a ,
1"l
t".

I

hv‘

5M1 .‘ -¢;.-.

.....

mg-‘wsammm:

Eu

w w" _' ':

ii

“a”...

are???

,4

A N «“341»?!

,5: a: '

[‘67

’5!

~5KLZM. nun“... . ...
“an

~I'u "' ,
". 3332*

(.,. .

nu...
. ‘5,

3 43
1 I: Eff
$9.23; ‘ l

:- a
‘

 

 

THEIS

515%?”

This is to certify that the
dissertation entitled

PARENTAL EFFICACY, ATTACHMENT, AND CHILD
SELF-CONTROL:
A COMPARISON OF PROCESSES AMONG
DEMOGRAPHIC GROUPS

presented by
M. ANGELA NIEVAR

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Family and Child Ecology

 

- . /"
i%azm M?‘

Major Profeésgr's Signature
51/9139!

Date

MSU is an Affimiative Action/Equal Opportunity Institution

 

 

 

'UBRARY I
Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

 

DATE DUE DATE DUE DAIEURUE
JAN 110 2096; “IL?“ 1‘ L»

w

 

FEB‘IZLSH 200$

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIRC/DateDuo.p65—p. 15

PARENTAL EFFICACY, ATTACHMENT, AND CHILD SELF-CONTROL:
A COMPARISON OF PROCESSES AMONG DEMOGRAPHIC GROUPS

By

M. Angela Nievar

A DISSERTATION
Submitted to
Michigan State University
in partial fitlfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Family and Child Ecology

2004

ABSTRACT

PARENTAL EFFICACY, ATTACHMENT, AND CHILD SELF -CONTROL:
A COMPARISON OF PROCESSES AMONG DEMOGRAPHIC GROUPS

By

M. Angela Nievar

According to Bandura’s (1982) self-efficacy theory, parents who feel competent
and effective, or who have high levels of parental efficacy, are more likely to invest time
and energy in parenting. In fact, parental efiicacy appears to be related to parental
resources and practices; fewer studies have examined the relation between parental
efficacy and children’s development (Bandura, Barbaranelli, Caprara, & Pastorelli, 2001;
Brazelton, 1983; Bugental & Shenum, 1984; Teti & Gelfand, 1991). This dissertation
examined linkages between parental efficacy, parenting, attachment security, and child
self-control.

The effects of parental efficacy and parenting may be different in various
environments. Bronfenbrenner (2000) theorized that the effects of parenting are
dependent on risks inherent within a child’s environment, such as poverty or racism. In
fact, research indicates that effects of parental efficacy vary by socioeconomic status,
race, and marital quality (Ardelt & Eccles, 2001; Corapci & Wachs, 2002; Machida,
Taylor, & Kim, 2002).

Using data from the National Institute of Child Health and Human Development
(N ICHD) Early Child Care Research Network, this dissertation empirically described
family processes related to parental efficacy within low-income and middle-income

families and within African American and Caucasian families. Structural equation

models indicated that parental efficacy and parental practices are predictive of
attachment security and children’s self-control. Linkages between parenting and child
outcomes were generally stronger among low-income groups than middle-income
groups, suggesting that proximal processes have a greater effect among children living
in disadvantaged families than among children with more resources.

Distinct effects were found within African American and Caucasian groups.
Different dimensions of parental efficacy, perceived competence and perceived control,
produced different effects on parenting style and child outcomes within demographic
groups. Marital conflict appeared to have a greater effect on boys than girls; other
gender differences raised interesting questions that warrant further investigation. In
sum, this dissertation added to our understanding of parental efficacy and its relation to
various aspects of children’s environments, attachment security, and child self-control

within diverse demographic groups.

ACKNOWLEDGMENTS

I would like to acknowledge the assistance I have received from the Graduate
School, the Department of Family and Child Ecology, and the professors and students
that have furthered my work. I would like to thank Dr. Luster for his assistance and
advice throughout my graduate program as my major professor, particularly in helping
me through the publication process on multiple occasions. I am also grateful to Dr. Luster
for submitting my admissions application to the University Distinguished Fellowship
program. I would like to thank the Graduate School for supporting me in that program
throughout my four years at Michigan State University. I would like to thank Dr. Holly
Brophy-Herb and Dr. Hiram Fitzgerald for their support and their encouragement in my
work as a research assistant. I am grateful for the travel support from Dr. Hiram
Fitzgerald, Dr. Holly Brophy-Herb, the Graduate School, the college, and the department.
I am also thankful for research support received from the Society for the Psychological
Study of Social Issues, the Graduate School, the College of Human Ecology, and the
Department of Family and Child Ecology. I want to thank Dr. Anne Soderman for her
support and encouragement in my research endeavors.

I would like to thank Dr. Marguerite Barratt for her encouragement and assistance
in obtaining data for the NICHD Early Child Care Study. I am grateful to Dr. Deborah
Johnson for her insight on that study. I am also grateful to the numerous researchers and
assistants that developed the study, collected the data, and completed previous works,

providing a foundation for this dissertation.

iv

Finally, I would like to thank my family. I am grateful to my parents for
providing a safe haven and fiee child care. I am thankful to them for giving me an
enriched home environment that allowed me to develop a love of learning. I am grateful
to my husband for his patience, assistance, and support. Finally, I would like to thank my

children for just being there and being wonderful.

TABLE OF CONTENTS

LIST OF TABLES ............................................................................... viii
LIST OF FIGURES .............................................................................. x
CHAPTER 1
INTRODUCTION ............................................................................... 1
Purpose of the Study ........................................................................ 3
General Research Questions ............................................................... 5
Attachment in Infancy .............................................................. 5
Social-emotional Development in Early Chilrfltood ............................ 5
Research Assumptions ....................................................................... 8
Importance of Parenting ........................................................... 8
Influence of Parental Beliefs ...................................................... 8
Ecological Perspective ............................................................ 9
Justification of the Research .............................................................. 9
Overview of the Dissertation ............................................................. 13
CHAPTER 2
LITERATURE REVIEW ...................................................................... 15
Ecological Theory ........................................................................... l 5
Social Cognitive Theory .................................................................. 18
Parental Eflicacy .................................................................. 20
Diversity and Parental Efl‘icacy ................................................. 22
Attachment Theory ......................................................................... 24
Correlates of A ttachment ........................................................ 26
Conclusion ................................................................................... 41
CHAPTER 3
METHOD ........................................................................................ 42
Hypotheses .................................................................................. 42
Dimensions of Parental Efficacy and Attachment in Infancy ............... 42
Ecological Model .................................................................. 45
Sample ....................................................................................... 46
Attrition ............................................................................. 51
Measurement ................................................................................. 52
Parental Characteristics .......................................................... 55
Contextual Characteristics ........................................................ 57
Data Analysis ............................................................................... 62

vi

CHAPTER 4

RESULTS ....................................................................................... 63

Preliminary Analyses ....................................................................... 63

Univariate Analyses ............................................................... 63

Bivariate Analyses ................................................................ 65

Tests of Hypotheses ......................................................................... 86

Conclusion ................................................................................... l 14
CHAPTER 5

DISCUSSION .................................................................................... l 15

Research Questions .......................................................................... 115

Dimensions of Parental Efficacy ................................................. l 16

Child Care Quality .................................................................. 123

Social-Emotional Development in Early Childhood......... .. . . .. 124

Marital Conflict ...................................................................... 126

Gender Diyfirences ................................................................ 129

Limitations ..................................................................................... 131

Future Directions ............................................................................ 133

Conclusion ...................................................................................... 135

APPENDICES .................................................................................... 137

REFERENCES ................................................................................... 148

vii

LIST OF TABLES

1 Income-to-needs ratios by time of data collection ....................................... 48
2 Descriptive statistics of study variables ................................................... 50
3 Measures, constructs, and age of child at administration ....................................... 53
4 Frequencies of dichotomous study variables by demographic groups ............... 67
5 Correlations between study variables for the total sample ............................. 69
6 Correlations for African American and Caucasian families ........................... 70
7 Correlations for low-income and middle-to-high income families ..................... 71
8 Correlations for study variables by child gender ........................................ 72
9 Tests for equality of means and equality of variances between African American

and Caucasian groups ....................................................................... 74
10 Study variables by economic group ........................................................ 76
l I Study variables by gender ................................................................... 78
12 Means of study variables by attachment classification with one-way analyses

of variance ..................................................................................... 81
13 Study variables by attachment security .................................................... 82
14 Hypotheses and tests related to perceived control ....................................... 86
15 Hypotheses and tests of perceived competence ........................................... 89

16 Parenting measures and strange situation classification with demographic

groups .......................................................................................... 92
17 Multinomial logistic regression model of attachment classification for the total

sample .......................................................................................... 94
18 Multinomial logistic regression model for Afiican American families ................ 95
19 Hypotheses and tests of moderators of child care quality ............................... 96
20 Hypotheses and corresponding significance tests of variable-level paths ........... 100

viii

21 Path coefl‘icients and estimates by child gender .......................................... 104

22 Tests for equality of means and equality of variances between low-income
Afiican American and Caucasian groups .................................................. 144

23 Tests for equality of means and equality of variances between midrfle-to-high
income Afiican American and Caucasian groups ........................................ 146

ix

LIST OF FIGURES

Hypothesized model of interactions with child care quality predicting

attachment security ........................................................................... 6
Hypothesized model of pathways leading to child outcomes: self-control

and attachment security ....................................................................... 7
Hypothesized model of perceived competence, responsive parenting, and

attachment classifications .................................................................... 44
Structural equation model with predictors of attachment and self-control ............ 102
Structural equation model with predictors of attachment and self-control,

excluding single-parent families ............................................................ 107
Structural equation model with low-income families .................................... 109
Structural equation model with low-income families, excluding single-parent

families ......................................................................................... 110
Structural equation model with middle-to-high income families ....................... 112

Structural equation model with middle-to-high income families, excluding
single-parent families ........................................................................ 113

CHAPTER 1

INTRODUCTION

The type of environment that parents provide for their children may be influenced
by their beliefs. Some parents may believe that they are efi‘ective teachers for their
children, although others do not. Their perception of their own effectiveness is defined as
parental efficacy (Bandura, 1982; Coleman & Karraker, 1997). Theoretically, parents
who feel that they are important and effective in their role will spend more time and
effort with their children. Research indicates that parental efficacy is related to parental
resources and practices, such as poverty, maternal depression, and harsh discipline, and
child characteristics such as temperament and attachment (Brazelton, 1983; Bugental &
Shenum, 1984; Nievar & Brophy-Herb, 2003; Luster & Kain, 1987; Teti & Gelfand,
1991). Experimental studies aimed at improving the home environment for young
children through parent education show related improvements among parental efficacy,
parenting, and child outcome (Hastings & Brown, 2002; Hoath & Sanders, 2002;
Sofronoff & Farbotko, 2002).

The effects of parental efficacy and parenting may be different in various
environments. Bronfenbrenner (2000) theorized that the effects of parenting are
dependent on risks inherent within a child’s environment, such as poverty or racism. In
fact, research indicates that effects of parental efficacy vary by socioeconomic status,
race, and marital quality (Ardelt & Eccles, 2001; Corapci & Wachs, 2002; Machida,

Taylor, & Kim, 2002).

Developmental processes related to parental practices may also vary due to
cultural differences (Johnson et al., 2003; McLoyd, Cauce, Takeuchi, & Wilson, 2000).
Cultural requirements may affect the way that parents think about parenting, and in turn
affect parenting practices (Ogbu, 1985). For example, certain cultures emphasize
children’s self-control more than others. Thus, parents in these cultures may provide
more opportunities for a child to learn to regulate their emotions.

Although ways of parenting may differ by culture, processes leading to child
outcome may be more universal. For example, power-assertive discipline appears to
inhibit the development of secure attachment and self-control (Davies, Harold, Goeke-
Morey, & Cummings, 2002; Kochanska & Knaack, 2003). Children’s attachment
security is related to their self-control as well as to the type of parenting they receive
(Londerville & Main, 1981; Braungart-Rieker, Garwood, Powers, & Wang, 2001;
Cummings & Davies, 1996; Cassidy, 1994)TI€PaCIr)'ents who are insensitive and
unresponsive to their child’s signals are more likely to have children who have difficulty
controlling their emotions or behavior (Rothbart & Bates, 1998). Parents who are
insensitive and unresponsive are also less likely to have children who develop secure
attachment relationships (De Wolff & van IJzendoorn, 1997).

Although relations between sensitivity, attachment, and self-control have been
researched, the study of parental efficacy and attachment is a relatively new area. A
search of the literature revealed no studies of parental efficacy and children’s self-control.
Literature that exists on parental efficacy, parenting, and attachment reports inconsistent
findings, particularly in the area of perceived parental competence (del Carmen,

Pedersen, Huffman & Bryan, 1993; Spieker & Booth, 1988; Teti & Gelfand, 1991).

Discrepancies in the literature may be the result of a misconception that a strong
sense of competence is always preferred. Some experimental work suggests that overly
confident parents who have a strong sense of competence may not be as sensitive as those
who have a more moderate perception of their abilities (Donovan & Leavitt, 1989;
Donovan, Leavitt, & Walsh, 1997; Donovan, Leavitt, & Walsh, 2000). Descriptive
studies indicate that clinically depressed mothers who have a weak sense of competence
are more likely to have problems with parenting their children effectively (T eti,
O’Connell, & Reiner, 1996). Thus, high levels of perceived competence and low levels
of perceived competence may both be related to inadequate parenting; however, most
measm'es are based on a linear scale that assumes high levels of competence to be
optimal. A second explanation for discrepant findings may be the way that parental
eflicacy is conceptualized. Parental efficacy is composed of multiple dimensions,
including perceived control over situations and perception of personal competence.
Perceived competence and perceived control may represent two separate belief systems

that have different effects on parenting and child outcomes.

Purpose of the Study
This dissertation focused on parental efficacy and its relation to parenting and
mother-child attachment, particularly among low-income families. Separate analyses
investigated the process of attachment within two periods of development: (a) in infancy,
and (b) in early childhood. Correspondence between parental efficacy, family processes

in infancy, and attachment classification at 15 months was analyzed. Pathways between

study variables, attachment security at 24 months, and children’s self-control in early
childhood were also examined.

An examination of attachment quality at 15 months tests the hypothesis that
parental efficacy is related to parenting and attachment in predictable ways. Because
parental beliefs vary among demographic groups, dimensions of parental efficacy were
studied among low-income and middle-to-high income families and African American
and Caucasian families. Separate analyses for other ethnic or racial groups are not
included due to the limited number of Latinos, Native Americans, and Asians in the
study.

Second, a longitudinal model was used to test the pathways between parental
efficacy in infancy, parenting behavior, children’s attachment security at 24 months, and
children’s self-control in early childhood. Other factors that may affect child outcome,
such as marital conflict, are taken into account. Child outcome is measured by the
security of children’s attachment to their mother and performance on a self-control task.

Models compare processes between low-income and high-to-middle income
families only. The use of structural equation modeling requires a minimum sample size
of at least 150 families. Although the NICHD Study of Early Child Care began with 176
African American families, there were only 119 African American children who
participated in the procedure measuring child self-control at 36 months. This sample size
does not allow for separate models comparing developmental processes as stated above

between African American and Caucasian children.

General Research Questions

Attachment in Infancy

I. Can counterintuitive findings from earlier studies of the relation between parental
efficacy and attachment security be explained by reexamining dimensions of parental
efficacy, i.e., competence and control?

2. Does parental efficacy have different effects on parenting and attachment among
different demographic groups, specifically, low-income and middle-to-high income
families and African American and Caucasian families?

3. Does child care quality moderate the effects of maternal depression and parenting

practices on mother-child attachment as shown in Figure 1?

Social-emotional Development in Early Childhood

4. Are the data fiom the National Institute of Child Health and Development Study of
Early Child Care consistent with the model linking parental efficacy, parenting
practice, and children’s social-emotional outcomes as shown in Figure 2?

5. Does the model fit vary as a function of demographic status? For example, do the
relations among the variables differ for low-income and middle-to-high income

families?

Positive
Parenting

  
 
    
 

  

Child Care
Quality

Attachment
Security

 
     
    

Maternal
Depression

Figure I . Hypothesized model of interactions with child care quality predicting
attachment security.

Attachment
_ --------- Security

CUIMU ": \
Racism PM” m; Child
'~ " Self-Control

Child
Tern

 

 

 

 

Figure 2. Hypothesized model of pathways leading to child outcomes: self-control and
attachment security.

Research Assumptions

Importance of Parenting

Parenting practices are central to children’s developing competencies in many
different areas (Bomstein, 2002). In recent years, there has been some debate over the
relative importance of parenting practices in comparison to inborn traits and peer
influences (Harris, 2002; Scarr, 1998). Some child characteristics, such as various
dimensions of temperament, have shown heritability; however, secure attachment to
parental figures, an important factor in early personality development, has not shown
significant heritability (Vaughn & Best, 1999). Thus, in the development of mother-

child attachment, nurture or parenting is more essential than genetic endowment.

Influence of Parental Beliefs

In addition to parenting practices, it is also important to consider parental beliefs
about parenting. There has been some debate concerning differences between beliefs and
behavior. In both psychology and anthropology, some have argued that the relationship
between beliefs and behavior is difficult to detect (Harkness & Super, 1992). A relation
between cultural group values and beliefs is not disputed; however, individual variation
in practice within cultures may result from either selective attention to societal norms,
personal practices that differ from publicly expressed values, or outright nonconforrnist
behavior (Gjerde, 2001; Grusec, Hastings, & Mammone, 1994). However, the
assumption that parenting practices are influenced by beliefs about what constitutes
acceptable behavior is generally upheld (Sigel, McGillicuddy—DeLisi, & Goodnow,

1992). It is of particular interest that varying parental beliefs within cultural groups often

explain differences in parenting behavior, which may, in turn, improve the prediction of

children’s developmental outcomes (Harkness & Super, 1992).

Ecological Perspective

Other factors besides parenting also influence children’s development. Systems
theory suggests that children are affected by multiple experiences and settings, directly
and indirectly. Emotional regulation, fathering, marital conflict, and the physical
environment have been associated with children’s attachment behavior in recent
empirical research (Braungart-Rieker, Garwood, Powers, & Wang, 2001; Davies, Harold,
Goeke-Morey, & Cummings, 2002; Diener, Nievar, & Wright, 2003). From a broader
perspective, the National Research Council’s report on early childhood noted that
multiple factors—including parents’ psychological adjustment, marital relationships, the
child care environment, neighborhood, and culture--affect parenting, parent-child

interactions, and child development (Shonkofi’ & Phillips, 2000).

Justification of the Research
The relation between parenting behavior, the parent-child attachment relationship,
and child behavior has been previously researched. A meta-analysis of 66 studies from
diverse cultures indicated that maternal behavior styles influence the quality of parent-
child attachment relationships (De Wolff & van IJzendoorn, 1997). The quality of the
attachment relationship, in turn, is related to children’s emotional regulation, an

important factor in developing social skills (Calkins, 1994).

Although parenting behavior, attachment, and children’s self-control have been
researched, fewer studies have looked at how parental beliefs, such as parental efficacy,
relate to child outcomes. Among studies relating parental efficacy to attachment, results
are inconsistent (e.g., del Carmen, Pedersen, Huffman, & Bryan, 1993; Spieker & Booth,
1988; Teti & Gelfand, 1991). High parental efficacy is related to several parenting
activities, ranging from improving parenting skills through self-education (Spoth &
Conroy, 1993) to teaching children injury protection (Peterson, Farmer, & Kashani,
1990); however, some studies of parental efficacy and parenting behavior have produced
unexpected results and noted confounding effects (del Carmen, Pedersen, Huffman &
Bryan, 1993; Luster & Rhoades, 1989; Mash & Johnston, 1983; Spieker & Booth, 1988).
An overly strong perception of self-competence and a weak sense of self-competence
may both be related to suboptimal parenting behavior. However, a curvilinear relation
between perceived competence and parenting behavior has not been specifically tested.

Bandura (1982, 1986, 1994) has led the field in investigations of efficacy. His
early work acknowledged multiple facets of self-efficacy beliefs, differentiating between
perceived competence and control over outcomes, while acknowledging that they were
interwoven (Bandura, 1982; Ozer & Bandura, 1990). Applying these principles to the
domain of parenting, a self-perceived efficacious parent would need to believe: (a) in
their personal competence and knowledge about parenting, and (b) that parenting makes a
positive difference in children’s outcomes. Differentiation between perceived
competence and perceived control might also clarify inconsistent findings in the research

literature.

10

This dissertation investigates the mediators and moderators of the theoretical
process represented pictorially in Figure 2. Distal factors, such as the quality of marital
relationships, may affect attachment and children’s subsequent adjustment, through
pathways such as parenting behavior (Davies, Harold, Goeke-Morey, & Cummings,
2002). Other research groups have investigated the development of parental efficacy,
attachment, and child behavior; however, typically such research has used smaller
samples and focused on a small number of variables. Hence, fewer studies examine
interactions and mediational processes.

A notable exception is Teti and colleagues’ work with high-risk families. In this
series of studies, spousal support and child temperament moderated the relation between
parent-child characteristics and parental efficacy, and parental efficacy mediated the
relation between maternal depression and sensitivity (Teti, O’Connell, & Reiner, 1996).
A model using a larger and more heterogeneous group of participants may extend these
findings, allowing for the examination of processes among low-risk as well as high-risk
groups.

Additional research is needed to further define parental efficacy and explore its
relation to parenting practice and child outcomes. In particular, research is necessary to
untangle differences in developmental processes across diverse groups (Bronfenbrenner,
1999). What is normative in one cultural group or sub-group may not be appropriate in
another. Cultures differ in some of their core beliefs, making certain types of parenting
beliefs and behavior and certain types of child behavior more acceptable in some groups
than in others. A scarcity of resources may also make certain types of behavior more

acceptable or useful; some researchers have theorized that certain parenting behavior,

11

such as strict discipline, is adaptive for parents and children in African American families
(Deater-Deckard, Dodge, Bates, & Pettit 1996; Gonzales, Cauce, Friedman, & Mason,
1996). Alternatively, this effect may be accounted for by socioeconomic status (Polaha,
1999). Yet other processes in child development are theorized to be constant across
cultures or settings.

In order to investigate diverse perspectives in parenting, it is important to have a
sample that is similar to the population of interest. Although some research has been
done on parenting processes among African Americans in the United States, most of this
research has been among low-income inner-city families involved in intervention
programs (Mcond, Cauce, Takeuchi, & Wilson, 2000). Although studies of prevention
programs can teach us much about a subgroup of the population, it is important to
recognize that an over-generalization of such research may create confusion. Research
on single-parent African American families has been particularly emphasized in the last
decade, with two-parent families simply used as a comparison. Furthermore, studies of
parenting in Caucasian families often focus on middle-class samples that are also
unrepresentative.

The sample of participants in this dissertation reflects the population to a greater
extent than studies using local convenience samples. Specifically, data were collected
from families who participated in the NICHD Study of Child Care. The demographics of
these participants reflect national income levels and racial proportions (N ICHD, 1999a).
Thus, descriptions of family processes should be more representative of American

families than most previous work.

12

In addition, varied environmental conditions may create circumstances wherein
certain parenting practices are important for children’s well-being. Given that a family’s
available resources influence child development (Shonkoff & Phillips, 2000), this
dissertation will examine both low-income and middle-to-high income families. Given
that racism and ethnic differences in parenting behavior may influence family processes
(McLoyd, Cauce, Takeuchi, & Wilson, 2000), a model evaluating the influence of
parental efficacy on attachment security will contrast African American and Caucasian
groups. An ecological perspective assumes that factors such as ethnicity, neighborhood
quality, and marital quality may influence child characteristics (Bronfenbrenner, 1992).
Thus, distal factors, such as family income, and proximal factors, such as parenting
practices and child characteristics, may alter the strength of the relation between parental
efficacy and child outcome.

In sum, this research will add to our current understanding of attachment
relationships within diverse environments. Its emphasis on parental efficacy may explain
the motivation behind parents’ actions and the connection of parents’ oognitions to
parenting practice and child behavior. From an applied perspective, findings may assist

in the development of culturally appropriate parent education curriculum or home visiting

programs.

Overview of the Dissertation
Chapter 1 has presented a brief description of the research questions and the
significance of this research. The following chapter presents a literature review

describing the theoretical and empirical basis for this research. A discussion of

13

ecological theory provides a framework for understanding social cognitive theory and
attachment theory. Applications of these theories as found in the empirical research
literature are discussed. Group differences and similarities are also reviewed.

Chapter 3 presents the methods and measures used to test the research questions
described in Chapter 1. Chapter 4 presents empirical results, including frequencies,
correlations, analyses of variance, logistic regressions, and structural equation models.
Finally, Chapter 5 discusses these results, describes limitations, and suggests future

research directions.

14

CHAPTER 2

LITERATURE REVIEW

This dissertation draws on three areas of social science literature as a basis for its
model: (a) ecological theory, (b) social cognitive theory, and (c) attachment theory. A
description of these three theories and pertinent literature is followed by a review of
literature on the correlates of attachment, including characteristics of the individual child,
the home environment, the child care enviromnent, and distal factors. Literature
discussing the process of children’s social-emotional development within diverse cultures

is also presented.

Ecological Theory

Bronfenbrenner’s (1979) early ecological model focused on the environment and
its effect on the developing individual. Historically, psychologists in the I9608 and
19703 spent more time in laboratory settings than in natural environments. In fact,
clinical and research psychologists tended to discount effects of the total family, working
conditions, and neighborhoods, focusing on individual perceptions apart from their
environment. In contrast, Bronfenbrenner’s ecological model emphasized total
environmental effects.

Bronfenbrenner designated various levels of a child’s total environment as
interactive systems, including the microsystem, mesosystem, exosystem, and

macrosystem (Hamilton & Luster, 2003). The system likely to have the greatest effect on

15

a child was named the microsystem. By definition, the microsystem included the
individual child. The home environment is a microsystem with parents, siblings, toys,
and the individual child. Issues involving the home as a microsystem include home
safety and parent-child interaction. The child care environment or the neighborhood play
group are other microsystems that include the child at diflemnt times.

Intersections between microsystems were designated as the mesosystem.
Connections between home and school or school and child care are important to the
developing child. For example, a parent who has difficulties communicating with a
caregiver may miss important information that would help them as parents in their child’s
home environment.

