
friends
P2-Nobody to talk to

Sample
N

Mean

Standard
Deviation
(SD)

Range

Weighted
Mean

Standard
Error
(SE)

696
695
695
686
695
694
695
695
694
695

1.85
1.79
2.20
1.98
2.00
1.52
1.19
1.74
2.00
2.55

0.838
0.901
0.930
0.963
0.926
0.866
0.537
0.942
0.965
1.135

1-4
1-4
1-4
1-4
1-4
1-4
1-4
1-4
1-4
1-4

1.94
1.81
2.18
1.95
1.98
1.54
1.22
1.71
1.96
2.40

0.047
0.059
0.063
0.054
0.059
0.060
0.044
0.057
0.058
0.076

687

3.44

0.872

1-4

3.43

0.050

688

3.28

0.925

1-4

3.43

0.058

689

3.41

0.891

1-4

3.38

0.062

688

3.53

0.823

1-4

3.48

0.042

688

3.60

0.803

1-4

3.56

0.051

688

3.24

0.967

1-4

3.18

0.069

680
686

2.71
3.28

1.153
0.930

1-4
1-4

2.68
3.27

0.076
0.049

683
684

3.02
3.39

1.045
0.901

1-4
1-4

2.96
3.34

0.075
0.058

685

3.32

1.006

1-4

3.27

0.059

673

3.14

1.089

1-4

3.16

0.060

696
696

3.86
4.34

1.111
1.080

1-5
1-5

3.81
4.35

0.084
0.068

66

Table 11 (cont’d)
P3-Good working with
other kids
P4-Hard to make friends
P5-Have lots of friends
P6-Feel alone
P7-Can find a friend
when needed
P8-Hard to get kids to
like me
P9-Don't have anyone to
play with
P10-Get along with other
kids
P11-Feel left out
P12-No kids to go to
when need
P13-Don't get along with
other kids
P14-Feel lonely at school
P15-Well liked by kids
P16-Don't have any
friends
Academic Achievement
Letter-Word
Identification SS
Applied Problems SS
Academic Engagement
1-Enjoy being in school
2-Hate being in school
3-Try to do best work
4-Find work too hard
5-Find class interesting
6-Fail to complete
assignments
7-Sent to office
8-Get along with teachers
9-Listen carefully
10-Get homework done
11-Get along with other
students

696
696
694
696

3.82
4.08
4.19
4.37

1.062
1.185
1.126
1.043

1-5
1-5
1-5
1-5

3.79
4.02
4.22
4.35

0.070
0.082
0.073
0.069

695

4.12

1.163

1-5

4.04

0.074

695

4.07

1.190

1-5

3.96

0.076

695

4.37

1.120

1-5

4.33

0.062

695
696

4.08
4.18

1.007
1.085

1-5
1-5

4.09
4.13

0.065
0.083

695

4.13

1.258

1-5

4.07

0.074

694
694
694

3.91
4.44
3.94

1.178
0.978
1.075

1-5
1-5
1-5

3.80
4.37
3.85

0.080
0.081
0.075

692

4.47

1.080

1-5

4.40

0.063

556
564

92.79
87.52

17.881
13.889

6-137
2-125

91.85
87.78

1.223
1.026

507
507
507
507
506

2.75
2.90
3.41
2.82
2.68

.926
.832
.748
.684
.877

1-4
1-4
1-4
1-4
1-4

2.70
2.85
3.42
2.90
2.67

0.065
0.058
0.054
0.049
0.069

506
507
506
506
501

2.96
3.56
3.23
3.14
3.09

.803
.707
.868
.762
.924

1-4
1-4
1-4
1-4
1-4

2.95
3.62
3.21
3.21
3.07

0.056
0.044
0.084
0.062
0.081

506

3.21

.839

1-4

3.24

0.058

67

Table 11 (cont’d)
Internalizing Behaviors
Youth Self Report (Int)
Child Behavior Checklist
(Int)
Externalizing Behaviors
Youth Self Report (Ext)
Child Behavior Checklist
(Ext)

530

48.36

11.458

26-81

48.37

0.789

551

54.20

10.991

33-90

53.22

0.727

530

53.15

11.62

25-88

52.35

1.042

551

56.44

11.46

34-91

55.14

0.801

Structural Model
Research Question 1. Does social support serve as a protective factor, moderating the
relation between trauma and academic and social-emotional well-being?
The first goal of this study was to determine whether social support moderated the
relation between trauma and academic and social-emotional well-being. Structural equation
models were assessed for caregiver support and peer support separately (Appendix A). Each
proposed model hypothesized that the latent variables of Trauma, social support, and the
moderator variable (trauma*social support) would significantly predict the outcome variables of
Achievement, Engagement, Internalizing Behaviors, and Externalizing Behaviors. Moderator
variables were defined as an interaction term of the latent variable for Trauma and the latent
variable for each type of support. The proposed models also hypothesized that the observed
variables of Type of Maltreatment, Severity of Maltreatment, Substantiation, Age, and Gender
would significantly predict Trauma and Caregiver Support. Regression coefficients and model fit
indices of the SEM were used to evaluate hypotheses.
Caregiver Support. Contrary to the proposed hypothesis, SEM analyses did not indicate
that caregiver social support moderates the relation between trauma and academic and socialemotional well-being in the current sample. Factor loadings and significance levels for the model
assessing caregiver social support as a moderator are presented in Table 12. The interaction term

68

Trauma*Care was not significantly related to achievement (b=-0.09, p = 0.14), engagement
(b=0.01, p = 0.79), internalizing behaviors (b=0.04, p = 0.11), or externalizing behaviors
(b=0.04, p = 0.051). The estimate for the interaction term and externalizing behaviors, however,
was at the borderline of being statistically significant.
Table 12
Unstandardized and Standardized Factor Loadings and Significance Levels for Model Assessing
Caregiver Social Support
Parameter Estimate
Unstandardized (SE)
Standardized
p
Trauma → Achievement
-0.22 (0.17)
-0.38
0.14
Trauma → Engagement
-0.06 (0.14)
-0.07
0.65
Trauma → Internalizing
0.45 (0.10)
0.83
<0.000
Trauma → Externalizing
0.27 (0.10)
0.44
<0.01
Caregiver → Achievement
0.05 (0.07)
0.14
0.43
Caregiver → Engagement
0.13 (0.08)
0.20
<0.05
Caregiver → Internalizing
-0.06 (0.07)
-0.16
0.34
Caregiver → Externalizing
-0.09 (0.07)
-0.20
0.20
Trauma*Caregiver → Achievement
-0.09 (0.06)
-0.31
0.06
Trauma*Caregiver → Engagement
0.01 (0.04)
0.02
0.79
Trauma*Caregiver → Internalizing
0.04 (0.02)
0.16
0.10
Trauma*Caregiver → Externalizing
0.04 (0.02)
0.13
0.052
Peer Support. Contrary to the proposed hypothesis, SEM analyses did not indicate that
peer social support moderated the relation between trauma and academic and social-emotional
well-being in the current sample. Factor loadings and significance levels for the model assessing
peer social support as a moderator are presented in Table 13. The interaction term Trauma*Peer
was not significantly related to achievement (b=-0.01, p = 0.50), engagement (b=0.01, p = 0.81),
internalizing behaviors (b=0.03, p = 0.17), or externalizing behaviors (b=0.006, p = 0.83). As the
interaction was not significant, peer social support does not appear to moderate the relation
between trauma and outcomes.

69

Table 13
Unstandardized and Standardized Factor Loadings and Significance Levels for Model Assessing
Peer Social Support
Parameter Estimate
Unstandardized (SE)
Standardized
p
Trauma → Achievement
-0.01 (0.06)
-0.02
0.88
Trauma → Engagement
-0.10 (0.01)
-0.11
0.31
Trauma → Internalizing
0.27 (0.06)
0.54
<0.000
Trauma → Externalizing
0.19 (0.08)
0.32
<0.01
Peer → Achievement
0.05 (0.05)
0.12
0.19
Peer → Engagement
0.10 (0.09)
0.12
0.24
Peer → Internalizing
-0.17 (0.04)
-0.40
<0.000
Peer → Externalizing
-0.03 (0.05)
-0.06
0.59
Trauma*Peer → Achievement
-0.02 (0.03)
-0.06
0.46
Trauma*Peer → Engagement
0.01 (0.04)
0.02
0.82
Trauma*Peer → Internalizing
0.03 (0.02)
0.10
0.14
Trauma*Peer → Externalizing
0.01 (0.03)
0.02
0.83
Final Model. Although moderation relations were not observed for either caregiver
support or peer support, direct relations were identified between predictors and outcomes. A final
model including trauma and both caregiver and peer social support was assessed. Factor loadings
and significance levels for this final model including predictors of trauma, caregiver support, and
peer support are presented in Table 14. Trauma was significantly related to Internalizing
Behaviors (b=0.59, p < 0.000) and Externalizing Behaviors (b=0.37, p<0.000). More significant
trauma symptoms were related to poorer social-emotional well-being. A direct relation was not
observed between Trauma and academic outcome variables. Caregiver support was significantly
related to Engagement in academic activities (b=0.11, p<0.05). Greater perceived caregiver
social support was related to greater engagement in academic activities. Caregiver support was
also significantly related to Externalizing Behaviors (b=-0.19, p<0.05). Greater perceived
caregiver social support was related to fewer behavior problems. The relation between caregiver
support and academic achievement (b=0.04, p=0.54) or internalizing behaviors (b=-0.05, p =
0.53) were not statistically significant. Caregiver social support predicted academic engagement

