University of Southern California Dissertations and Theses

Attachment, maltreatment and autonomic nervous system responsivity as predictors of adolescent anxiety and depression

(USC Thesis Other)

Attachment, maltreatment and autonomic nervous system responsivity as predictors of adolescent anxiety and depression

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content ATTACHMENT, MALTREATMENT AND AUTONOMIC NERVOUS SYSTEM
RESPONSIVITY AS PREDICTORS OF ADOLESCENT ANXIETY AND
DEPRESSION

by
Nashla Feres

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PSYCHOLOGY)

August 2010

Copyright 2010 Nashla Feres
ii

Table of Contents
List of Tables iii
List of Figures iv
Abstract v
Chapter One: Introduction 1
Attachment 5
Attachment & Internalizing Problems 6
Attachment and Physiology 9
Physiology and Internalizing Problems 13
Child Maltreatment 14
Child Maltreatment & Internalizing Problems 16
Child Maltreatment & Physiology 17
The Current Study 19

Chapter Two: Methods
Participants 25
Measures 26
Procedure 30

Chapter Three: Results
Preliminary Analyses 35
Specific Aim 1 51
Specific Aims 2-4 52

Chapter Four: Discussion 67
Limitations and Future Research 72
Strengths 73

References 75

iii

List of Tables
Table 1: Demographics for Maltreated and Comparison Groups at T2 and T3 27
Table 2: Descriptive Statistics for Continuous Variables 36
Table 3: Intercorrelations among Final Variables with Covariances in the
Diagonal 37

Table 4: T-tests for Males vs. Females 45
Table 5: Post-hoc Results for Age using LSD 45
Table 6: Post-hoc Results for Ethnicity using LSD 48
Table 7: Post-hoc Results for Income using LSD 49
Table 8: Univariate ANCOVA results for Comparison vs. Maltreated
Adolescents with Adjusted Means and Standard Errors 53

Table 9: Summary of Results 68

iv

List of Figures
Figure 1: Conceptual model for Hypothesis 1 21
Figure 2: Conceptual model for Hypothesis 2 22
Figure 3: Conceptual model for Hypothesis 3 23
Figure 4: Conceptual model for Hypothesis 4 23
Figure 5: Measurement model for mother and father attachment as measured
by the Security Scale 39

Figure 6: Measurement model for parasympathetic and sympathetic nervous
systems as measured by respiratory sinus arrhythmia recovery (RSAREC)
and skin conductance level recovery (SCLREC) 42

Figure 7: Measurement model for anxiety and depression as measured by the
Child Depression Inventory (CDI) and Multidimensional Anxiety Scale for
Children (MASC) 43

Figure 8: RSA Baseline, Conflict and Recovery for Four Groups 54

Figure 9: SCL Baseline, Conflict and Recovery for Four Groups 55

Figure 10: Father and mother attachment predicts internalizing problems 58

Figure 11: SNS and PNS predicts internalizing problems 59

Figure 12: Attachment predicts ANS responsivity 61
Figure 13: SNS mediates attachment and depression 63
Figure 14: Conceptual models for each of the four groups with significant
parameters 65

v

Abstract
The current study investigated four hypotheses. Hypothesis 1: Low attachment
security will predict high levels of anxiety and depression; maltreatment group may
moderate these relationships. Hypothesis 2: Low RSA and high SCL will predict high
levels of both anxiety and depression; maltreatment group may moderate these
relationships. Hypothesis 3: Low attachment security will predict low RSA and high
SCL responsivity; maltreatment group may moderate these relationships. Hypothesis
4: ANS responsivity will mediate the association between attachment and
internalizing problems. Specifically, low attachment security will predict low RSA
responsivity and high SCL responsivity which in turn will predict high levels of
anxiety and depression; maltreatment group may moderate these relationships.
Data were collected from 359 maltreated and comparison adolescents who
identified as African American, Hispanic or Caucasian. Attachment Security with
both mother and father, depression and anxiety symptoms were self-reported by
adolescents. Skin conductance level and respiratory sinus arrhythmia were collected
during a video presentation intended to induce emotional arousal.
Structural equation modeling was used to evaluate a multiple group mediational
model. Results indicated that attachment security with mother predicted depressive
symptoms 2 years later, whereas attachment security with father predicted anxiety
symptoms 2 years later. Furthermore, both RSA and SCL predicted anxiety
symptoms 2 years later. Finally, results showed that SNS mediated the relationship
between father attachment security and anxiety but only among comparison females.
1

Chapter One: Introduction
Over the past few decades a substantial amount of data has indicated that insecure
parent-child attachment relationships and child maltreatment are highly negative
experiences that put children at risk for developing later behavioral problems such as
anxiety and depressive disorders. There is a separate literature which substantiates a link
between these parent-child experiences and dysregulated physiology, and there is even a
body of research supporting a link between dysregulated physiology and internalizing
problems. Until now however, no research has tested physiology as a mediating factor
between these early traumatic experiences and later internalizing problems; yet it has
been proposed (Luecken, Appelhans, Kraft, & Brown, 2006). If there is support for
physiology as a mediator between parent-child experiences and internalizing problems,
researchers and clinicians may want to examine possible methods of incorporating
physiology into their interventions and therapeutic techniques.
National prevalence rates for a persistent depressive or anxiety disorder in the
U.S. have been reported at 4.7% (Young, Klap, Shoai, & Wells, 2008) and 15%
respectively (Beesdo, Knappe, & Pine, 2009). Furthermore, women are approximately
1.6 times more likely than men to report a lifetime history of depression or anxiety
(Ialongo et al., 2004; Kessler, McGonagle, Swartz, Blazer, & Nelson, 1993; Martin-
Merino, Ruigomez, Wallander, Johansson, & Garcia-Rodriguez, 2010) and three times
more likely in adolescence (Beesdo, Knappe, & Pine, 2009).
The ability to regulate one’s emotions, particularly negative emotions, is a skill
that is learned early in development. Emotion regulation is linked with the right brain,
2

which is the first hemisphere to undergo rapid development during the first 18 months of
life (Schore, 1994). This timeframe also coincides with the development of an initial
attachment with a primary caregiver. Neurological studies show that the brain is
physically altered by social interactions during the critical period in which strategies for
emotion regulation are developing (Schore, 2000, 2001; Siegel, 1999, 2001). Both
negative and positive interactions with the primary attachment figure(s) influence the
manner in which the infant and young child form neural connections and interact with the
environment. Thus, it is the role of caregivers during this critical period that has the
greatest impact on a child’s developing brain, specifically for learning emotion
regulation. A child’s survival depends on the ability of his caretaker to provide him with
care, and “those who are nurtured best survive best” (Cozolino, 2006, p. 14). Thus,
depending on a variety of parent-child experiences, such as attachment style, the
experience of maltreatment, and other family stressors, a child may not learn adaptive
coping strategies in order to deal with stress or negative moods and emotions.
It is suggested that depression and anxiety directly result from counterproductive
or ineffective attempts to “regulate acute affective episodes that lead to the exacerbation
and persistence of unwanted emotion” (Campbell-Sills & Barlow, 2007 p. 543).
Depressed and anxious individuals are most likely to suppress their unwanted emotions
rather than cope with them which only results in short term relief (Baker, Holloway,
Thomas, Thomas, & Owens, 2004; Campbell-Sills, Barlow, Brown, & Hofmann, 2006;
Levitt, Brown, Orsillo, & Barlow, 2004). Because coping strategies are learned early in
life before a language system develops, infants process this social information primarily
3

through emotional encoding. This emotional encoding, learned from caregivers, alters
right brain physiology which makes maladaptive responses difficult to change later in life
(Schore, 2000). Thus, early maladaptive social-emotional experiences (i.e. the learning of
maladaptive coping skills) that have altered physiology will increase the liklihood of
developing an anxiety or depressive disorder. One physiological system that may be
altered is the autonomic nervous system (ANS).
The ANS consists of two branches called the parasympathetic (PNS) and
sympathetic (SNS) nervous systems which help maintain homeostasis in the body. Porges
(2001) has advocated a tri-part autonomic nervous system which he calls polyvagal
theory. According to polyvagal theory, each branch of the ANS developed at different
points in our evolutionary history. Developing first was the PNS, responsible for
metabolic needs, rest, and repair of the body. For example, when heart rate is too high,
the PNS increases frequency of signals to the heart in order to decrease heart rate and this
is considered a normal, adaptive response. The part of the PNS said to be in charge of
regulating primitive functions, is what Porges calls the vegetative vagus, which regulates
heart rate and blood pressure (Hugdahl, 1995).
Second to evolve was the SNS, responsible for triggering action in states of
emergency or stress, also known as the fight/flight response. Reacting to real and present
danger with the assistance of SNS activation is a highly adaptive and necessary response
for survival. In contrast to the PNS, when the SNS increases frequency of signals, bodily
functions such as heart rate increase, allowing more blood flow to the limbs of the body.
4

Most recent to develop was the right branch of the vagus, part of the PNS, which
Porges refers to the as the “smart” vagus. The smart vagus is said to be associated with
higher level processing of emotion and communication. Thus, Porges commonly refers to
the smart vagus as the social engagement system because the development of this system
depends on and is aimed at social interactions. For example, modern social situations
may trigger the fight/flight system (SNS), but to react with anger or fear to a non-fatal
threat is typically not socially appropriate; this is when the smart vagus intercedes. The
purpose of the smart vagus is to put a “brake” on the SNS so that a person can experience
anger, but in a socially appropriate and regulated way. In other words, the smart vagus is
said to be a “social” brake applied to the impulses that are automatically triggered (when
our body experiences danger, threat, sexual desire, etc.) when it is not socially
appropriate to respond to them. The clear benefit of this system is to allow social
connections to continue by suppressing the behavior suggested by our initial
physiological indicators.
Many studies have provided consistent data showing that abnormal ANS
responsivity (hypo- or hyper-arousal), is related to maladaptive outcomes such as
aggression, conduct disorder and antisocial personality (Raine, Lenczk, Bihrle, LaCasse,
& Colletti, 2000; Raine, Reynolds, Venables, & Mednick, 1997; Raine, Venables, &
Williams, 1990). It is suspected that if the social engagement system is underdeveloped
because of insecure attachments or trauma, then one may come to depend on the more
primitive ANS subsystems within relationships and lack the social brake provided by the
more evolved smart vagus. This is a problem because such frequent use of primitive ANS
5

subsystems results in dramatic and long-lasting emotional and bodily reactions to
interpersonal stress (Porges, 2001). For example, high SNS reactivity to non-threatening
stimuli in the case of hyper-vigilant individuals (usually as a response to trauma) can
overburden and weaken the body’s immune system. Similarly, low PNS reactivity during
states of minor stress will hinder a person from returning to a calm state, potentially
leaving an individual in a constant state of low-level anxiety.
The following sections will examine the extant research on the relationships
between attachment, internalizing problems, physiology and maltreatment in order to
provide support for hypotheses regarding the nature of these relationships.
Attachment
A great deal of attachment research has been conducted over the past 4 decades,
beginning with the groundbreaking work of John Bowlby and Mary Ainsworth. John
Bowlby first laid the foundation for attachment theory, presenting evidence that early
parent-child interactions crucially impact a child’s view of self and others in regard to
relationships. Ainsworth and colleagues developed the first laboratory experiment which
examined the separation and reunion behavior of mothers and their 1-year old children
(Ainsworth, Blehar, Waters, & Wall, 1978). Much of the research today continues to use
Ainsworth and colleagues’ classification of secure and insecure attachment styles.
The development of attachment is a universal, evolutionary-based mechanism that
increases an infant’s likelihood of survival (Bowlby, 1969; Bretherton & Munholland,
1999). A child’s behavioral-motivational system monitors the availability and proximity
of an attachment figure and organizes behavior according to the perceived intentions and
6

behaviors of that attachment figure (Bretherton & Munholland, 1999). Whereas the
function of attachment is evolutionary-based and universal, the quality of attachment is
learned and varies from person to person (Crittenden, 1995). Based on interactions with a
caregiver who is either supportive, rejecting, or inconsistent, infants will develop
strategies for coping with daily stressors that may be adaptive or maladaptive (Bretherton
& Munholland, 1999; Crittenden, 1995; Kobak & Sceery, 1988).
Ainsworth’s paradigm, The Strange Situation, assessed an infant’s reactions
during several separation and reunion sessions with the primary attachment figure
(usually the mother). Coded videotapes of the infants’ behaviors allowed Ainsworth and
colleagues to develop an infant classification system to discriminate healthy (secure)
from unhealthy (insecure) parent-child relationships. The three distinct attachment styles
observed were classified as secure, insecure-avoidant, and insecure-preoccupied.
Researchers throughout the years have altered and expanded these classifications, and a
fourth category (disorganized) has been added to the classification system (Main &
Cassidy, 1988).
Attachment and Internalizing Problems
Early attachments are to a large extent the mechanism behind the regulation of
affect, in which caregivers help infants manage emotional tension that the infants do not
have the capacity to regulate on their own (Fury, Carlson, & Sroufe, 1997; Kochanska,
2001). For example, internalizing problems such as anxiety or depressive disorders may
be due to dysregulation of the fear system (Cole, Zahn-Waxler, & Smith, 1994;
Kochanska, 2001). Kochanska (2001) found that insecurely classified infants showed a
7

significantly higher increase of negative emotions from infancy through toddlerhood and
more distress in situations designed to elicit joy, suggesting that insecurely attached
children have learned to view the environment more harshly and fearfully than secure
children. It is these negative expectations about people and the environment that directly
influence later social and emotional success (Kerns, Klepac, & Cole, 1996).
Over the past two decades, hundreds of studies have been published on the
relationship between attachment and internalizing problems. Generally, the literature on
attachment has described a higher incidence of internalizing problems among those who
are insecurely attached compared with those who are securely attached. Many studies
have only discriminated secure from insecure, either because the questionnaire used was
dimensional or the sample size was insufficient to discriminate insecure types (Costa et
al., 2009; Essau, 2004; Gullone, Ollendick, & King, 2006; Margolese, Markiewicz, &
Doyle, 2005; Muris, Mayer, & Meesters, 2000; Muris, Meesters, & van den Berg, 2003;
Richaud de Minzi, 2006; Roelofs, Meesters, Huurne, Bamelis, & Muris, 2006; Ronnlund
& Karlsson, 2006; van Brakel, Muris, Bogels, & Thomassen, 2006).
Some studies however, have examined insecure types more specifically, which
usually identify either disorganized or preoccupied attachment as most predictive of
internalizing disorders or suicidal tendencies (Adam, Sheldon-Keller, & West, 1996;
Allen, Moore, Kuperminc, & Bell, 1998; Cooper, Shaver, & Collins, 1998; Salzman,
1996; Wright, Briggs, & Behringer, 2005). Of note are some of the findings from
longitudinal studies, such as those of Warren, Huston, Egeland & Sroufe (1997) who
found that infant anxious attachment predicted anxiety disorders at age 17.5 years
8

(Bosquet & Egeland, 2006). In contrast, a few studies have pointed at avoidant
attachment as most predictive of depression (Anan & Barnett, 1999; Cyranowski et al.,
2002). Interestingly, a 37-year longitudinal study found only indirect effects of early
attachment (through later romantic relationships) on depression and anxiety at age 37
(Overbeek, Stattin, Vermulst, Ha, & Engels, 2007).
Although a variety of measurement instruments (i.e. self-reports, interviews,
observations) and techniques (i.e. prospective and retrospective) were employed across
these studies, they all agree on the same point: Those with insecure attachments are at
greater risk for developing an anxiety or depressive disorder compared with their secure
counterparts. Also of note is that most studies which did examine subtypes of attachment
found that preoccupied attachment was most predictive of anxiety problems, whereas
avoidant attachment was most predictive of depression. These findings indicate that
insecure and disorganized types are at greater risk for developing internalizing problems
(Carlson, 1998; Egeland & Sroufe, 1981; Finzi, Cohen, Sapir, & Weizman, 2000;
Weinfield, Sroufe, & Egeland, 2000). Although attachment style is not examined in the
current study, the dimensional measure of security that is used has been validated against
a categorical measure of attachment (i.e. The Separation Anxiety Test), indicating that
insecure types were also low on the dimension of security (Kerns, Tomich, & Kim,
2006). It is likely that those adolescents in the current study who are low in security
would therefore fall into one of the insecure or disorganized subtypes; it is also likely that
they have developed dysregulated physiological responses.