The exosystem includes areas affecting the child indirectly. For example,
parental work demands may affect their volunteer work at their children’s preschool, thus
affecting the child. The microsystem, or the environment in contact with the child, is
affected by the mesosystem, which in turn is affected by the exosystem. The
macrosystem includes broader attitudes and ideologies, i.e., culture, socioeconomic
status, values, and government. In the previous example, the parent may be reluctant to
quit a demanding job because of a desire to maintain the family’s socioeconomic status.
The macrosystem may influence this parent’s beliefs to value spending time developing
their career over spending time volunteering in their child’s preschool.

Bronfenbrenner’s (1992, p. 197) later “person-process-context model”
emphasized the dynamic nature of development through interaction between people and
their environments. It is recognized that the environment affects children; however,

children make changes in their environment also. A more sociable child is more likely to

16

attract attention from teachers and peers. Researchers were encouraged to include both
the environment and the individual in their work, rather than focusing solely on the
influence of settings or the attributes of pe0ple.

Bronfenbrenner’s (2000, p. 129) current model, termed the “bioecological model”
accounts for heritable traits and the effect of the environment on the individual over time.
The environment may include objects, symbols, and people; it may be in proximal
contact with the individual as in the microsystem of the ecological model or it may be
more remote, such as the neighborhood or the government. The individual develops in
the context of the environment through proximal processes that vary as a function of the
person, the nature of developmental outcomes, and the continuities and discontinuities
that occur over time, i.e., process-person-context-time.

Furthermore, developmental processes are enhanced when they involve another
individual who has developed a close relationship with the child. Thus, this dissertation
will focus more on parent-child relationships and the home environment than child care
provider-child relationships and the child care environment. Family resources, child
temperament, and other characteristics of the person and environment will be included.

Although Bronfenbrenner’s (1979) emphasis on the study of development in
context was followed by his colleagues, he found that his overemphasis on the
environment as a context for development created a new problem, “a surfeit of studies on
context without development.” (Bronfenbrenner & Morris, 1998, p. 994). Thus, this
dissertation examines not only the context and the process of parenting but the
development of children. In particular, a focus on social-emotional development includes

attachment security and child self-control as outcome measures.

17

Ecological theory provides a framework for this dissertation; attachment theory
and social cognitive theory provide a basis for investigation. Attachment theory attempts
to explain the relation between the child’s environment, parental behavior, and child
behavior (Cassidy & Shaver, 1999); social cognitive theory suggests that parental
efficacy mediates between the environment and parental behavior or between the
environment and attachment (Bandura, 1995; Donovan & Leavitt, 1989; Gondoli &
Silverberg, 1997; Luster & Rhoades, 1989; Teti & Gelfand, 1991). The following section

discusses social cognitive theory and attachment theory as a basis for this research.

Social Cognitive Theory

Rotter, an early leader in the study of efficacy, developed the idea of a locus of
control (1990). An internal locus of control was defined as the perception that a person’s
behavior affected his environment. An external locus of control was defined as the
perception that luck, chance, or external forces beyond personal control afl‘ected a
person’s environment.

Instead of using the language of locus of control, Bandura (1982) explained
personal control, or the perception of a connection between efforts and outcomes, as self-
efficacy. A major tenet of Bandura’s social cognitive theory is that learning occurs by
91Mngggdeled_behayi9r; gertainly observation (of other individuals: effectiveness
~1;e_l_2_1_te_ks_t(_iself-motivationaryiiieficflgacy. However, personal experiences of success or
failure are more relevant to motivation. According to Bandura, individuals’ beliefs about
the connection between their efforts and the desired outcomes affect their motivation to

act. Theoretically, those who have higher levels of self-efficacy are more likely to be

18

motivated to succeed at a given task. for example, parents who believe that they can
affect their children’s actions by their efforts (e.g., help them learn to talk by talking to
them) are more likely to spend time and effort in parenting. A second dimension of self-
efficacy, also discussed by Bandura (1982, 1989), involves perceived competence. For
example, parents who felt that they were skilled and capable in handling their children
are more apt to spend time engaged with their children because of the personal
satisfaction they receive.

In his later work, Bandura cited attribution theory to differentiate the relation
between perceived control and perceived competence: “Self-efficacy beliefs [perceived
competence] influence causal attributions [perceived control]. People who regard
themselves as highly efiicacious attribute their failures to insufficient effort, those who
regard themselves as inefficacious attribute their failures to low ability” (Bandura, 1994).
Although Bandura has redefined the relation between perceived competence and
perceived control, research continues to use varied conceptualizations of efficacy and its
dimensions.

In addition to problematic conceptualizations within the field, a debate about the
boundaries of self-efficacy exists (Coleman & Karraker, 1997). Is self-efficacy a
personality trait, or is it a state that varies over time? Self-efficacy theory developed by
Bandura describes efficacy as a domain-specific construct; Harter (1978) prefers a global
orientation. According to Bandura (1989), efficacy in one area does not apply to efficacy
in another; however, a multi-faoeted approach to efficacy measurement may improve
measurement capabilities. In a previous study, a latent construct of multiple measures of

child efficacy was a significant predictor in structural equation modeling, although the

19

measures of efficacy were not intereorrelated (Bandura, Barbaranelli, Caprara, &
Pastorelli, 1996). Individuals may feel competent in one area but may feel deficient in
another; similarly they may feel that they have control in some situations but not in
others. Bandura has suggested that measures be constructed from task-specific items,
such as “I feel capable of soothing my child when he/she is upset,” rather than global

items, such as “I feel like I am a good mother.”

Parental Eflicacy
Self-efficacy that encompasses parenting is defined here as parental efficacy.

Parental eflicacy, or the belief in one’s ability to parent effectively, is composed of
various domains/I.<Bandura and colleagues have noted that efficacy may be divided into
multiple domains, such afmotivationfcognitive resources, and thetxercise of control
over events (Ozer & Bandura, 1990). This review of parental efficacy is largely
concerned with perceived control and perceived competence.

Although conceptualizations and specificity vary between studies, the literature

suggests that parental efficacy is related to parenting and child outcome. Early research
J- Vfi- KM _

 

 

found that low parental efficacy was related to children’s difficult temperament
(Bugental & Shennum, 1984). A review of research noted that parental self-efficacy has
emerged as a powerful correlate of parenting skills and a mediator of the effects of
varied constructs related to child outcome, including maternal depression, child
temperament, social support, and poverty (Coleman & Karraker, 1997). Applied
research has also underlined the importance of parental efficacy. Recent interventions

with clinical populations of young children (e.g., ADHD, autism, Asperger’s syndrome)

20

have noted associated improvements among parental efficacy, parenting, and child
outcome (Hastings & Brown, 2002; Hoath & Sanders, 2002; Sofronoff & Farbotko,
2002).

Although developmental research continues to show a trend toward examining
parental beliefs as well as parental behavior (Smetana, 1994), research on parental
beliefs is still an emerging field. As sucth varied definitions and operationalizations of
parental efficacy exist (Lovejoy, Verday & Hays, 1997). The construct of parental
efficacy has been alternatively described as parent attributions (Bugental, Blue, &
Cruzcosa, 1989; Deutsch, Ruble, Fleming, Brooks-Gunn, & Stangor, 1988), parental
locus of control (Carnpis, Lyman, & Prentice-Dunn, 1986; del Carmen, Pedersen,
Huffman, & Bryan, 1993), parental sense of competency (Abidin, 1986; Johnston &
Mash, 1989), and parenting self-agency (Durnka, Stoerzinger, Jackson, & Roosa, 1996).

From Bandura’s perspective on perceived competence, it may be hypothesized

k that parents who feel competent in specific tasks are more likely to derive satisfaction
from parenting. Thus, they may have a higher level of motivation than those parents
who do not feel competenthhis motivation is assumed to improve their parenting
abilities. Parents’ perceived control over child outcomes, their expectations for
themselves as parents, and prior experiences in teaching their children all affect personal
perceptions regarding competence.

Yef?’ parents who‘feel competent in their roles as mothers may not actually
behave in asensitive and responsive manner to their children’s needs (Luster &
Rhoades, 1989). In this study, adolescent mothers who rated themselves highly on

parental competency actually provided a less positive environment for their children

21

than mothers who viewed themselves as less competent parents. Those who felt less
competent, according to the researchers, may have had a greater understanding of the

difficulty in providing a nurturing environment as a young single parent.

Diversity and Parental Efi‘icacy

Stresses that occur in the daily life of parents who are unsuccessful in providing
subsistence to their families are likely to affect their expectations and their sense of
efficacy. Low-income parents may have low self-efficacy because of adverse
environmental conditions or depression (Olds, 1997). Previous efforts to change their
behaviors or circumstances may have met with resistance, thus teaching them low self-
efficacy by experience. In fact, parents are more likely to have negative beliefs about
parental involvement and efficacy in lower socioeconomic classes (Luster & Kain,
1987). )K

Parental efficacy may also be a stronger correlate of parenting and child outcome
in less advantaged families. Theoretically, parental efficacy may make more of a
difference among children who experience risks or disadvantages. Risk could come
through multiple routes, either through the macrosystem (e.g., racism, socioeconomic
status), the exosystem (e.g., dangerous neighborhoods) or through the microsystem (e.g.,
witnessing marital conflict). Strong parental efficacy, or the associated responsive
parenting, may then act as a protective factor against risk for some children.

Bronfenbrenner (2000) noted that proximal processes have differential effects
dependent on the level of risk in the environment. He also theorized that parental

attention related to developmental dysfunction would be greater in deprived

22

environments, assuming that parents are less vigilant in low-risk environments, where the
probability of poor child outcome is lower. Although the reasons for this effect have not
been established, differential effects of parental efficacy within varied contexts have been
researched.

A few recent studies support the relative importance of parental efficacy. A study
of parental efficacy among middle-class families showed no relation between efficacy

and parenting practgsgs (Corapci & Wachs, 2002). Among Head Start families, parental

 

efficacy mediated the relation between children’s difficult temperament and Mg
1W (Machida, Taylor, & Kim, 2002?? In a study of families living in
inner-city neighborhoods, parental efficacy was a predictor of positive parenting among
African American families; however, there was no relation between parental eflicacy and
parenting among Caucasian families (Ardelt & Eccles, 2001). It was noted that within
this disadvantaged group, African American families tended to live in more dangerous
neighborhoods than Caucasian families. In addition, parental efficacy was a stronger
correlate of academic success in African American single-parent families and married
families reporting husband-wife conflict than in Caucasian families and Afiican
American families who reported having a compatible marriage relationship. In the study
of parental efficacy, it may be important to view processes separately within
socioeconomic and cultural groups in order to fully understand parental belief systems,

parenting, and children’s social-emotional development.

23

Attachment Theory

Attachment theory provides a framework for understanding children’s social-
emotional development in the context of the parent-child relationship. Bowlby’s
attachment theory proposed that children use their caregivers as a secure base that
provides comfort in times of stress (Bowlby, 1969). From an evolutionary perspective,
children who remained close in proximity to their parents were more likely to be
protected from danger. A strong emotional attachment with the caregiver deve10ped as a
means of protecting the infant, and attachment behavior is organized around this parent-
child relationship.

Two measures of attachment security are presumed to be valid determinants of
attachment security (De Wolff & van IJzendoorn, 1997; Vaughn et al., 1992). First, the
strange situation classifies different types of attachment: 1) secure, 2) insecme—resistant,
and 3) insecure-avoidant. In later years, a disorganized classification has been included
in some studies (Main & Solomon, 1990). The strange situation is a scripted laboratory
procedure that induces mild stress for the infant; subsequent child behavior in the
presence of the caregiver, most often the mother, is observed and classified (Ainsworth,
Blehar, Waters, & Wall, 1978). To support of the validity of the strange situation,
qualitative differences in maternal behavior are related to strange situation
classifications (Isabella, Belsky, & von Eye, 1989). Specifically, mothers with insecure-
avoidant children tended to be more intrusive than mothers with secure children or
mothers with insecure-resistant children. Mothers with insecure-resistant children

tended to be emotionally or physically unavailable. Mothers with secure children are

24

more sensitively responsive; however, these mothers have less frequent interaction with
their children than mothers with insecure-avoidant children.

The second attachment measure, the Attachment Q-Set, has more recently been
established as a valid measure of attachment security (Solomon & George, 1999). This
measure rates secure behavior on a continuous scale; qualitative difi‘erences between
insecure attachment types are not measured (Waters, 1995). The validity of mother
ratings of her own child has been questioned; however, trained observer ratings of
attachment security are considered to measure the same construct as the Strange
Situation (De Wolff& van IJzendoorn, 1997).

It is important to note that both measures assess child behavior. Although
attachment is theoretically based on the mother-child relationship, these measures assess
child behavior with respect to that relationship. Hence, the child’s security is measured
separately from the mother’s input to the relationship.

Although research has largely verified the theoretical relationship between
parenting and attachment security, ongoing debate regarding the scope of attachment
theory has raised interesting questions. Some researchers have viewed attachment as a
process rather than a state; qualities of attachment and attachment behavior may change
over time. Others have suggested that infant-attachment security is not a child outcome
because children may have more than one attachment relationship with multiple
caregivers; for example, a child may be insecurely attached to the mother and securely
attached to the father or child care provider (Sagi et al., 1986; Sroufe, 1985).

Insecure attachment styles were often viewed as dysfunctional; however, later

research included a disorganized attachment designation. With the development of a

25

disorganized attachment class, it was noted that when attachment strategies exist,
whether secure or insecure, they provide a way of organizing behavior that allows for a
functional relationship. Alternatively, disorganized attachment is theorized to be
maladaptive; it is associated with parental depression, infant maltreatment, and later
hostile and aggressive behavior in preschoolers (Lyons-Ruth, 1996; Main & Solomon,
1990; Teti, Gelfand, Messinger, & Isabella, 1995).

Preferred attachment strategies may difl‘er between cultures; in fact, some
researchers believe that insecure-resistant attachment may be more adaptive than secure
attachment among other cultures (Crittenden, 2000; Hinde & Stevenson-Hinde, 1990;
Kondo-Ikemura, 2001 ). It is important to note, however, that insecure attachment is
often associated with child behavior problems in studies of children in Western cultures.
One such study found that pathways between insecure attachment and externalizing and
internalizing behavior problems showed a strong effect of attachment, even when
controlling for parenting processes and marital conflict (Davies, Harold, Goeke-Morey,

& Cummings, 2002).

Correlates of A ttachment

Over the past fifty years, researchers have elaborated on Bowlby’s initial work
on attachment correlating anti-social adolescent behavior and maternal separation. This
literature review discusses research on correlates of attachment, beginning with child
characteristics, including temperament and self-control, and parental factors, including

beliefs, practices, and parental psychopathology. More distal elements of the

26

environment, including marital relationships, culture or race, and socioeconomic class
are also considered.

Temperament. Some aspects of children’s behavior may be based on their
temperament, which includes traits such as negative mood, fearfulness, emotionality,
proneness-to—distress, and sociability (Goldsmith et al., 1987). Thomas and Chess
(1977) define temperament as behavioral style. Temperamental qualities they studied
include rhythmicity, activity level, attention span-persistence, distractibility,
adaptability, approach-withdrawal, threshold, intensity of reaction, and mood. In
addition, they equate temperament with the style of behavior, or the “how” something is
done. One enduring portion of their temperament theory is the goodness-of-fit model
(French, Rodgers, & Cobb, 1974). In this model, positive child outcomes occur when
the temperament of a child is consonant with the demands of the environment.

Empirical studies of the direct relationship between infant temperament and
attachment result in mixed findings (Bretherton, 1985). Two studies indicated that the
relationship between attachment and some aspects of temperament was stronger than the
relationship between attachment and maternal sensitivity (Seifer, Schiller, Sarneroff,
Resnick & Riordan, 1996; Teti, Nakagawa, Das, & Wirth, 1991). Seifer et al. (1996)
used mother reports and observational measures of infant temperament. Observation
measures of negative mood and total difficulty were both associated with low scores on
the Attachment Q-set (AQS), indicating insecure attachment. Other aspects of
temperament such as approach, activity and intensity were not significantly related to

attachment.

27

In a meta-analysis of 18 temperament and attachment studies, proneness-to-
distress was significantly related to insecure-resistant attachment, but not to insecure-
avoidant attachment (Goldsmith & Alansky, 1987). The authors theorized that infant
proneness-to-distress (e. g. fussiness, crying) could be a reflection of maternal behavior.
Caregiving practices may modify the behavioral expression of traits such as proneness-
to-distress or negative mood. Other contextual factors, such as infant health, are likely
to influence this aspect of temperament. Some studies reviewed found significant
relationships between attachment security and other aspects of temperament, such as
fearfulness (e.g., Thompson & Lamb, 1983). Theoretically, a fearful child may be less
likely to explore the surroundings, thus interfering with secure base behavior.

Teti et al. (1991) used scales on the Parenting Stress Index to assess various
aspects of child temperament. High scores on the child domain were strongly related to
low security as measured by the Attachment Q-Set. Related subscales measuring
various aspects of temperament include adaptability, demandingness, mood, and
distractibility. High scores in these areas, indicating low adaptability, demanding
infants, negative mood, and distractible or hyperactive infants, predicted low attachment
security. An additional study found a significant relationship between attachment type
determined by the Strange Situation and a combination of mother and observer reports
of temperament measuring sadness, anger, and negative emotionality (Izard, Haynes,
Chisholm, & Baak, 1991). Thus, some studies indicate a relationship between
temperament and attachment quality.

The previously discussed goodness—of-fit model can be applied to the

relationship between temperament and attachment. Two studies found that the

28

interaction between an inadequate environment and infant irritability related to insecure
infant attachment (Crockenberg, 1981; Mangelsdorf et al., 1990). The earlier study
found that low social support combined with high newborn irritability was significantly
related to insecure attachment (Crockenberg, 1981). Yet, temperament alone was not
significantly related to security of attachment. A second empirical test of the goodness-
of-fit model showed that the interaction between maternal personality, representing part
of the infant’s social environment, and infant proneness-to-distress temperament was
associated with security of attachment (Mangelsdorf et al., 1990). The correlation
between infant temperament and attachment quality without the addition of maternal
personality was not significant.

Other studies also formd that temperament determined the type of attachment
classification rather than security of attachment (Braungart & Stifter, 1991; Calkins &
Fox, 1992; Kochanska, 1998; Susman-Stillman et al., 1996). Susman-Stillman et al.
(1996) used path analysis in their study of 267 dyads. Infant sociability tended to
predict insecure-avoidant attachment. There was a trend for infant irritability to predict
insecure-resistant attachment. Mediator and moderator models indicated a complex
relationship between temperament and attachment. Maternal sensitivity mediated the
effects of infant irritability in this study. In addition, an interaction between irritability
and maternal sensitivity at 3 months was significantly related to security of attachment at
12 months. Theoretically, sensitive mothering would assist in the infant’s regulation of
temperamental extremes. Without a sensitive, responsive mother, proneness—to-distress

temperament may interfere with mother-child interactions.

29

Interestingly, alternative pathways to the development of conscience were found
to involve temperament as a moderator (F owles & Kochanska, 2000). Fearful children
were found to develop conscience through discipline; fearless children may develop
conscience through an alternate pathway of attachment. Later anti-social behavior may
be related to a fearless temperament combined with poor attachment relationships in
early childhood.

Some aspects of the relationship between temperament and attachment remain
unresolved. Yet, researchers have found direct and indirect relationships between
various attachment measures and negative mood, fearfulness, emotionality, proneness-
to-distress, and sociability. A growing consensus has developed that both caregiver
influences and child temperament are expressed in children’s behavior (Kochanska,
1998). Although some behavioral tendencies may represent relatively stable traits,
children’s behavior is presumed to be changeable over time and influenced by
environmental characteristics.

Self-control. One behavioral characteristic that may be influenced by the
environment is self-control. In the rubric of temperament, self-control may be
designated as a component of self-regulation; however, it is often not considered to be
temperament (Rothbart & Bates, 1998). Some cultures value self-regulation more than
others. Emotional regulation is important to social competence in the United States;
other societies are more open about the expression of feelings (Rothbaum, Pott, Azuma,
Miyake, & Weisz, 2000). It is of interest to note that fathers may fulfill an important
role in the development of children’s emotional regulation in Western cultures

(Braungart-Rieker, Garwood, Powers, & Notaro, 1998; Parke, 2002).

30

Researchers have identified two components to self-control: (a) passive self-
control associated with temperamental traits, e.g., fearfulness, inhibition, or caution, and
(b) active or effortful control associated with the development of conscience (Kagan,
1998; Kochanska, Murray, & Harlan, 2000). Parental discipline styles may be related to
the development of self-control. For example, maternal power assertion, such as
physical punishment or restriction, inhibited the development of effortful control in one
longitudinal study (Kochanska & Knaack, 2003).

Numerous other studies have used laboratory tests measuring self-control,
emotional regulation, impulsivity, or behavioral regulation by presenting the child with
an attractive stimulus, e. g., a toy, and subsequently asking the child to avoid or delay
interaction with the stimulus (for review, see Kochanska et al., 2000). Such studies
often report associations between performance on self-control tasks in early childhood
and later social-emotional behavior, including adjustment and attentiveness (Rothbart &
Bates, 1998).

Self-control in various forms has been associated with attachment security.
Early research linked secure attachment to greater compliance with maternal requests
(Londerville & Main, 1981). More recently, emotional regulation has been found to be
associated with attachment (Braungart-Rieker, Garwood, Powers, & Wang, 2001;
Cummings & Davies, 1996; Cassidy, 1994). In a longitudinal study of 94 families,
Braungart-Rieker and colleagues (2001) found that affect regulation discriminated
between types of insecure attachment classification. Infants with avoidant attachment

showed more affect regulation; infants with resistant attachment showed less affect

31

regulation. In addition, infant affect regulation mediated the association between
maternal sensitivity and attachment security.

Parental efi‘icacy. Insecure attachment quality has been found to have a counter-
intuitive relation to parental efficacy when it is measured as self-perceived competence
at parenting tasks. In one study, parents who felt extremely competent were more likely
to have an insecurely attached child (del Carmen, Pedersen, Huffman & Bryan, 1993).
Parental self-efficacy predicted the type of insecure attachment in an at-risk sample of
low-income families with mothers of insecure-avoidant children and disorganized
children having the most perceived competence (Spieker & Booth, 1988). Mothers
whose children were classified as insecure-resistant had less perceived competence, as
did mothers whose children were securely attached.

Among a sample of mostly clinically depressed mothers, the expected relation
between less perceived competence and attachment insecurity existed (T eti & Gelfand,
1991). This sample would have included few mothers who were intrusive; in fact, most
mothers would have been withdrawn. Those who felt least competent in this sample
would be more likely to be depressed and have lower self-efficacy rather than to have
externalizing problems coupled with dismissing questions of self-competence. The fact
that the counterintuitive relation between attachment and parental beliefs did not exist in
this sample supports the hypothesis that differences in perceived competence may
distinguish between two styles of parenting. In fact, differences in perceived
competence may differentiate between mothers of insecure-avoidant or disorganized

children and mothers of insecure-resistant children. Interestingly, when parental

32

efficacy was defined as perceived control, a stronger sense of parental efficacy was
related to secure attachment (Casady, Diener, Isabella, & Wright, 2001).

% Sensitivity. One correlate of attachment that received the most attention during
the early years of attachment research is the construct of sensitivity (De Wolff & van
IJzendoorn, 1997). The sensitive mother is defined as being attuned to her child’s
behavior, allowing her to accurately interpret, and respond appropriately to her child’s
signals (Bowlby, 1988). The sensitive mother is also able to observe the effect of her
behavior on her child and modify her actions to fit the child's needs. For example, the
infant touches the mother’s hand, and she extends a finger. The infant curls his hand
around the finger, and the mother moves her finger up and down. Another factor in the
construct of sensitivity involves the timing of response to the infant. Sensitive maternal
behavior has been compared to a dance. If the mother responds too quickly, she may
startle the infant. If she responds too slowly, the dance is interrupted. As in any social

_% interaction, the synchrony of social cues such as smiling or touch make a difference in
the amount of pleasure experienced. The same interaction could be viewed as either
playful or intrusive, depending on the continuity or discontinuity of the dyadic behavior.

Ainsworth, a leading figure in early attachment research, found a strong
association between maternal sensitivity and secure attachment (Ainsworth, Blehar,
Waters, & Wall, 197 8). Meta-analyses have confirmed that. sensitivity is related to
attachment; however, the grand mean correlations are in the small effect size range (De
Wolff & van IJzendoorn, 1997; Goldsmith and Alansky, 1987). The relative effect of
sensitivity may depend upon the method of measuring maternal behavior. Two studies

using home observation and measurements of maternal sensitivity that reflected

33

reciprocal and synchronous responsive behavior found a strong association between
sensitivity and attachment (Isabella, 1993; Pederson & Moran, 1996). Correlations
between maternal sensitivity and attachment in these studies were equal to or greater
than .50. Results from a group of longitudinal studies emphasized the importance of
measuring the appropriateness of maternal behavior rather than simple frequency of
response (Belsky, 1999). In addition, the strength of correlations between sensitivity
and attachment may depend on the context or culture.

Recently, questions have been raised about the appropriateness of the theoretical
sensitivity-attachment relation in non-Westem cultures (Rothbaum, Weisz, Pott,
Miyake, & Morelli, 2000). This article, however, did not present empirical evidence;
Rothbaum et al.’s hypothesis was contradicted in a later empirical test comparing
middle-class Columbian families and middle-class families from the United States
(Posada et al., 2002). A meta-analysis of 15 studies from countries other than the United
States found that sensitivity was related to attachment security regardless of the culture
(van Uzendoom & Sagi, 1999). Yet, additional tests of the cross-cultural
generalizability of the theoretical link between sensitivity and attachment may add to the
literature.

Discipline. Generally, physical punishment has been associated with negative
outcomes. However, a study of families living in poverty suggests that Afiican
American parents who engage in corporal punishment do not negatively impact their
children’s behavior (Deater-Deckard, Dodge, Bates, & Pettit, 1996). It is possible that a
child’s view of his parent’s behavior may make a difference in his outcome. A recent

article discussing the ecology of attachment suggests that different family members view

34

discipline and attachment-related issues in difiemnt ways (Hill, Fonagy, Saffier, &
Sargent, 2003). Some may see controlling behavior as frightening or playful, dependent
on the context as well as their own individual perspective. However, overtly harsh and
unpredictable discipline has been associated with poor child self-control and attachment
disorders, particularly disorganized attachment (Kart-Morse & Wiley, 1997; Gauthier,
2003).