70

and externalizing behaviors, but does not appear to be related to other outcome variables
assessed.
Peer support was significantly related to Internalizing Behaviors (b=-0.41, p<0.000).
Greater perceived social support from peers was related to fewer internalizing behavior
problems. The relation between peer support and academic achievement (b=0.09, p=0.21),
academic engagement (b=0.10, p=0.13), and externalizing behaviors (b=-0.04, p=0.73) were not
statistically significant. Peer social support predicts internalizing behaviors, but does not appear
to be related to other outcome variables assessed.
Table 14
Unstandardized and Standardized Factor Loadings and Significance Levels for Final Model
Parameter Estimate
Unstandardized (SE)
Standardized
p
Trauma → Achievement
-0.04 (0.10)
-0.03
0.70
Trauma → Engagement
-0.09 (0.80)
-0.10
0.26
Trauma → Internalizing
0.59 (0.11)
0.52
<0.000
Trauma → Externalizing
0.37 (0.13)
0.29
<0.001
Caregiver → Achievement
0.04 (0.07)
0.05
0.54
Caregiver → Engagement
0.11 (0.05)
0.14
<0.05
Caregiver → Internalizing
-0.05 (0.08)
-0.05
0.53
Caregiver → Externalizing
-0.19 (0.09)
-0.18
<0.05
Peer → Achievement
0.09 (0.07)
0.08
0.21
Peer → Engagement
0.10 (0.07)
0.11
0.13
Peer → Internalizing
-0.41 (0.08)
-0.40
<0.000
Peer → Externalizing
-0.04 (0.10)
-0.03
0.73
Note: RMSEA = 0.016, with lower bound of the 90% CI = 0.013; CFI = 0.930; TLI = 0.927
Research Question 2. Is caregiver social support or peer social support a more
important moderator between trauma and outcomes within the overall conceptual model?
The second goal of this study was to determine which source of social support (i.e.,
caregivers or peers) had a greater relative role on the relation between trauma and academic and
social-emotional well-being. As previously discussed, neither caregiver social support nor peer
social support was a significant moderator within the conceptual model. Rather, caregiver social
support and peer social support were statistically significant predictors of outcomes through a

71

direct relation. Several indicators were used to assess the relative importance of each of these
predictors using three conceptual models: (1) caregiver social support and peer social support
predicting outcomes, (2) caregiver social support predicting outcomes, and (3) peer social
support predicting outcomes.
First, model fit indices were compared between the three conceptual models (see Table
15). The RMSEA index was less than 0.05 for each model, indicating the models fit the data
well. The CFI and TLI indicators were greater than 0.90 for each model, indicating the models fit
the data well. Differences in model fit indices were observed between the three models, but these
differences were marginal. On the RMSEA indicator, the model including both caregiver social
support and peer social support fit the data most closely (RMSEA=0.016), followed by the
model with only caregiver social support (RMSEA=0.019) and the model with only peer social
support (RMSEA=0.021). On the CFI and TLI indicators, the model including only caregiver
social support fit the data most closely (CFI=0.933, TLI=0.928), followed by the model with
both caregiver social support and peer social support (CFI=0.930, TLI=0.927) and the model
with only peer social support (CFI=0.926, TLI=0.921). Comparing model fit indices suggests
that a model including both caregiver social support and peer social support may provide the
most accurate representation of the data.
Table 15
Model Fit Indices Comparing the Relative Impact of Caregiver Support and Peer Support
Model
RMSEA (CI)
CFI
TLI
Caregiver Social Support
0.019 (0.014-0.023)
0.933
0.928
Peer Social Support
0.021 (0.017-0.024)
0.926
0.921
Caregiver and Peer Social Support
0.016 (0.013-0.019)
0.930
0.927
Second, chi-square difference testing was conducted to compare the model including both
caregiver social support and peer social support with the model including only caregiver social
support (see Table 16). This analysis allows for testing of the statistical significance of

72

differences between model fit indices. The full model hypothesized that trauma, caregiver social
support, and peer social support predicted outcome variables. This full model was compared to a
nested model in which the path between peer social support and outcome variables was
constrained to zero. The difference test compared the model fit of the nested model to the full
model. This test resulted in a Chi-Square value of 14.48 (p < 0.01), indicating that the model fit
worsened when peer social support was removed from the model. A model including both
caregiver social support and peer social support predicting outcome measures better fit the data,
indicating that peer social support is an important factor within the hypothesized model.
Although caregiver social support and peer social support were not significant
moderators within the conceptual model, they did significantly predict some outcome measures.
Caregiver social support significantly predicted greater engagement and fewer externalizing
behaviors, while peer social support significantly predicted fewer internalizing behaviors.
Results indicated a differential relation between the type of social support and the outcome
variables assessed. Therefore, neither form of support can be considered more important than the
other. Difference testing, however, indicated that a full model containing both caregiver social
support and peer social support better fit the data compared to a nested model containing only
caregiver social support. Therefore, both forms of support are necessary to best understand the
complex relation between trauma, social support, and academic and social-emotional well-being.
Table 16
Results of Chi-Square Test for Difference Testing
Model
Chi-Square
df
p
Full Model: Trauma Caregiver Peer
1663.18
1408
<0.000
Nested Model: Trauma Caregiver
1677.77
1412
<0.000
Difference Testing
14.48
4
<0.01
Note: Significant p value of Difference Testing indicates that the fit of the Nested Model is
significantly worse than Full Model; therefore, the Full Model is retained.

73

Research Question 3. Do children’s perceived trauma symptoms vary based on type of
maltreatment, severity of maltreatment, substantiation, age, or gender?
The third goal of this study was to identify predictors of trauma symptoms. Based on
previous research, it was hypothesized that type of maltreatment, severity of maltreatment,
substantiation, age, and gender may predict the level of perceived trauma symptoms. To evaluate
this hypothesis, these demographic and maltreatment history variables were included in the
conceptual model as manifest variables predicting trauma symptoms. Type of maltreatment,
substantiation, and gender were each dichotomous variables. Severity of maltreatment was
assessed as two dummy coded variables assessing “mild” and “moderate to severe” levels of
harm. Age was assessed as a continuous variable. Factor loadings and significance levels for
hypothesized predictors of trauma symptoms are presented in Table 17.
Type. Trauma symptoms did not vary based on type of maltreatment. The manifest
variable for type of maltreatment did not significantly predict trauma symptoms (p = 0.09).
When assessing type of maltreatment as a dichotomous variable with the categories of “abuse”
and “neglect,” no significant differences were observed on the level of trauma symptoms
reported.
Severity. Trauma symptoms did not vary based on the level of severity. The dummy
variables representing “mild” level of harm (p = 0.25) and “moderate to severe” level of harm (p
= 0.82) did not significantly predict trauma symptoms. When assessing severity as three
categories of “none,” “mild,” and “moderate to severe,” no significant difference was observed
between the reported level of harm and the level of trauma symptoms reported.
Substantiation. Trauma symptoms did not vary based on determination of substantiation.
The manifest variable for substantiation did not significantly predict trauma symptoms (p =

74

0.95). The determination of whether there was substantial evidence to conclude that the
maltreatment likely occurred was not related to children’s’ reported trauma symptoms.
Age. Trauma symptoms did not vary based on participants’ age. The continuous variable
for age did not significantly predict trauma symptoms (p = 0.88). Contrary to the proposed
hypothesis, children’s age was not related to the level of trauma symptoms exhibited.
Gender. Trauma symptoms did not vary based on gender. The dichotomous variable for
gender did not significantly predict trauma symptoms (p = 0.24). While it was hypothesized that
females would exhibit greater trauma symptoms compared to males, gender differences were not
observed.
Table 17
Predictors of Trauma Symptoms
Parameter Estimate
Type of Maltreatment
Severity (Mild)
Severity (Moderate to Severe)
Substantiation
Age
Gender

Unstandardized (SE)
-0.16 (0.09)
-0.11 (0.09)
-0.03 (0.12)
0.01 (0.10)
0.00 (0.03)
0.14 (0.12)

Standardized
-0.12
-0.07
-0.02
0.00
0.01
0.10

p
0.09
0.25
0.82
0.95
0.88
0.24

Research Question 4. Do children’s perceived social support vary based on type of
maltreatment, severity of maltreatment, substantiation, age, or gender?
The fourth goal of this study was to identify differences in perceived caregiver and peer
social support based on demographic and maltreatment history variables. Previous research
among normative populations suggests that age and gender are related to perceived social
support. Existing literature has not examined potential differences in perceived social support
based on maltreatment history. However, maltreatment history including the type of
maltreatment experienced, the severity of that maltreatment, and the determination of
substantiation may result in significantly different experiences with caregivers and peers. To

75

evaluate this hypothesis, these demographic and maltreatment history variables were included in
the conceptual model as manifest variables predicting perceived caregiver social support and
predicting perceived peer social support. Factor loadings and significance levels for hypothesized
predictors of social support are presented in Table 18.
Type. It was hypothesized that children would report greater caregiver support if they
were the subject of an investigation due to neglect compared to an investigation due to abuse.
This hypothesis was supported by the analyses. Type of maltreatment significantly predicted
perceived caregiver social support (b = 0.22, p < 0.01). Children with investigations due to
neglect were more likely to report greater perceived caregiver support compared to children with
investigations due to abuse. Conversely, it was hypothesized that children would report greater
peer support if they were the subject of an investigation due to abuse compared to an
investigation due to neglect. This hypothesis was not supported by the analyses. Type of
maltreatment did not significantly predict peer social support (p = 0.18). There were no
differences on the level of perceived peer social support between children that primarily
experienced abuse and those that primarily experienced neglect.
Severity. For both caregiver social support and peer social support, it was predicted that
children with moderate to severe harm would perceive less social support compared to children
with mild or no harm as reported by caseworkers. This hypothesis was supported with caregiver
social support, but not peer social support. Children with reports of mild harm perceived
significantly less support from caregivers compared to those with reports of no harm (b = -0.28,
p <0.05). Similarly, children with reports of moderate to severe harm perceived significantly less
support from caregivers compared to those with reports of no harm (b = -0.25, p <0.05). Severity
of maltreatment significantly predicted the level of perceived social support from caregivers.