9

Attachment and Physiology
Extant research on the relationship between physiology and attachment has only
examined physiology in a descriptive capacity- that is, as a method of differentiating
attachment styles but not as a consequence or predictor of attachment style. Among this
research, mostly infant and adult populations have been examined, leaving out middle
childhood and adolescence. Despite these omissions however, the attachment-physiology
literature has provided evidence to suggest that different attachment experiences are
indicative of varying physiology; the next step is to test whether the attachment
experiences actually play a role in altering physiology.
It has been theorized that insecure-preoccupied children attempt to engage in an
intimate relationship with the caregiver, but because the affection is not reciprocated the
child develops a diminished trust in relationships (Vivona, 2000). More specifically,
preoccupied children receive inconsistent attention from a caregiver, who is typically
overwhelmed and concerned only with her own needs (Howe, 2005). Because
preoccupied children are more concerned about their caregiver’s feelings than the
caregiver is about the child’s, the child will develop maladaptive coping strategies.
Therefore these children are not successful in dealing with stressful situations and
subsequently, have difficulties being calmed or comforted. Such difficulties in stressful
situations have been supported by researchers such as Kobak and Sceery (1988), who
found that preoccupied children were more likely to display high levels of anxiety and
low levels of self-confidence than other styles. Initial experiences with an attachment
figure in combination with higher levels of anxiety and low levels of self-esteem likely
10

cause preoccupied children and adolescents to withdraw from social situations with peers
due to their feelings of inadequacy, fear of rejection (or reciprocation), and helplessness
(Finnegan, Hodges, & Perry, 1996). It is therefore suspected that preoccupied individuals
have an ANS biased toward SNS arousal, as they are often irritable, impulsive, act out,
and have a decreased ability to recover from stress (Cozolino, 2006).
In contrast, insecure-avoidant individuals are likely to have an ANS biased toward
PNS arousal in that they typically show low levels of emotional expression, which results
in reducing the proximity of others and exploring the environment. Research has found
that these individuals have a lower resting heart rate and decreased physical activity and
are likely to be depressed and unmotivated (Cozolino, 2006). This response is likely due
to the experience with a primary caregiver as someone who is not interested in her child’s
emotional needs, and may in fact punish the child for showing emotion (Howe, 2005).
These children have learned that they must soothe themselves and that it is not acceptable
to show any level of emotional need or response.
Research with infants in the Strange Situation has found that avoidant infants
showed decreased respiratory sinus arrhythmia after separation distress compared with
secure infants, supporting the theory of a dysregulated PNS (Hill-Soderlund et al., 2008).
However, a study on a low-income minority sample of adolescent mothers and their
infants found that heart rate was non-distinguishable amongst attachment groups in the
Strange Situation and only the infants’ behavior was different and notable (Zelenko et al.,
2005). Specifically, Zelenko and colleagues found that avoidant infants did not show any
outward signs of behavioral distress, whereas preoccupied infants were the most
11

behaviorally distressed. This finding may suggest that heart rate is not a distinguishing
marker for the SNS (as it is dually innervated by both the PNS and SNS) and that another
indicator, such as skin conductance would be a more appropriate measure.
In contrast to the infant research, more unique paradigms have been used in child
and adolescent studies. One study in the Netherlands examined skin conductance
responsivity during the Trier Social Stressor Task for 7-year olds and found that secure
attachment and the presence of a double long allele on the serotonin transporter gene (5-
HTT) appeared to buffer stress reactivity, resulting in a lower skin conductance response
(Gilissen, Bakermans-Kranenberg, van IJzendoorn, & Linting, 2008). This research
supports the theory that an insecure attachment style is related to greater activation of the
SNS- even in situations with only minor stress.
Another study with 14-16 year-old high school students, monitored adolescents’
blood pressure (BP) and heart rate (HR) over two days while at school (Gallo &
Matthews, 2006). Results differentiated preoccupied and avoidant adolescents during
their interactions with close friends. Specifically, preoccupied individuals had higher HR
and BP during interactions with friends whereas avoidantly attached adolescents had
higher HR and BP only during conflict with friends. The authors suggest that these
findings support the theory that preoccupied individuals are in a constant state of anxiety
and preoccupation about abandonment when it comes to close interpersonal relationships,
which is shown by their elevated sympathetic arousal. In contrast, avoidant individuals,
who attempt to dismiss the need for close relationships, have difficulties when faced with
conflict, as shown by their elevated sympathetic arousal.
12

A final set of attachment-physiology studies examined romantic relationships in
adults, the majority of which were sampled from undergraduate psychology courses. The
general findings from these studies indicate that avoidantly attached individuals suppress
their negative emotions behaviorally (similar to studies of avoidant infants), which is
related to higher skin conductance and/or heart rate during attachment or conflict related
tasks (Diamond, Hicks, & Otter-Henderson, 2006; Dozier & Kobak, 1992; Feeney &
Kirkpatrick, 1996; Holland & Roisman, 2010; Roisman, 2007; Roisman, Tsai, & Chiang,
2004) and that anxious attachment is related to lower levels of resting respiratory sinus
arrhythmia (Diamond & Hicks, 2005). These studies of young adults, while limited to
those in a college population, are still consistent with other findings indicating an over-
reactive SNS and under-reactive PNS in response to conflict among insecurely attached
individuals.
Taken together, these studies indicate some potentially significant findings. First,
depending on the task in which physiology is measured (i.e. whether it is evokes fears of
separation or conflict in close relationships) avoidant and preoccupied individuals both
respond with heightened physiological distress due to dysregulation of the sympathetic
nervous system. Furthermore, there is evidence in both infancy and adulthood that an
insecure attachment is related to a lower-level of RSA responsivity, indicative of
dysregulation of the parasympathetic nervous system.
Physiology and Internalizing Problems
The research on physiology and internalizing problems has been relatively
consistent over the past few decades. Generally, the literature shows that anxiety
13

disorders are associated with low RSA responsivity (Watkins, Grossman, Krishnan, &
Sherwood, 1998), higher blood pressure (Yeragani, Tancer, Seema, Josyula, & Desai,
2006) and higher skin conductance reactivity to a startle task (Bakker, Tijssen, van der
Meer, Koelman, & Boer, 2009). Like anxiety, research on depression has also found that
respiratory sinus arrhythmia is lower than normal, but in contrast to anxiety, those with
depression also have lower than normal heart rates (Rottenberg, Wilhelm, Gross, &
Gotlib, 2003; Tonhajzerova et al., 2010; Udupa et al., 2007) and resting skin conductance
level (Ward & Doerr, 1986). Unfortunately, studies which did not differentiate anxiety
from depression have yielded confusing and sometimes contradictory results (Argyle,
1991; Dietrich et al., 2007; Hildrum, Mykletun, Holmen, & Dahl, 2008).
Two studies of note examined not only baseline RSA but changes from baseline
to a stressful task. Hastings and colleagues (2008) examined RSA in preschool children
and noted that those who showed an increase in RSA (denoting healthy activation of the
parasympathetic system) from baseline to social stressor had less internalizing problems
compared to those who didn’t change. Similarly, in a longitudinal study of African
American and Caucasian 9-year olds, Hinnant and El-Sheikh (2009) found that the
interaction between low baseline RSA and high RSA suppression during a stressful task
was associated with the highest levels of internalizing problems 2 years later.
One issue with these physiological studies is the paradigms in which physiology
is measured. Baseline or resting states measurements will not necessarily illustrate how a
person copes with stress or conflict, whereas measurements during a stimuli or stress-
inducing tasks will. Depending on when the physiological measurements are taken,
14

results could vary. For example, depressed individuals have shown lower resting skin
conductance levels, yet if these individuals have preoccupied attachments, a relationship
task may show that these individuals have a high skin conductance response. Another
issue with previous research is that many studies do not discriminate anxiety from
depressive disorders, which are likely to exhibit differing physiology (i.e. hyper-aroused
versus hypo-aroused). Even when some studies do specify a particular disorder (e.g.
depression), they often fail to mention which disorders within that category are included
(e.g. dysthymia, anhedonia, seasonal-affective disorder, etc.) and whether they are
including those with severe or minor symptoms, or both.
Although such methodological issues present some difficulties in interpreting
previous findings, these studies do provide support for differing physiology among those
with and without internalizing problemst. The next step is to investigate what factors may
influence the relationship between dysregulated physiology and internalizing problems.
There is a significant amount of research which suggests that in addition to parent-child
attachment relationships, the experience of child-maltreatment may play a significant
role.
Child Maltreatment
Although child maltreatment has been a problem for quite some time, researchers
and clinicians have struggled with defining and classifying maltreatment experiences.
This is a difficult task because of the multitude of variables that can drastically alter the
impact of a maltreatment experience. Symptoms and outcomes of maltreatment often
depend on various factors involved in the maltreatment such as the type, severity, age of
15

onset, chronicity, and abuser type (i.e. family member, acquaintance or stranger; Trickett
& McBride-Chang, 1995). In many cases a child will experience more than one type of
maltreatment (i.e. physical and sexual), which further increases the likelihood and
severity of symptoms and disorders to follow.
Generally, child maltreatment has been shown to affect a child’s ability to
regulate emotions and behave appropriately in social situations (Bolger & Patterson,
2001). Abused children typically have less intimate relationships, more negative affect
and more aggression with peers than non-abused children (Parker & Herrera, 1996;
Salzinger, Feldman, Hammer, & Rosario, 1993). Chronic maltreatment has been
associated with heightened risk of peer rejection in childhood and adolescence (Cole,
Zahn-Waxler, & Smith, 1994). Without opportunities to learn appropriate perspective
taking skills with parents or peers, maltreated children often have difficulties
understanding appropriate affective responses to interpersonal situations and thus have
limited social problem-solving skills.
Research shows that physical abuse drives aggressive child behaviors, whereas
neglectful parenting steers children toward withdrawn behavior (Cicchetti, Lynch, Shonk,
& Manly, 1992; Rubin, LeMare, & Lollis, 1990; Salzinger, Feldman, Hammer, &
Rosario, 1993). Physical abuse leads to a variety of negative outcomes for a child such as
increased impulsivity and irritability, hypervigilance and paranoia, curtailed recognition
of pain in self and others, and increased attribution of hostile intent (McCarty &
McMahon, 2003). In contrast, neglected and emotionally maltreated children are more
likely to experience depression and attempt suicide (Kaplan et al., 1999). Child
16

maltreatment is associated with many clinical symptoms and disorders such as
depression, suicidality, self mutilation, dissociative identity disorder, borderline
personality disorder, substance abuse, and others (Putnam & Trickett, 1997; Trickett &
McBride-Chang, 1995).
Child Maltreatment and Internalizing Problems
The abundance of research on maltreatment and internalizing problems is similar
to that of attachment; however there are a few methodological issues in the existing
literature. One issue is that many studies have examined general maltreatment, whereas
others have specified types of maltreatment. Furthermore, there have been many
discrepancies as to the criteria by which maltreatment groups have been defined and in
how maltreatment has been verified (i.e. by case records or reported retrospectively). For
a detailed description of measurement and definitional issues see (Feerick, Knutson,
Trickett, & Flanzer, 2006). Despite these issues however, the literature has generally
concluded that the experience of child maltreatment increases the risk of developing
internalizing problems compared with those who have not been maltreated (Brown,
Craig, & Harris, 2008; Polanczyk et al., 2009; Schultz, Tharp-Taylor, Haviland, &
Jaycox, 2009; Toth & Cicchetti, 1996).
In those studies examining subtypes of maltreatment, some have claimed that
physical abuse is most predictive of internalizing problems (Kim & Cicchetti, 2006;
Salzinger, Rosario, Feldman, & Ng-Mak, 2007) while others have claimed that emotional
(i.e. psychological) maltreatment is most predictive of internalizing problems (Liu, Alloy,
Abramson, Iacoviello, & Whitehouse, 2009; McLewin & Muller, 2006; Powers, Ressler,
17

& Bradley, 2009; Spinhoven et al., 2010; Stein, Schork, & Gelernter, 2008; Stuewig &
McCloskey, 2005) and a third group has pointed to sexual abuse (Buzi, Weinman, &
Smith, 2007; Spinhoven et al., 2010; Turner, Finkelhor, & Ormrod, 2010).
Clearly, more precise data should be collected (with standard definitions and
criteria of maltreatment across studies) so that we may better understand the etiology,
underlying mechanisms, and developmental course of distinct maltreatment experiences.
Specifically, the unique and shared variance of different types of maltreatment needs to
be examined in light of the fact that multiple types of maltreatment are often experienced
simultaneously (Trickett, Mennen, Kim, & Sang, 2009).
Beyond a lack consistency of measurement and definitions, few have been able to
explain exactly how and why some maltreated individuals develop internalizing problems
whereas others do not. Several cognitive models have been proposed, suggesting that a
poor self-concept, negative attitudes, and a negative attributional style are the mediators
of this relationship. While cognitions must certainly play a role in later behavioral
outcomes, it is likely that altered physiology may also have a developmental impact.
Child Maltreatment and Physiology
The maltreatment-physiology literature is similar to the attachment-physiology
literature in that physiology has primarily been examined as a method of distinguishing
patterns among groups. Interestingly, no research has examined ANS as a consequence of
maltreatment; rather, researchers have focused on hormonal changes due to maltreatment.
In fact, a great deal of evidence has been presented which supports altered and
dysregulated hormonal physiology as a result of maltreatment (Gordis, Granger, Susman,
18

& Trickett, 2007; Linares et al., 2008; Putnam & Trickett, 1997). In particular,
researchers have examined cortisol, salivary amylase and other indicators of
neuroendocrine and hypothalamic-pituitary-adrenal (HPA) axis functioning to inform
these findings, which indicate that maltreated children have elevated cortisol levels and
delayed growth hormones (Carpenter et al., 2009; Kaplan et al., 1999; MacMillan et al.,
2009; Tarullo & Gunnar, 2006; Tyrka et al., 2009; van der Vegt, van der Ende,
Kirschbaum, Verhulst, & Tiemeier, 2009).
Surprisingly little research has examined autonomic nervous system differences in
maltreated samples. However, the research which has been conducted has generally
found that maltreated individuals are physiologically distinct from their non-maltreated
counterparts. For example, Gordis, Feres, Olezeski, Rabkin, & Trickett (2009) found that
RSA and SCL moderated the relationship between maltreatment and aggression for
adolescents aged 9-16 years, which varied by gender. Further, in a study of sexually and
physically abused children aged 7-13 years, skin conductance response was significantly
lower across eight cognitive tasks (approximately 27 minutes) for abused children
compared with control children (Carrey, Butter, & Persinger, 1995). These particular
findings indicate an inability of the maltreated children to adequately engage in tasks
(e.g. math and concept formations) which require focus and concentration.
In a sample of women with a history of childhood sexual abuse, those who
developed post-traumatic stress disorder (PTSD) had a significantly higher heart rate
during a personal story task compared to those who did not develop PTSD (Orr et al.,
1998). Interestingly, a recent study examining women with a history of maltreatment but
19

without PTSD symptoms discovered that although heart rate did not vary across groups,
RSA did. In particular, women with an abuse history had lower RSA than non-abused
women during a minimal exercise task; they also took longer to recover back to baseline
levels compared with non-abused women. These results indicate an altered
parasympathetic response, likely due to the experience of maltreatment (Dale et al.,
2009).
On the whole, previous research on maltreatment and physiology indicates that
growth hormones may be delayed and that there are likely to be higher levels of stress
hormones among maltreated individuals when presented with moderate stressors.
Furthermore, there is some evidence of sympathetic and parasympathetic dysregulation
within maltreated persons. Still, it is unclear whether there is a particular pattern of
physiological dysfunction which varies by type, length or severity of maltreatment.
While these studies present evidence for the link between maltreatment and physiology,
autonomic nervous system indicators (i.e. skin conductance reactivity and respiratory
sinus arrhythmia) have not been commonly examined.
The Current Study
There are several ways in which research on attachment, physiology, internalizing
problems and maltreatment could be improved. First, attachment research has historically
ignored the importance of the father-child attachment relationship. It has only been
relatively recently that some studies have begun to include measures of father-child
attachment. Unfortunately, some of these studies have had the mother report on the
father-child relationship rather than the father. Especially when many children and
20

adolescents are often raised without fathers or with substitute fathers, examination of
these relationships could be critical. One strength of the current study is the examination
of father-child attachment in addition to mother-child attachment. In examining these
relationships separately it will be possible to determine the potentially unique pathways
of each relationship.
There has been a lack of research examining the ANS in maltreated samples, as
the majority of research on maltreatment and physiology has investigated hormonal
differences. Additionally, due to inconsistent definitions and methodologies, research
examining the links between maltreatment and internalizing problems has been limited in
its conclusions regarding maltreatment and later outcomes. One strength of the current
study is the criteria by which the maltreated sample was selected (i.e. case records from
the Department of Child and Family Services rather than retrospective self-reports).
Furthermore, detailed information about each maltreatment case (e.g. type of abuse,
relationship to abuser, number of reports of abuse, etc.) is available for comparison with
other studies.
Finally, there has been a surprising dearth of literature connecting attachment,
physiology, internalizing problems and maltreatment within one model. Although a
substantial amount of research has shown that insecure attachments and maltreatment
experiences increase the likelihood of an anxiety or depressive disorder, the role of
physiology has only been examined as a method of differentiating these experiences, and
not as a mediator. One goal of the current study is to examine ANS responsivity as a
21

mediator of the relationship between attachment and internalizing problems with a
maltreated sample.
In order to extend current research on attachment, physiology, internalizing
problems and maltreatment, four specific aims will be examined in the current study:
1) To describe the differences between maltreated and comparison groups on attachment
security, ANS responsivity, and internalizing problems; 2) To examine attachment
security and ANS responsivity as predictors of internalizing problems among maltreated
and comparison groups; 3) To examine attachment as a predictor of ANS responsivity
and 4) To examine ANS responsivity as a mediator of attachment security and
internalizing problems among maltreated and comparison groups.
To examine Specific Aim 2, two hypotheses will be tested:
Hypothesis 1: Attachment security will predict internalizing problems. Specifically, low
attachment security will predict high levels of anxiety and depression; maltreatment
group may moderate these relationships.