Parental psychopathology. Clinical depression, particularly severe depression,
has been found to be associated with insecure attachment (Teti, Gelfand, Messinger &
Isabella, 1995). A recent meta-analysis of seven studies fiom European American
families confirmed this association (Martins & Gaffan, 2000). Overall, infants of
depressed mothers showed a reduced likelihood of secure attachment and a higher
likelihood of disorganized attachment.

One study found an interesting relation between depression, maternal behavior,
and child attachment security. Depressed and nondepressed mothers with infants who
were securely attached had similar emotional expression patterns in the dyadic
relationship (Radke-Yarrow, 1991 ). Thus, depressed mothers with securely attached
children appear to limit their expression of depression within the mother-child
relationship. In contrast, mothers with unipolar depression whose children were
insecurely attached exhibited high frequencies of anxiety, sadness, and downcast mood
when observed with their children. Their emotions were significantly different from
emotions of mothers in the control group who had insecurely attached children.

Another study measuring mother-child interaction among 70 depressed mothers

found that children with disorganized attachments were more frequently ignored by their

35

mothers than any other attachment group (T eti et al., 1995). Depressed mothers of
children with resistant attachments were more likely to blame their children for
misbehavior than depressed mothers of securely attached children. Mothers who self-
reported depressive symptoms in addition to having a psychiatric diagnosis were more
likely to show negative affect with their infants (Cohn & Campbell, 1992). There was
no significant relationship between attachment security and maternal depression when
mothers recovered fiom postpartum depression within 4 months. Yet, there was a
tendency toward attachment insecurity when mothers experienced longer lasting clinical
depression.

A major longitudinal study linking maternal depression and insecure attachment
was conducted through the Harvard Family Support Project (Lyons-Ruth, 1992a). This
project began as a study of risk factors for low-income families. Later, it was discovered
that 74% of the high-risk mothers referred to the intervention and 27% of the control
group met criteria for a diagnosis of clinical depression (DSM-III-R). Security of
attachment at 18 months was associated with lower scores on the Center for
Epidemiological Studies Depression Scale (CBS-D). Home observations at 12 and 18
months indicated that increased hostile and intrusive behavior towards the infant was
related to maternal depression.

It is apparent that various types of maternal depression tend to be related to
insecure attachment; however, depression does not always have the same effect on the
mother-child attachment. Maternal behavior appears to be a mechanism that impacts the
dyadic relationship. The chronicity, severity, and type of depression may also affect

attachment quality. In addition, maternal depressive symptoms, such as negative affect,

36

combined with stressful situations are more likely to result in insecure attachment than
an actual diagnosis of an affective disorder (Lyons-Ruth, 1992a). Lyons-Ruth (1992b)
suggests that stress may activate maternal depression as a manner of coping with adverse
situations. Stress and depression may also relate to feelings of hopelessness and a lack
of self-eflicacy.

Marital Quality. Several studies have found a positive association between
attachment security and marital quality (Belsky & Isabella, 1988; Crnic, Greenberg, &
Slough, 1986; Goldberg & Easterbrooks, 1984; Spieker & Booth, 1988; Teti et al.,
1991). Interactions with marital quality and other factors may also be important. For
example, maternal depression combined with a lack of spousal support is more likely to
result in child psychopathology than maternal depression combined with a supportive
spouse (Cicchetti & Toth, 1990).

The presence of a supportive spouse may be less important to children’s security
in cultures that place less emphasis on marital happiness (McLoyd, Cauce, Takeuchi, &
Wilson, 2000). Yet, it is well established that in Western cultures, conflict between the
father and mother, or even the absence of a father, is often related to children’s behavior
problems possibly due to an increase in maternal stress (Davies, Harold, Goeke-Morey,
& Cummings, 2002). Involved fathers may also add unique methods of interacting with
children to the family environment.

Shared caregiving. The process of mother-child attachment may differ
according to cultural practices of the family. For example, some cultmes practice shared
caregiving. In these cultures, the mother may not be the person most easily available to

the child. Among the Dogon of Mali, children are rarely left alone; however, multiple

37

caregiving arrangements exist (True, Pisani, & Oumar, 2001). Traditionally, the
grandmother rather than the mother cares for first-bom sons. In this study, about half of
the infants were considered to be cared for primarily by their mother. The Strange
Situation assessment indicated that no Dogon infants had avoidant attachment styles
with their mother; however, 13% of children had resistant attachment to their biological
mother (n = 47). As mentioned previously, resistant attachment is associated with
maternal unavailability. These children may have developed secure attachment to their
grandmother or another primary caregiver while developing resistant attachment to their
mother who was less frequently available to them.

Despite the issue of shared caregiving, a significantly higher proportion of the
infants in this study were secure when compared with the distribution of infant security
in a meta-analysis of studies from multiple countries and cultures (van IJzendoom &
Kroonenberg, 1988). This may be due to an emphasis on fiequent human contact and
instant response to distress among caregivers, an emphasis that is shared by other
Afiican tribes (Levine et al., 1994; Tronick, Morelli, & Ivey, 1992).

Although parallels may exist, findings from Afiican studies are not directly
applicable to Afiican Americans. Historically, African American mothers shared
caregiving with family and fiiends; they are still more likely to turn to family and
friendship networks for child care than Whites (Johnson et al., 2003). Yet, it is evident
that society in the United States places demands on mothers and children that differ fiom
Afiican culture. However, it is important to note that shared caregiving does not appear
to relate to attachment disorders within the Afiican culture; this suggests that shared

caregiving in and of itself does not lead to insecure attachment.

38

Shared caregiving in the United States is also experienced by children in child
care. Quality of child care may also be associated with attachment; however, the
complexity of the child’s total environment must be reviewed. Time in child care and
the quality of child care may be dependent on parents’ attitudes shaped by their culture.
In addition, the amount and quality of child care may impact mother-child attachment
security or children’s behavior, particularly when risk factors, such as insensitive
maternal behavior or low socioeconomic status are present (NICHD, 1997; Ritter &
Moyi, 2003). Child care appears to be more stressful for young children with a fearful
temperament than children with an easy-going temperament (Watamura, Donzella,
Alwin, & Oumar, 2003). Characteristics of children, parents, and the child care setting
all contribute to the effects of child care on child development.

Culture, ethnicity and race. Some cultures have different approaches to raising
children. For example, German families emphasize early independence. In a study of
German families, a higher proportion of German children displayed avoidant attachment
when compared to other cultural groups (Grossman et al., 1985). Japanese families
emphasize community and interdependence. A meta-analysis comparing proportions of
insecure-resistant children indicated higher percentages of resistant attachment among
studies of children in Japan than studies of Caucasian American children (van
IJzendoom & Sagi, 1999). The literature suggests that certain parenting styles are
related to children’s attachment behavior; hence, cultures that emphasize different
parenting styles may have children who respond differentially within caring

relationships.

39

Although multiple pathways to the development of close relationships may exist
based on cultural differences, the process of attachment, including proximity seeking,
maintaining contact, and separation discomfort is theorized to be universal and based on
biological needs (Crittenden, 2000; Rothbaum, Pott, Azuma, Miyake, & Weisz, 2000;
Rothbaum et al., 2001; van IJzendoom & Sagi, 1999). A cross-cultural comparison of
hypothetical ideal secure behavior used reports from child experts in seven countries
(Posada et al., 1995). Results indicated that the concept of attachment behavior is
similar across cultures. However, subtle differences may exist. For instance,
comparison of correlates of attachment in Colombia and the United States found that
mothers of secure Colombian infants were more likely to be concerned with their child’s
physical appearance; mothers of secure American infants were more likely to be
concerned with creating an interesting environment (Posada et al., 2002). Proportions
of attachment classifications may also vary between cultures and subcultures, and
certainly, the strength of correlation between certain constructs and attachment security
may vary between cultures and socioeconomic groups.

Socioeconomic status. In some cultures, certain aspects of the family ecology
may be important to attachment quality; these same conditions may be irrelevant to
attachment in other cultures. In the aforementioned study of Dogon families, income
was not related to attachment quality (True, Pisani, & Oumar, 2001). Studies of families
within the United States suggest that a greater proportion of insecurely attached dyads
exists in low-income neighborhoods than in middle class neighborhoods; however,
resources available to the securely attached Dogon children were fewer than the amount

of resources available to average low-income families in the United States.

40

Conclusion

Many domains of children’s environments may differ according to cultural or
socioeconomic orientation. Definitions of appropriate parenting styles, preferred
attachment styles, and the number of involved caregivers may vary; however, some
processes of development are theorized to remain consistent across groups.

In sum, parental efficacy, parenting strategies, and children’s social-emotional
development are reciprocal processes embedded within the context of the total
environment. Social cognitive theory suggests that caregivers who perceive that they are
in control of how their children turn out would be more motivated to be responsive to
their children. Attachment theory suggests that sensitive, appropriate responsiveness on
the part of the caregiver would be related to secure attachment. Insecure attachment may
relate to difficulties in self-control. Marital conflict, child care quality, and child
temperament may all relate to children’s development, which will be operationalized as
home and laboratory measures of observed behavior. The following chapter will describe
measurement of factors related to the attachment process, including parental beliefs,

practices, and environmental context within the United States.

41

CHAPTER 3

METHOD

This chapter discusses the methods used to test an ecological model of parent-
child attachment through secondary data analysis. First, specific hypotheses are
presented in two sections: a) hypotheses related to dimensions of parental efficacy, and b)
hypotheses related to the ecological model. A description of the sample is presented

followed by a description of the instruments used in this particular dissertation.

Hypotheses

Dimensions of Parental Efiicacy and Attachment in Infancy

In reviewing the literature, it appeared that two dimensions of parental
efficacy—perceived control and perceived competence-were related to parenting
behavior and attachment. In addition, it is possible that the different dimensions of
efficacy that have been used as a basis for past research are responsible for the mixed
results found among previous tests of parental self-efficacy theory. Tests of the
following hypotheses will attempt to clarify these dimensions and to examine correlates
of perceived control and perceived competence, particularly parenting practices, which in
turn may lead to attachment security and children’s self-control.

Hypothesis 1a: Parents with higher levels of perceived control are less likely to

exhibit detached or intrusive behavior.

42

Hypothesis 1b: Parents with higher levels of perceived control are more likely
than other parents to have infants rated as securely attached.

Hypothesis 1c: The relation between perceived control and secure attachment
will be stronger among low-income families than middle-to-high income families.

Hypothesis 1d: The relation between perceived control and attachment security
will be stronger among Afiican American families than Caucasian families.

Hypothesis 2: Perceived competence and parenting style will be related to
attachment classification.

Hypothesis 2a: Children displaying an insecure-avoidant attachment are likely to
have mothers with high levels of perceived competence.

Hypothesis 2b: Children displaying an insecure-resistant attachment are likely to
have mothers with low levels of perceived competence.

Hypothesis 2c: Children displaying a secure attachment are likely to have
mothers with an average level of perceived competence, greater than that of mothers in
insecure-resistant dyads and less than that of mothers in insecure-avoidant dyads.

Hypothesis 2d: Children displaying insecure attachment styles are likely to have
mothers who are less sensitive.

Hypothesis 2e: Relations between parental efficacy, parenting style, and
attachment vary between demographic groups.

For fiuther clarification of Hypothesis 2, a model of parental competency

perceptions, parental behavior, and attachment classification is displayed in Figure 3.

43

 

Averag
Percep'nns of
Com

 

 

Pawn bus of
Competen

 

Figure 3. Hypothesized model of perceived competence, responsive parenting, and
attachment classifications.

44

Hypothesis 3: For children in child care, child care quality may moderate the
effect of other variables on attachment.

Hypothesis 3a: The quality of child care may buffer the likelihood of insecure
attachment for depressed mothers.

Hypothesis 3b: The quality of child care may buffer the likelihood of insecure

attachment for mothers with insensitive parenting.

Ecological Model

Next, the model of parental efficacy, parenting, child self-control, and mother-
child attachment will be tested. Specific measures used to test this model (Figure 2) will
be determined by the results of preliminary variable tests. Different subsamples, such as
families with two parents, low-income families, and middle-to-high income families, will
also be tested to compare processes within diverse demographic groups.

Hypothesis 4: The model describing the development of mother-child attachment
and child self-control as displayed in Figure 2 will fit the data.

Hypothesis 4a-1: Mothers who have higher perceived control are more likely to
display positive parenting practices.

Hypothesis 4a-2: Mothers who have higher perceived control are less likely to be
depressed.

Hypothesis 4b: Mothers who are less depressed are more likely to display positive
parenting practices.

Hypothesis 4c: Families who experience less marital conflict are more likely to

have positive parenting practices.

45

Hypothesis 4d-1: Parents who display positive parenting practices (e.g.,
sensitivity) are more likely to have children with secure attachments.

Hypothesis 4d-2: Parents who display positive parenting practices are likely to
have children with greater self-control.

Hypothesis 4e-1: Children who are more securely attached are likely to display
greater self-control.

Hypothesis 4e-2: Children who have easier temperaments are likely to display
greater self-control.

Hypothesis 4f: Relations between parenting practices, attachment, and self-control

may vary between demographic groups.

Sample

The sample in this dissertation comes fiom the National Institute of Child Health
and Development’s Study of Early Child Care and Youth Development, commonly
referred to as the NICHD Study of Early Child Care. This longitudinal study was
developed to answer questions about child care experiences for young children. A
scientific review began in 1989 to establish a team of researchers from diverse areas of
expertise. Domains of study included children’s health, language, cognitive growth and
social-emotional development. Family environments and parental attitudes were also
part of the research design. In 1991, participants were enrolled at 10 sites throughout the
United States, including the University of Washington, the University of California,

Irvine, the University of Kansas, the University of Arkansas at Little Rock, the University

46

of Wisconsin, the University of Pittsburgh, the Western Carolina Center, the University
of Virginia, Temple University, and Wellesley College.

Although this is not a nationally representative sample, demographic status of
participants is roughly comparable to that of the nation as a whole. Yet, during sample
selection, mothers living in neighborhoods considered to be dangerous were excluded
from the study. Teen mothers were excluded, and families with children born with a
serious disability were excluded. Therefore, the sample is probably lower-risk than a truly
random sample of families.

Of the 1364 children in the database, 13% were African American and 80% were
Caucasian. The remaining 7% of the parents identified their children as belonging to
other racial groups. Of the total sample, only 6% reported Hispanic ethnicity. The 1097
Caucasians consisted of 1042 non-Hispanic whites and 75 Hispanic whites; 3 of 176
Afiican Americans reported Hispanic ethnicity. Children of Hispanic ethnicity were
removed fi'om the database for analyses using Caucasians and African Americans only.

The sample consisted of 705 boys and 659 girls; 45% of the children were first-
born. Mother’s age ranged from 18 to 46, with an average age of 28 (SD = 5.63).
During the first month after the child’s birth, 85% of the fathers lived with the family.
Other adult relatives in addition to the mother and father lived in 14% of the homes.
There were no missing data in the above demographic information. Less than 1% of the
data were missing for marital status and mother’s education. Regarding marital status,
77% of the mothers were married and 9% were living with a partner. Mothers reported a
10% high school dropout rate; 31% of the total group completed high school but

received no additional education beyond high school; 35% had completed four or more

47

years of additional education. Of the 92% of partners reporting education levels, 8%
were high school dropouts; 31% completed high school but received no further
education; 18% had completed four or more years of additional education.

Of the 93% of families reporting income information for the first month, 24%
were living in poverty based on an income-to—needs ratio; 11% of the families reported
no income. The percentage of families in poverty decreased over time; only 15% of the
families reported poverty-level incomes when the participating child was 3 years old.
This may be related to increased attrition among families in poverty as discussed below.
Income-to-needs ratios, the continuous measure used in this dissertation, represent total
income divided by the poverty level based on individual family size. Average ratios and

related statistics by time of data collection are displayed in Table 1.

Table 1

Income-to-needs ratios by time of d_ata collection

 

 

 

Standard Sample
Timepoint Meg devigtion Median Rme size
1 month 2.76 2.66 2.17 0 — 25.08 1274
6 months 3.53 3.19 2.88 0 — 27.89 1271
15 months 3.60 3.29 2.84 0 — 35.64 1234
24 months 3.62 3.11 3.01 0 — 27.30 1190
36 months 3.52 3.12 2.93 0 — 28.50 1208

 

48

The first wave of data was used for this dissertation, including multiple data
collection points occurring from the time of the child’s birth to 36 months. Table 2
provides descriptive statistics for selected measures administered during the first wave of
data collection. Permission to complete a secondary data analysis was obtained through
the Research Triangle Institute, which handles data access and dissemination for the
NICHD Study of Early Child Care. The Institute of Children, Youth, and Families at
Michigan State University under the direction of Dr. Marguerite Barratt had previously
purchased access to the data; however, additional permission to ensure confidentiality of

the participants was necessary for this dissertation.

49

Table 2

Descriptive sta_tr;stics of SIM variables

 

Measure Mean SD Range n
Parental Locus of 3.36 .49 1.64 — 5.00 1363
Control
Sense of Competence 4.00 .53 1.89 -— 5.00 1363
CBS-Depression
1-6 months 1 1.26 9.02 0 — 53 1363
15 months 9.05 8.18 0 — 54 1241
Global sensitivity 9.21 1.78 3 — 12 1272
Global stimulation 2.59 .66 1 — 4 1272
HOME
Enrichment 6.91 1.45 0 — 8 1234
Positive parenting 5.53 .89 0 - 6 1234
Lack of negativity 5.04 1.11 0 - 6 1234
Marital conflict 2.32 .93 1 - 6.78 1271
(mean item score)
Average of child care 22.12 4.81 8 — 37.5 773
quality assessments
Strange Situation 62.4% secure 1190
Procedure
Attachment Q-Set .29 .21 -.49 — .75 l 197
Self Control Procedure 0 .74 -1.11 — 2.62 1093
(composite z-score)
Temperament
Caregiver global 2.39 .60 1 — 3 504
Caregiver mean 3.29 .46 1.81- 4.67 397
Maternal global 2.35 .56 1 — 3 1362
Maternal mean 3.31 .65 1.00 - 5.36 1364

 

50

Attrition

Analyses of the data were divided into two major sets. The first part focuses on
measures in infancy. The final timepoint in the first set of longitudinal analyses is 15
months, when the Strange Situation Procedure was used to assess attachment
classification. A comparison of statistics pertaining to children participating in this
assessment (n = 174) and children not participating (n = 1190) yielded few results. Of all
study variables used for analyses in infancy, including income, maternal education,
parental efficacy measures, maternal depression, maternal sensitivity, positive parenting,
and child care quality, only income and education were significantly different for families
not participating at 15 months. Families not participating had lower income-to-needs
ratios (M = 2.21) than families participating (M = 3.42). Families not participating had
less education (M = 13.17 years) than families participating (M = 14.39 years). Attrition
was significantly higher for Afiican Americans (21%) than Caucasians (11%).

For the second set of longitudinal analyses, only children participating in the 24-
month attachment assessment or the 36-month self-control procedure were included.
Attrition within this sample (9.9%) was smaller than attrition at the 15-month assessment
(12.8%). Tests of all study variables, including income, maternal education, parental
efficacy measures, maternal depression, maternal sensitivity, positive/negative parenting,
enrichment, attachment at 24 months, and child self-control at 36 months, produced
significant differences between participants and non-participants on parenting and
enrichment scales. However, these differences were based on a small sample of families
(n = 34) who participated in 15-month assessment but did not participate in 24-month or

36-month assessments. These differences may not be representative of the total number

51

of non-participants in the later assessments (n = 162).

All other measures except for income and education were not significantly
different. Families not participating had lower income-to-needs ratios (M = 2.35) and
less education (M = 13.40 years) than families participating (income-to-needs ratios, M =
3.37; education, M = 14.33 years). Attrition was significantly higher for Afiican

Americans (17.0%) than Caucasians (8.6%).

Measurement
The measures in this dissertation were selected by investigators of the NTCHD
Study of Early Child Care. The Research Triangle Institute (RTT) is acting as data
managers for NICHD. Easily available data consist of constructed variables; however,
files of raw data at the item level were accessible through use of a codebook. Measures
were collected at multiple timepoints; more than one measure may be used for each
construct. For clarification, Table 3 lists the name of measure, related construct, and the

age(s) of children at the time of administration.

52

Table 3
Measures, constructs, and age ofchilflrt Ldministration
Measure Construct Ag of child in months

Parental Locus of perceived control 1

Control
Parenting Stress perceived 1

Index, Sense of competence

Competence scale
CES-D depression 1 6 15
Mother-child sensitivity 6

interaction ratings
HOME enrichment,

positive/negative
parenting 15

Love and marital quality 1

Relationships
Family Finances income-to-needs l 6 15 24 36
Assessment Profile child care quality 6 15

for Early Childhood

Program
Assessment Profile child care quality 6 15

for Family Day Care
Strange Situation attachment quality 15

Procedure
Attachment Q-Set attachment security 24
Self Control child self-control 36

Procedure

 

53

Perceived control. The Parental Locus of Control scale, a measure of perceived
control, was administered to mothers when children were 1 month old. A factor analysis
of the original 47-item scale indicated five factors (parental efficacy, parental
responsibility, child control of parent's life, parental belief in fate/chance, and parental
control of the child's behavior). Mothers rated statements, such as, “I have often found
that when it comes to my child, what is going to happen, will happen,” and, “I feel like
what happens in my life is mostly determined by my baby,” on a 5-point Likert scale.
The 20 items included in the NICHD data set were the four items that loaded highest on
each of the five factors. Cronbach’s alpha for the shortened 20—item version of the
Parental Locus of Control scale was .62.

For this dissertation, a subscale was created focusing on factors of parental belief
in fate/chance, child control of parent’s life, and parental efficacy. Items related to
parental responsibility and control of child’s behavior, such as “Parents who can’t get
their children to listen to them don’t understand how to get along with children,” were not
included. These items had a secondary focus of parental attitudes toward discipline; in
addition, they appeared to evaluate the parent’s stance on the nature vs. nurture issue.
Thus, they were not considered wholly appropriate for defining the locus of control
concept. Appendix A lists the items included in this subscale. A high score on this
subscale represented an internal locus of control or high levels of parental efficacy on the
dimension of perceived control. Cronbach’s alpha for the l4-item subscale was .69.

Perceived competence. Parenting Stress Index (PSI), a well-established measure

of parenting-related stress, contains a subscale named Sense of Competence (Abidin,

54

1986). The Sense of Competence subscale consists of nine items measuring parental
efficacy as perceived competence. Cronbach’s alpha was .76 at 1 month and .71 at 6
months. The l-month data were used in this dissertation in order to accurately compare
this measure to the Parental Locus of Control scale, which was completed at the 1-month
collection time only. Mothers rated items such as “Being a parent is harder than I
thought it would be” on a 5-point Likert scale (See Appendix A for additional items).
Items were reverse scored so that a high score reflected a positive sense of competence.
These two measures, Parental Locus of Control and the PSI, will allow for comparisons
of two types of efficacy, perceived control and perceived competence, and their relation

to other constructs.

Parental Characteristics

Maternal depression. The Center for for Epidemiological Studies Depression
Scale (CES-D) is used to measure maternal depression. The CES-D is a widely-used
measure of depression in non-clinical samples. The NICHD Study of Early Child Care
repeated this measure at all timepoints. In this dissertation, two measures of depression
were used. First, an average of the 1-month and 6-month score was tested in relation to
attachment quality at lS-months. Second, the 15-month score was tested in relation to
attachment security at 24 months.

Maternal depression is measured by self-report of the occurrence of 20 symptoms
over the past week. Sample items are: “I felt sad” and “I was bothered by things that
usually don’t bother me.” Additional items are included in Appendix A. Points of the

item scale included 0 (rarely or none of the time-Jess than 1 day), 1 (some or a little of

55

the time—1 - 2 days), 2 (occasionally or a moderate amount of time—3 - 4 days), and 3
(most or all of the time—5 - 7 days). A total score is computed by summing individual
item responses. Cronbach’s alpha was .85 in the general population (Radlofi‘, 1977,
Roberts, 1980). No group differences in internal reliability or factor structure were found
in scales completed by African Americans and Caucasians (Roberts, 1980).

Sensitive maternal behavior. For this dissertation, sensitive maternal behavior in
infancy was measured with global ratings of maternal behavior. Trained observers
completed ratings from videotapes of two semi-structured play activities. The first
activity was filmed in the laboratory at 6 months; a similar assessment was made during a
home visit at 15 months. Global 4-point rating scales devised by Owen and Vandell were
adapted to children at different ages (NICHD, 1999b). Scales were reduced to two binary
measures of negative parenting and detachment, as well as two continuous ratings of
sensitivity and stimulation. Sensitivity was a composite based on 4-point ratings of
sensitivity to nondistress, positive regard, and intrusiveness (reverse scored). Stimulation
was a global 4—point measure. High scores indicated more sensitive behavior and more
maternal stimulation. Ratings of negative parenting, detachment, sensitivity, and
stimulation at 6 months of age were used for preliminary analyses.

Parenting. The Home Observation for Measurement of the Environment
(HOME) inventory measures parenting practices, such as child discipline (Caldwell &
Bradley, 2000). The infant-toddler version given at 15 months was used to represent
parenting in the model describing the development of self-control and attachment in early
childhood. The HOME, a combination of observation and self-report survey items, is one

of the most widely used measures of the home envinonment. It is composed of 45 items

56

clustered into six subscales, including Parental Responsivity, Acceptance of Child,
Organization of the Environment, Learning Materials, Parental Involvement, and Variety
in Experience. Cronbach’s alpha for these subscales ranged from .38 to .89.

The Research Triangle Institute created three subscales from the HOME to
enhance reliability, including Home Enrichment, Positive Parenting, and Lack of
Negativity (NICHD, 1999a). These three subscales were used to represent different
aspects of the home environment. Each item is scored in binary fashion (yes/no).
Examples of items from the Positive Parenting scale include, “Parent’s voice conveys
positive feelings toward child,” and, “Parent responds to child’s vocalizations.”
Cronbach’s alpha was .56. Items from the Lack of Negativity scale include, “Parent does
not shout at child,” and “Parent is not hostile.” Cronbach’s alpha was .54. A high score

reflects a more enriched home environment; Cronbach’s alpha was .69.

