THE ROLE OF SOCIAL CAPITAL ON CHILDREN’S BEHAVIOR IN THE CONTEXT OF
MATERNAL DEPRESSION AND NEIGHBORHOOD POVERTY
By
Sunhee Moon

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Human Development and Family Studies
2010

ABSTRACT
THE ROLE OF SOCIAL CAPITAL ON CHILDREN’S BEHAVIOR IN THE CONTEXT OF
MATERNAL DEPRESSION AND NEIGHBORHOOD POVERTY
By
SUNHEE MOON

A most fundamental function of family is the parenting process, which demonstrates how
parents socialize their children. A child’s processes of growing as a member of society are
specified by parenting practices in daily life. Parents experiencing psychological disorder or
poverty are likely to be more involved in conflictual interactions with their children, thus
occasioning less satisfactory social development of those children.
This study’s purpose is to examine the paths for household income and child behavior
problems among categorized groups, based on neighborhood poverty and a mother’s depression
level, by exploring the operating mediators for each group. This study examines how these
operating mediators are moderated by a mother’s depression and by neighborhood poverty, by
comparing the groups. Last, this study provides policy and clinical implications that can improve
a child’s psychological well-being.
Drawing on social capital theory, this study examines the connection between household
income and child’s internalizing and externalizing behavior problems in the context of mother’s
depression and neighborhood poverty. Emotional interactions and discipline responses with
mother-child from family social capital, and social cohesion and interaction with neighborhoods
from neighborhood social capital, are employed to determine their mediating effect on a child’s
behavior problems.

The study utilizes the Los Angeles Family and Neighborhood Survey (L.A.FANS) and
secondary data analysis to describe and make inferences about hypothesized relationships.
Methods describing the primary analytical technique of Structural Equation Modeling (SEM)
with multiple groups and complex samples are used for the study.
Results demonstrate that there are mediation effects of discipline responses, emotional
interaction and social cohesion between household income and child’s behavior problems for
multiple groups, supporting the hypotheses for mediation effects. However, there is no statistical
significance of moderation effects by comparing multiple groups, despite the significance of
meditation effects across groups.
Findings imply that negative parenting attitudes are associated with increasing a child’s
overall behavior problems, including anxiety, depression, dependence, peer conflicts,
hyperactivity, stubbornness, and antisocial conduct. In addition, as important assets for a
deprived neighborhood, social cohesion and mutual trust may have a protective effect on child
problematic behavior. Policymakers should provide intervention programs addressing the need
for increasing neighborhood social capital as resources for families residing in poor
neighborhoods. Additionally, clinical intervention should be considered through a cooperative
approach between family and neighborhood, so as to establish both healthy families and better
neighborhoods.

Copyright by
Sunhee Moon
2010

ACKNOWLEDGMENTS

To all those who have made this dissertation possible and because of whom my graduate
experience has been one I always will cherish, I express my sincere thanks.
My deep gratitude is to my advisor, Dr. Esther Onaga who gave me freedom to explore
on my own and guidance to recover when I faltered. Her patience and support helped me
overcome many crises and finish this dissertation. I hope that one day I would become as good
an advisor to my students as she has been to me.
Thank you, sincerely, to my committee members Drs. Robert Griffore, John Schweitzer,
and Kathleen Jager for their guidance over the years. Dr. Robert Griffore always patiently and
listened to and advised me soundly, especially regarding technical details of my work. Thank
you for his expertise, many suggestions, and insightful feedback. Dr. John Schweitzer’s
insightful comments and constructive criticisms at every stage of my study were thoughtprovoking and helped me focus my ideas; in particular, he held me to a high research standard
and rightly insisted on strict validations for developing and polishing my research ideas. I am
thankful to Dr. Kathleen for commenting on my views and helping me enrich my ideas.
Dr. Youngil Cho of the Department of Psychology at Arizona State University has my
appreciation for helping the method section of my dissertation and for technical advice and
assistance. Also thank you to Dr. Steven Pierce, Associate Director of CSTAT (Center of
Statistical Training and Consulting) at Michigan State University for his insight and advice to me
in developing the statistical method of my study. Thank you also to Michael Morris, Michigan
State University’s faculty member inter alia specializing in academic text, for carefully reading,
and discussing with me revisions to, my dissertation manuscript.

v

Eunjin Han, Sumin Oh, Joungah Lee, Youngeun Lee, and Juyeun Kim were my fine
colleagues at Michigan State University. Thank you to them and for the sustaining friendship of
Kyungsook Lee Also my appreciation for their love and devotion is due to my friends in the
"Sisters of Lydia" prayers meetings, namely Eunsil Lee, Malsoon Lee, Hyunju Ha, Heygu Lee,
Sangeun Lee, Miran Kim, Yumi Chin, Suhyun Park, Mikyung Kim, Jaemin Cha, and Seunghyun
Kim, who through my doctoral years shared with me so many moments of happiness and
unhappiness.
Most important, none of this would have been possible without the love, support and
patience of my family, my parents and parents-in-law, my sisters, sisters-in-law, brother,
brothers-in-law, and nephews. Of course, extra-special thanks to my immediate family: my
husband, Kyungshik Yang, who sacrificed endlessly for my graduate study, and to my only son,
Taeyeun, the special joy of my life, who for many years in the United States shared all joyful and
sorrowful experiences with me, and well-deserves my deepest appreciation for completion of my
study.
My highest thanks to God Who rendered all this possible and blessed me throughout.

vi

TABLE OF CONTENTS

LIST OF TABLES ...............................................................................................................x
LIST OF FIGURES ........................................................................................................... xi
CHAPTER I
INTRODUCTION ............................................................................................................1
Problem Statement........................................................................................................1
The Purpose of this Study .............................................................................................5
CHAPTER II
LITERATURE REVIEW .................................................................................................6
Maternal Depression and Child Development .............................................................6
The depressed mother and the child.........................................................................6
Why are mothers depressed? ...................................................................................9
Neighborhood characteristics and maternal depression .........................................10
The effects of parent-child interaction in families of mothers with depression ....11
Maternal discipline response to child behavior .....................................................12
Children of mothers with depression and behavior problems ...............................13
Theoretical and Conceptual Foundation .....................................................................16
Key concepts of social capital................................................................................16
Economic status and social capital ........................................................................19
Family social capital and home environment ........................................................20
Neighborhood social capital and child social development ...................................23
Effects of Neighborhood Poverty ...............................................................................24
Neighborhood poverty and maternal psychological effects...................................24
Neighborhood poverty and the social development of children ............................25
CHAPTER III
METHOD .......................................................................................................................29
Description of Dataset ................................................................................................29
Sample for the Study...................................................................................................31
Measurements .............................................................................................................32
Dependent variable ................................................................................................32
Child behavior problems........................................................................................32
Independent variables ............................................................................................33
Household income..................................................................................................33
Social cohesion ......................................................................................................33
Interaction with neighborhoods .............................................................................34
Mother-child emotional interaction quality ............................................................34
Discipline responsiveness of mother .......................................................................35
Moderator variables ................................................................................................35
Neighborhood poverty level ....................................................................................35
vii

Feeling depression of mother...................................................................................36
Hypothesized Model ...................................................................................................36
Research Questions and Hypotheses ..........................................................................38
Data Analysis ..............................................................................................................40
Analytical method ..................................................................................................40
Analysis procedures ................................................................................................41
Treating missing data ..............................................................................................45
CHAPTER IV
RESULTS.......................................................................................................................47
Descriptive Statistics ..................................................................................................47
Exploratory Factor Analysis and Item Parceling on Dependent Variables ................52
Correlations among Variables for Groups ..................................................................54
Measurement Equivalence of Confirmatory Factor Model across Four Groups ........59
Sequence of equivalence testing ............................................................................59
Structural Equation Modeling .....................................................................................63
Mediation testing ...................................................................................................64
Group with depressed mothers/poor neighborhoods .............................................65
For internalizing symptoms ...................................................................................65
For externalizing symptoms ...................................................................................65
Group with depressed mothers/non-poor neighborhoods ......................................66
For internalizing symptoms ...................................................................................66
For externalizing symptoms ...................................................................................66
Group with non-depressed mothers/poor neighborhoods ......................................70
For internalizing symptoms ...................................................................................70
For externalizing symptoms ...................................................................................70
Group with non-depressed mothers/non-poor neighborhoods ...............................71
For internalizing symptoms ...................................................................................71
For externalizing symptoms ...................................................................................71
Significance of Mediation Effects ..............................................................................74
Multi-Group Comparison (Moderated Mediation Testing) ........................................77
CHAPTER V
DISCUSSION AND IMPLICATION ............................................................................82
Discussion ..................................................................................................................82
Discipline response of mothers ..............................................................................83
Emotional interaction of mother and child ............................................................86
Social cohesion ......................................................................................................87
Implications ................................................................................................................90
Implications of family social capital theory...........................................................90
Policy implications of neighborhood social capital theory ....................................91
Implications for clinical intervention .....................................................................93
CHAPTER VI
LIMITATION AND FUTURE RESEARCH ................................................................96
Summary ....................................................................................................................96
viii

Limitation ...................................................................................................................98
Future Research ..........................................................................................................99
APPENDIX .....................................................................................................................102
REFERENCES ...............................................................................................................107

ix

LIST OF TABLES

Table 1 Variable Structure ................................................................................................42
Table 2 Descriptive Statistics of Sample Population ........................................................50
Table 3 Parceling from EFA Results ................................................................................52
Table 4 EFA Results for Dependent Variables .................................................................53
Table 5 Correlation among Variables for Depressed Mothers/Poor Neighborhood
Group ...................................................................................................................55
Table 6 Correlation among Variables for Depressed Mothers/Non-poor Neighborhood
Group ...................................................................................................................56
Table 7 Correlation among Variables for Non-depressed Mothers/Poor Neighborhood
Group ...................................................................................................................57
Table 8 Correlation among Variables for Non-depressed Mothers/Non-poor
Neighborhood Group ...........................................................................................58
Table 9 Measurement Equivalence of Confirmatory Factor Models of Six-Factor
Solutions across Four Groups ..............................................................................61
Table 10 Results of Measurement Equivalence ................................................................63
Table 11 Significance of Mediation ..................................................................................76
Table 12 Equivalence of Structural Models for Internalizing Behavior ...........................79
Table 13 Equivalence of Structural Models for Externalizing Behavior ...........................81
Table 14 Behavior Problems Index Subscales .................................................................102

x

LIST OF FIGURES

Figure 1 Hypothesized Model ..........................................................................................37
Figure 2 Measurement Model ..........................................................................................62
Figure 3 Structural Model for Depressed mothers/Poor neighborhood Group................68
Figure 4 Structural Model for Depressed mothers/Non-poor neighborhood Group........69
Figure 5 Structural Model for Non-depressed mothers/Poor neighborhood Group ........72
Figure 6 Structural Model for Non-depressed mothers/Non-poor neighborhood Group 73

xi

CHAPTER I
Introduction

Problem Statement
One of the most fundamental functions of family is the parenting process, i.e., how
parents socialize their children. The processes a child has as he or she is growing as a member of
society are specified by parenting practices experienced in daily lives. Parents’ influential power
is extensive in terms of socialization of a child within a family. Parenting strategies generally
depend on their personal traits, including personality characteristics, socioeconomic status, and
mental health, because it is not possible for a child to determine the quality of his or her
relationship with his or her parents. Especially, parents’ secure mental status may play a key role
in good parenting.
Parents who experience psychological disorder or poverty are likely to be less healthy,
both emotionally and psychologically, than parents who are not poor (Adler, Boyce, Chesney,
Folkman, & Kahn, 1994). Primary caregivers - specifically mothers - with depressive symptoms
become more involved in conflictual interaction with their children, the result being less
satisfactory social development of those children (Hammen, Brennan, & Shih, 2004). In fact,
although poverty is most detrimental to children, positive parenting practice in a family
disadvantaged by poverty can serve as a protective factor to help prevent problematic behaviors
in their children (Brooks-Gunn & Duncan, 1997; Kiernan & Huerta, 2008; Pachter, Auinger,
Palmer, & Weitzman, 2006).
Families in disadvantaged neighborhoods often do not have the resources to meet the
needs in a daily life. Although government and private organization provide support and social

1

service to assist the “worthy” poor (Brisson & Usher, 2005), the effects of neighborhood poverty
on families and children have been associated with multiple risk factors that accompany that
condition. Especially, mothers and children are more likely to be exposed to risk factors that
cause psychological distresses because the environments in which both live are mostly
disadvantaged environments.
Continuous exposure to an impoverished disorganized neighborhood context can be a
critical risk factor of maternal depression. The prevalence of maternal depression among families
living in poor neighborhoods is consistent across ethnic groups with a high incidence of
behavioral problems occurring in the children (Pachter et al., 2006). Further, depressive
symptoms in mothers and emotional and behavioral disturbances in children tend to coexist
(Elgar, McGrath, Waschbusch, Stewart, & Curtis, 2004). Psychological disorders of mothers not
only influence parenting practices negatively, but also produce in those mothers difficulty when
trying to form relationships with others such as friends and neighbors.
Additionally, the incidence of disadvantaged neighborhoods is a deleterious factor that
affects children’s socialization. Many studies have investigated the impact of poverty
neighborhoods on a child’s behavior problems (Brooks-Gunn & Duncan, 1997). The most
common factors in explanation of neighborhood impact are such indicators as income, residential
turnover, and racial/ethnic heterogeneity (Simons, Simons, Burt, Brody, & Cutrona, 2005).
However, some research shows that associations between neighborhood characteristics and
children’s behavioral problems are modest and weaker than the effects of family and individual
characteristics of family members (Leventhal & Brooks-Gunn, 2000; Pebley & Sastry, 2003,
Sampson, Morenoff, & Gannon-Rowley, 2002). Additionally, there are few or minor direct
effects of neighborhood on adolescent outcomes (Furstenberg, Cook, Eccles, Elder, & Sameroff,

2

1999). Family processes and neighborhood characteristics, however, do act as mediators between
neighborhood context and children’s behavior and are interwoven in the actual environment.
To complement these weak research findings, several studies have investigated whether
there are more powerful effects from the neighborhood context, but produced indirectly through
family-level factors. These studies’ focus is on possible interaction between neighborhood
context and internal family processes. For example, parent-child conflict and neighborhood
disadvantage have been significantly associated with anti-social behavior in children (Ingoldsby,
Shaw, Winslow, Schonberg, Gilliom, & Criss, 2006). This trend of studies, however, has
produced mixed results (Simon et al., 2005), some studies finding a significant effect of
interaction between family process and neighborhood context on child’s developmental
outcomes (Ingoldsby et al., 2006; Leventhal & Brooks-Gunn, 2004), while others have not
(Klevanov, Brooks-Gunn, McCarton, & McCormick, 1998). Thus, the impact of neighborhood
on children’s development has remained controversial because it is difficult to disentangle
neighborhood factors from family processes (Leventhal & Brooks-Gunn, 2003).
In fact, family processes, specifically, parental, may be the strongest factor affecting
children’s socialization. However, some potential factors can also affect on family and child’s
development in certain contexts, such as living in a poor neighborhood or experiencing a
negative life event. Understanding a child’s socialization process should be explained through
exploring more complex contexts and dynamic interactions of the family. To date, most nonexperimental neighborhood research has been conducted via census-based measures of structural
characteristics in data collected on children and families (Leventhal & Brooks-Gunn, 2003).
These streams of recent research require moving beyond a simple consideration of the static

3

socio-demographic characteristics of neighborhoods (Simon et al., 2005) to examine a child’s
socialization process.
Children’s social development is not only explained by structural factors, such as
socioeconomic status of parents and neighborhoods, but also interpreted by applying the more
complicated mechanism which families experience. Investigation of child behavioral outcomes
needs to address the actual context of interactional processes that a child and an adult experience
in daily lives. Because the quality of a person-to-person relationship is too often regulated by
routine events and interactions in daily life, continuous interactions within the same living
boundary provide additional opportunities to accumulate a higher quality of trust and a greater
sense of closeness. Thus, the indicator of individual perception when measuring the quality of
interaction within a family or a neighborhood can produce different outcomes, beyond simply
examining given static conditions that the family has in different contexts. Indeed, children’s
problematic behaviors are affected by stressful situations inside and outside a family, but also
can differ in terms of reactions and outcomes such as how parents interact with their children and
the cohesive relationships parents and children share with neighbors.
In response to this research flow, the current study is concerned with the precise
mechanisms of how mother-child interactions, and perceptions of the interaction process with
neighbors, affect child behaviors across different contexts. Maternal depression and family
poverty can be definite determinants of child behavioral problems, but additional operating
mediators can differ according to the condition of the contexts that a mother and child actually
experience.

4

Purpose of this Study
The purpose of the current study is to examine the paths for neighborhood income level
and child behavior problems among categorized groups based on neighborhood poverty and a
mother’s depression level. Through exploring these paths, this study will examine the impact on
a child’s psychological well-being by examining precise operating mediators for each
characterized group.
Next, by comparing certain categorized groups, this study will examine how these
operating mediators are moderated by a mother’s depression and neighborhood poverty level.
Last, through mediating and moderating the effects of relations among these designated variables,
this study will offer policy- and clinical- implications that can improve child psychological wellbeing, i.e., specifically can ameliorate behavior problems in the context of neighborhood poverty
and maternal depression.

5

CHAPTER II
Literature Review

This chapter includes a literature review with particular reference to discussions of
depression and poverty as the main two contexts of this study, namely the theoretical background
for the study and the application to a social capital perspective. The chapter includes three
sections. The first covers the processes and characteristics of maternal depression and how it
influences a child’s developmental outcomes. The second discusses major concepts of social
capital, the importance of family social capital and of the home environment, the importance of
neighborhood social capital and its influence on the psychological health of mother and child.
The third discusses how neighborhood poverty influences a mother’s psychological health and
child’s social development.
Maternal Depression and Child Development
The depressed mother and the child
Depression is the most prevalent psychiatric disorder in the adult population (American
Psychiatric Association [APA], 1994). Especially within the general public, an estimated 10 to
25 % of women at some point in their lives are affected by depression symptoms (American
Psychiatric Association [APA], 2000). Thus, it is important to understand how a mother’s
depression symptoms affect her child’s social development because depression is one of the most
common psychological disorders, exposing children to the fragile mental status of a mother.
Myriad of research already has examined this and found that maternal depression is a critical risk
factor in the psychological development of the child (Burke, 2003; Goodman & Gotlib, 1999;
Gelfand & Teti, 1990).