Figure 1. Conceptual model for Hypothesis 1. Attachment=attachment security for mother and father as
measured by the Security Scale. Internalizing problems=anxiety as measured by the Multidimensional
Anxiety Scale for Children and depression as measured by the Child Depression Inventory.

Previous research indicates that an insecure attachment is predictive of
internalizing problems, regardless of exposure to maltreatment; therefore it is expected
that low attachment security will predict high levels of anxiety and depression among
Low
Attachment
Security
High
Internalizing
Problems
Group
22

adolescents in the entire sample. Since there is a limited amount of research examining
the role of father attachment, there are no predictions regarding which parent may have a
greater influence on internalizing problems.
Hypothesis 2: ANS responsivity will predict internalizing problems. Specifically, low
RSA and high SCL will predict high levels of both anxiety and depression; maltreatment
group may moderate these relationships.

Figure 2. Conceptual model for Hypothesis 2. PNS= parasympathetic nervous systems as measured by
respiratory sinus arrhythmia responsivity. SNS= sympathetic nervous system as measured by skin
conductance level responsivity. Internalizing problems= anxiety as measured by the Multidimensional
Anxiety Scale for Children and depression as measured by the Child Depression Inventory.

As there is some previous research indicating that there is differing physiology
among those with and without internalizing problems, in addition to the fact that emotion
regulation is developed early in childhood, it is hypothesized that ANS responsivity is a
predictor of internalizing problems. More specifically, adolescents with low RSA and
high SCL may be more likely to have high levels of both anxiety and depression.
To address Specific Aim 3, one hypothesis will be examined:
Hypothesis 3: Attachment security will predict PNS and SNS responsivity. Specifically,
low attachment security will predict low RSA and high SCL responsivity; maltreatment
group may moderate these relationships.

Low PNS &
High SNS
High
Internalizing
Problems
Group
23

Figure 3. Conceptual model for Hypothesis 3. Attachment= mother and father attachment security as
measured by the Security Scale. PNS= respiratory sinus arrhythmia responsivity. SCL= skin conductance
level responsivity.

Based on the fair amount of research examining the physiological differences
between insecure and secure individuals, it is hypothesized that low security adolescents
will have dysregulated physiology. Specifically, it is hypothesized that low attachment
security will predict low RSA and high SCL.
To examine Specific Aim 4, one hypothesis will be tested.

Hypothesis 4: ANS responsivity will mediate the association between attachment and
internalizing problems. Specifically, low attachment security will predict low RSA
responsivity and high SCL responsivity which in turn will predict high levels of anxiety
and depression; maltreatment group may moderate these relationships.

Figure 4. Conceptual model for Hypothesis 4. Group= maltreated or comparison. Attachment= attachment
security for mother and father as measured by the Security Scale. RSA= respiratory sinus arrhythmia
responsivity. SCL= skin conductance level responsivity. Anxiety & Depression= as measured by the
Multidimensional Anxiety Scale for Children and the Child Depression Inventory.

Finally, in an attempt to unite the research on attachment, maltreatment, ANS
responsivity and internalizing problems, a moderated mediational model will be
examined. Specifically, that PNS and SNS mediate the relationship between attachment
security and internalizing problems, which may be moderated by maltreatment group.
Low
Attachment
Security
Low RSA &
High SCL
High
Internalizing
Problems
Group
Low
Attachment
Security
Low PNS &
High SNS
Group
24

In summary, four specific aims will be examined in the current study:
Specific Aim 1: To describe the differences between maltreated and comparison groups
on attachment security, ANS responsivity, and internalizing problems.
Specific Aim 2: To examine attachment security and ANS responsivity as predictors of
internalizing problems among maltreated and comparison groups.
• Hypothesis 1: Attachment security will predict internalizing problems.
Specifically, low attachment security will predict high levels of anxiety and
depression; maltreatment group may moderate these relationships.
• Hypothesis 2: ANS responsivity will predict internalizing problems. Specifically,
low RSA and high SCL will predict high levels of both anxiety and depression;
maltreatment group may moderate these relationships.
Specific Aim 3: To examine attachment as a predictor of ANS responsivity.
• Hypothesis 3: Attachment security will predict PNS and SNS responsivity.
Specifically, low attachment security will predict low RSA and high SCL
responsivity; maltreatment group may moderate these relationships.
Specific Aim 4: To examine ANS responsivity as a mediator of attachment security and
internalizing problems among maltreated and comparison groups.
• Hypothesis 4: ANS responsivity will mediate the association between attachment
and internalizing problems. Specifically, low attachment security will predict low
RSA responsivity and high SCL responsivity which in turn will predict high
levels of anxiety and depression; maltreatment group may moderate these
relationships.
25

Chapter Two: Methods
Participants
The data collected for the proposed project were part of a National Institute of
Child Health and Development (NICHD) funded three-year longitudinal study on the
effects of maltreatment on adolescent development (a fourth year of data collection is
currently underway). The recruitment of all subjects was approved by the Los Angeles
Department of Child and Family Services (LADCFS), The Juvenile Court of Los Angeles
County, and the University of Southern California Institutional Review Board. In the
current study, only data from Time 2 (T2) and Time 3 (T3) were analyzed.
For recruitment of the maltreated sample (N=217 at T2), monthly updates were
coordinated with LADCFS for referrals which met the criteria for inclusion into the
study. These criteria were: 1) A case opened with LADCFS in the past month, 2) Child
between the ages of 9-12 years old, 3) Child’s ethnicity reported as Latino, African
American, or Caucasian (non-Hispanic), and 4) Child residing in one of 10 predetermined
zip codes within Los Angeles County. The selection of the 10 zip codes was based on
cases of maltreatment LADCFS within the three targeted ethnic groups. Further, it is
important to select children from the same zip codes to ensure that they have similar
community and neighborhood experiences.
For the comparison sample (N=142 at T2), a list of families with children ages 9-
12 years, from within the same 10 zip codes as those from the maltreated sample, were
provided by a service that offers names of local residents for direct marketing. Similar to
the maltreated sample, inclusion variables included: 1) A child aged 9-12 years old,
26

2) Child’s ethnicity reported as Latino, African American, or Caucasian (non-Hispanic),
and 3) Child residing in one of 10 predetermined zip codes within Los Angeles County.
Demographics for the maltreated and comparison groups are shown in Table 1. Families
were first contacted through the mail, with a letter describing the study and inviting their
participation. Caregivers were encouraged to return the enclosed pre-addressed and pre-
stamped postcard, indicating their interest in volunteering for the study. If families did
not respond within 10 days, a second contact was initiated via phone.
For the current study, 15% of the maltreated participant’s were excluded from
analyses – those whose case records were coded as “at risk”, “incapacitated caregiver”, or
“at risk sibling” (i.e. no record of any type of abuse or neglect; n= 24); participants who
were never maltreated by either biological parent (n=9); and participants who did not
have any detailed information available regarding their maltreatment experience (n=5).
These exclusions were made in an attempt to homogenize the sample so that the
maltreatment group included only those who had experienced actual maltreatment at the
hands of biological parents.
Measures
Attachment. The Security Scale (Kerns, Klepac, & Cole, 1996) was developed as
a dimensional measure of security for late childhood. For each of the 15 items, children
were instructed to choose which of two descriptions was most like them, and then to
further indicate whether that description was “really true” or “sort of true” for them.
Thus, for each item, four different answers were possible, indicating a range of security
from 1 (lowest score) to 4 (highest score). Scores are typically summed for mother and
27

Table 1.

Demographics for Maltreated and Comparison Groups at T2 and T3

Comparison Maltreated
T2 T3 T2 T3

Participation Rate 94% 83% 82% 52%
Final Sample 142 126 217 159
Age (SD) 12.28 (1.26) 13.82 (1.45) 11.99 (1.18) 13.46 (1.35)
Male/Female Ratio 85/57 73/53 101/116 69/90
African American 32.4% 33.3% 38.2% 44%
Latino 45.1% 42.9% 37.8% 32.1%
Caucasian 10.6% 11.1% 12% 8.8%
Biracial 12% 12.7% 12% 15.1%
Speak English only at home 74.1% 75% 66.4% 70.4%
Speak Spanish only at home 16.5% 15.3% 28.1% 25.8%
Speak both at home 9.4% 9.7% 5.1% 3.8%
Household Income <$30k 39.5% 41.1% 66.8% 64.1%
Household Income $30-60k 38.9% 37.9% 24.9% 28.3%
Household Income >$60k 21.6% 21% 6.9% 6.3%

Note. There were significantly (p<.05) more boys in the comparison group than the maltreated group.
Comparison children were significantly (p<.05) older at both T2 and T3. Comparison children came from
households with significantly (p<.01) higher household incomes than maltreated children at both T2 and
T3.

28

father scales separately, resulting in one score of security for each relationship. A range
between 15 and 60 points are possible for each scale, with lower scores indicating lower
levels of security and higher scores indicating higher levels of security.
One of the main strengths of the Security Scale compared to the other available
paper and pencil measures of attachment is that the items appropriately reflect the
dimensions outlined by attachment theory. After an intense review of potential measures,
the Security Scale was one of only two measures that met the criteria for examining
responsiveness and availability of the attachment figure. Additionally, several studies
have reported high reliability and validity of the Security Scale (Contreras, Kerns,
Weimer, Gentzler, & Tomich, 2000; Kerns, Klepac, & Cole, 1996; Kerns, Tomich,
Aspelmeier, & Contreras, 2000; Lieberman, Doyle, & Markiewicz, 1999). Internal
consistency reliability over 27 studies has been reported between α= .64 - .93, with the
majority in α=.80 range. In the current study, alpha reliability was high for both mother
(α=.82) and father (α=.90).
The Security Scale has also been validated against other measures of attachment,
which again, cannot be said for many of the other existing attachment measures;
specifically, the doll-play interview (Granot & Mayseless, 2001; Kerns, Abraham,
Schlegelmilch, & Morgan, 2007; Kerns, Brumariu, & Abraham, 2008) and Separation
Anxiety Test (Kerns, Tomich, & Kim, 2006; Resnick, 1993). Finally, test-retest
reliability over 3-years with father (r=.37, p<.001) and mother (r=.37, p<.001) was found
to be moderate (Verschueren & Marcoen, 2005).
29

Due to the nature of the current sample, many adolescents had experienced
multiple caregivers or were currently residing with foster parents. Of note is that a
significant number (11%) of adolescents in the current study had never known a father
figure and thus could not answer the Security Scale about a father. We therefore thought
it relevant to ask children prior to answering the Security Scale, a few preparatory
questions, including: 1) ‘Have you ever had more than one mom (or dad)?’ 2) If yes,
‘Who was your first mom? That is, the mom that raised you when you were a baby until
at least around age 5’ 3) ‘Do you currently live with this mom?’ 4) If no, ‘Do you still see
your mom?’ 5) If no, ‘When was the last time you saw your mom?’ Based on the
answers to these questions, interviewers then instructed the child to answer the
questionnaire about the mom who was their “first” mom, only if they still had contact
with her within the past year. If multiple parent figures were equally involved in the
child’s life, children were instructed to answer the questions about the first
(chronological) mother if the child was also still in contact with both.
Depression. Adolescents’ depressed mood was examined with the Child
Depression Inventory (CDI), a 27-item self report measure adapted from the adult Beck
Depression Inventory (Beck, Steer, & Garbin, 1988) for children ages 7-17 years
(Kovacs, 1985, , 1992). Children were asked to choose one of four statements for each
item which best represented how they felt over the past two weeks. Items were summed
for an overall rating of depressed mood with higher ratings indicating greater depressed
mood.
30

The author reports good internal consistency reliability (α= .70 - .86) and test-
retest reliability (α= .87 - .71). Further, the scale has been correlated with global ratings
of depression by clinicians and diagnoses of depression from a structured psychiatric
interview (Hodges, 1990). The measure includes five subscales (negative mood,
interpersonal problems, ineffectivity, anhedonia and negative self-esteem) and also yields
a total depression score. Alpha reliability for all 27 items in the current sample was
adequate (α =.83). Three subscales (negative mood, anhedonia and negative self-esteem),
which best fit a measurement model in the current sample, were used for analysis in the
present study and yielded alpha internal consistency reliability of .81.
Anxiety. Adolescents’ level of anxiety was measured with the Multidimensional
Anxiety Scale for Children (MASC), a 39-item self report measure designed for children
ages 8-17 years (March, Parker, Sullivan, & Stallings, 1997). The measure includes four
subscales called physical symptoms, social anxiety, harm avoidance, and separation
anxiety, in addition to a Total Anxiety score. The MASC has shown good internal
consistency ranging from .70 - .89, good test-retest reliability, and invariance across
gender and age (March, Parker, Sullivan, & Stallings, 1997). The Total Anxiety score
yielded a high alpha reliability (α=.91.) in the current sample, however the best fitting
measurement model excluded the harm avoidance subscale, which did not change the
alpha reliability (α=.91.).
Procedure
Children and their caregivers came into the project office and participated in a 4-5
hour interview at Time 2 (T2) and Time 3 (T3). During the interview, children and
31

caregivers were given cognitive, physiological, and mental health questionnaires.
Children and caregivers were paid for their time at the conclusion of each visit and were
not penalized if they did not complete the interview. Paper and pencil measures were
translated for both English and Spanish speakers, however all children preferred to
complete the interview in English.
At T2, the psychophysiological measure was introduced 1-2 hours into the four-
hour procedure, depending on the time of day the family visited the lab. A trained
interviewer explained the procedures to the child, assuring the child that no pokes or
prods would occur. The child was then directed to place the sticky side of the disposable
ECG electrodes onto his skin (one on each side of the torso and one above the navel), and
told that these would measure heart rate during the video clips. Next, a respiration
bellows was placed around the child’s midsection, and the child was told that this device
would measure how fast she was breathing during the video clips. Finally, Ag/AgCl
electrodes (filled with isotonic citrate salt electrode gel, limited to a 1-cm diameter circle
contact area by double-sided adhesive collars) were placed on the volar surfaces of the
child’s distal phalanges (on the non-dominant hand) and she was told that these would
measure how much she sweats during the video.
After the equipment was tested for proper functioning, the child was given
headphones to listen to the video. The child was told that he would view 20 minutes of
video clips that show a child in different situations with his/her parent. The interviewer
remained in the room with the child and sat behind a curtain during the presentation. The
child was viewed on a monitor and the interviewer noted any large movements that
32

would affect the readings. After the video was over, the interviewer removed the sensors
from the child and asked the child how the video clips made him feel.
Before any content clips were shown, the child sat quietly in front of a black
screen for three minutes in order to assess a baseline reading. After the baseline the child
watched two video clips without parent-child conflict, which were then followed with
four parent-child conflict clips. Immediately following the last conflict clip the child
again observed a blank screen for three minutes. Finally, the child watched a last clip in
which a parent-child conflict was resolved. The order of the clips was the same for every
child.
The conflict clips were selected to represent a range of maltreatment scenarios
that adolescents may have experienced. For example, clips include mother-daughter,
father-daughter, mother-son, and father-son interactions representing neglect or physical
or emotional abuse. Conflict clips were carefully selected for intense, but age-appropriate
scenarios. All video-clips were PG-rated, and were movies shown either on national TV
or were films created by film students at the University of Southern California.
Furthermore, caregivers were given the option to view the video clips before the
procedure if they were concerned about the content.
Although other measures of autonomic functioning were collected (i.e. heart rate
and blood pressure), the current study used only two of these indices to evaluate the
functioning of each branch of the autonomic nervous system: Respiratory Sinus
Arrhythmia (RSA) - a purely parasympathetic indicator, and Electrodermal Activity
(EDA) - a purely sympathetic indicator. A bioamplifier (James Long Company, Caroga
33

Lake, NY) recorded psychophysiological activity continuously while participants
watched the videos.
PNS-linked Responsivity. Respiratory Sinus Arrhythmia (RSA) is a naturally
occurring relationship between the variations in respiration rate and the timing between
successive R waves (Beauchaine, 2001). It is used to measure the fluctuations in heart
rate that coincide with respiration and is used as an indicator of vagal activity.
A standard strain gauge respiration bellows was wrapped around each subject’s
midsection and connected to the same bioelectric amplifier for continuous transduction,
amplification, and digitization. ECG data were sampled and digitized at 1 kHz and R-
wave times were extracted from the ECG channel and edited manually via ECGRWAVE
software (James Long, Caroga Lake, NY). The RSA program computed RSA as the
difference between the minimum interbeat interval (IBI) during inspiration and the
maximum IBI during expiration (calculated in seconds). For the current study, RSA was
calculated by averaging across two minutes of the recovery period after the last conflict
clip. The final minute of recovery was not used because it was disrupted by an
assessment of blood pressure.
SNS-linked Responsivity. Electrodermal activity, also referred to as skin
conductance, is recorded as changes in electrical resistance in the skin. It is well known
that as a person becomes emotionally stressed, sweat gland activity increases (Dawson,
Schell, & Filion, 2000). This increase in hydration makes the skin better able to conduct
an electric current. The change in current is recorded as a change in electrodermal
activity.
34

EDA can be measured with electrodes placed on the surface of the skin- usually in
areas dense with eccrine glands (i.e. hands and feet). In the current study, two non-
polarizable silver/silver chloride electrodes were placed on the first two fingertips of each
subject’s non-dominant hand, in order to capture the small current passed across the
electrodes. The James Long bioamplifer used a 500 mV, 30 Hz sinusoidal excitation
waveform to collect tonic skin conductance level (SCL); that is, the baseline activity
which reflects sustained attention and heightened arousal over time. Phasic responses,
(rapid responses reflecting the immediate impact of a stimulus) also known as skin
conductance reactivity (SCR) was also collected but not used in the current study. Skin
conductance level output was measured at 10 μs/V with an A/D converter with a 16-bit
resolution and a +/- 2.5 V input range and data were digitized at 1 kHz. Like RSA, SCL
was calculated by averaging across the first two minutes of the recovery period after the
last conflict clip.