Contextual Characteristics

Marital conflict. The Love and Relationships questionnaire is a self-report
measure assessing marital quality and conflict (Braiker & Kelley, 1979). A 25-item
questionnaire consists of four subscales, including love, maintenance, ambivalence, and
marital conflict. Sample items include “To what extent do you have a sense of
‘belonging’ with your partner?” and “How often do you and your partner argue?” This
questionnaire was administered to mothers and fathers when their child was 1 month old.
In a previous study, internal reliability for the four subscales averaged .76 (Belsky, Lang,
& Rovine, 1985). Questions are anchored on a 9—point scale, ranging from very little or

not at all to very much or extremely. Both fathers and mothers completed this measure.

57

A subscale reflecting marital conflict was computed for this dissertation. The 9 items in
this subscale included such questions as, “How often do you feel angry or resentful
toward your partner?” (See Appendix A for all subscale items.) A high score reflected
higher levels of marital conflict. Cronbach’s alpha was .82 for mothers and .80 for
fathers. Partners did not have to be married to complete this measure.

Family income. The Family Finances measure is a self-report measure created for
this study. This measure includes questions about income that may not always be
reported, such as public assistance and total family income. Based on four questions,
NICHD calculated poverty income-to-needs ratios for all five data collection timepoints.
For some analyses, two samples were compared consisting of low-income and middle-to-
high income families. An average income-to-needs ratio across timepoints was used to
construct these two groups. In previous research, low-income has been defined as 150%
or even 200% of poverty (e.g., Hofferth, Smith, McLoyd, & Finkelstein, 2000). This
dissertation used a more conservative figure of 150% of poverty (n = 358). Middle-to-
high income was defined as more than 300% of poverty (n = 600).

Child care. The Assessment Profile for Early Childhood Program and The
Assessment Profile for Family Day Care, measurements of child care quality, were
administered when children were 6 and 15 months of age. Programs were observed only
if the target child was in care for an average of at least 10 hours a week. High scores
indicate good quality of care.

The Assessment Profile for Early Childhood Program (APECP) uses observation,
documentation, and child care provider reports to assess center quality (Abbott-Shim &

Sibley, 1992). Multiple dimensions, such as the dimension of Safety and Health, are

58

rated on a series of specific items coded as present or absent. For the NICHD Study of
Early Child Care, the number of dimensions on the standard APECP was reduced to three
(Safety and Health, Learning Environment, and Individualizing) for the 6 and lS-month
assessments. These measures have been used in multiple studies of child care center
quality. In standardizing the measure, only items with part-whole correlations of at least
.40 were retained, reducing the number of items to 41.

The Assessment Profile for Family Day Care (APFDC) is the counterpart to the
APECP for children in home-based child care. This measure was also shortened by
matching its items to similar or identical items on the shortened version of the APECP.
For this study, an average of the 6-month and 15-month assessments was used to
represent quality of early care. Because of the similarity of measures, family child care
scores and center child care scores on quality assessments were used interchangeably.
Child characteristics

Temperament. Temperament measurement consists of two parent/caregiver self-
report measures. These include the Early Infancy Temperament Questionnaire (Medoff-
Cooper, Carey, & McDevitt, 1993) and the Infant Temperament Questionnaire Revised
(Carey & McDevitt, 1978). The Early Infancy Temperament Questionnaire was in the
process of development when the measures were being selected for this study. It is
similar in construction and reliability to the more widely used Infant Temperament
Questionnaire. A test of 40 l-month-old infants indicated the following reliabilities for
the dimensions of the Early Infancy Temperament Questionnaire as measured by
Cronbach’s alpha: activity (.48), approach (.44), adaptability (.65), intensity (.43), and

mood (.70). Parents completed a shortened version of the Early Infancy Temperament

59

Questionnaire (39 items) when their child was 1 and 6 months old; child care providers
completed a shortened version of the Infant Temperament Questionnaire Revised (43
items) when the child was 6 months old.

Attachment quality in infancy. Mother-child attachment was measured using the
Strange Situation Procedure at 15 months (Ainsworth, Blehar, Waters, & Wall, 1978).
This procedure consists of a series of 3-minute episodes involving the mother, child, and
a person tmfamiliar to the child. In the first two episodes the mother and child enter the
playroom and explore the toys. In the third episode a stranger is introduced into the
playroom. The mother leaves the child alone with the stranger for three minutes. Ifthe
child is distressed when the mother returns, the procedure is discontinued. Three minutes
later the mother leaves again, and the child is left completely alone for flrree minutes.
Finally, the child is reunited with the mother. Trained coders rated videotapes of the
procedure. Based largely on the child’s behavior at reunion with the mother, attachment
classifications were assigned as insecure-avoidant, secure, or insecure-resistant.
Disorganized behaviors were also coded; children with scores greater than 6 were
classified as disorganized.

Attachment security in early chilflrood. The Attachment Behavior Q-Set (Waters
& Deane, 1985) measured attachment security at 24 months. This instrument includes
90 items that are used to assess the quality of the mother-child relationship. Sample
items include: “When child is near mother and sees something he wants to play with, he
fusses or tries to drag mother over to it,” and “Child readily shares with mother or lets her
hold things if she asks to.” During a 2-hour home visit, trained observers made notes on

child behavior viewed in naturally occurring or semi-structured situations. Semi-

60

structured activities included a small book with surprise windows, a snack, and a hide-
and-seek game.

Immediately following the observation, observers divided the 90 items on cards
into nine piles with a fixed distribution. The placement of the pile indicated which items
were least like the child, somewhat like the child, and most like the child, representing a
nine-point Likert scale. Items that were most like the child were scored as 9. To obtain a
total score, the placement of each item by the observer was correlated with a criterion sort
consisting of expert opinion of an ideal sort of the “most secure child.” Developmental
psychologists provided ideal security sorts for children. If the subject was more similar
to the hypothetical “most secure child,” the mean of the correlations between items was
higher. The resulting number ranges fi'om -—1 to +1, with +1 representing the most secure
attachment possible.

Self-control. Self-control was measured as a laboratory assessment at 36 months.
Researchers videotaped and coded children’s performance on a self-control task that has
been used in previous research. In this task, a researcher asks a child to delay handling
an attractive toy while the mother is present. Behaviors that were coded included (a) the
length of time the child refi'ains fiom touching the toy following initial instructions fi'om
the experimenter, (b) the amount of time the child spends actively playing with the toy
after being asked not to touch it, (c) the amount of time the child spends tentatively
and/or furtively touching the toy, and (d) the child’s focus of attention (social, elsewhere)
when not touching the toy (NICHD, 1999a). A composite measure was created
averaging z-scores of time engaged in various behaviors, including latency to first active

engagement, total away time, focus on toy, social focus, and a reverse-scored measure of

61

time spent actively playing with the toy. This five-item measure had internal reliability

of .79.

Data Analysis

The objective of this dissertation was to examine relations between parental
efficacy, child environment, and children’s social-emotional development. Preliminary
analyses determined demographic characteristics of the sample, the amount of data
available, and the reliability of certain measures. Chi-square tests, MANOVAs, and
correlations among variables tested some preliminary hypotheses. A multinomial logistic
regression examined the relation between dimensions of parental efficacy and attachment
quality. Finally, structural equation modeling tested the model depicting precursors of
social-emotional development, using a latent construct of children’s self-control and a
continuous measure of early childhood attachment behavior as outcomes. The next
chapter presents the results of these analyses. Applications are made for each of the

previously presented hypotheses.

62

CHAPTER 4
RESULTS

The results of statistical analyses are presented in three sections. First,

preliminary analyses include bivariate tests of relations between measures. Second,

user

 

“no”

 

multivariate analyses for the total sample and subgroups are presented. Finally, a series
of structural equation models tested the hypothesized model of social-emotional

development.

Preliminary Analyses
Univariate Analyses

Before proceeding with other analyses, the distribution and number of missing
data were examined for individual variables. The amount of missing data was
problematic for father assessments and child temperament. A description of missing
data for these variables and a discussion of corrections made in order to approximate the
normal distribution for additional variables follow.

Temperament. Missing data on child temperament assessments by child care
workers were problematic. Numbers of caregiver global and mean temperament
assessments for low-income families and African American families ranged from 43 to
70. In comparison, child care quality was assessed for 82 Afiican American children
(47% of the total African Americans in the study) and 140 children from low-income

families (39% of the total low-income families in the study). Because more than half of

63

the caregiver temperament assessments were not present and the present data were
insufficient to allow for imputation with traditional means, caregiver assessments of
temperament will not be used in further analyses.

Father data. A preliminary assessment of the Love and Relationships
Questionnaire indicated that 1271 mothers had reported their level of conflict with their
husband or partner; 446 fathers (35% of fathers) reported their level of conflict with
their wife or partner. Further investigation indicated that this measure was only
collected at 6 out of 10 possible sites. Of these 6 sites, data were present for 60% of
fathers. Preliminary tests indicated that father data were not missing at random.

Specifically, families with fathers who responded and families with fathers who
did not respond were examined for group differences in mother's marital conflict, child
care quality, income-to-needs ratio, mother—child attachment, total HOME scores, and
maternal belief/psychological measures (locus of control, depression, sense of
competence). T-tests showed no differences on the 3 maternal belief/psychological
measures at one month. However, mothers whose partners did not participate reported
significantly higher marital conflict at one month (p = .002). The HOME scores for
families whose fathers did not participate were significantly lower at 36 months (p<
.001 ). The attachment q-sort scores for children whose fathers did not participate were
significantly lower at 36 months (p = .001). Trends indicated a tendency for lower child
care quality (6-15 months) and lower income (1 month) among those families who did
not have a father participating in the study. In sum, father data was not randomly missing
and thus not representative of the total sample. For this dissertation, only maternal

reports of marital conflict will be used.

Normality of distribution. Extreme values ofien alter the covariance matrix
(Schumaker & Lomax, 1996). Thus, the distribution of the observed variables and tests
for skewness and kurtosis was performed before proceeding with analyses. Income-to-
needs ratios and depression scales were significantly skewed. Trimming the right tail of
both sets of variables for statistical analyses improved the normality of the distributions.
Income was trimmed at 7 times the level of poverty. The CBS-Depression measure has a
clinical cut-off point of 16 with higher scores representing more depression; this measure

was trimmed at 25 to decrease the impact of outliers.

Bivariate Analyses

Odds ratios, correlations, and AN OVAs were used as preliminary tests of the
hypotheses. Bivariate comparative tests also assisted in determining controls used in the
multinomial regression. Racial comparisons in test scores along with validity testing
examined the possibility of culturally biased measures. Income group comparisons were
also performed, with low-income defined as less than 150% of the poverty level (n = 358)
and middle-to-high income defined as over 300% of the poverty level (n = 600).

Before proceeding, the power of the test was examined to minimize Type 11 error
rate (false negative statistical inference). In general, bivariate tests of the total sample
had sufficient power to detect small effect sizes (r = .10 - .30) at a p—value of .01 . For
example, correlations required a sample size of 1,163 to detect these effect sizes at a
power of .80 (Cohen, 1992). Correlations with self-control had a slightly lower range of

participants (1061 - 1093); however, this was still more than sufficient to detect small

65

 

efl‘ects at the .05 significance level. Power is somewhat more limited within analyses of
the African American sample. The sample size required to detect medium effect sizes (r
= .30 - .50) at a significance level of .05 requires 177 participants; data were only
collected for 173 African Americans.

Odds ratios. Odds ratios compared dichotomous mother-child interaction
variables (detachment, negative parenting) by demographic groups. Odds ratios are a
common measure of effect size. For additional information, frequencies are displayed in
Table 4.

The odds of observing negative parenting were 3.16 times greater among low-
income families than among middle-to-high income families (2 = 4.71, n = 888). Low-
income mothers were 3.54 times more likely to be rated as detached than middle-to-high
income mothers (2 = 5.33, n = 888). Afiican Americans were 3.39 times more likely to
be rated as detached than Caucasians (z = 5.15, n = 1137). Afiican Americans were 2.89
times more likely to be rated as displaying negative parenting than Caucasians (z = 4.09,
n = 1137).

Odds ratios of dichotomous study variables with dichotomized attachment
security (secure and insecure) were also performed. Disorganized attachment
relationships were classified as insecure. Caucasians were 1.50 times more likely to be
rated as securely attached than Afiican Americans (2 = 2.21, n = 1061). Children fi'om
middle-to-high income families were 1.57 times more likely to be securely attached than
children fiom low-income families (2 = 3.03, n = 832). It is of interest that observer

ratings of detached and negative parenting were unrelated to attachment security. Tests

66

 

 

of gender differences of dichotomous study variables (detachment, negative parenting,

security) were not significant.

 

 

 

 

Table 4

Frequencies ofdichotomous stud}; variables by demographic gzougs

Variable Low-income Mid/high Income African American; Caucasian
Detached 52 32 31 68
Not observed 253 551 123 915
Negative 46 31 24 59
Not observed 259 552 130 924
Secure 146 355 74 590
Insecure 130 201 63 334

 

Correlations. In addition to correlations for the total sample, separate
correlations were performed by demographic variables, race, economic status, and child
gender. Correlations tested for significant relations among continuous variables in the
study, including perceived control, perceived competence, depression, stimulation,
sensitivity, positive parenting, lack of negativity, marital conflict, child care quality,
attachment security, and self-control. Correlations for the total sample are presented first

in Table 5. Correlations for Afiican Americans and Caucasians are presented in Table 6.

67

Similar tables by economic groups and gender groups are displayed in Table 7 and Table
8, respectively.

Correlations of study variables were consistent with expectations. Medium sized
effects were found in inter-correlations among perceived competence, perceived control,
and less maternal depression (CES-D, l — 6 mo.); the following effect sizes of significant
correlations were small. Perceived control was positively related to maternal sensitivity
and home enrichment. Perceived control and perceived competence were positively
related to children’s attachment security and self-control. Parenting measures, including

stimulation, sensitivity, and subscales of the HOME, were also predictive of children’s

attachment security and self-control in the expected directions.

It" A, '9‘“ ‘ M‘ .
T“ ""‘HD-upp. rum
aw...“

Generally, correlations between parenting and child outcomes were stronger for

.FfivaJ-P— —..—.’ “‘7-
. e...—

Afri an Americans than for Caucasians. The link between negative parenting and child
self-control, however, was significant only among Caucasians. Correlations between
parenting and child outcomes were also generally stronger within the sample of low-
income families than among middle-to-high income families. Correlations of perceived

control and perceived competence with attachment security are reviewed as tests of

hypotheses later in this chapter.

68

 

 

 

2.. v a.

 

 

- .2. .2. .:: .2. .2. .8. .2. .8. .8.- .2: .8. .8. .8. ewes-:02:
- .2. .2: .2. .2. .8. .2. .2. .2.- .2: .2. .2. .2. aoafié... .2

- 8.- .2. .8. .8. .8. .2. 8.- .2: 8.- 8. .8. 28220 .2

- .2.- .2: .8.- .2: .2.- .8. .3. .8: .8.- .8: 66:83.52 .2

- .2. .8. .8. .2. .2.- .2: 8. 8. .8. 8888.28.82 .2

- .8. .8. .2. .2: .2: 8 8. .8. 88:83 228.. .8

- .8. .8. .8.- .8: 8. .2. .8. .5585 .8

- .3. .2: .2: 8. .8. .8. 232.8% .8

- E: .2.- 8. .8. .8. 833% .8

- .8. .8.- .8: .8.- .8. 22.26 .m

- .2..- .8.- .8: .2: 25.26 .2.

- .mm. 3. 3:82.88 eozoaom .m

- .2. 35:8 82.2... .8

- 2:85 ._

:2 2 2 : S 8 w 8 e m e m w 2 83E

 

m 2.3.

69

 

 

22528:. 323 8.8 8:32-28:22 22822.2 .28 8882282228 22288:. 26828 8.28 8:283:80 20.2 282222.80 .882

 

we. V E.

 

- 88. 88. 2 - 222. 2. .82. 82. 82. 2.- .88.- 2. 22. 2. 222-228-8288 .22
.22. - 88. .88.- 82. .88. .88. 2.2. 8. .8: .82: .28. .88. 2. 28252822 .2
.88. .2 2. - 2. 88. 82. .88. 2. 822. 22. .8: 8. 8.- .88. 28 8222228 .2
88.- .8.- 28: - 88. 22.- .2: .88.- 22.- .28. .22.. .2..- .88: 88.- 2822.8 252282 .22
.2 2. .2. .8 2. .2 2.- - .88. .88. 8. 8: 2 .- 22: 8: 2. .2 2. 822228822 .28 28.2 .2
.88. .2. .82. .88.- 2. - .88. .8. 822. 88.- 2.- .2. 8. .8. 82.22828 8.22288 .8
2.2. 2. .88. .2.- .88. .8. - .88. .8. .82: .82: 8. .88. .88. 2222528222282 .8
2. .2. .2. 88.- .88. .2. .8. - .88.. 88.- .82.- 22. .2. .28. 8228222888 .8
8. .88. 8. .2.- .2. .2. .228. .82.. - 2.- 2.- 8. 2. .28. 2282282222228 .8
8.- .22: 88.- .88. .22.- .88: .82.- .88.- .8.- - .22.. .8.- .8: .88.- .8222 2 82-8820 .8
.2.- .2: 8.- .82.. .22.- .2.- .82: .2.- .2.- .88. - .88.- .28.- .82: .222 8-2 82-88-28 .2.
8. .88. .28.- .88.- .88. 8. 88. 22. 28. .228.- .2..- - .2.. 8. 82282222228 228828.288 .8
.8. 88. 22. .88.- 8. 8. 88. .8. 8. .88.- .28..- .22.. - 2.2. 222228 8228288 .8
.2. .22. .88. .2.- .88. .28. .88. .88. .88. .88.- .88.- 88. .2. - 288.22 .2
12.2 2 2 22 2 8 8 8 8 8 2. 8 18 2 828283

 

88.28.2228 28.288289 8828 28.8.2858 288.2 2.. .8 28.223856

8 858.2-

70

 

 

no. v 8... 22222282 2.2622. 822 82:25..— 0882252 .28 222222.28 22283228 338 88 88:22:28.2 2228822 23222626228222. 20.2 2288222280 .882

 

8o _ .
«6.
Na...

.2.:.

2.3.
2.2.

8b.

9%.
2.0—.

.2.—N.

5N:

LN.

.2.—N.

3N.

cc. .gbom .m—
:. u:o§o§< .N~
S.- 23 220 .:
smmr 85:3 .852 .2

8. 822822882. 88 228-2 .8

S.- wfififlfl. 038282 .a

88. 2882528282282 .8
.82.- 82282222288 .8
88. 88228228228 .8
.88.- .8222 82 82-8828 .2.
.88.- .828 8-2 82-8828 .8

Ne. 3:322:28 28023208— .m

- 28.2288 88.28.88 .2

 

 

2 8032.223

 

 

88.28.2258 85822.2 28 362-322.2222 8888 228.822.2288 8&822238289

8. 8528.2.

71

no. v Q. .228w222 2226222 0-228 8022: 282822.28 22283228 o>e828 8222 82828.2 8.2 8282222222280 .882

 

 

- .52. em 2. .2.. eo2. .2. eoN. .2.» 2. coo. 2.2 2.- .n2.- .oo. ooo. 2922-2238223 .m2
2. - oo. sm2.- *oN. eon. acm- sa2. .2. 2.02.- ..2.- swo. *2. 282028282032 .2
2o. 22o. - mot 2.02. 8mm. awn. 828m. 8N2. an 2.- .802 .- mo. oo- 23 2.822220 .2 2
Not 5.. No. - 2&2.- ..oor eon: somr 2.3.- ...nm. .58. .236 .nm.- 82:28 282.332 .3

8.2 2. .82. no. no.- - an2. «nN. 8on- 822. 2.3.- 8.02.- ..o2. 8N2. 802282338220 V228-2 .o
no. co. vo. not goo. - 2.228. .222. no. m2.- 2; 2o. no. 3222229382 32828082 .w
8N2. .822. 2.92. no.- *2. com. - enm. *NN. anmr awn: no. 822. 2208:2202-222m .n
.2. .82. mo. Nor .oo. *oo. «2. - 2.9.. .n2.- soNr No. .2. b2>222mcom .o
no. mo. oo. mot emo. no. .8222. 8mm. - *m2.- 2&2.- vo. no. 85228222822m .m
3. cor 2o. .3. wot oo.- «no: 5. mo. - ion. {emf *nm.- .02: m2 ammo .2.
mo: 2.2.. Sr now. wot no.- no.- 2o.- vor 8mm. - 2.228.- 2.3...- .oE v.2 ammo .m
28o. mo. 2o.- .52.- mo. vo. mo: 3. mo: _..n2.- 8.98.- - 2.3.. 0082208288 2898208082 .N
No. .52. mo. 2&2.- mor No. goo. mo. not 2.2 2.- *mNr enm. - 295:8 28282092082 .2
m2 .2 22 3 o m n o m 2. m N 2 40243.ch

 

amen-IN» 3228.8 25 8.832.292

 

w 828228-2-

72

Analyses of variance. Analyses of variance tested relations between demographic
groups and continuous variables. These tests were used as preliminary contrasts of
processes. Multiple comparison procedures, such as MANOVAs, were not used in these
preliminary analyses. Multiple comparison procedures followed by post-hoe tests delete
participant information that is incomplete. These procedures may result in false
negatives; therefore, some statisticians recommend such methods only for determining
sets of variables, i.e., with multiple regression (Ludbrook, 1998). Although Type I errors
may result fi'om multiple comparisons, my goal is not solely to determine significance but
rather to perform preliminary tests of effects.

Comparisons of study variables between Afiican Americans and Caucasians are
displayed in Table 9. In order to ensure that differences in process were not due to
unequal variances within groups, Levene’s test of equality of variance was included in
this table. Comparisons by economic groups are displayed in Table 10. (See Appendix
B for comparisons by race and income.) Tests of gender differences are included in
Table 11. Parent characteristics, including parental perceived control, perceived
competence, maternal depression, ratings of mother-child interaction, parenting as
measured by subscales of the HOME, and marital quality were tested for significant
differences between groups. Child-related variables, including child care quality, 36-

month attachment, and self-control, were also tested for differences.

73

.28. V Q8:

 

on.2 9.6 So 5828-32280

2 2. own 2.... coo- No.2 2 2.22 3 2 2282-2258 22822.2(
H822822282288

88. SN 8... 228228280

ow. we. 2.... ood- oo. and 2.82 22.82.2252. 22822:.
2282822283

no.8. mod oeo 5828322280

mo. 2o.m .:.. o8; Nmn 8.3 3.2 22822252. 52...:
2.8222 82 Q-mmu

nno n~.o 2vo2 2828322280

no. mn.m .:.. 228 eon 822.2 mn2 502.208... 828222.82
2.8282 0.2 Q-mmo

mn- 2o.v ~32 582882280

2n. v2. 5. 2m. mm. Nod mn2 2282-22528 22822.32.
238282288 282823-2082

we. ond ~32 282832280

.:.... mm.» mo. oo. mm. on...“ ~n2 22822252. 582.52.
22825... 289828-2882

2o.~ nm.m 2.82 52882280

.2.: 8.28 :3 $.82- Nm.2 «2.2 mn2 2232-2252 22828-222.
2088222
.8248- 38138 28 2228-2 828 8 2. 8. 28.88I. >

 

8 822288

74

48. V Q2...

 

 

 

mu. 3. new 523250

9.. 2.. .:.. mod- 3. 5..- a2 2 2232.85 §E<
”20.52.00.223

cm. 2m. 93 52.8280

mm. on. 2:. mod. 2. cm. 2: 592.2223. 5352‘
”2220822852.

«3. Kama moo 5228250

mm. 3. .:.. $6. 3.2V 3.2 mm 2232.232 §E<
”E23: 28 22.6

m». 2 N. N mg 52.0.3280

9: NYMM .22.. no.0 2.2 end mv— anew—oi §E<
H822.2223 282.82

8.2 o2.m 223 5298280

3.... £242 .:.... N 2 .m- am; NW? 9..— gov—05 §E<
53283: .3 V203

an. 20d 03 528030

.:.... owdv I... cod- 2m._ mad m: 2282.85 5054‘
“$2225.23 3222832

2 .2 222 .5 03 2823250

.:.... Nvdm— .2:... 3».— 2- ocN mN.m m: 502.205 §E<
”uaogotcm
ma 0%3 42m 2.29.» a H a 2 23>

8822 um E22923

 

9.282 a 2.3

75

Table 10

Stu_dz variables by economic groyg

 

 

Variable n M SD t-value Sig.

Perceived control:
Low-income 357 3.68 .55 -4.40 *"
Middle income 600 3.83 .47

Perceived competence:
Low-income 357 3.95 .52 -1 .96 .05
Middle income 600 4.02 .53

CES-D 1-6 mo.:
Low-income 357 13.74 7.31 13.60 * "
Middle income 600 7.97 5.70

CBS-D 15 mo.:
Low-income 292 12.20 7.76 10.22 *"
Middle income 573 6.91 5.96

Stimulation:
Low-income 305 2.33 .62 -9.74 ***
Middle income 583 2.77 .64

Sensitivity:
Low-income 305 8.25 1.75 -12.90 ""
Middle income 583 9.80 1.59

Enrichment:
Low-income 288 5.82 1.97 -13.1 1 *"
Middle income 573 7.42 .87

Positive parenting:
Low-income 288 5.10 1.21 -8.61 “W
Middle income 573 5.75 .57

Lack of negativity:
Low-income 288 4.56 1.33 -8.81 "*
Middle income 573 5.32 .91

Marital conflict:
Low-income 288 2.70 1.13 7.71 *"
Middle income 595 2. 13 .75

"* p < .001.

76

Table 10 (cont’d)

 

 

Variable n M SD t-me Sig.
Child care quality:
Low-income 140 1 8.91 5.27 -8.92 "' “ "'
Middle income 399 23.35 4.44
Attachment:
Low-income 277 .22 .20 -6.77 ”"
Middle income 555 .32 .20
Self-control:
Low-income 251 -.25 .71 -6.93 **‘"
Middle income 506 .13 .71
"‘ p < .001.