6

In general, a person with depression tends to be more self-focused and have negatively
biased self-perception, high-levels of self-punishment, and low levels of self-reinforcement. Also,
such an individual is more likely to exhibit negative behavior in interactions with others
(Goodman & Gotlib, 1999). Further, since depression is a disorder extensively affecting a
person’s social and interpersonal functioning (Luoma, Taminen, Kaukonen, Laippala, Puura, &
Salmelin, 2001), a mother’s depression may display and communicate negativity in terms of
psychosocial development of her children. Luoma and colleagues (2001) examined maternal
depressive symptoms connected with high levels of internalizing and externalizing behavioral
problems as well as low levels of psychosocial functioning in school-age children, and
determined that maternal depression can be a predictor of a child’s well-being.
In addition, the mother’s depression can be transmitted to the child. The study compared
depressed adolescents in terms of mothers with and without depression (Hammen & Brennan,
2001) and determined that mothers with depression often produce youngsters who have
depression. Depressed children of depressed mothers were also more likely to have elevated rates
of dysfunction in their interpersonal relationships than were children of women without a display
of depression mood. Also, family stress and conflict, such as negative life events or
interpersonal difficulties played a more important role in the development of depressive
disorders in children of women with depression than it did in children of women without
depression (Hammen, Brennan, & Shih, 2004). When the family stress rate was low, there was a
significant difference, suggesting less of a problem among the offspring of both depressed and
non-depressed mothers.
Eventually, a child with a depressed mother becomes exposed to poor parenting practice
because of that mother’s depressive disorder (Gelfand & Teti, 1990; Goodman & Gotlib, 1999;

7

Luoma, et al., 2001). Children reared by a mother with depressive symptoms can have deficits or
delays in the development of their behavioral or cognitive skills (Goodman & Gotlib, 1999). This
finding indicates that maternal depression is connected to undesirable parenting practices, such
as unresponsiveness, inattention, or inappropriate discipline (Gelfand & Teti, 1990), thus induces
problematic parenting and poor social and educational outcomes for their children.
Many studies have explored this association to a variety of parenting behaviors and
ongoing problems in child development (Downey & Coyne, 1990; Goodman, 1993; Dix &
Meunier, 2009). According to Dix and Meunier (2009), depressive symptoms predict maternal
withdrawal, defined as low responsiveness and lack of ongoing involvement with children, and
also will predict flat and negative expression of emotion to children. Furthermore, maternal
depressive symptoms predict ineffective discipline, specifically discipline that is harsh,
inconsistent, manipulative, and/or indulgent (Leung & Slep, 2006), eventually leading to poor
developmental outcomes for children (Louma et al.,2001). Children’s externalizing behavior
problems leading to aggression and acting out may be due to less sensitive parenting by
depressed mothers.
Poor parenting practice of mothers with depression, and poor development outcomes for
their children, suggest there are mutual influences that connect maternal depression and a child’s
development (Elgar et al., 2004). Compared to mothers without depression, mothers with
depression are less likely to interact with their children and more likely to respond harshly in
terms of daily parenting practice. Since mothers with depression are more likely to have high
levels of self-punishment (Goodman & Gotlib, 1999), their parenting style may also be more
punitive than that of mothers without depression. Mothers with depression not only treat their
children more harshly, but also feel guilt and anxiety about that behavior at the same time

8

(Gelfand & Teti, 1990). Indeed, harsh punitive parenting style is a factor that weakens the stable
and continuing emotional relationship between parent and child.
Previous research mainly focused on the emotional interaction between a mother with
depression and her infant (Murray, Cooper, Wilson, & Romaniuk, 2003; Stein, Gath, Bucher,
Bond, Day & Cooper, 1991). Children of mothers with depression during early childhood may
not form the secure emotional attachment to mothers that they need and want. Such insecure
emotional attachment negatively affects their social adjustment in later childhood or adolescence.
Studies of a mother with depression focused more on developmental parenting
disturbance between a mother and infancy- and early childhood- periods than on school-age
childhood behavior problems. Regarding negative effects of maternal depression on
developmental outcomes in parent-child emotional and behavioral interaction, to date relatively
few studies of school age children have been undertaken (Gelfand & Teti, 1990).
Why are mothers depressed?
Mothers are depressed for various reasons. A major portion of previous research has
focused on the association between mothers’ postpartum depression symptoms and their
children’s developmental tasks (Sinclair & Murray, 1998; Stein, et al., 1991). In general,
maternal depression accompanies the process of child rearing over time; however, potential
variables that dividing mothers with depression from mothers without depression include
mothers’ characteristics and negative life events (Webster-Stratton & Hammond, 1988). It is not
possible to change a person’s natural characteristics, but a number of life events due to social and
environmental variables influence extensively the process of a mother’s psychological transition.
It is also clear that mothers who experience more negative life events within their families or in
the workplace are likely to be more depressed. If a mother experiences negative events, such as

9

poverty or unemployment, these stressful events can produce depressive symptoms. Being poor
accompanies deficit of resources needed for childrearing. Thus, maternal depression when a
child is of school age tends to expose not only hardships in parenting practice but also hardships
derived from poverty, and eventually leads to child’s social mal-adjustment.
Neighborhood characteristics and maternal depression
Low socioeconomic status - SES, lack of resources, and physical stressors found in a
disadvantaged environment can predict maternal depression (Cutrona, Wallace, & Wesner, 2006).
Families do have some choice about where they live (Duncan & Brooks-Gunn, 1997; Leventhal
& Brooks-Gunn, 2004), but poor families are often limited in their residential choices and may
be unable to choose a good school district and safe neighborhood. Residential segregation by
neighborhood SES characteristics predicts as substantial a causal impact on child developmental
outcomes as that on a mother’s mental health. Previous studies have documented a relationship
between the association of neighborhood characteristics and the mother’s mental health (Balaji,
Claussen, Smith, Visser, Morales, & Perou, 2007; Cutrona, et al., 2006; Hill & Herman-Stahl,
2002; Klebanov, Brooks-Gunn, & Duncan, 1994; Pachter et al., 2006; Ross, 2000) and child
outcomes (Caughy, O'Campo, & Muntaner, 2003; Duncan & Brooks-Gunn, 2000, 1997;
Leventhal & Brooks-Gunn, 2000, 2004).
The stress imposed by unfavorable neighborhoods can increase a mother’s depression as
well as the child’s negative social development, because that stress can be greater than the effects
of individual personal stress. A neighborhood with few resources and safety lapses, weakens
parenting strategies and eventually will affect child development(Leventhal & Brooks-Gunn,
2000). Hill and Herman-Stahl (2002) examined the causal relation between neighborhood
characteristics and parenting, using maternal depression as a mediator, and suggested that

10

negative neighborhood characteristics will impede effective parenting practice by exposing
mother’s depressive symptoms.
Since parenting practice is a daily process between parent and child, neighborhoods in
which family daily life occurs can affect parenting quality by its abundance or dearth of
resources and environments necessary for raising children. If the frequency of social contact
within safer neighborhoods among mothers increases, mothers will encounter less stress in
supervising their children. As such, mothers’ parental stress will reduce, and this will connect to
more positive behavioral development of the child.
On the other hand, Steinberg and colleagues (1992) examined the mediating effect of the
parenting practice of social integration on adolescent behavior. Children residing in
neighborhoods with a high level of collective socialization have an increased possibility of prosocial behaviors and positive developmental outcomes.
The effects of parent-child interaction in families of mothers with depression
School-aged children are at the development stage where they build academic skills, and
establish good peer relations and sound social skills (Gelfand & Teti, 1990). However, children
of mothers with depression may lack intimate parental supervision because it may be difficult for
a depressed mother to provide appropriate guidance in learning positive social skills (Goodman
& Gotlib, 1999). Similarly to mothers with depression who have negative self-concepts, children
of a mother with depression will also develop a lower level of self-perception. For this reason,
the developmental achievement of these children may be disturbed or take longer.
However, the quality of mother-child interactions can be a mediator between maternal
depression and a child’s psychological outcomes (Harnish, Dodge, & Valente, 1995). Harnish
and his colleagues (1995) examined the mediation effect of mother-child interactions in each

11

lowest and highest socio-economic groups in both Caucasian and African-American families.
Their findings indicate that poor mother-child interaction quality was mediated by maternal
depression and increased risk for child externalizing problems for Caucasian families. In a low
SES group, if the quality of mother-child interaction was poor, child externalizing behavior
problems were worse, suggesting that socio-economic status affects child behavior problems via
the mother-child interaction in depressed mother families. In terms of a punitive parenting style,
a mother’s behavioral response to her child is part of that mother-child interaction.
Maternal discipline response to child behavior
Physical discipline for misbehavior has been examined for effectiveness of consistent
parenting practice (Darling & Steinberg, 1993; Fletcher, Walls, Cook, Madison, & Bridge, 2008).
Mothers with psychological disorders such as depression or aggression are more likely to apply
harsh physical punishment in parenting practice than are mothers with a non-psychological
disorder. Too much punitive discipline as a parenting strategy may cause children to have
problematic behaviors. Previous studies have found that the relation between a mother’s
aggressive attribution tendencies and her children’s externalizing behavior problems is mediated
by the mother’s harsh discipline responses to children’s misbehavior (Nix, Pinderhughes, Dodge,
Bates, Pettit, & McFadyen-Ketchum, 1999; Patterson, Reid & Dishion, 1992; Simons, Johnson,
& Conger, 1994).
Exposure to physical punishment discipline has been linked to negative behavior for
adolescents (Bender, Allen, McElhaney, Antonishak, Moor, & Kelly, 2007). Adolescents who
experienced harsh maternal discipline were less likely to maintain close relationships with their
mothers. They had difficulty expressing warmth and engagement during interaction with their
mothers, suggesting that the parenting might be characterized by high parental control, verbal

12

hostility, restrictiveness, and other punitive discipline strategies (Robinson, Mandleco, Olsen, &
Hart, 1995). These adolescents who revealed less intimate relationships with their mothers
demonstrated deficits in establishing autonomy, one of the main developmental tasks during
adolescence (Bender et al., 2007). From early childhood to adolescence, the negative parenting
practice of a mother with depression may lead to a less autonomous and more dependent child,
as well as predict increasing problematic behaviors in a child as a long-term psychological
outcome.
Additionally, use of harsh discipline by both parents may also relate to greater
internalizing and externalizing behavior in children. Bender and colleagues (2007) found that
maternal and paternal disciplines predict adolescent problematic behaviors, such as depression
and externalizing. Specifically, paternal discipline was linked to anxiety. It could be
hypothesized that a father’s healthy psychological status and involvement might be associated
with more positive outcomes for adolescents in the context of maternal depression (Brennan, Le
Brocque, & Hammen, 2003)
In conclusion, negative parenting attitudes involving physical and non-physical
punishment by a mother with a psychological disorder may be associated with the child’s
psychological and social development.
Children of mothers with depression and behavior problems
Children’s behavior and their development are influenced by multidimensional factors,
such as parenting practice, parents’ beliefs and mental health, family socioeconomic status,
children’s genetic-biological factors, and neighborhood characteristics. According to
Bronfenbrenner’s theory (1986), an individual family does not exist by itself, but rather involves

13

inter-relationships between a micro-system, meso-system, and macro-system or the hierarchical
environment surrounding that family.
Thus, parental depression, especially in the mother who is the primary caretaker, may be
not always a problem due to the mother alone, but also dependent on the complicated reasons in
the full family structure, home environment, and community. For the same reason, child
developmental outcomes are often affected ecologically by a number of factors at the individual-,
family-, and community- levels. These multidimensional factors directly or indirectly will
influence child developmental issues regarding cognitive, academic, psychological and social
achievements (Pachter, et al., 2006).
Depressive symptoms in a mother and behavioral problems in her children, commonly
tend to accompany each other (Elgar, et al., 2004). Relationships between a depressed mother
and her child are likely to be unengaged, since depressive symptoms are regarded as a
psychological disorder that extensively can influence all interpersonal relationships. Compared
to children with mothers who have no depression, children with depressed mother are fussier
(2004), receive lower cognitive academic scores (Kurstjens & Wolke, 2001) , and show less
sensitivity to their own needs (Ehrle, Moore, & Brown, 1999).
Downey and Coyne (1990) examined children’s problems as well as those of the entire
family and suggested that they derived directly from living with a depressed parent, especially
with a mother with depression. Children with depressed mothers are more likely to reveal
behavioral and emotional problems, social-skill deficits, and lower academic competence,
compared to their peers living with non-depressed mothers. These authors further suggest that
children’s behavior problems should be understood through addressing the more complex
mechanism within each family caused by preexisting conditions, such as family stress and

14

disorder, although there remains an association between maternal depression and children’s
behavior problems(Downey & Coyne, 1990).
While it has been reported that maternal depression is associated with insecure
attachments for infants and toddlers (Goodman & Gotlib, 1999), depression of a mother with
school-aged children is linked mainly to deficits in accomplishing tasks relevant to, and needed
in, the developmental stage in children as well as school problems (Cummings & Davis, 1994).
Thus, developmental outcomes for different children can reveal different coping strategies that
work differently at varying developmental stages (Luoma, et al., 2001). The timing and the
recurrence of maternal depressive symptoms may be vulnerable at a certain developmental stage
due to the unstable psychological status of a primary-caregiver mother.
Positive parent-child relationship qualities can be protective factors for children of a
depressed mother. Brennan and colleagues (2003) found that a high level of perceived maternal
warmth and acceptance was associated with higher levels of positive behavior outcomes for
children with depressed mothers. As perceived, maternal psychological control and emotional
over-involvement levels were lower, and positive behavior outcomes in children were also
higher, suggesting that the mother-child relationship acts as a resource factor relating to higher
functioning outcomes for children.
Mothers with depressive symptoms often have other psychological disorders, such as
antisocial problems (Kim-Cohen, Moffitt, Taylor, Pawlby, & Caspi, 2005) and transmit these to
their children (Hammen & Brennan, 2001; Kim-Cohen et al., 2005). Depression of a child tends
to occur more frequently in a family with a mother who has depression associated with
interpersonal impairment, than in a child in a family with a mother without depression (Hammen
& Brennan, 2001). Additionally, maternal depression is associated with negative parenting

15

practice that influences children’s internalizing and externalizing of behavior problems (Pachter
et al., 2006).
However, any psychological disturbance between the depressed mother and the child is
not limited by only the mother’s psychological disorders and negative parenting practice, and
their effects on children, but may involve a more complex and structural mechanism. For
example, unstable economic status, such as temporal unemployment or long-term chronic
poverty and change of family structure, such as death or divorce, may cause a depressed
condition for parents as well as directly or indirectly a child’s developmental outcome. Thus,
families with stable economic status are likely to have less depressed parents and fewer problem
behaviors with their children, compared to families experiencing poverty.
Theoretical and Conceptual Foundation
Key concepts of social capital
The concept of social capital has been studied extensively in the fields of sociology,
economics, politics, and family studies since achieving general agreement on the plausibility of
the accumulation of social capital affecting the development of civic society, economic
performance, and the development of the child (Arrow, 2000; Coleman, 1988, 1990; Putnam,
2000).
The basic meaning of social capital starts as a source of family support (Portes, 1998;
Coleman, 1988). James Coleman (1988; 1990) described social capital as a resource that families
can draw upon to promote the cognitive abilities and academic success of their children.
Coleman, specifically, emphasizes the functional effects of social capital. He regards social
capital as resources that are to family relations and community social organization (Coleman,
1990). In addition to functional effects of social capital is the concept of reciprocity, which refers

16

to giving and receiving certain valued items, such as information or resources that can increase
positive cooperation among multidimensional social contexts.
According to Coleman (1990), applying social capital theory is particularly useful when
examining the cognitive or social development of a young person and emphasizing the advantage
of social capital in the development of human capital. He elaborated on his definition of social
capital to demonstrate the contribution of social capital to the development of human capital.
Social capital of both family and community plays a key role in formulating human capital for
the next generation which in turn affects children within the family. Although the family has
much human capital, if there is little social capital formed by interactions within the family, and
the family and outside society, it is difficult to expect positive growth of the next generation
which in turn desires successful development for its children.
Alternatively, Bourdieu, the sociologist, defined social capital as the sum of potential
resources that link to the possession of a continuing network whether more of an institutionalized
relationship or less of one. (Baron, Field, & Schuller, 2000). According to Bourdieu (1986), the
amount of social capital a certain individual can use depends on the size of the network and the
capital that the individual possesses, i.e., economic and/or cultural capital. Various types of
capital can be transformed to economic capital, which Bourdieu refers to as the roots of other
types of capital. As economic capital enables individual life to be richer, social capital also
provides benefits to individuals or groups via mutual exchange (Baron, Field & Schuller, 2000).
In this vein, as the size of the network and the amount of capital that a certain family can use
becomes larger, the power of the social and economic capital that affects the family’s and
children’s well-being will also expand more or vice versa.

17

Human development ecology is a complex web of personal relationships, social settings
and institutions that influence developmental trajectories (Crosnoe, 2004). Social capital can
facilitate a better understanding not only of interactions within families, but also of mezzo-level
interactions between families and their surrounding communities and of how both affect the
well-being of children (Ferguson, 2006).
Therefore, empirical research on social capital should be understood in a
multidimensional context of social resources that facilitate positive outcomes as well as reduce
negative behaviors and maltreatment in terms of the development of children and adolescents
(Furstenberg & Hughes, 1995). The individual family does not exist alone, but through interrelationships among micro-system, meso-system, and macro-system, which represents the
hierarchical environments surrounding the family (Bronfenbrenner, 1986).
The ecological perspective examines the concepts of interpersonal relationships with
family and its surrounding environment as shared with the concepts of social capital. Furstenburg
and Hughes (1995) provided a perspective that assumes the integration of both the sociologicaland the psychological- concepts. Concepts of social capital overlap with concepts of
developmental psychology, such as buffering or resiliency, and concepts of sociology, such as
institutional resources and social networks (Benson & Deal, 1995). While Bronfenbrenner’s
ecological perspective focuses on how individuals, specifically children, develop successfully
within the hierarchical system of human development, social capital perspective emphasizes
collectively-shared benefits gained through resources generated by interactional networks that
can bring successful outcomes to children and their families. As such, the direction of the social
capital perspective should focus on characteristics of social capital as group- rather than
individual- property (Mckenzie, Whitley, & Weich, 2002).

18

Economic status and social capital
A number of previous studies have explored how household income is associated with
child development (Duncan & Brooks-Gunn, 1997; Duncan, Yeung, Brooks-Gunn & Smith,
1998). Compared to families with fewer resources, parents with a higher socioeconomic status
can provide more and varied resources to attain greater social and educational achievement for
their children. Furthermore, the stable economic status of parents allows their children to
experience enriching life experiences, various activities, and safe neighborhoods, including a
good school district. Children born into affluent families are raised by parental social and
economic investments that promote their children’s development.
As the amount of parental investment for children is low, their developmental outcomes
would be low as well. The logic of capital regarding which outcomes is directly proportional to
the amount of investment applied by parents to children within the family. Economic investment
of parents in their children enables them to meet desirable needs and opportunities by accessing
other resources to produce better outcomes.
Such parental investments, however, do not always associate with parents’
socioeconomic status (Parcel & Menaghan, 1994). Parcel & Menaghan (1994) examined the
argument that parental socioeconomic standing influences a child’s subjective developmental
outcome. However, if parents do not desire their children to be raised successfully through
money, time, and effort, one would not expect positive development of their children. Thus,
investing possible resources for children’s better outcomes corresponds with providing suitable
home environments for them. Bradley and colleagues (1988) point out that indicators of
socioeconomic standing associate only weakly to quality of children’s home environments. Of
course, if parents do not use their socioeconomic resources for their children, those children’s

19

development will be delayed. However, if parents without enough socioeconomic resources still
use what they do have to the extent that they invest their time, effort, and emotional support in
the child-rearing process, eventually there will be positive social development of those children.
Thus, factors affecting a child’s socialization should consider the dimensions of individual
structure and socioeconomic standing and its by-products generated by the relational process
from the parents.
Negative effects of parents’ low socioeconomic status are transmitted to children
through the home environment, specifically through fewer interactions with parents and lesser
access to material resources in the home. Thus, the quality of the home environment has
important significance for children’s cognitive abilities and academic achievement, as well as for
their psychological well-being and behavior or behavioral problems (Parcel & Menaghan, 1994).
Safe home environments are generated by parents, and the beneficiaries are always their children.
Family social capital and home environment
Coleman (1988) defined family social capital as a byproduct based on the relationship
between parents and their children. It includes time, efforts, resources, and energies that parents
invest in their children. Social capital determines whether parents’ financial and human capital
can influence children’s socialization and is not simply another variable to be added to the
human and financial capital. For example, household income or marital status is a human capital
that may influence socializing of children and their families. However, if parents do not have
enough human capital, those who create appropriate environments that contain intellectual
stimulation by age level, do not provide sufficient levels of parent-child bonding, in particular,
enough levels of maternal warmth. It suggests the importance of home environment in
facilitating both cognitive and social child outcomes (Parcel & Menaghan, 1994). Essentially,

20

social capital plays an important role in the transmission of human capital, specifically, in a
child’s social development.
Stable home environments provided by parents who have sufficient economic resources
yield a cognitively stimulating environment as well as an emotionally warm atmosphere.
Consequently, the family environment strongly impacts children’s cognitive abilities,
educational attainments, and social development (Parcel & Menaghan, 1990, 1994). In terms of
a child’s social development, a strong home environment has been connected to fewer behavior
problems and better emotional growth in children (Bradley et al., 1988; Parcel & Menaghan,
1994). The socioeconomic characteristics of parents affect the child’s social development
through accumulation of the continuous relationship of parent-child as family social capital. That
relates to how parents and children interact mutually, not simply one direct impact on another.
Thus, family financial instability and insufficient material resources may create a feeling of
psychological distress in parents and promote negative parent-child relationships (Parcel &
Menaghan, 1994).
Coleman emphasized that family social capital influences children’s academic
achievements (Coleman, 1990); however, empirical research into the influence of family social
capital focused mainly on children’s cognitive and academic achievement relative to parenting
styles and parent characteristics (Pong, Hao, & Gardner, 2005; Israel et al., 2001).
Children’s problematic behaviors are linked mainly to inappropriate parenting style and
strategy. More positive parenting styles result in fewer problem behaviors of children (Peterson
& Rollins, 1987). In addition, parents under stress are more likely to use harsh discipline and be
both inconsistent and uninvolved parents (Conger, Ge, Elder, Lorenz, & Simons, 1994). These

21

negative parenting behaviors hinder child development and increase the possibility that children
will exhibit behavioral problems (McLeod & Nonnemaker, 2000).
Parenting strategies as a family resource are one of the strongest factors that facilitate
children’s social development. In the maternal mental health or poverty context, there is yet
another predictor for how parents interchange information with other parents to promote
children’s positive outcomes. Parents with school-aged children need relationships with other
parents to exchange important information about school activities and child rearing.
Studies have documented that quality of parent-child interaction and maternal mental
health affect the problematic behavior of children for multiple social contexts (McNeal, 1999;
Harnish, Dodge, & Valente, 1995; Moss, Rousseau, Parent, St-Laurent, & Saintonge, 1998;
Pachter et al, 2006).
McNeal’s study (1999) indicated that parental involvement is a salient factor in
explaining behavioral-, but not cognitive-, outcomes through greater support for parent-child
discussion and involvement in parent-teacher organizations. Also, a specific dimension of
involvement that has greater effect for affluent and white students is the greater level of social
capital. Additionally, Pachter et al., (2006) found that there were ethnic differences when
indicating that chronic poverty affects child behavioral problems indirectly through maternal
depression and neighborhood, and directly via parenting practice.
People, in general, if they can afford to, tend to choose where they live, and parents who
are living within the same neighborhoods are likely to aggregate with similar socioeconomic
status or same ethnic groups. Families are likely to be affected by sharing daily-life experiences
with other families within similar social contexts. Child behavior problems, when used as an

22

indicator of child socialization especially, are actually influenced by parenting strategy within
families and the family relationship with its outside environments.
Neighborhood social capital and child social development
Coleman (1988) defines social capital as a dynamic process of interactions among closed
networks of parents and their children in the family, the school or neighborhood, which foster
agreement on the children’s behavioral norms as well as enforce those norms in relation to the
children’s well-being.
Parents of families with high neighborhood social capital are more likely to be embedded
in surrounding social networks, comprising both immediate- and extended-family support, as
well as participating in local social institutions (Ferguson, 2006). Levels of trust and safety with
respect to neighborhood residents and environments do play a role in determining positive child
socialization. However, it would be difficult to expect children raised in impoverished
environments to achieve the same levels of developmental outcomes in terms of socialization,
academics, and opportunity, as do affluent children.
Previous studies of low-income and minority neighborhoods have hypothesized that
parents will form cohesive social networks via mutual agreement on children’s behavioral norms
and collaborative action and will try to implement those norms for all neighborhood children
(Hank, 2008). A sense of common membership and social norm is shared by these parents
through participation in the social networks they form. Children benefit from social connections
and interactions parents have with others, including neighbors, schools, and work colleagues
(Crosnoe, 2004; Dufur, Parcel, & McKune, 2008; Parcel & Dufur, 2001). These connections and
interactions solidly constitute neighborhood social capital; the stronger the connections, the
greater the resources that children can access (Parcel, Dufur, & Zito, 2010).