76

This pattern was not observed when examining perceived peer social support. Severity of
maltreatment did not significantly predict the level of social support perceived from peer
relationships (Mild: p = 0.14, Moderate to Severe: p = 0.95).
Substantiation. It was hypothesized that children with substantiated investigations of
maltreatment would perceive less caregiver and peer social support compared to children with
unsubstantiated investigations of maltreatment. Contrary to the proposed hypothesis,
substantiation did not predict caregiver social support (p=0.42) or peer social support (p=0.59).
This result suggests that, on average, there were no differences in perceived level of social
support between children whose cases of alleged maltreatment investigations were substantiated
and those whose cases were not.
Age. It was hypothesized that younger children would report greater social support from
caregivers while older children would report greater social support from peers. This hypothesis
was supported by the data; age significantly predicted both caregiver social support and peer
social support in the expected directions. Older children reported less social support from
caregivers compared to younger children (b=-0.07, p < 0.01). Older children reported greater
social support from peers compared to younger children (b=0.05, p < 0.05). This data suggests
that children rely on different sources of social support at different developmental stages.
Gender. It was hypothesized that males would report greater social support from
caregivers while females would report greater social support from peers. Based on previous
research, females tend to have more social relationships with peers which may lead to greater
perceived emotional support. Contrary to the proposed hypothesis, gender was not significantly
related to caregiver social support (p=0.09) or peer social support (p=0.36). The level of social
support did not vary based on this characteristic.

77

Table 18
Predictors of Social Support
Parameter Estimate
Caregiver Social Support
Type of Maltreatment
Severity (Mild)
Severity (Moderate to Severe)
Substantiation
Age
Gender
Peer Social Support
Type of Maltreatment
Severity (Mild)
Severity (Moderate to Severe)
Substantiation
Age
Gender

Unstandardized (SE)

Standardized

p

0.29 (0.11)
-0.28 (0.11)
-0.25 (0.14)
0.09 (0.11)
-0.07 (0.03)
-0.14 (0.09)

0.19
-0.16
-0.13
0.05
-0.16
-0.09

<0.01
<0.05
0.06
0.42
<0.01
0.09

0.10 (0.07)
-0.14 (0.10)
-0.01 (0.12)
-0.06 (0.11)
0.05 (0.03)
0.09 (0.09)

0.07
-0.09
-0.01
-0.04
0.13
0.06

0.18
0.14
0.95
0.59
<0.05
0.36

78

CHAPTER 5
DISCUSSION
The goal of this study was to examine the role of social support among maltreated youth
within the context of resilience. The study aimed to determine whether social support moderated
the relation between trauma and outcomes, identify differences between the role of social support
from caregivers and the role of social support from peers, and understand how demographic
variables and characteristics of maltreatment may predict perceived trauma symptoms and social
support. These findings were interpreted within the framework of resilience to better understand
how children develop positive outcomes in the presence of adversity.
Resilience
To understand the role of social support among maltreated youth, structural equation
modeling was used to assess whether social support served as a moderator within the conceptual
model. Resilience is a process in which individuals exposed to significant adversity are able to
maintain positive adjustment or outcomes (Masten, 2001; 2011). Among populations of
maltreated youth, this has been conceptualized as achieving a normal range of competence across
multiple domains despite exposure to abuse or neglect (Walsh et al., 2010). Resilience has been
primarily conceptualized as the interaction between competence and stress (Garmezy, Masten, &
Tellegen, 1984). The impact of adversity changes as a result of a protective attribute. These
positive attributes are generally referred to as protective factors (Masten, 2001; 2011). While the
process of resilience is conceptualized as an interaction, the majority of studies assessing
resilience have measured main effects of protective factors; few studies have utilized moderation
analyses (Wilson & Scarpa, 2014). A main effect indicates that an independent variable is related
to a dependent variable. In other words, it can show that a protective attribute is related to

79

positive outcomes. However, it cannot demonstrate that the impact of adversity is reduced as a
result of the protective attribute. To assess whether the negative impact of adversity is reduced as
a result of protective factors, the interaction of these variables must be assessed.
Therefore, moderation analyses were used to determine whether the impact of trauma on
outcomes decreases as a function of social support among maltreated youth. The statistical
significance of the interaction terms Trauma*Caregiver and Trauma*Peer were evaluated in
addition to the main effects of Caregiver and Peer social support. These analyses provided an
indication of whether the relation between trauma and academic and social-emotional well-being
depends on children's perceived emotional social support.
Social support was not found to moderate the relation between trauma and academic and
social-emotional well-being. Neither caregiver social support nor peer social support were
statistically significant moderators within the conceptual model (see Figures 2 and 3). These
variables were, however, direct predictors of positive outcomes. While this finding does not
support an interaction model of resilience, it does support a promotive model of competence in
which an attribute has an additive effect on outcomes in the presence of adversity (Garmezy et
al., 1984). Children with the attribute of social support were more likely to maintain positive
adjustment (academic and social-emotional well-being) following adversity (trauma resulting
from maltreatment).

80

Figure 2
Standardized Factor Loadings of Model Assessing Caregiver Social Support

Figure 3
Standardized Factor Loadings of Model Assessing Peer Social Support

81

Social support has been described as a protective factor within many studies assessing
trauma and maltreatment. This literature conceptualizes and measures resilience as either an
outcome or as a process. Studies that conceptualize resilience as an outcome generally assess
protective factors through main effects. Social support from family and friends has been shown
to directly predict positive outcomes among individuals with a history of maltreatment (e.g.,
Afifi & MacMillan, 2011; McLewen & Muller, 2006). Results of the current study are consistent
with these findings. Other studies conceptualize resilience as a process and assess protective
factors through mediation or moderation analyses. These studies have found that social support
changes the relation between maltreatment and outcomes (e.g., Pepin & Banyard, 2005;
Vranceanu et al., 2007). The current study conceptualized resilience as a process, describing how
contextual factors interact with adverse experiences to promote positive adaptation. This
conceptualization, however, was not supported by the study findings. Social support may
promote positive outcomes, but it appears do so as an additive support rather than interacting
with risk factors.
Alternatively, interaction effects may not have been observed due to the time of
measurement of trauma symptoms, social support, and outcome variables. As previously
discussed, most studies of social support and resilience utilize retrospective self-report methods
to assess childhood trauma history and adult outcomes. These studies are conducted years after
the occurrence of maltreatment. Resilience develops over time and may be observed at different
points following adversity (Rutter, 2013). Prior research utilizing retrospective measures have
demonstrated that social support serves as a protective factor that directly (e.g., McLewen &
Muller, 2006) and indirectly (e.g., Muller et al., 2008) promotes positive outcomes. The current
study found that social support directly predicted outcomes 18 months after a maltreatment

82

investigation, but that social support did not indirectly relate to outcomes as an interaction with
trauma symptoms during this time frame. A greater amount of time (i.e., years to decades) may
be required for social support to facilitate resilience and moderate the impact of maltreatment.
This could explain why retrospective studies have identified moderation effects and the current
study did not. Conversely, studies that assess social support through retrospective studies may
lack validity and project resilience through participants' perceptions of past trauma and current
functioning. A longitudinal study that assesses these variables from the time of maltreatment
through adulthood may be required to fully understand the role of social support within the
framework of resilience.
Although social support did not moderate the relation between trauma and outcomes, the
observed direct relation is an important finding. Caregiver support was observed to predict
positive academic engagement and fewer externalizing behavior problems. Peer social support
was observed to predict fewer internalizing behaviors. These direct relations are presented in the
final conceptual model (Figure 4). While causal effects cannot be determined through this study,
this finding suggests that social support promotes positive academic and social-emotional
outcomes following maltreatment.
Social support provides many benefits which may promote these positive outcomes.
Perceived support creates a sense of stability and facilitates positive experiences with others
(Cohen & Willis, 1985). Cohen and Wills (1985) proposed a model of the relation between
social support and well-being. They hypothesized that social support affects the relation between
stressful events and well-being during two stages: the appraisal of the event and the response to
the event. Perceived social support may prevent a stress response following the event by creating
a sense of security. An individual exposed to a potentially stressful event may perceive that their

83

supportive network will be able to provide them with the necessary resources and supports to
cope with the experience (Cohen & Wills, 1985). Therefore, instead of appraising the experience
as traumatic, the individual may perceive less stress following the event. Children may believe
that they have adequate support from caregivers and peers and therefore perceive the experience
as less severe than they would if they had little support. Additionally, social relationships may
provide necessary support following a potentially traumatic event which help the individual cope
with the situation. Supportive relationships may aid in providing a solution to the problem,
reducing the importance of the problem, and by facilitating healthy behaviors (Cohen & Wills,
1985). After experiencing maltreatment, children may receive emotional support from adults and
friends which helps them cope with the experience. The model proposed by Cohen and Willis
(1985) suggests that social support may benefit individuals exposed to adversity in multiple ways
at different times following the adverse experience.
Overall, social support appears to promote positive outcomes among maltreated youth,
regardless of the degree of trauma. The presence of perceived emotional social support is
associated with more positive academic and social-emotional outcomes following maltreatment.
Children and adolescents with supportive relationships are more likely to experience positive
outcomes following abuse or neglect.
Differential Effects
Caregiver and peer social support were differentially related to outcomes. Social support
from caregiver relationships was associated with fewer externalizing behaviors. Social support
from peer relationships was associated with fewer internalizing behaviors. To understand this
finding, it is necessary to identify the relation between these sources of support and their unique
roles in adolescents' lives.