77

Table 11

 

 

Stu variables b ender

Variable n M SD t—value Sig.

Perceived control:
Male 704 3.79 .49 1 .23 .22
Female 659 3.76 .50

Perceived competence:
Male 704 3.99 .55 -1.00 .32
Female 659 4.02 .52

CBS-D 1-6 mo.:
Male 705 10.02 6.83 .22 .83
Female 658 9.94 6.78

CES-D 15 mo.:
Male 637 8.52 7.10 -.72 .47
Female 604 8.81 7.06

Stimulation:
Male 658 9.15 1.81 -1.25 .21
Female 614 9.28 1.75

Sensitivity:
Male 658 2.55 .65 -2.37 .02
Female 614 2.64 .66

Enrichment:
Male 633 6.87 1.41 -.96 .34
Female 601 6.95 1 .49

Positive parenting:
Male 633 5.47 .98 -2.66 .01
Female 601 5.60 .77

Lack of negativity:
Male 633 4.91 1.17 -4.45 ”“
Female 601 5.19 1.03

"’p < .001.

78

Table 11 (cont’d)

 

 

Variable n M SD t-value Sig.

Marital conflict:
Male 658 2.29 .89 -1.34 .18
Female 613 2.36 .97

Child care quality:
Male 404 22.12 4.98 -.04 .97
Female 369 22.13 4.62

Attachment:
Male 614 .27 .21 -4.50 ”"
Female 583 .32 .20

Self-control:
Male 564 -.06 .79 -2.97 ”*
Female 529 .07 .67

"‘ p < .001.

79

Results indicated that African American mothers experienced more depression
and marital conflict. African American families displayed less sensitive mother-child
interaction and had lower HOME scores. There were no differences between racial
groups on the parental efficacy measures of perceived control and perceived competence.
However, low-income mothers had lower scores on both dimensions of parental efficacy
than middle-to-high income mothers. Low-income mothers were also significantly more
depressed, less sensitive, and stimulating with their children. There were significant
differences in the expected direction between economic groups on all other study
variables (HOME scores, marital conflict, child care quality, 24-month attachment
security, and self-control).

Tests of gender differences indicated that boys received more positive parenting
and negative parenting (i.e., more positive attention and negative attention) than girls.
Mothers behaved more sensitively toward girls than boys during videotaped mother-child
interaction. Girls were more securely attached than boys and exhibited more self-control.

Attachment Security. One-way analyses of variance with quality of attachment
(avoidant, secure, resistant) as the factor were conducted for continuous variables. Only
variables measured in infancy (either concurrent with or previous to strange situation
measurement) were included in these analyses of variance. These include parental
perceived control, perceived competence, global ratings of sensitivity and stimulation,
depression, marital quality, and child care quality. Results are presented in Table 12. T-
tests were completed for attachment as a 2-way factor (security/insecurity). For these
comparisons, disorganized attachment was included in the insecure group. Relations

between study variables and attachment security follow in Table 13.

80

Table 12

Means ofstudy variables by attachment classification with one-way analyses of variance

 

Variable A B C F Contrasts
Income 2.48 3.12 3.30 9.32* A < B, A < C
(195) (817) (175)
Perceived control 3.97 4.01 3.94 .10
(195) (819) (176)
Perceived competence 3.97 4.01 3.95 1.33
(195) (819) (176)
CESD 1-6 mo. 10.34 9.67 10.47 1.53
(195) (819) (176)
Stimulation 2.50 2.62 2.66 3.21 *
(189) (808) (173)
Sensitivity 8.90 9.26 9.50 5.52“ A < B, A < C
(189) (808) (173)
Enrichment 6.65 6.95 7.04 4.13“ A < B, A < C
(193) (813) (174)
Positive parenting 5.34 5.59 5.51 6.58" A < B
(193) (813) (174)
Lack of negativity 4.80 5.11 5.13 6.74" A < B, A < C
(193) (813) (174)
Marital conflict 2.43 2.29 2.30 1.71
(174) (780) (168)
Child care 20.68 22.63 21.90 9.04"' A < B
(130) 4483) (102)

 

Notes: Ns are in parentheses. " p < .05. A = avoidant, B = secure, C = resistant.

81

Table 13

Stm variables by attachment securig

 

 

 

Variable n M SD t-vw Sig

Perceived control:
Secure 742 3.78 .50 -.05 .96
Insecure 448 3.78 .47

Perceived competence:
Secure 742 3.78 .52 1 .76 .08
Insecure 448 3.96 .55

CBS-D 1-6 mo.:
Secure 742 9.63 6.59 -1.73 .08
Insecure 448 10.34 6.97

CES-D 15 mo.:
Secure 738 8.71 7.07 .1 1 .91
Insecure 448 8.66 7.08

Stimulation:
Secure 732 9.30 1.74 1.55 .12
Insecure 438 9.13 1.81

Sensitivity:
Secure 732 2.62 .64 1.21 .23
Insecure 438 2.58 .67

Enrichment:
Secure 736 6.99 1 .39 2.26 .02
Insecure 444 6.79 1.53

Positive parenting:
Secure 736 5.60 .78 3.20 ‘“
Insecure 444 5.42 1.02

Lack of negativity:
Secure 736 5.13 1.04 2.51 .01
Insecure 444 4.95 1.19

m p < .001.

82

Table 13 (cont’d)

 

 

Variable n M SD t—value Sig.

Marital conflict:
Secure 707 2.26 .87 -2.44 .02
Insecure 415 2.40 .99

Childcare:
Secure 437 22.79 4.66 4.48 ***
Insecure 278 21.19 4.70

Attachment:
Secure 71 1 .30 .20 2.13 .03
Insecure 431 .28 .21

Self-control:
Secure 657 .03 .72 1.12 .26
Insecure 384 -.03 .76

m p < .001.

83

Analyses with study variables and attachment by classification and by a
dichotomous test of security/insecurity yielded interesting results. First, one-way
analyses of variance indicated that income, all parenting variables, and child care quality
were related to attachment classification. Insecure-avoidant children came fiom families
with lower income, received less sensitive and positive parenting but more negative
parenting, had less enriched home environments, and received lower quality child care
than securely attached children. Children classified as insecure-resistant received more
sensitive parenting, less negative parenting, and had more enriched home environments
than insecure-avoidant children, but there were no differences in mother-child interaction
or other study variables between insecure-resistant children and secure children.

The second set of analyses (secure/insecure attachment) indicated that home
enrichment, positive parenting, and quality child care were associated with secure
attachment. Negative parenting and marital conflict were associated with insecure
attachment. Children who were securely attached at 15 months had significantly higher
scores on the Attachment Q—Set at 24 months.

Sensitivity. Preliminary tests of mother-child interaction ratings indicated some
interesting differences between demographic groups. To review the statistical
differences, African American mothers were 3 times more likely to be rated as detached
or as displaying negative parenting than Caucasians. Low-income mothers were 3 times
more likely to be classified as detached and 4 times more likely to be rated as displaying
negative parenting than middle-to-high income mothers. Neither rating was significantly
related to a dichotomized measure of attachment security measured at 15 months;

however, both ratings were significantly related to attachment security at 24 months.

84

 

Sensitivity and stimulation were related to attachment classification within the Caucasian
group but not within the African American group. Afiican Americans received
significantly lower ratings than Caucasians on stimulation and on the composite measure
of sensitivity (See Table 9).

The HOME inventory subscales also measured parent-child interaction. It is
possible that the Positive Parenting subscale provides a more effective alternative to the
global sensitivity rating. The two measures have only a small association (r = .21). The
Positive Parenting subscale was significantly associated with the Attachment Q-Set in the
expected direction for both African Americans and Caucasians. The two measures
approach the concept of sensitivity from different perspectives (NICHD, 1997); therefore,
global sensitivity and the Positive Parenting subscale are both used in subsequent
analyses.

Temperament. Maternal assessments of temperament using the Early Infancy
Temperament Questionnaire were initially included in the correlation matrix; however,
these tests indicated that the measures were problematic. Correlations with self-control
ranged from .03 to .06. The correlation between global and mean assessments was -.30
(p < .01). Documentation noted that the Early Infancy Temperament Questionnaire had
not been previously validated (NICHD, 1999a). Although initial plans called for use of
the Infant Temperament Questionnaire, a more established measure, at the 6-month
assessment, maternal assessments of child temperament were discontinued after the 1-
month data collection. The Infant Temperament Questionnaire was not used at the 1-
month timepoint because it was considered to be developmentally inappropriate for 1-

month-old infants. Conceptually, a measure of temperament would have been useful in

85

evaluating these data; however, the existing measure proved to be inadequate and was not

used in this dissertation.

Tests of Hypotheses
Hypotheses are grouped in four major areas. First, tests of hypotheses related to
perceived control are displayed in Table 14. Results are then discussed. Second, tests of
hypotheses related to perceived competence as presented in Table 15 are discussed.
Third, results of the logistic regressions are reviewed (Hypothesis 2 — 3). Fourth, results

of the structural equation models are reviewed.

Table 14

Hypotheses aid tests related to perceiveaL control

 

 

Hyp_othesis Empirical Test

1a: Parents with higher levels of perceived Significant correlations between
control are more likely to exhibit sensitive Parental Locus of Control and global
behavior. ratings of mother-child interaction

1b: Parents with higher levels of perceived Analysis of variance on Parental
control are more likely to have infants rated Locus of Control with attachment
as securely attached. security/insecurity as the factor

1c: The relation between perceived control and Contrast analysis comparing effects
secure attachment will be stronger among between low-income families and
low-income families than middle-to-high middle-to-high income families; 2-
income families. score significance test.

1d: The relation between perceived control and Contrast analysis comparing effects
attachment security will be stronger among between African American and
African American families than Caucasian Caucasian families; z-score
families. significance test.

 

86

Perceived control. For the total sample, maternal sensitivity at six months was
significantly correlated to perceived control at one month (p = .001, n = 1272).

However, this relation bordered on a small effect size (r = .09). Cohen’s (1988)
convention for correlational effect sizes indicates the range for small effect sizes as .10 to
.30. Although it is possible to reject the null hypothesis for Hypothesis 1a as stated, the
practical significance may be overestimated because of the large sample size. Tests using
a large sample size may infer a significant effect that is only trivial. Among African
Americans, maternal perceived control showed a more substantial relation to maternal
sensitivity (r = .19, p = .016, n = 154). The relation between perceived control and
sensitivity was not significant among Caucasians. Among low-income families, maternal
perceived control was also related to sensitivity (r = .17, p = .003, n = 305). However,
no significant association was found between perceived control and sensitivity among
middle-to-high income families.

Hypothesis 1b was not confirmed by significance testing. Mothers with securely
attached children did not have significantly higher perceived control scores than mothers
with insecurely attached children. Insecure attachment was defined as avoidant, resistant,
or disorganized attachment. Although perceived control was not related to strange
situation classification at 15 months, perceived control was related to attachment security
at 24 months as measured by the Attachment Q-Set (r = .11,p < .001, n = 1197).

Hypothesis 1c tests the relation between perceived control and attachment for
African Americans and Caucasians. No significant relation existed when attachment was
measured at 15 months for either group. However, perceived control predicted

attachment security at 24 months among African Americans (r = .26, p = .002, n = 141).

87

Furthermore, there was no significant relation between perceived control and attachment
security among Caucasians at 24 months (r = .06, p = .06, n = 930). To test for
significant differences between the two correlations, the coefficients were transformed
with the Fisher-Z transform (Papoulis, 1990). The difference of the two z-scores was
adjusted for sample size by dividing by the square root of the sum of inverse sample sizes
minus 3 [1/(N1-3) + 1/(N2-3)]. The resulting z-value was significant (p < .02). This test
provided partial support for concluding that the relation between perceived control and
attachment is stronger among African Americans than among Caucasians.

Hypothesis 1d tests the relation between perceived control and attachment by
income groups. No significant relation existed when attachment was measured at 15
months for either group. In low-income families, perceived control tended to predict
attachment security at 24 months (r = .1 1, p < .08, n = 277). In middle-to-high income
families, the correlation effect size was similar (r = . 10, p < .02, n = 555). The difference
in significance level between the two groups was related to the sample size. The Fisher-Z
contrast was not significant (p < .98). Thus, it is not possible to reject the null
hypothesis; findings indicate no real difference in the effect of perceived control on
attachment between economic groups. Tests of perceived competence follow in Table

15.

88

Table 15

Hypotheses and tests 01 aerceived competence

anothesia

 

 

2: Perceived competence and parenting
style will be related to attachment
classification.

2a: Children displaying an insecure-avoidant
attachment are likely to have mothers
with high levels of perceived competence.

2b: Children displaying an insecure-resistant
attachment are likely to have mothers with
low levels of perceived competence.

2c: Children displaying a secure attachment are
likely to have mothers with an average level
of perceived competence, greater than that of
mothers in insecure-resistant dyads and less
than that of mothers in insecure-avoidant dyads.

2d: Children displaying insecure attachment
are likely to have mothers who are less
sensitive.

2e: Relations between parental efficacy,
parenting style, and attachment will vary
between ethnic and income groups.

Empirical Test

Tests 2 (a - f) and Model )(2 indicates
fit of a logistic regression

Significance of contrasts in
analysis of variance

(preliminary test)

Significance of contrasts in
analysis of variance

(prelimimry test)

Significance of contrasts in
analysis of variance

(preliminary test)

Significance for ratings
of parenting behavior, analysis of
variance (preliminary test)

Within-group tests of Hypotheses 2a
through 2d for African Americans,
Caucasians, and economic groups

 

Perceived competence. Preliminary tests of perceived competence, corresponding

to Hypotheses 2a — 2d, are discussed in this section.

Results of the logistic regression of

perceived competence and parenting on attachment security (Hypothesis 2) follow.

Preliminary AN OVAs indicated that perceived competence did not predict

strange situation classification for the total sample. For African Americans, however,

perceived competence was a significant predictor of attachment classification, F (2, 134)

89

= 3.34, p < .04. Greater variance within the African American sample may have
explained these results; however, a test for equality of variance with Caucasians was not
significant, F(2, 1212) = .14,p < .71.

Post-hoe tests indicated that Afi'ican American children with insecure-avoidant
attachments did not have mothers with the highest levels of competence as suggested in
Hypothesis 2a. However, insecure-avoidant children did have mothers with high scores
on this measure (M = 4.03) that compared favorably with scores of mothers with secure
children (M = 4.07). Children with insecure-resistant attachments did have mothers with
the lowest mean scores on this measure (M = 3.74), as suggested in Hypothesis 2b.
Differences on this measure between insecure-resistant and secure groups reached
significance, indicating that children displaying a secure attachment were more likely to
have mothers with a level of perceived competence greater than that of mothers in
insecure-resistant dyads. This provided partial support for Hypothesis 2c.

Within-group tests of Caucasian, low-income, and middle-to—high income
families produced no significant relations between perceived competence and strange
situation classification. Thus, Hypotheses 2(a — c) were supported only among the
African American sample.

The Positive Parenting subscale of the HOME as measured at 15 months was also
used as a measure of parenting style in the following analyses. For the total sample, the
positive parenting scale was a significant predictor of attachment classification, F (2,
1177) = 6.58, p < .001. Post-hoe tests indicated that mothers with insecure-avoidant

children displayed significantly less positive parenting than mothers with securely

90

attached children. In general, preliminary tests indicated that positive parenting was
related to strange situation classification at 15 months (Hypothesis 2d).

Table 16 below presents results for the measures of parenting style, including
global sensitivity and positive parenting. Sensitivity was a significant predictor of
attachment classification for the total sample (n = 1169). Post-hoe tests indicated that
mothers with insecure-avoidant children displayed significantly less sensitivity than
mothers with securely attached children. Means of positive parenting scores and tests of
significance by demographic groups are also displayed in Table 16. Both sensitivity and
positive parenting were associated with attachment security (Hypothesis 2d). There
were, however, differences by demographic groups (Hypothesis 2e). For example,
positive parenting was not a significant predictor of strange situation classification among
low-income families. Sensitivity did not predict strange situation classification among
African American or low-income groups. Furthermore, mothers with insecure-resistant
children were rated as more sensitive than mothers with secure children among middle-

to-high income families.

91

Table 16

Parenting measures and strange sthuation classification with demogaahic gzouas

 

 

 

Meaas
A B C F Contrasts
Sensitivity:
African Americans 8.03 8.22 7.56 1.01
(30) (86) (18)
Caucasians 9.16 9.46 9.83 5.49“ A < C
(138) (634) (138)
Low-income 8.07 8.21 8.11 .79
(60) (155) (35)
Mid/high income 9.41 9.76 10.25 5.73“ A < C, B < C
(59) (358) (84)
Positive parenting:
African Americans 4.58 5.24 4.37 5.64" A < B, B > C
(31) (85) (19)
Caucasians 5.44 5.64 5.67 4.43" A < B, A < C
(140) (638) (138)
Low-income 4.95 5.17 5.18 .79
(65) (169) (38)
Mid/high income 5.62 5.77 5.80 2.401
(74) (389) (91)

 

Note: A = avoidant, B = secure, C = resistant. ns for groups are in parentheses.
"p < .01, ‘p< .05, *p < .10

Logistic regressions. Logistic regressions were used to test hypotheses for

attachment security in infancy (Hypotheses 2 - 3). First, a multinomial regression

including parental efficacy as perceived competence, parenting style, and income as a

control variable tested Hypothesis 2. In multinomial logistic regression, the dependent

92

variable is nominal but not restricted to two categories (Hosmer & Lemeshow, 1989). In
addition to model fit, significance of individual variables tested Hypothesis 2a — 2e.
Second, logistic regressions including child care interactions and control variables
examine Hypothesis 3. These regressions examine differences in the environments of
children with secure and insecure attachments.

Based on preliminary tests, income was included as a control variable. Gender
differences were tested as a control because of previous findings related to attachment
security (NICHD, 2001). However, in these analyses, no effects were found. A measure
of income-to-needs ratio averaging three timepoints (1 month, 6 months, 15 months) was
constructed to exclude inappropriate information (24 months, 36 months) from these
analyses of 15-month attachment security.

The fit of the multinomial logistic regression model with perceived competence,
positive parenting, and income was significant for the full sample, 12 (6) = 28.52, df= 6,
p = .000, n = 1147). However, the effect size index of Nagelkerke, a pseudo r-square
measure, was trivial (.03). In simpler terms, this model explained only 3% of the

variance in attachment quality. Table 17 displays significant betas for this regression.

93

Table 17

Mdtinomial logistic regression model of attachment classification for the total samale

 

Avoidant Resistant
Income -.13 .06
(.17) (.24)
Perceived Competence -. 10 -.27T
(.15) (.53)
Positive Parenting -.21 * -.14
(.08) (.10)
Intercept .46 .12*
(.74) (.79)
n 193 173

 

Notes: N = 1177. Reference category is secure attachment. Standard errors in
parentheses. * p < .05, Tp < .10

A test of this regression among Caucasians was similar to the test of the total
sample, with significant model fit but insubstantial effect sizes. The only significant beta
was that of income. The fit of the same model using data from Afiican Americans was
considerably improved. The likelihood ratio test indicated model fit, 12 (6, N = 135) =
15.94, df= 6, p = .005). The Nagelkerke index reached .14, indicating that the model
explained 14% of the variance in attachment classification. Table 18 below displays

significant betas for this regression.

94

Table 18

Multinomial logi_stic regression model for African American families

 

Avoidant Resistant
Income .04 -.13
(.17) (.24)
Perceived Competence -.06 -1.1 1*
(.42) (.53)
Positive Parenting -.42* -.46*
(.17) (.20)
Intercept 1 .27 5 .24"
(1.86) (2.19)
n 31 19

 

Notes: n = 135. Reference category is secure attachment. Standard errors in parentheses.
*p < .05, Tp < .10

Within the African American sample, income was not a predictor of attachment
classification. Mothers with lower perceived competence were more likely to have
children with insecure-resistant classifications but not insecure-avoidant classification.
Mothers whose children had insecure-avoidant classification did not have lower scores
than mothers whose children were rated as secure. Positive parenting predicted secure
classification.

Models divided by income groups did not reach a significant fit. In sum, findings
provide partial support for Hypothesis 2, which states, in simple terms, that mothers who
are insensitive and have children with insecure-avoidant classification (Type A) may

overestimate their competence as parents. This effect is most noticeable among African

95

Americans, who had greater variance in parenting style than Caucasians (Levine’s test for
equality of variance, F = 48.89, p < .001).

The following logistic regressions, testing hypotheses related to child care quality,
( Table 19) examines only the total sample because of the reduced number of participants
related to child care participation. Specific interactions that are mentioned in Hypothesis

3 include depression and parenting style.

Table 19

Hmtheses an_d tests of moderators of child care quality

 

 

 

Hypothgsi_s Empirical Test

3: For children in child care, child care quality With interaction terms, Model X
may moderate the effect of other variables indicates fit of the logistic
on attachment. regression

3a: The quality of child care may buffer the Wald 12 statistic for interaction term
likelihood of insecure attachment for of child care quality (APECP or
depressed mothers. APFDC) x depression (CBS-D)

3b: The quality of child care may buffer the Wald 1.2 statistic for interaction term
likelihood of insecure attachment for mothers of child care quality (APECP or
with insensitive parenting. APF DC) x average global ratings

 

Two logistic regressions with a dichotomous insecure/secure dependent variable
were performed on the total sample. The first regression tested child care as a moderator
of the relation between depression and attachment security. In this equation, only child
care quality was a significant predictor (.12 = 19.02, p < .001). Thus, Hypothesis 3a is not
confirmed. The second regression tested the interaction of positive parenting x child care

as a moderator of the relation between child attachment security. In this model, the

96

interaction term was significant (.12 = 17.62, p < .001). The model fit was also adequate
(f = 18.32, p < .001, n = 705). Thus, children who experience positive home
environments and quality child care are more likely to be securely attached than those
who rate highly on only quality child care or the positive home environment measure.
These results provide support for Hypothesis 3b. The following section describes

. structural equation models testing relations with attachment security at 24 months and
child self-control at 36 months.

Structural equation modeling. Structural equation models specify a causal order
among a set of variables (Klein, 1998) and then assess the extent to which the predicted
interrelationships differ from the interrelationships actually observed in the data.
However, structural equation modeling does not imply causality. Causality is more
probable if certain conditions are met, such as temporal ordering of path variables and the
elimination of other possible causes through experimental controls.

If a model adequately describes actual relations between variables, then the model
is said to fit the data, as in Hypothesis 4, “The model describing the development of
mother-child attachment and child self-control as displayed in Figure 2 will fit the data.”
A chi-square statistic is used to test whether the differences between the obtained and the
predicted data are greater than zero. A significant chi-square indicates an implausible
model because the differences are too large. To reduce the tendency of chi-square to be
inflated by sample size, the chi-square is divided by the degrees of freedom. If the
quotient meets the commonly-accepted criteria of being less than 3.0, the data will be

said to fit the model (Kline, 1998).

97

A number of other fit indices are available to evaluate the degree of congruence
between a model and data. The goodness-of-fit index (GFI) considers the proportion of
variance in the data accounted for by a given model (Ullman, 1996). It can range from
0.0 (no fit at all) to 1.0 (a perfect fit). A GFI of 0.9 is generally accepted as indicating a
good fit (Bentler & Bonett, 1980). The Root Mean Square Error of Approximation
(RMSEA) reflects the differences between the observed and the model-implied
covariance. RMSEA of 0.0 would indicate a perfect fit of the model to the data. In
practice, 0.05 or less is considered to indicate a close fit (Browne & Cudeck, 1993).

Structural equation modeling requires a set of complete data. Missing data are
likely to result in less accurate computations than when data are replaced by estimation of
maximum likelihood (Little & Rubin, 1989). Missing data also make inferences to the
general population less meaningful. Thus, we created substitutions for missing values
with a maximum likelihood method based on the Estimation Maximization (EM)
algorithm (Dempster, Laird, & Rubin, 1977). Maximum likelihood methods estimate the
means and covariance matrix of study variables, rather than simply estimating single
missing values (Little & Rubin, 1987).

Only cases that had a child outcome variable, either self-control or attachment
security, were included in the missing data analyses. Although the total sample size is
more than adequate, division of the sample into groups makes certain analyses difficult.
The number of African American families (n = 176) that had one or more child variables
(n = 144) is too small for an accurate analysis using structural equation modeling. Thus,

this method was not used for comparisons of processes across racial groups.

98

Structural equation models tested a modified model similar to Figure 2.
Preliminary tests eliminated temperament as a variable; thus temperament was not
included in the final model. Marital quality was included in separate structural equation
models including only participants who had data on the marital conflict measure. Paths

linking variables were used as tests for Hypotheses 4a — 4e, displayed in Table 20.

99

Table 20

Hmtheses and corresaandingsignificance tests of variable-level path;

 

Hypothesis Test of path significance
4a-1: Mothers who have higher perceived Parental Locus of Conn-019
control are more likely to display positive Attachment Q-Set

parenting practices.

4a-2: Mothers who have higher perceived
control are less likely to be depressed.

4b: Mothers who are less depressed are
more likely to display positive parenting
practices.

40: Families who experience less marital
conflict are more likely to have positive
parenting practices.

4d-1: Parents who display positive
parenting practices (e. g., sensitivity) are
more likely to have children with
secure attachments.

4d-2: Parents who display positive
parenting practices are likely to
have children with greater self-control.

4e: Relations between parenting practices,
attachment, and self-control may vary
between ethnic and economic groups.

Parental Locus of Control9 CBS-D

CES-D 9 HOME

Love and Relationships 9 HOME

HOME 9 Attachment Q-Set

HOME 9 Self Control Procedure

Comparisons of structural equation
models

 

100

Eight models describing the linkages among characteristics of maternal perceived
control and maternal depression, children’s home environment, and child outcomes are
presented. The first model used data from the total sample; the second model, which
included a variable of marital conflict, excluded single parent families. The next two
models compared differences in process by child gender. The last four models described
low-income and middle-to-high income samples, with and without the marital conflict
variable.