23

Residential segregation by family socioeconomic status is more likely to facilitate
intergenerational transmission of poverty (Pebley & Sastry, 2003). Especially, if children are
growing up in a disadvantaged neighborhood with fewer family resources, children’s social and
behavioral outcomes as well as their opportunities for success are restricted by the poor
neighborhood. However, despite low-socioeconomic status of parents, which suggests negative
child development, if parents can create enough neighborhood social capital through intimate
relationships, adult friendships, and common neighborhood values, successful children’s social
development can result.
Sampson, Raudenbush and Earls (1997) emphasize that shared expectations and
involvement of neighborhood residents in active support and social control is key to a positive
neighborhood environment for child social development. Eventually, neighborhood social capital
created by these parents is then transmitted to their children and their outcomes in socialization.
Using family social capital, parents can create various neighborhood social capital entities as a
family resource, and let these resources benefit their children.
Effects of Neighborhood Poverty
Neighborhood poverty and maternal psychological effects
To mothers who are at the child-rearing stage, extra familial support, whether financial or
emotional, plays a salient role in reducing parenting stress and the effects of psychological
disorders (Klebanov et al., 1994).
Recent research shows that depression in mothers may be associated with characteristics
of neighborhoods in which these mothers live (Cutrona et al., 2006). Although neighborhoods do
not affect all people similarly, adverse neighborhoods can intensify the harmful impact of
personal stressors and can interfere with the forming of bonds between people, thus, increasing

24

risks of depression (2006). Ross’ study (2000) on the distressing effects of female headship and
neighborhood poverty examined the stresses in daily life in disadvantaged neighborhoods, and
indicated these are associated with the depression of the family head who is a mother. The
effects of chronic poverty multiply child behavioral problems through experiencing
neighborhood environments and parenting practices and the mother’s mental health (Klebanov et
al., 1994; Pachter et al., 2006).
Studies have examined the association of neighborhood effects and maternal depression
(Cutrona et al., 2006; Hill & Herman-Stahl, 2002; Klebanov et al., 1994; Mitchell & LaGory,
2002; Ross, Reynolds, & Geis, 2000). According to the perspective of cohesiveness, living in
economically poor neighborhoods has negative effects on residents’ psychological well-being
(Ross et al., 2000).Mother’s psychological well-being influences parenting strategies in unsafe
neighborhoods and thus leads to children’s negative development. Physical condition of
neighborhood is crucial for rearing children. In addition, neighborhood disorder may generate
more a daily stress in mothers compared with mothers who reside in non-poor and safe
neighborhoods (Hill & Herman-Stahl, 2002).
While there are few studies about the consequences of neighborhood conditions for adult
psychological well-being on poverty condition, many studies have concentrated on child
developmental achievement through investigating the school environment and peer groups
(Bowen, Bowen, & Ware, 2002; Caughy & O’Campo, 2006; Crosnoe, 2004; Ginter, Havemen
&Wolfe, 2000; Woolley & Grogan-Kaylor, 2006).
Neighborhood poverty and the social development of children
Previous research related to neighborhood effects on child development has examined
children and adolescents living in poor neighborhoods and has revealed that there are fewer

25

positive developmental outcomes in various directions, than for those for children from more
advantaged neighborhoods (Brooks-Gunn & Duncan, 1997; Leventhal & Brooks-Gunn, 2004).
Research also has documented that many mothers living in neighborhoods characterized by
mainly poor families resident there, have less positive psychological outcome for their children
than do families in non-poor neighborhoods (Klebanov et al., 1994). However, little research
has investigated the possible effects of neighborhood context on psychological health in children,
such as depression, anxiety, withdrawal, and other externalizing behavior (Brooks-Gunn, Duncan,
Klebanov, & Sealand, 1993; Kupersmidt, Griesler, DeRosier, Patterson, & Davis, 1995;
Martinez, 1999; Pachter, et al., 2006; Xue, Leventhal, Brooks-Gunn, & Earls, 2005).
Regarding effects of maternal depression, poverty, and neighborhoods on children’s
behavior problems, Pachter and colleagues (2006) explored whether the processes through which
parenting practices, maternal depression, neighborhood, and chronic poverty influence child’s
behavioral problems, produce similar or different outcomes in minority- and nonminoritychildren. Chronic poverty, neighborhood, maternal depression, and parenting practices do have
an effect on children’s behavior problems. Neighborhood effects were present in both the White
and Black samples but were not significant for the Latino sample, suggesting a difference due to
a social stratification mechanism as well as the socio-cultural differences present in family- and
childrearing- practices.
Xue and colleagues (2005) examined whether the level of children’s mental health is
associated with neighborhood characteristics and whether neighborhood social process in terms
of collective efficacy and organizational participation underscores these effects. Disadvantaged
neighborhood condition was associated more with psychological health problems and
neighborhood collective efficacy, while organizational participation was associated with better

26

mental health. Many children in disadvantaged neighborhoods have psychological health
problems, and that mechanism was operated by neighborhood social control and cohesion,
suggesting, presumably suggesting that the collective power of a neighborhood improves the
psychological health of an individual
On the other hand, some studies have implied that a high density of neighborhood social
capital is connected to a high level of behavior problems in children. Caughy and colleagues
(2003) examined the association between attachment to neighborhood as an indicator of social
capital, and the presence of behavior problems of children in a sample of African -American
parents. In their study, general sense of neighborhood was not strongly related to children’s
behavioral problems; instead, knowing neighbors was associated with these behavioral problems.
Specifically, among neighborhoods with low levels of impoverishment, or wealthy
neighborhoods, behavioral problems were highest for those children whose parents did not know
many neighbors, i.e., the parents had low social capital. This finding suggests that excessive
social cohesion may influence children’s behavior problems negatively.
For school-aged children and their parents, resources related to the school a child attends
may be the type of neighborhood social capital that affects the children’s social development.
Parcel and Dufur (2001) examined parental involvement in school activities and indicated that it
likely contributes to positive social adjustment for children, further suggesting that there may be
significant links between family inside and outside the home. Parental involvement for the
betterment of children in school activities is necessary for sharing an interchange of information
among all parents.
With the parenting practice of mothers in poverty, participation in the social network in
which these mothers live may be beneficial for ameliorating the depression level of mothers and

27

help overcome stress due to long- term poverty. The sense of belonging, in social networks,
promotes the bonding of mothers. For example, community participation increases the likelihood
of constructing and maintaining interactive ties in these social networks (Kawachi & Berkman,
2001). Indeed, an individual’s social ties are contingent on structural characteristics, and this is a
key insight for understanding social capital theory. Therefore, an important task for such
empirical investigation is to demonstrate social capital’s contextual influence on individual
networks and their support, and on level mental health outcomes for individuals.

28

CHAPTER III
Method

This chapter provides a brief description of the dataset, sample selection criteria, model
design, variable description, and analytical procedures.
Description of Dataset
The present study used a secondary data analysis, drawing data from the Los Angeles
Family and Neighborhood Survey - L.A. FANS. By collecting longitudinal data on
neighborhoods, families, and children, as well as on residential choice and neighborhood change,
L.A. FANS was designed to treat many of the problems that have limited previous research on
neighborhood effects (Peterson, Sastry, Pebley, Ghosh-Dastidar, Williamson, & Lara-Cinisomo,
2004). The dataset was conducted from April 2000 to January 2002. Adults and children living
in 3085 households were interviewed from a diverse, stratified probability sample of 65
neighborhoods based on a census tract throughout Los Angeles County.
L.A. FANS questionnaires for the survey included modules for household, adult, parents,
primary caregiver, child, and cognitive assessment. The adult questionnaire collected detailed
information on neighborhood definition, participation, interaction, and perception of
characteristics. L.A. FANS subsequently developed and incorporated a Behavior Problem Index
and included parent-child interaction into the parent questionnaire. In addition, this dataset
includes the depression- and self-assessment- questionnaire in the primary caregiver module.
L.A. FANS data are based on a stratified random sample using census tracts because
tracts generally include children attending two or more elementary schools. For examining both

29

neighborhood and schools’ effects on children’s development, using census tracts as sampling
units provides a broader perspective for researchers (Peterson et al., 2004)
Interestingly, the stratified sampling design adapted to this dataset is helpful to obtain an
over-sample of poor- and very poor- tracts which provide a relatively large number of
respondents in poor households. Thus, L.A. FANS data is based on the percentage of the tract’s
population in poverty, classifying neighborhoods corresponding to tracts that are very poor
(those in the top 10% of the poverty distribution), poor (tracts in the 60th-89th percentiles), and
non-poor (tracts in the bottom 60% of the distribution) (Peterson et al., 2004). In addition, by
including a sample of neighborhoods across the entire income range, the data provide for
comparisons of parents and their children between poor areas and non-poor areas.
Fifty households within each tract were selected from the 65 sample tracts. In each
sampled household, one adult respondent was sampled at random (i.e., randomly selected adult
[RSA]) as was one child respondent (i.e., randomly selected child [RSC]). In households with
children, the mother of the RSC was selected as the respondent and was designated the primary
caregiver (PCG). Mothers among selected RSAs within each household were also regarded as
PCGs. If the RSC had one or more siblings, ages 17 or younger, who shared the same biological
or adoptive mother and the same PCG, one randomly selected sibling was selected for interview
(designated as sibling [SIB]) (Peterson et al., 2004). A total of 9378 addresses were randomly
selected from the L.A. FANS sample; 4110 households were released while 3085 cases were
completed for rosters, except when the respondent could not be contacted. Of the 3085
households with a completed roster, 25% (777 cases) were households without children and 75%
(2308 cases) were households with children (Peterson et al., 2004).

30

Sample for the Study
Data gathered from the primary caregivers (PCG) are generally mothers from the selected
RSAs or RSAs who are different people within the RSA, but remain as primary caregivers for
the current study. The total number of primary caregiver respondents was 1641, but this study
only included respondents who completed the PCG module and women primary caregivers.
After excluding male primary caregivers and respondents who did not complete the PCG module,
the number of primary caregivers was 1546. Relating to the questionnaire for parent-child
interaction and child behavior problems, participants for the current study are limited to women
primary caregivers who have school-aged children. In addition, children over 15 are excluded
from the range of the child module because the same type of questionnaire for parent-child
interaction was designed only for children aged 3 to 15. Thus, PCGs of children who are not of
school-age, and those over 15, were excluded from the current study. Overall, through the
process of data reduction, the total number of selected participants was 1062.
Since one crucial part of this study is focused on how child behavior problems are
moderated by a mother’s depression and neighborhood poverty provided by census tract
information, all participants were categorized in four groups according to depression and
neighborhood poverty level. For depression level, every woman primary caregiver was asked the
question “During the past 12 months, was there ever a time when you felt sad, blue, or depressed
for two weeks or more in a row?”, for which 316 (29.8%) answered “yes” and 746 (70.2%)
answered “no.” In addition, for neighborhood poverty level, the number of PCGs who are in
poor neighborhood was 726 (68.4%) and those not in poor neighborhood was 336 (31.6%)
through the classification from census tracts. Estimates of census tract level for calculating
percentage in poverty were developed by Los Angeles County’s Urban Research Division

31

(URD) using state- and county- administrative data. (Peterson et al., 2004) By these two
variables, which refer to depression and poverty level, all participants were divided into 4 groups.
Of these women primary caregivers, 229 PCGs who are mothers living with depression
and in poor neighborhoods (21.6% of total sample), 87 are living with depression but in nonpoor neighborhoods (8.2%) 495 are not living with depression but in poor neighborhoods
(46.6%), and 249 are living without both depression and in poor neighborhoods (23.5%).
Measurements
Dependent variable
Child behavior problems The Behavior Problems Index (BPI) was designed to assess
children’s behavior problems, including anxiety, depression, and aggression (Peterson & Zill,
1986). The BPI instrument includes 26 items divided into two subscales: internalizing and
externalizing. The internalizing score provides a measure of the presence of withdrawn- and sadbehaviors and the externalizing score provides a measure of the presence of aggressive- and
other related- behaviors. The total score provides an overall measure of problem behaviors,
including responses of both internalizing and externalizing. The L.A. FANS dataset included
four additional items pertaining to older children from the National Longitudinal Survey of
Youth - NLSY. However, to narrow the scope of this study, child behavior problems were
limited to the internalizing and externalizing index.
Primary caregivers were asked to provide information on their school-aged children’s
behavior, including whether children revealed a particular behavior problem (BPI). Primary
caregivers responded to the BPI questions using a three-point Likert scale which indicated how
true each statement was of their child. Items for child behavior problems were rated ranging 0 =
not true to 2 = often true; a lower number indicates fewer behavior problems. To compute the

32

three primary BPI scales, the coding of items was reverse-recoded before categorizing
measurement items to create scores. In addition, child behavior problem indices were grouped
together by conducting exploratory factor analysis and item parceling to create scales of
internalizing (withdrawn and sad etc.) and externalizing (aggressive and angry etc.) behaviors.
The reliability of the two primary scales from the original dataset is as follows: internalizing
= .73, externalizing = .85.
Independent variables
Independent variables include household income level as a predictor, social cohesion and
interaction with neighbors from neighborhood social capital, and parent-child emotional
interaction and behavioral responsiveness from family social capital as mediators. In addition,
mother’s depression and poverty levels were employed for comparison across the four groups.
Household income: Household income includes the sum of the earning of the respondent,
the respondent’s spouse/partner and any co-resident children of the respondent and
spouse/partner. It is reported by the amount that they had earned during the previous year.
Social cohesion: The social cohesion measure was informed by Sampson and colleagues
(1997) as a conception of collective efficacy through the perception of mutual trust and shared
willingness to help others. By previous research, collective efficacy consists of both willingness
to intervene and expectations for social control, social cohesion and trust. The current study
includes five items related to individuals’ perceptions of social cohesiveness and trust with
neighborhoods (Cohen, Inagami, & Finch, 2008). These five include “This is a cohesive or
unified neighborhood,” “People around here are willing to help their neighbors.” “People in this
neighborhood generally don’t get along with each other,” “People in this neighborhood do not
share the same values,” and “People in this neighborhood can be trusted.” Indicators for

33

measuring social cohesion and trust are rated on a 5-point Likert scale ranging from 1 = strongly
agree to 5 = strongly disagree. Once the items are recorded, they will be summed to determine
social cohesion and trust. While the original L.A. FANS data did not report reliability of the
items related to neighborhood collective efficacy, the validity and reliability of these items were
examined by previous research (Cohen, Inagami, & Finch, 2008). The five measuring indicators
were used in this study after reducing the number of indicators by combining conceptually
similar items.
Interaction with neighborhoods: The indicator for interaction with neighborhoods is
defined as how often people in their neighborhood do favors for one another, such as watching
each other’s children, helping with shopping, or lending household tools. In addition, this
measurement reflects, conceptually, the respondent’s perceived sense of membership, shared
emotional connection, and degree of mutual influence with her neighborhood (Caughy, O'Campo,
& Muntaner, 2003). This variable includes three items such as “about how often do you and
people in your neighborhood do favors for each other? For example, watch each other’s children,
help with shopping, and lend gardening or house tools, would you say?” “When a neighbor is not
at home, how often do you and other neighbors watch over his or her property, would you say?,”
and “How often do you and other people in the neighborhood ask each other’s advice about
personal things such as child rearing or job openings, would you say?”
Items for interaction with neighborhoods were rated on a 4-point Likert scale ranging
from 1 = often to 4 = never. These indicators of perception for interaction with neighborhoods
were used by recording them.
Mother-child emotional interaction quality: Quality of interaction between primary
caregiver and child was assessed from the home environment inventory, which specifies

34

questions for degrees of emotional interaction by responses of primary caregivers for their
school-aged children. Three questions for mothers’ emotional interaction were employed for this
measurement; “In the past week, about how many times have you praised your child for doing
something worthwhile?,” “In the past week, have you shown your child physical affection (for
example: kisses, hugs, stroking hair, etc.)?” and “In the past week, how many times have you
told another adult (for example: spouse, friend, co-worker, visitor, relative) something positive
about your child?” The indicators were rated on a 4-point Likert scale ranging from 1= never to
4=almost every day and also used by recording them.
Discipline responsiveness of mother: Quality of behavior responsiveness between
primary caregiver and child was assessed from the home environment inventory, which specifies
questions for degrees of behavioral responsiveness by the responses of primary caregivers for
their school-aged children. Three questions for mothers’ discipline responsiveness were
employed for this measurement; “In the past week, how many times have you had to spank your
child?” “In the past week, how many times have you grounded your child?” and “In the past
week, how many times have you sent your child to his/her room or another room as a
punishment?” Indicators were rated on a 4-point Likert scale ranging from 1=never to 4= almost
every day and also used by recording them.
Moderator variables
To explore differences among the four groups, the primary caregivers poverty and
depression levels were employed as criteria to categorize participants with four groups.
Neighborhood poverty level: Neighborhood poverty stems from economic stressors such
as unemployment and lack of affordable housing. Living in a poor neighborhood is more likely
to lead to psychological illnesses, such as depression and anxiety; thus, it is considered an

35

important risk factor for psychological illness (Kuruvilla & Jacob, 2007). Neighborhood poverty
level of family was divided into three strata which refer to very poor, poor and non-poor from
census tracts information about poverty based on the percentages of the tract’s population in
poverty. The current study categorized two levels of neighborhood poverty based on the
classification for neighborhood poverty.
Feeling depression of mother: The depression variable served as an indicator of whether
a mother had major depression and was informed from the Composite-International Diagnostic
Interview-Short Form (CIDI-SF), a diagnostic instrument that L.A. FANS used and that is an
international protocol adopted by the World Health Organization (Kessler, Andrews, Mroczek,
Ustrun, & Hu, 1998). For the current study, a question asking whether or not the primary
caregivers felt any symptoms of depression such as being sad, blue, or depressed during the past
two weeks was employed for the criterion of grouping.
Hypothesized Model
Figure 1 presents a hypothetical framework driven from the theoretical model, based on
literature review for this study, which illustrates the causal relationships of economic statuses of
mothers on the child’s psychological well-being through neighborhood- and family social
capital- factors.
Household income level is associated with child’s behavior problems which consist of
internalizing and externalizing behaviors, through parent-child emotional interaction and
mother’s discipline responsiveness as well as neighborhood variables in terms of social cohesion
and interaction with the neighborhood. It is hypothesized that each group categorized by
maternal depression and neighborhood poverty level will be mediated by mediating factors,

36

mother’s depression and neighborhood poverty level. In addition, causal relationships among
variables will be moderated by mother’s depression and neighborhood poverty levels.

Figure 1 Hypothesized Model

Depression level,
Neighborhood poverty level

Social cohesion
Interaction with neighbor.

Household
Income level

Internalizing
Externalizing

Emotional interaction
Behavioral response

Based on the proposed hypothesized model, the main purpose of the current study is to
examine paths from the household income level to behavior problems of their children. A child’s
behavioral development is affected by the stressful status within the family, but the effects would
be varied by different mediating factors. It would depend on which mediating factor and how the
amounts of interaction among mothers, children, as well as neighborhoods are accessed within

37

each group membership. Thus, since this study has hypothesized that paths from household
income level to child’s behavior problems differ, according to the mother’s psychological status
and neighborhood poverty level. By exploring these pathways, the current study promotes
understanding of how these relationships predict a child’s positive development.
Research Questions and Hypotheses
Based on the stated purpose and literature, the current study aims to address the following
research questions and hypotheses.
Research Question 1: Examine the mediation effects of neighborhood and family factors
in causal effects of household income level on a child’s behavior problems within the four
groups. Are there significant indirect effects of the four mediator variables between household
income and children’s internalizing behavior problems within the four groups?
Hypothesis 1.1. A child’s internalizing problems in a mother with depression group living
in poor neighborhood will be mediated by the quality of discipline responsiveness of the mother.
As the quality of discipline responsiveness of mother is higher, the child’s internalizing problems
will be lower.
Hypothesis 1.2. A child’s internalizing problems in a mother with depression group living
in non-poor neighborhood will be mediated by the amount of emotional interaction between
mother and child. As the amount of emotional interaction between mother and child is higher,
the child’s internalizing problems will be lower.
Hypothesis 1.3. A child’s internalizing problems in a mother without depression group
living in poor neighborhood will be mediated by the amount of social cohesion of the mother
whereas the amount of social cohesion of mother is higher, the child’s internalizing problems
will be lower.