84

Figure 4
Standardized Factor Loadings of Final Model

Within this study, caregiver social support and peer social support were moderately
correlated, indicating that they are related, but separate, constructs. Similarly, an analysis of
adolescents in the general population assessed the relation between caregiver social support and
peer social support (van Beest & Baerveldt, 1999). The goals of the study were to evaluate
whether adolescents compensate for a lack of parental support through peer support (negative
correlation), whether parental support cannot be compensated by peer support (no correlation), or
whether adolescents with a lack of parental support would also have low peer support (positive
correlation). Within this population of adolescents with generally intact parental support, a
perceived lack of parental support was not compensated for by support from peers, or was not
strongly correlated (van Beest & Baerveldt, 1999). This suggests that social support from
caregivers and social support from peers are independent of each other and therefore may relate
to outcomes differently. This supports an additive model of social support proposed by Brittain

85

(1968). This model theorizes that both caregivers and peers influence adolescents, but influence
different situations. This model therefore suggests that caregiver and peer social support would
each contribute to adolescents' well-being separately.
Relationships with caregivers and relationships with peers are inherently different and
serve different roles in the lives of adolescents. Peer relationships are generally voluntary and
built on mutual interests, cooperation, and trust (Gifford-Smith & Brownell, 2003). Caregiver
relationships are generally developed through early social experiences such as acting as a secure
base and as reassuring play partners (Al-Jagon, 2016). Relationships with caregivers among
maltreated youth, however, may be significantly different. Following maltreatment, an
adolescent's primary caregiver may continue to be the perpetrator of abuse or neglect. Or their
caregiver may be an extended family member or non-familial caregiver that they may or may not
have known for an extended period of time. These varied experiences may create significant
differences across adolescents' relationships with their current caregiver. Additionally, the role of
caregiver relationships and peer relationships tend to change during adolescence. During
childhood, caregivers function as the center of children's social networks (Helsen, et al., 2000).
During adolescence, the significance of caregiver relationships decreases or remains constant
(Helsen, et al., 2000). Adolescents report valuing and seeking emotional support from
relationships that are trusting, familiar, mature, approving, and with whom they have a good
relationship (Camara, Bacigalupe, & Padilla, 2014).
The current study demonstrated that perceived social support from caregivers predicted
fewer externalizing problems within a sample of adolescents that were the subject of child
welfare investigations. Similar patterns have been observed between caregiver relationships and
mental health outcomes among normative populations. In a correlational analysis examining

86

externalizing behaviors within a diverse sample of adolescents, perceived parental social support
predicted fewer externalizing behavior problems (White & Renk, 2012). Similarly, a longitudinal
study assessing changes in adolescent behavior across one year and their relation to parenting
behaviors found that parenting behaviors predict fewer externalizing behavior problems (Reitz,
Dekovic, & Meijer, 2006). Several aspects of parenting behaviors were assessed including:
involvement (perceived parental emotional responsiveness and affection), emotional autonomy
(perceived parental encouragement of emotional autonomy), decisional autonomy (adolescent
decision making on a variety of choices), and strictness (perceived level of strictness of parents).
Higher parental involvement and lower decisional autonomy were related to fewer externalizing
behavior problems (Reitz et al., 2006). As parental involvement was measured similarly to
perceived caregiver social support, this relation is consistent with the current finding that greater
perceived emotional social support is associated with fewer externalizing behaviors.
This finding can be explained by the interaction between parenting behaviors and
adolescent behaviors. Caregiver behaviors may change based on perceived adolescent behaviors
and vice versa. This relationship is dynamic, particularly during adolescence. Reitz et al., (2006)
identified significant interactions between parental involvement and adolescents’ externalizing
behaviors. Among adolescents with high levels of externalizing behaviors, perceptions of
parental involvement tended to change after a one-year period to the opposite direction.
Adolescents with high levels of externalizing behaviors who initially perceived their parents as
highly involved later perceived them as less involved. Adolescents with high levels of
externalizing behaviors who initially perceived their parents as less involved later perceived
them as more involved. Overall, parenting behaviors seem to adjust based on the level of

87

adolescents’ behavioral needs and this adjustment is associated with fewer externalizing
behavior problems among typical adolescents.
Perceived caregiver support did not predict internalizing behaviors among youth that
were the subject of child welfare investigations or non-referred youth. As in the current study,
Reitz et al. (2006) did not find a significant direct relation between parenting behaviors and
adolescents’ internalizing behaviors. A significant interaction effect was identified, however,
between parent involvement and internalizing behaviors. Among a sample of non-referred
adolescents with higher levels of internalizing problems, higher levels of parent involvement
were associated with even greater internalizing problems one year later (Reitz et al., 2006). This
suggests that for adolescents who express high levels of emotional difficulties, perceived
parental involvement may increase internalizing behavior problems. This could be due to a
perception by adolescents that caregivers are over-involved or do not understand the adolescents’
difficulties accurately. Highly involved parents may also convey concerns and fears that amplify
the adolescents’ anxieties and insecurities resulting in higher levels of internalizing problems.
The current study examined the interaction between perceived trauma symptoms and mental
health outcomes, but did not assess the potential interaction between perceived caregiver social
support and adolescents’ behaviors including internalizing problems. Caregiver social support
may not directly relate to adolescents’ internalizing behavior problems due to adolescents’
perception that caregivers are not able to provide the appropriate level and type of support
needed. When parental involvement is overly high or overly low, adolescents may experience
increased internalizing behavior problems.
The current study identified a significant direct relation between social support from
peers and internalizing behaviors, but not externalizing behaviors. Previous research assessing

88

peer relationships among normative populations, however, have suggested different patterns.
Among non-referred populations, adolescents tend to prefer friends with similar externalizing
problems, but not internalizing problems as measured by peer nomination (Fortuin, van Geel, &
Vedder, 2015). In addition to choosing a peer group that is similar to themselves in terms of
externalizing behaviors, adolescents also become more similar to their peer group over time
through socialization (Fortuin et al., 2015). This suggests that peer relationships may be
associated with greater externalizing behaviors. Mixed findings have been identified within the
relation between peer relationships and internalizing behaviors. Some studies have shown that
positive peer relationships promote emotional well-being (Fortuin et al., 2015) while others have
shown that for females, peer relationships can contribute to increased internalizing behaviors
(Rose, Carlson, & Waller, 2007). Both females and males often engage in co-rumination, or
excessively discussing problems, with friends. Among females, co-rumination predicts increased
internalizing behavior problems including symptoms of depression and anxiety (Rose et al.,
2007). While co-rumination was not assessed in the current study, greater perceived social
support from peers was related to fewer internalizing behavior problems. This suggests that for
youth that were the subject of child welfare investigations, support from peer relationships may
provide effective emotional support contributing to positive functioning in relation to
internalizing problems such as anxiety and depression.
Overall, the findings from the current study assessing the role of caregiver support on
externalizing and internalizing behaviors were consistent with research within normative
populations. Caregiver support can provide effective physical security and structure, which may
reduce behavior problems (Al-Jagon, 2016). In contrast, findings from the current study related
to peer support were not consistent with research assessing normative populations. While some

89

studies have suggested that peer relationships have a negative effect on internalizing behaviors
through co-rumination (Rose et al., 2007), the current study suggested that social support from
peers is associated with fewer internalizing behavior problems among maltreated youth.
Adolescents seek emotional support from peer relationships (Gifford-Smith & Brownell, 2003).
This may provide feelings of trust and connectedness which reduces emotional distress,
especially among maltreated youth. Maltreated youth may have their emotional needs and
behavioral needs met through these different types of relationships. Relationships with caregivers
may support behavioral needs while relationships with peers may support emotional needs. As
both needs are important for healthy functioning, both relationships are important in promoting
positive outcomes.
Another explanation for this finding could be that the relation between social support and
outcomes differs based on demographic or maltreatment characteristics. In particular, the role of
caregiver and peer relationships may differ between males and females. Assessing group
differences related to gender, maltreatment type, substantiation, etc. could aid in further
understanding the relations observed within this study. Previous research (e.g., Tolin & Foa,
2006) suggests that the observed differential relation between caregiver social support and peer
social support may be partially explained by gender differences, but due to multicollinearity
between peer social support and internalizing behaviors, this could not be examined. Cross
tabulation analyses suggest that differences may exist in the relation between social support and
outcomes on the basis of maltreatment characteristics. Future research that includes the
assessment of group differences is indicated to better understand the role of social support within
this heterogeneous population.

90

Significance of Peer Social Support
Another important finding of this study was the significance of peer social support within
the conceptual model. The role of peer social support has been examined in some prior research,
but the majority of studies examining normative populations or maltreated populations have
assessed social support through caregiver relationships (e.g., Dingfelder et al., 2010; Muller et
al., 2000). There has been a lack of research examining the role of peer support. The current
study demonstrated that social support from peer relationships is important for several reasons.
First, peer relationships appear to have a unique role in promoting emotional well-being.
The differential relation previously discussed demonstrates that relationships from peers
contributes to well-being separately from caregiver social support. Specifically, greater social
support from peers was associated with fewer internalizing behavior problems, while social
support from caregivers was not related to this outcome variable.
Second, consideration of peer social support provides a more complete model of trauma
and resilience among maltreated youth. The conceptual model including both caregiver social
support and peer social support as predictors better explained the sample data compared to a
model including only caregiver social support, as measured by goodness of fit testing. Consistent
with an additive model of social support, caregiver support and peer support independently
contribute to outcomes (Brittain, 1968). Previous studies that have examined only caregiver
social support may have missed important aspects of the relation between risk and resilient
outcomes. Overall, these findings suggest that consideration of multiple sources of social support
is needed to understand the complex relations between trauma, social support, and academic and
social-emotional outcomes.