The first model tested the predictors of child attachment and self-control with
marital conflict omitted. Maternal perceived control and maternal depression were co-
varying exogenous variables. Positive mother-child interaction, negative mother-child
interaction, enrichment, and attachment were the four endogenous variables. Paths
between positive and negative mother-child interaction indicated that these variables had
a bi-directional effect, resulting in a nonrecursive model. This nonrecursive subset had a
stability index of .36. Indices below 1 are considered to be stable. Figure 4 displays the
model with path coefficients. All paths were significant excepting the path between
perceived control and positive mother-child interaction. The path between positive
mother-child interaction and self-control indicated a trend only (p < .10).

Goodness-of-fit measures indicated that the data fit the model, confirming
Hypothesis 4. The [I test was not significant, indicating a good fit, 12 (3) = 3.03, p = .39.
Two additional indices also indicated that the interpretation of the estimated paths is

acceptable (RMSEA = .003, GFI = .999).

101

 

28803228 223 «5822083 .20 $290an .23 20258 cows—225 gogm .2» 953k

 

 

 

 

 

 

;/.,

 

 

 

 

 

\\ ./

 

 

 

 

 

.228 .
828.8 l

 

 

 

 

 

 

C..-

102

Two additional models examined gender differences. The structure of these
models was identical to the first model. Goodness of fit was acceptable for boys, [2 (3) =
5.55, p = .14, RMSEA = .037, GFI = .998, n = 631; similar fit was achieved for girls, )8
(3) = 5.66, p = .13, RMSEA = .039, GFI = .997, n = 598. Path coefficients by gender
are presented in Table 21. An enriched home environment and attachment security were
related to self-control for boys only. Negative mother-child interactions were related to

self-control for girls only. Perceived control predicted home enrichment only for girls.

103

Table 21

Path coeficients and estimates by child gender

 

 

 

 

m Umdized
Control 9 Positive

Male -.16

Female -.06
Depression 9 Negative

Male -.05

Female -.04
Control 9 Enrichment

Male .05

Female .25
Positive 9 Enrichment

Male .53

Female .56
Lack of Negativity 9 Enrichment

Male .23

Female .24
Depression 9Enrichment

Male -.04

Female -.02
Control 9 Attachment

Male .03

Female .03
Positive 9 Attachment

Male .02

Female .03
Enrichment 9 Attachment

Male .02

Female .01

Standardized p
-.08 .22
-.04 .36
-.31 .02
a26 ***

.02 .64
.08 .03
36 ##ll
.29 *$*
.19 ***
.16 ttt
-.19 m
-.09 .02
.07 .06
.09 .03
.10 .02
.1 1 .01
.16 ***
.11 .01

 

Notes: For boys, n = 631. For girls, n = 598. "*p < .001.

104

Table 21 (cont’d)

 

 

 

 

Paths Unstandardized Standardized D
Lack of Negativity 9 Attachment
Male .02 .12 ***
Female .01 .06 .12
Depression 9 Attachment
Male .00 -.07 .09
Female .00 -.08 .04
Lack of Negativity 9 Self-control
Male .02 .03 .45
Female .07 .11 .01
Positive 9 Self-control
Male .04 .04 .29
Female .04 .05 .28
Enrichment 9 Self-control
Male .08 .15 ***
Female .08 .19 ***
Attachment 9 Self-control
Male .52 .14 ***
Female .21 .06 .12
Lack of Negativity 9 Positive
Male 1.03 1.26 .01
Female .29 .52 .01
Positive 9 Lack of Negativity
Male -1.68 -1.38 .08
Female -.51 -.38 .09
m p < .001.

105

A third structural equation tested the model with marital conflict included. For
this model test, I created a separate data set limited to children whose mothers were in a
married or partnered relationship when the child was 1 month old (n = 1159). The
structure of this model was similar to the first model; only marital conflict was added as
an endogenous variable. This model (Figure 5) showed additional effects of marital
conflict. All path coefficients linked to marital conflict were significant. The direct path
between perceived control and home enrichment was no longer significant, but perceived
control was indirectly related to home enrichment through marital conflict. All indices
indicated a good fit, )8 (5) = 4.62, p = .47, RMSEA = .000, GFI = .999. Maternal
depression was related to marital conflict, which in turn, was related to a lack of
negativity and home enrichment. Marital conflict had a direct effect on child self-control
and indirect effects mediated by parenting variables. Marital conflict also indirectly
affected attachment security through its relation to lack of negative parenting.

A similar model for boys had an improved fit, f(5) = 3.64, p = .60, n = 595,
RMSEA = .000, GFI -—- .998 ; however, the model did not fit for girls, i (5) = 15.84, p =
.01, n = 564. Comparative processes by gender are not examined for marital conflict;
however, it is interesting that marital conflict appears to have more effect on the

development of attachment and self-control among boys.

106

32233.2 .:qubiEm 9.225286 2888228 98 Eon—F233.“ .20 $582253 22223 20258 823525 28322.5 .m. Eami

 

 

25562.5 till

nae .- 19
a... 2

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

107

The model was also tested in both economic groups. These models were not
tested separately by gender due to the sample size. First, low-income families were
included in a model with two-parent and single-parent families (n = 292). Although the
model exhibited goodness-of-fit, 12 (3) = 1.25, p = .74, fewer paths were significant than
those found in the first model. Figure 6 displays significant paths between variables for
the model of the low-income group. Paths between perceived control and study variables
were no longer significant. Less depression predicted a lack of negative parenting, which
in turn predicted more secure attachments. Positive mother-child interaction was related
to attachment both directly and indirectly through home enrichment. Thus, positive
parenting predicted home enrichment; both home environment measures were positively
related to attachment security. Other paths linking endogenous variables were not
significant. The model with marital conflict was also tested (n = 242). The model
exhibited goodness-of-fit, f (5) = 1.67, p = .89. However, among low-income families,
there were no significant paths between marital conflict and any study variables. This

model is displayed in Figure 7.

108

32288.2 6885.302 523 26258 52.5226 gogm 6 Eami

 

 

 

 

 

 

 

 

 

 

 

 

KMWV
\\ 22.08.202.25 l//
\\\ ///,.
\100) \\ . . ,/-/
A v \\\\ 8 8 2 - m - 0°.I /////
2 - . x,
no *5 o / / /Mr
N 21
22:28.28 . 32%
8 K8 / 2.- 8 cc. 2. T x
2 95:22. W/ \\

 

 

 

 

 

 

 

109

2.6222225 883-09% 92225283 .82228am 0885.32 523 26238 concave 28325m

 

 

 

 

 

 

 

 

 

 

 

 

N 2:5

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

110

In the final two models, only families whose income-to-needs ratios equaled or
exceeded 3.0 were included (n = 567). The model exhibited goodness-of-fit, f (3) =
2. 12, p = .55, RMSEA = .000, GFI = .999. Perceived control was related to home
enrichment, which was related to child self-control. Perceived control was also directly
related to attachment security. Negative mother-child interactions predicted less home
enrichment and less secure attachments, which in turn predicted less child self-control.
Path coefficients are displayed in Figure 8.

In the model of middle-to-high income families with single parents excluded (n =
563), the f test confirmed the null hypothesis, indicating no difference between the
predicted and observed data, {(5) = 4.27, p = .51. Two additional indices confirmed the
goodness of fit (RMSEA = .000, GFI = .998). Figure 9 displays the standardized
regression weights for tested paths. Mothers who were depressed were more likely to
report marital conflict; mothers with more perceived control were less likely to report
marital conflict. Within this group, however, marital conflict was not significantly

related to other endogenous variables.

111

8:23. cases 52.282828 523 secs 82.828 2832.5 .m seat

 

 

 

 

 

 

 

 

 

 

112

 

 

 

 

 

 

 

 

«Gaga 32,2098

 

 

 

 

 

322:5 528-0236 92225288 .8222Eé 088:2 $363628 523 20258 cone—226 gogw d 959%

 

 

MU

232.22.:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

02v

 

 

 

 

 

 

 

113

In sum, the structural equation model in Figure 4 supported Hypotheses 4 (a — d).
Perceived control was directly related to attachment (Hypothesis 4a) and associated with
maternal depression (Hypothesis 4b). In Figure 5, marital conflict was related to negative
parenting, and families experiencing marital conflict were less likely to have an enriched
home environment (Hypothesis 4c). Figure 4 displayed significant relations between
aspects of the home environment and attachment (Hypothesis 4d-1) and between the
home environment and child self-control (Hypothesis 4d-2). The path between
attachment at 24 months and self-control at 36 months indicated a small effect, even
when controlling for other study variables (Hypothesis 4e-1). There were some
differences between demographic groups, e. g., gender, economic groups (Hypothesis 41).
However, limited variance within economic groups may have made it difficult to detect

certain effects.

Conclusion
The results of data analyses indicated partial support for the hypotheses. Some
findings related to global sensitivity were somewhat unexpected; also findings related to
gender differences were unanticipated. These findings will be discussed in the final

chapter. In addition, limitations and future directions for research will be included.

114

CHAPTER 5

DISCUSSION

The previous chapter included results of analyses, tests of hypotheses, and
findings related to parental efficacy, home environments, children’s attachment
relationships, and child self-control at 3 years of age. This chapter will discuss these
findings. The first section of this chapter presents findings in relation to the research
questions from Chapter 1. Additional findings follow, including gender differences in
developmental processes. A section addressing limitations of the study and suggestions

for future research conclude the chapter.

Research Questions

The NICHD Study of Early Child Care investigated the relation between infant
child care and the quality of attachment (NICHD, 1997). One of the goals of this
dissertation was to extend their study by including dimensions of parental efficacy and
later child outcomes. In addition, this dissertation tested the universal application of
hypotheses across demographic groups. The first set of research questions focused on
parental self-efficacy and its relation to attachment. Second, an investigation of child
care quality partially replicated findings from the NICHD (1997) study. Third, predictors

of attachment and self—control in early childhood are discussed.

115

Dimensions of Parental Efi‘r‘cacy

1. Can counterintuitive findings from earlier studies of the relation between parental
efficacy and attachment security be explained by reexamining dimensions of parental
efficacy, i.e., competence and control?

2. Does parental efficacy have different effects on parenting and attachment among
different demographic groups, specifically, low-income and middle-to-high income
families and African American and Caucasian families?

A differentiation between the dimensions of parental eflicacy, perceived control
and perceived competence, yielded interesting results. Although past research has
referred to both dimensions as representative of parental efficacy, the results of this
research indicate that they are two distinct parental beliefs. Perceived competence was
associated with attachment classification as measured by the Strange Situation Procedure
at 15 months, particularly among Afiican Americans; however, perceived control was not
related to attachment classification. With the total sample, the correlations of perceived
control and perceived competence with attachment security (Attachment Q—Set) at 24
months were equivalent, representing a small effect size. Yet, when the sample was
broken into separate demographic groups, different associations emerged for perceived
control and perceived competence as discussed below.

Perceived competence. A review of the literature formd that past studies have
produced counterintuitive findings related to perceived competence and attachment (e.g.,
Spieker & Booth, 1988). In these studies, mothers of children with insecure-avoidant

attachment (Type A) felt that they were competent mothers. However, insecure-avoidant

116

attachment has been associated with harsh discipline and intrusive parenting. Children
with insecure-avoidant attachments were more likely to have behavior problems in school
than children with secure or insecure—resistant attachment classifications (Renken,
Egeland, Marvinney, Mangelsdorf, & Sroufe, 1989; Waters, etal., 1979). Thus, the high,
and in some studies, higher levels of perceived competence among mothers with children
classified as insecure-avoidant were not explained by the theory of self-efficacy.

One possible explanation of inconsistent effects is the way that parental efficacy
is measured. Bandura believed that perceived competence should be measured by
specific behaviors, e.g., I am competent at soothing my baby when he/she is upset. In
contrast, the Sense of Competence scale used in this dissertation measures global feelings
about parental competence, e.g., I feel capable and on top of things when I am caring for
my baby. Yet, some researchers have found associations with global scales of parental
competence (e.g., del Carmen, Pedersen, Huffman, & Bryan, 1993); others have found no
effects with measures of specific behaviors (e.g., Corapci & Wachs, 2002). A recent
study comparing the Sense of Competence scale with the Maternal Self-Efficacy
Questionnaire, a domain-specific measure (Teti & Gelfand, 1991), found that only the
domain-specific measure was a predictor of child outcome among White middle-class
mothers.

Although the literature suggests that domain-specific measures are preferable
measures of competence, the purpose of this dissertation is not to compare domain-
specific and global measures, but to examine why global measures of perceived
competence are sometimes ineffective. In other studies, perceived competence has been

associated with positive parenting and secure attachment. Findings in the current

117

dissertation suggest that a global measure of perceived competence may not be related to
positive parenting and attachment security within all groups and cultures.

In this dissertation, the strength of associations between global perceived
competence and attachment differed between demographic groups. Researchers have
speculated that the stronger associations with parental efficacy found among diverse
groups may be related to greater variance in parental efficacy scores within
disadvantaged groups (Coleman & Karraker, 2003). Yet, a test comparing variances
between groups indicated that there was no significant difference in perceived
competence scores between Afiican Americans and Caucasians or between low- and
middle-to-high income families.

Among low-income families, mothers who perceived themselves as competent
when their children were one month old were more likely to have children who were
securely attached at 24 months. However, this effect was not significant among higher
income families. It is possible that risks associated with economic disadvantage make
assets, such as a mother who has positive feelings about her parenting abilities, more
essential to the child’s development of attachment security. In higher income families,
other resources, such as quality child care, a supportive community, or a supportive
father, may compensate for a mother who feels insecure in her parental role.

Similarly, differential risks among African American and Caucasian families
affect the hypothetical process of parental efficacy influencing parenting, which in turn
influences attachment development. The bivariate correlation between perceived
competence at one month and attachment security at 24 months was 0.24 among African

Americans; the same correlation was .08 among Caucasians. According to

118

Bronfenbrenner’s (2000) theory, Afiican American or low-income families should have
stronger links between perceived competence and attachment security than Caucasian or
middle-class families.

Parenting and other processes that directly affect the child may be more important
to children who experience more risk. For example, families from middle-class homes
are able to draw on other resources when a mother is depressed or neglectful of her
parenting role. Neighborhood playgroups, a supportive school system, and medical
insurance covering the cost of treatment are likely to be available to assist the family
through difficult times. Thus, access to other environments that promote successful child
outcomes may compensate somewhat for deficits in the home environment, and links
between the home environment and child outcomes are theorized to be weaker among
more advantaged groups.

There were also differential associations of perceived competence and attachment
classification at 15 months. In a regression of income, perceived competence, and
parenting on attachment classification, only income was a significant predictor of
attachment classification among Caucasians. Within the African American sample,
mothers who were less sensitive and had children with insecure-avoidant classification
appeared to think of themselves as competent parents. Although they were less
responsive and sensitive to their children than mothers with secmely attached children,
they perceived themselves to be equally competent. Mothers who had children with
insecure-resistant classification did not report feeling competent as parents.

This finding duplicates those of a study of attachment among at-risk families,

with mothers who had insecure-avoidant children scoring as high on measures of

119

perceived competence as mothers of securely attached children, and mothers with
insecure-resistant children being significantly lower on perceived competence (Spieker &
Booth, 1988). Although their study did not directly explain the unexpected findings
related to perceived competence, they did point out that mothers of avoidant children in
at—risk families tend to idealize their relationships. Perhaps mothers of avoidant children
in families disadvantaged by the social and economic structure of our society (i.e.,
racism) would also idealize their parenting abilities. Furthermore, the racial composition
of their sample is not cited; thus, it is not possible to draw conclusions about effects of
racism within their study.

Analyses of the NICHD data also indicated that Afiican American families were
rated as less sensitive or positive parents. Links between sensitivity and attachment
classification at 15 months were only present among the Caucasian group. It should be
noted that global measures of sensitivity were assessed by mostly White middle-class
graduate students who were, in general, untrained to understand Afiican American
parenting styles. Global sensitivity measures have been used with success for decades
beginning with Ainsworth et al.’s (1978) classic attachment study, but it is only in recent
years that issues related to observer bias have been raised.

Although measurement methods may have been culturally inappropriate, it is also
likely that differences in the general style of mother-child interaction existed between
African Americans and Caucasians. It may be that African Americans place less of an
emphasis on positive parenting and responsiveness, resulting in different links between
perceived competence, parenting, and attachment. The literature suggests that Afiican

Americans value parents who teach their children to survive in a challenging environment

120

(Murry, 2000). Afiican American parents who teach their children to be obedient and
hard workers may be considered competent parents. Cultural differences contributing to
variations in parenting style and motivations may explain differences in the relations
among attachment and perceived competence among Afiican American and Caucasian
parents. While positive attitudes and emotional security are certainly valued, Afiican
American parents have been shown to approach discipline in a different way than
Caucasian families (Murry, 2000; McLoyd, Cauce, Takeuchi, & Wilson, 2000). Thus, it
may be considered preferable to parent in accordance with cultural imperatives,
socializing Afiican American children to succeed within their environment.

Perceived control. In a structural equation model of the total sample, perceived
control was a significant predictor of attachment security as measured by the Attachment
Q-Set, even when controlling for maternal depression and parenting practices. This
indicated the importance of parents’ beliefs about their effects on child outcomes.

Although processes within African American and Caucasian families were not
modeled because of the higher attrition among African Americans, other analyses
indicate some differences in process. In a bivariate correlation, perceived control was
associated with attachment security among African American families (r = .26). This is
particularly notable considering that perceived control was measured when the child was
1 month old; attachment security was measured when the child was 24 months. Among
Caucasian families, perceived control was not associated with attachment security at 24
months (r = .06). This may be due to a differential variance in perceived control scores.
African American mothers had a significantly higher variance in perceived control scores

than Caucasians.

121

Another possible explanation is that a sense of control would be a stronger asset
in a less predictable environment. Many African Americans who experience unfair
treatment may not view their environment as something that they can control. One study
found that Afiican Americans who perceive that they are unfairly treated due to racism
passed teachings and messages about mistrust on to their children through racial
socialization (Hughes & Johnson, 2001). The literature suggests that these teachings
occur at a later developmental stage; however, it is possible that subtle non-verbal
messages may occur in early childhood. Furthermore, among African Americans who
reported higher levels of discrimination, linkages between stress and psychological
problems and between psychological problems and mother-child interaction were
stronger (Murry, Brown, Brody, Cutrona, & Simons, 2001). Thus, negative aspects of
children’s environments had a stronger influence when mothers perceived their daily
experiences as being affected by capricious discriminatory acts. Parents who perceive
that they are in control of their firture and their children’s firture may use this perception
as a positive coping mechanism.

It is of interest that there were no significant group differences in the association
of perceived control and attachment security at 24 months between low-income families
and middle-to-high income families in the structural equation model or within bivariate
correlations. However, there is a significant amount of difference in variance between
scores of the two groups, with low-income families having more variance in levels of
perceived control. In past research, effects of Afiican American status have been
confounded with income effects (Murry, 2000). By viewing economic groups and racial

groups separately, it is possible to make distinct inferences about Afiican American

122

families apart from their economic status. Although it was not possible to include
comparisons of processes between racial groups using structural equation models due to
the limited sample size, ad-hoc analyses (see Appendix B) assisted in disentangling
factors of class and race. Even though the sample size was small for middle-to-high
income African Americans (n = 16), t-tests indicated group differences when contrasted
with middle-to-high income Caucasians. This robust finding suggests strong cultural
effects that supercede class differences, and further research is certainly indicated.

In sum, three factors may result in counter-intuitive effects related to parental
efficacy and attachment. First, perceived control produced different effects than
perceived competence. Second, in families with more risk factors, efficacy and parenting
styles appear to be more strongly related to children’s outcome. For instance, a lack of
perceived control among Afiican Americans may be more detrimental to their children’s
environments than among groups who do not experience discrimination. Third, Afiican
American mothers with insecure-avoidant children were likely to overestimate their
parental competence in comparison with their parenting skills, as defined by the measure
of positive parenting. A trend toward this effect was noted among the total sample, but

not among Caucasian families.

Child Care Quality
3. Does child care quality moderate the effects of maternal depression and parenting
practices on mother-child attachment?
An earlier study with these data found no main effects of child care quality or the

amount of child care on attachment classification at 15 months (NICHD, 1997). Because

123

of these earlier results, only interactions were tested in this study. Tests indicated that
child care quality was not a significant moderator of maternal depression and attachment.
However, child care quality did moderate the relation between positive parenting and
attachment. Although this dissertation used different measures of child care quality and
mother-child interaction, results from the earlier NICHD (1997) study were replicated.
Child care quality did not directly predict attachment security as measured by the Strange
Situation Procedure at 15 months; however, mothers who chose good quality child care
and behaved sensitively toward their children were more likely to have securely attached
children.

Bivariate correlations with child care quality in infancy suggest that a different
approach to modeling may yield other results. Higher child care quality was significantly
related to income, less maternal depression, sensitivity, positive/negative parenting, an
enriched home environment, attachment security at 24 months as measured by the
Attachment Q-Set, and children’s self control at 36 months (See Table 5). Other methods
of modeling these data will be discussed in the section describing future research

directions.

Social-emotional Development in Early Chiltflrood
4. Are the data from the National Institute of Child Health and Development Study of
Early Child Care consistent with the model linking parental efficacy, parenting

practice, and children’s social-emotional outcomes as shown in Figure 2?

124

5. Does the model fit vary as a function of demographic status? For example, do the
relations among the variables differ for low-income and middle-to-high income
families?

In general, the model shown in Figure 2 fit the data, suggesting that maternal
perceived control and mater-ml depression predict parenting practices and attachment,
which in turn predict children’s self-control. Models including marital conflict also
showed goodness of fit, excepting for the model with girls only. Gender differences are
discussed later in the chapter.

Maternal characteristics. In accordance with the literature, maternal depression
predicted more negative mother-child interaction. Mothers who were more depressed
were less likely to provide an enriched home environment. Through parenting and the
home environment, depression affected children’s later attachment security and self-
control. There was also a small but significant direct effect between depression and
children’s attachment security. Thus, depressed mothers were more likely to have
children with insecure attachment, even when positive/negative mother-child interaction
was controlled.

Additional effects were found between perceived control, a dimension of parental
efficacy, and study variables. As expected, perceived control was associated with
depression. This association was substantially larger among low-income families. There
was a small but significant direct effect of perceived control on attachment. Effects of
perceived control on the home environment were generally not significant, except that
middle-to-high income mothers with more perceived control were likely to provide a

more enriched home environment. The effect size for the path between perceived control

125

and enrichment among middle-to-high income mothers was actually equivalent to the
effect size for the same path among low-income mothers. Significance testing varied
only due to sample size.

Parenting and child characteristics. In turn, an enriched home environment was
related to attachment. Mother-child interactions of positive and negative parenting also
predicted attachment security. Many of these associations were reduced when the sample
was divided into low-income and income families. A possible cause of reduced effects
may be the lower variance within economic groups.

An enriched home environment at 15 months was related to child self-control
measured 1‘/2 years later at 36 months. The measure of enriched home environments
included regular experiences, both formal and informal, that teach young children as well
as the provision of toys and other materials for teaching children. It is possible that
mothers who spend time teaching and playing with their children in developmentally
appropriate ways also help them to develop self-control. Mothers who spend time
teaching their children may enhance their social behaviors as well as assisting their
cognitive development. As expected, attachment also predicted children’s self-control.
A positive attachment relationship assists children in their behavioral regulation (Davies,

Harold, Goeke-Murray, & Cummings, 2002).

Marital Conflict
The model including marital conflict indicated some minimal effects. Among
middle-to-high income families but not low-income families, marital conflict was a

significant addition to the model. The variance in marital conflict was, in fact,

126

significantly higher within the low-income group than among the middle-to-high income
group. Thus, differential effects are not explained by a restricted range.

There are other possible explanations for the limited effects of marital conflict
within low-income families, besides differential variance. It may be that low-income
mothers depend more on other family members and friends for support in raising their
children than on their spouse. Men in low-income families may be less egalitarian in
gender roles, if only because of the additional pressures to succeed in the role of
breadwinner. Thus, marital conflict may have less of an effect on parenting and child
outcome if the father is not actively engaged in childrearing. Although this study did not
examine the effects of marital conflict within ethnic groups, ethnicity may be an
important factor. Both gender roles and the effects of marital conflict differ within
cultural context (McLoyd, Cauce, Takeuchi, & Wilson, 2000). For example, Afiican
American men are more egalitarian in gender roles but not in attitudes towards women.
It is also important to note that a higher proportion of African Americans are single in
comparison with Caucasians, particularly among low-income families. In this study,
56% of the African American mothers were no longer living with the father of the child
at 15 months old in comparison to only 9% of the Caucasians.

Marital conflict was measured when the child was 1 month old; other measures
were collected from 15 to 36 months later. Both cohabiting and married partners
completed these measures. In high-income households, 2% of the fathers were no longer
living at home at 15 months, but 38% of the fathers were no longer present at 15 months
among low-income families. Hence, it may be that differences in low-income and high-

income families related to the importance of partner conflict are a reflection of unstable

127

family structure, which is generally more prevalent in low-income households. Marital
conflict that existed with the father of the child may not affect the home environment as
much if the father is no longer present.

Among middle-to-high income families, mothers with more perceived control
were less likely to experience marital conflict. Mothers reporting depression were likely
to experience more marital conflict. Marital conflict was related to negative parenting;
marital conflict indirectly affects attachment through this path. Thus, expected
associations existed between marital conflict and study variables, but only for higher-
income families.

In sum, differences in process between low-income families and middle-to-high
income families existed in this sample. Although effects of perceived control on
attachment did not differ between groups as expected, effects of maternal depression on
negative parenting were stronger among low-income families. In turn, effects of
parenting, particularly positive parenting and enrichment, on attachment and self-control
were stronger among low-income families. Overall, most proximal processes had a
stronger effect on child outcome within low-income families.

¥The importance of parenting practices to children living in low-income families
suggests that parent education programs would be particularly beneficial for this group.
Presently, family programs, such as home visiting forlowlincome families, arenet well-
supported at a federal level in the United States. Government support for experimental
demonstration programs has improved in the past decade, particularly with positive
results coming fi'om certain programs (Olds & Korfrnacher, 1997). However, few parent

education programs are available nationwide for families with children in the preschool

128

K2,.