38

Hypothesis 1.4. A child’s internalizing problems in a mother without depression group
living in non-poor neighborhood will be mediated by the amount of interaction with the
neighbors of the mother whereas the amount of interaction with neighbors of the mother is
higher, the child’s internalizing problems will be lower.
Research Question 2: Examine the mediation effects of neighborhood and family factors
in causal effects of household income level on a child’s behavior problems within the four
groups. Are there significant indirect effects of the four mediator variables between household
income and child’s externalizing behavior problems within the four groups?
Hypothesis 2.1. A child’s externalizing problems in a mother with depression group
living in poor neighborhood will be mediated by the quality of discipline responsiveness of the
mother whereas the quality of discipline responsiveness of the mother is higher, the child’s
externalizing problems will be lower.
Hypothesis 2.2. A child’s externalizing problems in a mother with depression group in
non-poor neighborhood will be mediated by the quality of emotional interaction between the
mother and the child whereas the amount of emotional interaction between the mother and the
child is higher, the child’s externalizing problems will be lower.
Hypothesis 2.3. A child’s externalizing problems in a mother without depression group in
poor neighborhood will be mediated by the amount of social cohesion of the mother whereas the
amount of social cohesion of the mother is higher, the child’s externalizing problems will be
lower.
Hypothesis 2.4. A child’s externalizing problems in a mother without depression group
living in non-poor neighborhood will be mediated by the amount of interaction with neighbors of

39

the mother whereas the amount of interaction with neighbors of the mother is higher, the child’s
externalizing problems will be lower.
Research Question 3: Examine moderated mediation effects of family income level and
mediators on child’s behavior problems; internalizing and externalizing. Do the mediating effects
of neighborhood and family factors significantly differ across the four groups?
Hypothesis 3.1. The effects of family income and mediators on a child’s internalizing
problems will be moderated by the mother’s depression and neighborhood poverty levels.
Hypothesis 3.2. The effects of family income and mediators on a child’s externalizing
problems will be moderated by the mother’s depression and neighborhood poverty levels.
Data Analysis
Analytical method
The current study was based on a multiple mediator path analysis with structural equation
model - SEM (MacKinnon, 2008). The SEM using Mplus software (ver. 5.2), was used as the
primary analytical method for the present study. This method consists of a measurement model
to define hypothetical latent constructs in terms of measured variables and a structural model to
draw causal relationships among latent constructs. SEM is a multivariate statistical technique
considered to be a combination of factor analysis, regression, and path analysis (Kline, 2005).
Unlike other multivariate statistical methods, an important characteristic of SEM is its focus on
latent variables. Latent variables are not directly observed or measured, but studied indirectly
through indicators (e.g., observed variables).
SEM consists of two dimensions: measurement model, and structural model. The
measurement model involves specifying the relationship between indicators and the latent
variable they represent. As an integral part of SEM, the development of the measurement model

40

assesses how well the observed variables (indicators) reflect the theoretical constructs (latent
variables). Thus, testing the measurement model using confirmatory factor analysis (CFA) helps
assess measurement issues such as validity and reliability. SEM involves specifying the testing
of causal relationships among latent variables. As such, it is similar to path analysis, but with
latent variables (Kline, 2005).
For several reasons, SEM is particularly useful for testing and refining a conceptual
model and its accompanying proposed hypotheses. First, SEM does not examine the relationship
solely between independent and dependent variables, but among all latent variables
simultaneously. SEM also is advantageous in that it allows for more flexible assumptions over
multiple regressions, especially when multicollinearity is suspected. Finally, SEM allows
researchers to assess measurement error in the model estimation process. Unlike other
multivariate statistic methods, which do not separate measurement error from hypothesized
relationships among constructs, SEM can examine the relationship among constructs that are not
influenced by measurement errors.
Analysis procedures
As the hypothesized model is already outlined, the observed and latent variables selected
for the initial measurement model are presented in Table 1.

41

Table 1
Variable Structure
Latent variable

Observed variable

Internalizing

IN1: felt complaining, too dependent
IN2: felt depressed, inferior, withdrawn, too much crying
IN3: easily confused, anxious, fearful, clinging to adult

Externalizing

EX1: stubborn, tells lies, highly tense, overly active, lack of attention
EX2: strong temper, impulsive, bully, trouble with peers, not sorry
EX3: too much argue, disobedient, lack of attention
SO1: cohesiveness with neighborhoods
SO2: willing to help each other, get along with each other
SO3: sharing same value, trust in each other
NE1: do favors for each other (watching child, shopping, lending tools)
NE2: watch over each other’s property when not at home, advice about personal
things (child rearing or job opening)
EI1: praise child for doing something
EI2: physical affection (kiss, hug, stroking hair)
EI3: telling other adults something positive about child
BR1: spanking during last week
BR2: grounding child
BR3: sending child to a room

Social cohesion

Interaction with neighbor

Emotional interaction

Behavior responsiveness

Total household income

INC: household income level

42

First, a multiple confirmatory factor analysis (CFA) with latent variables was conducted
to determine the validity of each subscale indicator. Before conducting multiple CFA,
Exploratory Factory Analysis (EFA) and item parceling were conducted for two dependent
variables by using 26 observed indicators of internalizing and externalizing symptoms. In
addition to EFA, item parceling refers to combining measured variables into a set of several
variables, either by summing or by averaging items (Bandolas, 2002) to apply for more
parsimonious measurement model.
A multiple CFA was tested to examine whether 17 observed indicators access to the same
construct across groups. Six latent variables in terms of social cohesion, interaction with
neighbors, emotional interaction between parent and child, behavioral responsiveness with parent
and child, and children’s behavioral problems of internalizing and externalizing were included in
the process of a multiple CFA.
A full structural model was tested to examine mediation effects between household
income level and children’s internalizing and externalizing symptoms according to the
categorized four groups. A design that I employed for this analysis is multiple mediator path
analysis (Mackinnon, 2008). Including several mediators in the same model has the advantage of
determining the relative consequence of the specific mediation effects associated with all
mediators, rendering the possibility to compare theoretical backgrounds employed for the
mediation model (Preacher & Hayes, 2008).
In the current study, indirect effects of the four mediators between household income
level and children’s behavior problems were tested. If the direct effect is not significant, but the
indirect effect is significant, then the proposed latent variables are supposed as mediators. The
complete mediation effect is defined as the case in which the independent variable does not

43

affect the dependent variable after the mediator has been controlled. Finally, the direct path is
zero, although theoretically the amount of reduction is also called the indirect effect. In this point,
a multiple mediator analysis was used, but conceptually, based on Baron and Kenny’s method
(Baron & Kenny, 1986).
To ensure significance of model fit, several indices for overall model fit were assessed.
The goodness of fit between the hypothesized model and the sample data were assessed using
several goodness-of-fit indices as individual fit indices can each be associated with distinct
biases. Chi-square statistics are among the most commonly used techniques to examine overall
2

model fit. A non-significant goodness-of-fit statistic (χ ) is needed as this indicates that the
implied covariance matrix is nearly identical to the observed data (Kline, 2005). Model fit
indices have been developed to supplement the Chi-square statistics due to the extreme
2

sensitivity of χ related to a sample size. Second, the Root Mean Square Error of
Approximation (RMSEA) was used to determine fit since it is less likely affected by sample size
(Raykov & Marcoulides, 2006). RMSEA may be an appropriate cut off, suggesting less than or
equal to .05 for decision of good fit (Hu & Bentler, 1999), but acceptable between.05 and .1 as
model-data fit (Deng, Doll, Hendrickson, & Scazzero, 2005). In addition, there is the
Comparative Fit Index (CFI), the Tucker Lewis Index (TLI) and the Standardized Root Mean
squared Residual (SRMR). Fit is considered acceptable if the CFI and TLI are greater than or
equal to .90, and the SRMR is below .05 (Kline, 2005). The M-plus software provides several fix
2

indices—specifically, such as Chi-square (χ ), SRMR, CFI, TLI, and RMSEA.

44

Treating missing data
In the analysis with the second dataset, one of the most sensitive and considerable
limitations is how to treat missing data. On a given measured variable used in this study, missing
rate in selected data are between .2% and 36.8% through all categorized groups. For this analysis,
the current study used Full Information Maximum Likelihood (FIML) to address issues with
missing data.
The goal of analysis with missing data is to minimize statistical bias while
simultaneously maximizing use of available information (Allison, 2008; Barnes, Mallinchrodt,
Lindborg, & Carter, 2008). According to Allison (2008), data may be missing for several
different reasons, specifically, missing completely at random (MCAR), missing at random
(MAR), or not missing at random (NMAR). Data are MAR if the probability that Y is missing is
not dependent on the value of Y and are randomly distributed upon observed data. Although it is
the best possible case because it indicates that the data may be regarded as a random subset of
the original target sample in the case of missing completely at random, most research assumes
missing at random.
In multivariate analysis, listwise deletion is used to analyze missing data. However,
listwise deletion is problematic for many reasons. Deleting a large number of cases may result in
a loss of power in analyzing. In addition, the listwise method will likely bring statistical bias if
the data are not missing completely at random, but missing at random (Barnes et al., 2008).
Full Information Maximum Likelihood (FIML) is designated with the default method for
addressing issues with missing at random (MAR) data, when data are run using the Mplus
program. All approaches will bias results if missing data are not ignorable. However, FIML
estimates tend to be less biased and also are “the recommended parameter estimation method

45

when data are missing in structural equation model analyses” (Schumacker, 2004). Since FIML
estimates use all available information from data to specify model parameters, this is the best
possible way to address missing at random (MAR) data.

46

CHAPTER IV
Results

This chapter includes the results from descriptive statistics of the sample, data reduction
analysis for dependent variables and bivariate correlations to ensure characteristics of the four
mother’s groups as a primary analysis. According to the literature review and the hypotheses, a
structural equation model was used to examine significance for the current study to examine
research questions 1, 2 and 3 and their hypotheses. Testing the effects of moderators based on
research question 3 and its hypotheses was conducted as a last step of this analysis.
Descriptive Statistics
Descriptive analysis of the sample population across the four groups was conducted
(Table 2). Number of total sample was 1,062 but 2 of them were excluded from the analysis due
to system missing in the process of grouping. Of the total sample of 1,060, 21.6% - i.e., 229
mothers - were depressed and living in poor neighborhood, 8.2% - i.e., 87 - were depressed but
not living in poor neighborhood, 46.6% - i.e., 496 - were non-depressed but living in poor
neighborhood, and 23.5% - i.e., 249 - were both non-depressed and living in non poor
neighborhood. The average of mothers and their children indicates that they are in a similar
range across the four groups. However, socio-economic status and race among the four groups
indicate that mother’s characteristics differ across the groups. L.A. FANS dataset provides an
indicator to determine neighborhood poverty level by the percentage of the census tract’s
population in neighborhood poverty, classifying population corresponding to tracts that are nonpoor (coded 1), poor (coded 2), and very poor (coded 3). For this study, three levels of
neighborhood poverty provided by the dataset were dichotomized with two levels of non-poor

47

(coded 1) and poor (coded 0). Mother’s depression was also provided with two levels of
depressed (coded 1) and non-depressed (coded 5).
Mother’s educational levels also have revealed a difference among the four groups. The
percentage of mothers who earned a college, professional or graduate school level was higher in
non- poor neighborhood groups (41.4% in depressed/non-poor neighborhood; 50.2% in nondepressed/non-poor neighborhood) than in poverty groups (24.0% in depressed/poor
neighborhood; 27.7% in non-depressed/poor neighborhood).
Mothers especially who are not in depression and living in non-poor neighborhood group
showed the highest education level. The percentage of mothers who graduated at the secondary
school level was higher in mother’s groups living in poor neighborhood (59.0% in
depressed/poor neighborhood ; 58.2% non-depressed/poor neighborhood) than groups in nonpoor neighborhood (37.9% in depressed/non-poor neighborhood; 16.9% in non-depressed/nonpoor neighborhood), suggesting that mothers living in poor neighborhood possess a lower level
of formal education.
Racial distribution among the four groups indicates that there is a difference by groups:
around half the total sample population included Latinos (50.2%). The ratio of Asian/Pacific
Islanders and Whites were more dominant in the mother’s group who are depressed living in
non-poor neighborhood (25.3%, 31.0%) and the mother’s group who are not in depression living
in non-poor neighborhood (21.7%, 34.5%), while the majority of mothers in depressed/poor
neighborhood and non-depressed/poor neighborhood groups consisted of much more Latina
(74.7%, 68.1%), suggesting that racial distribution is related to neighborhood poverty status for
this data set.

48

Interestingly, the percentage of mothers with 3 or more children was higher in the group
living in poor neighborhood (31.0% in depressed/poor neighborhood; 30.1% in nondepressed/poor neighborhood) - regardless whether they were depressed or not - than in the
group living in non-poor neighborhood (12.6% in depressed/ non-poor neighborhood; 18.2% in
non-depressed/non-poor neighborhood). In sum, while the ratio of secondary school is higher in
mothers living in poor neighborhood, relatively higher level of education such as college,
professional and post graduate school was higher in mothers living in non-poor neighborhood.
That means mothers and their children living in poor neighborhood are in deficit of resources
which make more positive development and family functioning possible, compared to mothers
and children living in non-poor neighborhood. Thus, factors connecting economic status with
children’s behavior development would differ across groups, because of these different
characteristics of the groups.

49

Table 2
Descriptive Statistics of Sample Population
Total

N(%)
Average age of respondent
Average of child
Ratio of more than 3
children
Household income
Mean
SD
Education
Below elementary
Elementary
Secondary
College or professional
school
Post graduate level
Marital status
Married
Living with a partner, not
married
Not living with a partner,
not married
Race
Asian/Pacific Islander
White
Latino + Black

1,060 (100%)
37.09
9.81
26.2%

Depressed/
Poor
neighborhood
229(21.6%)
36.35
9.38
31.0%

Depressed/
Non-poor
neighborhood
87(8.2%)
38.39
9.61
12.6%

Non-depressed/
Poor
neighborhood
495(46.6%)
37.31
10.03
30.1%

Non-depressed/
Non-poor
neighborhood
249(23.5%)
36.93
9.82
18.2%

$49,405.92
$35,641.50

$43,268.93
$23,161.64

$54,179.04
$43,043.89

$43,858.23
$25,919.93

$51,040.04
$37,487.70

47(4.5%)
137(12.9%)
498(47.0%)
279(26.3%)

10(4.4%)
29(12.7%)
135(59.0%)
35(15.3%)

4(4.6%)
14(16.1%)
33(37.9%)
28(32.2%)

18(3.6%)
52(10.5%)
288(58.2%)
93(18.8%)

15(6.0%)
42(16.9%)
42(16.9%)
123(49.4%)

99(9.3%)

20(8.7%)

8(9.2%)

44(8.9%)

27(10.8%)

668(62.9%)
109(10.3%)

151(65.9%)
17(7.4%)

51(58.6%)
6(6.9)

307(62.0%)
61(12.3%)

158(63.5%)
25(10.0%)

285(26.8%)

61(26.6%)

30(34.5%)

127(25.7%)

66(26.5%)

94(8.9%)
124(11.7%)
159(15.0%)

4(1.7%)
5(2.2%)
42(18.3%)

22(25.3%)
27(31.0%)
0(0.0%)

14(2.8%)
6(1.2%)
115(23.2%)

54(21.7%)
86(34.5%)
0(0.0%)

50

Table 2 (cont’d)
Descriptive Statistics of Sample Population
Latino
533(50.2%)
White + Other
152(14.3%)

171(74.7%)
7(3.1%)

51

8(9.2%)
30(34.5%)

337(68.1%)
23(4.6%)

17(6.8%)
92(36.9%)

Exploratory Factor Analysis and Item Parceling on Dependent Variables
Before using the 26 observed indicators in terms of internalizing and externalizing
behavior problems, Exploratory Factor Analysis (EFA) and item parceling were conducted to
reduce the number of parameters estimated for observed items. Item parceling refers to
combining measured variables into a set of composite measures by either summing or averaging
the items (Bandalos, 2002).
Item parceling should be considered when a construct has a large number of measured
variable indicators. The best parcels are formed by items that display approximately the same
covariance, which should lead them to have approximately the same factor loading estimates
(Holt, 2004). When the number of extracted factor is one, the factor can have, at least, 3 items
(Holt, 2004), thus all measurement items are divided into 3 items evenly by ordering the value of
factor loading.
Using EFA with the original 11 indicators for internalizing behaviors and 15 indicators
for externalizing items (Table 4), each factor was extracted from two dependent variables. Item
parceling was conducted by combining measured variables into 3 items (Table 3). (For detail
contents for items, see appendix.)

Table 3
Parceling from EFA Results
Combined items by parceling
IN1: p2, p15, p26
Internalizing IN2: p20, p14, p24, p21
IN3: p16, p8, p5,p23
EX1: p18, p4, p3, p17, p25
Externalizing EX2: p19, p13, p9, p12, p11
EX3: p6, p10, p1, p7, p22
Note: Each indicator by parceling was computed by having similar factor loadings

52

Table 4
EFA result for Dependent Variables
Internalizing

Externalizing

Internalizing

Eigenvalue

Eigenvalue

Indicator

Factor loading for
retained factor

1.978
.610
.173
.135
.093
-.021
-.063
-.101
-.156
-.254
-.296

4.391
.460
.386
.228
.164
.048
.007
-.037
-.043
-.089
-.120
-.135
-.172
-.222
-.240

#2
#5
#8
#14
#15
#16
#20
#21
#23
#24
#26

Externalizing

.558
.329
.456
.484
.379
.485
.546
.288
.260
.417
.349

Indicator

Factor loading for
retained factor

#1
#3
#4
#6
#7
#9
#10
#11
#12
#13
#17
#18
#19
#22
#25

.490
.603
.617
.668
.481
.534
.629
.368
.453
.618
.474
.685
.664
.270
.415

53

Correlations among Variables for Groups
Correlations among variables were conducted by SPSS 17.0 software program (Tables 5
~ 8). Parent-child emotional interaction and behavioral response of mothers were highly
associated with children’s both internalizing and externalizing behavior problems through all
groups, suggesting that as the amount of emotional and behavioral interaction increases,
children’s behavior problems in terms of internalizing and externalizing symptoms reduce. For
only the mothers who are depressed and living in poor neighborhood (r =.168, p<.05; r = .137,
p<.05) household income level was associated with both internalizing and externalizing behavior
problems.
Social cohesion with mothers and neighborhoods was associated with household income
level in depressed/poor neighborhood and non-depressed/poor neighborhood mothers.
Interestingly, while there was positive association between social cohesion and household
income level in the group of mothers who are depressed and living in poor neighborhood
mothers (r = .138, p<.05), there was negative association between social cohesion and
household income in the group of mothers who are not depressed but living in poor
neighborhood (r = -.086, p<.05) which means, as the amount of social cohesion is weaker,
household income level is higher. In addition, social cohesion had a negative association with a
child’s internalizing behavior problems (r = -.237, p<.05), suggesting that weaker social
cohesion reduces internalizing behavior problems.