35

Chapter Three: Results
Preliminary Analyses
Data Cleaning for RSA and SCL. For RSA the James Long system used for
psychophysiological data collection monitored and highlighted abnormal heart readings;
thus a trained staff member only needed to oversee the cleaning procedure, which the
system automatically conducted. Any abnormal readings were deleted (12% for
comparison and 8% for maltreated).
SCL data was also cleaned, though, without the help of any computer program.
Interviewer notes during the psychophysiological procedure were compared against the
raw SCL data. Any increase of SCL that was caused by any noted external noise,
participant movement, or equipment error was removed. This accounted for anywhere
between 10-200 seconds of missing data, depending on the type of problem. This deletion
meant that only valid data points were averaged for the final score.
Descriptive Statistics. Descriptive statistics (see Table 2) for all variables were
first examined for outliers, skewness, kurtosis and multicollinearity. As confirmed by the
5% trimmed means and boxplots, only RSA evidenced outliers. In order to conform to
the assumptions necessary for use of parametric tests 17 participants’ RSA scores were
trimmed to within the normal range of scores as indicated by the boxplot. Specifically,
the outlying cases were assigned the value one unit larger (or smaller) than the next most
extreme score in the distribution (Tabachnik & Fidell, 2007). All final variables were
then assessed for multicollinearity as indicated by the variable intercorrelations and
36

Table 2.
Descriptive Statistics for Continuous Variables

Variable Valid N Mean
5% Trimmed
Mean SD
Security Scale Mother 359 47.94 48.48 8.47
Security Scale Father 317 43.35 43.78 11.69
Respiratory Sinus Arrhythmia Recovery 298 0.1 0.09 0.06
Skin Conductance Level Recovery 296 19.78 19.35 8.74
Child Depression Inventory 282 8.42 7.99 6.45
Multidimensional Anxiety Scale for Children 285 38.67 37.72 19.01

Table 2. continued
Descriptive Statistics for Continuous Variables
Variable Minimum Maximum Skewness Kurtosis
Security Scale Mother 21 60 -0.87 0.37
Security Scale Father 10 60 -0.5 -0.72
Respiratory Sinus Arrhythmia Recovery 0 0.29 1.34 1.58
Skin Conductance Level Recovery 3.68 49.12 0.56 0.47
Child Depression Inventory 0 35 0.97 0.72
Multidimensional Anxiety Scale for Children 0 102 0.86 0.84

37

Table 3.
Intercorrelations Among Final Variables with Covariances in the Diagonal

1 2 3
1
Maltreatment Group 0.24
2 Security Scale Mother
-0.17** 71.76
3 Security Scale Father
-0.14** 0.35** 136.87
4 Respiratory Sinus Arrhythmia Recovery
0.04 0.01 0
5 Skin Conductance Level Recovery
-0.06 -0.02 0.16**
6 Child Depression Inventory
0.14** -0.33** -0.23**
7 Multidimensional Anxiety Scale for Children
0.08 -0.07 -0.06
Note. **p<.01

Table 3. continued
Intercorrelations Among Final Variables with Covariances in the Diagonal

4 5 6 7
1
Maltreatment Group
2 Security Scale Mother

3 Security Scale Father

4 Respiratory Sinus Arrhythmia Recovery
0
5 Skin Conductance Level Recovery
-0.01 76.33
6 Child Depression Inventory
0.06 -0.07 41.71
7 Multidimensional Anxiety Scale for Children
-0.02 -0.1 0.23** 348.7
Note. **p<.01

38

tolerance statistics. As shown in Table 3, none of the variables indicated evidence of
multicollinearity.
Power Analysis. A power analysis for the most stringent model (the mediated-
moderation model) was calculated. Using a hypothesis testing framework for root mean
square error approximation (RMSEA) allows for a statistical procedure for calculating
power and a minimum sample size for covariance structure modeling (MacCallum,
Browne, & Sugawara, 1996). Thus, a confidence interval for testing a model’s closeness
of fit, rather than a potentially misleading point estimate of exact fit, was used for the
current analysis. MacCallum et al. provide support for testing the hypothesis of “not close
fit” as opposed to the hypothesis of “close fit”, showing that the former allows for
stronger support of the alternative hypothesis if the null hypothesis is rejected. With
MacCallum et al’s recommendation of .05 as the cutoff for a close fit model, the null
hypothesis is defined as H
o
: ≥.05 and the alternative hypothesis as H
a
: <.05. Testing this
hypothesis with 270 degrees of freedom, 300 participants and α= .05, yields power of .99
for the current study. Thus, power for the most stringent model is very good; indicating
that all other tests of nested models (which all have greater degrees of freedom) will be
equal or greater in power.
Measurement Models. Before testing the main hypotheses with a full structural
model, the measurement models for each of the latent variables were first examined.
Assuring that each construct is well indicated by its manifest variables allows for an
appropriate test of the structural model (Byrne, 2010).
39

Figure 5. Measurement model for mother and father attachment as measured by the Security
Scale. Standardized regression weights are shown. All factor loadings were significant at p<.01.
Dad attach= father attachment; mom attach= mother attachment. DadAttach_P1 and
MomAttach_P1= first parcel of items from the Security Scale (i.e. items 1-3). DadAttach_P2 and
MomAttach_P2= second parcel (items 4-6). DadAttach_P3 and MomAttach_P3= third parcel
(items 7-9). DadAttach_P4 and MomAttach_P4= fourth parcel (items 10-12). DadAttach_P5 and
MomAttach_P5= fifth parcel (items 13-15).

χ
2
=74, DF=26, CFI=.96, RMSEA=.07, CI=.05-.09, PCLOSE=.03
40

Attachment. Attachment security was tested with two latent variables, one each
for mother and father as supported by current theory (Kerns, Klepac, & Cole, 1996). The
15 items were represented by five continuous manifest variables (parcels with three items
each) for both mother and father as shown in Figure 5. Parcels were assigned items
randomly- that is, no a priori test was run to combine items that were highly related into
parcels. The initial results indicated a poor fit (χ
2
=531.43, DF=34, CFI=.73,
RMSEA=.20, CI=.19-.22, PCLOSE=.00) and after investigating the factor loadings, it
was discovered that parcel 1 (i.e. items 1-3) for the father security scale had a poor
loading. Upon removing this variable and re-running the model, a more adequate fit was
obtained (χ
2
=74, DF=26, CFI=.96, RMSEA=.07, CI=.05-.09, PCLOSE=.03).
Physiological indicators. There were several possibilities for investigating
respiratory sinus arrhythmia (RSA) and skin conductance level (SCL) since baseline,
responsivity and recovery have all been examined in the literature using various
paradigms. Thus all three types of indicators were examined for the best fit to the current
data.
Baseline PNS and SNS were first examined. PNS was indicated by two manifest
variables, each indicating a minute of baseline respiratory sinus arrhythmia. Similarly,
SNS was also indicated by two manifest variables, each representing a minute of baseline
skin conductance level. The resulting model fit was good (χ
2
=8.77, DF=9, CFI=1.00,
RMSEA=.00, CI=.00-.06, PCLOSE=.90).
Conflict PNS and SNS were next examined. The latent variable, PNS, was
indicated by four manifest variables representing average respiratory sinus arrhythmia
41

responsivity during each of the four conflict clips (clips varied in length from 71 to 191
seconds each). Similarly, the latent variable, SNS, was indicated by four manifest
variables representing average skin conductance level responsivity during the same four
conflict clips. To account for baseline levels, minute two of baseline was controlled for.
The responsivity model fit poorly (χ
2
=167.98, DF=35, CFI=.97, RMSEA=.10, CI=.09-
.12, PCLOSE=.00) even after adjusting for RSA1 and RSA4, which had low factor
loadings.
Next, the measurement model for recovery PNS and SNS was examined. PNS and
SNS were each indicated by two manifest variables which represented recovery
respiratory sinus arrhythmia and skin conductance level, respectively, during each minute
of recovery. To account for any differences at baseline, the second minute of RSA and
SCL baseline was controlled for. The resulting model fit well (χ
2
=7.50, DF=9, CFI=1.00,
RMSEA=.00, CI=.00-.05, PCLOSE=.95).
It was concluded that both the baseline and recovery models were good, whereas
the conflict model was not. However, to determine whether some combination of
baseline, conflict, and recovery would fit better, each of these six model combinations
were also run; however none were better than the baseline or recovery models. Although
the model fit and beta weights were virtually equivalent between the two models, it was
decided that the recovery model would be the most interesting and novel model to use for
the current study (see Figure 6).
42

Figure 6. Measurement model for parasympathetic and sympathetic nervous systems as measured
by respiratory sinus arrhythmia recovery (RSAREC) and skin conductance level recovery
(SCLREC). Standardized regression weights are shown. All factor loadings were significant at
p<.01. RSABASE2=RSA baseline minute two. SCLBASE2=SCL baseline minute 2. PNS=
parasympathetic nervous system; SNS= sympathetic nervous system. RSAREC1 and SCLREC1=
first minute of recovery. RSAREC2 and SCLREC2= second minute of recovery.

χ
2
=7.50, DF=9, CFI=1.00, RMSEA=.00, CI=.00-.05, PCLOSE=.95
43

Figure 7. Measurement model for anxiety and depression as measured by the Child Depression
Inventory (CDI) and Multidimensional Anxiety Scale for Children (MASC). Standardized
regression weights are shown. All factor loadings were significant at p<.01. CDI_NegMood=
negative mood subscale of the CDI (items 1,6,8,10,11,13). CDI_Anhedonia= anhedonia subscale
of CDI (items 4,16,17,18,19,20,21,22). CDI_NegSelfEst= negative self esteem subscale of CDI
(items 2,7,9,14,25). MASC_Physical= tense (items 1,8,15,20,27,35) and somatic (items 5,12,18,
24,31,38) physical symptoms subscale of the MASC. MASC_SocAnx= humiliation fears (items
3,10,16,22,29) and performance fears (items 14,33,37,39) social anxiety subscale of the MASC.
MASC_SepPan= the separation/panic subscale of the MASC (items 4,7,9,17,19,23, 26,30,34).
χ
2
=26.8, DF=8, CFI=.97, RMSEA=.08, CI=.05-.11, PCLOSE=.06
44

Internalizing problems. Due to the fact that a good deal of literature on
internalizing problems has found that depression and anxiety disorders are highly
comorbid (Hirschfeld, 2001; Kessler et al., 2008) and are in fact significantly correlated
(r=.35, p<.01) in the current study, the two variables were run together in one model.
The latent construct of depression was initially indicated by five manifest variables,
which represented the five subscales of the Child Depression Inventory (CDI). Similarly,
anxiety was indicated by four manifest variables, which represented the subscales of the
Multidimensional Anxiety Scale for Children (MASC). The resulting model fit
adequately (χ
2
=26.8, DF=8, CFI=.97, RMSEA=.08, CI=.05-.11, PCLOSE=.06) but only
after two subscales were removed from depression and one subscale was removed from
anxiety due to poor model fit and low factor loadings (see Figure 7).
Invariance Models. To examine whether each measurement model fit across
both maltreatment and comparison groups, and males and females, factorial invariance
(i.e. measures are measuring the same construct in each group) was tested with Amos 18.
The four groups were: comparison males, comparison females, maltreated males, and
maltreated females. Evaluation of invariance consists of examining the chi-square
difference and probability value derived from the unconstrained model (which allows all
factor loadings between groups to freely vary) and the constrained model (which forces
all factor loadings between groups to be equal). A probability value that is non-significant
indicates that there are no significant differences between the constrained and
unconstrained models and thus supports measurement invariance.
45

Table 4.
T-tests for Males vs. Females

Males Females
p t df N Mean SD N Mean SD

Security Scale Mother 0.12 -1.57 357 186 47.26 7.88 173 48.67 9.03
Security Scale Father 0.01 2.54 315 162 44.96 11.05 155 41.64 12.15
CDI 0.16 -1.40 280 140 7.88 5.94 142 8.96 6.91
MASC 0.00 -2.82 283 142 35.53 18.52 143 41.80 19.05

Table 5.
Post-hoc Results for Age using LSD

Means and Standard Deviations
N 9-11 years N 11-13 years N 13-15 years

Security Scale Mother 77 49.23 (7.87) 198 47.57 (8.73) 81 47.58 (8.47)
Security Scale Father 69 46.33 (10.76)
a
175 43.40 (11.59) 70 40.61 (12.30)
a

RSA Recovery* 61 .09(.01) 166 .10(.01) 69 .09(.01)
SCL Recovery* 66 20.43(.49) 159 19.90(.32) 68 18.82(.49)
CDI 63 8.33 (6.71) 159 8.17 (6.31) 57 9.33 (6.77)
MASC 63 39.40 (22.00) 162 37.59 (18.30) 57 40.79 (18.05)
Note. Groups with the same superscripts (a, b) are significantly different at p<.01.
* Denotes ANCOVA results reported with adjusted means and standard errors.