91

Trauma Symptoms
Although participants did report experiencing symptoms of trauma following alleged
maltreatment, the majority of the sample did not report clinically significant levels of traumatic
stress following a child protective services investigation of alleged maltreatment. This suggests
that the event experienced may not have been perceived as traumatic or that children were able to
cope effectively and did not develop symptoms of traumatic stress. In the current study, 8.7% of
the sample exhibited clinically significant trauma symptoms as measured by the Trauma
Symptom Checklist for Children (TSCC). Previous studies have reported that 11.7% of
maltreated youth had significant levels of traumatic stress, as measured by the TSCC (Kolko et
al., 2010). The prevalence observed in the current study is slightly lower than previous
assessments. This difference may be due to differences between the populations sampled. The
current study assessed trauma symptoms among children ages 11 to 17 years old while Kolko et
al. (2010) sampled children ages 8 to 14 years old. While each of these prevalence rates are low
to moderate, they represent a significant number of children in the population that experience
significant levels of trauma symptoms following maltreatment.
Contrary to prior research, the type of maltreatment reported, the severity of
maltreatment, substantiation of the investigation, and demographic characteristics of age and
gender did not predict the level of perceived trauma symptoms. Previous research has
demonstrated that children that experience less severe maltreatment develop fewer symptoms of
trauma compared to those that experience more severe maltreatment (Hyman & Williams, 2011).
Regarding demographic characteristics, males tend to develop fewer trauma symptoms compared
to females after experiencing a potentially traumatic event (Tolin & Foa, 2006). Mixed findings
have been observed related to the impact of the other hypothesized predictors. Findings from the

92

current study suggest that none of these maltreatment or demographic variables directly predict
perceived traumatic stress. This null finding may indicate that trauma symptoms do not vary
based on the type of maltreatment, subjective severity, substantiation, or child characteristics
including age and gender. Children's own perceptions of the potentially traumatic event may
influence their level of distress independent of these other factors.
A likely explanation of this finding is that a combination of factors predicts trauma
symptoms following maltreatment, with no one variable significantly predicting trauma alone.
Previous studies have assessed a variety of hypothesized predictors of trauma symptoms with
mixed findings. A recent meta-analysis compiled 34 longitudinal studies assessing predictors of
trauma symptoms (Alisic, Jongmans, van Wesel, & Kleber, 2011). They found that none of the
assessed predictors explained a large part of long-term posttraumatic stress symptoms;
significant predictors explained up to 31% of variance in trauma symptoms (Alisic et al., 2011).
This suggests that an individual's response to a traumatic event may be influenced by the
interaction or combination of a variety of factors, rather than by individual predictors as assessed
within the current study.
Alternatively, this null finding may be the result of limited variance in trauma symptoms
and measurement of the hypothesized predictors. Type of maltreatment was based on the
caseworker’s report of the primary type of maltreatment under investigation. This measure does
not capture other forms of maltreatment that the participant may have experienced during the
incident under investigation or in previous incidents of abuse or neglect. Additionally, response
options were dichotomized to represent forms of maltreatment that were primarily abuse (i.e.,
physical, emotional, sexual abuse) and those that were primarily neglect (i.e., physical neglect)
due to limited sample sizes within each individual category. Combining these categories may

93

have created excessive variance, or “noise,” within the variable or may have masked differences
between individual categories. Therefore, distinct patterns may not be observed. Similarly,
severity of maltreatment was assessed using the caseworker’s report of the subjective level of
harm experienced by the child. This rating may not be a valid representation of the actual
severity of maltreatment. The variable was also categorized into three response options by
combining the responses of “moderate” and “severe” due to small sample sizes to create the third
category of “moderate to severe.” Differences may have been observed in participants' perceived
trauma symptoms between these two categories if they had not been combined.
Consistent with previous research, trauma symptoms were associated with elevated
internalizing and externalizing behavior problems. Trauma symptoms had a moderate positive
relation to internalizing behaviors and to externalizing behaviors. Following maltreatment,
children and adolescents may experience significant trauma caused by the abuse and neglect.
They may also experience trauma and dysfunction related to the child protective services
investigation processes and changes to their primary caregiver and living situation. These
symptoms of trauma were related to negative social-emotional outcomes. This demonstrates that
this population is in need of support to prevent these negative outcomes.
Trauma symptoms were not directly related to academic achievement or academic
engagement. This finding was unexpected as trauma has been shown to interfere with cognitive
processes required for learning such as sustained attention. There may be several reasons for this
null finding. First, trauma symptoms may cease to influence academics after several years have
passed since the traumatic event. In the current study, trauma symptoms were assessed at the
start of a child protective services investigation and academic achievement was assessed
approximately 18 months after the investigation. During that time, the negative effects of trauma

94

may have subsided. Additionally, children may have received supplemental supports from
caregivers, social services, or educational personnel which improved educational performance.
Second, clinical levels of trauma symptoms may affect academic outcomes while non-clinical
levels of trauma may not significantly affect academics. The current study included children that
were the focus of an investigation of maltreatment, regardless of the level of trauma symptoms.
If children with clinically significant trauma symptoms were assessed separately, a significant
relation between trauma and academic outcomes may have been observed at the time of
assessment.
Furthermore, the methods of assessing academic achievement within this study may not
adequately represent students’ actual school performance. Standardized measures of academic
achievement, such as the Woodcock Johnson III, are valid indicators of academic knowledge and
skills and are often used in research studies. Achievement within the school setting, however, is
generally assessed through homework assignments and in-class tests. Students affected by
trauma may possess adequate knowledge and skills, but may have difficulty demonstrating these
skills through homework assignments and tests. The finding within this study that trauma
symptoms do not predict academic achievement may represent participants’ academic
knowledge and skills rather than classroom achievement. This finding does not indicate whether
trauma affects students’ abilities to achieve academic success as measured by typical classroom
assessments.
Limitations
The current study utilized data from the NSCAW II data set. Use of a secondary data set
has many advantages, particularly access to a large sample of maltreated youth. However, there
are several limitations to the design and procedures of this study using measures from the

95

NSCAW II data set. This study utilized non-experimental analyses to understand the relation
between variables of interest. Variables of interest were not manipulated and participants were
not randomly assigned to conditions. Therefore, this study does not allow for causal explanations
of these relations. Results of this study suggest that trauma and social support statistically predict
outcomes. However, these results do not indicate whether these or other variables cause changes
in academic and social-emotional outcomes across time. Furthermore, this study sampled
children that were the subject of a child welfare investigation. A comparison sample of nonreferred children was not included. It is therefore not known whether the results of this study are
specific to maltreated individuals or whether these results characterize the general population.
This study is also limited by several measurement factors. First, the operationalization of
social support measures did not fully assess the conceptualization of this construct. As previously
discussed, social support was conceptualized within the current study as perceived emotional
support from caregivers and peers. This involves the perception that emotional support is
available if needed. The measures used to assess social support from caregiver and peer
relationships were each developed by different researchers for different purposes and were each
designed to measure different constructs. The caregiver measure was designed to assess
perceived relatedness with caregivers. It assessed participants’ emotions while with their
caregiver and their perceptions regarding the caregiver’s feelings toward the child. These
statements may contribute to or be secondary to children’s perceived emotional support, but they
do not directly assess children’s perceptions of the level of emotional support available from
caregivers. The measure used to assess peer social support was designed to assess the absence of
strong peer relationships. It assessed participants’ assessment of the availability of friends at
school and whether they got along with peers at school. This measure also focused on peer

96

relationships in the school setting and excluded interactions with peers in other settings. Similar
to the caregiver measure, the availability and compatibility of friendships in the school
environment may contribute to or may be secondary to children’s perceived emotional support,
but it does not directly assess peer social support. As the operationalization of the caregiver and
peer social support measures do not directly assess perceived emotional social support, this limits
the interpretation of these findings.
Second, the measures used to assess caregiver and peer social support were not parallel,
or related scales. The measures included a different number of items, different item questions,
and a different range of response options. Additionally, each of these measures included many
items that were negatively phrased (e.g., “I have nobody to talk to at school” and “When I’m
with my caregiver, I feel unhappy”). These negatively phrased statements may have caused
participants to overemphasize negative interactions with caregivers and peers compared to
neutral or positive interactions. Equal measurement of negative, neutral, and positive statements
regarding these relationships may have provided a more valid representation of perceived social
support. Differences in children’s responses to these items may be a function of these
characteristics of the instruments rather than a result of quantifiable differences in perceived
social support.
Third, the relationship to the caregiver and length of time participants had known the
caregiver may have also influenced ratings of perceived social support. Participants were asked
to respond to items relating to their current caregiver. The caregiver may have been the child’s
biological parent, a relative, a foster parent, or other guardian. The length of time the child lived
with the caregiver prior to data collection varied. This causes several potential limitations. If the
child was not removed from the home where maltreatment occurred, social support ratings may

97

be directed toward the perpetrator of abuse or neglect. If the child was removed from the home,
social support ratings may be directed toward a caregiver the child may not have known for a
significant period. These differences may have influenced children’s ratings of perceived social
support from the current caregiver.
Fourth, social-emotional well-being was assessed through rating scales completed by
children and caregivers rather than a direct measure of behavior. Children and caregivers
reported their perceived observations of internalizing and externalizing problems which may not
be as valid as a direct measure of children's behavior. Caregiver ratings may have error due to
possible frustration caused by the child’s trauma symptoms and behavior problems, or due to not
knowing the child long enough to accurately rate behavior problems. Children’s self-report
ratings may have error due to experiencing significant behavior problems and lacking sufficient
awareness of their own behaviors. Children, especially those with significant mental health
concerns, may not accurately identify elevated externalizing and internalizing behavior
problems. Therefore, self-reported outcomes may not represent children’s true challenges. Since
caregiver and child ratings may each over- or under-estimate children's internalizing and
externalizing behaviors, both measures were included in the latent variables assessing
internalizing and externalizing problems. Due to these design and measurement limitations,
results from this study should be interpreted with caution.
Implications for Practice
Results from this study suggest that the presence of emotional social support may
promote more positive outcomes among maltreated youth. Therefore, it is important for
professionals working with this population to assess sources of emotional support to identify
potential areas for interventions. Social support may be obtained through home-based supportive

98

relationships and school-based relationships. Results of this study indicated that social support
from multiple sources benefit different areas of adolescents' functioning. Following assessment
of social support, professionals should actively promote social support and foster supportive
relationships for children with a history of maltreatment through relationships at home and at
school. As children with a history of maltreatment may continue to be in the primary care of the
perpetrator of maltreatment, professionals should consider children's relationships with their
primary caregiver and other family members. In some cases, it may be more beneficial to seek
support from an older sibling or extended family member. Caseworkers should teach individuals
effective caregiving practices and provide strategies to create a warm and supportive
environment for children. Helping children and caregivers develop a positive relationship is
especially important when children begin living with a new or temporary caregiver during a
maltreatment investigation. Teachers and other school professionals may help students with a
history of maltreatment connect with their peers and help them develop appropriate and
supportive relationships. As maltreated youth often display externalizing behavior problems that
may be aversive to peers, school professionals should help maltreated youth develop appropriate
social skills through direct instruction, modeling, and feedback. While school professionals
cannot create friendships, they can facilitate positive interactions between students that may lead
to friendships. Teachers may provide students with a partner to complete class activities with.
They may assign seats where the student with a history of maltreatment is near friendly peers.
Schools may even create peer support groups that may be a safe space for students to discuss
challenging issues. Fostering emotional social support in any way may help promote more
positive outcomes in the presence of trauma.

99

Future Research
Caregiver and peer social support are uniquely related to social-emotional outcomes and
should therefore both be considered when promoting social support and assessing social support.
Research studies examining the role of social support among maltreated youth should measure
social support from multiple sources such as current caregiver, peers, and teachers. As previously
discussed, most research in this area has examined social support through caregiver relationships.
This limited view of social support does not provide a complete understanding of the role of
social support among maltreated youth. Results of this study indicated that trauma is more
strongly related to poor internalizing behaviors than poor externalizing behaviors. Maltreatment
may have a greater negative impact on children and adolescents' emotional well-being. The
examination of social support demonstrated that social support from peer relationships is related
to fewer internalizing behaviors problems. Therefore, peer social support may address the
primary effects of trauma and should be considered in both research and practice with this
population.
Further research is needed to understand the effect of trauma on academic outcomes and
the potential role of social support as a protective factor in promoting positive academic
outcomes. Neither trauma nor social support predicted academic achievement as measured by
standardized achievement tests within this study. Caregiver social support predicted academic
engagement, but peer social support and trauma symptoms were not related to academic
engagement. Future research should examine other indicators of academic success such as grades
or grade point averages, homework submission, summative assessments, and graduation rates.
Researchers should assess whether trauma symptoms are related to these measures of academic
achievement.

100

Finally, longitudinal research is needed to assess causal relationships between trauma,
social support, and academic and social-emotional well-being. Without longitudinal analyses,
conclusions regarding causal relations cannot be determined. The process of resilience occurs
over time. Therefore, variables of interest should be examined over time. While measures of
social support and academic and social-emotional well-being may not be available prior to the
occurrence of maltreatment, these measures can still be analyzed over time following
maltreatment. The current study demonstrated that trauma is significantly related to poorer
social-emotional well-being and that social support is related to better social-emotional wellbeing and academic engagement. These results suggest that the presence of social support may
lead to more positive outcomes among victims of maltreatment.
To further assess these patterns, resilience could be conceptualized and measured as
maintenance or improvement in outcome variables over time. Specifically, analysis of residuals
could be used to assess change over time within outcome variables. Non-negative residuals
would indicate participants who exhibited resilience between two time points. Assessing the
relation between predictor variables such as social support and residual outcomes could more
accurately assess the role of social support in promoting resilience. Additionally, a model-based
approach may be used to assess differences between groups. Latent profile analyses could be
used to identify groups of participants that exhibit maintained or improved academic and socialemotional outcomes. In other words, this methodology could be used to identify participants that
exhibit resilience. Differences between these groups or classes could then be assessed to
determine characteristics, or protective factors, that are associated with a greater likelihood of
resilient outcomes. These analysis methods would provide more conclusive evidence that

101

caregiver and peer social support may serve as protective factors, promoting positive outcomes
in the presence of risk.

102

APPENDIX

103

Figure 5
Structural Equation Model Assessing Caregiver Social Support

104

Figure 6
Structural Equation Model Assessing Peer Social Support

105

REFERENCES

106

REFERENCES

Aber, J. L., Allen, J., Carlson, V., & Cicchetti, D. (1989). The effects of maltreatment on
development during early childhood: Recent studies and their theoretical clinical and
policy implications. In D. Cicchetti & V. Carlson (Eds.), Child maltreatment: Theory and
research on causes and consequences, 579-619. Cambridge University Press, Cambridge.
Achenbach, T. M. (1991). Manual for Child Behavior Checklist/ 4-18 and 1991 Profile.
Burlington, VT: University of Vermont, Dept. of Psychiatry
Action. (2008, August). Child Safety and Substantiation of Maltreatment. In Action for Child
Protection. Retrieved May 1, 2016, from
http://action4cp.org/documents/2008/doc/August_Child_Safety_and_Substantiation_of_
Child_Maltreatment_08042008.doc
Afifi, T. O., & MacMillan, H. L. (2011). Resilience following child maltreatment: A review of
protective factors. Canadian Journal of Psychiatry, 56(5).
doi:10.1177/070674371105600505
Alisic, E., Jongmans, M. J., van Wesel, F., & Kleber, R. J. (2011). Building child trauma theory
from longitudinal studies: A meta-analysis. Clinical Psychology Review, 31(5), 736-747.
doi:10.1016/j.cpr.2011.03.001
Allen, R. E., & Oliver, J. M. (1982). The effects of child maltreatment on language development.
Child Abuse & Neglect, 6, 299-305.
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders
(5th ed.). Washington, DC: Author.
Asher, S., & Wheeler, V. (1985). Children’s loneliness: A comparison of rejected and neglected
peer status. Journal of Consulting and Clinical Psychology, 53(4), 500-505.
Banyard, V. L., & Williams, L. M. (2007). Women’s voices on recovery: A multi-method study
of the complexity of recovery from sexual abuse. Child Abuse & Neglect, 31(3), 275-290.
doi:10.1016/j.chiabu.2006.02.016
Barrera, M. (1981). Social support in the adjustment of pregnant adolescents: Assessment issues.
In B. H. Gottlieb (Ed.), Social networks and social support (pp. 69-96). Beverly Hills,
CA: Sage.
Barrera, M. (1986). Distinctions between social support concepts, measures, and models.
American Journal of Community Psychology, 14, 413-445. doi:10.1007/BF00922627
Boden, J. M., Horwood, L. J., & Fergusson, D. M. (2007). Exposure to childhood sexual and
physical abuse and subsequent educational achievement outcomes. Child Abuse &
Neglect, 31, 1101-1114. doi:10.1016/j.chiabu.2007.03.022
107

Brewin, C. R., Andrews, B., & Valentine, J. D. (2000). Meta-analysis of risk factors for
posttraumatic stress disorder in trauma-exposed adults. Journal of consulting and
Clinical Psychology, 68, 748-766. doi:10.1037//0022-006X.68.5.748
Briere, J. (1996) Trauma Symptom Checklist for Children: Professional Manual. Florida:
Psychological Assessment Resources Inc.
Brittain, C. V. (1968). An exploration of the bases of peer-compliance and parent-compliance in
adolescence. Adolescence, 13, 445-458.
Camara, M., Bacigalupe, G., & Padilla, P. (2014). The role of social support in adolescents: Are
you helping me or stressing me out? International Journal of Adolescence and Youth,
doi:10.1080/02673843.2013.875480
Child Welfare Information Gateway. (2011). About CAPTA: A legislative history. Washington,
DC: U.S. Department of Health and Human Services, Children’s Bureau.
Child Welfare Information Gateway. (2014). Definitions of child abuse and neglect. Washington,
DC: U.S. Department of Health and Human Services, Children's Bureau.
Cicchetti, D., Rogosch, F., Lynch, M., & Holt, K. (1993). Resilience in maltreated children:
Processes leading to adaptive outcome. Development and Psychology, 5, 629-647.
Cohen, S., & Willis, T. H. (1985). Stress, social support, and the buffering hypothesis.
Psychological Bulletin, 98(2), 310-357.
Connell, J. P. (1990). Context, self, and action: A motivational analysis of self-system processes
across the life span. In D. Cicchetti & M. Beeghly (Eds.), The self in transition (pp. 6197). Chicago: University of Chicago Press.
Connell, J. P., & Wellborn, J. G. (1991). Competence, autonomy, and relatedness: A
motivational analysis of self-system processes. In M. Gunnar & L. A. Sroufe (Eds.),
Minnesota Symposium on Child Psychology: Vol. 23. Self processes in development.
Chicago: University of Chicago Press.
Crittenden, P. M. (1992). Children’s strategies for coping with adverse home environments: An
interpretation using attachment theory. Child Abuse & Neglect, 16, 329–343.
Dingfelder, H. E., Jaffee, S. R., & Mandell, D. S. (2010). The impact of social support on
depressive symptoms among adolescents in the child welfare system: A propensity score
analysis. Children and Youth Services Review, 32(10), 1255-1261.
doi:10.1016/j.childyouth.2010.04.016
DuMont, K. A., Widom, C. S., & Czaja, S. J. (2007). Predictors of resilience in abused and
neglected children grown-up: The role of individual and neighborhood characteristics.
Child Abuse & Neglect, 31, 255-274.

108

English, D. J., Graham, J. C., Litrownik, A. J., Everson, M., & Bangdiwala, S. I. (2005).
Defining maltreatment chronicity: Are there differences in child outcomes? Child Abuse
& Neglect, 29(5), 575-595. doi:10.1016/j.chiabu.2004.08.009
Evans, S. E., Steel, A. L., & DiLillo, D. (2013). Child maltreatment severity and adult trauma
symptoms: Does perceived social support play a buffering role? Child Abuse & Neglect
37, 934-943. doi:10.1016/j.chiabu.2013.03.005
Finkelhor, D., Turner, H. a, Shattuck, A., & Hamby, S. L. (2013). Violence, crime, and abuse
exposure in a national sample of children and youth: an update. JAMA Pediatrics, 167(7),
614–21. doi:10.1001/jamapediatrics.2013.42
Flett, G.L., Druckman, T., Hewitt, P.L. et al. J Rat-Emo Cognitive-Behav Ther (2012) 30: 118.
doi:10.1007/s10942-011-0132-6
Folger, S. F. & Wright, M. O. (2013). Altering risk following child maltreatment: Family and
friend support as protective factors. Journal of Family Violence, 28, 325-337.
Ford, J. D., Wasser, T., & Connor, D. F. (2011). Identifying and determining the symptom
severity associated with polyvictimization among psychiatrically impaired children in the
outpatient setting. Child Maltreatment, 16(3), 216-226.
Fortuin, J., van Geel, M., & Vedder, P. (2015). Peer influences on internalizing and externalizing
problems among adolescents: A longitudinal social network analysis. Journal of Youth
and Adolescence, 44(4), 887. doi:10.1007/s10964-014-0168-x
Fuller, T., & Nieto, M. (2009). Substantiation and maltreatment reporting: A propensity score
analysis. Child Maltreatment, 14(1), 27-37.
Garmezy, N., Masten, A. S., & Tellegen, A. (1984). The study of stress and competence in
children: A building block for developmental psychopathology. Child Development,
55(1), 97-111. doi:10.1111/j.1467-8624.1984.tb00276.x
Geiser, S., & Santelices, M. V. (2007). Validity of high-school grades in predicting student
success beyond the freshman year (Research and Occasional Paper Series
No.CSHE.6.07). University of California, Berkeley, Center for Studies in Higher
Education.
Gifford-Smith, M. E. & Brownell, C. A. (2003). Childhood peer relationships: Social acceptance,
friendships and peer networks. Journal of School Psychology, 41(4), 235-284.
Gleason, M. E. J. & Iida, M. (2015). Social support. In M. Mikulincer and P. R. Shaver (Eds.),
APA Handbook of Personality and Social Psychology: Vol 3. Interpersonal Relations
(pp. 351-360). Washington, DC, US: American Psychological Association.
Gottlieb, B. H. (1978). The development and application of a classification scheme of informal
helping behaviors. Canadian Journal of Behavioral Science, 10, 105-115.

109

Gottlieb, B. H. (1983). Social support strategies: Guidelines for mental health practice. Sage,
Beverly Hills, CA.
Gottlieb, B. H., & Bergen, A. E. (2010). Social support concepts and measures. Journal of
Psychosomatic Research, 69(5), 511-520. doi:10.1016/j.jpsychores.2009.10.001
Haber, M. G., Cohen, J. L., Lucas, T., & Baltes, B. B. (2007). The relationship between selfreported received and perceived social support: A meta-analytic review. American
Journal of Community Psychology, 39, 133-144.
Helsen, M., Vollebergh, W., & Meeus, W. (2000). Social support from parents and friends and
emotional problems in adolescence. Journal of Youth and Adolescence, 29(3), 319-335.
doi:10.1023/A:1005147708827
Herrenkohl, E. C., Herrenkohl, R. C. & Egolf, B. (1994). Resilient early school-age children
from maltreating homes: Outcomes in late adolescence. Journal of Orthopsychiatry,
64(2), 301-309.
Hildyard, K. L., & Wolfe, D. a. (2002). Child neglect: Developmental issues and outcomes.
Child Abuse and Neglect, 26(6-7), 679–695. doi:10.1016/S0145-2134(02)00341-1
Hofstra, M. B., Van Der Ende, J., & Verhulst, F. C. (2002). Child and adolescent problems
predict DSM-IV disorders in adulthood: A 14-year follow-up of a dutch epidemiological
sample. Journal of the American Academy of Child & Adolescent Psychiatry, 41(2), 182189. doi:10.1097/00004583-200202000-00012
House, J. S., Kahn, R. L., McLeod, J. D., & Williams, D. (1985). Measures and concepts of
social support. In S. Cohen & S. Syme (Eds.), Social support and health (pp. 83–108).
San Diego, CA: Academic Press.
Hussey, J. M., Marshall, J. M., English, D. J., Knight, E. D., Lau, A. S., Dubowitz, H. & Kotch,
J. B. (2005). Defining maltreatment according to substantiation: Distinction without a
difference? Child Abuse & Neglect, 29, 479-492. doi:10.1016/j.chiabu.2003.12.005
Hussey, J. M., Chang, J. J., & Kotch, J. B. (2006). Child maltreatment in the United States:
Prevalence, risk factors, and adolescent health consequences. Pediatrics, 118(3), 933942.
Hyman, B., & Williams, L. (2001). Resilience among women survivors of child sexual abuse.
Affilia, 16(2), 198-219. doi:10.1177/08861090122094226
IBM Corp. Released 2013. IBM SPSS Statistics for Windows, Version 22.0. Armonk, NY: IBM
Corp.
Jaffee, S. R., & Maikovich-Fong, A. K. (2011). Effects of chronic maltreatment and
maltreatment timing on children’s behavior and cognitive abilities. Journal of Child
Psychology and Psychiatry and Allied Disciplines, 52(2), 184–194. doi:10.1111/j.14697610.2010.02304.x
110

Juvonen, J., Espinoza, G., & Knifsend, C. (2012). The role of peer relationships in student
academic and extracurricular engagement. (2012th ed., pp. 387-401). Boston, MA:
Springer US. doi:10.1007/978-1-4614-2018-7_18
Kaplow, J. B., & Widom, C. S. (2007). Age of onset of child maltreatment predicts long-term
mental health outcomes. Journal of Abnormal Psychology, 116(1), 176-187.
doi:10.1037/0021-843X.116.1.176
Kaufman, J., & Cicchetti, D. (1989). Effects of maltreatment on school-age children’s
socioemotional development: Assessments in a day camp setting. Developmental
Psychology, 25, 516-524.
Kinard, E. M. (1999). Psychosocial resources and academic performance in abused children.
Children and Youth Services Review, 21(5), 351-376.
Kolko, D. J., Hurlburt, M. S., Zhang, J., Barth, R. P., Leslie, L. K., & Burns, B. J. (2010).
Posttraumatic stress symptoms in children and adolescents referred for child welfare
investigation: A national sample of in-home and out-of-home care. Child Maltreatment,
15(1), 48–63. doi:10.1177/1077559509337892
Lakey, B., & Drew, J. B. (1997). A social-cognitive perspective on social support. New York,
NY: Plenum Press.
Lamis, D. A., Wilson, C. K., King, N. M., & Kaslow, N., J. (2014). Child abuse, social support,
and social functioning in African American children. Journal of Family Violence, 29,
881-891. doi:10.1007/s10896-014-9639-9
Langford, C., Bowsher, J., Maloney, J., & Lillis, P. (1997). Social support: A conceptual
analysis. Journal of Advanced Nursing, 25(1), 95-100. doi:10.1046/j.13652648.1997.1997025095.x
Leeb, R. T., Paulozzi, L., Melanson, C. Simon, T., & Arias, I. (2008). Child maltreatment
surveillance: Uniform definitions for public health and recommended data elements,
version 1.0. Atlanta, GA: Centers for Disease Control and Prevention, National Center
for Injury Prevention and Control.
Li, F., & Godinet, M. T. (2014). The impact of repeated maltreatment on behavioral trajectories
from early childhood to early adolescence. Children and Youth Services Review, 36, 22–
29. doi:10.1016/j.childyouth.2013.10.014
Litty, C. G., Kowalski, R., & Minor, S. (1996). Moderating effects of physical abuse and
perceived social support on the potential to abuse. Child Abuse & Neglect, 20, 305-314.
Loth, A. K., Drabick, D. A. G., Leibenluft, E., & Hulvershorn, L. A. (2014). Do childhood
externalizing disorders predict adult depression? A meta-analysis. Journal of Abnormal
Child Psychology, 42(7), 1103-1113. doi:10.1007/s10802-014-9867-8

111

Luthar, S. S. (2006). Resilience in development: A synthesis of research across five decades. In
D. Cicchetti, & D. Cohen (Eds.) Developmental psychopathology: Risk, disorder, and
adaptation (vol. 3, 2nd edn, pp. 739-795). New York: Wiley.
Luthar, S. S., Cicchetti, D., & Becker, B. (2013). The construct of resilience: A critical
evaluation and guidelines for future work. Child Development, 71(3), 543–62.
Lynch, M., & Cicchetti, D. (1991). Patterns of relatedness in maltreated and nonmaltreated
children: Connections among multiple representational models. Development and
Psychopathology, 3, 207-226.
Martinez-Torteya, C., Anne Bogat, G., von Eye, A., & Levendosky, A. A. (2009). Resilience
among children exposed to domestic violence: The role of risk and protective factors.
Child Development, 80(2), 562-577. doi:10.1111/j.1467-8624.2009.01279.x
Masten, A. S. (2001). Ordinary magic. Resilience processes in development. The American
Psychologist, 56(3), 227–238. doi:10.1037/0003-066X.56.3.227
Masten, A. S., Garmezy, N., Tellegen, a, Pellegrini, D. S., Larkin, K., & Larsen, a. (1988).
Competence and stress in school children: The moderating effects of individual and
family qualities. Journal of Child Psychology and Psychiatry, and Allied Disciplines,
29(6), 745–764. doi:10.1111/j.1469-7610.1988.tb00751.x
Masten, A. S. (2007). Resilience in developing systems: Progress and promise as the fourth wave
rises. Developmental Psychopathology, 19(3), 921-930.
Masten, A. S. (2011). Resilience in children threatened by extreme adversity: Frameworks for
research, practice, and translational synergy. Developmental Psychopathology, 23(2),
141-154.
Masten, A. S., & Obradovic, J. (2006). Competence and resilience in development. Anny N. Y.
Acad. Sci 1094, 13-27.
McLewin, L. A., & Muller, R. T. (2006). Attachment and social support in the prediction of
psychopathology among young adults with and without a history of physical
maltreatment. Child Abuse & Neglect, 30, 171-191. doi: 10.1016/j.chiabu.2005.10.004
Muller, R. T., Goebel-Fabbri, A. E., Diamond, T., & Dinklage, D. (2000). Social support and the
relationship between family and community violence exposure and psychopathology
among high risk adolescents. Child Abuse & Neglect, 24(4), 449-464.
doi:10.1016/S0145-2134(00)00117-4
Muthén, L.K. and Muthén, B.O. (1998-2012). Mplus User’s Guide. Seventh Edition. Los
Angeles, CA: Muthén & Muthén
National Survey of Child and Adolescent Well-being (2007). NSCAW Combined Waves 1–5
Data File User’s Manual – Restricted Release. Ithaca, NY: National Data Archive on
Child Abuse and Neglect.
112

Ogle, C. M., Rubin, D. C., & Siegler, I. C. (2013). The impact of the developmental timing of
trauma exposure on PTSD symptoms and psychosocial functioning among older
adults. Developmental Psychology, 49(11), 2191-2200. doi:10.1037/a0031985
Pearlin, L. I., & McCall, M. E. (1990). Occupational stress and marital support: A description of
microprocesses. In J. Eckenrode & S. Gore (Eds.), Stress Between Work and Family.
New York: Plenum.
Pepin, E. N., & Banyard, V. L. (2006). Social support: A mediator between child maltreatment
and developmental outcomes. Journal of Youth and Adolescence, 35(4). 617-630.
doi:10.1007/s10964-006-9063-4
Reef, J. (2010). Children's problems predict adults' DSM-IV disorders across 24 years. J Am
Acad Child Adolesc Psychiatry, 49(11), 1117-1124. doi:10.1016/j.jaac.2010.08.002
Reitz, E., Deković, M., & Meijer, A. M. (2006). Relations between parenting and externalizing
and internalizing problem behaviour in early adolescence: Child behaviour as moderator
and predictor. Journal of Adolescence, 29(3), 419-436.
doi:10.1016/j.adolescence.2005.08.003
Rose, A. J., Carlson, W., & Waller, E. M. (2007). Prospective associations of co-rumination with
friendship and emotional adjustment: Considering the socioemotional trade-offs of corumination. Developmental Psychology, 43(4), 1019-1031. doi:10.1037/00121649.43.4.1019
Rosenthal, S., Feiring, C., & Taska, L. (2003). Emotional support and adjustment over a year’s
time following sexual abuse discovery. Child Abuse & Neglect, 27, 641-661.
Rutter, M. (2013). Annual research review: Resilience – clinical implications. Journal of Child
Psychology and Psychiatry, 54(4), 474-487.
Scott, L. A., Ingels, S. J., & National Center for Education Statistics. (2007). Interpreting 12thgraders' NAEP-scaled mathematics performance using high school predictors and
postsecondary outcomes from the national education longitudinal study of 1988 (NELS:
88). Statistical Analysis Report. NCES 2007-328 Distributed by ERIC Clearinghouse.
Seeds. P. M., Harkness, K. L., & Quilty, L. C. (2010). Parental maltreatment, bullying, and
adolescent depression: Evidence for the mediating role of perceived social support.
Journal of Clinical Child & Adolescent Psychology, 39(5), 681-692.
doi:10.1080/15374416.2010.501289
Shonk, S. M., & Cicchetti, D. (2001). Maltreatment, competency deficits, and risk for academic
and behavioral maladjustment. Developmental Psychology, 37, 3-17.
Snyder, S. M. & Smith, R. E. (2015). Do youth with substantiated child maltreatment
investigations have distinct patterns of delinquent behaviors? Children and Youth
Services Review, 58, 82-89. doi:10.1016/j.childyouth.2015.09.008

113

Kendall-Tackett, K. A., & Eckenrode, J. (1996). The effects of neglect on academic achievement
and disciplinary problems: A developmental perspective. Child Abuse & Neglect, 20(3),
161-169.
Thompson, R., & Tabone, J. K. (2010). The impact of early alleged maltreatment on behavioral
trajectories. Child Abuse & Neglect, 34(12), 907-916. doi:10.1016/j.chiabu.2010.06.006
Tolin, D. F., & Foa, E. B. (2006). Sex differences in trauma and posttraumatic stress disorder: A
qualitative review of 25 years of research. Psychological Bulletin, 132(6), 959-992.
Toth, S. L., & Cicchetti, D. (1996). The impact of relatedness with mother on school functioning
in maltreated children. Journal of School Psychology, 34(3), 247-266.
Types of Child Abuse. (2015). In Adults Surviving Child Abuse. Retrieved October 21, 2015,
from http://www.asca.org.au/WHAT-WE-DO/Resources/General-Information/Types-ofchild-abuse#emotional
U.S. Department of Health and Human Services, Administration for Children and Families,
Administration on Children, Youth and Families, Children’s Bureau. (2015). Child
maltreatment 2013. Available from http://www.acf.hhs.gov/programs/cb/research-datatechnology/statistic....
Ungar, M., Ghazinour, M., Richter, J. (2013). Annual research review: What is resilience within
the social ecology of human development. Journal of Child Psychology and
Psychiatry, 54(4), 348-366. doi:10.1111/jcpp.12025
van Beest, M., & Baerveldt, C. (1999). The relationship between adolescents' social support from
parents and from peers. Adolescence, 34(133), 193.
Vranceanu, A., Hobfoll, S. E., & Johnson, R. J. (2007). Child multi-type maltreatment and
associated depression and PTSD symptoms: The role of social support and stress. Child
Abuse & Neglect, 31, 71-84. doi:10.1016/j.chiabu.2006.04.010
Walsh, W. A., Dawson, J., & Mattingly, M. J. (2010). How are we measuring resilience
following childhood maltreatment? Is the research adequate and consistent? What is the
impact on research, practice, and policy? Trauma, Violence, and Abuse, 11, 27-41.
Wechsler‐Zimring, A., & Kearney, C. A. (2011). Posttraumatic stress and related symptoms
among neglected and physically and sexually maltreated adolescents. Journal of
Traumatic Stress, 24(5), 601-604. doi:10.1002/jts.20683
White, R., & Renk, K. (2012). Externalizing behavior problems during adolescence: An
ecological perspective. Journal of Child and Family Studies, 21(1), 158-171.
doi:10.1007/s10826-011-9459-y
Wildeman, C., Emanuel, N., Leventhal, J. M., Putnam-Hornstein, E., Waldfogel, J., & Lee, H.
(2014). The Prevalence of Confirmed Maltreatment Among US Children, 2004 to 2011.
JAMA Pediatrics, 06520(8), 1–8. doi:10.1001/jamapediatrics.2014.410
114

Wilson, L. C., & Scarpa, A. (2014). Childhood abuse, perceived social support, and
posttraumatic stress symptoms: A moderation model. Psychological Trauma: Theory,
Research, Practice, and Policy, 6(5), 512–518. doi:10.1037/a0032635
Woodcock, R.W., McGrew, K.S, & Mather, N. (2004). Woodcock-Johnson III Tests of
Achievement. Itasca, IL: Riverside Publishing Company.
Woodruff, K., & Lee, B. (2011). Identifying and predicting problem behavior trajectories among
pre-school children investigated for child abuse and neglect. Child Abuse and Neglect,
35(7), 491–503. doi:10.1016/j.chiabu.2011.03.007

115