54

Table 5
Correlation among Variables for Depressed Mothers/Poor Neighborhood Group
1
Internalizing
1
Externalizing
.645***
Social cohesion
.017
Neighborhood
.018
interaction
Emotional interaction
.181**
Discipline response
.267***
Household income
.168*
Mean
14.06
SD
4.440
Note: p<.05* p<.01** p<.001***

2

3

1
.015

1

.000

.501***

.197**
.387***
.137*
18.08
2.722

4

5

6

7

1

-.035
.067
.138*
7.13
1.051

-.035
.004
.004
4.36
1.579

55

1
.082
-.049
2.09
1.494

1
.112
1.12
1.225

1
43,268.930
23,161.638

Table 6
Correlation among Variables for Depressed Mothers/Non-poor Neighborhood Group
1
Internalizing
1
Externalizing
.499***
Social cohesion
-.237*
Neighborhood
-.026
interaction
Emotional interaction
.221*
Discipline response
.167
Household income
.124
Mean
14.08
SD
3.071
Note: p<.05* p<.01** p<.001***

2

3

1
.052

1

.059

.426***

.302**
.527***
.000
17.94
4.583

4

5

6

7

1

.174
.128
.103
7.17
1.137

.030
.007
-.041
4.36
1.658

56

1
.193
.191
2.07
1.697

1
.005
1.06
1.340

1
54,179.040
43,043.888

Table 7
Correlation among Variables for Non-depressed Mothers/Poor Neighborhood Group
1
Internalizing
1
Externalizing
.624***
Social cohesion
-.052
Neighborhood
.001
interaction
Emotional interaction
.204***
Discipline response
.234***
Household income
.014
Mean
14.61
SD
3.230
Note: p<.05* p<.01** p<.001***

2

3

1
-.080

1

-.013

.378***

.203***
.443***
.014
18.30
4.267

-.079
-.013
-.086*
7.06
1.027

4

5

6

1
.022
-.007
2.18
1.476

1
.010
1.13
1.358

7

1
-.080
-.035
-.042
4.27
1.561

57

1
43,853.230
25,919.931

Table 8
Correlation among Variables for Non-depressed Mothers/Non-poor Neighborhood Group
1
Internalizing
1
Externalizing
.749***
Social cohesion
-.096
Neighborhood
-.036
interaction
Emotional interaction
.187**
Discipline response
.322***
Household income
-.065
Mean
14.23
SD
3.148
Note: p<.05* p<.01** p<.001***

2

3

1
-.048

1

-.031

.367***

.190**
.397***
-.008
18.30
4.573

-.023
-.059
.076
7.07
1.052

4

5

6

7

1
-.069
-.111
.026
4.48
1.586

58

1
.050
-.084
1.97
1.305

1
.003
1.13
1.281

1
51,040.040
37,487.695

Measurement Equivalence of Confirmatory Factor Model across Four Groups
Before fitting the structural equation model to examine direct and indirect causal
relationships, multi-group confirmatory factory analysis of the measurement model was
conducted. Multiple-sample confirmatory factor analysis (MCFA) is concerned with
measurement invariance, i.e., whether a set of indicators assesses the same constructs in different
groups (Kline, 2005). Testing a model’s structure or its individual parameters for equivalence
across subgroups using confirmatory factor analysis, provides explicit evidence whether
measures of a construct are consistent with the theoretical understanding for this study of the
nature of that construct. If the invariance test of the measurement model is not statistically
significant across subgroups, any examination of structural parameters could not be substantiated.
Thus, any research findings would be highly suspect, due to variance across subgroups.
In the current study, measurement invariance test was conducted for 17 observed items,
to assess the same constructs of all four subgroups which refer to depressed mothers/poor
neighborhood, depressed mothers/non-poor neighborhood, non-depressed mothers/poor
neighborhood, and non-depressed mothers/non-poor neighborhood (Figure 3).
Sequence of equivalence testing
Measurement equivalence using confirmatory factor analysis (CFA) is tested
hierarchically at several different levels. First, observed items load on the same constructs (latent
variables) across groups, which is commonly termed the configural model (Byrne, 2008). As a
baseline model against which all subsequent tests for equivalence are compared, a configural
model tests for equal measurement parameters of the observed variables, thus is linked to the
unobserved (latent) variables. These parameters include factor loadings, error variances, and

59

error covariances. In addition, testing all steps of measurement equivalence should include the
equality of the observed variable intercept.
The initial step in testing for group equivalence requires that the same number of factors
and their loading pattern be the same across groups (Byrne, 2008). As no equal constraints were
imposed on the parameters, the hypothesis is that each subgroup retains the same type of
measurement model. Overall, chi square statistic is required to assess congruence between data
and the theoretical model. It is sensitive to sample size and departures from multivariate
normality (Byrne, 2008; Deng, et al., 2005).
Comparing Model 3 with Model 2 produced a significant chi-square difference value, χ

2

(33) = 63.04, p < .001. A chi-square test is sensitive to sample size; however, other model fit
indices should be considered. Additionally, a ratio of chi-square values to degrees of freedom
was less than 2 indicating that Model 3 would better represented the observed data than Model 2
(Byrne, 1991) Although the chi-square difference test showed a significantly worse model fit
with Model 3 than Model 2, other model fit indices indicated that Model 3 well represented the
observed data, CFI=.925, TLI=.934, RMSEA=.049, and 90% CI= [.042, .055].
Comparing Model 3 with Model 4, Model 4 yielded a worse model fit to the data than
Model 3, χ

2

(51) = 179.93. The ratio was larger than 2, indicating that fitting Model 3 to the

observed data was better than fitting Model 4 (Byrne, 1991). Additionally, other practical fit
indices were considerably worse in Model 4 than in Model 3; TLI (.934 vs .906 for Model 3 and
Model 4, respectively), CFI (.925 vs. .904), and RMSEA (.049 vs. .055). Therefore, Model 3
better represented the observed data than Model 4.
By the results of multi-group CFA, since three hierarchical invariance tests yielded
significant results, measurement equivalence for the four groups used in the current study proved
60

that 17 observed indicators are significantly assessed to the same six constructs in the four
groups. It is also said to be possible to examine all parameters from the structural equation model.

Table 9
Measurement Equivalence of Confirmatory Factor Models of Six-Factor Solutions across
Four Groups
Model

χ

2

(d.f.)



χ2

(  d.f.)

p

TLI

CFI

RMSEA

90% CI

Model 1 680.70(416)

n/a

n/a

.942

.924

.049

[.042, .056]

Model 2 722.21(449)

41.51(33)

.147

.940

.928

.048

[.041, .054]

Model 3 785.25(482)

63.04(33)

.001

.934

.925

.049

[.042, .055]

Model 4 965.18(533)

179.93(51)

<.001

.906

.904

.055

[.050, .061]

Note: Model 1 included a model in which no equal constraints on parameter estimates were
imposed. Model 2 included a model in which factor loadings were assumed to be equal across
the four groups. Model 3 included a model in which factor loadings and intercepts were
constrained to be equal across the four groups. Model 4 represented a model in which factor
loadings, intercepts, and residual variances were constrained to be equal across the four groups.

61

SO1
Figure 2 Measurement Model

SO2

SO3

.577*** .767*** .318**
SOCH

NE1
.647***
*
.57***

NE2
.706***

.57***

INT

NEIG

.632*** IN1
*
.631***
IN2
*
.57***

.624***
INC

.793***
EXT
.781***
EMOT
.779***
.736***

.626***

.575***

EM1

EM2

EM3

RESP
.586***
.57***

*

RE1

.593***
*
RE3

.544***
RE2

.57***

62

IN3
EX1
EX2
EX3

Table 10
Results of Measurement Equivalence
Latent variable
Internalizing
IN1
IN2
IN3
Externalizing
EX1
EX2
EX3
Social cohesion
CO1
CO2
CO3
Interaction with neighbor
NE1
NE2
Emotional interaction
EI1
EI2
EI3
Behavior responsiveness
BE1
BE2
BE3

Factor loading

Intercept

.632
.631
.624

3.756
4.500
3.854

.793
.771
.789

3.353
3.699
3.592

.577
.767
.318

4.601
4.494
5.149

.647
.706

2.616
2.213

.736
.575
.626

1.242
.576
1.334

.586
.544
.593

.477
.700
.718

Note: Factor loadings and intercepts were drawn from Model 3.

Structural Equation Modeling
There are multiple steps to Structural Equation Modeling (SEM) including model
specification, identification, estimation, testing and modification (Schumacher and Lomax, 2004).
Since measurement equivalence of the theoretical model supported that all 17 observed
indicators of the four groups were assessed to the same constructs significantly, the next steps
include mediation analysis and multi-group comparison across the four groups with the full

63

structural model. In addition, to ensure the significance of mediation effects across the four
groups, a bootstrapping test was conducted.
Mediation testing
I hypothesized that each characterized mother group will have a different mediator from
the neighborhood and family factors and its mediation effects between household income and
child’s behavior problems will be significant within the four groups. I also hypothesized that
depressed mother groups will be more influenced by family social capital factors in which the
group of mothers who are living in poor neighborhood will be more influenced by neighborhood
social capital factors. To examine research questions 1 and 2, and their hypotheses based on
multiple mediator path model (MacKinnon, 2008), mediation effects of social cohesion,
interaction with neighbors, parent-child emotional interaction, and parent-child behavior
responsiveness between household income and children’s behavior problems relating to
internalizing and externalizing symptoms, were considered. In addition, the parameter estimated
in the multiple mediator path analysis is based on the concept of mediation developed by Baron
and Kenny (1986).
According to this method, the following conditions are necessary for establishing
mediation using SEM: Step 1) a model in which household income predicts the children’s
internalizing and externalizing symptoms fit the data; Step 2) a model in which household
income predicts mediators (social cohesion, interaction with neighbor, parent-child emotional
interaction, and parent-child behavior responsiveness) which in turn predicts children’s
internalizing and externalizing symptoms fit the data; Step 3) pathways between household
income and the four mediators and between the four mediators and internalizing and
externalizing symptoms are significant and in the directions predicted. Overall, household

64

income is hypothesized to exert indirect effects on children’s internalizing and externalizing
symptoms through the mediators. In the current study, mediation analysis with full structural
model was conducted within the four groups. Each of the proposed paths was individually tested
in SEM and should be significant as required to test for mediation steps 1 through 3.
Group with depressed mothers/poor neighborhoods
For internalizing symptoms: As hypothesized, results indicate that the quality of
discipline responsiveness of a mother and child in the group of depressed mothers living in poor
neighborhoods mediates between household income and both internalizing symptoms. (Step 1;
INCINT: β = .052, p=.537, Step 2; INCRESP: β = .164, p=.050 and Step 3; RESPINT: β
= .391, p =.000). Indices for model fit indicate significant results as χ

2

(90) = 111.492, CFI

=.956, TLI=.949, RMSEA=.032 and SRMR=.058, which means there are mediating effects of
discipline responsiveness of mothers between household income level and a child’s internalizing
behavior problems.
For externalizing symptoms: As hypothesized, results indicate that the quality of
discipline responsiveness with a mother and child in the group of depressed mothers living in
poor neighborhoods mediates between household income and both internalizing symptoms. (Step
1; INCEXT: β = .026, p=.720, Step 2; INCRESP: β = .169, p=.043 and Step 3;
RESPEXT: β = .612, p=.000). The indices for model fit indicate significant results as χ

2

(90)

= 106.488, CFI =.976, TLI=.972, RMSEA=.028 and SRMR=.055; this means there are
mediating effects of discipline responsiveness of mothers between household income level and a
child’s externalizing behavior problems.
Hypothesis 1.1. “A child’s internalizing problems in a mother with depression group
living in poor neighborhood will be mediated by the quality of discipline responsiveness of the
65

mother. As the quality of discipline responsiveness of mother is higher, the child’s internalizing
problems will be lower.” was supported. Hypothesis 2.1. “A child’s externalizing problems in a
mother with depression group living in poor neighborhood will be mediated by the quality of
discipline responsiveness of the mother whereas the quality of discipline responsiveness of the
mother is higher, the child’s externalizing problems will be lower.” was also supported.
Group with depressed mothers/non-poor neighborhoods
For internalizing symptoms: As hypothesized, the results indicate that the amount of
emotional interaction with a mother and child acts as a mediator between household income and
both internalizing symptoms for the group of depressed mothers living in non-poor
neighborhoods. (Step 1; INCINT: β = .071, p=.594, Step 2; INCEMOT: β = .280, p=.015
and Step 3; EMOTINT: β = .290, p =.043). Indices for model fit indicate significant results as

χ 2 (90) = 111.231, CFI =.916, TLI=.903, RMSEA=.052 and SRMR=.094, which means there
are mediating effects of mother-child emotional interaction between household income level and
a child’s internalizing behavior problems.
For externalizing symptoms: As hypothesized, results indicate that the amount of
emotional interaction with mother and child acts as a mediator between household income and
externalizing symptoms for the group with depressed mothers living in non-poor neighborhoods.
(Step 1; INCEXT: β = -.034, p=.741, Step 2; INCEMOT: β = .278, p=.016 and Step 3;
EMOTEXT: β = .244, p=.035). Although indices for model fit were a little lower than those of
other groups because of relatively small number of sample population, as

χ 2 (90)

= 129.202,

CFI =.892, TLI=.874, RMSEA=.071 and SRMR=.100, this means there are mediating effects of
mother-child emotional interaction between household income level and a child’s externalizing
behavior problems.
66

Hypothesis 1.2. “A child’s internalizing problems in a mother with depression group
living in non-poor neighborhood will be mediated by the amount of emotional interaction
between mother and child. As the amount of emotional interaction between mother and child is
higher, the child’s internalizing problems will be lower.” was supported. Hypothesis 2.2 “A
child’s externalizing problems in a mother with depression group in non-poor neighborhood will
be mediated by the quality of emotional interaction between the mother and the child whereas
the amount of emotional interaction between the mother and the child is higher, the child’s
externalizing problems will be lower.” was also supported.

67

Figure 3 Structural Model for Depressed mothers/Poor neighborhood Group
SO1
SO1

SO3
SO3

SO2
SO2

SOCH
.062

a

NE1

.096

NE2

b

.095

.010

a

-.021

NEIG

IN1

-.117
INT

b

.009

.119

.052
INC

IN3

.178

.026

EX1
EXT

a

.200**

-.009

-.011

b

IN2

EMOT

EX2
EX3

.391***
a

EM1

.164 *

EM2

EM3

.612***

11

b

.169 *
*

2

Internalizing: χ (90)=111.492
CFI= .956 TLI= .949 RMSEA=.032 SRMR=.058
2
Externalizing: χ (90)=106.488
CFI=.976 TLI=.972 RMSEA=.028 SRMR=.055
Note: a =internalizing b=externalizing

RESP

RE1

RE2

RE3
68

Figure 4 Structural Model for Depressed mothers/Non-poor neighborhood Group
SO1

SO2

SO3

SOCH
-.627*

a

NE1

.002

NE2
1

b

-.097

a

IN1

-.226

-.001

.399

NEIG

INT

IN2

b

-.097

.071

INC

.290*

-.034

EX1
EXT

.244*

b

.278 *
*

IN3

.248

EMOT

a

.280 *

EX2
EX3

.269

a

.016

EM1
b

.006

EM2

EM3

11

2

RESP

RE1

RE2

.756**
*

RE3
69

Internalizing: χ (90)=111.231
CFI= .916 TLI= .903 RMSEA=.052 SRMR=.094
2
Externalizing: χ (90)=129.202
CFI=.892 TLI=.874 RMSEA=.071 SRMR=.100
Note: a =internalizing b=externalizing

Group with non-depressed mothers/poor neighborhoods
For internalizing symptoms: It was hypothesized that the amount of social cohesion will
mediate between household income level and children’s internalizing problems but there were no
significant mediating effects. (Step 1; INCINT: β = .012, p=.812, Step 2; INCSOCH: β = .141, p=.011 and Step 3; SOCHINT: β = -.118, p =.214). Indices for model fit indicate
significant results as χ

2

(90) = 151.448, CFI =.942, TLI=.932, RMSEA=.037 and SRMR=.043,

which means there are no mediating effects of the social cohesion factor.
For externalizing symptoms: Results indicate that the amount of social cohesion with
mothers and their neighborhoods in acted as a mediator between household income and both
internalizing symptoms for non-depressed mothers living in non-poor neighborhoods. (Step 1;
INCEXT: β = -.024, p=.610, Step 2; INCSOCH: β = -.141, p=.011 and Step 3;
SOCHEXT: β = -.185, p=.024). Indices for model fit indicate significant results as χ

2

(90) =

165.184, CFI =.949, TLI=.941, RMSEA=.041 and SRMR=.040, which means there are
mediating effects of social cohesion between household income level and child’s externalizing
behavior problems.
Hypothesis 1.3. A child’s internalizing problems in a mother without depression group
living in poor neighborhood will be mediated by the amount of social cohesion of the mother
whereas the amount of social cohesion of mother is higher, the child’s internalizing problems
will be lower.” was not supported. Hypothesis 2.3. A child’s externalizing problems in a mother
without depression group in poor neighborhood will be mediated by the amount of social
cohesion of the mother whereas the amount of social cohesion of the mother is higher, the child’s
externalizing problems will be lower.” was supported.

70

Group with non-depressed mothers/non-poor neighborhoods
For internalizing symptoms: It was hypothesized that the amount of interaction with
neighborhoods will mediate between household income level and child’s internalizing problems,
but there were no significant mediating effects. (Step 1; INCINT: β = -.083, p=.250, Step 2;
INCNEIG: β = -.027, p=.713 and Step 3; NEIGINT: β = .065, p =.585). Indices for model
fit indicate significant results as χ

2

(90) = 107.947, CFI =.967, TLI=.962, RMSEA=.028 and

SRMR=.054, which means there are no mediating effects of interaction with neighborhoods
between household income level and child’s internalizing behavior problems.
For externalizing symptoms: It was hypothesized that the amount of interaction with
neighborhoods will mediate between household income level and child’s externalizing problems
but there were no significant mediating effects. (Step 1; INCEXT: β = -.010, p=.881, Step 2;
INCNEIG: β = -.028, p=.709 and Step 3; NEIGEXT: β = .075, p=.470). Indices for model
fit indicate significant results as

χ 2 (90)

= 107.596, CFI =.977, TLI=.973, RMSEA=.028 and

SRMR=.051, which means there are no mediating effects of interaction with neighborhoods
between household income level and child’s internalizing behavior problems.
Hypothesis 1.4. A child’s internalizing problems in a mother without depression group
living in non-poor neighborhood will be mediated by the amount of interaction with the
neighbors of the mother whereas the amount of interaction with neighbors of the mother is
higher, the child’s internalizing problems will be lower.” was not supported. Hypothesis 2.4. A
child’s externalizing problems in a mother without depression group living in non-poor
neighborhood will be mediated by the amount of interaction with neighbors of the mother
whereas the amount of interaction with neighbors of the mother is higher, the child’s
externalizing problems will be lower.” was not supported.
71

Figure 5 Structural Model for Non-depressed mothers/Poor neighborhood Group
SO1

SO2

SO3

SOCH
-.118

a

-.141 **

NE1

NE2
1

IN1

b

-.141 **
-.027

b

a

-.027

.140

NEIG
.012

INC

-.185*
INT

IN3

.144
.283**
*

-.024
b

EX1
EXT

.256***

.012

.013

EMOT

a

IN2

EX2
EX3

.390***

a

-.023

EM1

EM2

EM3

.613***

b

-.023

2

Internalizing: χ (90)=151.448
CFI= .942 TLI= .932 RMSEA=.037 SRMR=.043
2
Externalizing: χ (90)=165.184
CFI=.949 TLI=.941 RMSEA=.041 SRMR=.40
Note: a =internalizing b=externalizing

RESP

RE1

RE2

RE3
72

Figure 6 Structural Model for Non-depressed mothers/Non-poor neighborhood Group
SO1

SO2

SO3

SOCH
-.088

a

.070

NE1

NE2
1

IN1

.075

b

.070

-.027
b

-.028

.065

NEIG

a

INT

-.083

INC

IN3

.119
.240**

-.010

EX1
EXT

.204**

b

-.035

-.034

a

EMOT

a

IN2

EX2
EX3

-.004

.515***
b

-.014

EM1

EM2

EM3

.593***

11

2

Internalizing: χ (90)=107.947
CFI= .967 TLI= .962 RMSEA=.028 SRMR=.054
2
Externalizing: χ (90)=107.596
CFI=.977 TLI=.973 RMSEA=.028 SRMR=.051
Note: a =internalizing b=externalizing

RESP

RE1

RE2

RE3
73

Significance of Mediation Effects
In addition to testing mediation effects by examining the structural equation model, the
significance of mediation effects should be tested by examining the power with statistical
methods such as bootstrapping. The bootstrap method refers to random sampling with
replacement from the original sample number so that a new sample of number observations is
obtained (MacKinnon, 2008; Preacher & Hayes, 2008). The first step when using bootstrapping
is the creation of a sample of size N from the given sample using random sampling with
replacement. This means that values for individuals are randomly chosen from the existing data
set, repeatedly, without excluding already selected data, until a new sample of size N exists,
created by data from the original sample. This process should be conducted by repeating, at least
1000 times and thus the mediated effects are estimated.
Therefore, from the data, 1,000 bootstrap samples should be used to compute 95%
confidence intervals to determine whether indirect effects were significantly different from zero,
indicating statistically significant mediation (Shrout & Bolger, 2002). In other words, the null
hypothesis for mediation is that the indirect effect of the independent variable on the dependent
variable via mediator is zero. By bootstrapping, using confidence intervals, if zero is included in
between these confidence limits, it could be said that the null hypothesis is not rejected,
indicating statistically non-significant results.
According to bootstrapping procedure, testing significance for mediation effects was
conducted for the current analysis. Results indicated that the indirect effect of behavioral
responsiveness of mother on child’s internalizing and externalizing symptoms in the mother’s
who are depressed and living in poor neighborhood mother group was significant (95% CI =
[.001, .010] for internalizing , [.001, .022] for externalizing). (Table 11) The indirect effect of

74

parent and child emotional interaction on a child’s internalizing symptoms in the mothers who
are depressed but living in non-poor neighborhoods was significant (95% CI = [.001, .036]), but
not statistically significant for externalizing symptoms (92% CI = [-.002, .037]). Bootstrap
analysis also confirmed that indirect effects of social cohesion for externalization with mothers
who are not in depressed and living in poor neighborhoods were significant (95% CI =
[.001, .006]).
In sum, the mediation effect of a mother’s behavioral response between household
income level and child’s both internalizing and externalizing behavior problems with the mothers
who are depressed and living in poor neighborhood group was statistically significant. For the
group with depressed mothers living in non-poor neighborhoods, there were significant indirect
effects of household income level on a child’s internalizing behavior problems via parent-child
emotional interactions, but no significant effects for externalizing behavior problems. Last, for
the group with non-depressed mothers living in poor neighborhoods, there were significant
indirect effects of household income level on a child’s externalizing behavior problems via social
cohesion.

75

Table 11
Significance of Mediation
Group

Non-depressed/Poor neighborhood

CI (95%)

Significance

Internalizing

.003

[.001, .010]

Yes

.008

[.001, .022]

Yes

Internalizing

.009

[.001, .036]

Yes

Externalizing

Depressed/Non-poor neighborhood

Estimates

Externalizing

Depressed/Poor neighborhood

Dependent variable

.010

[-.002, .037]

No

Externalizing

.002

[.001, .006]

Yes

76

Multi-Group Comparison (Moderated Mediation Testing)
For answering research question 3, whether there are moderated mediation effects across
the four groups, multi-group analyses were used to compare the fit of the hypothesized mediation
model for the designated four groups - depressed mothers /poor neighborhoods, depressed
mothers/non-poor neighborhoods, non-depressed mothers/poor neighborhoods, and nondepressed mothers/non-poor neighborhood groups. These multiple group analyses each
compared the fit of two models (i.e., constrained vs. unconstrained) for each of the proposed
mediators (social cohesion, interaction with neighbors, mother-child emotional interaction,
mother’s behavioral responsiveness) and for each of the proposed moderators (mother’s
depression and neighborhood poverty level).
In the first step of these analyses, all parameters (path coefficients, intercepts, error
variances, etc.) were constrained across the moderators. This means that parameters for the level
of moderators were conducted the same for these analyses. The second step of the analyses
should be retained as the basic form of the model, but allowed the parameter values to differ
across the four groups as the level of hypothesized moderators (unconstrained). If moderation is
established, the next step is to examine the pathways among the variables across the four groups
to determine which differ. Overall, the procedure of these analyses is used to reveal any evidence
of moderation.
A chi-square difference test is then conducted to determine whether data fits significantly
better across constrained or unconstrained conditions. Evidence of a better fit for the
unconstrained model would provide support for the moderation hypothesis. As described in the
previous section on model fit statistics, the following indicators are considered the indices of

77

good model fit: RMSEA having a value of less than .05, SRMR having a value of .08 or less, and
a value of more than .90 of CFI and TLI.
Results of multiple group analyses, testing possible differences in moderated meditational
pathways for internalizing behavior problems, failed to examine the evidence of moderation. For
internalizing behavior problems, comparing Model 1 (constrained) and Model 2 (unconstrained)
produced significant chi square difference value, 

χ 2 (26) = 44.76, p < .007, indicating

unacceptable as a structural model with moderating effects (Table 12). Additionally, other model
fit indices were considerably worse in both models, CFI (.829 vs. 825) TLI (.830 vs. .837),
RMSEA (.060 vs. .061).

78

Table 12
Equivalence of Structural Models for Internalizing Behaviors
Model
Model 1
(constrained)
Model 2
(unconstrained)

χ

2

(d.f.)



χ2

(  d.f.)

p

TLI

CFI

RMSEA

90% CI

820.15(418)

n/a

n/a

.830

.829

.060

[.054, .066]

775.80(392)

44.76(26)

.007

.837

.825

.061

[.055, .067]

Note: Model 1 included a model in which equal constraints on regression coefficients were imposed across
the four groups. Model 2 represents a model in which regression coefficients among factors varied across
the four groups.

79

For externalizing behavior problems, comparing Model 1 (constrained) and Model 2
(unconstrained) produced insignificant chi square difference value, 

χ 2 (26) = 44.76, p < .146,

indicating acceptable as a structural model with moderating effects (Table 12). Because a chisquare test is sensitive to sample size, however, other model fit indices should be considered.
Other model fit indices from Model 1 and Model 2 were considerably worse in both models, CFI
(.874 vs. 866) TLI (.874 vs. .875), RMSEA (.061 vs. .063).
In sum, a mother’s depression and neighborhood poverty levels do not have moderation
effects of household income level on child’ behavior problems for both internalizing and
externalizing through four types of mediators, although it was supported that some groups have a
mediator different from the hypotheses of the previous section. That means the effects of a
different mediator within the four groups were examined by mediation analyses through full
structural model, but moderated mediation effects with constrained and unconstrained models to
compare the four groups by moderators were not supported.

80

Table 13
Equivalence of Structural Models for Externalizing Behaviors
Model
Model 1
(constrained)
Model 2
(unconstrained)

χ

2

(d.f.)



χ2

(  d.f.)

p

TLI

CFI

RMSEA

90% CI

833.63(418)

n/a

n/a

.874

.874

.061

[.055, .067]

804.21(392)

44.76(26)

.146

.875

.866

.063

[.057, .069]

Note: Model 1 included a model in which equal constraints on regression coefficients were imposed
across the four groups. Model 2 represents a model in which regression coefficients among factors
varied across the four groups.

81

CHAPTER V
Discussion and Implication

Discussion
This study examined the role of mediators between household income and children’s
internalizing and externalizing behaviors. Four groups categorized by mother’s depression and
neighborhood poverty level were also examined. Results for this study suggested that children’s
behavior problems depend upon the characteristics of each group.
The mother’s discipline response to the child’s misbehavior was a significant mediator of
household income and both internalizing and externalizing children behavior in the group of
mothers with depression living in poor neighborhoods. Emotional interaction between mother
and child was a significant mediator between household income level and child’s internalizing
behavior problems, in the group of mother with depression but living in non-poor neighborhoods.
Interestingly, both depressed mothers living in neighborhoods of poverty and depressed
mothers living in non-poor neighborhoods revealed similar results. In households with
depression, as their income rose, the child’s behavior problems were also high through more
negative discipline responses of mothers, such as spanking, time-out, and withdrawal of privilege
for child’s behaviors. Likewise, higher household income was associated with higher child
behavior problems through lower level of emotional interaction including hugging, praise, and
affection for children. That means despite their income, depressed mothers have fewer positive
emotional interactions with their children, suggesting that this condition leads to behavior
problems. For the non-depressed mother’s group in poor neighborhood, higher quality of social
cohesion acted as a mediator between lower household income and lower child’s externalizing

82

behavior problems. Overall, problematic behaviors of children with depressed mothers were
associated more with parenting related mediators. On the other hand, child’s behavior problems
of mothers living in poor neighborhood were mediated by neighborhood-related factors.
Discipline response of mothers
Results indicated that discipline response of mothers with depression living in poor
neighborhood significantly mediated between household income and children’s internalizing and
externalizing behaviors behavior. Children’s behavior problems depended on how their mothers
responded; spanking, time-out, and withdrawals of privileges increase children’s behavior
problems. This supports previous findings that negative parenting strategy using physical
punishment is not effective in promoting children’s positive behaviors (Barry, Dunlap et al.
2009; Straus & Mouradian 1998;). In addition, a depressed mother is likely to demonstrate
inadequate parenting and poor quality interactions which then facilitate behavior problems in
their children (Kim-Cohen, Moffitt et al. 2005).
A mother’s unstable psychological status disrupts the environment in which children are
raised. As such, children of mothers who are depressed have several behavior problems. Since a
mother with depression has a negative self-perception, a high level of self-punishment
(Goodman & Gotlib, 1999) as well as an antisocial personality (Kim-Cohen et al., 2005), she
may experience troubles in interactions with other family members, specifically, her children.
Thus, a mother who experiences depression is more likely to respond harshly to children’s
misbehavior; this results in more problematic or antisocial behavior (Kim-Cohen et al., 2005).
Antisocial behaviors1 are included as externalizing behavior indicators, including lying, never

1

L.A. FANS dataset have created six behavior subscales-antisocial, anxious/depressed,
hyperactive, headstrong, dependent, and peer conflicts- using internalizing and externalizing
83

apologizing, bullying, and disobedience. It may be difficult for a depressed mother to provide
emotionally consistent support for her children. A depressed mother is characterized by biased
self-perception and high level of self-punishment so that she interacts negatively with others
(Goodman & Gotlib, 1999). Children without emotional support from their parents may have
trouble adapting to school or to other situations. However, the children with enough emotional
support of parents are able to find a solution when confronted with conflicts (Davis & Woitach,
2008).
When using physical punishment, effective parenting should be consistent and firm
(Baumrind, 1996); otherwise it may evoke fear, anxiety, and anger in children (Kim-Cohen et al.,
2005). It would be difficult to construct a positive parent-child relationship when a child has a
fear and avoidance of parents. If physical punishment does lead children to avoid their parents,
the bonds of trust between the child and parent are weakened (Gershoff, 2002).
Finally, this study’s findings affirm research that shows children who experience harsh
discipline from a depressed mother show anti-social behaviors, such as cheating, lying, or being
mean to others. The child does not seem sorry after misbehaving. This study was limited in its
information to examine the severity of the mother’s depression to examine its relationship with
externalizing behavior. Likewise those disabilities like conduct disorder or antisocial personality
traits were not accessible in this study, however interesting such explorations would have been to
the study.
This study is not only about the association between children’s behavior problems and
their mother’s depression, but on depression as a mediator connecting the path between
household income level and children’s behaviors. The findings from this study suggest that

scales from NLSY data (Codebook, p.88). Antisocial behaviors include four items from
externalizing problems in this study.
84

children who are raised in economically deprived environments may generate more problematic
behaviors, but the problem behaviors may depend on how psychologically stable a mother is.
However, if a mother has severe depressive symptoms, thus making it difficult for her to rear her
child, practical assistance to improve psychological health of both a depressed mother and her
child would be needed. Additionally, predicting the degree of seriousness of depressive
symptoms in mothers would also be helpful to plan customized intervention programs.
It is important to understand how depression was defined in this dataset. From the L.A.
FANS dataset, the CIDI-SF questionnaire provides a probability-of caseness score ranging
from .0 to 1.0. One study (Lara-Cinisome & Griffin, 2007) considered a score which is equal to
or greater than .55 as “probability of caseness” which needs clinical treatment. Of 316 mothers
who answered “yes” for feeling a depressive symptom, 145 mothers were equal to or greater
than .55 which indicates “probability of caseness” for dysphoric depressive symptom. It suggests
that the range of maternal depression in the data include both chronic and mild symptoms. It is
also possible to assume that children’s behavior problems would differ between mothers who
needed clinical treatment and mothers who feel daily stresses, relatively less depressive. Because
psychiatric disorders in parents are transmitted biologically or environmentally to
psychopathology of young offspring (Najman, Williams, Nikles, Spence, Bor, & O’Callaghan,
2000), exposure to mothers’ depression is likely to be associated with childhood behavior
problems. It is plausible that children with mothers who need clinical treatment are at greater risk
of behavior problems. For example, mothers who have psychological disorders, such as
depression, aggression, or hostility are likely to use harsher parenting practices. A mother’s
aggressive and hostile attributions reinforce children’s negative behavior (Nix et al, 1999;
Patterson et al., 1992). Neighborhood poverty exacerbates the stress of a mother with depression,

85

thus makes her likely to practice negative parenting. As a mediator between household income
and children’s behavior problem, the discipline response of mother was effective in the group of
mothers with depression.
Mother’s depression is not always caused by economic hardships or difficult home
environments, but a harsh environment to rear their children. Neighborhood poverty is likely
associated with less maternal warmth and responsiveness (Klebanov, Brooks-Gunn, & Duncan,
1994). Living in an impoverished neighborhood might influence the warmth displayed to
children. Since mothers want their children to adjust to a harsh environment, their parenting style
may be stricter and more authoritarian, especially, for poor families (McLoyd, 1990). Thus,
harsh neighborhood combined with family poverty can trigger depressive symptoms, but in
general, it tends to be combined with other stressful life events. For instance, financial hardship
by unemployment (Cutrona et al., 2006; Kessler et al., 1998), widowhood and divorce (Aseltine
& Kessler, 1993; Umberson et al., 1992) and chronic marital difficulties (Gotlib & McCabe,
1990) provide the strongest evidence about the effects of stressful life events on depressive
symptoms.
In conclusion, the significant effect of mothers’ discipline response to children’s
misbehaviors supports that negative parenting attitudes were associated with increasing a child’s
overall behavior problems, anxiety, depression, dependence, peer conflicts, hyperactivity,
stubbornness, and antisocial conduct.
Emotional interaction of mother and child
In this study, in the group of depressed mothers living in non-poor neighborhoods,
emotional interactions between a mother and child significantly mediated the relationship
between household income and child’s internalizing behavior problems. This finding suggests

86

that as household income declined, the child’s internalizing problematic behaviors were also low
through less negative emotional interaction. This finding supports the importance of supporting
positive parenting practices.
Although the importance of the quality of parent-child interaction and maternal mental
health in behavior problems in school-aged children has emphasized (Hammen et al., 2004;
Hammen, Burge, & Stansbury, 1990; Harnish et al., 1995; Moss et al., 1998), there is less
attention about how depressed mothers communicate with their children. If a mother has chronic
depressive symptoms, her depression would affect her ability to have intimate interactions with a
child. Furthermore, if mothers can not receive sufficient emotional support from their spouses,
other family members or friends, the quality of emotional interactions with their children would
also be weaken in parenting.
In general, the quality of emotional interactions has been a significant factor in explaining
social development in infancy, and extra-familial resources and peer influence have been show to
affect social development of school aged children and adolescents. However, the finding of this
study suggests that the quality of the interactions between children and their parents is still an
important factor in promoting the socialization of school aged children, not only during early
childhood and infancy. Furthermore, mothers with depression are more likely to interact poorly
with children because they are more self-focused and have negative self-image (Goodman &
Gotlib, 1999) than mothers without depression. Finally, a mother’s depression will affect a
child’s psychological health relating to internalizing behavior problems.
Social Cohesion
For mothers who are non-depressed but living in poor neighborhood, social cohesion
(intimacy and closeness with their neighborhood) was a significant mediator between household

87

income and children’s externalization behavior problems. By the finding, as household income
level declines, social cohesion with neighborhoods increases, and thus was associated with fewer
externalizing behavior problems. This result supports previous studies (Leventhal & BrooksGunn , 2000; Odgers, Moffitt, Tach, Sampson, Taylor, & Matthews, 2009) that some
economically-deprived neighborhoods have important assets such as social cohesion and mutual
trust that may have a protective effect on children’s problematic behavior, although a poor
neighborhood can be a dangerous place for children. This finding supports institutionalization
of such activities as play groups for young children, wraparound services for at risk populations
in the natural setting, and utilization of natural supports for students’ Individual Education Plans
and Transition plans.
Neighborhood poverty imposes stress on families, and often includes social and
structural mechanisms. For instance, long-term unemployment often originates from nationwide
social and economic problems. Financial and human capital, such as assets, occupation, and
education level possessed by parents significantly influence their children’s socialization.
However, children of parents with high socioeconomic standing do not always generate a
high level of positive developmental accomplishment. Successful parenting depends on how a
primary caregiver attends to the successful developmental achievement of a child and how the
caregiver participates positively in parenting practice. In this vein, mothers’ self-efficacy or
esteem is important in facilitating positive parenting practice. A mother with a more confident
self-image will be more effective in promoting children’s social development.
A mother’s psychological health can be understood in the social context (Cutrona et al.,
2006; Pachter et al., 2006). These findings address the importance of positive parental interaction
with their child. Individual self-efficacy may change to group efficacy through relationships

88

and closeness as well as sharing value with significant others within a homogeneous group.
Parents with more positive self-efficacy would be more able to overcome stressful situations
although they experience the same level of stressful events. If people who live in a
homogeneous environment share collective trust and efficacy, they may achieve treatment effects
as a group (Odgers et al., 2009).
Children’s aggression, disobedience, or lying, specifically in a poor environment, may be
mitigated by neighborhood social capital accumulated by parents. According to Caughty and
colleagues (2003), families with a higher household income generate less neighborhood social
capital, thus their children have more behavior problems. It suggests that neighborhood social
capital is more needed for families living in poor environments. It is not important how many
neighbors parents know, but how regularly parents in the same neighborhood area meet and
watch each other’s children. In addition, parents need to construct relationships with their
neighbors so that they can share information on effective parenting practice. In this context, as a
family resides in the same place for a long time the cohesiveness and sense of belonging in the
neighborhood would increase within the family, suggesting the accumulation of neighborhood
social capital. This type of perspective is in keeping with the emergence of community
gardening in neighborhoods of poverty across the country, in particular Detroit, city with high
poverty. As a result, parents can expect positive social development in their children. In addition,
geographic and residential mobility can disrupt cohesion and connectedness at the extra- familial
level (Hagan, MacMillan, & Wheaton, 1996). The negative effects of residential mobility are
more significant for families with uninvolved fathers and unsupportive mothers because the
children experience simultaneously both the reduction of family social capital by unsupported

89

parents and loss of neighborhood social capital due to residential mobility. The role of fathers in
social capital in this study would have been interesting, although this was not explored.
One way to improve children’s behavioral problems in a poor/non depressed context is to
increase the amount of social capital by accumulating trust and closeness with neighborhoods. If
families living in poor neighborhoods have a collective potential to build a strong and healthy
community, it would contribute to the socialization of children.
Implications
Implications of family social capital theory
Family social capital variables for this study focus on the emotional attachment and
discipline behavior between mothers and children. Parents possess human capital, such as their
education, financial assets and family structure, which contribute to the growth of their children.
However, parents may not expect successful outcomes for children although they possess enough
human capital to support the child’s social development. Parents’ human capital should be
complemented by social capital which can be specified within parent-child relationships.
Parent’s high education level cannot assure child’s successful development and should be made
more substantial by the amount of interaction with parent-child relationships.
In a case of maternal depression, the stable psychological status of a father may be
extremely important in sustaining close relationships with children. Although the mother is
depressed, children’s problematic behaviors would be decreased if their relationships with the
father are strong and if he has good parenting practices. However, if both parents have
psychological problems, the children’s psychopathological problems would need clinical
treatment. Thus, adults within a family play an important role in accumulating family social
capital. If there were no adults within a family there might be a structural deficit of family social

90

capital. Children raised by a single parent could benefit from social support by extended family
and friends. Additionally, although there are adults within a family, a lack of family social
capital would exist if the relationships between parent and child are not strong. Therefore, it
would not be possible to generate social capital within a family without intergenerational
interactions, no matter how much human capital parents may possess. These ideas would be
interesting to explore in a future study.
In addition to increasing emotional interactions as a mediator of child’s behavioral
problems, improving the parental discipline response is another type of family social capital
generated by parents. If a mother is in depression, she may have difficulties to discipline her
child if misbehavior is due to deficit of interacting ability. Inappropriate communication with a
child and harsh punishment parenting may also reduce emotional interactions. This process
reduces family social capital as byproduct from inter-relationships with a mother and child and,
finally, contributes negatively to a child’s successful social development.
Policy implications of neighborhood social capital theory
Social capital is not an individual product, but should be considered a collective
dimension of society (Kawachi & Berkman, 2000). It should also be distinguished from social
support or social network that operates at the individual level.
The concept of social capital (Coleman, 1988) emphasizes functional efficiency as social
resources to facilitate children’s well-being. However, Coleman does not propose a strategy for
families in poverty. According to Coleman (1988; 1990), children’s successful development is
based on excellent material- and emotional- parental support. Furthermore, parents can exchange
useful information for children’s well-being in neighborhoods with a same socioeconomic status.
Because parents residing in poor neighborhoods may often feel that their children have

91

insufficient social resources to demonstrate their ability; they may experience more frequently
stressful situations in parenting practices. This study’s findings suggest that parents who live in
poor neighborhoods can share a sense of community belonging and trust with other parents
within the same community. Through the building of these trust relationships, parents can then
expect more effective outcomes of their children. This is a hopeful finding that low income
neighborhoods can create situations in which children can have positive behavioral outcomes.
Children raised with sufficient support from parents with affluent financial- as well as
human- capital acquire a high level of academic accomplishment. Bourdieu (1986) defined it as
reinvestment in education by highly-educated parents. Eventually, social capital will be
reproduced extensively through these circulating processes (Bourdieu, 1986). These reproduction
processes of capital are suited to middle-class parents and children, but Coleman’s explanation
for children’s successful development does not provide an alternative for families in poverty who
lack financial-, human-, and social- capital.
How do parents bond with their neighbors? How do mothers living in disadvantage
neighborhoods improve their mental health? What factors affect the social development of their
children? An obvious but practical alternative is to increase the amount of family social capital.
If it is difficult to increase social capital within a family to generate enough, extra-familial
resources, including social connections and support should be an alternative. Family and
neighborhood social capital should be considered as complementary to increase parental
psychological health and economic well being.
Participating in neighborhood activities, such as ethnic- or religious-based community is
important in building social capital as well as the psychological well-being of individual family
members. In addition, policies and programs that seek to build social capital should focus on

92

family participation. The Community Capacity Model of Bowen and colleagues (2000; cited in
Brisson & Usher, 2005) demonstrates how community members develop social capital by
sharing responsibility for community welfare and by meeting community needs.
Policies and intervention programs should address the need for increasing resources for
public schools, communities, and homeownership in low-income neighborhoods relating to
physical and mental well-being of both parents and children (Brisson & Usher, 2005). The most
ideal and practical alternative is to acquire the collective power to frame policy. In the broader
context, social capital affects mental health through political processes at the neighborhood level
(Kawachi, Kennedy, Wilkinson, 1999). It implies that practical actions are needed for better
quality of life in a homogenous group. While family social capital is accumulated by investing
time, materials, and energy in building relationships with parents and children, neighborhood
social capital is used to participate extensively in neighborhood activities, even to the extent of
establishing policies by local decision-makers, such as better social services (Lofors & Sundquist,
2007).
Implications for clinical intervention
Some studies have found that a mother’s psychological health is a strong predictor of her
children’s behavior problems. In the majority of these studies, all indicators employed as was for
the current study were rated and reported by the primary caregiver who is a mother. However,
there have been many arguments about the validity of findings depending on maternal reports as
indicators of children’s behavior (Achenbach, Howell, McConaughy, & Stanger, 1998;
Ackermann & De Rubeis, 1991; Chilcoat & Breslau, 1997; Ferguson, Lynskey, & Horwood,
1993; Kinard, 1995; McFarland, 1996; Richter, 1992; Tarullo, Richardson, Radke-Yarrow, &
Martinez, 1995).

93

For example, children’s behavior problems may be observed differently from one context
to another, which means behavior assessed at a home by either a mother or father may differ
from behavior observed at a school by a teacher. This study grouped mothers who were
depressed and mothers living in poor neighborhood, and children’s behavior problems were
reported by mothers as primary caregivers.
Najman and colleagues (2000) found that anxious or depressed mothers tend to be more
accurate in detecting and reporting child behavior pathology; non-depressed mothers tend to
under-report their children’s behavior problems. In addition, findings indicated that the higher
rate of behavior problems in children of depressed mothers was biased by over-reporting
behavior problems. They, however, also found that the children of depressed mothers actually do
manifest more behavior problems than do other children, and these behaviors are detected
accurately by depressed mothers; this suggests there is not necessarily bias in reports by mothers’
who are depressed.
Findings of the current study also indicate that the effects of children’s behavior
problems were more significant among depressed mothers, suggesting that the children of
depressed mothers were likely to have behavior problems. It implies that there was no bias in the
reports between mothers with depression and without depression. Several studies suggested that
mothers with depression were more accurate in their assessments of psychiatric problems of the
children than were mothers who did not report depression (Ackermann & De Rubeis, 1991;
Kinard, 1995). McFarland (1996) found that disagreements between mothers and other raters
were no greater for mothers who were depressed than for mothers who were not depressed,
implying that a mother’s depression affects her perceptions of herself more than those she has
about her child.

94

For detecting behavior problems of school-aged children, teachers presumably may be
more aware of a child’s behavior problems, specifically, externalizing behaviors rather than
internalizing, because they should maintain behavioral control in the classroom (Weiss, Dodge,
Bates, & Pettit, 1992). Thus, it may be possible to assume that teachers’ reports are more
accurate than are reports by mothers. However, there is no certainty that a teacher’s report can
capture all behavior problems, because of discrepancies in the child’s behavior at home and at
school. If children’s behaviors at home differ from those at school, the reports by depressed
mothers would capture more behavior problems (Achenbach et al., 1998).
Clinicians require more objective and accurate rating reports to reduce the bias of
observation for mothers with depression and their children who need clinical treatment. In
general, compared to mothers without depression, mothers with depression have more trouble in
controlling their children’s behaviors and may overstate their children’s behavior problems. In
addition, since using only a mother’s report may reduce the reliability of findings, it would
provide more accurate prescription and treatment if a clinician uses observational reports from
multiple raters before beginning treatment (Kim-Cohen et al., 2005; Najman et al., 1999).
If a mother’s depressive symptoms are critical, she is likely to be an ineffective parent
(Serbin & Karp, 2003). Clinical treatment, including medication or therapy to decrease
children’s behavior problems as well as to improve mother’s depressive symptoms, requires
positive and skilled parenting practice. Therefore, clinicians who treat behaviorally-disordered
children should consider screening their mothers by evaluating seriousness of depression to
identify treatment needs for the family (Kim-Cohen et al., 2005).

95

CHAPTER VI
Limitation and Future Research

The findings of the study are now summarized. Following this summary, limitations and
recommendations for future research are presented last.
Summary
This study investigated the relationship between household income and child behavior
problems, indicating internalizing and externalizing those behaviors, through examining different
mediating factors across groups designated by maternal depression and family poverty.
Research questions and the following hypotheses of the study were addressed by utilizing
structural equation modeling. Findings of this research are summarized below as they relate to
the three research questions in this study.
Research Question 1: Examine the mediation effects of neighborhood and family factors
on the causal effects of household income level on a child’s behavior problems within the four
groups. Are there significant indirect effects of the four mediator variables for household income
and children’s internalizing behavior problems within the four groups?
The mediation effect of a mother’s discipline response between household income level
and child internalizing behavior problems in the mothers who are depressed and living in poverty
neighborhoods was statistically significant. Emotional interaction between mother and child was
also a significant mediator between household income level and child internalizing behavior
problems in the mother with depression group, but not in the poor neighborhood group. There
were no mediating effects in relation to neighborhood social cohesion and interaction between

96

household income and child internalizing behaviors in both the non-depressed mothers living in
poor neighborhoods and the non-depressed mothers living in non-poor neighborhoods
Research Question 2: Examine the mediation effects of neighborhood and family factors
for causal effects of household income level on child behavior problems within the four groups.
Are there significant indirect effects of the four mediator variables for household income and
child externalizing behavior problems within the four groups?
The mediation effect of a mother’s discipline response for household income level and
child externalizing behavior problems in the mothers who are depressed and living in poor
neighborhoods was statistically significant. In addition, the mediation effect of emotional
interaction between mother and child was found in the group with depressed mothers living in
non-poor neighborhoods; however, that result was rejected by testing the significance of
mediation, indicating no statistical significance for a mediator in this group. Neighborhood social
cohesion between a mother and a child was also a significant mediator for household income
level and child externalizing behavior problems in the group for mother with non-depressed, but
not in poor neighborhoods. There were no mediating effects in relation to neighborhood
interactions between household income and child externalizing behaviors with non-depressed
mothers living in non-poor neighborhoods.
Research Question 3: Examine the moderated mediation effects of family income level
and mediators on child behavior problems that are internalizing and externalizing. Do the
mediating effects of neighborhood and family factors significantly differ across the four groups?
A mother’s depression and neighborhood poverty levels do not have the moderation
effects of household income level on child behavior problems for internalizing and externalizing
through the four types of mediators, although different characteristics across groups were

97

supported in that some groups had a different mediator from hypotheses presented in Research
Questions 1 and 2.
Limitation
This study hypothesized that there would be moderator effects of mother’s depression
and poverty level across the four groups. The difference between depressed and non-depressed
mother and between mothers living in poor neighborhood and non-poor neighborhood, in
children’s behavioral problems has been documented in previous studies. However, despite the
evidence provided by literature, results of the analysis for group comparisons were not
statistically significant.
These findings first may be explained by socio-cultural and childrearing style differences
among ethnic groups. In this study, around 50% of total participants were Latina. Furthermore,
from each subgroup, 74.7% of all mothers in depressed/poor neighborhood, and 68.1% of total
number of mothers in non-depressed/poor neighborhood, were Latina, indicating unbalanced
distribution in the data. This unbalanced population distribution might bring biased or
unexpected results because there are socio-cultural differences in lifestyle, ethnic characteristics,
and childrearing style among ethnic groups. One study examined how chronic poverty affected
child behavioral problems indirectly through parenting practices across White, Black, and Latino
children (Pachter et al., 2006). More concentration of certain ethnic group than other ethnic
groups may affect the direction of response of the questionnaire which designated selected
variables for this study, thus, finally, leading to weaken significant moderating effects.
Another possible reason is that the depression variable was rated by participants
(mothers), while the neighborhood poverty variable was based on the report by the poverty
estimation of census tract level (Peterson et al., 2004), which means the former was an individual

98

level of data and the latter was a group level of data. Because perception of neighborhood
poverty may vary by individual or ethnic characteristics, there may be a discrepancy in analyzing
all four groups using those two variables as moderators. It may also affect the results of group
comparison analysis.
However, overall results indicate that there are different mediating factors for three
groups (depressed/poor neighborhood, depressed/non-poor neighborhood, and nondepressed/poor neighborhood group), suggesting the existence of a distinct mediator according to
each characterized group. A mother’s depression and neighborhood poverty still influence the
causal relation among variables of the proposed structural model. Clinicians and policy makers
need the most appropriate treatment and intervention programs consistent with the characteristics
of each group.
Future Research
The significant effect of a mother’s discipline response to child misbehaviors supports the
finding that negative parenting attitudes are associated with increasing a child’s overall behavior
problems, including anxiety, depression, dependence, peer conflicts, hyperactivity, stubbornness,
and antisocial conduct. If a mother has chronic depressive symptoms, her depression will affect
her ability to have intimate interactions with a child. Overall, a mother’s depression affects a
child’s psychological health, as it relates to internalizing and externalizing behavior problems.
Poor neighborhoods have important assets to offer, such as social cohesion and mutual
trust, and these may have a protective effect on child problematic behavior. The way to improve
a child’s behavior problem in a neighborhood poverty context is to increase the amount of social
capital by accumulating trust and closeness within the neighborhood. If families with

99

neighborhood poverty have a collective potential for building a strong and healthy community,
then that potential also can contribute to better socialization of children.
Policy makers should provide intervention programs that can address the need for
increasing the amount of neighborhood social capital and using that capital as a resource in lowincome neighborhoods for better physical and mental health. Establishing policies by decisionmakers should start with the need for practical actions, such as participating in neighborhood
activities and acquiring influential power, such as collective efficacy and voting rights for a
better quality of life, especially in the poverty context.
With respect to future research, findings from this study would suggest taking a
cooperative approach between family and neighborhood, so as to establish both a healthy family
and better neighborhood. Getting more direct clinical services to parents with depression and
providing wrap around services for families where positive parenting could be supported along
with building reciprocal relationships of trust in the neighborhood.
For the interventions to address the problems associated with disadvantaged
neighborhoods, more synthesized social service programs focused on integrated focus on both
mother and child would be more effective (Pierre & Layzer, 1999). For the many existing
interventions, attention has been placed on either the child or only the parent. More integrated
intervention programs that can benefit both mother and child simultaneously should be
developed. Additionally, continuous treatment service for depressed mothers and parental
education programs to better support positive parenting are necessary. Future research should be
aimed at developing more effective intervention programs to maximize both parent’s
psychological stability and children’s positive social developments.

100

APPENDIX

101

Appendix: Latent Variable Composites

Table 14
Behavior Problems Index Subscales
Item from parent questionnaire
1. Has sudden changes in mood or feeling
2. Has felt or complained that no one loved him/her
3. Has been rather high strung, tense and/or nervous
4. Has been cheated or told lies
5. Has been too fearful or anxious
6. Has argued too much
7. Has had difficulty concentrating; has not been able to pay attention for long
8. Has been easily confused and/or has seemed to be in a fog
9. Has bullied or has been cruel or mean to others
10. Has been disobedient
11. Has not seemed to feel sorry after misbehaving
12. Has had trouble getting along with other children
13. Has been impulsive or has acted without thinking
14. Has felt worthless or inferior
15. Is not liked by other children
16. Has had much difficulty getting his/her mind off certain thoughts
17. Has been restless or overly active; has not been able to sit still
18. Has been stubborn, sullen, or irritable
19. Has a very strong temper and has lost it easily
20. Has been unhappy, sad or depressed
102

Internalizing

Externalizing
X

X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X

Total
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X

Table 14 (cont’d)
Behavior Problems Index Subscales
21. Has been withdrawn, and/or has not become involved with others
22. Has purposely broken items or deliberately destroyed his/her own or another’s
23. Has been clinging to adults
24. Has cried too much
25. Has demanded much attention
26. Has been too dependent on others
Number of Items
Alpha level

103

X
X
X
X
X
X
11
.73

15
.87

X
X
X
X
X
X
26
.89

Collective Efficacy
Set 1. Neighborhood cohesiveness
1. This a cohesive or unified neighborhood.

2. People around here are willing to help their neighbors.
3. People in this neighborhood generally don’t get along with each other.

Set 2. Collective efficacy
1. People in this neighborhood do not share the same values.

2. People in this neighborhood can be trusted.

Response range (1=strongly agree, 2=agree, 3=unsure, 4=disagree, 5=strongly disagree)

Interaction with neighborhoods
1. About how often do you and people in your neighborhood do favors for each other? For
example, watch each other’s children, help with shopping, lend gardening or house tools. Would
you say:

2. When a neighbor is not at home, how often do you and other neighbors watch over his or her
property? Would you say:
3. How often do you and other people in the neighborhood seek each other’s advice about
personal things such as child rearing or job openings? Would you say:

Response range (1= often, 2=sometimes, 3=rarely, 4=never)

104

Interaction with a mother-child
Set of emotional interactions
1. In the past week, about how many times have you praised your child for doing something
worthwhile?

2. In the past week, have you shown your child physical affection (for example: kisses, hugs,
stroking hair, etc.)?

3. In the past week, how many times have you told another adult (for example: spouse, friend,
co-worker, visitor, relative) something positive about your child?

Response range (1=never, 2=once, 3=several times, 4=almost every day)

Set of discipline responses
1. In the past week, how many times have you had to spank your child?

2. In the past week, how many times have you grounded your child?

3. In the past week, how many times have you sent your child to his/her room or another room as
a punishment?

Response range (1=never, 2=once, 3=several times, 4= almost every day)

105

REFERENCES

106

REFERENCES

Aber, J. L., Bennett, N. G., Conley, D. C., & Li, J. (1997). The effects of poverty on child health
and development. Annual Review of Public Health, 18, 463-483.
Achenbach, T. M., Howell, C. T., McConaughy, S. H., & Stanger, C. (1998). Six-year predictors
of problems in a national sample, IV: Young adults signs of disturbance. Journal of the
American Academy of Child and Adolescent Psychiatry, 37, 718-727.
Ackermann, R., & De Rubeis, R. J. (1991). Is depressive realism real? Clinical Psychological
Review, 11, 565-584.
Adler, N. E., Boyce, T., Chesney, M., Cohen, S., Folkman, S., Kahn, R. L., (1994).
Socioeconomic status and health. American Psychologist, 49(1), 15-24.
Allison, P. D. (2008). Missing Data. The Ohio State University. Sponsored by the Institute for
Population Research.
Alpern, L., & Lyons-Ruth, K. (1993). Preschool children at social risk: Chronicity and timing of
maternal depressive symptoms and child behavior problems at school and at home.
Development and Psychopathology, 5, 371-387.
Amato, P. R., & Rivera, F. (1999). Parental involvement and children's behavior problems.
Journal of Marriage and Family, 61, 3750384.
American Psychiatric Association. (1994). Diagnostic and Statistical Manual of Mental
Disorders (4th ed.). Washington, DC: Author.
American Psychiatric Association. (2000). Diagnostic and Statistical Manual of Mental
Disorders. Text Revision Washington, DC: Author.
Arrow, K. J. (Ed.). (2000). Social Capital: A Multifaceted Perspective. Washington, DC: The
International Bank for Reconstruction.
Aseltine, R. H., & Kessler, R. C. (1993). Marital disruption and depression in a community
sample. Journal of Health and Social Behavior, 34, 237-251.
Balaji, B. A., Claussen, H. A., Smith, C. D., Visser, N. S., Morales, J. M., & Perou, R. (2007).
Social support networks and maternal mental health and well-being. Journal of Women's
Health, 16(10), 1386-1396.

107

Bandalos, D. L. (2002). Is parceling really necessary? A comparison of results from item
parceling and categorical variables methodology. Structural Equation Modeling, 15, 211240.
Barnes, S., Mallinchrodt, C. H., Lindborg, S. R., & Carter, M. K. (2008). The impact of missing
data and how it is handled on the rate of false-positive result in drug development.
Pharmaceutical Statistics, 7(3), 215-225.
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.
Baron, S., Field, J., & Schuller, T. (Eds.). (2000). Social Capital: Critical perspectives. New
York.
Barry, T. D., Dunlap, S. T., Lochman, J. E., & Wells, K. C. (2009). Inconsistent discipline as a
mediator between maternal distress and aggression in boys. Child and Family Behavior
Therapy, 31(1), 1-19.
Baumrind, D. (1996). The discipline controversy revisited. Family Relations, 45, 405-415.
Belle, D. (1990). Poverty and women's mental health. American Psychologist, 45, 385-389.
Belle, D., & Doucet, J. (2003). Poverty, inequality, and discrimination as sources of depression
among U.S women. Psychology of Women Quarterly, 27, 101-113.
Bender, H. L., Allen, J. P., McElhaney, K. B., Antonishak, J., Moore, C. M., Kelly, H. O., et al.
(2007). Use of harsh physical discipline and developmental outcomes in adolescence.
Development and Psychopathology, 19(227-242).
Benson, M. J., & Deal, J. E. (1995). Bridging the individual and the family. Journal of Marriage
and the Family, 57, 561-566.
Bourdieu, P. (Ed.). (1986). The Forms of Capital. In J. G. Richardson: Handbook of theory and
research for the sociology of education. New York.
Bowen, N. K., Bowen, G. L., & Ware, W. (2002). Neighborhood social disorganization, families,
and the educational behavior of adolescents. Journal of Adolescent Research, 17, 468490.
Bowen, G. I., Martin, J. A., Mancini, J. A., & MNelson, J. P. (2000). Making Connections: A
neighborhood Transformation Family development Initiative. Retrieved from, November
21, 2002 http://aecf.org/initiatives/mc/.
108

Bradley, R. H., Caldwell, B. M., & Rock, S. R. (1988). Home environment and school
performance: A ten-year follow-up and examination of three models of environmental
action. Child Development, 59(4), 852-867.
Brennan, P. A., Le Brocque, R., & Hammen, C. (2003). Maternal depression, parent-child
relationships, and resilient outcomes in adolescence. Journal of the American Academy of
Child and Adolescent Psychiatry, 42(12).
Brisson, D. S., & Usher, C. L. (2005). Bonding social capital in low-income neighborhoods.
Family Relations, 54, 644-653.
Bronfenbrenner, U. (1986). Ecology of the family as a context for human development: Research
perspectives. Developmental Psychology, 22, 723-742.
Brooks-Gunn, J., & Duncan, G. (1997). The effects of poverty on children The Future of
Children, 7, 55-71.
Brooks-Gunn, J., Duncan, G. J., Klebanov, P. K., & Sealand, N. (1993). Do neighborhoods
influence child and adolescent development? American Journal of Sociology, 99(2), 353395.
Burke, L. (2003). The impact of maternal depression on familial relationships. International
Review of Psychiatry, 15(3), 243-255.
Byrne, B. M. (1991). The maslach Burnout inventory: Validating factorial structure and
invariance across intermediates, secondary, and university educators. Multivariate
Behavioral Research, 26(4), 583-605
Byrne, B. M. (2008). Testing for multigroup equivalence of a measuring instrument: A walk
through the process. Psicothema, 20 (4), 872-882.
Carlson, M. J. (2006). Family structure, father involvement, and adolescent behavioral outcomes.
Journal of Marriage and Family, 68(1), 137-154.
Caughy, M. O., & O'Campo, P. J. (2006). Neighborhood poverty, social capital, and cognitive
development of African American Preschoolers. American Journal of Community
Psychology, 37(I/2), 141-154.
Caughy, M. O., O'Campo, P. J., & Muntaner, C. (2003). When being alone might be better:
neighborhood poverty, social capital, and child mental health. Social Science and
Medicine, 57(2), 227-237.

109

Chilcoat, H. D., & Breslau, N. (1997). Does psychiatric history bias mothers' reports? An
application of a new analytic approach. Journal of the American Academy of Child and
Adolescent Psychiatry, 36, 971-979.
Cohen, D. A., Inagami, S., & Finch, B. (2008). The built environment and collective efficacy.
Health and Place, 14, 198-208.
Coleman, J. S. (1988). Social capital in the creation of human capital. American Journal of
Sociology, 94, S95-S120.
Coleman, J. S. (1990). Foundation of Social Theory. Cambridge, MA: Belknap Press of Harvard
University Press.
Conger, R. D., Ge, X., Elder, G. H., Lorenz, F. O., & Simons, R. L. (1994). Economic stress,
coercive family process and developmental problems of adolescents: Special issue on
children and poverty. Child Development, 65, 541-561.
Crosnoe, R. (2004). Social capital and the interplay of families and schools. Journal of Marriage
and the Family, 66, 267-280.
Cummings, E. M., & Davis, P. T. (1994). Maternal depression and child development. Journal of
Child Psychology and Psychiatry, 35(1), 73-112.
Cutrona, C. E., Wallace, G., & Wesner, K. A. (2006). Neighborhood characteristics and
depression: An examination of stress processes. Current Directions in Psychological
Science, 15, 188-192.
Darling, N., & Steinberg, L. (1993). Parenting style as context: An integrative model.
Psychological Bulletin, 113(3), 487-496.
Davis, P. T., & Woitach, M. J. (2008). Children’s emotional security in the interparental
relationship. Current Directions in Psychological Science, 17, 269-275.
Deng, X., Doll, E. J., Hendrickson, A. R., & Scazzero, J. A. (2005). A multi-group analysis of
structural invariance: an illustration using the technology acceptance model. Information
and Management, 42, 745-759.
Dix, T., & Meunier, L. N. (2009). Depressive symptoms and parenting competence: An analysis
of 13 regulatory processes. Developmental Review, 29, 45-68.
Downey, G., & Coyne, J. C. (1990). Children of depressed parents: in integrative review.
Psychological Bulletin, 108(1), 50-76.
110

Dufur, M. J., Parcel, T. L., & McKune, B. A. (2008). Capital and context: Using social capital at
home and at school to predict child social adjustment. Journal of Health and Social
Behavior, 49, 146-161.
Duncan, G., & Brooks-Gunn, J. (2000). Family poverty, welfare reform, and child development.
Child Development, 71(1), 188-196.
Duncan, G., & Brooks-Gunn, J. (Eds.). (1997). Consequences of growing up poor. New York:
Russell Sage Foundation.
Duncan, G. J., Yeung, W. J., Brooks-Gunn, J., & Smith, J. R. (1998). How much does childhood
poverty affect the life chances of children? American Sociological Review, 63, 406-423.
Ehrle, J., Moore, K., & Brown, B. (1999). 1997 Snapshots of America's families: Adults'
environment and behaviors-Mental health of parents. Retrieved November 9, 2003, from
http://www.urban.org
Elgar, F. J., McGrath, P. J., Waschbusch, D. A., Stewart, S. H., & Curtis, L. J. (2004). Mutual
influences on maternal depression and child adjustment problems. Clinical Psychology
Review, 24(4), 4410459.
Engle, L. P., & Black, M. M. (2008). The effect of poverty on child development and educational
outcomes. Annals of the New York Academy of Sciences, 1136, 243-256.
Ferguson, K. M. (2006). Social capital and children's wellbeing: A critical synthesis of the
international social capital literature. Journal of Social Welfare, 15, 2-18.
Fergusson, D. M., Lynskey, M. T., & Horwood, L. J. (1993). The effects of maternal depression
on maternal ratings of child behavior. Journal of Abnormal Child Psychology, 21, 245269.
Fletcher, A. C., Walls, J. K., Cook, E. C., Madison, K. J., & Bridge, T. H. (2008). Parenting style
as a moderator of associations between maternal disciplinary strategies and child wellbeing. Journal of Family Issues, 29(12), 1724-1744.
Furstenberg, E., Cook, T., Eccles, J., Elder, G. H., & Sameroff, A. (1999). Managing to Make It:
Urban Families and Adolescent Success. Chicago: University of Chicago Press.
Furstenberg, F. F. S., & Hughes, M. E. (1995). Social capital and successful development among
at-risk youth. Journal of Marriage and the Family, 57(3), 580-592.
Furstenberg, F. F. S., Morgan, P., & Allison, P. D. (1987). Parental participation and children's
well-being after marital dissolution. American Sociological Review, 52, 695-701.
111

Garbarino, J. (1992). The meaning of poverty in the world of children. American Behavioral
Scientist, 35(3), 220-237.
Gelfand, D. M., & Teti, D. M. (1990). The effects of maternal depression on children. Clinical
Psychology Review, 10, 329-353.
Gershoff, E. T. (2002). Corporal punishment by parent and associated child behaviors and
experiences: A meta-analytic and theoretical review. Psychological Bulletin, 128(4), 539579.
Ginter, D., Haveman, R., & Wolfe, B. (2000). Neighborhood attributes as determinants of
children's outcomes: How robust are the relationships. Journal of Human Resources,
35(4), 603-642.
Goodman, S. H. (2007). Depression in mothers. Annual Review of Clinical Psychology, 3, 107135.
Goodman, S., Brogan, D., Lynch, M., & Fielding, B. (1993). Social and emotional competence
in children of depressed mothers. Child Development, 64, 516-531.
Goodman, S. H., & Gotlib, I. H. (1999). Risk for psychopathology in the children of depressed
mothers: a developmental model for understanding mechanisms of transmission.
Psychological Review, 106(3), 458-490.
Gotlib. I., & MaCabe, S. B (1990). Marriage and psychology., In F. D. Fincham & T. N
Bradbury (Eds.), The Psychology of Marriage, 226-257. New York, Guilford Press.
Grace, S. L., Evindar, A., & Stewart, D. E. (2003). The effect of postpartum depression on child
cognitive development and behavior: A review and critical analysis of the literature.
Archives of Women's Mental Health, 6, 263-274.
Graham, C. A., & Easterbrooks, A. E. (2000). School-aged children's vulnerability to depressive
symptomatology: The role of attachment security, maternal depressive symptomatology,
and economic risk. Development and Psychopathology, 12(2), 201-213.
Hagan, J., MacMillan, R., & Wheaton, B. (1996). New kid in town: Social capital and the life
course effects of family migration on children. American Sociological Review, 61(3),
368-385.
Hammen, C., & Brennan, P. A. (2001). Depressed adolescents of depressed and nondepressed
mothers: Tests of an interpersonal impairment hypothesis. Journal of Consulting and
Clinical Psychology, 106(3), 458-490.

112

Hammen, C., Brennan, P. A., & Shih, J., H. (2004). Family discord and stress predictors of
depression and other disorders in adolescent children of depressed and nondepressed
women. Journal of the American Academy of Child and Adolescent Psychiatry, 43(8),
994-1002.
Hammen, C., Burge, D., & Stansbury, K. (1990). Relationship of mother and child variables to
child outcomes in a high-risk sample: A causal modeling analysis. Developmental
Psychology, 26, 24-30.
Hanks, C. A. (2008). Social capital in an impoverished minority neighborhood: Emergence and
effects on children's mental health. Journal of Child and Adolescent Psychiatric Nursing,
21(3), 126-136.
Harnish, J. D., Dodge, K., A, & Valente, E. (1995). Mother-child interaction quality as a partial
mediator of the roles of maternal depressive symptomatology and socioeconomic status
in the development of child behavior problems Child Development, 66(3), 739-753.
Harris, K., Furstenberg, F. F. S., & Marmer, J. K. (1998). Paternal involvement with adolescents
in intact families: The influence of fathers over the life course. Demography, 35(2), 201216.
Hill, E. N., & Herman-Stahl, A. M. (2002). Neighborhood safety and social involvement:
Associations with parenting behaviors and depressive symptoms among African
American and Euro-American mothers. Journal of Family Psychology, 16(2), 209-209.
Holt, J. K. (2004). Item parceling in structural equation models for optimum solutions. Paper
presented at the Annual Meeting of the Mid-western Educational Research Association.
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis:
Conventional criteria versus new alternations. Structural Equation Modeling, 6, 1-55.
Ingoldsby, E. M., Shaw, D. S., Winslow, E., Schonberg, M., Gilliom, M., & Criss, M. M. (2006).
Neighborhood disadvantage, parent-child conflict, neighborhood peer relationships, and
early antisocial behavior problem trajectories. Journal of Abnormal Child Psychology,
34(3), 293-309.
Israel, G. D., Beaulieu, L. J., & Hartless, G. (2001). The influence of family and community
social capital on educational achievement. Rural Sociology, 66(1), 43-68.
Kawachi, I., & Berkman, F. L. (2001). Social ties and mental health. Journal of Urban Health,
78(3), 458-468.
Kawachi, I., & Berkman, L. F. (2000). Social Epidemiology. New York: Oxford University Press.
113

Kawachi, I., Kennedy, B. P., & Wilkinson, R. G. (1999). Income Inequality and Health: A
Reader. New York: The New Press.
Kessler, R. C., Andrews, G., Mroczek, D., Ustrun, B., & Hu, W. (1998). The World Health
Organization Composite International Diagnostic Interview short-form(CIDI-SF).
International Journal of Methods in Psychiatric Research, 7(172-185).
Kiernan, K. E., & Huerta, M. C. (2008). Economic deprivation, maternal depression, parenting
and children's cognitive and emotional development in early childhood. The British
Journal of Sociology, 59(4), 783-804.
Kim-Cohen, J., Moffitt, T. E., Taylor, A., Pawlby, S. J., & Caspi, A. (2005). Maternal depression
and children's antisocial behavior. Archives of General Psychiatry, 62, 1730181.
Kinard, E. M. (1995). Mother and teacher assessments of behavior problems in abused children.
Journal of the American Academy of Child and Adolescent Psychiatry, 34, 1043-1054.
Klebanov, P. K., Brooks-Gunn, J., & Duncan, J. G. (1994). Does neighborhood and family
poverty affect mother's parenting, mental health, and social support? Journal of Marriage
and Family, 56, 441-455.
Klebanov, P. K., Brooks-Gunn, J., McCarton, C. M., & McCormick, M. C. (1998). The
contribution of neighborhood and family income upon developmental test scores over the
first three years of life. Child Development, 69, 1420-1436.
Kline, R. B. (Ed.). (2005). Principles and Practice of Structural Equation Modeling. New York:
NY: The Guilford press.
Kupersmidt, J. B., Griesler, P. C., DeRosier, M. E., Patterson, C. J., & Davis, P. W. (1995).
Childhood aggression and peer relations in the context of family and neighborhood. Child
Development, 66, 360-375.
Kurstjens, S., & Wolke, D. (2001). Effects of maternal depression on cognitive development of
children over the first 7 years of life. Journal of Child Psychology and Psychiatry, 42(5),
623-636.
Kuruvilla, A., & Jacob, K. S. (2007). Poverty, social stress and mental health. Indian Journal of
Medical Research, 126, 273-278.
Lara-Cinisomo, S., & Griffin, B. H. (2007). Factors associated with major depression among
mothers in Los Angeles. Women's Health Issues, 17(5), 316-324.

114

Leung, D. W., & Slep, A. M. S. (2006). Predicting inept discipline: The role of parental
depressive symptoms, anger, and attributions. Journal of Consulting and Clinical
Psychology, 74(3), 524-534.
Leventhal, T., & Brooks-Gunn, J. (2000). The neighborhoods they live in: The effects of
neighborhood residence on child and adolescent outcomes. Psychological Bulletin,
126(2), 309-337.
Leventhal, T., & Brooks-Gunn, J. (2003). Children and youth in neighborhood context. Current
Directions in Psychological Science, 12, 27-32.
Leventhal, T., & Brooks-Gunn, J. (2004). A randomized study of neighborhood effects on lowincome children's educational outcomes. Developmental Psychology, 40(4), 488-507.
Lofors, J., & Sundquist, K. (2007). Low-linking social capital as a predictor of mental disorders:
A cohort study of 4.5 million Swedes. Social Science and Medicine, 64, 21-34.
Luoma, I., Taminen, T., Kaukonen, P., Laippala, P., Puura, K., Salmelin, R., (2001).
Longitudinal study of maternal depressive symptoms and child well-being. Journal of the
American Academy of Child and Adolescent Psychiatry, 40(12), 1367-1374.
Lyons-Ruth, K., Connell, D. B., Grunebaum, H. U., & Botein, S. (1990). Infants at social risk:
Maternal depression and family support services as mediators of infant development and
security of attachment. Child Development, 61(1), 85-98.
MacKinnon, D. P. (2008). Introduction to Statistical Mediation Analysis. New York, NY:
Lawrence Erlbaum Associates.
Marsiglio, W., Amato, P., Day, R. D., & Lamb, M. E. (2000). Scholarship on fatherhood in the
1990s and beyond. Journal of Marriage and Family, 62(1173-1191).
Mayer, S. (1997). What Money Can't Buy: The Effect of Parental Income on Children's
Outcomes. Cambridge, MA: Harvard University Press.
McFarland, M. L. (1996). Mother's depression and the treatment of disruptive child behavior
disorders. PhD thesis. The University of Queensland, Australia.
McKenzie, K., Whitley, R., & Weich, S. (2002). Social capital and mental health. British Journal
of Psychiatry, 181, 280-283.
McLeod, J. D., & Shanahan, M. J. (1993). Poverty, parenting, and children's mental health.
American Sociological Review, 58(3), 351-366.
115

McLeod, J. D., & Nonnemaker, J. M. (2000). Poverty and child emotional and behavioral
problems: Racial/ethnic differences in processes and effects. Journal of Health and
Social Behavior, 41, 137-161.
McNeal, R. B. (1999). Parental involvement as social capital: Differential effectiveness on
science achievement, truancy, and dropping out. Social Force, 78(1), 117-144.
Mitchell, C. U., & LaGory, M. (2002). Social capital and mental distress in an impoverished
community. City and Community, 1(2), 199-222.
Moss, E., Rousseau, D., Parent, S., St-Laurent, D., & Saintonge, J. (1998). Correlates of
attachment at school age: Maternal reported stress, mother-child interaction, and behavior
problems. Child Development, 69(5), 1390-1405.
Murray, L., & Cooper, P. J. (1997). Effects of postnatal depression on infant development.
Archives of Disease in Childhood 77, 99-101.
Murray, L., Cooper, P. J., Wilson, A., & Romaniuk, H. (2003). Controlled trial of the short- and
long-term effect of psychological treatment of post-partum depression: Impact on the
mother-child relationship and child outcome. British Journal of Psychiatry, 182, 420-427.
Murray, L., Halligan, S. L., Adams, G., Patterson, P., & Goodyer, I. M. (2006). Socio-emotional
development in adolescents at risk for depression: The role of maternal depression and
attachment style. Development and Psychopathology. 18, 489-516.
Najman, J., Williams, G. M., Nikles, J., Spence, S., Bor, W., O'Callaghan, M., et al. (2000).
Mothers' mental illness and child behavior problems: Cause-effect association or
observation bias? Journal of the American Academy of Child and Adolescent Psychiatry,
39(5), 592-602.
Nix, R. L., Pinderhughes, E., Dodge, K. A., Bates, J. E., Pettit, G. S., & McFadyen-Ketchum, S.
A. (1999). The relation between mothers' hostile attribution tendencies and children's
externalizing behavior problems: The mediating role of mother's harsh discipline
practices. Child Development, 70(4), 896-909.
Odgers, C. L., Moffitt, T. E., Tach, L. M., Sampson, R. J., Taylor, A., & Matthews, D. L. (2009).
The protective effects of neighborhood collective efficacy on British children growing up
in deprivation: A developmental analysis. Developmental Psychology, 45(4), 942-957.
Pachter, M. L., Auinger, P., Palmer, R., & Weitzman, M. (2006). Do parenting and the home
environment, maternal depression, neighborhood, and chronic poverty affect child
behavioral problems differently in different racial-ethnic groups? Pediatrics, 117(4),
1329-1338.
116

Parcel, L. T., & Menaghan, G., E. (1993). Family social capital and children's behavior problems.
Social Psychology Quarterly, 56(2), 120-135.
Parcel, T. L., & Dufur, M. J. (2001). Capital at home and at school: Effects on child social
adjustment. Journal of Marriage and the Family, 63, 32-47.
Parcel, T. L., Dufur, M. J., & Zito, R. C. (2010). Capital at home and at school: A review and
synthesis. Journal of Marriage and Family, 72, 828-846.
Parcel, T. L., & Menaghan, E. G. (1994). Early parental work, family social capital, early
childhood outcomes. The American Journal of Sociology, 99(4), 972-1009.
Patterson, G., Reid, J., & Dishion, T. (1992). Antisocial Boys: A Social Interactional Approach.
Eugene, OR: Castalia Publishing.
Pebley, A. R., & Sastry, N. (2003). Neighborhood, Poverty and Children's Well-being: A Review.
RAND: Labor and Population.
Peterson, G. W., & Rollins, B. C. (Eds.). (1987). Parent-child Socialization as Symbolic
Interaction, In M. Sussman and S. K. Steinmetz: Handbook of marriage and the family.
New York: Plenum.
Peterson, C. E., Sastry, N., Pebley, A. R., Ghosh-Dastidar, B., Williamson, S., & Lara-Cinisomo,
S. (2004). The Los Angeles Family and Neighborhood Survey Codebook. LAND: Labor
and Population.
Peterson, J. L., & Zill, N. (1986). Marital disruption, parent-child relationships, and behavior
problems in children. Journal of Marrigae and Family, 48(2), 295-307.
Pong, I., Hao, L., & Gardner, E. (2005). The roles of parenting styles and social capital in the
school performance of immigrant Asian and Hispanic adolescents. Social Science and
Quarterly, 86(4), 928-950.
Portes, A. (1998). Social capital: its origins and applications in modern sociology. Annual
Review of Sociology, 24, 1-24.
Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling strategies for assessing and
comparing indirect effects in multiple mediator models. Behavior Research Methods,
40(3), 879-891.
Putnam, R. D. (2000). Bowling Alone: The collapse and revival of American community. New
York: Simon and Schuster.
117

Raykov, T., & Marcoulides, G. A. (2006). A First Course in Structural Equation Modeling.
Mahwah, NJ: Lawrence Erlbaum Associates.
Richters, J. E. (1992). Depressed mothers as informants about their children: A critical review of
the evidence for distortion. Psychological Bulletin, 112, 485-499.
Robinson, C. C., Mandleco, B., Olsen, S. F., & Hart, C. H. (Eds.). (2001). Parenting style and
dimensions questionnaire. In J. Touliatos & B. Handbook of Family Measurement
Techniques: Instruments and Index. Thousand Oaks, CA: Sage.
Ross, C. E. (2000). Neighborhood disadvantage and adult depression. Journal of Health and
Social Behavior, 41, 177-187.
Sampson, R. J., Morenoff, J. D., & Gannon-Rowley, T. (2002). Assessing "neighborhood
effects": Social processes and new directions in research. Annual Review of Sociology, 28,
443-478.
Sampson, R. J., Raudenbush, S. W., & Earls, F. (1997). Neighborhoods and Violent Crime: A
multilevel study of collective efficacy. Science, 277(15), 918-924.
Schumacher, R., & Lomax, G. (2004). A Beginner's Guide to Structural Equation Modeling (2nd
ed.). London: Lawrence Erlbaum Associates.
Schumacker, R. E. (2004). A Beginner's Guide to Structural Equation Modeling. Mahwah, NJ:
Lawrence Erlbaum Associates.
Serbin, L., & Karp, J. (2003). Intergenerational studies of parenting and the transfer of risk from
parent to child. Current Directions in Psychological Science, 12, 138-142.
Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and non-experimental studies:
New procedures and recommendations. Psychological Methods, 7, 422-445.
Silver, J. E., Heneghan, M. A., Bauman, J. L., & Stein, E. K. R. (2006). The relationship of
depressive symptoms to parenting competence and social support in inner-city mothers of
young children. Maternal and Child Health Journal, 10(1), 105-112.
Simons, R. L., Johnson, C., & Conger, R. D. (1994). Harsh corporal punishment versus quality
of parental involvement as an explanation of adolescent maladjustment. Journal of
Marriage and Family, 56(3), 591-607.
Simons, R. L., Simons, L. G., Burt, C. H., Brody, G. H., & Cutrona, C. (2005). Collective
efficacy, authoritative parenting and delinquency: A longitudinal test of a model
integrating community and family level process. Criminology, 43(4), 989-1028.
118

Sinclair, D., & Murray, L. (1998). Effects of postnatal depression on children's adjustment to
school: Teacher's reports. The British Journal of Psychiatry, 172, 58-63.
St. Pierre, R. G., & Layzer, J. I. (1999). Using home visit for multiple purposes: The
comprehensive child development program. The Future of Children, 9(1), 134-151.
Stein, A., Gath, D. H., Bucher, J., Bond, A., Day, A., & Cooper, P. J. (1991). The relationship
between post-natal depression and mother-child interaction. British Journal of Psychiatry,
158, 46-52.
Steinberg, L., Lamborn, S. D., Dornbusch, S. M., & Darling, N. (1992). Impact of parenting
practices on adolescent achievement: Authoritative parenting, school involvement, and
encouragement to succeed. Child Development, 63, 1266-1281.
Tarullo, L. B., Richardson, D. T., Radke-Yarrow, M., & Martinez, P. E. (1995). Multiple sources
in child diagnosis: Parent-child concordance in affectively ill and well families. Journal
of Clinical Child Psychology, 24, 173-183.
Teti, D. M., Gelfand, D. M., Messinger, D. S., & Isabella, R. (1995). Maternal depression and
the quality of early attachment: An examination of infants, preschoolers, and their
mothers. Developmental Psychology, 31(3), 364-376.
U. S. Census Bureau, Statistical Abstract of the United States. (2002). Washington, DC.
U. S. Census Bureau, Statistical Abstract of the United States. (2008). Washington, DC.
U. S. Census Bureau, Statistical Abstract of the United States. (2010). Washington, DC.
Umberson, D. (1992). Gender, marital status, and the social control of health behavior. Social
Science and Medicine, 34(8), 907-917.
Webster-Stratton, C., & Hammond, M. (1988). Maternal depression and its relationship to life
stress, perceptions of child behavior problems, parenting behaviors, and child conduct
problems. Journal of Abnormal Child Psychology, 16(3), 299-315.
Weiss, B., Dodge, K. A., Bates, J. E., & Pettit, G. S. (1992). Some consequences of early harsh
discipline: Child aggression and a maladaptive social information processing style. Child
Development, 63(6), 1321-1335.
Woolley, M. E., & Grogan-Kaylor, A. (2006). Protective factors in the context of neighborhood:
Promoting positive school outcomes. Family Relations, 55(1), 93-104.

119

Xue, Y., Leventhal, T., Brooks-Gunn, J., & Earls, F. (2005). Neighborhood residence and mental
health problems of 5 to 11 year olds. Archives of General Psychiatry, 62, 554-563.

120