46

First, invariance of attachment security was tested among the four groups using
the final measurement model indicated above (see Figure 5). Results indicated a ∆χ
2
=
17.36, DF=21 (p= .69), suggesting full factorial invariance of mother and father
attachment security across all four groups. Next, ANS was tested for invariance, using the
final model shown in Figure 6. Results for SNS and PNS showed that ∆χ
2
= 8.07, DF=6
(p= .23), suggesting full factorial invariance for SCL and RSA across the four groups.
Lastly, invariance of internalizing problems was tested using the final measurement
model indicated in Figure 7. Results indicated a ∆χ
2
= 17.46, DF=12 (p= .13), suggesting
full factorial invariance of anxiety and depression across all groups.
Covariates. To investigate demographic variables that may moderate the
relationships between attachment, physiology and internalizing problems, comparisons
were investigated among gender, age, ethnicity, and annual household income.
Gender. Comparisons between males and females on attachment security,
physiology, and internalizing problems were explored. Independent samples t-tests
suggested that males reported significantly (p<.01) higher attachment security with
fathers than females (see Table 4). Females however, reported significantly (p<.01)
higher anxiety than males. Univariate ANCOVA with RSA and SCL baseline as
covariates, respectively, indicated gender differences on SCL recovery (F=10.93(296),
DF=1, p<.01). Specifically, males (M
adjusted
=20.50, SD=.32) had significantly higher SCL
than females (M
adjusted
=18.96, SD=.34).
Age. Next, comparisons among age groups were examined for potential
differences on attachment security, physiology, and internalizing problems. Age at T2
47

was broken into three age groups: 1) 9-11 years, 2) 11-13 years and 3) 13-15 years. The
results from a one-way ANOVA suggested significant group differences for father
attachment security (F=3.44(317), DF=3, p<.05). Specifically, 9-11 year-olds reported
significantly (p<.01) higher attachment security with father compared with 13-15 year-
olds (see Table 5).
Ethnicity. Group differences among ethnicity were examined next. Results from a
one-way ANOVA showed significant differences in father attachment security among
African American and Hispanic participants (F=3.65(315), DF=3, p<.01). Interestingly,
African American children had the lowest mean scores whereas Latino children had the
highest mean scores on father attachment security among all ethnicities (see Table 6).
The second finding among ethnicity was for skin conductance level (SCL; F=7.67(296),
DF=3, p<.01). The most striking difference on SCL was between African American and
Latino children, with the former scoring the lowest and the latter scoring the highest.
White and Biracial participants scored somewhere in between, however the results
indicate that Latino participants were significantly higher on SCL than all other ethnic
groups (see Table 6).
Annual Household Income. Finally, comparisons of annual household income
were examined (see Table 7). Six annual household income brackets were defined: 1) $0-
15k, 2) $15-30k, 3) $30-45k, 4) $45-60k, 5) $60-75k, and 6) over $75k. The results from
a one-way ANOVA indicated significant group differences on father attachment security
(F=2.28(316), DF=6, p<.05). Specifically, the $15-30k and $30-45k groups were
significantly lower in father attachment security compared with the $45-60k group
48

Table 6.
Post-hoc Results for Ethnicity using LSD

Means and Standard Deviations
N African American N Latino
Security Scale Mother 129 48.82 (7.90) 146 47.79 (8.21)
Security Scale Father 109 40.94 (11.00)
a
132 45.50 (11.58)
a

RSA Recovery* 108 .10 (.01) 122 .09 (.01)
SCL Recovery* 107 19.01 (.39)
a
120 21.11 (.36)
abc

CDI 112 9.03 (7.04) 103 8.35 (5.99)
MASC 112 37.76 (20.81) 105 40.65 (17.61)
Note. Groups with the same superscripts (a,b,c) are significantly different at p<.05.
* Denotes ANCOVA results reported with adjusted means and standard errors.

Table 6. continued
Post-hoc Results for Ethnicity using LSD

Means and Standard Deviations
N Biracial N White
Security Scale Mother 43 47.18 (9.42) 41 46.49 (9.97)
Security Scale Father 38 41.37 (12.52) 38 44.68 (11.99)
RSA Recovery* 38 .09 (.01) 32 .10 (.01)
SCL Recovery* 35 18.43 (.66)
b
34 18.91 (.67)
c

CDI 39 8.02 (6.50) 28 6.82 (5.58)
MASC 40 38.72 (20.15) 28 34.86 (14.44)
Note. Groups with the same superscripts (a,b,c) are significantly different at p<.05.
* Denotes ANCOVA results reported with adjusted means and standard errors.

49

Table 7.
Post-hoc Results for Annual Household Income using LSD

Means and Standard Deviations
N $0-15K N $15-30K N $30-45K
Security Scale Mother 89 48.43 (8.41) 111 46.96 (8.73) 65 48.47 (8.51)
Security Scale Father 75 43.29 (12.47) 94 41.52 (12.05)
ab
59 41.39 (11.26)
cd

RSA Recovery* 75 .10(.01)a 93 .09(.01) 58 .10(.01)b
SCL Recovery* 72 19.84 (.48) 91 19.59 (.43) 52 20.34 (.56)
CDI 62 9.13 (7.20) 90 9.50 (6.66) 55 8.09 (6.16)
MASC 63 40.33 (22.13) 90 39.12 (17.72) 56 36.55 (19.11)
Note. Groups with the same superscripts (a,b,c,d) are significantly different at p<.05.
* Denotes ANCOVA results reported with adjusted means and standard errors.

Table 7. continued
Post-hoc Results for Annual Household Income using LSD

Means and Standard Deviations
N $45-60K N $60-75K N >$75K
Security Scale Mother 43 48.25 (8.70) 24 47.12 (8.61) 21 48.86 (7.98)
Security Scale Father 41 47.39 (10.01)
ac
23 43.48 (11.79) 21 47.81 (9.80)
bd

RSA Recovery* 36 .09(.01)c 18 .07(.01)abcd 13 .11(.01)d
SCL Recovery* 38 19.68 (.66) 18 20.63 (.96) 20 18.04 (.91)
CDI 36 7.53 (5.70) 16 5.62 (6.16) 19 6.10 (4.62)
MASC 36 41.53 (20.84) 17 32.29 (13.85) 19 38.31 (14.71)
Note. Groups with the same superscripts (a,b,c,d) are significantly different at p<.05.
* Denotes ANCOVA results reported with adjusted means and standard errors.
50

(p<.05) and over $75k group (p<.05). These results are difficult to interpret because there
is no logical pattern to the findings. Examination of the group means does not suggest
that father attachment security increases with income. For instance, there were no
significant findings among the lowest income group ($0-15k) or the second highest
income group ($60-75k).
Univariate ANCOVA results suggested that RSA was also significantly different
across income (F=3.31(316), DF=6, p<.01). Interestingly, participants in the $60-75k
group had significantly lower RSA compared to all other groups (see Table 7). In sum,
these group comparisons suggest that gender, age, ethnicity and annual household income
should be entered as covariates in the subsequent analyses.
Missing data imputation. Data were imputed only at the second order for the
Child Depression Inventory (CDI) and Multidimensional Anxiety Scale for Children
(MASC). First order missing data are those in which an adolescent did not respond to a
particular item on a scale (e.g. in the CDI or MASC); there were no such cases in the
current data. Second order missing data are those in which there were no responses for an
entire set of items (e.g. if an adolescent did not take the CDI). At T3, 20% of adolescents
did not return and therefore had no CDI and MASC data. Based on the suggestion of
Tabachnick & Fidell (2007) and previous research supporting missing data replacement
from a previous time point (Noll, Trickett, & Putnam, 2003) T2 CDI and MASC data
were investigated.
T2 and T3 CDI scores were highly correlated (r=.60, p<.01) as were T2 and T3
MASC scores (r=.40, p<.01). Independent samples t-tests were used to examine potential
51

differences between those who did and did not return at T3. There were no significant
differences between groups on attachment security, depression, anxiety, RSA or SCL.
Further, no demographic information (i.e. gender, age, ethnicity, annual household
income) discriminated those who did and did not return at T3. These results suggest that
attrition rates at T3 were not systematic or indicated by any particular qualities of the
participants. Next, T2 CDI replacement scores were compared against T3 CDI scores; no
significant differences (p=.45) emerged between the replacement scores and the T3 data.
Similarly, T2 MASC replacement scores were compared against T3 MASC scores, again
showing no significant differences (p=.44) between groups. Thus, T2 replacement scores
were used for participants who did not return at T3. It should be noted that these
replacement scores were only used for analyses in SPSS and not Amos, as Amos
estimates missing data using full information maximum likelihood (Kline, 2005; Loehlin,
2004).
Specific Aim 1
In order to address Specific Aim 1, maltreatment group comparisons were
investigated in order to describe potential differences among members of the current
sample. Maltreatment versus comparison groups were examined for differences among
attachment security, physiology, and internalizing symptoms. Results from univariate
ANCOVAs are shown in Table 8. In summary, comparison children had significantly
higher attachment security with both mother and father compared with maltreated
children (p<.05). There were no other significant group differences on physiology or
internalizing symptoms although group differences were marginally significant for
52

depression (p=.09) and anxiety (p=.07). Of note, is that after maltreatment group and
gender are accounted for, the only significant covariate among mother attachment, father
attachment and SCL is age; RSA and the internalizing variables have no other significant
covariates. Thus, only age will be modeled as a covariate in the subsequent multiple
group analyses. Maltreatment x gender differences on patterns of ANS responsivity was
also examined. Figures 8 and 9 show the patterns of responsivity from baseline to conflict
to recovery (17 minutes) for RSA and SCL in all four groups. While the differences
among groups appear much more varied for RSA, the differences in SCL among groups
is a bit more subtle.
Specific Aims 2-4
Structural equation modeling (SEM) has several benefits over other analytical
techniques such as regression. First, all variables of interest can be examined
simultaneously whereas multiple steps are needed to test the same hypotheses in
regression. Second, error is estimated for parameters in the model which are not
estimated with regression. Next, latent constructs are examined as indicated by manifest
variables which are either items indicating the model or based on a theoretical construct.
Another benefit is that models of mediation and moderation are tested more reliably in
SEM than with regression (Hopwood, 2007).
Because of these benefits, SEM was the preferred method of analysis in the
proposed study. Amos Statistical Program, version 18, was used to test the hypothesized
models and all values are reported as standardized values. Further, the variance of all

53

Table 8.

Univariate ANCOVA results for Comparison vs. Maltreated Adolescents with Adjusted Means and Standard Errors

p F DF Covariates N Comparison N Maltreated

Security Scale Mother 0 13.24 1 2 139 50.01 (.72) 217 46.60 (.57)

Security Scale Father 0.02 5.37 1 1, 2 133 45.12 (1.00) 183 42.04 (.84)

RSA Recovery 0.96 0.33 1 99 .09(.01) 199 .10(.01)

SCL Recovery 0.99 0 1 1, 2 118 19.75 (.37) 176 19.75 (.37)

CDI 0.07 3.18 1 122 7.65 (.59) 158 9.07 (.52)

MASC 0.11 2.5 1 1 124 36.69 (1.73) 159 40.40 (1.52)

Note. For covariates, 1= gender, 2= age, 3= ethnicity, 4= annual household income

54

Figure 8. RSA baseline, conflict and recovery for all four groups. RSABASE1 and
RSABASE2 = average baseline RSA during minutes 1 and 2 respectively. RSACON1,
RSACON2, RSACON3, and RSACON4= average RSA for clips 1-4, respectively.
RSAREC1 and RSAREC2= average recovery RSA during minutes 1 and 2 respectively.

55

Figure 9. SCL baseline, conflict and recovery for all four groups. SCLBASE1 and
SCLBASE2= average baseline SCL during minutes 1 and 2 respectively. SCLCON1,
SCLCON2, SCLCON3, and SCLCON4= average SCL for clips 1-4, respectively.
SCLREC1 and SCLREC2= average recovery SCL during minutes 1 and 2 respectively.

56

exogenous and endogenous variables as well as the residuals of endogenous variables
was estimated for all models.
The suggestions by MacCallum and colleagues were used as a guideline for
considering the fit statistics for each model (MacCallum, Browne, & Sugawara, 1996).
Specifically, a RMSEA fit statistic of <.05 was considered a close fit, whereas a fit of
between .05 and.08 was considered adequate; .08-.10 mediocre, and >.10 poor. In
conjunction with RMSEA it is important to look also at the confidence interval (CI) and
probability of the value fitting within the population (PCLOSE), and thus these statistics
were also examined (Byrne, 2010). Other commonly reported fit statistics, including chi-
square and degrees of freedom (χ
2
, DF) and the close fit index (CFI), were also
examined.
Furthermore, Byrne (2010) suggests that the statistical significance of each
parameter be evaluated by the critical ratio (CR) test. Upon examination of the CR for
each parameter a value greater than ±1.96 allows for the rejection of the null hypothesis
(i.e. nonsignificance). If a parameter does not meet the CR of > ±1.96, Byrne suggests
deleting this path and rerunning the model for greatest parsimony. This advice will be
followed when evaluating the subsequent models.
To examine Specific Aim 2, two hypotheses were tested:
Hypothesis 1: Attachment security will predict internalizing problems. Specifically, low
attachment security will predict high levels of anxiety and depression; maltreatment
group may moderate these relationships.
57

Hypothesis 2: ANS responsivity will predict internalizing problems. Specifically, low
RSA and high SCL will predict high levels of both anxiety and depression; maltreatment
group may moderate these relationships.
Testing Hypothesis 1. To test the impact of father and mother attachment
security on anxiety and depression, a multiple group model was examined in Amos 18
(see Figure 10). The paths from dad attachment to anxiety and depression as well as the
paths from mom attachment to anxiety and depression were constrained across all four
groups (comparison males, comparison females, maltreated males and maltreated
females) and age was entered as a covariate. Results indicated a good overall model fit
(χ
2
=714.12, DF=431, CFI=.86, RMSEA=.04, CI=.04-.05, PCLOSE=.98) and that the
structural weights were not significantly different across the four groups (∆χ
2
=4.99,
DF=6, p=.54). This suggests that neither maltreatment group nor gender moderate the
effect of attachment of internalizing problems. Examination of the regression weights
indicated that only mother attachment security was significant in predicting level of
depression (β=-.36, p<.01); neither mother or father attachment security predicted
anxiety.
Testing Hypothesis 2. To test the hypothesis that PNS and SNS responsivity
predicts anxiety and depression, a multiple group model was run in Amos 18. The paths
from PNS and SNS to anxiety and depression were constrained across all four groups;
age was entered as a covariate (Figure 11). Although the model fit well, (χ
2
=414.90,
DF=270, CFI=.95, RMSEA=.04, CI=.03-.04, PCLOSE=.99) and was not significantly

58

Figure 10. Father and mother attachment predicts internalizing problems. Values indicate standardized
regression weights for comparison males. All parameters were significant at p<.01 except those indicated
ns, which were not significant. Dad attach= father attachment; mom attach= mother attachment.
DadAttach_P1 and MomAttach_P1= first parcel of items from the Security Scale (i.e. items 1-3).
DadAttach_P2 and MomAttach_P2= second parcel (items 4-6). DadAttach_P3 and MomAttach_P3= third
parcel (items 7-9). DadAttach_P4 and MomAttach_P4= fourth parcel (items 10-12). DadAttach_P5 and
MomAttach_P5= fifth parcel (items 13-15). CDI_NegMood= negative mood subscale of the CDI (items
1,6,8,10,11,13). CDI_Anhedonia= anhedonia subscale of CDI (items 4,16,17,18,19,20,21,22).
CDI_NegSelfEst= negative self esteem subscale of CDI (items 2,7,9,14,25). MASC_Physical= tense (items
1,8,15, 20,27,35) and somatic (items 5,12,18,24,31, 38) physical symptoms subscale of the MASC.
MASC_SocAnx= humiliation fears (items 3,10,16,22,29) and performance fears (items 14,33, 37,39) social
anxiety subscale of the MASC. MASC_SepPan= the separation/panic subscale of the MASC (items
4,7,9,17,19,23,26,30,34).
ns
ns
ns ns
ns
ns
χ
2
=714.12, DF=431, CFI=.86, RMSEA=.04, CI=.04-.05, PCLOSE=.98
59

Figure 11. SNS and PNS predicts internalizing problems. Values indicate standardized regression weights
for comparison females. All parameters were significant at p<.01 except those indicated ns, which were not
significant. SNS=sympathetic nervous system. PNS=parasympathetic nervous system. RSABASE2 and
SCLBASE2=baseline minute 2. RSAREC1 and SCLREC1= first minute of recovery. RSAREC2 and
SCLREC2= second minute of recovery. CDI_NegMood= negative mood subscale of the CDI (items
1,6,8,10,11,13). CDI_Anhedonia= anhedonia subscale of CDI (items 4,16,17,18,19,20, 21,22).
CDI_NegSelfEst= negative self esteem subscale of CDI (items 2,7,9,14,25). MASC_Physical= tense (items
1,8,15, 20,27,35) and somatic (items 5,12,18,24,31, 38) physical symptoms subscale of the MASC.
MASC_SocAnx= humiliation fears (items 3,10,16,22,29) and performance fears (items 14,33, 37,39) social
anxiety subscale of the MASC. MASC_SepPan= the separation/panic subscale of the MASC (items
4,7,9,17,19,23,26,30, 34).
ns ns
ns ns
ns
ns
χ
2
=414.90, DF=270, CFI=.95, RMSEA=.04, CI=.03-.04, PCLOSE=.99
60

different across the four groups, neither PNS nor SNS were predictive of anxiety or
depression.
To examine Specific Aim 3, one hypothesis was tested:

Hypothesis 3: Attachment security will predict PNS and SNS responsivity. Specifically,
low attachment security will predict low RSA and high SCL responsivity; maltreatment
group may moderate these relationships.
Testing Hypothesis 3. To test the impact of father and mother attachment
security on SNS and PNS responsivity, a multiple group model was examined in Amos
18. Paths were constrained from both mother and father security to PNS and SNS,
respectively, and age was entered as a covariate (see Figure 12). Model fit results showed
that the structural weights were not significantly different across the four groups and that
the model fit well (χ
2
=722.86, DF=439, CFI=.92, RMSEA=.04, CI=.04-.05,
PCLOSE=.98). The only significant parameter however, was from father attachment to
SNS responsivity, which although significant was not meaningful (β=.08, p<.05).
To examine Specific Aim 4, one hypothesis was tested:

Hypothesis 4: ANS responsivity will mediate the association between attachment and
internalizing problems. Specifically, low attachment security will predict low RSA
responsivity and high SCL responsivity which in turn will predict high levels of anxiety
and depression; maltreatment group may moderate these relationships.
Testing Hypothesis 4. To test the hypothesis that PNS and SNS would mediate
the relationship between attachment security and internalizing problems, a multiple group
model was run in Amos 18 (see Figure 13). Based on the results of hypotheses 1-3 and
61

Figure 12. Attachment predicts ANS responsivity. Values indicate standardized regression weights for
Maltreated males. All parameters were significant at p<.01 except for paths indicated ns, which were not
significant. Dad attach= father attachment; mom attach= mother attachment. MomAttach_P1= first parcel
of items from the Security Scale (i.e. items 1-3). DadAttach_P2 and MomAttach_P2= second parcel (items
4-6). DadAttach_P3 and MomAttach_P3= third parcel (items 7-9). DadAttach_P4 and MomAttach_P4=
fourth parcel (items 10-12). DadAttach_P5 and MomAttach_P5= fifth parcel (items 13-15). SNS=
sympathetic nervous system. PNS= parasympathetic nervous system. RSABASE2 and SCLBASE2=
baseline minute 2. RSAREC1 and SCLREC1= first minute of recovery. RSAREC2 and SCLREC2= second
minute of recovery.
ns
ns ns
ns
ns
χ
2
=722.86, DF=439, CFI=.92, RMSEA=.04, CI=.04-.05, PCLOSE=.98
62

elimination of paths whose critical ratio was less than ±1.96, five paths were constrained
across the four groups: 1) mother attachment security to depression, 2) father attachment
security to anxiety, 3) father attachment security to SNS, 4) PNS to anxiety and 5) SNS to
anxiety. The results indicated that the structural weights were significantly different
across groups (∆χ
2
=326.16, DF=15, p=.00).
In order to investigate which parameters were significantly different across
groups, and for which groups, each parameter was constrained separately. The first
significant parameter was from mother attachment to depression. Results showed that
maltreated females were significantly different from comparison females (p<.01). More
specifically, the parameter was nonsignificant for comparison females whereas it was
significant for maltreated females (β=-.26, p<.05). This parameter was also
nonsignificant for comparison and maltreated males.
The second significant parameter was between father attachment and anxiety.
Results showed that comparison females were significantly different from both
comparison males (p<.05) and maltreated females (p<.01). Furthermore, maltreated
females were significantly (p<.05) different from maltreated males. In particular, the
parameter was significant for both comparison females (β=-.67, p<.01) and maltreated
males (β=-.67, p<.01) but not for comparison males and maltreated females.
The third significant parameter was between father attachment and SNS. In
particular, maltreated females were significantly different from comparison males (p<.05)
and comparison females (p<.01). The parameter was significant for both comparison
63

Figure 13. ANS mediates attachment and internalizing problems. Values indicate standardized regression
weights for maltreated males. All parameters were significant at p<.01. Mom attach= mother attachment;
Dad attach= father attachment. MomAttach_P1= first parcel of items from the Security Scale (i.e. items 1-
3). MomAttach_P2 and DadAttach_P2=second parcel (items 4-6). MomAttach_P3 and DadAttach_P3=
third parcel (items 7-9). MomAttach_P4 and DadAttach_P4=fourth parcel (items 10-12). MomAttach_P5
and DadAttach_P5= fifth parcel (items 13-15). PNS=parasympathetic nervous system; SNS=sympathetic
nervous system. RSABASE2 and SCLBASE2= second minute of baseline. RSAREC1 and SCLREC1=first
minute of recovery; RSAREC2 and SCLREC2=second minute of recovery; RSAREC3 and
SCLREC3=third minute of recovery. CDI_NegMood= negative mood subscale of the CDI (items 1,6,8,10,
11,13). CDI_Anhedonia= anhedonia subscale of CDI (items 4,16,17,18,19,20,21,22). CDI_NegSelfEst=
negative self -esteem subscale of CDI (items 2,7,9,14,25).
χ
2
=440.07, DF=329, CFI=.96, RMSEA=.03, CI=.03-.04, PCLOSE=1.00
64

males (β=.17, p<.01) and comparison females (β=.16, p<.01) whereas it was not
significant for maltreated males or maltreated females.
The fourth significant parameter was between SNS and anxiety. Specifically,
comparison females were significantly different than comparison males (p<.01),
maltreated males (p<.01) and maltreated females (p<.01). Additionally, maltreated males
were significantly different from comparison males (p<.01) and maltreated females
(p<.01). Finally, maltreated females were significantly different than maltreated males
(p<.01). The parameter was significant for comparison females (β=.82, p<.01) and
maltreated males (β=-.41, p<.05), whereas it was not significant for comparison males
and maltreated females. Of note is that the parameter was significant for comparison
females and maltreated males in opposite directions.
Finally, the fifth significant parameter was between PNS and anxiety. Maltreated
males were significantly different from comparison males (p<.01), comparison females
(p<.01) and maltreated females (p<.01). Additionally, comparison males were
significantly different from comparison females (p<.01) and maltreated females (p<.01).
In particular, comparison males (β=.98, p<.01) and maltreated females (β=-.62, p<.05)
were significant on this parameter whereas comparison females and maltreated males
were not. Interestingly comparison males and maltreated females were significant in
opposite directions. In sum, these results suggest moderation of group and gender on the
relationships between attachment, ANS, and internalizing problems.
In examining the conceptual models (see Figure 14) there is a potential mediation
of SNS between father attachment and anxiety for comparison females. The testing of
65

Comparison Males (1) Comparison Females (2)

Maltreated Males (3) Maltreated Females (4)

Figure 14. Conceptual models for each of the four groups with significant parameters. Values
indicate standardized beta weights. Dad attach= father attachment security as indicated by the
Security Scale. Mom attach= mother attachment security as indicated by the Security Scale.
SNS= sympathetic nervous system responsivity as measured by skin conductance level during
three minutes of rest. PNS= parasympathetic nervous system responsivity as measured by
respiratory sinus arrhythmia during three minutes of rest. Anxiety= anxiety problems as measured
by the Multidimensional Anxiety Scale for Children. Depression= depressive problems as
measured by the Child Depression Inventory.

Dad Attach
Dad Attach
Dad Attach
SNS
SNS
SNS
PNS
PNS Mom Attach
Anxiety Anxiety
Anxiety
Anxiety Depression
.17
.98
.16
.82
-.62 -.26
-.41
-.67
-.67
66

mediation in an SEM framework is different from regression, and thus does not follow
the same rules as those laid out by Baron & Kenny (1986) for regression. The main
difference is that in SEM the predictor variable need not predict the outcome variable, as
there is often not enough power to detect the relationship (MacKinnon, Lockwood,
Hoffman, West, & Sheets, 2002; Shrout & Bolger, 2002). As with regression however,
using Sobel’s test (1982) is a requirement for supporting mediation. Sobel’s test uses the
equation: z-value = a*b/SQRT(b
2
*s
a
2
+ a
2
*s
b
2
), where a is the unstandardized regression
coefficient between the IV and the mediator; s
a
is the standard error of a; b is the
unstandardized regression coefficient between the mediator and the DV; and s
b
is the
standard error of b. A significant value- indicating mediation- is a test statistic of greater
than ±1.96 with p<.05. To test whether mediation exists between father attachment, SNS
and anxiety, the following values were entered: a=.58, s
a
=.19, b=.32, s
b
=.08. The results
indicated a significant (z=2.43, SE=.07 p<.05) mediation of SNS between father
attachment and anxiety for comparison females.

67

Chapter Four: Discussion
There were four specific aims of the current study: 1) To describe the differences
among maltreated and comparison groups on attachment security, ANS responsivity, and
internalizing problems; 2) To examine attachment security and ANS responsivity as
predictors of internalizing problems; 3) To examine attachment security as a predictor of
ANS responsivity; and 4) To examine ANS responsivity as a mediator of attachment
security and internalizing problems. Table 9 provides a summary of the results.
Attachment Security. Taken together, these results suggest several things. First
and foremost is that mother and father attachment have different developmental
pathways. In many previous studies a measure of father attachment has been lacking and
the current research points to some interesting differences between mother and father
attachment. In the current study results showed that the attachment relationship with
mother predicts depressive symptoms, whereas the attachment relationship with father
predicts anxiety symptoms. These results indicate that: 1) the greater the security with
mother, the lower the depressive symptoms reported two years later and 2) the greater the
security with father, the lower the anxiety symptoms reported two years later. In
particular, the parameter from mother attachment to depression was significant for
maltreated females whereas the parameter from father attachment to anxiety was
significant for both comparison females and maltreated males.
One theory for the mother attachment and depression finding is that the
attachment relationship with mother may be particularly important for maltreated
females, who perhaps experience maltreatment differently than maltreated males.
68

Table 9.
Summary of Results

Confirmed Significant Parameters
Hyothesis 1:
Attachment predicts internalizing Yes 1) Mother attachment to depression
Maltreatment moderates No
Gender moderates No
Hypothesis 2:
ANS predicts internalizing Yes 1) SNS to anxiety, 2) PNS to anxiety
Maltreatment moderates Yes
Gender moderates Yes
Hypothesis 3:
Attachment predicts ANS Yes 1) Father attachment to SNS
Maltreatment moderates Yes
Gender moderates Yes
Hypothesis 4:
ANS mediates attachment &
internalizing Yes Comparison Females only
Maltreatment moderates Yes
Gender moderates Yes

69

This finding could also indicate that females identify more with or depend more on their
mothers compared with males, such that the relationship with mother impacts their
emotional functioning. In regard to comparison adolescents, it may be that without the
developmental disruptions of maltreatment, peers rather than parents are more important
during adolescence and thus relationships with peers would be more indicative of
depressive symptoms than the relationship with mother. This makes sense since on whole
comparison adolescents reported greater security with mother and father than maltreated
adolescents, suggesting that security with parents is not a great concern.
Regarding the findings for father attachment and anxiety, these results are a little
more interesting. The first question that comes to mind is how comparison females and
maltreated males are alike? Specifically, what factors make the relationship with father
predictive of anxiety symptoms in these seemingly disparate groups? It could be that
attachment with father predicts anxiety for two distinct reasons and that these groups are
not similar. One explanation could be that for females who do not experience
maltreatment, the relationship with father is important for emotional development during
the teenage years, more so than for males. For maltreated males, it may be important, but
for different reasons. Because more fathers were absent in the maltreated group and
fathers are more likely to be the abuser, it could be that those maltreated boys who do
have a good relationship with their father are protected from anxiety symptoms. These of
course are just a couple of possibilities and should be further tested.
The next important finding regarding attachment was that father attachment
predicts SNS responsivity, whereas mother attachment does not; and neither predicts PNS
70

responsivity. The results showed specifically that the higher the attachment security with
father the higher the SNS responsivity during recovery. These results are very interesting
because they suggest that fathers impact sympathetic functioning whereas mothers do
not, at least for comparison adolescents. However, because maltreated males were not
significantly different from comparison males and females, it is unclear whether this
relationship may also apply to comparison males. It is certain however, based on the
current findings, that maltreated females are distinctly different from comparison males
and females. In examining Figure 9 it is clear that maltreated females have a much lower
SCL responsivity throughout the video clips. In sum, these results suggest that the
adaptive response during recovery is to continue in a state of high sympathetic arousal.
Thus in the presence of maltreatment (especially for females) there may be other, more
important predictors of sympathetic functioning. These results clearly need further
investigation and should be examined in future research.
Autonomic Nervous System Responsivity. In addition to differing pathways for
mother and father attachment relationships, a second major finding of the current study is
the varying pathways of the PNS and SNS among maltreatment and gender groups.
Interestingly, neither PNS nor SNS indicators predicted depressive symptoms in any of
the four groups. However both PNS and SNS indicators predicted anxiety symptoms.
In particular, PNS responsivity predicted anxiety symptoms for comparison males
and maltreated females, but in opposite directions. For comparison males, higher RSA
responsivity during recovery (indicating activation of the parasympathetic nervous
system) predicted higher reported anxiety two years later. However, for maltreated
71

females, lower RSA responsivity during recovery (indicating withdrawal of the
parasympathetic nervous system) predicted higher reported anxiety two years later. The
finding for maltreated females is what would be expected based on the literature
indicating that vagal withdrawal is maladaptive during times when activation would
assist in normal recovery. Based on the pattern of responsivity shown in Figure 8, the
finding for comparison males indicates that high activation of the PNS is maladaptive
during recovery.
In regard to the parameter from SNS to anxiety, this relationship was significant
for both comparison females and maltreated males, but in opposite directions. Thus for
comparison females the higher their SNS responsivity the higher their reported anxiety;
for maltreated males, the lower their SNS responsivity the higher their reported anxiety.
These findings are fascinating, suggesting that comparison females and maltreated males
have unique pathways from sympathetic arousal to anxiety. A univariate ANCOVA
showed that comparison females (M
adjust
=19.40) were significantly (F=5.08(296), DF=3,
p<.01) lower on SCL responsivity during recovery compared to maltreated males
(M
adjust
=19.40). Based on these results, it seems that it is adaptive for comparison females
to activate their PNS during recovery and simultaneously decrease activity of the SNS.
Thus, comparison females who are still very high on SCL responsivity during recovery
will show a maladaptive response (i.e. a more anxious individual). For maltreated males,
who are higher than females in SCL across both groups, seem to have a different norm.
Because it is normal for them to have a higher SCL, a lower SCL is related to anxiety
72

symptoms. These findings are fascinating and should be further investigated in future
research.
SNS Mediation. A third major finding in the current study was that SNS
responsivity, as indicated by skin conductance level during the recovery period, mediated
the relationship between father attachment security and anxiety problems. Interestingly,
this was only true for the comparison females. What this suggests is that attachment
security with fathers and the sympathetic nervous system play a unique role in predicting
anxiety problems for comparison females that is not evident in either males or maltreated
females. Due to the fact that anxiety was significantly predicted by either PNS or SNS
responsivity in all four groups, it is clear that the ANS is a predictor of later anxiety. It
may be that other qualities of early parent child relationships are a better predictor of
responsivity for comparison males and the maltreated group. Future research should
examine this possibility and continue to investigate ANS responsivity as a potential
mediating factor.
Limitations and Future Research
There are some limitations of the current study such as: 1) Attachment and ANS
indicators being collected at the same time point, 2) Lack of multi-method multi-
informant data for attachment, anxiety and depression (i.e. only self-report data) 3) RSA
and SCL missingness, and 4) High attrition from T2 to T3 (20%).
Although parent-child attachments are said to endure throughout the lifetime
(Bowlby, 1969) and follow mostly stable patterns (Waters, Merrick, Treboux, Crowell, &
Albersheim, 2000), earlier measures of parent attachment would be helpful in supporting
73

these claims. In particular, measures of attachment from infancy through adolescence
would be helpful in determining the stability of attachment in addition to documenting
when ANS regulation can be discriminated among attachment groups.
Another present limitation is the reliance on self-report data for attachment,
anxiety and depression. Teacher, parent reports and clinical diagnoses could be extremely
helpful in triangulating on the true values of these constructs. Similarly, multiple
measurements of physiological data could have improved the current study tremendously.
Due to equipment errors and confounding movements, a good amount of data was
unusable. Future research examining parent-child conflict should incorporate
physiological readings during real life conflicts and perhaps during other simulations of
conflict or emotionally salient interpersonal situations. Multiple physiological readings
could be useful, not only in assisting with missing data, but in creating a much more
powerful measure of physiology. Unlike the physiological measures, depression and
anxiety were measured at multiple time points and although data from T2 was used to
replace missing T3 data, it clearly would have been preferable to have all participants
return at T3.
Strengths
Despite these limitations, there were several strengths of the current study.
Attachment research has historically ignored the importance of the father-child
attachment relationship and it has only been relatively recently that some studies have
begun to include measures of father-child attachment. One strength of the current study
is the examination of father-child attachment in addition to mother-child attachment. In
74

examining these relationships separately it was possible to determine the unique
pathways of each relationship. This study found that mother-attachment security is an
important factor in predicting adolescent depression two years later whereas father-
attachment security is not. However, father-attachment was predictive of anxiety two
years later, whereas mother was not. Furthermore, father-attachment was predictive of
both PNS and SNS responsivity during recovery from conflict clips, whereas mother-
attachment was not.
Finally, an initial model examining the ANS as a mediator of attachment and
internalizing problems was investigated. This study provided some initial evidence
suggesting that SNS mediates the relationship between father attachment-security and
anxiety, which has not previously been examined. Future research should continue to
investigate this link, as it may illuminate alternative methods for interventions with
adolescents who are at risk for developing anxiety problems.