4‘
f t ~

years. Previous research indicates that early childhood is the stage when family resources
are most influential (Duncan & Brooks-Gum, 2000). Thus, parent education programs
and programs that provide additional resoru'ces, such as healthcare, nutrition, and learning

activities, need to find more avenues to reach young children before they enter school.

Gender Differences

A division of the sample by gender suggested different processes for boys and
girls. The literature suggests that gender differences in behavior emerge during later
years, when children are 4 or 5 years of age (Achenbach, Edelbrock, & Howell, 1987;
Zahn-Waxler, Ianotti, Cummings, & Denham, 1990). Boys tend to have more
externalizing behavior problems than girls in later years. Yet, some gender differences
related to emotionality in infancy have been found, with boys being more negatively
emotional during mother-child interaction than girls (Bralmgart-Rieker, Courtney, &
Garwood, 1999). This study found that girls exhibited more self-control than boys at 36
months.

In this dissertation, the Strange Situation Procedure at 15 months also favored
girls; girls were significantly more likely to be securely attached than boys. In previous
research, significant gender differences with respect to attachment quality are not often
evident. Yet, a study of low-income African American families also found that boys
were significantly less likely to be securely attached than girls (Barnett et al., 1998). This
study measured attachment during the preschool years and not in infancy.

Results from the present data indicated that attachment at 24 months was related

to children’s self-control at 36 months for boys but not for girls. Other studies have

129

found that early attachment security and later behavioral expression varied by gender. A
study of middle-class families found an association for boy infants but not for girl infants
between insecure attachment and later psychopathology (Lewis, Feiring, McGuffog, &
Jaskir, 1984). Another study of toddlers found that early attachment was related to later
autonomy for girls; a similar association was not found for boys (Aber & Baker, 1990).

Other significant gender differences include effects of negative parenting and
marital conflict. First, the path between negative parenting and child self-control was
significant for girls but not for boys. Social learning theory suggests that girls are more
likely to model negative maternal behavior than boys (Crockenberg & Langrock, 2001).
For example, if parents show lack of self-control in their behavior with their children, it is
possible that girls are more likely to model that behavior. Thus, assuming that negative
parenting, such as angry or aggressive behavior toward the child, represents low maternal
self-control, girls would be more likely than boys to have lower self-control in response
to viewing maternal uncontrolled behavior. In the present study, a model including
marital conflict fit for boys but not for girls; other studies have found gender differences
related to children’s reactions to marital conflict (e.g., Davies, Harold, Goeke-Morey, &
Cummings, 2002).

It is of interest that significant gender differences in attachment security have
been noted among dual-earner households, but not among single-eamer households, with
more negative effects present for boys (Belsky & Rovine, 1988; Chase-Lansdale &
Owen, 1987). The oversampling of dual-earner households for this study may have
exaggerated gender differences in comparison with the national population. Further

study of gender differences and child care usage may clarify these questions.

130

Limitations

Certain limitations are inherent in secondary data analysis. In this particular
study, the diversity of measures and overall quality of measurement methods is above
average. Yet, primary research allows for more focus on a particular area of interest. In
this study, the use of partial measures rather than complete measures may have reduced
reliability, particularly on the measures of perceived control and temperament. The use
of the Early Infancy Temperament Questionnaire, which was under development at the
time of the study, rather than an established measure made it difficult to reach adequate
reliability. It would have been useful for the purposes of this study to have additional
measures of perceived competence, such as the domain-specific Maternal Self-Efficacy
Questionnaire.

Although the total sample size is more than adequate, division of the sample into
groups makes certain analyses difficult For example, the number of African American
families (n = 176) was too small for an accurate analysis using structural equation
modeling, especially considering attrition or missing data. An individual researcher
collecting primary data could have oversampled Afiican American families in order to
compare processes in African American and Caucasian families. The number of low-
income families and middle class families was sufficient for two separate structural
equation models; however, additional separate models comparing processes by child
gender were not possible.

The sample itself, although largely representative of demographics within the
United States, is not a truly random sample. A purposeful sampling plan reflecting

demographic diversity included mothers who planned to use child care full-time (60%)

131

or part-time (20%) and mothers who planned to stay at home (20%). The sample is
limited to ten localities, mostly in the Eastern United States. Of greater importance, the
sample excludes families living in unsafe neighborhoods, infants with perinatal
problems requiring extensive hospitalization, school-age mothers, and mothers with
insufficient English skills. In addition, those who declined to participate (42%) may
have had more difficulties in their families than those who agreed to participate. Thus,
the sample of low-income families is actually more advantaged than samples of low—
income families in other national studies. The exclusion of high-risk families may have
made it more difficult to detect differences between low-income and middle-to-high
income economic groups.

In addition to the exclusion of high-risk families, the NICHD Study of Early
Child Care excluded high-risk child care settings. Exclusion was not necessarily an
effect of predefined qualifications for the study; it appeared that child care settings that
were not optimal did not participate. A comparison between nationally representative
data on child care quality and the NICHD sample suggested that national child care
quality is lower and more variable (N ICHD, 2000). Exclusion of low quality child care
and high-risk families may have resulted in lower effect sizes of child care quality,
assuming that quality of care would have stronger effects among higher-risk groups.

Attrition was relatively low for this study; however, father data were missing at a
much higher rate. Documentation from NTCHD (1999a) cites l-month father data as
being collected at all sites; however, data were only available for 6 of the 10 sites.
Furthermore, data were not randomly missing even within the 6 sites, with fathers from

higher-risk families being less likely to participate. Although a separate study could be

132

made of father’s beliefs and involvement, it was not advisable to include father

information in a study of the total sample due to the lack of consistent data collection.

Future Directions

Earlier studies from this data set found limited effects of child care quality on
attachment security. However, there were significant interactions between parenting
practices and child care quality on attachment. Although this study replicated the finding
that child care quality combined with positive parenting predicted attachment security, it
is possible that child care quality by itself is a predictor of mother-child attachment and
child self-control. A structural equation model including only children in child care may
answer questions about the significance of child care quality to children’s developing
attachment relationships that may not be detected with the use of a regression model
(Newcombe, 2003).

Recent evidence suggests that significant amounts of time in child care during
the early years is related to socioemotional maladjustment (NICHD, 2003). Additional
effects of time in child care could also be investigated with this data set. Controls for
income may artificially isolate the effects of child care (Newcombe, 2003). Fmther
investigation of time in child care, attachment security, and behavior problems could use
structural equation modeling to examine pathways between variables.

Further study is also suggested to examine effects of child care quality and time
in child care among children of color. Child care quality, as measured in this
dissertation, was only available for 82 African American children. Limited analyses

could certainly be performed with this sample size. Yet, the NICHD has limited

133

information that may be relevant to the effect of child care on African American
children. Discrimination and segregation within the child care setting are not measured;
ethnic identity and perceived discrimination would also be helpful measures to ascertain
the effects of racial socialization. Additional primary studies are needed to further
investigate children’s environments, including child care, and their meaning for children
of color (Johnson, Jaeger, Randolph, Cauce, Ward, & NICHD, 2003).

In this dissertation, dimensions of parental efficacy had different effects on
family process among Afiican American families than among Caucasian families.
Although it is possible to speculate about reasons for differential effects, further research
is necessary to explain these differences. A study measuring ethnic identity, perceptions
of discrimination, and their relation to parental efficacy among families of color may
assist in understanding how these processes work. Certainly, more work could be done
among Latino, Asian, and Native American families to gain further cross-cultural

perspectives.

X Unexpected findings related to gender differences could be investigated with
ommwathpfimary research. It wouldbe particularlyusefill to investigate
father involvement when studying gender differences with self-control. Earlier research
has indicated that fathers play a unique role in children’s development, particularly in
the development of self-control (Braungart—Rieker, Garwood, Powers, & Wang, 2001).
Parenting styles of both fathers and mothers may differ between ethnicities; hence, it
would be useful to include diverse groups in this study.

It is of interest that we found few effects of perceived control on parenting itself,

although perceived control was associated with child outcomes. One study of mostly

134

middle—class Caucasian families found that neither global perceived competence or
domain-specific perceived competence was related to parenting, yet a domain-specific
measure of perceived competence was related to toddler behavior (Coleman & Karraker,
2003). Other studies have found more effects of perceived control on parenting among
low-income mothers than among higher-income mothers; however, these studies may
have confounded ethnicity with income. Further research among diverse groups may

explain these effects.

Conclusion

In summary, findings indicated
N" T
environments are predictive of attachment security and children’s self-control. Separate

that parental efficacy and children’s proximal
2W_

 

 

 

descriptive analyses of African American and Caucasian families added to our
understanding of children’s social-emotional development within different contexts. In
particular, findings regarding parental efficacy indicated distinct effects within Afi'ican
American and Caucasian groups. Furthermore, the traditional measurements of
sensitivity and attachment indicated possible cultural differences in the ways and
meanings of parenting. Dimensions of perceived competence and perceived control also
produced different effects on parenting style and child outcomes. Linkages between
parenting and child outcomes were generally stronger among low-income groups than
middle-to-high income groups, suggesting that proximal processes have a greater effect
among children living in disadvantaged families than among children with more
resources. Findings on gender, ethnicity, and class raised interesting questions that

warrant further investigation. In sum, this dissertation added to our understanding of

135

parental efficacy and its relation to various aspects of children’s environments,

attachment security, and child self—control within diverse demographic groups.

136

APPENDICES

137

APPENDIX A

MEASURES

138

Sense of Competence Subscale
1. When my baby came home from the hospital, I had doubtful feelings about my
ability to handle being a parent.
2. Being a parent is harder than I thought it would be.
3. I feel capable and on top of things when I am caring for my baby.
4. I can’t make decisions without help.
5. I have had many more problems caring for my baby than I expected.
6. I enjoy being a parent.

7. I feel that I am successful most of the time when I try to get my baby to do or not
do something.

8. Since I brought this baby home from the hospital, I find that I am not able to take
care of this baby as well as I thought I could. I need help.

9. I often have the feeling that I cannot handle things very well.

139

Perceived Control Subscale

8.

9.

. If a child has tantrums, no matter what you try you might as well give up.

My baby influences the number of friends I have.

I have often found that when it comes to my child, what is going to happen, will
happen.

My baby’s behavior is sometimes more that I can handle.

Being a good parent often depends on being lucky enough to have a good baby.
When something goes wrong with my baby there is little I can do to correct it.
It is often easier to let my baby have his/her way than to put up with a tantrum.
My baby does not control my life.

I feel like what happens in my life is mostly determined by my baby.

10. Fate was kind to me; if I had had a bad baby I don’t know what I would have

done.

11. What I do has little effect on my baby’s behavior.

12. I’m just one of those parents who happened to have a good child.

13. Sometimes I feel my baby’s behavior is hopeless.

14. My life is chiefly controlled by my baby.

140

Center for Epidemiological Studies — Depression

1. I was bothered by things that usually don’t bother me.
2. I felt that everything I did was an effort.

3. I felt I was just as good as other people.

4. I had trouble keeping my mind on what I was doing.
5. I felt sad.

6. I felt fearful.

7. I felt lonely.

8. I had crying spells.

9. I talked less than usual.

10. My sleep was restless.

11. I enjoyed life.

12. I felt that I could not shake off the blues even with the help of my family/friends.
13. I thought my life had been a failure.

14. I was happy.

15. I could not get “going”.

16. I felt hopeful about the future.

17. People were unfriendly to me.

18. I did not feel like eating; my appetite was poor.

19. I felt depressed.

20. I felt that people dislike me.

141

Marital Conflict Subscale

1. How often do you and your partner argue with one another?

2. To what extent do you try to change things about your partner that bother you (e. g.
behaviors, attitudes, etc.) ?

3. How confused are you about your feelings toward your partner?

4. How much do you think or worry about losing some of your independence by being
involved with your partner?

5. How often do you feel angry or resentful toward your partner?

6. How ambivalent or unsure are you about continuing in the relationship with your
partner?

7. To what extent do you feel that your partner demands or requires too much of your
time and attention?

8. To what extent do you feel “trapped” or pressured to continue in this relationship?

9. When you and your partner argue, how serious are the problems or arguments?

142

APPENDIX B

AD-HOC ANALYSES OF INCOME AND RACE

143

.28. vac:

 

 

 

 

no. 2 9.6 $2 523230

mo. 36 a: 3.6. 3.2 mm... 3 2282.205 582.22
35822055

.3. 2 «.2..» an 2 528250

8. m2. 3. ma. 2- 222.2 8.22 82 5°55 §E<
Hosanna

hm. «Om an 2 52.3030

mm. 2.”. an. 22 2 .2- we. mmd 82 22823:? 2.82.2.2
"2228222825

$6 2.2 2 232 528030

3. ma. 2.. mm. 22. 222.2 cm 5828252 5854‘
”.88 n2 ammo

22. v02 32 523250

222. 32 2a. 2 2.- 8.2. 3.2 02 2 22.823222 2282...:
”.88 v.2 ammo

9.. mm. m cm 2 523250

co. Sam 2m. 32 on. 222.24 2 2 502.55 2282.22
”002282288 2.928.682

2 m. 2.9m cm 2 523250

S. pen on. em. an. 24m n 22 2232.85 2282.22
”2888 2828.52
.flw w222|a>d~ ram use: QM :2 2 o2naum>

llldlqmooeetuar .20 2282 m Jana—822 .2 22 m

 

m. 30 fifimauafiu 3:3 :QUNLNEV 23.9% V. NEOSQJSQN 2NMENQ hmufifimkfiap 9

mm 2an

144

.28. V Q2...

 

 

 

 

 

2.. 2 N.- m m 2 52.8230

.:.. on.» 2.22. Q2- 3. on: 2.2. Serena §E<
”20.52.8225

om. em. 2222 528230

no. 8. 8. www- om. m 2. ma 2282.62.22 582.2,:
38822082

220m 362 mm 2328030

ea. 86 222. SN- 234 21.2 ma 582.2222 502.5.
52232. as 225

cc. 2 and an 2 523250

on. 2.28. No. and n 2 .2 and an 2282.225, §E<
H8222228 282.32

32 and 32 23286st

22a. 8. mm. mm: v.02 214 3 502.2222 2282.223.
”E2339: .3 e203

cm. 2 m~.n 32 523230

R. mm. 8. and- 32.2 3.2. 3 2282.55 §E<
$265.28 3282
am .4352 1am can: a. 2 a czar;

3% flag

9.2802 mm 033.

145

.28. V Re:

 

mm. 3.6 now 523250

.:.... a2 .2 no. and- 3.2 2.2.6 2 502.5223 §E<
”Bogotcm

on. 2 and n 2 w 2823230

22. no. 222. no. 2- 2.02 and 2 2232.05 §o£<
Essex

3. and n 2 w 5230.50

mm. 2:. M2. 32- Nm. mnN 2 582.282 §E<
”2823222825

36 3.6 new 528260

co. 9: 2 2. 20.2 no.» 56 2 2282.205 2.82.2.2
”.88 n2 Q-mmo

$26 32.2. awn 2823250

3.. on. 222. 3.2 36 NW2: 2 502.232 §E<
”.88 0.2 O-mmo

mm. 8.2. own 522830

2. 2mm mm. 3.. ov. 3.». 2 2232.225. §E<
6882288 2333.292

we. and am 528250

3. mm. an. 3.. mv. 3.2m 02 532.2222 2.88222
"22:28 2.920292
.3 I lacs 5 raw Elsi no. :2 2 23%|. >

alone-jg. a: m a: an 22% m

 

 

 

e. 3e agencaeb 236 toutoEV 58.52.. 2:822 fiME-S-eESE $953 $33.2me Q2233» 238 entomEWe b.2233? SN 28$

mm 2an

146

.28. V use...

 

2.22.. m 2. «9. 528030

2.N. mm. 2 8. m m.- 2.. mm: m 2 582.5252 §E<
”20bcoo-22om

om. mm. 229. 523230

mm. 2.2 22. 20.2- 2. mm. 2 582.2222 §E<
“22225208222

31. 09mm 3m 523250

cm. 8. .:.... and- 36 3.62 N2 582.2252 §oE<
522.32. 88 226

3. o2.m nan 528250

3. cum 8. 2.28 ea. $.m o2 2.82.2222. 22822:,
2202—2228 282.32

2 a. and non 2825830

we. 52. 2.2. 5.2- mo. 86 2 532.2222. 50$?
553mg .20 V23

an. m2..n non 282.8250

mm. 3.2 8. mm; mm. 2.96 n 2 502.2222 §E<

”922228.322 322282

am 834.82 ram oat-a: mm :2 a azaleal>
llhqlm-aeaza 223. m

 

ease 8 asap

147

REFERENCES

148

Aber, J ., & Baker, A. (1990). Secruity of attachment in toddlerhood: Modifying
assessment procedures for joint clinical and research purposes. In M. Greenberg,
D. Cicchetti, & E. Cummings (Eds), Attachment in the Preschool Years: Theory,
research, and intervention (pp. 427-460). Chicago, IL: University of Chicago
Press.

Abbott-Shim, M. & Sibley, A. (1992). Assessment profile for early childhood programs:
Research version Atlanta, GA: Quality Assist, Inc.

Abidin, R. R. (1986). Parenting stress index. Charlottesville, VA: Pediatric Psychology
Press.

Achenbach, T. M., Edelbrock, C., & Howell, C. (1987). Empirically-based assessment
of the behavioral/emotional problems of 2-3 year old children. Journal of
Abnormal Child Psychology, 15, 629-650.

Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of
attachment: A psychological stuay of the strange situation. Hillsdale, NJ:
Lawrence Erlbaum Associates.

\/ Ardelt, M., & Eccles, J. S. (2001). Effects of mother’s parental efficacy beliefs and
promotive parenting strategies on inner-city youth. Journal of Family Issues, 22,
944-972.

Bandura, A. (1982). Self-efficacy mechanism in human agency. American
Psychologist, 3 7, 122-147.

\/ Bandura, A. (1986). Social foundations of thought and action: A social cognitive theory.
Englewood Cliffs, NJ: Prentice Hall.

Bandura, A. (1989). Regulation of cognitive processes through perceived self-efficacy.
Developmental Psychology, 25, 729-735.

Bandura, A. (1994). Self-efficacy. In V. S. Ramachaudran (Ed), Encyclopedia of human
behavior (V 01. 4, pp. 71-81). New York: Academic Press.

Bandura, A. (1995). Self-eflicacy in changing societies. New York: Cambridge
University Press.

Bandura, A., Barbaranelli, C., Caprara, G. V., & Pastorelli, C. (1996). Multifaceted
impact of self-efficacy on academic functioning. Child Development, 67, 1206-
1222.

Barnett, D., Kidwell, S., & Leung, K. (1998). Parenting and preschooler attachment

among low-income urban African American families. Child Development, 69,
1657-1671.

149

Belsky, J. (1999). Child psychology: A Wk of contemporary issues. Philadelphia,
PA: Psychology Press.

Belsky, J ., & Isabella, R A. (1988). Maternal, infant, and social-contextual determinants
of attachment security. In J. Belsky & T. Nezworski (Eds.), Clinical implications
of attachment (pp. 41-94). Hillsdale, NJ: Lawrence Erlbaum.

Belsky, J ., Lang, M. & Rovine, M. (1985). Stability and change in marriage across the
transition to parenthood: A second study. Journ_a1 of Marriage and the Family. 47.
855-865.

Belsky, J., & Rovine, M. (1988). Nonmaternal care in the first year of life and the
security of infant-parent attachment. Child Development, 59, 157-167.

Bentler, P. M., & Bonett, D. G. (1980). Significance tests and goodness of fit in the
analysis of covariance structures. Psychological Bulletin, 88, 588-606.

Bornstein, M. H. (2002). Handbook of parenting. Mahwah, NJ: Erlbamn.
Bowlby, J. (1969). Attachment and loss, Vol. 1: Attachment. New York: Basic Books.

Bowlby, J. (1988). A secure base: Parent-child attachment and healthy human
development. New York: Basic Books.

Braiker, H. & ey, H. (1979). Conflict in the development of close relationships. In
- . urgess and T. Huston (Eds.), Social exchange and developing relationships.
New York: Academic Press.

Braungart, J. M., & Stifter, C. A. (1991). Regulation of negative reactivity during the
strange situation: temperament and attachment in 12-month-old infants. Infant
Behavior and Development, 1 4, 349-364.

Braungart-Rieker, J., Courtney, S., & Garwood, M. M. (1999). Mother- and father-
infant attachment: Families in context. Journal of Family Psychology, 13, 535-
553. .

Braungart-Rieker, J ., Garwood, M. M., Powers, B. P., & Notaro, P. C. (1998). Infant
affect and affect regulation during the still-face paradigm with mothers and
fathers: The role of infant characteristics and parental sensitivity.
Developmental Psychology, 34, 1428-1437.

Braungart-Rieker, J. Garwood, M. M., Powers, B. P., & Wang, X. (2001). Parental

sensitivity, infant affect, an affect regulation: Predictors of later attachment.
Child Development, 72, 252-270.

150

Brazelton, T. B. (1983). Assessment techniques for enhancing infant development. In J.
Call, E. Galenaon, & R. Tyson (Eds.), Frontiers of infant psychiatry. New York:
Basic Books.

Bretherton, I. (1985). Attachment theory: retrospect and prospect. In I. Bretherton & E.
Waters (Eds.) Growing points of attachment theory and research. Monographs of
the Society for Research in Child Development, 50 (1-2, Serial No. 209).

Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature
and design. Cambridge, MA: Harvard University Press.

Bronfenbrenner, U. (1992). Ecological systems theory. In R. Vasta (Ed), Six theories of
child development (pp. 187-249). London: Jessica Kingsley Publishers, Ltd.

Bronfenbrenner, U. (1999). Environments in developmental perspective: Theoretical and
operational models. In S. L. Friedman & T. D. Wachs (Eds.), Measuring
environment across the life span: Emerging methods and concepts. (pp. 3-2 8).
Washington, DC: American Psychological Association.

Bronfenbrenner, U. (2000). Ecological systems theory. In A. E. Kazdin, (Ed.)
Encyclopedia of psychology, Vol. 1. London: Oxford University Press.

Bronfenbrenner, U., & Morris, P. A. (1998). The ecology of developmental processes.
In W. Damon (Series Ed.) & R. M. Lerner (V 01. Ed.), Handbook of child
psychology: Vol. 1. Theory. (5"I ed., pp. 993-1028). New York: Wiley.

Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model fit. In K. A.
Bollen & J. S. Long (Eds.), Testing structural equation models (pp. 136-162).
Newbury Park, CA: Sage.

Bugental, D. B., Blue, J ., & Cruzcosa, M. (1989). Perceived control over caregiving
outcomes: Implications for child abuse. Developmental Psychology, 25, 532-539.

Bugental, D. B., & Shenum, W. A. (1984). “Difficult” children as elicitors and targets of
adult communication patterns: An attributional-behavioral transactional analysis.
Monographs of the Society for Research in Child Development, 49(], Serial No.
205).

Caldwell, B. M., & Bradley, R. H. (2000). The home inventory for the observation of
the environment. Little Rock, AK: University of Arkansas at Little Rock.

Calkins, S. D. (1994). Origins and outcomes of individual differences in emotion
regulation. Monographs of the Society for Research in Child Development, 59 (1 -
2, Serial No. 240).

Calkins, S. & Fox, N. (1992). The relations among infant temperament, security of

attachment, and behavioral inhibition at twenty-four months. Child Development,
63, 1456-1472.

151

Campis, L. K., Lyman, R. D., & Prentice-Dunn, S. (1986). The Parental Locus of
Control Scale: Development and validation. Journal of Clinical Child
Psychology, 15, 260-267.

Carey, W. B., & McDevitt, S. C. (1978). Revision of the infant temperament
questionnaire. Pediatrics, 61, 735-739.

Casady, M. A., Diener, M., Isabella, R., & Wright, C. (2001, April). Attachment security
among families in poverty: Maternal, child, and contextual characteristics. Poster
session presented at the biennial meeting of the Society for Research in Child
Development, Minneapolis, MN.

Cassidy, J. (1994). Emotion regulation: Influences of attachment relationships. In N. A.
Fox (Ed.), Emotion regulation: Behavioral and biological considerations.
Monographs of the Society for Research in Child Development (288-249).
Chicago: University of Chicago Press.

Cassidy, J ., & Shaver, P. R. (1999). Handbook of attachment: Theory, research, and
clinical applications. New York: Guilford Press.

Chase-Lansdale, P. L., & Owen, M. T. (1987). Maternal employment in a family
context: Effects on infant-mother and infant-father attachments. Child
Development, 58, 1505-1512.

Cicchetti, D., & Toth, S. (1990). Developmental Perspectives on Depression. Rochester,
NY: University of Rochester Press.

Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences. Hillsdale, NJ:
Lawrence Erlbaum Associates.

Cohen, J. (1992). A power primer. Psychological Bulletin, 112, I, 155-159.

Cohn, J ., & Campbell, S. B. (1992). Influence of maternal depression on infant affect
regulation. In D. Cicchetti and S. L. Toth (Eds.), Developmental perspectives on
depression (pp. 103-130). Rochester, NY: University of Rochester Press.

\/ Coleman, P. K., & Karraker, K. H. (1997). Self-efficacy and parenting quality: Findings
and future applications. Developmental Review, 18, 47-85.

\/ Coleman, P. K., & Karraker, K. H. (2003). Maternal self-efficacy beliefs, competence in
parenting, and toddlers' behavior and developmental status. Infant Mental Health
Journal, 24, 126-148.

/ Corapci, F ., & Wachs, T. D. (2002). Does parental mood or efficacy mediate the

influence of environmental chaos upon parenting behavior? Merrill-Palmer
Quarterly, 48, 182-201.

152

Crittenden, P. M. (2000). Introduction. In P. M. Crittenden and A. H. Claussen (Eds.)
The organization of attachment relationships: Maturation, culture, and context.
Cambridge, UK: Cambridge University Press.

Crnic, K. A., Greenberg, M. T., & Slough, N. M. (1986). Early stress and social support
influences on mothers’ and high-risk infants’ functioning in late infancy. Infant
Mental Health Journal, 7, 19-33.

Crockenberg, S. (1981). Infant irritability, mother responsiveness, and social influences
on the security of mother-infant attachment. Child Development, 52, 857-865.

Crockenberg, S. B., & Langrock, A. (2001). The role of specific emotions in children’s
responses to interparental conflict: A test of the model. Journal of Family
Psychology, 15, 163-182.

Cummings, E. M., & Davies, P. (1996). Emotional security as a regulatory process in
normal development and the development of psychopathology. Development and
Pscyhopathology, 8, 123-139.

Davies, P. T., Harold, G. T., Goeke-Morey, M. C., & Cummings, E. M. (2002). Child
emotional security and interparental conflict. Monographs of the Society for
Research in Child Development, 6 7(3, Serial No. 270).

De Wolff, M., & van IJzendoorn, M. H. (1997). Sensitivity and attachment: A meta-
analysis on parental antecedents of infant attachment. Child Development, 68,
571-591 .

Deater-Deckard, K., Dodge, K. A., Bates, J. E., Pettit, G. S. (1996). Physical discipline
among African American and European American mothers: Links to children’s
externalizing behaviors. Developmental Psychology, 32, 1065-1072.

del Carmen, R., Pedersen, F. A., Huffman, L. C., & Bryan, Y. E. (1993). Dyadic
distress management predicts subsequent security of attachment. Infant
Behavior and Development, 16, 131-147.

Dempster, A. P., Laird, N. M., & Rubin, D. B. (1977). Maximum likelihood fi'orn
incomplete data via the EM algorithm. Journal of the Royal Statistical Society,
Series B 39, 1-38.

Deutsch, F. M., Ruble, D. N., Fleming, A., Brooks-Gunn, J., & Stangor, G. S. (1988).
Information seeking and maternal self-definition during the transition to
motherhood. Journal of Personality and Social Psychology, 55, 420-431.

Diener, M., Nievar, M. A., & Wright, C. (2003). Attachment security between

disadvantaged young children and their mothers: Associations with maternal,
child and contextual characteristics. Merrill Palmer Quarterly, 49, 154-182.

153

Donovan, W., & Leavitt, L. (1989). Maternal self-efficacy and infant attachment:
Integrating physiology, perceptions, and behavior. Child Development, 60, 460-
472.

Donovan, W. L., Leavitt, L. A., & Walsh, R. O. (1997). Cognitive set and coping
strategy affect mothers’ sensitivity to infant cries: A signal detection approach.
Child Development, 68, 760-772.

Donovan, W. L., Leavitt, L. A., & Walsh, R. O. (2000). Maternal illusory control
predicts socialization strategies and toddler compliance. Developmental
Psychology, 36, 1-10.

Dumka, L. E., Stoerzinger, H. D., Jackson, K. M., & Roosa, M. W. (1996).

Examination of the cross-cultural and cross-language equivalence of the
Parenting Self-Agency Measure. Family Relations, 45, 216-222.

Duncan, G. J ., & Brooks-Gunn, J. (2000). Family poverty, welfare reform, and child
development. Child Development, 71, 188-196.

F owles, D. C., & Kochanska, G. (2000). Temperament as a moderator of pathways to
conscience in children: The contribution of electrodermal activity.
Psychophysiology, 3 7, 788-795.

French, J ., Rodgers, W., & Cobb, S. (1974). Adjustment as person-environment fit. In
G. Cochlo, D. Harnberg, & J. Adams (Eds.), Coping and adaptation. New York:
Basic.

Gauthier, Y. (2003). Infant mental health as we enter the third millennium: Can we
prevent aggression? Infant Mental Health Journal, 24, 296-308.

Gjerde, P. F. (2001). Attachment, culture, and amae. American Psychologist, 56, 826-
827.

Goldberg, W., & Easterbrooks, M. (1984). Role of marital quality in toddler
development. Developmental Psychology, 20, 504-514.

Goldsmith, H. H., & Alansky, J. A. (1987). Maternal and infant temperamental
predictors of attachment: A meta-analytic review. Journal of Consulting and
Clinical Psychology, 55, 805-816.

Goldsmith, H. H., Buss, A. H., Plomin, R., Rothbart, M. K., Thomas, A., Chess, 8.,

Hinde, R. A., & McCall, R. R. (1987). Roundtable: What is temperament? Four
approaches. Child Development, 58, 505-529.

154

Gondoli, D., & Silverberg, S. (1997). Maternal emotional distress and diminished
responsiveness: The mediating role of parenting efficacy and parental perspective
taking. Developmental Psychology, 33, 861 -868.

Gonzales, N. A., Cauce, A. M., Friedman, R. J., & Mason, C. A. (1996). Family, peer,
and neighborhood influences on academic achievement among African-American
adolescents: One-year prospective effects. American Journal of Community
Psychology, 24, 365-387.

Grossman, K., Grossman, K. E., Spangler, G., Suess, G., & Uzner, L. (1985). Maternal
sensitivity and newborns’ orientation responses as related to quality of attachment
in northern Germany. In I. Bretherton & E. Waters (Eds.) Growing points in
attachment theory and research. Monographs of the Society for Research in Child
Development, 50(1-2, Serial No. 209).

Grusec, J. E., Hastings, P., & Mammone, N. (1994). Parenting cognitions and
relationship schemas. In W. Damon (Series Ed.) & J. G. Smetana (V 01. Ed.),
Beliefs about parenting: Origins and developmental implications. New Directions
for Child Development, 66, 5-20. San Francisco: Josey-Bass.

Hamilton, S., & Luster, T. (2003). Ecological systems theory. In J. M. Miller, R. M.
Lerner, L. B. Schiarnberg, Anderson (Eds). Human ecology: An encyclopedia of
children, families, communities, and environments. Santa Barbara, CA: ABC-
Clio.

Harkness, S., & Super, C. M. (1992). Parental ethnotheories in action. In I. E. Sigel, A.
V. McGillicuddy-Delisi, & J. J. Goodnow (Eds.). Parental belief systems.
Hillsdale, NJ: Lawrence Erlbaum.

Harris, J. R. (2002). Beyond the nurture assumption: Testing hypotheses about the child's
environment. In J. G. Borkowski & S. L. Rarney (Eds.), Parenting and the child's
world: Influences on academic, intellectual, and social-emotional development.
Monographs in parenting. (pp. 3-20). Mahwah, NJ: Lawrence Erlbaum.

Harter, S. (1978). Effectance motivation reconsidered: Toward a developmental model.
Human Development, 21, 34-64.

Hastings, R. P., & Brown, T. (2002). Behavior problems of children with autism,
parental self-efficacy, and mental health. American Journal on Mental
Retardation, 107, 222-232.

Hill, J ., Fonagy, P., Safier, E., & Sargent, J. (2003). The ecology of attachment in the
family. Family Process, 42, 205-221.

Hinde, R. A., & Stevenson-Hinde, J. (1990). Biological, cultural, and individual
desiderata. Human Development, 33, 62-72.

155

Hoath, F. E., & Sanders, M. R. (2002). A feasibility study of enhanced group triple P -
Positive Parenting Program for Parents of Children with Attention-
deficit/Hyperactivity Disorder. Behavior Change, 19, 191 -206.

Hofferth, S. L., Smith, J ., McLoyd, V. C., & Finkelstein, J. (2000). Achievement and
behavior among children of welfare recipients, welfare leavers, and low-income
single mothers. Journal of Social Issues, 5 6, 747-774.

Hosmer, D. W., & Lemeshow, S. (1989). Applied logistic regression. New York:
Wiley.

Hughes, D., & Johnson, D. (2001). Correlates in children’s experiences of parents’
racial socialization behaviors. Journal of Marriage and the Family, 63, 981-
985.

Isabella, R. A. (1993). Origins of attachment: Maternal interactive behavior across the
first year. Child Development, 64, 605-621.

Isabella, R. A., Belsky, J ., & von Eye, A. (1989). Origins of infant-mother attachment:
An examination of interactional synchrony during the infant’s first year.
Developmental Psychology, 25, 12-21.

Izard, C. E., Haynes, O. M., Chisholm, G., & Baak, K. (1991). Emotional determinants
of infant-mother attachment. Child Development, 62, 906-917.

Johnson, D., Jaeger, E., Randolph, S. M., Cauce, A. M., Ward, J ., & NICHD Early Child
Care Research Network (2003). Studying the effects of early child care

experiences on the development of children of color in the United States:
Toward a more inclusive research agenda. Child Development, 74, 1227-1244.

Johnston, C., & Mash, E. J. (1989). A measure of parenting satisfaction and efficacy.
Journal of Clinical Child Psychology, 18, 167-175.

Kagan, J. (1998). Biology and the child. In W. Damon (Series Ed.) & N. Eisenberg
(V 01. Ed.), Handbook of child psychology: Vol. 111. Social, emotional, and
personality development (5th ed., pp. 177-235). New York: Wiley.

Kart-Morse, R., & Wiley, M. S. (1997). Ghosts fiom the nursery: Tracing the roots of
violence. New York: Atlantic Monthly Press.

Klein, L. (1998). Path analysis. In L. G. Grimm and P. R. Yamold (Eds.) Reading and
understanding multivariate statistics (pp. 65-96). Washington, D. C.: American
Psychological Association.

Kline, R B. (1998). Principles and practice of structural equation modeling. New York:
Guilford.

156

Kochanska, G. (1998). Mother-child relationship, child fearfulness, and emerging
attachment: A short-term longitudinal study. Developmental Psychology, 3 4,
480-490.

Kochanska, G., & Knaack, A. (2003). Effortfirl control as a personality characteristic of
young children: Antecedents, correlates, and consequences. Journal of
Personality, 71, 1087-11 12.

Kochanska, G., Murray, K. T., & Harlan, E. T. (2000). Effortful control in early
childhood: Continuity and change, antecedents, and implications for social
development. Developmental Psychology, 3 6, 220-232.

Kondo-Ikemura, K. (2001). Insufficient evidence. American Psychologist, 5 6, 825-826.

Levine, R.A., Dixon S., Levine, S., Richman, A., Liederrnan, P.H., Keefer, C.,
Brazelton, T. (1994). Child care and culture: Lessons fiom Afi'ica. Cambridge,
UK: Cambridge University Press.

Lewis, M., Feiring, C., McGuffog, C., & Jaskir, J. (1984). Predicting psychopathology
in six-year-olds from early social relations. Child Development, 55, 123-136.

Little, R. J. A., & Rubin, D. B. (1987). Statistical analysis with missing data. New
York: Wiley.

Little, R. J, & Rubin, D. B. (1989). The analysis of social science data with missing
values. Sociological Methods and Research, 18, 292-326.

Londerville, S., & Main, M. (1981). Security of attachment, compliance, and maternal
training methods in the second year of life. Developmental Psychology, 1 7, 289-
299.

Lovejoy, M. C., Verda, M. R., & Hays, C. E. (1997). Convergent and discriminant
validity of measures of parenting efficacy and control. Journal of Clinical Child
Psychology, 26, 366-376.

Ludbrook, J. (1998). Multiple comparison procedures updated Clinical and
Experimental Pharmacolog and Physiology, 25, 1032-1037.

Luster, T., & Kain, E. (1987). The relation between family context and perceptions of
parental efficacy. Early Child Development and Care, 29, 201-311.

Luster, T., & Rhoades, K. (1989). The relation between child-rearing beliefs and the

home environment in a sample of adolescent mothers. Family Relations, 38,
317-322.

157

Lyons-Ruth, K. (1992a). Maternal depressive symptoms, disorganized infant-mother
attachment relationships and hostile-aggressive behavior in the preschool
classroom: A prospective longitudinal view from infancy to age five. In D.
Cicchetti & S. Toth (Eds.), Developmental Perspectives on Depression (pp. 131-
171). Rochester, NY: University of Rochester Press.

Lyons-Ruth, K. (1992b). Maternal depression. In D. Cicchetti & S. Toth (Eds.),
Developmental Perspectives on Depression Rochester, NY: University of
Rochester Press.

Lyons-Ruth, K. (1996). Attachment relationships among children with aggressive
behavior problems: The role of disorganized early attachment patterns. Journal
of Consulting and Clinical Psychology, 64, 64-73.

Machida, S., Taylor, A. R., & Kim, J. (2002). The role of maternal beliefs in predicting
home learning activities in head start families. Family Relations, 51, 176-184.

Main, M., & Solomon, J. (1990). Procedures for identifying infants as
disorganized/disoriented during the Ainsworth strange situation. In M.
Greenberg, D. Cicchetti, & E. M. Cummings (Eds.) Attachment in the preschool
years: Theory, research, and intervention (pp. 121-160). Chicago, IL: University
of Chicago Press.

Mangelsdorf, S., Gunnar, M., Kestenbaum, R., Lang, S., & Andreas, D. (1990). Infant
proneness-to-distress temperament, maternal personality, and mother-infant
attachment: Associations and goodness of fit. Child Development, 61, 820-831.

Martins, C., & Gaffan, E. A. (2000). Effects of early maternal depression on patterns of
infant-mother attachment: A meta-analytic investigation. Journal of Child
Psychology and Psychiatry and Allied Disciplines, 41, 73 7-746.

Mash, E. J., & Johnston, C. (1983). The prediction of mothers’ behaviors with their
hyperactive children during play and task situation. Child and Family Behavior
Therapy, 5, 1-14.

McLoyd, V. C., Cauce, A. M., Takeuchi, D., & Wilson, L. (2000). Marital processes and
parental socialization in families of color: A decade review of research. Journal
of Marriage and Family, 62, 1070-1093.

Medoff-Cooper, B., Carey, W. B., & McDevitt, S. C. (1993). Early Infancy
Temperament Questionnaire. Journal of Developmental and Behavioral
Pediatrics, 14, 230-235.

Murry, V. M. (2000). Challenges and experiences of Black American families. In P. C.

McKenry & S. J. Price (Eds.), Families and change: Coping with stressful events.
(2"d ed., pp. 333-358).

158

Murry, V. M., Brown, P. A., Brody, G. H., Cutrona, C. E., & Simons, R. L. (2001).
Racial discrimination as a moderator of the links among stress, maternal
psychological functioning, and family relationships. Journal of Marriage and
Family, 63, 915-926.

Newcombe, N. S. (2003). Some controls control too much. Child Development, 74,
1050-1052.

NICHD Early Child Care Research Network (1997). The effects of infant child care on
infant-mother attachment security: Results of the NICHD Study of Early Child
Care. Child Development, 68, 860-879.

NICHD Study of Early Child Care (1999a). Phase I Instrument Document.
http://secc.rti.org=

NICHD Early Child Care Research Network (1999b). Child care and mother-child
interaction in the first 3 years of life. Developmental Psychology, 35, 1399-1413.

NICHD Early Child Care Research Network (2000). Characteristics and quality of child
care for toddlers and preschoolers. Journal of Applied Developmental Sciences,
4, 116-135.

NICHD Early Child Care Research Network (2001). Child care and family predictors of
preschool attachment and stability from infancy. Developmental Psychology, 3 7,
847-862.

NICHD Early Child Care Research Network (2003). Does amount of time spent in child
care predict socioemotional adjustment during the transition to kindergarten?
Child Development, 74, 976-1005.

NICHD Early Child Care Research Network, & Duncan, G. J. (2003). Modeling the
impacts of child care quality on children’s preschool cognitive development.
Child Development, 74, 1454-1475.

Nievar, M. A., & Brophy-Herb, H. (2003, April). Parental Eflicacy and School
Readiness for Kids: What Are the Connections? Symposium presented at the
Biennial Conference of the Society for Research in Child Development, Tampa,
FL.

Ogbu, J. (1985). A cultural ecology of competence among inner-city blacks. In M. B.
Spencer, G. K. Brookins, & W. R. Allen (Eds.), Beginnings: The social and
aflective development of black children. Lawrence Erlbaum: Hillsdale, NJ.

Olds, D. (1997). The prenatal/early infancy project: Fifteen years later. In G. Albee and

T. Gullotta (Eds.), Primary prevention works. Issues in Children ’s and Families’
Lives, 6. (pp. 41-67). Sage Publications: Thousand Oaks, CA.

159

Olds, D., & Korfmacher, J. (1997). The evolution of a program of research on prenatal
and early childhood home visitation: Special issue introduction. Journal of
Community Psychology, 25, 1-7.

Ozer, E. M., & Bandura, A. (1990). Mechanisms governing empowerment effects: A
self-efficacy analysis. Journal of Personality and Social Psychology, 58, 472-
486.

Papoulis, A. (1990). Probability and Statistics. Pearson Higher Education. Upper
Saddle River, NJ.

Parke, R. D. (2002). Fathers and families. In M. H. Bomstein (Ed), Handbook of
parenting: Vol. 3: Being and becoming a parent. (pp. 27-73). Mahwah, NJ:
Lawrence Erlbaum.

Pederson, D. R., & Moran, G. (1996). Expressions of the attachment relationship outside
of the Strange Situation. Child Development, 6 7, 915-927.

Peterson, L., Farmer, J ., & Kashani, J. (1990). Parental injury prevention endeavors: A
function of health beliefs? Health Psychology, 9, 177-191.

Polaha, J. A. (1999). The relationship between physical discipline and child behavior
problems: A study of group differences. Dissertation Abstracts International, 59
(08), 451 1B.

Posada, G., Gao, Y., Wu, F ., Posada, R., Tascon, M., Schoelmelich, A., Sagi, A.,
Kondo-Ikemura, K., Ylaland, W., & Synnevaag, B. (1995). The secure-base
phenomenon across cultures: Children's behavior, mother's preferences, and
experts' concepts. In E. Waters, B. Vaughn, G. Posada, & K. Kondo-Ikemura
(Eds.), Caregiving, cultural, and cognitive perspectives on secure-base behavior
and working models: New growing points of attachment theory and research
(pp.27-48). Monographs of the Society for Research in Child Development, 60
(2-3, Serial No. 244).

Posada, G., Jacobs, A., Richmond, M. K., Carbonell, O. A., Alzate, G., Bustamante, M.
R., & Quiceno, J. (2002). Maternal caregiving and infant security in two
cultures. Developmental Psychology, 38, 67-78.

Radke-Yarrow, M. (1991). Attachment patterns in children of depressed mothers. In C.
M. Parkes, J. Stevenson-Hinde, & P. Marris aids), Attachment Across the Life
Cycle (pp.115-126). London: Routledge.

Radloff, LS. (1977). The CES-D scale: A self-report depression scale for research in the
general population. Applied Psychological Measurement, 1, 385-401 .

160

Renken, B., Egeland, B., Marvinney, D., Mangelsdorf, S., & Sroufe, L. A. (1989). Early
childhood antecedents of aggression and passive withdrawal in early elementary
school. Journal of Personality, 5 7, 257-281.

Ritter, G., & Moyi, P. L. (2003). The impact of day care on school readiness: Using new
data to examine the controversy. Paper symposium presented at the Biennial
Conference of the Society for Research in Child Development, Tampa, FL.

Roberts, RE. (1980). Reliability of the CES-D scale in different ethnic contexts.
Psychiatry Research, 2, 125- 1 34.

Eisenberg (V 01. Ed), Hancbook of child psychology: Vol. 111. Social, emotional,

(Rothbart, M. K., & Bates, J. E. (1998). Temperament. In W. Damon (Series Ed.) & N.
and personality development (58- ed., pp. 105-176). New York: Wiley.

Rothbaum, F., Pott, M., Azuma, H., Miyake, K., & Weisz, J. (2000). The development
of close relationships in Japan and the United States: Paths of symbiotic harmony
and generative tension. Child Development, 71, 1121-1142.

Rothbaum, F ., Weisz, J ., Pott, M., Miyake, K., & Morelli, G. (2000). Attachment and
culture: Security in the United States and Japan. American Psychologist, 55,
1093-1104.

Rothbaum, F., Weisz, J ., Pott, M., Miyake, K., & Morelli, G. (2001). Deeper into
attachment and culture. American Psychologist, 5 6, 827-829.

Rotter, J. B. (1990). Internal versus external control of reinforcement: A case history of
a variable. American Psychologist, 45, 489-493.

Sagi, A., Lamb, M. E., Lewkowicz, K. S., Shoharn, R., Dvir, R., & Estes, D. (1986).
Security of infant-mother, -father, and —metapelet attachments among kibbutz-
reared Israeli children. In I. Bretherton & E. Waters (Eds.), Growing points in
attachment theory and research. Monographs of the Society for Research in Child
Development, 50(1-2, Serial No. 209).

Scarr, S. (1998). American child care today. American Psychologist, 53, 95-108.

Schumacker, R. E., & Lomax, R. G. (1996). A Beginner ’s Guide to Structural Equation
Modeling. Lawrence Erlbaum: Mahwah, NJ.

Seifer, R., Schiller, M., Samerofl‘, A. J., Resnick, S., & Riordan, K. (1996). Attachment,

maternal sensitivity, and infant temperament during the first year of life.
Developmental Psychology, 32, 12-25.

161

Shonkoff, J. P. & Phillips, D. A. (2000). From neurons to neighborhoods: the science of
early chilcflrood development. Washington, DC: National Academy Press.

Sigel, I. E., McGillicuddy-DeLisi, A. V., & Goodnow, J. J. (1992). Parental belief
systems: The psychological consequences for children. Hillsdale, N.J.: Erlbaum.

Smetana, J. G. (1994). Beliefs about parenting: Origins and developmental implications.
New Directions for Child Development, 66.

Solomon, J. & George, C. (1999). The measurement of attachment security in infancy
and childhood. In J. Cassidy & P. Shaver (Eds.), Handbook of Attachment:
Theory, Research and Clinical Applications (pp. 287-316). New York: Guilford
Press.

Sofronoff, K., & F arbotko, M. (2002). The effectiveness of parent management training
to increase self-efficacy in parents of children with Asperger syndrome. Autism,
6, 271-286.

Spieker, S. J ., & Booth, C. L. (1988). Maternal antecedents of attachment quality: What
makes social risk risky? In J. Belsky & T. Nezworski (Eds.), Clinical
implications of attachment (pp. 95-135). Hillsdale, N.J.: Erlbaum.

Spoth, R., & Conroy, S. (1993). Survey of prevention-relevant beliefs and efforts to
enhance parenting skills among rural parents. The Journal of Rural Health, 9,
227-239.

Sroufe, L. A. (1985). Attachment classification from the perspective of infant-caregiver
relationships and infant temperament. Child Development, 5 6, 1-14.

Susman-Stillman, A., Kalkoske, M., Egeland, B. & Waldman, I. (1996). Infant
temperament and maternal sensitivity as predictors of attachment security. Infant
Behavior and Development, 19, 33-47.

Teti, D. M., & Gelfand, D. M. (1991). Behavioral competence among mothers of infants
in the first year: The mediational role of maternal self-efficacy. Child
Development, 62, 918-929.

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, 364-3 76.

Teti, D. M., Nakagawa, M., Das, R., & Wirth, O. (1991). Security of attachment between
preschoolers and their mothers: Relations among social interaction, parenting
stress, and mothers’ sorts of the Attachment Q-Set. Developmental Psychology,
27, 440-447.

162

Teti, D. M., O’Connell, M. A., & Reiner, C. D. (1996). Parenting sensitivity, parental
depression and child health: The mediational role of parental self-efficacy. Early
Development and Parenting, 5, 237-250.

Thomas, A., & Chess, S. (1977). Temperament and Development. New York:
Brunner/Mazel.

Thompson, R A., & Lamb, M. E. (1983). Security of attachment and stranger sociability
in infancy. Developmental Psychology, 1 9, 184-191.

Tronick, E., Morelli, G. A. & Ivey, P. [C (1992). The infant forager infant and toddler’s
patterns of social relationships: Multiple and simultaneous. Developmental
Psychology, 28, 568-577.

True, M. M., Pisani, L., & Oumar, F. (2001). Infant-mother attachment among the
Dogon of Mali. Child Development, 72, 1451-1466.

Ullman, J. (1996). Structural equation modeling. In B. Tabachnik & L. Fidel] (Eds.),
Using Multivariate Statistics (3rd ed) (pp. 709-811). New York: HarperCollins.

van Uzendoom, M. H., & Kroonenberg, P. M. (1988). Cross-cultural patterns of
attachment: A meta-analysis of the strange situation. Child Development, 59, 147-
1 56.

van IJzendoorn, M. H., & Sagi, A. (1999). Cross-cultural patterns of attachment:
Universal and contextual dimensions. In J. Cassidy and P. R. Shaver’s (Eds.)
Handbook of attachment: Theory, research, and clinical applications (pp. 713-
734). New York: Guilford Press.

Vaughn, B., & Bost, K. K. (1999). Attachment and temperament: Redundant,
independent, or interacting influences on interpersonal adaptation and personality
development? In J. Cassidy and P. R. Shaver’s (Eds.) Handbook of attachment:
Theory, research, and clinical applications. (pp. 713-734). New York: Guilford
Press.

Vaughn, B., Stevenson-Hinde, J., Waters, E., Kotsafiis, A., Lefever, G. B., Shouldice, A.,
Trude], M., & Belsky, J. (1992). Attachment security and temperament in infancy
and early childhood: Some conceptual clarifications. Developmental Psychology,
28, 463-473.

Watamura, S. E., Donzella, B., Alwin, J., & Gunnar, M. R. (2003). Morning-to-

afiemoon increases in cortisol concentrations for infants and toddlers at child care:
Age differences and behavioral correlates. Child Development, 74, 1006-1020.

163

Waters, E. (1995). Appendix A: The attachment Q-set (version 3.0). In E. Waters, B.
Vaughn, G. Posada, & K. Kondo-Ikemura (Eds.), Caregiving, cultural, and
cognitive perspectives on secure-base behavior and working models: New
growing points of attachment theory and research. Monographs of the Society for
Research in Child Development, 60 (2-3, Serial No. 244).

Waters, E., Wippman, J ., & Sroufe, L. A. (1979). Attachment, positive affect, and
competence in the peer group: Two studies in construct validation. Child
Development, 50, 821-829.

Waters, E. & Deane, K. (1985). Defining and assessing individual differences in
attachment relationships: Q-methodology and the organization of behavior in
infancy and early childhood. In I. Bretherton & E. Waters (Eds.), Growing points
of attachment theory and research. Monographs of the Society for Research in
Child Development.

Zahn-Waxler, C., Ianotti, R. J., Cummings, E. M., & Denham, S. (1990). Antecedents

of problem behaviors in children of depressed mothers. Development and
Psychopathology, 2, 271 -291 .

164

IIIIIIIIIIIIIIIIIIIIIIIIIIIIII

will111111111111W

1293 0250