75

References

Adam, K. S., Sheldon-Keller, A. E., & West, M. (1996). Attachment organization and
history of suicidal behavior in clinical adolescents. Journal of Consulting and
Clinical Psychology, 64(2), 264-272.
Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of
Attachment: A Psychological Study of The Strange Situation. Hillsdale, NJ:
Lawrence Erlbaum Associates, Publishers.
Allen, J. P., Moore, C., Kuperminc, G., & Bell, K. (1998). Attachment and adolescent
psychosocial functioning. Child Development, 69(5), 1406-1419.
Anan, R. M., & Barnett, D. (1999). Perceived social support mediates between prior
attachment and subsequent adjustment: A study of urban African American
children. Developmental Psychology, 35(5), 1210-1222.
Argyle, N. (1991). Skin conductance levels in panic disorder and depression. The Journal
of Nervous and Mental Disease, 179(9), 563-566.
Baker, R., Holloway, J., Thomas, P. W., Thomas, S., & Owens, M. (2004). Emotional
processing and panic. Behaviour Research and Therapy, 41, 1271-1287.
Bakker, M. J., Tijssen, M. A. J., van der Meer, J. N., Koelman, J. H. T. M., & Boer, F.
(2009). Increased whole-body auditory startle reflex and autonomic reactivity in
children with anxiety disorders. Journal of Psychiatry and Neuroscience, 34(4),
314-322.
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in
social psychological research: Conceptual, strategic, and statistical considerations.
Journal of Personality and Social Psychology, 51(6), 1173-1182.
Beauchaine, T. (2001). Vagal tone, development, and Gray’s motivational theory: toward
an integrated model of autonomic nervous system functioning in
psychopathology. Development and Psychopathology, 13, 183-214.
Beck, A. T., Steer, R. A., & Garbin, M. G. (1988). Psychometric properties of the Beck
Depression Inventory: Twenty-five years of evaluation. Clinical Psychology
Review, 8(1), 77-100.
Beesdo, K., Knappe, S., & Pine, D. S. (2009). Anxiety and anxiety disorders in children
and adolescents: Developmental issues and implications for DSM-V. Psychiatric
Clinics of North America, 32, 483-524.
76

Bolger, K. E., & Patterson, C. J. (2001). Developmental pathways from child
maltreatment to peer rejection. Child Development, 72(2), 549.
Bosquet, M., & Egeland, B. (2006). The development and maintenance of anxiety
symptoms from infancy through adolescence in a longitudinal sample.
Development & Psychopathology, 18, 517-550.
Bowlby, J. (1969). Attachment and Loss, Volume 1: Attachment (Vol. 1). New York:
Basic Books Inc. Publishers.
Bretherton, I., & Munholland, K. A. (1999). Internal working models in attachment
relationships: A construct revisited. In J. Cassidy & P. R. Shaver (Eds.),
Handbook of attachment: Theory, research, and clinical applications (pp. 89-
111). New York, NY: Guilford Press.
Brown, G. W., Craig, T. K. J., & Harris, T. O. (2008). Parental maltreatment and
proximal risk factors using the childhood experience of care and abuse
instrument: A life-course study of adult chronic depression. Journal of Affective
Disorders, 110, 222-233.
Buzi, R. S., Weinman, M. L., & Smith, P. B. (2007). The relationship between adolescent
depression and a history of sexual abuse. Adolescence, 42(168), 679-688.
Byrne, B. M. (2010). Structural equation modeling with AMOS (2nd ed.). New York:
Routledge.
Campbell-Sills, L., & Barlow, D. H. (2007). Incorporating emotion regulation into
conceptualizations and treatments of anxiety and mood disorders. In J. Gross
(Ed.), Handbook of Emotion Regulation (pp. 542-559). New York: The Guilford
Press.
Campbell-Sills, L., Barlow, D. H., Brown, T. A., & Hofmann, S. G. (2006). Effects of
suppression and acceptance on emotional responses of individuals with anxiety
and mood disorders. Behaviour Research and Therapy, 44, 1251-1263.
Carlson, E. A. (1998). A prospective longitudinal study of attachment
disorganization/disorientation. Child Development, 69(4), 1107-1128.
Carpenter, L. L., Tyrka, A. R., Ross, N. S., Khoury, L., Anderson, G. M., & Price, L. H.
(2009). Effect of childhood emotional abuse and age on cortisol responsivity in
adulthood. Biological Psychiatry, 66(1), 69-75.
Carrey, N. J., Butter, H. J., & Persinger, M. A. (1995). Physiological and cognitive
correlates of child abuse. Journal of the American Academy of Child and
Adolescent Psychiatry, 34(8), 1067-1075.
77

Cicchetti, D., Lynch, M., Shonk, S., & Manly, J. T. (1992). Family-peer relationships:
Modes of linkage. In R. D. Parke & G. W. Ladd (Eds.), (pp. 345-383). Hillsdale,
NJ: Lawrence Erlbaum Associates, Inc.
Cole, P. M., Zahn-Waxler, C., & Smith, K. D. (1994). Expressive control during a
disappointment: Variations related to preschoolers’ behavior problems.
Developmental Psychology, 30, 835-846.
Contreras, J. M., Kerns, K. A., Weimer, B. L., Gentzler, A. L., & Tomich, P. L. (2000).
Emotion regulation as a mediator of associations between mother-child
attachment and peer relationships in middle childhood. Journal of Family
Psychology, 14, 111-124.
Cooper, M. L., Shaver, P. R., & Collins, N. L. (1998). Attachment styles, emotion
regulation, and adjustment in adolescence. Journal of Personality and Social
Psychology, 74(5), 1380-1397.
Costa, B., Pini, S., Gabelloni, P., Abellie, M., Lari, L., Cardini, A., et al. (2009).
Oxytocin receptor polymorphisms and adult attachment style in patients with
depression. Psychoneuroendocrinology, 34, 1506-1514.
Cozolino, L. (2006). The neuroscience of human relationships: Attachment and the
developing social brain. New York: W.W. Norton and Company.
Crittenden, P. M. (1995). Attachment and Psychopathology. In S. Goldberg, R. Muir & J.
Kerr (Eds.), Attachment Theory: Social, Developmental, and Clinical
Perspectives (pp. 367-406).
Cyranowski, J. M., Bookwala, J., Feske, U., Houck, P., Pilkonis, P., Kostelnik, B., et al.
(2002). Journal of Social and Clinical Psychology, 21(2), 191-217.
Dale, L. P., Carroll, L. E., Galen, G., Hayes, J. A., Webb, K. W., & Porges, S. W. (2009).
Abuse history is related to autonomic regulation to mild exercise and
psychological well-being. Applied Psychophysiological Biofeedback, 34, 299-308.
Dawson, M. E., Schell, A. M., & Filion, D. L. (2000). The electrodermal system. In J. T.
Cacioppo, L. G. Tassinary & G. G. Berntson (Eds.), Handbook of
Psychophysiology (2nd ed ed., pp. 200-223). NY: Cambridge University Press.
Diamond, L. M., & Hicks, A. M. (2005). Attachment style, current relationship security,
and negative emotions: The mediating role of physiological regulation. Journal of
Social and Personal Relationships, 22(4), 499-518.
78

Diamond, L. M., Hicks, A. M., & Otter-Henderson, K. (2006). Physiological evidence for
repressive coping among avoidantly attached adults. Journal of Social and
Personal Relationships, 23(2), 205-229.
Dietrich, A., Riese, H., Sondeijker, F. E., Greaves-Lord, K., van Roon, A. M., Ormel, J.,
et al. (2007). Externalizing and internalizing problems in relation to autonomic
function: A population-based study in preadolescents. Journal of the American
Academy of Child and Adolescent Psychiatry, 46(3), 378-386.
Dozier, M., & Kobak, R. R. (1992). Psychophysiology in attachment interviews:
Converging evidence for deactivating strategies. Child Development, 63 (6),
1473-1480.
Egeland, B., & Sroufe, L. A. (1981). Attachment and early maltreatment. Child
Development, 52(1), 44-52.
Essau, C. A. (2004). The association between family factors and depressive disorders in
adolescents. Journal of Youth and Adolescence, 33(5), 365-372.
Feeney, B. C., & Kirkpatrick, L. A. (1996). Effects of adult attachment presence of
romantic partners on physiological response to stress. Journal of Personality and
Social Personality, 70(2), 255-270.
Feerick, M. M., Knutson, J. F., Trickett, P. K., & Flanzer, S. M. (Eds.). (2006). Child
Abuse and Neglect: Definitions, classifications, and a framework for research.
Baltimore: Paul H. Brookes Publishing Co.
Finnegan, V. E., Hodges, E., & Perry, D. G. (1996). Preoccupied and avoidant coping
during middle childhood. Child Development, 67, 1318-1328.
Finzi, R., Cohen, O., Sapir, Y., & Weizman, A. (2000). Attachment styles in maltreated
children: A comparative study. Child Psychiatry & Human Development, 31(2),
113-128.
Fury, G., Carlson, E. A., & Sroufe, L. A. (1997). Children's representations of attachment
relationships in family drawings. Child Development, 68(6), 1154-1164.
Gallo, L. C., & Matthews, K. A. (2006). Adolescents' attachment orientation influences
ambulatory blood pressure responses to everyday social interactions.
Psychosomatic Medicine, 68(2), 253-261.
Gilissen, R., Bakermans-Kranenberg, M. J., van IJzendoorn, M. H., & Linting, M.
(2008). Electrodermal reactivity during the trier social stress test for children:
Interaction between the serotonin transporter polymorphism and children's
attachment representation. Developmental Psychobiology, 615-625.
79

Gordis, E. B., Feres, N., Olezeski, C. L., Rabkin, A. N., & Trickett, P. K. (2009). Skin
conductance reactivity and respiratory sinus arrhythmia among maltreated and
comparison youth: relations with aggressive behavior. Journal of Pediatric
Psychology.
Gordis, E. B., Granger, D. A., Susman, E. J., & Trickett, P. K. (2007). Salivary alpha
amylase-cortisol asymmetry in maltreated youth. Hormones and Behavior.
Granot, D., & Mayseless, O. (2001). Attachment security and adjustment to school in
middle childhood. International Journal of Behavioral Development, 25, 530-541.
Gullone, E., Ollendick, T. H., & King, N. J. (2006). The role of attachment representation
in the relationship between depressive symptomatology and social withdrawal in
middle childhood.
Hastings, P. D., Nuselovici, J. N., Utendale, W. T., Coutya, J., McShane, K. E., &
Sullivan, C. (2008). Applying the polyvagal theory to children's emotion
regulation: Social context, socialization, and adjustment. Biological Psychology,
79(3), 299-306.
Hildrum, B., Mykletun, A., Holmen, J., & Dahl, A. A. (2008). Effect of anxiety and
depression on blood pressure: 11-year longitudinal population study. The British
Journal of Psychiatry, 193, 108-113.
Hill-Soderlund, A. L., Mills-Koonce, W. R., Propper, C., Calkins, S. D., Granger, D. A.,
Moore, G. A., et al. (2008). Parasympathetic and sympathetic responses to the
strange situation in infants and mothers from avoidant and securely attached
dyads. Developmental Psychobiology, 361-376.
Hinnant, J. B., & El-Sheikh, M. (2009). Children's externalizing and internalizing
symptoms over time: The role of individual differences in patterns of RSA
responding. Journal of Abnormal Child Psychology, 37, 1049-1061.
Hirschfeld, R. M. A. (2001). The comorbidity of major depression and anxiety disorders:
Recognition and management in primary care. Primary Care Companion to the
Journal of Clinical Psychiatry, 3, 244-254.
Hodges, K. (1990). Depression and anxiety in children: A comparison of self-report
questionnaires to clinical interview. Psychological Assessment, 2, 376-381.
Holland, A. S., & Roisman, G. I. (2010). Adult attachment security and young adults'
dating relationships over time: Self-reported, observational, and physiological
evidence. Developmental Psychology, 45(2), 552-557.
80

Hopwood, C. J. (2007). Moderation and mediation in structural equation modeling:
Applications for early intervention research. Journal of Early Intervention, 29(3),
262-272.
Howe, D. (2005). Child Abuse and Neglect: Attachment, Development and Intervention:
Palgrave Macmillan.
Hugdahl, K. (1995). Psychophysiology: The Mind-Body Perspective. Cambridge:
Harvard University Press.
Ialongo, N., McCreary, B. K., Pearson, J. L., Koenig, A. L., Schmidt, N. B., Poduska, J.,
et al. (2004). Major depressive disorder in a population of urban African-
American young adults: Prevalence, correlates, comorbidity and unmet mental
health service need. Journal of Affective Disorders, 79, 127-136.
Kaplan, S. J., Pelcovitz, D., Salzinger, S., Mandel, F., Weiner, M., & Labruna, V. (1999).
Adolescent physical abuse and risk for suicidal behaviors. Journal of
Interpersonal Violence, 14(9), 976.
Kerns, K. A., Abraham, M. M., Schlegelmilch, A., & Morgan, T. A. (2007). Mother-child
attachment in later middle childhood: Assessment approaches and associations
with mood and emotion regulation. Attachment and Human Development, 9(1),
33-53.
Kerns, K. A., Brumariu, L. E., & Abraham, M. M. (2008). Homesickness at summer
camp: Associations with the mother-child relationship, social self-concept, and
peer relationships in middle childhood. . Merrill-Palmer Quarterly, 54(4), 473-
498.
Kerns, K. A., Klepac, L., & Cole, A. (1996). Peer relationships and preadolescents’
perceptions of security in the child-mother relationship. Developmental
Psychology, 32(3), 457-466.
Kerns, K. A., Tomich, P. L., Aspelmeier, J. E., & Contreras, J. M. (2000). Attachment-
based assessments of parent-child relationships in middle childhood.
Developmental Psychology, 36.
Kerns, K. A., Tomich, P. L., & Kim, P. (2006). Normative trends in children's
perceptions of availability and utilization of attachment figures in middle
childhood. Social Development, 15(1), 1-22.
Kessler, R. C., Gruber, M., Hettema, J. M., Hwang, I., Sampson, N., & Yonkers, K. A.
(2008). Co-morbid major depression and generalized anxiety disorders in the
National Comorbidity Survey follow-up. Psychological Medicine, 38(3), 365-374.
81

Kessler, R. C., McGonagle, K. A., Swartz, M., Blazer, D. G., & Nelson, C. B. (1993).
Sex and depression in the National Comorbidity Survey I: Lifetime prevalence,
chronicity and recurrence. Journal of Affective Disorders, 29, 85-96.
Kim, J., & Cicchetti, D. (2006). Longitudinal trajectories of self-system processes and
depressive symptoms among maltreated and nonmaltreated children. Child
Development, 77(3), 624-639.
Kline, R. B. (2005). Principles and practice of structural equation modeling (20d ed.).
New York: The Guilford Press.
Kobak, R. R., & Sceery, A. (1988). Attachment in late adolescence: Working models,
affect regulation and representations of self and others. Child Development, 59,
135-146.
Kochanska, G. (2001). Emotional development in children with different attachment
histories: The first three years. Child Development, 72(2), 474.
Kovacs, M. (1985). The Children's Depression Inventory (CDI). Psychopharmacology
Bulletin, 21, 995-998.
Kovacs, M. (1992). Child Depression Inventory Manual. Toronto: Multi-Health System.
Levitt, J. T., Brown, T. A., Orsillo, S. M., & Barlow, D. H. (2004). The effects of
acceptance versus suppression of emotion on subjective and psychophysiological
response to carbon dioxide challenge in patients with panic disorder. Behavior
Therapy, 35, 747-766.
Lieberman, M., Doyle, A., & Markiewicz, D. (1999). Developmental patterns in security
of attachment to mother and father in late childhood and early adolescence:
Associates with peer relations. Child Development, 70(1), 202-213.
Linares, L. O., Stovall-McClough, K. C., Li, M., Morin, N., Silva, R., Albert, A., et al.
(2008). Salivary cortisol in foster children: A pilot study. Child Abuse & Neglect,
32, 665-670.
Liu, R. T., Alloy, L. B., Abramson, L. Y., Iacoviello, B. M., & Whitehouse, W. G.
(2009). Emotional maltreatment and depression: Prospective prediction of
depressive episodes. Depression and Anxiety, 26, 174-181.
Loehlin, J. C. (2004). Latent variable models (4th ed.). London: Lawrence Erlbaum
Associates.

82

Luecken, L. J., Appelhans, B. M., Kraft, A., & Brown, A. (2006). Never far from home:
A cognitive-affective model of the impact of early-life family relationships on
physiological stress responses in adulthood. Journal of Social and Personal
Relationships, 23(2), 189-203.
MacCallum, R. C., Browne, M. W., & Sugawara, H. M. (1996). Power analysis and
determination of sample size for covariance structure modeling. Psychological
Methods, 1(2), 130-149.
MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., & Sheets, V. (2002).
A comparison of methods to test mediation and other intervening variable effects.
Psychological Methods, 7(1), 83-104.
MacMillan, H. L., Georgiades, K., Duku, E. K., Shea, A., Steiner, M., Niec, A., et al.
(2009). Cortisol response to stress in female youths exposed to childhood
maltreatment: Results of the youth mood project. Biological Psychiatry, 66(1),
62-68.
Main, M., & Cassidy, J. (1988). Categories of response to reunion with the parent at age
6: Predictable from infant attachment classifications and stable over a 1-month
period. Developmental Psychology, 24, 415-526.
March, J. S., Parker, J. D. A., Sullivan, K., & Stallings, P. (1997). The Multidimensional
Anxiety Scale for Children (MASC): Factor structure, reliability, and validity.
Journal of the American Academy of Child and Adolescent Psychiatry, 36(4),
554-565.
Margolese, S. K., Markiewicz, D., & Doyle, A. B. (2005). Attachment to parents, best
friend, and romantic partner: Predicting different pathways to depression in
adolescence. Journal of Youth and Adolescence, 34(6), 637-650.
Martin-Merino, E., Ruigomez, A., Wallander, M., Johansson, S., & Garcia-Rodriguez, L.
A. (2010). Prevalence, incidence, morbidity and treatment patterns in a cohort of
patients diagnosed with anxiety in the UK primary care. Family Practice, 27, 9-
16.
McCarty, C. A., & McMahon, R. J. (2003). Mediators of the relation between maternal
depressive symptoms and child internalizing and disruptive behavior disorders.
Journal of Family Psychology, 17(4), 545-556.
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 and Neglect, 30, 171-191.
83

Muris, P., Mayer, B., & Meesters, C. (2000). Self-reported attachment style, anxiety, and
depression in children. Social Behavior and Personality, 28(2), 157-162.
Muris, P., Meesters, C., & van den Berg, S. (2003). Internalizing and externalizing
problems as correlates of self-reported attachment style and perceived parental
rearing in normal adolescents. Journal of Child and Family Studies, 12(2), 171-
183.
Noll, J. G., Trickett, P. K., & Putnam, F. W. (2003). A prospective investigation of the
impact of childhood sexual abuse on the development of sexuality. Journal of
Consulting and Clinical Psychology, 71(3), 575-586.
Orr, S. P., Lasko, N. B., Metzger, L. J., Berry, N. J., Ahern, C. E., & Pitman, R. K.
(1998). Psychophysiologic assessment of women with posttraumatic stress
disorder resulting from childhood sexual abuse. Journal of Consulting and
Clinical Psychology, 66(6), 906-913.
Overbeek, G., Stattin, H., Vermulst, A., Ha, T., & Engels, R. C. M. E. (2007). Parent-
child relationships, partner relationships, and emotional adjustment: A birth-to-
maturity prospective study. Developmental Psychology, 43(2), 429-437.
Parker, J. G., & Herrera, C. (1996). Interpersonal processes in friendship: A comparison
of abused and nonabused children's experiences. Developmental Psychology,
32(6), 1025-1040.
Polanczyk, G., Caspi, A., Williams, B., Price, T. S., Danes, A., Sugden, K., et al. (2009).
Protective effect of CRHR1 gene variants on the development of adult depression
following childhood maltreatment. Archives of General Psychiatry, 66(9), 978-
985.
Porges, S. W. (2001). The polyvagal theory: Phylogenetic substrates of a social nervous
system. International Journal of Psychophysiology, 42, 123-146.
Powers, A., Ressler, K. J., & Bradley, R. G. (2009). The protective role of friendship on
the effects of childhood abuse and depression. Depression and Anxiety, 26, 46-53.
Putnam, F. W., & Trickett, P. K. (1997). Psychobiological effects of sexual abuse. Annals
of the New York Academy of Sciences, 821, 150-159.
Raine, A., Lenczk, T., Bihrle, S., LaCasse, L., & Colletti, P. (2000). Reduced prefrontal
gray matter volume and reduced autonomic activity in APD. Archives of General
Psychiatry, 57, 119-131.

84

Raine, A., Reynolds, C., Venables, P. H., & Mednick, S. A. (1997). Biosocial bases of
aggressive behavior in childhood: Resting heart rate, skin conductance orienting,
and physique. In A. Raine, D. Brennan, D. Farrington & S. A. Mednick (Eds.),
Biosocial Bases of Violence (pp. 107-126). New York: Plenum.
Raine, A., Venables, P. H., & Williams, M. (1990). Relationships between central and
autonomic measures of arousal at age 15 years and criminality at age 24 years.
Archives of General Psychiatry, 47, 1003-1007.
Resnick, G. (1993). Measuring attachment in early adolescence: A manual for the
administration, coding and interpretation of the separation anxiety test for 11-14
year olds.Unpublished manuscript, Rockville, MD.
Richaud de Minzi, M. C. (2006). Loneliness and depression in middle and late childhood:
The relationship to attachment and parental styles. Journal of Genetic Psychology,
167(2), 189-210.
Roelofs, J., Meesters, C., Huurne, M. T., Bamelis, L., & Muris, P. (2006). On the links
between attachment style, parental rearing behaviors, and internalizing and
externalizing problems in non-clinical children. Journal of Family and Child
Studies, 15(3), 331-344.
Roisman, G. I. (2007). The psychophysiology of adult attachment relationships:
Autonomic reactivity in marital and premarital interactions. Developmental
Psychology, 43(1), 39-53.
Roisman, G. I., Tsai, J. L., & Chiang, K. H. S. (2004). The emotional integration of
childhood experience: Physiological, facial expressive, and self-reported
emotional response during the adult attachment interview. Developmental
Psychology, 40(5), 776-789.
Ronnlund, M., & Karlsson, E. (2006). The relation between dimensions of attachment
and internalizing or externalizing problems during adolescence. The Journal of
Genetic Psychology, 167(1), 47-53.
Rottenberg, J., Wilhelm, F. H., Gross, J. J., & Gotlib, I. H. (2003). Vagal rebound during
resolution of tearful crying among depressed and nondepressed individuals.
Psychophysiology, 40, 1-6.
Rubin, K. H., LeMare, L. J., & Lollis, S. (1990). Social withdrawal in childhood:
Developmental pathways to peer rejection. In S. R. Asher & J. D. Coie (Eds.),
Peer Rejection in Childhood. Cambridge Studies in Social and Emotional
Development (pp. 217-249). New York, NY: Cambridge University Press.
85

Salzinger, S., Feldman, R. S., Hammer, M., & Rosario, M. (1993). The effects of
physical abuse on children's social relationships. Child Development, 64(1),
169(119).
Salzinger, S., Rosario, M., Feldman, R. S., & Ng-Mak, D. (2007). Adolescent suicidal
behavior: Associations with preadolescent physical abuse and selected risk and
protective factors. Journal of the American Academy of Child and Adolescent
Psychiatry, 46(7), 859-866.
Salzman, J. P. (1996). Primary attachment in female adolescents: Association with
depression, self-esteem, and maternal identification. Psychiatry: Interpersonal &
Biological Processes, 596(1), 20-33.
Schore, A. N. (2000). Attachment and the regulation of the right brain. Attachment and
Human Development, 2, 23-47.
Schore, A. N. (2001). Effects of a secure attachment relationship on right brain
development, affect regulation, and infant mental health. Infant Mental Health
Journal, 22, 7-66.
Schultz, D., Tharp-Taylor, S., Haviland, A., & Jaycox, L. (2009). The relationship
between protective factors and outcomes for children investigated for
maltreatment. Child Abuse & Neglect, 33, 684-698.
Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and nonexperimental
studies: New procedures and recommendations. Psychological Methods 7(4),
422-445.
Siegel, D. J. (1999). The Developing Mind. New York: Guilford Press.
Siegel, D. J. (2001). Toward an interpersonal neurobiology of the developing mind:
Attachment relationships, mindsight, and neural integration. Infant Mental Health
Journal, 22, 67-94.
Sobel, M. E. (1982). Asymptotic confidence intervals for indirect effects in structural
equation models. In S. Leinhardt (Ed.), Sociological Methodology (pp. 290-312).
Washington DC
American Sociological Association.
Spinhoven, P., Elzinga, B. M., Hovens, J. G. F. M., Roelofs, K., Zitman, F. G., van
Oppen, P., et al. (2010). The specificity of childhood adversities and negative life
events across the life span to anxiety and depressive disorders. Journal of
Affective Disorders.
86

Stein, M. B., Schork, N. J., & Gelernter, J. (2008). Gene-by-environment (serotonin
transporter and childhood maltreatment) interaction for anxiety sensitivity, an
intermediate phenotype for anxiety disorders. Neuropsychopharmacology, 33,
312-319.
Stuewig, J., & McCloskey, L. A. (2005). The relation of child maltreatment to shame and
guilt among adolescents: Psychological routes to depression and delinquency.
Child Maltreatment, 10(4), 324-336.
Tabachnik, B. G., & Fidell, L. S. (2007). Using Multivariate Statistics (5th ed.). Boston:
Pearson Education, Inc.
Tarullo, A. R., & Gunnar, M. R. (2006). Child maltreatment and the developing HPA
axis. Hormones and Behavior, 50(4), 632-639.
Tonhajzerova, I., Ondrejka, I., Javorka, K., Turianikova, Z., Farsky, I., & Javorka, M.
(2010). Cardiac autonomic regulation is impaired in girls with major depression.
Progress in Neuro-Psychopharmacology & Biological Psychiatry.
Toth, S. L., & Cicchetti, D. (1996). Patterns of relatedness, depressive symptomatology,
and perceived competence in maltreated children. Journal of Consulting and
Clinical Psychology, 64(1), 32-41.
Trickett, P. K., & McBride-Chang, C. (1995). The developmental impact of different
forms of child abuse and neglect. Developmental Review, 15(3), 311-337.
Trickett, P. K., Mennen, F. E., Kim, K., & Sang, J. (2009). Emotional abuse in a sample
of multiply maltreated, urban young adolescents: Issues of definition and
identification. Child Abuse and Neglect, 33, 27-35.
Turner, H. A., Finkelhor, D., & Ormrod, R. (2010). The effects of adolescent
victimization on self-concept and depressive symptoms. Child Maltreatment,
15(1), 76-90.
Tyrka, A. R., Price, L. H., Gelernter, J., Schepker, C., Anderson, G. M., & Carpenter, L.
L. (2009). Interaction of childhood maltreatment with the corticotropin-releasing
hormone receptor gene: Effects on hypothalamic-pituitary-adrenal axis reactivity.
Biological Psychiatry, 66(7), 681-685.
Udupa, K., Sathyaprabha, T. N., Thirthalli, J., Kishore, K. R., Lavekar, G. S., Raju, T. R.,
et al. (2007). Alteration of cardiac autonomic functions in patients with major
depression: A study using heart rate variability measures. Journal of Affective
Disorders, 100, 137-141.
87

van Brakel, A. M. L., Muris, P., Bogels, S. M., & Thomassen, C. (2006). A multifactorial
model for the etiology of anxiety in non-clinical adolescents: Main and interactive
effects of behavioral inhibition, attachment and parental rearing. Journal of Child
and Family Studies, 15, 569-579.
van der Vegt, E. J. M., van der Ende, J., Kirschbaum, C., Verhulst, F. C., & Tiemeier, H.
(2009). Early neglect and abuse predict diurnal cortisol patterns in adults: A study
of international adoptees. Psychoneuroendocrinology, 34(5), 660-669.
Verschueren, K., & Marcoen, A. (2005). Perceived Security of Attachment to Mother and
Father: Developmental Differences and Relations to Self-Worth and Peer
Relationships at School. In K. A. Kerns & R. A. Richardson (Eds.), Attachment in
middle childhood. (pp. 212-230). New York City: Guildford Press.
Vivona, J. M. (2000). Parental attachment styles of late adolescents: Qualities of
attachment relationships and consequences for adjustment. Journal of Counseling
Psychology, 47(3), 316-329.
Ward, N. G., & Doerr, H. O. (1986). A potentially sensitive and specific marker for
depression. The Journal of Nervous and Mental Disease, 174(9), 553-559.
Warren, S. L., Huston, L., Egeland, B., & Sroufe, L. A. (1997). Child and adolescent
anxiety disorders and early attachment. Journal of the American Academy of
Child and Adolescent Psychiatry, 36(5), 637-644.
Waters, E., Merrick, S., Treboux, D., Crowell, J., & Albersheim, L. (2000). Attachment
security in infancy and early adulthood: A twenty-year longitudinal study. Child
Development, 71(3), 684-689.
Watkins, L. L., Grossman, P., Krishnan, R., & Sherwood, A. (1998). Anxiety and vagal
control of heart rate. Psychosomatic Medicine, 60, 498-502.
Weinfield, N. S., Sroufe, L. A., & Egeland, B. (2000). Attachment from infancy to early
adulthood in a high-risk sample: Continuity, discontinuity, and their correlates.
Child Development, 71(3), 695-702.
Wright, J., Briggs, S., & Behringer, J. (2005). Attachment and the body in suicidal
adolescents: A pilot study. Clinical Child Psychology and Psychiatry, 10(4), 477-
491.
Yeragani, V. K., Tancer, M., Seema, K. P., Josyula, K., & Desai, N. (2006). Increased
pulse-wave velocity in patients with anxiety: implications for autonomic
dysfunction. Journal of Psychosomatic Research, 61, 25-31.
88

Young, A. S., Klap, R., Shoai, R., & Wells, K. B. (2008). Persistent depression and
anxiety in the United States: Prevalence and quality of care. Psychiatric Services,
59(12), 1391.
Zelenko, M., Kraemer, H., Huffman, L., Gschwendt, M., Pageler, N., & Steiner, H.
(2005). Heart rate correlates of attachment status in young mothers and their
infants. Journal of the American Academy of Child and Adolescent Psychiatry,
44(5), 470-476.

Abstract (if available)

Abstract The current study investigated four hypotheses. Hypothesis 1: Low attachment security will predict high levels of anxiety and depression

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Untangling the developmental relations between depression and externalizing behavior among maltreated adolescents

PDF

The impact of child maltreatment on the mental health of adolescents: a longitudinal study of social anxiety and self-perception

PDF

Trajectories of internalizing problems among maltreated girls and boys: differences by maltreatment type and developmental timing

PDF

Examining the longitudinal relationships between community violence exposure and aggressive behavior among a sample of maltreated and non-maltreated adolescents

PDF

Autonomic inertia as a proximal risk marker for moments of perseverative cognition in everyday life in remitted depression

PDF

Modeling of cardiovascular autonomic control in sickle cell disease

PDF

Maltreated adolescents and their families: a longitudinal examination of family functioning, parenting attitudes, & youth mental health

PDF

Examining the associations of respiratory problems with psychological functioning and the moderating role of engagement in pleasurable activities during late adolescence

PDF

The relationship between social anxiety and verbal creativity in the context of induced mindfulness

PDF

Using observed peer discussions to understand adolescent depressive symptoms and interpersonal interactions

PDF

The link between maternal depression and adolescent daughters' risk behavior: the mediating and moderating role of family

PDF

Familial and cultural variables as predictors of retention of Latino engineering students

PDF

Social attributes of adolescents’ problem-talk partners predict changes in depression

PDF

The old ball and linkage: couples’ prenatal conflict behavior, cortisol linkage, and postpartum depression risk

PDF

Adolescent life stress and the cortisol awakening response: the moderating roles of emotion regulation, attachment, and gender

PDF

Allocation of educational resources to improve student achievement: Case studies of non-title I schools

PDF

The role of monoamine oxidase in behavioral plasticity

PDF

Prenatal predictors of parental sensitivity in first-time mothers and fathers

PDF

A series of longitudinal analyses of patient reported outcomes to further the understanding of care-management of comorbid diabetes and depression in a safety-net healthcare system

PDF

Addressing federal pain research priorities: drug policy, pain mechanisms, and integrative treatment

Asset Metadata

Creator Feres, Nashla (author)

Core Title Attachment, maltreatment and autonomic nervous system responsivity as predictors of adolescent anxiety and depression

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Psychology

Degree Conferral Date 2010-08

Publication Date 08/10/2010

Defense Date 05/12/2010

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag anxiety,attachment,child depression inventory,Depression,Fathers,maltreatment,mediational model,multidimensional anxiety scale for children,neglect,OAI-PMH Harvest,physical abuse,psychophysiology,respiratory sinus arrhythmia,security scale,Sexual Abuse,skin conductance level,structural equation modeling

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Trickett, Penelope K. (committee chair), Dawson, Michael E. (committee member), Manis, Franklin R. (committee member), Mennen, Ferol E. (committee member)

Creator Email feres@usc.edu,nashla.feres@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m3375

Unique identifier UC1452894

Identifier etd-Feres-3944 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-381870 (legacy record id),usctheses-m3375 (legacy record id)

Legacy Identifier etd-Feres-3944.pdf

Dmrecord 381870

Document Type Dissertation

Rights Feres, Nashla

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu