\
VI
3... .x
. 3%.“.
A r2 3.3.:
:2 ; {~sz ..

c.

.w 2E2.
If

 

wiumhbvuw-ﬂ:

. . h}; V. V:
$35.41;... w
3. 5:3

I I,
Bi... ‘. ('1. .,

 

for "l LIBRARY
.' ,. -‘ Michigan State
University

This is to certify that the
dissertation entitled

EXAMINING THE EFFECTS OF COPING ON THE
RELATIONSHIP BETWEEN FAMILY STRESS AND
MOTHER-CHILD INTERACTIONS ACROSS TIME

presented by
LORRAINE M. McKELVEY

has been accepted towards fulﬁllment
of the requirements for the

PhD. degree in Pslcholgqy

 

 

Oﬂ¢ﬂm¥

/ Major, Professﬁ’s Signature

12/11/03

 

Date

MSU is an Afﬁrmative Action/Equal Opportunity Institution

.0-I---I-n-a-h-I-n-I-n-o-I-t-I-c-o-I_-. -

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

DATE DUE

 

state/W

 

MAY 1 02009

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/ClRC/DateDue.p65—p. 1 5

 

 

EXAMINING THE EFFECTS OF COPING ON THE RELATIONSHIP BETWEEN
FAMILY STRESS AND MOTHER-CHILD INTERACTIONS ACROSS TIME

BY

Lorraine M. McKelvey

A DISSERTATION

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

DOCTOR of PHILOSOPHY
Department of Psychology, College of Social Sciences

Interdepartmental Graduate Specialization
in Applied Developmental Science

2003

ABSTRACT

EXAMINING THE EFFECTS OF COPING ON THE RELATIONSHIP BETWEEN
FAMILY STRESS AND MOTHER-CHILD INTERACTIONS ACROSS TIME

BY
Lorraine M. McKelvey

This study investigated the relation between stress and mother-child
interaction in low-income, Early Head Start eligible families. The sources of
stress that were examined were twofold: global adversity, which included
conditions that are chronic and less amenable to change, such as teenage
parenthood and premature birth; and speciﬁc family stress, which was quantiﬁed
by reports of family conﬂict and parenting stress. Adversity, parenting stress,
and family conﬂict have been repeatedly found to negatively impact the
interactions between parents and their young children. Family support seeking
and cognitive reframing coping strategies were examined as possible moderators
of the relationship between family stress and mother-child interaction.

Results from path analyses demonstrated that relationships among target
variables differed across time with the single consistent predictor of mother-child
interaction being parenting stress. This construct negatively impacted mother-
child interaction at both 24 and 36 months. Family conﬂict did not directly impact
dyadic interactions, but did inﬂuence mothers’ reports of parenting stress;

mothers who reported more conﬂict also experienced greater amounts of

parenting stress. Family adversity did not predict parenting stress, family conﬂict
or mother-child interactions as assessed at either age of the child.

Results also supported effects of coping on mother-child interaction.
Cognitive reframing explained a greater percentage of variance in mother-child
interaction at 24 months, while support seeking explained a greater percentage
of variance in the same measure at 36 months. Support seeking coping reported
at 24 and 36 months demonstrated direct effects on mother-child interaction at
36 months, unexpectedly in different directions. Cognitive reframing, on the
other hand did not directly predict mother-child interaction. Parenting stress and
cognitive reframing were related in that heightened stress related to lesser
cognitive reframing. Parenting stress and cognitive reframing also interacted to
signiﬁcantly predict mother-child interactions assessed at 24 months.

There were no demonstrated impacts of the total number of EHS program
hours devoted to the families on mother-child interaction. A more thorough
examination of data concerning families’ perceptions of need, family engagement
in the program and staff turnover within family may prove to be potential

predictors of differential outcomes for mothers and children.

Copyright by
LORRAINE M. MCKELVEY
2003

Dedicated in loving memory of Eugenia Neuner

ACKNOWLEDGEMENTS

This dissertation would not have been possible without the assistance and
support from many individuals. First of all, I must give thanks to the families who
gave of themselves for the sake of this research. I am indebted to them for their
gift of time, patience, and trust. I hope that the work presented in this
dissertation and that of the future will not only accurately represent those
families, but also work to inform programs and communities about their strengths
and needs.

I owe a debt of gratitude to the Michigan State University Early Head Start
Research Team. I could not have asked for a better environment in which to
develop and to learn about collaboration. From the time I started working with
the members of the team, I have been welcomed as a colleague and my
thoughts and interpretations have always been acknowledged. The individuals
on the research team have taught me many lessons about respect and collective
effort, for which I would like to thank each of them.

In addition to the team of individuals who have invested time and
commitment to the research study, I would also like to thank those who have
been inﬂuential in my professional development. I would like to extend special
thanks to my advisor, Hiram Fitzgerald, for his many years of support,
encouragement, and guidance. His dedication to my success went beyond the

ordinary and I owe him much appreciation. I would also like to thank Rachel

vi

Schiffman. She has provided me with every opportunity to collaborate with the
MSU and the national EHS research teams over the years. She has been
instrumental in fostering my professional career and has dedicated much time
and effort to my development. I cannot possibly thank the both of them enough
for all of their support.

The members of my committee offered their unconditional time and
dedication to this dissertation, for which I am appreciative. Alex von Eye has
spent much time teaching me statistics and answering the questions that
resulted. Never once in the many years that l have asked him unsophisticated
questions about methods did he reply in a manner in which I might be made
aware of exactly how basic my question might have been. I owe him much
gratitude for his patience and understanding. I would also like to thank Cris
Sullivan and Holly Brophy—Herb for their time and insights. Their efforts to
provide feedback from alternative lenses have been so valuable. Thank you to
you all!

Finally, acknowledgements could not be complete without thanking my
family and friends for their enduring support. They have offered me more care
and encouragement than they can ever know. l can never thank them enough,
and I offer them heartfelt gratitude! I love you!

This study was supported by grant # 90YF0010, Pathways Proiect:
Research intoﬁDirectjons for Family Health and Service Use, from the

Administration on Children, Youth, and Families, Department of Health and

vii

Human Services, Rachel F. Schiffman, PhD, RN, Principal Investigator,
Michigan State University. The cross-site national evaluation data were collected
under subcontract to Mathematica Policy Research, Inc, Princeton, New Jersey,
which was responsible for the national Early Head Start program evaluation
under contract 105-95-1936 with the Administration for Children and Families,
US. Department of Health and Human Services. Key MPR staff were John M.
Love (Project Director), Ellen Kisker (Principal Investigator), and Jeanne Brooks-
Gunn (Principal Investigator), under the supervision of the project manager for
ACYF, Helen H. Raikes. Community Action Agency in Jackson, Michigan, and
Applied Developmental Science Graduate Programs, Michigan State University,

also provided supplemental funding for this research.

viii

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................. xii
LIST OF FIGURES .......................................................................................... xiii
LIST OF APPENDICES ................................................................................... xv

CHAPTER 1: Examining the Effects of Coping on the Relationship between
Family Stress and Mother-Child Interactions across Time
INTRODUCTION

Models of Parenting Behaviors ................................................................... 1

Family Stress .............................................................................................. 7

Global Adversity ......................................................................................... 9
Parent-Related Adversity ...................................................................... 1O
Child-Related Adversity ......................................................................... 10
Family-Related Adversity ...................................................................... 11
Models of Risk and Adversity ................................................................ 13
Risk Threshold ...................................................................................... 15
Speciﬁc Family Stress ................................................................................ 17
Interpersonal Conﬂict .................................. 19
Parenting Stress .................................................................................... 22
Family Coping ............................................................................................. 24
Cognitive Reframing .............................................................................. 26
Support Seeking .................................................................................... 27
Conclusion and Critique ........................................................................ 29
CHAPTER 2: Study Aim and Hypotheses .................................................... 31

CHAPTER 3: Methods

Intervention and Data Collection ................................................................. 36
Sample: Dissertation Subsample ................................................................ 39
Sample characteristics .......................................................................... 4O
Instrumentation ........................................................................................... 42
Program Dosage ................................................................................... 42
Global Adversity .................................................................................... 44
Parenting Stress Index ......................................................................... 45
The Family Environment Scale ............................................................. 46
Family Crisis Oriented Personal Scale .................................................. 47

TABLE OF CONTENTS (Cont)

CHAPTER 3: Methods (Cont.)
Instrumentation (Cont)

The Nursing Child Assessment Satellite Training — Teaching Scale ..... 48
Data Analysis
Power analysis and sample size ........................................................... 49
Missing data estimation ......................................................................... 50
Outlier analysis ...................................................................................... 53
Analytic Methods ................................................................................... 53
CHAPTER 4: Results
Model Fitting: Base Model - Stress and Mother-Child Interaction ................... 56
Model adaptations ...................................................................................... 57
Model ﬁt ...................................................................................................... 57
Observed effects ........................................................................................ 61
Modeling the Direct Effects of Coping .............................................................. 61
Model adaptations ...................................................................................... 61
Model ﬁt ...................................................................................................... 63
Observed effects ........................................................................................ 63
Coping as a Moderator — Support Seeking ...................................................... 63
Model adaptations ...................................................................................... 65
Model ﬁt ...................................................................................................... 67
Observed effects ........................................................................................ 67
Coping as a Moderator — Cognitive Reframing ................................................ 69
Model adaptations ...................................................................................... 70
Model ﬁt ...................................................................................................... 71
Observed effects ........................................................................................ 71
Supporting Analyses ........................................................................................ 76
Family Adversity as a Predictor of Parenting Stress and Conﬂict ............... 76
Adversity as a Predictor of Mother-Child Interaction ................................... 77
Family Conﬂict and Mother-Child Interaction .............................................. 78
Program Dosage ........................................................................................ 80

CHAPTER 5: DISCUSSION
Family Adversity as a Predictor of Speciﬁc Stress and

Mother-Child Interaction ........................................................................ 82
The Effects of Speciﬁc Stresses on the Family ................................................ 86
Support Seeking - Direct and Moderating Effects ........................................... 91
Cognitive Reframing — Direct and Moderating Effects ..................................... 96
Program Impacts .............................................................................................. 101
Study Strengths and LimitatiOns ...................................................................... 104
Summary and Closing Comments ................................................................... 107

X

TABLE OF CONTENTS (Cont)

REFERENCES ................................................................................................ 111

xi

LIST OF TABLES

Table 1. Global Adversities and Percentages in Study Population ................. 44

Table 2. Univariate Descriptive Statistics:
Pre-Imputation for Missing Data ................................................................. 51

Table 3. Univariate Descriptive Statistics:
Post-Imputation for Missing Data ................................................................ 52

Table 4. Correlation Matrix for Variables Modeled .......................................... 58

xii

LIST OF FIGURES

Figure 1. Conceptual model 1 predicting mother-child interaction
in low-income families ................................................................................ 6

Figure 2. Conceptual model 2 testing stress and coping
using manifest variables to predict mother-child interaction ....................... 34

Figure 3. Path analysis ﬁndings testing base conceptual model
using manifest variables regarding stress and estimating
parameters with maximum likelihood approach .......................................... 59

Figure 4. Path analysis ﬁndings testing conceptual model
using manifest variables regarding stress and the direct effects
of coping and estimating parameters with
maximum likelihood approach .................................................................... 64

Figure 5. Path analysis ﬁndings testing conceptual model using
manifest variables regarding stress and the moderating effects
of support seeking and estimating parameters with
maximum likelihood approach .................................. 66

Figure 6: Plots for Nearly Signiﬁcant Interaction Effect -
Parenting Stress and Support Seeking predicting
Mother-2-year old Child Interaction ............................................................ 68

Figure 7: Plots for Nearly Signiﬁcant Interaction Effect —
Family Conﬂict and Support Seeking predicting
Mother-2-year old Child Interaction ............................................................ 69

Figure 8. Path analysis ﬁndings testing conceptual model using
manifest variables regarding stress and the moderating effects
of cognitive reframlng and estimating parameters with
maximum likelihood approach .................................................................... 72

Figure 9: Plots for Signiﬁcant Interaction Effect -
Parenting Stress and Cognitive Reframing predicting
Mother-Z-year old Child Interaction ............................................................ 73

Figure 10: Plots for Nearly Signiﬁcant Interaction Effect —

Family Conﬂict and Support Seeking predicting
Mother-Z-year old Child Interaction ............................................................ 74

xiii

LIST OF FIGURES (Cont’d)
Figure 11: Plots for Signiﬁcant Interaction Effect —

Family Conﬂict and Cognitive Reframing predicting
Parenting Stress while parenting a 2-year old child .................................... 75

xiv

LIST OF APPENDICES

Appendix A: Missing Data Analysis .................................................................. 131
Appendix B: Family Characteristics by Data Presence .................................... 134
Appendix C: Outlier Analysis ........................................................................... 135

Appendix D: F it lndices for Base Model —
Independently predicting ............................................................................ 138

Appendix E: Fit lndices for Base Model —
Relationships between Predictors .............................................................. 139

Appendix F: Fit lndices for Base Model —
Relationships between Predictors, Parsimonious Model ............................ 140

Appendix G: Fit lndices for Modeling Coping -

Independent Predictors .............................................................................. 141
Appendix H: Fit lndices for Modeling Coping —

Modeling Relationships Between Predictors .............................................. 142
Appendix I: Partial Correlation Coefﬁcients Support Seeking .......................... 143
Appendix J: Fit lndices for Modeling Moderating

Effects of Support Seeking ......................................................................... 144
Appendix K: Fit lndices for Modeling Moderating

Effects of Support Seeking - Parsimonious Model ..................................... 145
Appendix L: Fit lndices for Modeling Moderating

Effects of Cognitive Reframing ................................................................... 146
Appendix M: Fit lndices for Modeling Moderating

Effects of Cognitive Reframing — Parsimonious Model ............................... 147
Appendix N: Risk Thresholds ........................................................................... 148
Appendix 0: Examining Conﬂict across Time .................................................. 149
Appendix P: Exploring EHS Program Dosage ................................................. 155

XV

CHAPTER 1
EXAMINING THE EFFECTS OF COPING ON THE RELATIONSHIP BETWEEN
FAMILY STRESS AND MOTHER-CHILD INTERACTIONS ACROSS TIME

INTROD_UCTION

Past research has consistently demonstrated the inﬂuence of the mother-
child relationship on child developmental outcomes (Benzies, Harrison, & Magill-
Evans, 1998; Hadadian & Merbler, 1996; Harris, Brown, & Bifulco, 1986;
Maccoby & Martin, 1983; MagilI-Evans & Harrison, 1999; Rutter, 1995b; Wallace,
Roberts, & Lodder, 1998). The expression of the relationship between parents
and their children is observable through their interaction, during which each
member of the dyad has responsibilities to maintain (Barnard, 1997; Tronick,
1986; Sumner & Spietz, 1994; Zeanah, Larrieu, Heller, & Valliere, 2000). Within
the context of healthy mother-child interaction, mothers must interpret and be
responsive to a child’s cues, and provide opportunities for the interaction to be
reciprocal; concurrently, young children must provide clear cues and be
responsive to the parent with whom they are interacting (Barnard, 1997; Sumner
& Spietz, 1994; Tronick, 1986). In addition, interactions between mothers and
children entail contingent responses, in which a child’s behavior is dependent
upon a response from his mother and the inverse (Jameson, Glefand, Kulcsar, &
Teti, 1997; Sumner & Spietz, 1994; Tronick, 1986).

There are many circumstances for families that may disrupt the formation

and maintenance of healthy relationships, and thus mother—child interactions.
1

According to Belsky’s (1984) process model of parenting, the factors that are
crucial in determining the risk of problematic parenting are encompassed within
the subsystems of the parent, child, and environment. Belsky (1984) underlined
the psychological and personality resources of the parent as being paramount,
followed by the social-contextual circumstances of the family, and ﬁnally, the
characteristics of the child for predicting parenting behavior. On the basis of this
tenet, he argued that when all three domains of inﬂuence are adaptive, the
behaviors exhibited by parents are optimal, but with the compromised integrity of
each subsystem comes the likelihood of less positive interactions between
parents and children.

The model of parenting behavior as proposed by Belsky (1984) is a
cumulative-risk model and, as such, suggests a monotonic relationship between
risk and parenting (Phelps, Belsky, & Cmic, 1998). Belsky (1984) argued that as
the number of risk factors increases, so does the likelihood of poor parenting.
Other models of parenting, such as that posited by Abidin (1992) also suggest
that parenting behaviors are predicted by stress experiences, but also by
appraisals of the stressors, mechanisms for coping, and other resources
available to the parent, such as social support and parenting skills. Abidin’s
model for determining parenting behaviors, unlike the model proposed by Belsky
(1984), includes the parent’s motivational systems and perceptions of stress in
the parenting role. According to Abidin (1992) there are resources that support

optimal parenting, which include social support, parenting skills, the parenting
2

alliance and cognitive coping. Thus within the framework of Abidin’s model,
stress, perceptions of stress as related to one’s role as a parent, and coping
(cognitive coping and support) are all necessary elements to predict parenting
behavior.

Models of maladaptive parenting behaviors such as those developed to
describe the etiology of child maltreatment also lend credence to the necessity of
considering coping and appraisals of stress (Hillson & Kuiper, 1994; Milner,
1993). The model proposed by Milner (1993) focused on parental cognition as a
mediator of child physical maltreatment. Similarly, a model ascribed by Hillson
and Kuiper (1994) regarded parental perceptions of stress key to understanding
the etiology of maltreatment. Additionally, much like the tenets of Abidin (1992),
Hillson and Kuiper (1994) also proposed that understanding “appraisals of
resources, and dispositional coping strategies may help facilitate a more
complete explanation of the onset and maintenance of child neglect and abuse”
(pp.262)

A consistency across all models of parenting behavior reviewed is that stress
has been suggested to be one of the principal taxes to the psychological
resources of the parent. Indeed, research on stress has consistently been
shown to detrimentally impact on the quality of relationships among members of
families and mother—child interactions (Coyl, Roggman, & Newland, 2002; Lavee,
& Olson, 1991; McKelvey, Fitzgerald, Schiffman, & von Eye, 2002; Webster-

Stratton, 1990) and has been repeatedly ascribed to predicting child
3

maltreatment (Browne & Saqi, 1988; Holden & Banez, 1996; Holden, Willis, &
Foltz, 1989). Stress speciﬁcally related to adversity, ﬁnancial need, parenting,
and difﬁculties with adult relationships have been most consistently implicated as
negatively impacting the interactions between mothers and children (Conger,
McCarty, Yang, Lahey, & Kropp, 1984; Cmic, Greenberg, Ragozin, Robinson, &
Basham, 1983; Deater-Deckard & Scarr, 1996; Feldman, Leger & Walton-Allen,
1997; McKelvey et al. 2002; Peterson & Hawley, 1998; Rodriguez & Green,
1997; Schiffman, Omar & McKelvey, 2003; Weinraub & Wolf, 1983).

These models of parenting behavior also suggested that a thorough
examination of coping responses to stress may help researchers more fully
understand parenting behavior (Abidin, 1992; Hillson &'Kuiper, 1994; Milner,
1993). Coping is of particular interest in this study as the efﬁcacies of such
strategies have been posited to buffer the effects of stress on the quality of
parenting behavior and mother-child interactions. Furthermore, understanding
the ways in which stress and coping interact to predict parenting behaviors may
be critical for intervening with families. Many intervention programs have been
aimed at helping families develop effective strategies for coping (Reppucci,
Britner, & Woolard, 1997), which included the extension of support systems,
increased utilization of support, and enhanced problem solving skills of
individuals within families (Barth, Hacking, & Ash, 1988; Black, Nair, Kight,
Wachtel, Roby, & Schuler, 1994; Blair, Ramey, & Hardin, 1995; Campbell, 1994;

Cowen, 2001; Dawson, van Doornick, & Robinson, 1989; Gray & Ruttle, 1980;
4

Marcenko & Spence, 1994; Olds, Henderson, & Kitzman 1994; Olds, Henderson,
Tatelbaum, & Chamberlain, 1988; Reppucci, Britner, & Woolard, 1997; Seitz,
Rosenbaum, & Apfel, 1985).

Most interventions for families with young children, including Early Head
Start (EHS), aim to promote the optimal growth of infants and toddlers in many
domains, which include but are not limited to areas of cognitive and social-
emotional development (Raikes & Love, 2002). Therefore, many EHS programs’
target objective is to enhance the quality of mother-child relationships and
interactions. In particular, the Jackson, Michigan Community Action Agency’s
EHS program was developed based on an Infant Mental Heath (IMH) Model.
The IMH Model focuses on relationships between the mother and infant/toddler
as central to the establishment of healthy patterns of development for children
(Brophy-Herb, Schiffman, McKelvey, Cunningham-DeLuca, & Hawver, 2001;
Fitzgerald & Barton, 2000). Within this framework, relationships between other
members of the family and between the family and the intervention, in this case a
home visitor, are also viewed as signiﬁcant inﬂuences on positive child
developmental outcomes (Brophy-Herb et al. 2001; Fitzgerald & Iarton, 2000).
As a result, EHS and other prevention or early intervention programs could
greatly beneﬁt from gaining insight into the processes working within the family
that affect the quality of the parent-infant relationship (Coyl, Roggman, &
Newland, 2002; McKelvey et al. 2002).

 

Figure 1: Base Conceptual Model

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Family Stress: Family Stress: Family Parent-Child
Global Speciﬁc Coping Interaction
Number of
Adverse
Conditions
(Family,
Parent and Parent-
.' A
Child) / Child
Parenting ‘ Interaction
1 Stress
if
/ Cognitive
Reframing
Family
Conﬂict
SuppOrt
Seeking

 

 

 

 

 

The focus of the current study was to examine the relationship between
stress and mother-child interactions in low-income, Early Head Start eligible
families. The sources of stress that were examined in this study were twofold:
global adversity, which included chronic conditions that were less amenable to
change, such as teenage parenthood, premature birth of the target child, and
poverty; and speciﬁc family stress, which was quantiﬁed by reports of family
conﬂict and parenting stress. These family stressors, namely adversity,
parenting stress and family conﬂict, have been repeatedly found to negatively

impact parent-child interaction (Almeida, Wethington, & Chandler, 1999; Coyl,
6

 

Roggman, & Newland, 2002; McKelvey et al. 2002; O’Brien & Badahur, 1998).
Family support seeking and cognitive reframlng coping strategies were examined
as possible moderators of the relationship between family stress and mother-
child interaction (see Figure 1) as predicated by the models of Abidin (1992) and
Hillson & Kuiper (1994). The effects of these stress constructs and of coping on
the mother-child relationship were examined for mothers through observations of

their interactions with their infants.

FAMILY STRESS

There are many types of stressful events that may occur in the lives of
families. Stress can occur as a result of a change in one’s life circumstances
(Ostberg & Hagekull, 2000), as a result of the daily hassles (Almeida,
Wethington, & Chandler, 1999; Crnic & Acevedo, 1995), or as a result of
chronically stressful life circumstances or adversities (Lupien, King, Meaney, &
McEwen, 2001; Quittner, Gleuckauf, & Jackson, 1990). Family adversities, such
as those Identiﬁed as risks for less optimal child developmental outcomes, are
often conceptualized as family stressors, but the stress literature embraces two
general classiﬁcations of stress (Crnic & Acevedo, 1995). The ﬁrst type of family
stress pertains to that which accompanies new or daily demands and life events.
The second category of family stress pertains to chronic conditions that persist
over time (Quittner, Gluekauf, & Jackson, 1990; Rodgers, 1998; Weinraub &

Wolf, 1983). Researchers have used a variety of labels to delineate these two
7

types of stressors; acute versus chronic (McEwen, 1998), proximal versus distal
(McGrath & Beehr, 1990), event versus everyday (Crnic & Acevedo, 1995),
transitional versus enduring (Cicchetti & Rizley, 1981 ), and speciﬁc versus global
(Quittner, Glueckauf, & Jackson, 1990).

Some researchers argue that, due to their stable nature, chronic stressors
are more likely to have pervasive consequences (Quittner, Gluekauf, and
Jackson, 1990), while others argue that everyday stresses are the most salient
predictors of parenting outcomes (Crnic & Acevedo, 1995). The distinction
between these two types of stressors is necessary in that it has been argued that
chronic stress or adversity may place demands on a system, which bring about
more difﬁculty in coping with speciﬁc stresses or events (McEwen, 1998; Meyer,
Chrousos, & Gold, 2001) and that exposure to adversity predisposes individuals
to experiencing other negative life events (Patterson, 2002). It is important to
note, however, that not all researchers believe that stress experience ultimately
leads to the compromise of the individual, whether physically or in terms of one’s
mental health. Speciﬁcally, within the coping literature it has also been posited
that stress experiences and coping behaviors may strengthen the system by
making people more able to cope with subsequent insults or threats (Skinner,
Edge, Altman, & Shenrvood, 2003).

In an attempt to determine the relative contributions of past and present
stress experiences, a study of the temporal aspects of stress (Bar-Tal, Cohen-

Mansﬁeld, & Golander, 1998) found stress related to current daily hassles
8

directly and signiﬁcantly predicted current psychological distress. Past stress did
not directly impact current psychological distress, but rather did so indirectly
through its effect on the appraisal of future stressors (Bar-Tal, Cohen-Mansﬁeld,
& Golander, 1998). Given these two classiﬁcations of family stress within the
literature and their potential effects on the family, both were considered for the
current study (see Figure 1). Global adversity represented those sources of
stress that were more chronic in nature and less amenable to change; being a
teen at the time of a child’s birth, for example. Speciﬁc family stress, such as
that related to family conﬂict and to the parenting role, represented sources of
stress that were likely to change across time and which may be aggravated by a
person’s experience of global or chronic adversity (Lupien et al., 2001; McEwen,
1998).
GLOBAL ADVERSITY

There is a general consensus among social scientists that parenting
quality depends on a multitude of stress and support factors that exist both within
and outside the family (Abidin, 1992; Belsky, 1984, 1993; Green & Rodgers,
2001; Hillson & Kuiper, 1994; Milner, 1993; Patterson, 2002; Reiss, Hetherington,
Plomin, Sowe, Simmens, Henderson, & O’Connor, 1995; Woodward &
Fergusson, 2002). These stress and support factors act in combination to
determine risks of less optimal parenting. The child development literature is
replete with risk factors related to the likelihood of inadequate parenting and, at

the extreme, child maltreatment. These factors have been theorized as
9

belonging to three major groups of predictors, those pertaining to the parent, the
child, and the context within which those individuals exist (Abidin, 1992; Belsky,
1984; Hillson & Kuiper, 1994; Sandler, 2001; Woodward & Fergusson, 2002).
Parent-Related Adversig

There are multiple psychological characteristics of parents that have been
implicated as risks for less optimal parenting. Among these are being depressed
or having other mental health problems (de Chateau, 2000; Downey & Coyne,
1990; Fergusson, Lynskey, & Honrvood, 1993; Jameson, Gelfand, Kulcsar, &
Teti, 1997; McCurdy, 1995; Teti & Gelfand, 1991; Seifer & Dickstein, 2000),
abusing drugs and/or alcohol (Kelley, 1998; Lester, Boukydis, & Twomey, 2000;
Luthar, Cushing, Merikangas, & Rounsaville, 1998), and having a history of being
abused or neglected themselves (Egeland, Jacobvitz, & Papakola, 1987;
Kaufman & Zigler, 1987; Woodward & Fergusson, 2002). There are also
demographic characteristics of the parent such as being adolescent (Barratt &
Roach, 1995; Luster & Brophy-Herb, 2000; Sumner & Spietz, 1994) and having
minimal levels of education (Brayden, Altemeier, Tucker, Dietrich, & Vietz, 1992;
Cicchetti & Lynch, 1993; Woodward & Fergusson, 2002) that have been
repeatedly related to parenting.
Child-Related Adversiy

The child is not a passive recipient of interactions with parents and child-
level characteristics play a role in risk for less optimal mother-child interactions

as well. Some of the characteristics of children that can make interactions with
10

parents more difﬁcult are not being carried to full term pregnancy (Magill-Evans &
Harrison, 1999; Younger, Kendell, & Pickler, 1997), being low birth weight,
having medical complications or being born with a developmentally challenging
condition (Minde, 2000), and having a difﬁcult pattern of behaviors or
temperament (Barnard, 1997; Hansen, 1996; Jameson et al., 1997; Podolski &
Nigg, 2001; Rutter, 1990; Rutter, Dunn, Plomin, Simonoff, Pickles, Maughan,
Orrnel, Meyer, & Eaves, 1997; Woodward & Fergusson, 2002).

Family-Related Adversmr'

As well as characteristics of the individuals directly involved in the
interaction, the structural and psychological environment of the family plays a
role in determining parenting (Pinderhughes, Nix, Foster, & Jones, 2001; Rutter
& Quinton, 1977). Poverty is one of the most studied predictors of family
functioning and child developmental outcomes that has been repeatedly linked to
decreased levels of functioning for family systems (Conger, Conger, Elder,
Lorenz, Simons, & Whitbeck, 1992; Conger & Elder, 1994; Conger, Ge, &
Lorenz, 1994; Fisher, Fagot, & Leve, 1998; Garabino, 1976; Gomel, Tinsley,
Parke, & Clark, 1998; Ostberg & Hagekull, 2000; Peterson & Hawley, 1998;
Seccombe, 2002; Simons, Whitbeck, Conger, & Melby, 1990). For example,
acute economic difﬁculties, such as those resulting from job loss or economic
recession, have been linked to increased family conﬂict, both in the marital dyad

and between parents and children (Conger et al., 1994).

ll

In a recent longitudinal study of the predictors of child physical
punishment, Woodward and Fergusson (2002) stated, “...high and chronic levels
of contextual stress arising because of poverty, unemployment, parental
immaturity, inter-parental conﬂict and violence place strains on a parent’s care-
giving abilities, which in turn, increases their susceptibility to problematic
parenting practices...” (p. 216). In addition to the contextual stressors listed
above, single parenthood, high family density, lack of stability in residence, and
homelessness have also been demonstrated as risks for less optimal parenting
(Brooks-Gunn & Duncan, 1997;Gorzka, 1999; Liaw & Brooks-Gunn, 1994;
Rutter, 1995b; Rutter & Quinton, 1977; Webster-Stratton, 1990; Woodward &
Fergusson, 2002).

These systemic characteristics have been implicated in, but do not
necessitate, less optional parenting or parent-child relationships. Indeed, some
empirical studies have found these adversities and mother-child interaction
unrelated. Work related to the risk of teen parenthood has demonstrated mixed
results. For example, Roosa and colleagues found a suppression effect between
teen parenthood and socioeconomic status (SES), with SES playing a more
critical role in explaining mothering behaviors (Roosa, Fitzgerald, & Carlson,
1982). Another example, found in the work of Garcia Coll (2002) and Field and
colleagues (1990) demonstrated that Latina mothers who were teenaged
engaged in very nurturing parenting behavior with their infants. Field and

colleagues (1990) also found differences in mother-infant interactions based on
12

family constellation, with ethnically diverse teen mothers in the sample having
more positive interactions within the context of a nuclear family living
arrangement. These authors concluded that the support available within the
home environment may help to counter the risks of less optimal teen mother-
infant interactions.

Models of Risk and Adversig

Many models within many disciplines have been developed for predicting
the effects of exposure to risks. Within the medical literature, the effects of stress
have been well described. The Allostasis Model as proposed by McEwen (1998)
elucidates the accumulation of stress over time on the human system by
attempting to explain the additive detriment of repeatedstress exposure over
time. Within this theory of physiological functioning, allostatic load (McEwen,
1998) refers to the tension placed on the body that results with repeated
exposure to stress. The model suggests that repeated exposure to stress exacts
changes in the system and makes each additional stress exposure more difﬁcult
for the body to manage. McEwen argues that the additive insults of stress can
inevitably lead to less optimal functioning.

The Allostatis Model is not unlike psychological models of stress and
adversity. Researchers (McCubbin & Patterson, 1983; Patterson, 2002) have
described models of cumulative risk as “pile-up” and propose that exposure to
risk can often cascade, with one risk leading to another. Cumulative risk models

for predicting outcomes are vulnerability models which suggest that there is a
13

monotonic relationship between risks and negative outcomes (Phelps, Belsky, &
Crnic, 1998). Such models fundamentally suggest that as the number of risks to
which one is exposed increases so does the likelihood of a negative outcome
(Phelps, Belsky, & Crnic, 1998). Furthermore, a cumulative risk model of
development posits that a combination of risk factors, rather than any single
factor, can better predict adverse developmental outcomes (Liaw & Brooks-
Gunn, 1994).

Researchers have demonstrated that the number of risk indicators or
adversities, rather than the speciﬁc types, is essential for predicting child
developmental outcomes (Biederrnan, Milberger, Faraone, Kiely, Guite, Mick
Ablon, Warburton, & Reed, 1995; Fitzgerald, Puttler, Mun & Zucker, 2000; Kolvin,
Miller, Fleeting, & Kolvin, 1988a, 1988b; Liaw & Brooks-Gunn, 1994; Rutter,
1979, 2000; Sandler, 2001 ). Sandler (2001) stated that “the effects of exposure
to any single adversity need to be interpreted in the context of the cumulative
effect of exposure to multiple adversities, in that several studies demonstrate
particularly strong effects of exposure to multiple adversities and that the
cumulative adversities account for much of the reported effect of any single
adversity” (p. 22). Some examples of this research can be found in the work of
Sameroff and colleagues (Sameroff, Bartko, Baldwin, Baldwin, & Seifer, 1998;
Sameroff & Seifer, 1983; 1995; Sameroff, Seifer, & Bartko, 1997) who
demonstrated that risk load scores, as opposed to individual indicators, were

better predictors of child development outcomes, like IQ and mental health.
14

gig Thretsholg

Within studies that have examined the cumulative effects of stress on
development, some researchers have demonstrated effects of a risk threshold.
For example, the work of Rutter and colleagues (Rutter, 1979; 1981; 1985; 1987;
1990; 1995b; 1996; Rutter, Champion, Quinton, Maughan, & Pickles, 1995;
Rutter & Quinton, 1977; 1984) demonstrated that children exposed to four or
more risks were likely to have poorer child development outcomes. Rutter (1979)
examined the likelihood of child psychiatric diagnosis based on the contribution
of six risk factors; (1) SES as measured by limited ﬁnancial resources, education,
and employment, (2) family density, (3) marital distress, (4) maternal depression,
(5) paternal antisociality, and (6) removal of child from the family. His work
demonstrated that the risk of diagnosing a psychiatric disorder in children rose
tenfold for children in families with four or more risk factors, as opposed to those
children with one or no factors.

Although many studies have demonstrated a risk threshold,
inconsistencies in the number of risks studied are common. For example, in a
study of family stress and parenting, Peterson and Hawley (1998) found that a
threshold of three stressors; such as those related to ﬁnances, meeting basic
family needs, and social support, differentiated between impacts on parenting
behavior. Families with fewer than three stressors reported signiﬁcantly healthier
parenting attitudes, more family cohesion and less family conﬂict. In addition,

families with no reported stressors scored signiﬁcantly better than higher stress
15

groups on measures of empathic parenting and negative attitudes toward
physical punishment. Although risk thresholds will not be utilized in the current
study, it is important to note that past studies have demonstrated impacts on
child developmental outcomes predicted by greater numbers of stressors to
which the family is exposed.

Cumulative risk models have been explained in relation to adversity, but
there are other models to explain the impacts of risk or stress on family systems.
Developmental systems models (Ford & Lerner, 1992), for example, allow for the
dynamic interplay of demands upon the system and capabilities or strengths
(Patterson, 2002). Within the dynamic systems model, one can describe the
potential for continuity and discontinuity across development (Fitzgerald, Zucker,
& Yang, 1995; Ford & Lerner, 1992). Developmental systems models do not
assume an additive effect of stress on individuals and family systems. The
models ascribed by Abidin (1992) and Hillson and Kuiper (1994) fall within
developmental systems models in which a transaction between stress, appraisals
and coping are assumed to describe behavior. Within the present study both
types of models are considered; global adversities are summed and assumed to
have equal value in predicting mother-child interaction. Although it is accepted
that some of the adversities that are contained within the index are malleable
(living below 100% of poverty for example), they are treated as constant from

reports at enrollment for the sake of modeling. Parenting stress, family conﬂict,

l6

coping behaviors and mother-child interaction are expected to change across

time.

SPECIFIC FAMILY STRESS

Acute family stress and global adversity have been demonstrated to have
negative effects on humans, both biologically and interpersonally (Dorn, Bergess,
Susman, von Eye, De Bellis, Gold & Chrousos, 1996; Fisher et al., 1998;
McEwen, 1998; Meyer, Chrousos, & Gold, 2001; Rudolph, Hammen, Burge,
Lindberg, Herzberg, & Daley, 2000). Overall, family stress speciﬁc to the role of
parenting and family conﬂict has been found associated with parenting behaviors
and family interactions that are less supportive and morepunitive in nature (Crnic
et al. 1983; Coyl, Roggman, & Newland, 2002; Deater-Deckard & Scarr, 1996;
Erel & Burman, 1995; Feldman, Leger, & Walton-Allen, 1997; Longfellow,
Zelkowitz, & Saunders, 1982; McKelvey et al. 2002; Peterson & Hawley, 1998;
Rodriguez & Green, 1997; Weinraub & Wolf, 1983). Parenting stress and family
conﬂict have been shown to negatively effect mother-child interactions in both
single- and two-parent households (Deater-Deckard & Scarr, 1996; Weinraub &
Wolf, 1983), in families with lesser as well as greater ﬁnancial resources (Conger
et al., 1984; Coyl, Roggman, & Newland, 2002; Dix, 1991; McKelvey et al, 2002;
Rodriguez & Murphy, 1997), and regardless of the age of the child (Buehler,
Anthony, Krishnakumar, Stone, Gerard, & Pemberton, 1997; Buehler & Gerard,

2002)
I7

The cross-cutting effects of stress have been demonstrated empirically.
For example, with a sample of older, middle income, well-educated single and
married mothers, Weinraub and Wolf (1983) demonstrated that mother-child
interactions were positively enhanced by the availability of support and worsened
by increases in stress associated with life events. Differences between single
mothers and married mothers in the likelihood of experiencing life changes and in
the amounts of support were also found; single as opposed to married mothers
encountered a greater number of life changes (9.4 versus 6.6 per year) and also
reported receiving less emotional support, speciﬁcally related to the role of
mothering.

Within low-income families, stresses related to difﬁculties with adult
relationships and to managing limited ﬁnancial resources have been found to be
related to parenting behaviors. Utilizing a sample of mothers eligible for Early
Head Start, Coyl and colleagues (2002) demonstrated a relation between stress
and increases in occasions of physical punishment (Coyl, Roggman, 8 Newland,
2002). Another study with this sample of low-income, EHS eligible mothers
found links between family stress, as related to family conﬂict, satisfaction with
romantic relationship, and perceived ﬁnancial need, and less positive mother-
infant interaction (McKelvey et al. 2002). This study examined the mothers’
sensitivity to infant cues, responsiveness to infant distress, and in fostering of
social-emotional development within the context of mother-child interactions

(McKelvey et al., 2002).

The speciﬁc stressors that were examined in the current study were
disharmony or conﬂict between members of the family and stress that was
related to parenting itself, such as the parent’s experience of distress in the role
and concerns for the development and expectations of the child. These stress
sources were conceptualized as more malleable and have been theorized and
demonstrated as having direct consequences for predicting parenting behaviors
and mother-child interaction (Abidin, 1990, 1992; Almeida, Wethington, &
Chandler, 1999; Hillson & Kuiper, 1994; O’Brien & Badahur, 1998). Stress
related to family socioeconomic variables were included in the global adversity
construct (see Table 1; page 44).
lnterptersonalginﬂict

Dynamic systems and ecological models suggest that anger generated by
interparental conﬂict may spread to the mother-child relationship and thus
generate overarching feelings of anger and tension within the family (Almeida,
Wethington, & Chandler, 1999; Buehler & Gerard, 2002; Larson & Almeida,
1999; O’Brien & Badahur, 1998). In support of that tenet, stress and conﬂict in
the marital (or romantic) relationship has been found consistently associated with
less favorable interactions between family members (Almeida, Wethington, &
Chandler, 1999; Belsky, Youngblade, Rovine, & Volling, 1991; Buehler et al.,
1997; Erel & Burman, 1995; O’Brien & Badahur, 1998). Parental conﬂict and
hostility may create an aversive home environment that is less optimal for a

developing child (Buehler et al., 1997; Cummings & Davies, 1994; Erel &
19

Burman, 1995; Grych & Fincham, 1994; Maccoby & Martin, 1983). For example,
family conﬂict has been found related to increases in harsh, punitive and
inconsistent parenting behaviors (Jekeilek, 1998; Jouriles, Pﬁffner, & O’Leary,
1988; Webster-Stratton, 1990). Men and women who are dissatisﬁed with their
adult romantic relationships are likely to report increases in stress associated
with parenting than those who are happily situated (Webster-Stratton, 1990). In
fact, there seems to be relatively consistent correlations between marital and
family conﬂict and parenting stress (Crnic & Acevedo, 1995; Deater-Deckard &
Scarr, 1996; Gelfand, Teti, & Radin Fox, 1992).

Studies of families with lesser (Coyl, Roggman, & Newland, 2002; McKelvey
et al. 2002) and with greater ﬁnancial capital (Fisher, Fagot, & Leve, 1998;
Garabino, 1976; Gomel et al. 1998; Ostberg & Hagekull, 2000; Peterson &
Hawley, 1998; Simons, Whitbeck, Conger, & Melby, 1990) have both found
complications in the adult romantic relationship may result in less optimal mother-
child interaction. Buehler and Gerard (2002) demonstrated that in families where
inter-parental conﬂict was prevalent, speciﬁcally disagreement and verbal
aggression, “parents also tend to spank, slap, or yell at the children, as well as
argue more frequently with their child” (p. 88). They found that parents in
families marked by inter—parental conﬂict were less likely to praise, read to, play
with, and spend time engaged in social activities with their children. Interestingly,

the study sample consisted of families with children of varying ages (ranging from

20

2 to 11) and the authors concluded that children seem to be equally vulnerable to
the effects of conﬂict regardless of age.

Physical conﬂict within the family has also been shown to impact the quality of
parenting. Researchers have found that women who were involved in
aggressive conﬂict in the spousal relationship were more likely to be aggressive
when interacting with their children (Almeida, Wethington, & Chandler, 1999), to
use power assertion discipline techniques, display anger and were less likely to
show warmth when interacting with their children (O’Brien & Bahadur, 1998).
Other studies have shown that mothers’ parenting behaviors became less
consistent, more restrictive and more punitive when in families with physical
conﬂict (Dix, 1991; Jouriles, Pﬁffner, & O’Leary, 1987; Ritchie & Holden, 1998).

Family conﬂict may alter the ways in which mothers interact with their
children. What seems to be a consistent ﬁnding is family conﬂict also may
increase a mother’s stress related to the role of parenting. It is important to note
that verbal aggression and physical aggression are different processes that may
yield alternative outcomes. Spousal abuse and domestic violence may tax the
family in different ways than is the case with verbal aggression. The conﬂict
measure within the current study was developed to measure conﬂict in the entire
family and not just within the adult romantic relationship. Therefore, the measure
of conﬂict that was included in the present study was more broadly deﬁned and
may represent conﬂict between mothers and children as much as between

mothers and domestic partners.
21

Parenting Stress

Parenting stress is a complex construct that involves behavioral, cognitive,
and affective components. According to Abidin (1992) “Parenting stress is the
result of a series of appraisals made by each parent in the context of his or her
level of commitment to the parenting role. Parenting stress is viewed as a
motivational variable which energizes and encourages parents to utilize the
resources available them to support their parenting” (p.410). Abidin (1992)
proposed that the total stress experienced by a parent is a combination of child
and parent characteristics, and family situational components as they relate to
the person’s appraisal of his or her role as a parent. He also proposed a
relationship between high parenting stress and dysfunctional parenting (1992).
Within the context of a dynamic family systems model, Crnic and Acevedo (1995)
posited that the parenting stress may work to alter the processes that operate
within the family. They suggested that stress may hinder the development of
positive and supportive relationships within the family and generate greater
tension among family members.

Effects of parenting stress on interactions within family members have been
demonstrated. Parenting stress has been implicated in less positive interactions
between parents and children (Coyl, Roggman, & Newland, 2002; Nitz,
Ketterlinus, & Brandt, 1995), in increasing the likelihood of child maltreatment
and abuse (Holden & Banez, 1996; Holden, Wills, & Foltz, 1989; Mash &

Johnston, 1990; Rodriguez & Green, 1997; Whipple & Webster-Stratton, 1991)
22

and in increasing conﬂict between family members (Almeida, Wethington, &
Chandler, 1999; Buehler & Gerard, 2002; Crnic & Acevedo, 1995; Deater-
Deckard & Scarr, 1996; Gelfand, Teti, & Radin Fox, 1992; Larson & Almeida,
1999; O’Brien & Badahur, 1998; Webster-Stratton, 1990). With a sample of
mothers and children in Head Start, Rodgers (1998) demonstrated effects of
parenting stress on parenting behaviors. In this study, parenting stress had
direct effects on punishment, parental coldness, sensitivity, inconsistency, and
rejection. Other research has demonstrated the relationship as well. Ritchie and
Holden (1988) found that as levels of parenting stress increase not only do
negative interactions increase, but positive qualities of interactions such as
physical affection decreased.

Based on ﬁndings from their study of sources of parenting stress, Gelfand,
Teti, and Radin Fox (1992) stated, “highly stressed parents may lose their ability
to care for their children in a warm, sensitive, and competent manner” (p. 262). It
appears that the effects of global and speciﬁc stress may undermine a family’s
functioning. An ability to cope with stressful events and circumstances could
potentially moderate (or buffer) the effects of such on family interactions. In fact,
there are several family interventions that have sought to alter family coping
strategies through the extension of support systems and problem solving skills

(Cowen, 2001; Olds et al., 1994; Repucci, Britner, & Woolard, 1997).

23

FAMILY COPING

“The study of coping is fundamental to an understanding of how stress
affects people, for better or for worse” (Skinner at al., 2003, p.216). Coping has
been deﬁned as any strategy intended to aid in the management of stressful
events or circumstances (Judge, 1998; Lazarus & Folkman, 1984; Pearlin &
Schooler, 1978). There are generally two forms of coping strategies discussed in
the relevant literature: emotion-focused and problem-focused (Lazarus &
Folkman, 1984). Emotion-focused coping is directed toward regulation of the
emotional components of response to stress (Judge, 1998; Lazarus & Folkman,
1984) and problem-focused coping is directed toward managing or altering stress
conditions (Judge, 1998; Lazarus & Folkman, 1984).

Lazarus and Folkman (1984) argue that emotion-focused coping will be
more likely when the stress condition is appraised to be resistant to modiﬁcation
and problem focused strategies more likely in situations that are appraised as
amenable to change. Emotion-focused strategies for alleviating stress range
from making positive comparisons and controlling one’s feelings regarding the
stressor, to detaching from the problem (Hillson & Kuiper, 1994). The second
category of strategies, problem-focused coping, is thought to make the distress
causing situation less threatening, which can be achieved either through altering
one’s environment or oneself. Problem-focused strategies include acquiring

social support, changing external pressures, and changing oneself through

24

learning new behaviors and acquiring new skills (Hillson & Kuiper, 1994; Judge,
1 998).

In addition to these categorizations for strategies of coping, Billings and
Moos (1982) conceptualized coping behaviors in terms of active versus avoidant
strategies. Within this framework, avoidant strategies are used to remove the
person from the source of stress, whereas active coping strategies are thought to
cognitively or behaviorally impact the stressor. Using the Billings and Moos
(1982) deﬁnitions of coping, emotion-focused strategies, as deﬁned by Lazarus
and Folkman (1984), assume avoidant characteristics, whereas problem-focused
strategies would be assumed to be active behaviors.

Family coping strategies can potentially strengthen or maintain family
resources that serve to protect the family from stressful situations (McCubbin,
Joy, Cauble, Comeau, Patterson, & Needles, 1980; Skinner et al., 2003).
According to models of parenting as ascribed by Abidin (1992) and Hillson &
Kuiper (1994), coping is proposed to effect the relationship between stress and
parenting behavior. Wills, Blechman and McNamara (1996) suggested that
effective coping strategies for family systems remove the source of stress or
impact perceptions of coping ability. They further suggested that ineffective
strategies may alienate sources of support, lead to negative perceptions of
oneself and family, and fail to alleviate the source of stress itself. The family
coping strategies to be examined in the current study are cognitive reframing and

support seeking behaviors.
25

Many studies of coping have examined coping as a mediator of the
relationship between stress and an outcome, while others have examined coping
as a conceptual buffer or moderator of the same relationship (Morano, 2003).
Conceptually, the mediator describes a mechanism through which a relationship
between two variables is achieved, while a moderator is a variable that predicts
variation in the magnitude of a relationship (Baron & Kenny, 1986; Holmbeck;
1997; Kramer, Stice, Kazdin, Offord, & Kupfer, 2001; Kraemer, Wilson, Fairburn,
& Agras 2002; Morano, 2003). Methodologically and conceptually, there is much
debate as to the distinction between moderation and mediation and some have
proposed applying temporal deﬁnitions to the processes in effort to solve
potential ambiguity (Kraemer et al, 2001; Kraemer et al, 2002). It is also
important to note that most studies related to the concept of coping do so in
regards to illness or injury and few study stress and coping process in families
with typically developing children.

Family Coping: Cognitive Refrgming

Jarvis and Creasey (1991) examined the inﬂuence of ways of coping on
attachment between mothers and infants. They found that in coping with
parenting stress that cognitive positive reappraisal (which reﬂects a person’s
ability to think of stress as having some positive outcome) rather than avoidant
coping strategies, was associated with a reduction in stress and with securely
attached infant-parent relationships. The authors also found that cognitive

positive reappraisal mediated the relationships between parenting stress and
26

parental reports of attachment with their 18-month-old children. Another study of
family stress did not support the mediating effect of coping (McKelvey et al.,
2002). An interesting result from the latter study of low-income, Early Head Start
eligible families was that increases in family stress (as measured by family
conﬂict, marital dissatisfaction, and ﬁnancial need) were associated with
decreases in amount of coping. The authors concluded that the relationship
between stress and parent-infant interactions may be moderated rather than
mediated by coping (namely cognitive reframing and seeking support from family
and friends) as the magnitude of this relationship decreased with the inclusion of
coping in the model.
Family Coping: Support Seeking

Families cope with stress in many ways. Cognitively reframlng or
reappraising the stressor has been shown to impact parenting, as has social
support. A buffering (moderating) hypothesis has been supported in studies of
stress, social support, and parenting behaviors. Rodgers (1998) found that
participants’ perceptions of usefulness of social support discriminated the
magnitude of the relationship between parenting stress and parenting behaviors.
There was a signiﬁcant difference between the path coefﬁcients for parenting
stress and parenting behaviors in families dichotomized as having highly versus
lowly useful social support. Other studies have shown that social support
moderated the effects of parenting stress on discipline behaviors, parental

warmth, and sensitivity (Crnic et al. 1983; Deater-Deckard & Scarr, 1996; Kotch,
27

Browne, Ringwalt, Dufort, Ruina, Stewart, & Jung, 1997; Rodgers, 1998) and
mother-child interaction (Crnic & Greenberg, 1990; Weinraub & Wolf, 1983). Not
only has emotional support from friends been found to be important for mothers’
coping with parenting stress (Crnic & Acevedo, 1995), but also to be negatively
related to stress in the parenting role itself (Deater—Deckard & Scarr, 1996;
Melson, Windecker, Nelson, & Scwarz, 1998).

The moderating role of social support has not been consistently supported in
the literature, however. Quittner, Glueckauf and Jackson (1990) failed to ﬁnd
moderating effects of support. They concluded that social support functions
differently in chronic versus speciﬁc stress conditions. They suggested that
increases in support may be helpful for short-term stressors, while a sudden
infusion of support in the context of chronic conditions may be appraised as
“intrusive or suggestive of incompetence” or “critical and unhelpful” (p.1276). In a
similar light, Pearlin and Schooler (1978) found that self-reliance as a coping
strategy was more efﬁcacious than support from others in reducing the stress
associated with parenting and marriage. These authors also found that a
reduction in the stress associated with interpersonal relationships (marriage and
parenthood) was more likely when active coping strategies, as opposed to those
that were avoidant in nature, were utilized.

There are many studies that have examined the role of social support on
parenting behaviors. Those reviewed examined the presence versus absence or

usefulness of support, rather than the active seeking of support. It is important to
28

note that support in the current study includes the active seeking of support only,
and does not reﬂect the quality or amount of support as discussed above.
Support seeking behaviors and cognitive reframlng are the two particular coping

strategies to be examined in the current study.

Conclusion and Critigue

In sum, stress, whether the result of a global or speciﬁc source, has been
demonstrated to effect relationships among members of families (Coyl,
Roggman, & Newland, 2002; Liaw & Brooks-Gunn, 1996; Loukas,Twitchel,
Piejak, Fitzgerald, & Zucker, 1998), and mother-child interaction (McKelvey et al.
2002). Although the results of studies are mixed, many suggested the
relationship between stress and interactions between parents and children are
impacted by coping (Jarvis & Creasey, 1991; McKelvey et al. 2002). The
purpose of the current study was to determine if the relationship between family
stress, both global and speciﬁc, and mother-child interaction differed based on
the stress source. The study also sought to determine if ways of coping with
stress, either through cognitively reframlng the situation or through seeking
support, impacted the effects of stress on mother-child interactions.

What is less understood about the family processes described is in what way
patterns of stress and family interactions may change across time. Most of the
existing literature is cross sectional in nature, and as such, provides a limited

picture of how coping with stressors may impact relationships within the family.
29

The current study adds to the existing literature a more thorough examination of
these family processes; as assessed when parenting a two-year old child and
again when the child was three years old.

An examination of strategies for coping across time and in families with young
children is a useful addition to the existing literature as well. Researchers who
study coping have found that some coping strategies in certain situations may
not only fail to alleviate stress, but may indeed generate greater stress. For
example, Quittner, Glueckauf and Jackson (1990) posited that support from
others functions differently in chronic versus specific stress conditions with
increases in support considered helpful for short-term stressors, while in the
context of chronic conditions appraised as “intrusive or suggestive of
incompetence...critical and unhelpful” (p.1276). Understanding the ways in
which coping strategies may act to alleviate or exacerbate family stress over time
is a useful addition to our existing knowledge; especially when parenting an
infant and toddler. Little is known about how strategies may impact families with
young children and even less is known about how effective strategies may be
while parenting a two year old versus parenting a three year old.

Another contribution that this study yielded to the literature was found within
the nature of the sample. Many of the studies reviewed, despite attempts to ﬁnd
studies with a more similar sample, utilized middle- to high-income families. The
sample for the current study came from a low-income, Early Head Start eligible

population. Stress, coping, and interactions between mothers and toddlers in
30

this population have been less often studied. The families within this sample
were randomly assigned to EHS programming, which also permits the study of
these family processes within the context of early intervention and the greater
community. The comparison of these processes across early intervention is not

common in the existing studies reviewed.

CHAPTER 2
STUDY AIM and HYPOTHESES

The aim of the current study was to examine the moderating role of coping
on the relationship between stress and mother-infant interactions in low-income,
high-risk families. The sources of stress that were examined in this study were
twofold: global adversity and speciﬁc family stress. Included in the measure of
global adversity were conditions that were chronic and difﬁcult to change.
Examples of these global adversities include poverty, teenage parenthood, and
premature birth of the target child. Speciﬁc family stress was quantiﬁed by family
conﬂict and parenting stress as these stressors had been repeatedly found to
impact mother-child interaction (Almeida, Wethington, & Chandler, 1999; Coyl,
Roggman, & Newland, 2002; McKelvey et al. 2002; O’Brien & Badahur, 1998).
The coping constructs that were examined in the present study include support
seeking behaviors and cognitive reframlng. The effects of these stress
constructs and of coping on the mother-child relationship were examined for

mothers through observations of their interactions with their infants.
31

The current study sought to understand the relationships between stress,
coping, and mother-child interactions when the child in the study was age two
and age three. Changes in relationships among target variables may differ
across time not only because children and their families had developed, but
many of the families in the study would had been participating in a home
visitation program for an additional year. Participation in home visiting programs,
in general, has been demonstrated to change the nature of mother-child
interactions (Black, Nair, Kight, Wachtel, Roby, & Schuler, 1994; Duggan,
McFarland, Windham, Rohde, Salkever, Fuddy, Rosenberg, Buchbinder, & Sia,
1999). Indeed, one of the primary goals of the CAA’s EHS program was to
promote positive relationships between parents and their children. One of the
speciﬁc aims of the current study was to determine whether EHS program
participation (dosage) impacts mother-child interaction.

Figure 2 depicts the manifest variable model hypothesized in the current
study. Family stress, in the form of global adversity, parenting stress, and family
conﬂict, was expected to predict speciﬁc family stresses and mother-child
interaction. Family coping strategies were expected to moderate the
relationships between family stress, both global adversity and speciﬁc, and
mother-child interaction. The hypothesized relationships within this model

include:

32

Direct Relationships:

Baseline global adversity was expected to predict measures of speciﬁc family
stress (namely parenting stress and family conﬂict) and mother-child
interaction collected at or near the child’s 2nd and 3rd birthdays. It was
hypothesized that greater amounts of adversity would predict higher reports
of family conﬂict and parenting stress and less optimal mother-child
interaction.

Parenting stress measured at or near the child’s 2nd and 3rd birthdays was
predicted to impact mother-child interaction observed at or near the child’s
2nd and 3rd birthdays, respectively, with more stress in the parenting role
predicting less positive mother-child interaction.

Family conﬂict collected at or near the child’s 2nd and 3rd birthdays was
hypothesized to predict mother-child interaction observed at or near the
child’s 2nd and 3rd birthdays, respectively. It was expected that mother-child
interactions in families in which there was a great deal of family conﬂict would
be less positive.

Parenting stress and family conﬂict (as measures of speciﬁc family stress)

were expected to correlate with one another at both time points.

33

 

$585. 6m @

 

 

 

 

 

 

 

 

 

 

 

 

6 :6
.\.\..\\1.. A ﬁssﬁwmmg 5586 SEE
35:22 mm ©v .\. ..\\ 6>_=:moO
5:66:62. . a/ +
2EO¢6£22 65:22 on OS
666:5 m:=:6:6d

 

 

 

 

 

 

     

658.2 6m @ +
.1. m:_v_66w toaazm +

 

 

 

 

 

  

 

 

 

 

 

 

 

 

+ 358.2 em as +
95.6.61
/ 6>=_:moo
.2 a.
+ / + 2.: x. + 358.2 em OS
65.5.2 em @ . . _._ , \ 8:86 22:26“. 4

 

 

m:_x66w toaasw .2.

—’/. .\

 

 

 

 

« ,. ,. . 2. .. 35:22 em @
r / . + 666:5. m:=:6:6n_

 

 

 

 

 

 

35:22 «N @ ., :.
3:66:25 . ..

 

 

 

 

 

 

 

 

26.6%: Z
ii/
+ mommoo 56:92.”. £9656. .6620

656: 336.669: ......... 1
Boos. 9.300 cc: 6625 _6>6._ «662:6: "N 6.59“.

 

 

 

 

 

 

 

 

 

Moderating Relationships (depicted by segmented lines):

Coping strategies (support seeking behaviors and cognitive reframlng) would
moderate the family stress, global and speciﬁc, to mother-child interaction
relationships.

These two measures of coping were expected to correlate with one another at

both assessments as they represent overall coping strategies.

Autoregressive Relationships:

Mother-child interaction observed at 24 months of age would predict the same
at 36 months.

Parenting stress collected at or near 24 months of age was expected to
predict parenting stress at 36 months of age.

Family conﬂict collected at or near 24 months of age would predict family
conﬂict at 36 months of age.

Support seeking measured at or near 24 months of age was expected to
predict the same at 36 months of age.

Cognitive reframing assessed at or near 24 months of age would predict the

same at 36 months of age.

Program Impacts:

Program dosage was hypothesized to directly affect mother-child interaction
assessed at or near the child’s 2nd and 3rd birthdays. It was the goal of the
Jackson EHS program to improve parent-infant interaction from the program’s

35

conception and had been demonstrated in previous studies with this sample
based upon the intent to treat model (Van Egeren, McKelvey, Schiffman, &
Fitzgerald, 2002). This research demonstrated differences in interactions

across time using random assignment and not dosage of programming.

CHAPTER 3
METHODS
Intervention and Data Collection
The current study was a secondary analysis of an existing data source.
The original study, entitled “Pathways Project: Research into Directions for
Family Health and Service Use” (Schiffman, 1996) was a longitudinal study of
children eligible for Early Head Start. It is important to note that the data
collected from the families in the current study were done in conjunction with a
national evaluation of Early Head Start as conducted by Mathematica Policy
Research, Inc. (Raikes & Love, 2002). The eligibility requirements for the
national Early Head Start research and evaluation project included: (a) the child
had to have been born between 9/1/95 and 9/30/98 and had to have been
younger than one-year-of age, (b) the family could not have had participated for 3
months or longer in any comprehensive child development program during the 5
years immediately preceding the birth of the EHS child, (0) the family could not
have had participated for 3 months or longer in a parent child center, Head Start

or similar program within the last year, and (d) the family had to have met the
36

federal deﬁnition of families that fall below the 100 percent poverty line (10% of
families could fall above income if focus child was identiﬁed as having a
developmental delay). Families who met these eligibility requirements were
recruited through a local health care and social service agency. The eligible
families were asked to participate in the research by staff of a community health
center. Families were recruited during a visit for prenatal care, postpartum
check, or for a health-related visit for the child. A total of 196 families were
enrolled in the study, nearly one-third of the women were pregnant at enrollment.
Early Head Start program staff at the Community Action Agency (CAA) in
Jackson, Michigan collected demographic and needs data during application to
the Early Head Start program. Families were then randomly assigned into the
EHS program or into a comparison group at the national level by Mathematica
Policy Research, Inc. (Raikes & Love, 2002). Of the 196 families enrolled in the
research and evaluation study, 189 were retained in the sample after random
assignment. Those families who were not retained in the study were withdrawn
due to the death of the focus child, a loss of custody, or at the family’s request to
not be contacted. There were no differences between the groups at random
assignment at the national level (Raikes & Love, 2002; US. Department of
Health and Human Services, 2002). Additionally, there were no differences in
the full local sample of which the sample for the current study was a subset with

regards to family demographic characteristics and family related measured

37

collected at or near enrollment into the research and evaluation study (Schiffman
& McKelvey, 2002).

Data collection for mothers and infants participating in the study began
with an interview and observation at or near enrollment (Infant age: M = 5
months, SD = 3.8 months) and continued at multiple intervals through the focus
child’s third birthday. Upon application to Early Head Start programming and 6,
17, and 28 months after enrollment into the study, participants completed a
demographic questionnaire, which included measures of annual income, age,
ethnicity, marital status, and educational level. Shortly after enrollment and at or
near the child’s 14-, 24-, and 36-month birthdays, study participants also
completed a set of instruments upon home visitation by research project
personnel.

Data collected from participating families at or near enrollment, the child’s
24th month (M = 25.3 months, so = 1.5 months) and 36th month (M = 37.6
months, SD = 1.9 months) were utilized for the current study. Collection of data
for the current study occurred within the home primarily through interview
techniques, although there were also observations of mother-child interaction by
trained data collectors. Parent interview data, collected from mothers, focused
on relationship quality, perceived ﬁnancial needs and resources, family coping
techniques, and family demographics. The source of data for mother-child

interaction was an observed teaching episode (Sumner & Speitz, 1994).

38

Program dosage was attained during a thorough review of home visitor
documentation in the EHS program records. Employees of the “Pathways
Project” were trained to extract critical information from the families’ EHS
program charts. Among the information attained from program records were the
length of time of each individual visit, the number of persons present, the
location, the focus of the visit, and the engagement of the participants during the
visit. In addition to information about individual visits, information regarding the
referrals made on behalf of the family and regarding other services offered by the
program and agency was obtained.

m

The sample consisted of 166 of the total 189 mothers in the “Pathways
Project” study. Data from the mothers collected at the 24- or 36-month birthday
related interview were included in the current study. The sample consisted of
those mothers who were contacted and from whom data was collected at the
child’s 24-month birthday related assessment, 36-month birthday related
assessment or both. The remaining 23 mothers in the sample who were
recruited for the study and who failed to complete these later assessments were
found to be demographically different from the 166 mothers who were
interviewed. Given the potentially non-random nature of the missing values,
those 23 mothers were excluded from the study (see Missing Data Estimation for

greater detail).

39

Of the 166 families included in the current study, 82 families were
assigned to the comparison group and 84 were assigned to the Jackson CAA’s
Early Head Start Program. Those families in the program received an average of
82 hours of services (SD = 67 hours, range = 5 - 343) and the average length of
time in the program was 23 months (SD = 12, range = 2 - 45). The comparison
families were free to seek other services available in the community; therefore
families within that group may had had home visiting services from other
agencies, but not from the local EHS program and most likely not of the same
duration or intensity. The exact amount of community programming attained by
families assigned to the comparison group was unknown in the current study and
as such will be assumed null.

Sample Characteristics

At enrollment, the mothers in the sample were an average 22 years (SD =
4.8) of age. Thirty-eight percent of the mothers in the sample were teenaged (19
or younger) at the birth of the study’s focus child. A majority of the women
reported being Caucasian (76.2%), 15.9 percent reported being African-
American, and 7.9 percent reported being of other ethnic backgrounds
[Mexican/ChicanolHispanic (2.7 percent), Asian (2 percent), American Indian (2
percent), and biracial (1.3 percent)]. In this sample, 42.6 percent did not
complete high school, 35.8 percent had a high school diploma, and 21.6 percent

reported attending college. Of those mothers that attended college, two (1.4%)

40

reported attaining an Associate’s degree and one (.7%) completed a Bachelor’s
degree.

The median annual household income1 at baseline for this sample was
$7,920 (range = $0 to $35,000) and the mean income was $9,400 (SD = $6,559).
Families in the study were living at 67 percent of poverty (computed using family
income, family size and federal poverty guidelines) on average (SD = .45).
Eighty-two percent of the families in the current study reported living at or below
100% of poverty at enrollment. To further indicate ﬁnancial status, nearly all of
the families (98.7%) report using at least one type of public assistance including
WIC, AFDC, Food Stamps, public housing, Social Security Income, and
Medicaid. The average number of assistance programs utilized by study
participants was 2.9 (SD = 1.3, range = 1 to 6) at enrollment. Only two families
reported not utilizing any public assistance programs. For maternal occupation,
28.2 percent were employed and 12.1 percent were in school or other training
programs, and 32.2% reported being unemployed. The options of being a
homemaker, retired, or unable to work due to disability were reported by 27.5%
of the sample.

Family composition of the 166 families at the time of enrollment indicated
most families were single parent families. Over half of the sample (62.7%)
reported being single parents, with 15.4 percent of those mothers reporting

having been married at one time and either separated, divorced, or widowed.

4]

The remaining 37.3% of the mothers reported being married or cohabiting and
thus in two-parent families. Furthermore, the average family size was 3.5
persons (range = 1 to 9; one representing the applicant being pregnant at time of
enrollment). The infants in the study averaged 5 months of age (SD = 3.8) at ﬁrst
assessment (range = 1 - 16 months), 25.3 months (SD = 1.5 months) at the 24-
month assessment and 37.6 months (SD = 1.9 months) at the 36—month
assessment. The maternal report of the infants’ ethnic breakdown was like that
of their mothers: 63.3 percent were Caucasian, 18.3 percent were African-
American, and 18.3 percent were reported being of other ethnicities.
Instrumentation
Data collected from the families in the current study were done in

conjunction with a national evaluation of Early Head Start as conducted by
Mathematica Policy Research, Inc. (Raikes & Love, 2002). As a result, much of
the data being used in the current study were entered, veriﬁed, and constructs
were computed by an external agency contracted to conduct the national
program evaluation. There were constructs being used in the current study that
were collected in tandem with the national battery because they were of speciﬁc
interest for local research questions. Both sources of data were represented in
the instruments of interest for this study.

Prpgram Dosage: The total number of hours of programming was attained

during a thorough review of home visitor documentation. Among the information

 

‘ Income reported at local baseline interview, does not reﬂect income used to determine eligibility.

attained from program records were the length of time of each individual visit, the
number of persons present, the location, the focus of the visit, and the
engagement of the participants during the visit. Data regarding the total amount
of time the EHS program staff were involved in visits with the family or making
referrals on the family’s behalf was used in the current study as a proxy for
program dosage. The variable was computed as the total number of hours that
program staff was engaged in activities with the family or on the family’s behalf.
The total number of program hours was signiﬁcantly correlated with the total
number of visits to the family and the length of time in the program (p < .05).
Those families randomized into the comparison group were assigned the
program dosage amount of zero. It was understood that these families likely
received other services that were available within the community, but they should
have had no EHS program services.

This program dosage variable is a continuous variable that represents the
total number of hours that families in the EHS program received services. This
program dosage variable existed for only half of the families in the study (i.e., the
randomized comparison group received no EHS services) and was assumed
zero for the comparison families in the sample, as such there is wide variation in
the variable itself. Univariate skewness and kurtosis statistics suggest that this
variable may yield problematic results (2 and 4, respectively). Estimates of
univariate kurtosis fall within the limits, but skewness is on the cusp of meeting

criteria for problematic results as suggested by Curran, West, and Finch (1996).
43

Global Adversig: The measure of global adversity was constructed using
variables collected at or near enrollment into the EHS Research Study. The
construct was operationalized as the total number of adverse conditions
experienced by a family, at the level of the individuals within the family and family

as a unit (see Table 1). The individual items within the scale were binary scored

Table 1: Global AdversitiesI Percentages in Population

 

CHILD RELATED: % PRESENT:
Low Birth Weight 8.4%
Preterm Birth 10.2%
Biological/Medical Risks 27.1%

PARENT RELATED: -

Single Parent 38.6%
Teenage Parent (age 19 or younger) 38.4%
Low School Achievement (High School not

Completed) 42.3%
Substance Abuse or Addiction 33.1%
Mental or Emotional Illness 5.2%

FAMILY RELATED:

Homelessness 14.9%
Unstable residence (2 or more moves in one year) 19.9%
Unemployment 59.7%
Poverty (at or below 100% Federal Guidelines) 81.8%

 

 

 

 

 

as either present (1) or absent (0) such that the total number of present factors
represents the Global Adversity score. This single composite score was based

on the non-speciﬁc hypothesis supported by the literature. It was assumed that
44

all adversities were equal and carry the same weight for the family in terms of
predicting outcomes. Therefore, higher scores on this measure represent a
higher number of adversity experiences for a family, but do not suggest that any
one condition was more or less of a risk for adverse outcomes than any other.

Parenting Stress Index/thort Form (Pﬂ_S_F): The PSI/SF contains 36
items divided into three 12-item subscales (Abidin, 1990): Parental Distress (PD),
Parent-Child Dysfunctional Interaction (P-CDI), and Difﬁcult Child (DC). The DC
subscale was not included in the national battery and, as such, will not be
included in the current study. The ﬁve-point Likert-type response scale ranged
from 1 (strongly disagree) to 5 (strongly agree).

The Parental Distress (PD) subscale was designed to quantify the distress
a person experiences, as a function of individual personal characteristics, in the
role as a parent. The principal component stressors of the PD subscale were
associated with stress associated with the demands of being a parent, and with
depressive symptomatology. Examples of items on the PD subscale were “You
often feel that you cannot handle things very well” and “You feel trapped by your
responsibilities as a parent.” The Parent-Child Dysfunctional Interaction (P-CDI)
subscale taps the parent’s perceptions that the child did not meet his or her
expectations. The subscale was also designed to capture whether or not
interactions with the child were reinforcing to the parent. An example item from
the P-CDI subscale was “Your child rarely did things that make you feel good.”

With a normative sample of 800 subjects, Abidin (1990) reports Cronbach alpha
45

reliability coefﬁcients of .87 for Parental Distress, and .80 for Mother-child
Dysfunctional Interaction. Using a sample of low-income Head Start families,
which was more like the sample in the current study, Reitman, Currier, 8. Stickle
(2002) report Cronbach’s alpha coefﬁcients of .88 for both the Parental Distress
and the Mother-Child Dysfunctional Interaction subscales. Reliabilities for the
current sample were high with Cronbach alpha reliability coefﬁcients of .83 for
Parental Distress collected at the child’s 24th and 36"1 months, .80 and .79 for
Mother-child Dysfunctional Interaction and .86 and .85 for the combined scales
as were used in the analyses at 24 and 36 months, respectively.

The Fa_milv mironment Scale (FES): Conﬂict Subscale. The FES is a
90-item scale created to measure relationship dimensions, personal growth
dimensions and system maintenance dimensions within the family environment
(Moos, 1974). The F ES Conﬂict subscale only was used to measure the
expression of aggression and anger as well as conﬂicted interactions of the
family (Moos, 1974). The conﬂict items included in the analyses were: 1) we ﬁght
a lot, 2) we hardly ever lose our tempers, 3) we sometimes get so angry we
throw things, 4) we often criticize each other, and 5) we sometimes hit each
other. The reported internal consistency for the subscale was .75 (Moos, 1974),
and for a previous study with the current sample the Cronbach alpha was .69
(McKelvey et al., 2002). At the 24 month child’s birthday related assessment, the
Cronbach alpha reliability coefﬁcient was .62 and at the 36 month birthday

related assessment was .70.
46

family Crisis Oriented Personal Scale (F-COﬂEﬁ). The FCOPES is a 30-
item scale designed to measure problem solving behaviors and attitudes that
families utilize to respond to problems (McCubbin, Olson, & Larson, 1987). The
FCOPES consists of ﬁve subscales that include Mobilizing the Family to Acquire
and Accept Help, Acquiring Social Support, Reframing, Seeking Spiritual Support
and Passive Appraisal. Due to low internal consistency reliabilities for our
sample on some of the original scales, a conﬁrmatory factor analysis was
conducted on the FCOPES measure and new scales were developed (see
McKelvey et al., 2002).

The new structure of the FCOPES for this study was comprised of one
original and three new subscales. The subscale from the original FCOPES was
Cognitive Reframing. The Refrarning subscale had a reported reliability of .82
(McCubbin et al., 1987) and a Cronbach’s alpha of .67 with the current sample.
Three new subscales were determined to be more appropriate for our low-
income high-risk sample; Seeking Support from Friends and Family, Seeking
Support from Neighbors, and Seeking Support from Service Providers. These
three subscales represent a realignment of items from the original Acquiring
Social Support (Cronbach’s alpha=.83) and Mobilizing Family to Acquire and
Accept Help (Cronbach’s alpha=.71) subscales (McCubbin et. al, 1987). The
Seeking Support from Friends and Family subscale consists of 6 items and had a
reliability of .79 for the current study sample. The Seeking Support from

Neighbors subscale consists of 3 items with a reliability of .73. The third of the
47

new subscales, Seeking Support from Service Providers, consists of 3 items with
a Cronbach’s alpha of .6. For the purpose of the current study, these three
scales were combined into a composite score to represent general support
seeking coping strategies. The zero-order correlations of the three subscales
ranged from r= .31 to r= .38 (p < .01). The Cronbach’s alpha for the resulting
support seeking scale was .81.

The Nursing Child Assessment Satellite Training (NCAST) — Teaching
Spaiep The NCAST — Teaching Scale is a 73-item observation measure created
to measure mother-child interactions during a teaching episode (Barnard &
Kelley, 1990). During the teaching episode, mothers were asked to teach their
child an age-appropriate skill they had not already acquired, such as visually
following a rattle for younger infants and stacking cubes for those who were
older. Data collectors were trained by a certiﬁed NCAST instructor and had to
reach a reliability of at least 85% on standardized videotapes to be qualiﬁed to
conduct and score the observation. Data collector reliability was renewed at
least yearly.

The scale was comprised of items scored as 1 (observed) or 0 (not
observed) across four parent behavior subscales (Sensitivity to Cues, Response
to Distress, Social-Emotional Growth Fostering and Cognitive Growth Fostering)
and two child behavior subscales (Clarity of Cues and Responsiveness to
Caregivers). The number of observed behaviors was summed for each subscale

and for the total scales, which include parent (50 items), child (23 items), and
48

dyadic total scales. For the purpose of this study, the total dyad score will be
used as the outcome. Internal consistency reliabilities reported by NCAST range
from .52 to .80 for the subscales and from .76 to .87 for the total scales (Sumner
& Spietz, 1994). Estimates of reliability for the full scale as observed at 24
months was .80 and .76 at the child’s 36 month birthday related assessment.
Power Apglvsis and Sample Size

Determination of sample size in Structural Equation Modeling is
dependent upon two distinct needs. Adequate sample size is important for the
accuracy with which one accepts or rejects a model, as many indices of model ﬁt
are sensitive to sample size (Muthén & Muthén, 2002; Tanaka, 1993). Sample
size is also an important factor in the accurate depiction of parameter estimates
(MacCullum & Austin, 2000). A power analysis for the model being proposed
yielded promising results. MacCullum, Browne, & Sugawara (1996) carefully
describe power analysis and determining sample size for structural equation
modeling using the distribution of one of the most commonly used indices of
model ﬁt, the root mean-square error of approximation (RMSEA). For the
proposed model, there were 12 manifest variables and 78 base degrees of
freedom. According to MacCullum and colleagues, to achieve a minimum power
of .80 for models of close ﬁt, with 80 degrees of freedom, the minimum sample
size must be 154 subjects. The sample for the current study (N = 166) falls

above this number for a minimum power of .80.

49

Missing Data Estimation

Of the 196 families enrolled in the research and evaluation study, 189
were retained in the sample after random assignment. There were no differences
between the groups at random assignment at the national level (Raikes & Love,
2002; US. Department of Health and Human Services, 2002). Furthermore, in
the local Michigan sample there were also no differences between those
assigned to the comparison and the program groups on demographic
characteristics and measures collected at or near enrollment into the research
study (Schiffman & McKelvey, 2002).

There were 140 families from whom data were collected at both the 24
and 36 month collection point. Data were imputed for the remaining 26 families
from whom information was collected at one time point, either the 24- or the 36-
month assessment. At the 24-month child birthday related interview, there was
little missing data. For each individual measure, 82% to 93% of the cases had
complete information. This was also the case at the 36-month interview. For
each individual measure in the model, cases with complete data ranged from
82% to 91% (see Appendix A for a detail of missing values).

At enrollment, there were no signiﬁcant differences between those families
included in the present set of analyses and those that were excluded based on
missingness of data (n = 23) with respect to age, race, education level,

employment or marital status of the mother, or to family income. However, a

50

greater percentage of mothers reported being employed at enrollment (28%)

versus unemployed (5%) when 24- and/or 36-month child birthday related data

Table 2: Univariate Descriptives Pre-Imputation

 

 

Std.
N Mean Deviation
Program Dosage 164 41.05 62.64
Global Adversity 166 3.62 1 .57
Parenting Stress Index (24 months) 154 46.46 14.71-
Parenting Stress Index (36 months) 151 46.21 13.81
Family Conﬂict (24 months) 144 1.78 0.53
Family Conﬂict (36 months) 143 1.85 0.62
Cognitive Reframing (24 months) 151 4.32 0.50
Cognitive Reframing (36 months) 144 4.39 0.49
Support Seeking (24 months) 151 2.92 0.79
Support Seeking (36 months) 144 3.09 0.79
Mother — Child Interaction (24 months) 145 55.77 6.55
Mother — Child Interaction (36 months) 141 54.77 6.24

 

were collected. There was also a difference in the primary language of the family
with 16.7% versus 4.7% of the sample from whom we collected 24 and 36 month

child birthday related data speaking a language other than English at enrollment

51

Table 3: Univariate Descriptives Post-Imputation IN = 166)

 

 

 

Std.
Mean Skew Kurtosis

Dev.
Program Dosage 40.42 57.81 1.33 0.63
Global Adversity 3.62 1.52 0.25 -0.58
Parenting Stress Index (24 months) 46.23 13.03 0.74 -0.12
Parenting Stress Index (36 months) 46.08 13.32 0.51 044
Family Conﬂict (24 months) 1.76 0.49 0.48 041
Family Conﬂict (36 months) 1.83 0.58 0.83 0.18
Cognitive Reframing (24 months) 4.31 0.45 -0.5 0.14
Cognitive Reframing (36 months) 4.41 0.45 -0.69 -0.16
Support Seeking (24 months) 2.89 0.77 0.14 -0.44
Support Seeking (36 months) 3.12 0.76 -0.09 03
Mother — Child Interaction (24 months) 55.13 6.44 -0.23 -0.17
Mother — Child Interaction (36 months) 54.68 5.75 -0.2 -0.07

 

(see Appendix B for a list of demographic characteristics compared). There were
no differences between families who were retained in the sample, but for whom
missing data were imputed with regards to these same demographic
characteristics. The univariate descriptive statistics of the data pre- and post-
imputation were very similar (see Table 2 for Pre-Imputation Descriptive

Statistics and Table 3 for Post-lmputation Descriptive Statistics).
52

Data were imputed using the expectation maximization (EM) method. The
EM method is a full information method of imputing missing values by iterating
through the existing data and ﬁtting the best values to the existing covariance
structure (Acock, 1997). This method of handing missing values produces less
bias in the results than deleting cases or using the sample mean for imputation
(Acock, 1997; McDonald & Ho, 2002; Rovine & Delaney, 1990). In order to
estimate missing values, one must be able to assert that data were missing at
random (MAR). For any single variable, data were MAR if the pattern of missing
values did not rely on the values of the same variable (Myrtveit, Stensrud, &
Olson, 2001; Rovine & Delaney, 1990).
Outlier Analysis

Upon examination of the distributions of the data, various outliers were
discovered. There were several values in the data that were out of range and
were determined likely to change the results of the analyses. These values were
truncated to the next nearest in range value, or windsorized (Foster, 1986). Data
were modeled after being windsorized (see Appendix C for a complete listing of
numeric replacements).
Analﬂic Methods

The hypothesized manifest variable model depicted in Figure 2 was tested
using structural equation modeling techniques with LISREL 8.54 (Jbreskog &
Sbrbom, 2003). The observed covariance matrix of the variables was compared

to what would had been expected given the set of interrelationships that were
53

proposed. The model was tested using a manifest variable model, which
explored the interrelationships between the observed variables, rather than
between latent constructs (Schumaker & Lomax, 1996). A careful examination of
the descriptive statistics and correlations between the variables being tested in
the manifest variable model was a crucial ﬁrst step before testing the structural
equation model. It is Important to be aware of any skewness or kurtosis in the
data prior to ﬁtting a model (McDonald & Ho, 2002). Curran, West, and Finch
(1996) recommend concern if univariate skewness and kurtosis criteria were
greater than 2 and 7, respectively. Many of the estimation methods for ﬁtting
structural equation models, such as maximum likelihood (ML) and generalized
least squares (GLS), require the assumption of multivariate normality (Bollen &
Long, 1993) and simulation studies had demonstrated that ML and GLS
estimation can yield biased results with data that are highly skewed (Hu &
Bentler, 1995). The effect of violating the assumption of multivariate normality is
that the chi-square statistic may be too large, which would yield a model being
rejected, and standard errors may be too small, such that signiﬁcance tests have
too much power (Hu & Bentler, 1995).

Identiﬁability of the path models was also taken into consideration.
Parameters of an estimated path model are identiﬁed when they can be uniquely
determined from the covariance matrix (Brito & Pear, 2002). McDonald and Ho
(2002) stress the importance of following the precedence and orthogonality rules

in determining the identiﬁability of a model. The precedence rule (McDonald,
54

1997) indicates that the covariances of the error terms of the causally ordered
variables in the model are zero. The orthogonality rule (McDonald, 1997) can be
met when the covariances of the error terms of all endogenous variables in the
model are zero.

Commonly reported statistics of global model ﬁt include the Satorra-
Bentler Scaled Chi-Square (Satorra & Bentler, 1988) statistic, the Root Mean
Squared Error of Approximation (RMSEA; Steiger, 1990), the Goodness of Fit
Index (GFI), and the Comparative Fit Index (CFI; Bentler, 1990). The GFI and
CFI were incremental ﬁt indices and measure the improvement of the model over
a null model. The GFI and CF I range in value from 0 to 1 with values above .90
a common cut for acceptable models and above .95 indicating an excellent ﬁt
(Hu & Bentler, 1995, 1999; Vandenberg & Lance, 2000). The RMSEA is
considered an absolute ﬁt index with values close to zero suggestive that the
model is a good ﬁt to the data. These measures of model ﬁt were considered in
tandem for determining if the data appropriately ﬁt the model tested as had been
suggested in the literature (Hu & Bentler, 1999). Methodologists also suggest
that researchers give theoretical grounds for the inclusion, or alternatively the
omission, of paths within the model (Boomsma, 2000; Hoyle & Panter, 1995;
McDonald & Ho, 2002). Adaptations to the hypothesized model were therefore
theoretically substantiated and modiﬁcation indices, as a result of model ﬁtting,
were considered for inclusion in the model only on the basis of theoretical

suppon.
55

CHAPTER 4
RESULTS
ModLel F itting;_Base Modil- Styress and Mother-Child Interaction

The ﬁrst set of structural equation models tested whether or not the stress
variables in the model independently predicted mother-child interaction (see
Table 4 for correlation matrix). The model that tested the independence of the
predictors ﬁt the data moderately well; the overall chi-square (df = 16, N = 166)
was 30.21, p = 0.02 (see Appendix D for model ﬁt indices). The chi-square
provides a measure of discrepancy between the covariance matrix of the sample
and the estimated matrix (Schumacker & Lomax, 1996). The Root Mean Square
Error of Approximation (RMSEA; Steiger, 1990) for the model was 0.07, and
other indices of model ﬁt were also indicative of moderate model ﬁt (Brown
&Cudeck, 1993).

The second model tested was less stringent and allowed the modeling of
relationships between stress predictor variables as well as with the outcome,
mother-child interaction. Overall, the ﬁt indices for the model suggest that the
model ﬁt the data well; the chi-square (df = 10, N = 166) was 6.84, p = 0.74, the
RMSEA for the model was 0.0, and the GFI and the CF! for the model were .99
and 1.00, respectively (see Appendix E for model ﬁt indices). Interestingly,
amounts of family adversity was not signiﬁcantly related to program dosage,
mother-child interaction assessed at the child’s 24'" month of age, parenting

stress, or family conﬂict at 36 months of age. A trend towards signiﬁcance was
56

apparent in the relationships between family adversity and mother-child
interaction at 36 months and with family conﬂict at the 24 month assessment.
Contrary to expectation, neither program dosage nor family conﬂict signiﬁcantly
predicted mother-child interaction at either observation.

Given the lack of relation between constructs in the model, an additional
analysis was run which resulted in a more parsimonious model (see Figure 3).
The model was adapted by excluding paths with non-signiﬁcant coefﬁcients.
There were two notable exceptions to these exclusions; 1) even though the
autoregressive effect of mother-child interaction collected when the child was 24
months to the same measure when the child was 36 months of age was not
signiﬁcant, it was retained in the model as the relationship should theoretically be
modeled and 2) the non-signiﬁcant relationships between program dosage and
mother-child interaction at both collection points were retained as they were of
central interest in the study. The resulting model yielded ﬁt indices that were
indicative of very good ﬁt between the data and the model. The chi-square for
the model (df = 16, N = 166) was 7.05, p = .97. The RMSEA, GFI and CF! for the

model were .00, .99 and 1.00, respectively (see Appendix F).

57

828.2 8.6. 88 65 8 288288 e 882686 ..
828-2 862 8.8 65 s 288288 2 88686 t

 

 

8. 8. :8. 8. 2. 8.- 2.- :8.- ......8.- 2.- 8.- 6288882
226-882
8.- 8. :. 8. 8.- 2.- 2.- ...2.- 2. 8. 8288882
220-882
:8. :8. 2. 8.- 8.- 2.- 8.- 8. 2. Ge
@2me tommsm
8.- 2. 8.- 8.- 8. 8.- 8. 3. $2
mac—66m tomasm
:8. ...2.- 2.- 1:.- ......8.- 8. 8. 3282882
8288
8.- 8.- :8.- ......8.- 8. 8. $28288:
63886
*nmv. 3.5m. .ILN. mo... S. 63
2:86 882
:8. :8. 2.- 8.- 82
2:86 28: s
:8. 8.- 2.- 62 s
mmobm @5262wa
8.- 8.- $2
mmobm 25:6an
8.- 28%.: 826
OWN SS SS 933033 313 DJ Sd S.d Vmut Gum
6mm 6.6. .8 6m... 6W8 Um um um um pq m.-
oo mo mo u. n. 3M 1.6. sn. sn 61 t
n... a u u 1.1. 1.1. w w ) ) I. a
mmm m 3.26 a 8 .8... m
L G {W {G {W
p

 

 

6:23.230 3 636...

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.6>6_ o... v o 65
“6 25.6568 56: .0 66:85:96 665:6: . 35:05. on ©v
.96. mo. v a 65 «05:00 >=Emu
.6 626.6568 56: .o 66:85:96 663:6: . to F. >
w
.8".- .8822 on Q
85:22 on ©v 666.5 m:=:6.6n. one.
w:o_>6:6m >
5:66.625 .6565.
_ .66. 358.2 «N O.
. 8:86 2:8“.
on.- - >
tou-
mo.-
.oscoz em O.
36.5 m:=:6:6n_ .
3. .9.

 

 

86.6%.: _ooo_o

 

 

 

 

 

$582 em O.
8o_>6:6m
5:66.625 6:85.

 

59

 

 

 

 

63660 E6365

 

 

 

.6uos. 666m ”n 6.59“.

When models are nested the chi-square difference test is most frequently
used for hypothesis testing (Bentler & Mooijaart, 1989). The differences between
the chi-square estimates of each model were known to be distributed as a chi-
square with degrees of freedom equal to the difference in the degrees of freedom
of the two models. When the difference in chi-square units is non-signiﬁcant, the
interpretation is that the two models do not differ signiﬁcantly in their modeling of
the data, and the more parsimonious model (i.e., the model with fewer estimated
parameters) is accepted. The change in chi-square units between the two
nested models described above was 0.21 for 6 degrees of freedom. This non-
signiﬁcant change indicated that the model presented in Figure 3 did not ﬁt the
data more poorly than the previous model.

The results. from the model suggest that mother-child dyads in which the
mothers report less parenting stress were observed to have had more positive
interactions when the child was both 24 and 36 months of age (I = -2.27, p < .05
at 24 months and t= -3.41, p < .05 at 36 months). The mothers in these dyads
were more sensitive to infant cues, more responsive to distress signals and
fostered social/emotional and cognitive growth. Likewise, the children in these
dyads were clearer in their cues and more responsive to their mothers. Further
results suggested that mothers who reported more family conﬂict also reported
more parenting stress when assessed at or near the child’s 24"1 (t = 3.66, p <

.05) and 36‘" month (t = 2.87, p < .05). As seen in the previous model, program

60

dosage did not directly predict mother-child interaction at either assessment.
Furthermore, family adversity did not predict mother-child interaction at 24
months, but a trend was demonstrated at the 36 month observation which may
be suggestive that greater amounts of family adversity were related to lower
interaction scores. Interestingly, family adversity nearly signiﬁcantly predicted
family conﬂict reported at the child’s 24th month of age (t = 1.69, p < .10). This
relationship was in the opposite direction as to what was predicted and may
indicate that mothers in families with greater adversity reported lesser amounts of
family conﬂict. Overall, the percentage of variance explained in mother-child
interaction at both ages was very small (at 24 months, R2=.05 and at 36 months
R2=.09)

Modeling the Direct Effects of Coping:

The next set of models was executed to determine the direct effects of
coping on mother-child interaction. Using the base model determined in Figure
3, the coping constructs, cognitive reframlng and support seeking behaviors,
were added to the model predicting mother-child interaction at the child’s 24th
and 36"’ month of age. Not modeling relations between the coping constructs
themselves or relations between coping variables and the remaining predictors in
the model yielded a model with poor ﬁt (see Appendix G). The overall chi-square
(df = 46, N = 166) was 96.04, p = .00. The RMSEA was .08, the GFI was .91 and

the CPI was .83, some of which were indicative of less than adequate ﬁt.

61

A second model was run which permitted the relationships between
coping constructs and the remaining predictors to be estimated. For the sake of
model parsimony, paths that were non-signiﬁcant were excluded from the model.
In addition to those paths that were not estimated from the base model, the paths
from cognitive reframlng and support seeking (reported at the child’s 24th month)
predicting mother-child interaction when the child was 24 months old were non-
signiﬁcant and therefore not modeled. Furthermore, the path from cognitive
reframing reported at 36 months of child’s age predicting mother-child interaction
at the same time point was also non-signiﬁcant and excluded from the ﬁnal
model. What resulted was a model that adequately represented the data
according to indices of ﬁt (see Figure 4 and Appendix H for model ﬁt indices).
The chi-square (df = 46, N = 166) was 24.6, p = .99, the RMSEA was .00, and
the GFI and CF I were .98 and 1.00, respectively.

In addition to the ﬁndings in the base model, results from this model
suggested that in families in which mothers reported more parenting stress they
also reported using less cognitive reframlng at 24 months (I = -4.72, p < .05) and
at 36 months (t = -3.42, p < .05) of age. There was also a positive relationship
between the coping constructs, with mothers who report more highly utilizing
cognitive reframing strategies also reporting using more support seeking
strategies (t = 1.98, p < .05 at 24 months and t = 4.35, p < .05 at 36 months).
Use of support seeking coping directly impacted mother-child interaction; this

effect was not seen with cognitive reframing. Mother-child interaction at 24
62

months was not predicted by support seeking, while the same measure at 36
months of child’s age was signiﬁcantly predicted by both the 24 and the 36 month
support seeking behavior coping scores. What was most notable in this analysis
was that support seeking behaviors at 24 months positively predict (t = 4.48, p <
.05) mother-child interaction at 36 months, while support seeking at 36 months of
age negatively predicts (t = -2.07, p < .05) interaction at the same age.
Interestingly, this negative relationship between 36 month support seeking and
dyadic interaction appeared to be a suppressor effect (see Appendix I). The zero
order correlation for these variables was small and positive (r=.03, p = .75),
however, when one controlled for the shared variance in support seeking at the
two time points, the partial correlation between the two concurrently assessed
variables, albeit non-signiﬁcant, became negative (r=-.12, p = .12). Another
unexpected ﬁnding was that although program dosage did not predict mother-
child interaction at 24 months, a negative trend was demonstrated at 36 months,
which may suggest that greater amounts of program dosage predicts lower
mother-child interaction scores. Overall, there was no change in the percentage
of variance in mother-child interaction scores at 24 months explained (R2=.05),
but the addition of the coping variables explained 11% more variance in mother-
child interactions observed at 36 months (R2=.20).
Coping as a Moderator — Support Seeking:

Interaction terms had to be computed to determine moderating effects of

coping; as such many additional variables were included in the model. Support
63

 

 

 

 

 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

66. or. v a 65 .6 25.65600 N.
58 .o oomoogcgo neocoo . 35:22 on O. .owcos. on O.
_o>o_ 8. v 65 .o 286568 m 2.: 582.6”.
. . 65:00 =E6n.
.7
«0N... «0F. .7
«m...- 50. e .3.
. .8. .2822 8 00.
35:05. on ©. LN - 666:5 m:_.:6._6n.
20_>m:6m t L». ,. - >
8. 88 c .o o .- $58.2 em O. 2 .6585. em O.
.. 5 _ 5 .2 NF 0:263 :03?» 0:.E656m.
«.m- 6>_._:moO «oo- Am5:0.2 VN @.
8:80 2:8:
hm.- tau.
4.
.8282 em O.
36.5 0:55:60 .2..-
so.
3.665.... .6020
«m...-
_. F.
.8285. em ©.
m.o_>m_.__wm .vc.
5:06.655 .6505. 68600 E6505

 

 

@5000 n.0 ﬂootm “06.5 05.6005 3. 6.59....

64

seeking and cognitive reframlng coping strategies were modeled separately as
the sample was not sufﬁcient to allow them to be modeled together. Results
from the previous model suggested that support seeking coping directly impacted
mother-child interaction when the child was 36 months of age. To determine
whether support seeking moderated the relationships between other predictors in
the model and mother—child interaction, a set of analyses were executed using
the interactions of predictors in the model and support seeking.

Using the relationships determined in the previous model and adding in
the original main effects of variables that were excluded for the sake of model
parsimony, eight additional interaction terms were added to predict mother-child
interaction. Interaction terms were created with the centered predictor variables
and were held time constant (i.e. product terms of stress and coping measures
collected at the same wave of data collection were computed). The ﬁt indices
suggested that the model ﬁts the data well. The RMSEA for the model was .00,
chi-square (df = 122, N = 166) was 120.58, p = .52, the GFI was .92 and the CF I
was .98 (see Appendix J). Within this model, many of the interactions terms
between predictors and support seeking were not related to mother-child
interaction.

The non-signiﬁcant pathways within the model were: 1) interactions of the
dosage with support seeking coping at 24 months or 36 months of child’s age did
not signiﬁcantly predict mother-child interaction, 2) neither of the adversity and

support seeking interactions signiﬁcantly predicted mother-child interaction at
65

 

Figure 5: Moderating Effects of Support Seeking

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

   
     

 

 

 

 

 

 

 

 

 

 

 

 

 

grogram Mother-
a
OS ge .05 Child
Global 12 Intféagt‘iton
Adversity - Months)
«13' -.17*
Parenting
Stress (24) “-02
A .03
.28"
“ . -.13‘ t
66,, Family .3
' Conﬂict (24)
-.11 '07
Parenting _ __1 2,
Stress (36) N 45.
l ”6* tr -.3o*, Mother-
Family Chlld.
Conﬂict (36) 1 lntaagtéon
Support Months)
Seeking (24)
.47* -.16*
Suppon
Seekin 36
g ( ) Support by
Support by Family
Parenting Conﬂict (24)
Stress (24)
\ '18*\
'21. Support by Support by Family
Parentm Conﬂict (36)
Stress (36)

 

 

 

 

 

 

 

* denotes signiﬁcance of path estimates at the p < .05 level
‘ denotes signiﬁcance of path estimates at the p < .10 level

 

 

66

either wave of collection, 3) the interaction term of parenting stress at the child’s
36th month with support seeking at the same age did not predict concurrently
observed mother-child interaction, 4) the interaction term of family conﬂict with
support seeking (both at 36 months) also did not predict mother-child interaction
at the same time. The model was reﬁned by excluding both of the adversity and
support seeking interaction terms and both of the program dosage and support
seeking interaction terms. The ﬁt indices for the resulting model suggested that
the model ﬁts the data moderately well. The RMSEA for the model was .03, chi-
square (df = 69, N = 166) was 79.47, p = .18, the GFI was .94 and the CF! was
.94 (see Figure 5 and Appendix K).

In addition to what was discovered in previous models, two product terms
nearly signiﬁcantly predicted mother-child interaction when the child was at or
near 2 years of age; parenting stress with support seeking at 2 years (1‘ = -1.69, p
< .10) and family conﬂict with support seeking at 2 years (t = 1.74, p < .10). The
interaction effect of parenting stress and support seeking may indicate that for
families with low amounts of parenting stress, support seeking did little to impact
mother-child interaction scores (see Figure 6, graph 1 depicts scores of mothers
who report low parenting stress). For the high stress parents, however, support
seeking may impact mother-child interaction in that within that group of mothers,
those who relied on support from others had higher mother-child interaction

scores than those mothers who did not report seeking as much support (see

67

Figure 6, graph 2 depicts mothers who report greater amounts of parenting

stress).

Figure 6: Plots for Nearly Signiﬁcant Interaction Effect — Parenting Stress and
Support Seeking predicting Motpeﬂ-vegrgld Child Interpction

 

 

 

1 2
5- ,_ 5-
V 4' <1- 4.
E 3--—-———-—— E 3-
% 2- ° .31-“CT % 2:
U) 1. J c0 1-
”W70 807011 70
W 68w 1% 68%
P8124 20 MTOTN—3 P804 20 WOT/5L3

The trend towards signiﬁcance for the family conﬂict and support seeking
interaction term may be suggestive of much the same as what was found for the
parenting stress and support seeking interaction. For families who reported
lesser conﬂict, support seeking strategies did not seem to impact mother-child
interaction. However, in the families in which mothers reported higher amounts
of conﬂict, those who reported seeking support from others tended to have higher

scores on the mother-child interaction measure (see Figure 7, graph 1 and graph
68

2 refer to the lower and higher conﬂict families, respectively). Overall, there was
little change in the explained variance in mother-child interaction scores at 24
months explained (R2=.08) and in mother-child interactions observed at 36

months (R2=.23) with the inclusion of the support seeking interaction terms.

Figure 7: Plots for Nearly SigniﬁcaLnt Interaction Effect — Fa_milv Conﬂict and
Support Seeking predicting Mother-g-vear old Child Interaction

 

1 2

5- 5-
v 4" 5 '6 4r J"
133.”... i2 .- 2....—
E ‘ oat-.0. o: J” E ________.———°' 3c. .
0- 2L ° 3' o. 2- . -.:.’
3 °' . D
(I) 1. - U) 1.

 

 

1
30.0w MTOTN—3 BMW wow

Coping as a Moderator - Cognitive Reﬁjming

According to the results of the direct effects of coping model, cognitive
reframlng coping strategies did not directly impact mother-child interaction.
Additionally, there was a relationship between parenting stress and reframing

such that those who reported more parenting stress reported less cognitive
69

reframlng. To determine if cognitive reframing moderated the relationships
between other predictors in the model and mother—child interaction, a set of
analyses were executed using the interactions of predictors in the model and
cognitive reframlng. As with the moderating effects of support seeking analyses,
eight additional interaction terms, computed in the same manner, were added to
predict mother-child interaction. The moderating model thus contained all main
effect terms and their coinciding relationship to mother-child interaction and the
relationship between the eight interaction effects and the same outcome. The ﬁt
indices suggested that the model ﬁts the data moderately well. The RMSEA for
the model was .04, chi-square (df = 120, N = 166) was 158.72, p = .01, the GFI
was .90 and the CF I was .85 (see Appendix L for indices of model ﬁt).
Adaptations to the model were made to arrive at a model with better ﬁt to
the existing covariance structure. The adaptations included the removal of non-
signiﬁcant path coefﬁcients, speciﬁcally the removal of the moderating effects of
dosage and cognitive reframlng as they did not predict mother-child interaction.
An addition to the model that was suggested by modiﬁcation indices and
supported by earlier models was the inclusion of the relationship between family
conﬂict and parenting stress. The interaction of family conﬂict with cognitive
reframing was found related to parenting stress each asSessed at 24 months.
There were also error covariances added between main effect of parenting stress
and its interactions with cognitive reframlng. The resulting model was a much

better ﬁt to the data.
70

The adapted model approximated the data moderately well (see Figure 8).
The RMSEA for the model was .03, chi-square (df = 93, N = 166) was 105.71, p
= .17. Further indices of model ﬁt include the GFI, which was .93 and the CFI,
which was .92 (see Appendix M). In addition to what was discovered in previous
models, one product term signiﬁcantly and one term nearly signiﬁcantly predicted
mother-child interaction when the child was at or near 2 years of age; parenting
stress with cognitive reframlng at 2 years (t = -2.82, p < .05) and family adversity
with cognitive reframlng at 2 years (t = 1.74, p < .10), respectively. There was
also a signiﬁcant prediction of parenting stress at 2 years by the interaction of
family conﬂict and cognitive reframlng at the same time (t = 2.27, p < .05).

The interaction effect of parenting stress and cognitive reframlng
appeared to indicate that for families with low amounts of parenting stress,
mother-child interaction scores were better for mothers who did more cognitive
reframlng (see Figure 9, graph 1 depicts scores of mothers who report low
parenting stress). For the high stress mothers there less of a relationship
between cognitive reframlng and mother-child interaction, those who did more
cognitive reframlng had similar mother-child interaction scores than those
mothers who did less cognitive reframing (Figure 6, graph 2 depicts mothers who

report greater amounts of parenting stress).

71

Figure 8: Modeling Moderating Effects of Cognitive Reframing

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Program
Dosage Mother-
.05 Child
Global 11 '"tggﬂon
Adversity ' Months)
«13‘ -.23*
Parenting
Stress (24) ~01
.03
F .28*
-.22* t
65* Famin .13
' Conﬂict .16“
-.08 .07
-.35* .
Parenting *1 2.
Stress (36) 3‘ 45*
, “16* -.27* Mother-
"26 Family Child.
Conflict -.o1 Interaction
(@ 36
Cognitive Months)
Reframe (24) ‘
.1e* --03
Cognitive
Reframe (36)
CR by FES CR by Adversity
( ) .11 : .21* \
-11\‘ CR by PSI CR by FES CR by
(36) (36) Adversity (36)

 

 

*p<.05,‘p<.10

 

CR = Cognitive Reframing, PSI = Parenting Stress, FES = Family Conﬂict

 

 

72

 

Figure 9: Plots for Signiﬁcant Interaction Effect — Parenting Stress and Cognitive
Reframing predicting Mother-Z-year old Child Interaction

 

 

 

The interaction of family adversity and cognitive reframing both assessed
at the child’s 24th month nearly signiﬁcantly predicted concurrently observed
mother-child interaction. This interaction may indicate that for families with both
low and high amounts of family adversity, mother-child interaction scores were
better for mothers who did more cognitive reframing (see Figure 10, graph 1
depicts mothers in families with lesser numbers of adversities and graph 2
depicts mothers in families with greater numbers of adversities). For mothers in
families with greater adversity, however, there was a stronger relationship
between cognitive reframing and mother-child interaction scores than for mothers

in families with lesser adversity.

73

Figure 10: Plots for Nearly Signiﬁcant Interﬂon Eject — Family Adversity gag
Cognitive Reframing predicting Mathew-veg old Child Interaction

REFRAM3
wbP-Sﬂ
use?

 

 

 

In addition to the direct relationship between parenting stress and
cognitive reframlng, there was a signiﬁcant moderator effect demonstrated to
predict parenting stress. The interaction of family conﬂict and cognitive reframing
signiﬁcantly predicted parenting stress at the child’s 2"“ birthday. This interaction
indicated that for families with both low and high amounts of family conﬂict,
parenting stress scores were lower for mothers who did more cognitive reframlng
(see Figure 11, graph 1 depicts mothers in families with lesser family conﬂict and
graph 2 depicts mothers in families with greater conﬂict). For mothers in families

with greater conﬂict, however, there was a weaker relationship between

74

Figure 11: Plots for Signiﬁcant lnterac_tion Effect — Cognitive Reframing and

Cognitive Reframing predicting Parenting Stress in Motiﬁrs with 2-year old
Children

1 2
5.5
u. u. 4.0- '- .
LlJ . -°.‘ '
lg ‘1 3.5- ' -

 

 

W

1
32”me O 30 P924

 

cognitive reframing and parenting stress scores than for mothers in families with
greater conﬂict. In terms of the percentage of variance explained by this model,
there was change in the explained variance in mother-child interaction scores at
24 months explained (R2 = .11) and in mother-child interactions observed at 36
months (R2 = .09) with the inclusion of the support seeking interaction terms.
Interestingly, a greater percentage of variance was explained for mother-child
interaction at 24 months when considering the role of cognitive reframlng, while

support seeking appeared to explain variance in the same measure at 36

months.

75

Supporting Analyses

Some of the hypothesized relationships were supported by the results
from the structural equation models tested; alternatively, many were not
supported. In addition to the structural equation models presented, supplemental
analyses were conducted to determine possible explanations for ﬁndings that
were contrary to prediction. To summarize the ﬁndings, the main hypothesis
from the study, the unpredicted ﬁndings from the structural equation models and
some subsequent supporting analyses are detailed in the following.
family Aggrsijr a: a Predictor of Parenting Stress and Conﬂict.

Baseline family adversity was expected to predict measures of parenting
stress and family conﬂict collected at or near the child’s 2nd and 3rd birthdays.

It was hypothesized that greater amounts of adversity would predict higher
reports of family conﬂict and parenting stress. These two hypotheses were not
supported. Baseline adversity did not predict levels of parenting stress assessed
at the child’s 24th or 36th month of age, nor did it predict family conﬂict at the
child’s 36th month of age. Adversity did nearly signiﬁcantly predict family conﬂict
at 24 months (path estimate was -.13, t = -1.71, p < .10), however this trend may
suggest that there was less family conﬂict for families who had a greater number
of adversities at baseline, which was contrary to what had been hypothesized.

Within studies of cumulative risk, it has been supported that the number of
risks to which a person or family is exposed is a more powerful predictor than

speciﬁc risks (Evans, 2003). Within this sample, dichotomizing the adversity
76

variable at the threshold of three or more versus fewer adversities resulted in no
mean differences in parenting stress or in family conﬂict collected at either
assessment (see Appendix N). The mean score on parenting stress for the low
risk (< 3 risks) group was 45.97 (SD = 12.25) and for the high risk group (_>_ 3
risks) was 46.31 (SD = 13.31) as assessed at the child’s 24th month. The mean
scores for parenting stress when assessed at 36 months are 46.36 (SD = 11.51)
and 45.99 (SD = 13.89) for the low and high groups, respectively. There were
also no differences demonstrated for family conﬂict with mean scores for the low
and high risk group being 1.76 (SD = .42) and 1.77 (SD = .51) at 24 months and
1.89 (SD = .56) and 1.82 (SD = .59) at 36 months. This risk threshold analysis
also suggested no differences for families in coping strategies.
Adversity as a Predictor of Mother-Chilg lnteractm

Baseline family adversity was expected to predict mother-child interaction
observed at or near the child’s 2"Cl and 3rd birthdays. It was hypothesized that
greater amounts of adversity would predict lower scores on the observation.
Results from the analyses indicated a trend towards signiﬁcance in the
relationships between family adversity and mother-child interaction at 36 months,
but not at 24 months of age. The direction of the trend was analogous to what
was predicted. The nearly signiﬁcant path estimate between adversity and
dyadic interaction at the child’s 36th month would suggest that as family adversity

increased scores on the mother-child interaction measure decreased.

77

This trend was also supported in the results of the threshold analysis.
Mother-child interaction scores in the low versus high risk groups were nearly
signiﬁcantly different (p = .06) at the child’s 36th month. The mean scores were
56.14 (SD = 4.78) for the low risk group and 54.22 (SD = 5.96) for the high risk
group. These ﬁndings may indicate that mother-child interaction scores were
lower in families with greater adversity; as categorized by reporting at least three
adversities (see Appendix N). Interestingly, although the path estimate between
adversity and interaction as measured at 24 months was not signiﬁcantly
different than zero, it was in the opposite direction than was found at the 36
month assessment. The mean scores in the threshold analysis, although not
signiﬁcant, were also in the opposite direction from what was described at 36
months, with a lower mean score for the low risk group (M = 53.71, SD = 6.55)
than for the high risk group (M = 55.57, SD = 6.37).

Eamilv Conﬂict and Mother-Child Interaction

Family conﬂict collected at or near the child’s 2nd and 3rd birthdays was
expected to predict mother-child interaction observed at the same point in time,
with mother-child interactions in families in which there was a great deal of family
conﬂict expected to be less positive. Results from the analyses did not support
this prediction. Family conﬂict did not directly predict mother-child interaction at
either assessment. A potential explanation for the lack of direct relationship
between family conﬂict and mother-child interaction may lie within the distribution

of the conﬂict variable. Very few mothers report higher amounts of conﬂict in
78

their families (see Table 2). The responses of the conﬂict measure ranged from
1 “strongly disagree” to 4 “strongly agree”. The mean of the conﬂict items
measured at the child’s 24th month of age was 1.76 (SD = .49) and at 36 months
was 1.83 (SD = .58) which indicates that most mothers disagreed with the
statements in the scale.

When the means of those mothers whose scores suggest that they agreed
(mildly or strongly) with the statements were examined, the conclusions were
different from those drawn by examining the structural equation models (see
Appendix 0). Sixteen of the mothers (9.6 percent) interviewed when their
children were 2 years old had scores indicative of agreement with the conﬂict
scale and by the child’s 3rd year, 24 (14.5 percent) of the mothers reported
agreement. Examining the scores across time yielded 133 mothers (80.1
percent) whose scores indicated disagreement with the conﬂict scale at both
assessments, 26 mothers (15.7 percent) who agreed at one time point, and 7
mothers (4.2 percent) who agreed with the conﬂict statements at both the 24 and
36 month assessments. Using the across time grouping to examine mean
differences in measures within the study and mother-child interaction at 36
months, signiﬁcant differences arose. Mean mother-child interaction scores at 36
months were signiﬁcantly different (F=5.36, p < .01) across conﬂict groups with
mother-child dyads in which mothers reported agreement with conﬂict also
having been observed as having less optimal interactions with one another. In

dyads in which mothers reported agreement with the family conﬂict scale at both
79

time points, interaction scores were signiﬁcantly lower (p = 48.36) than in dyads
in which the mothers disagreed with the items at one time (p = 54.73) and those
who reported disagreement at both interviews (p = 56.14). In addition to the
differences discovered in mother-child interaction, parenting stress was also
predicted by the across time conﬂict groupings. Mean parenting stress at 36
months was signiﬁcantly different (F = 5.05, p < .01) across conﬂict groups with
mothers who reported agreement with conﬂict also reporting signiﬁcantly higher
parenting stress ()1 = 59.86) as compared to the mothers who disagreed with the
items at both times (p = 44.84) and those who reported disagreement at one time
point (p = 48.67).
Program Dosage

Program dosage was hypothesized to directly affect mother-child
interaction assessed at or near the child’s 2nd and 3rd birthdays as it was a
primary goal of the Jackson EHS program to improve parent-infant interaction
from the program’s inception. An unexpected ﬁnding was that program dosage
did not predict mother-child interaction as observed when the child was near two
or three years old. Program dosage also did not relate to other variables within
the model, including the number of adversities reported by mothers at enrollment.

One possible reason for this unexpected result could be related to the
distribution of the dosage construct. As described in the measures, the program
dosage variable was a continuous variable that represents the total number of

hours that families in the EHS program received services. As such, the program
80

dosage variable existed for only half of the families in the study. The randomized
comparison group received no EHS services and was assigned zeros for this
construct. Perhaps, the inclusion of the comparison group in the analyses
diminished the effects of the EHS program as the comparison group was also
free to seek services from other agencies within the community and as such
families within the comparison group may had received similar services within
their community that might account for changes in mother-child interaction across
time. Unfortunately, there was insufﬁcient power to test models for the program
and comparison groups separately.

Examining the amount of program delivered to the families assigned to the
EHS program group was done to ascertain possible effects of EHS program
services independent of the comparison group. Correlations between the total
number of hours in the program and other variables within the model, while
controlling for family baseline adversity, were computed (see Appendix P).
Results from this set of analyses suggest a relationship between support seeking
behaviors reported by moms with a 2 year old and amount of hours in the
program (r=.26, p < .05). This was the only signiﬁcant ﬁnding demonstrated for

the families assigned to Early Head Start programming.

81

CHAPTER 5
DISCUSSION

The goal of this study was to elucidate the nature of mother-infant
interactions in low-income, early intervention eligible families while focusing on
stress and coping processes within the family. The sources of stress that were
examined in this study, family adversity, parenting stress, and family conﬂict,
were hypothesized to directly impact mother-child interaction. The coping
constructs that were examined included support seeking behaviors and cognitive
reframing strategies. These mechanisms for coping were suspected to moderate
the relationships between the stress constructs and mother-child interaction. The
models tested in this investigation provided a number of insights into the nature
of these processes as they occur within the family.
_F_a_milv Adversity aﬂPradictor of§pecmc Stress gld Mother-Child Interaction

Contrary to prediction, global family adversity did not predict measures of
parenting stress, family conﬂict, or mother-child interaction collected at or near
the child’s 2nd and 3rd birthdays. This ﬁnding is counter to the models of
parenting as ascribed by Belsky (1984) and from other models of cumulative risk,
both medical (McEwen, 1998) and psychological (McCubbin & Patterson, 1983).
This ﬁnding is also contradictory to what has been found in other studies of
families eligible for Early Head Start and other limited-income populations

(Brooks-Gunn & Liaw, 1994; Seabrook, 2000) which demonstrated that chronic

82

risk load was consequential for predicting both mother’s well-being and parenting
behavior.

Cumulative risk models for predicting outcomes are vulnerability models
which fundamentally suggest that as the number of risks to which one is exposed
increases so does the likelihood of a negative outcome (Phelps, Belsky, & Crnic,
1998). Additionally, cumulative risk models of development posit that a
combination of risk factors, rather than any single factor, can better predict
adverse developmental outcomes (Liaw & Brooks-Gunn, 1994; Rutter, 1979;
1981; 1990; 1996; Rutter, Champion, Quinton, Maughan, & Pickles, 1995; Rutter
& Quinton, 1977; 1984). Categorizing the families in this study based on amount
of family adversity resulted in identifying no mean differences between families
with more limited numbers of adversities than those with higher numbers (see
Appendix N).

Possible explanations for this ﬁnding could lie within the way in which the
construct was computed. Based on the suppositions ascribed within the models
to describe parenting behavior (Belsky, 1984), the measure of family adversity
was calculated to account for risks to the family in regards to parent, child and
environmental characteristics. Included in the construct was information
regarding family demographic characteristics, maternal health and well-being,
and possible threats to typical child development (see Table 2). These variables
were weighted equally and summed to reﬂect global family adversity which might

mask the effects of a single risk factor. Parenting a child with a developmental
83

delay might serve as a greater stressor for a mother than being unemployed, for
example, and treating these adversities as equally taxing could detract from the
importance of family stressors at the individual level.

Also potentially problematic is that this cumulative exposure construct was
developed using indicators collected at enrollment into the study. There were
several variables within the construct that could not change over time, such as
the mother’s being a teenager at the birth of the focus child or the focus child
having been born pre-term or low birth weight. There were, however, variables
for which change was potential; a mother’s having been single, having limited
school achievement, having been unemployed, and/or living at or below 100% of
poverty per national guidelines. Given that some of the variables within this
construct may have changed for families from enrollment, this might account for
the lack of prediction at later assessments. Because not all of the item-level risk
indicators were static conditions, families could have moved into and out of
different contexts between enrollment and the focus child’s 24-month and 36-
month birthday related interviews.

Apart from potential problems with the construct, this lack of effect could
also be because the sample was more homogeneous than cumulated risk could
differentiate. Every family in the sample was eligible for Early Head Start
services and as such had high levels of expected need relative to the other
families within their community. All of the families in the study were comparable

in terms of economic situation, as well, with median family income at enrollment
84

being $7,920 and nearly 100% of the families receiving at least one type of public
assistance. Given this overarching economic similarity and the eligibility
requirements for the study and program, effects of additional stressors or
adverse conditions may have been diminished by this restricted range of
participants.

This homogeneity of the sample brings up a potentially critical omission in
the study. Measuring cumulative risk alone excluded the factors related to
appraisals of the stressors that made up the adversity construct (Abidin, 1992;
Hillson & Kuiper, 1994; Patterson, 2002). Other models of parenting, such as
those posited by Abidin (1992) and Hillson and Kuiper (1994), propose that
parenting behaviors are predicted by stress experiences, but also by appraisals
of the stressor. What may be misleading about the adversity construct is that it is
impossible to ascertain whether these circumstances were actually stressful for
the mothers in the sample. Studies of economic adversity, for example, often
measure dealing with the loss of income (Conger & Elder, 1994) and not limited
income as normative for the family.

In previous work using this sample of families (McKelvey et al. 2002), a
stress construct was developed which included mothers’ reports of ability to meet
their families’ basic needs. Although median income for the families was less
than $8000 annually, mothers overwhelmingly stated that they were able to meet
their family’s needs, which included having enough money for food, clothing,

housing, utilities and needs related to their children, such as money for diapers.
85

One would expect families living on such a limited income to report difﬁculties in
meeting these basic needs, but the mother’s perception would not support that
premise. This is a piece of evidence that one could use to suggest that the
mothers in this sample might not experience the same amount of stress based
on the adversities facing their family as might be expected based on studies from
middle-income families (Prelow, Tein, Roosa, & Wood, 2000). Without
understanding whether mothers actually experienced stress related to their
family’s adversities, making assumptions about how one might cope and how
relationships within the family might be altered may be misguided.
T_he_l1ects of Speciﬁc Stresses on the Family

Parenting stress has been implicated in less positive interactions between
parents and children (Coyl, Roggman, & Newland, 2002; Nitz, Ketterlinus, &
Brandt, 1995) and in increasing conﬂict between family members (Almeida,
Wethington, & Chandler, 1999; Buehler & Gerard, 2002; Crnic & Acevedo, 1995;
Deater-Deckard & Scarr, 1996; Gelfand, Teti, & Radin Fox, 1992; Larson &
Almeida, 1999; O’Brien & Badahur, 1998; Webster-Stratton, 1990). In the
current study, parenting stress measured at or near the child’s 2nd and 3rd
birthdays predicted mother-child interactions observed at concurrent
assessments. The patterns of ﬁndings were negative and signiﬁcant and
indicated that stress in the parenting role related to less positive dyadic

interactions.

86

Parenting stress and family conﬂict were expected to correlate with one
another at both time points and this was the case in all analyses conducted.
Family conﬂict assessed at the child’s 24th and 36th month predicted parenting
stress at simultaneous collection points. The path coefﬁcient from family conﬂict
to parenting stress at 24 months was positive and signiﬁcant and the same
pattern held true when examining scores at 36 months. These patterns indicated
that increases in family conﬂict were met by increases in parenting stress, which
is consistent with existing literature as there seem to be relatively consistent
correlations between marital and family conﬂict and parenting stress (Crnic &
Acevedo, 1995; Deater—Deckard & Scarr, 1996; Gelfand, Teti, & Radin Fox,
1992)

Many studies examine family conﬂict as a result of parenting stress. For
example, Crnic and Acevedo (1995) argued that the parenting stress works to
alter the processes that operate within the family. They suggested that stress
hinders the development of positive and supportive relationships within the family
and generates greater tension among family members. In these analyses, the
relationships between the variables were supported in either direction, but path
estimates and indices of model ﬁt were stronger when conﬂict predicted
parenting stress and, theoretically, parenting stress as described by its creator
should be the endogenous construct. Parenting stress as designed by Abidin
(1990) is a construct that involves behavioral, cognitive, and affective

components. He argued that parenting stress is the result of a series of
87

appraisals made by each parent in the context of his or her level of commitment
to the parenting role (Abidin, 1992). Abidin (1992) proposed that the total stress
experienced by a parent is a combination of child and parent characteristics, and
family situational components as they relate to the person’s appraisal of his or
her family and role as a parent.

With a sample of mothers and children in Head Start, Rodgers (1998)
demonstrated effects of parenting stress on parenting behaviors. In Rodgers’
(1998) study, parenting stress had direct effects on parenting behaviors, such as
punishment, parental coldness, sensitivity, inconsistency, and rejection. Other
research has demonstrated the relationship as well. Ritchie and Holden (1988)
found that as levels of parenting stress increase not only do negative interactions
increase, but positive qualities of interactions such as physical affection
decreased. Based on ﬁndings from their study of sources of parenting stress,
Gelfand, Teti, and Radin Fox (1992) state, “highly stressed parents may lose
their ability to care for their children in a warm, sensitive, and competent manner”
(p. 262).

The ﬁndings from the present and previous studies would lead one to posit
that parenting stress has negative consequences for parent-child relationships.
What is important to remember is that parenting stress, as described by Abidin
(1992), is not a unidimensional construct. It is comprised of many dimensions,
including depression, stress as it relates to the responsibilities of raising children,

a mother’s perception of her child’s abilities relative to other children and of her
88

Yang-:1... ‘. ' .‘ :—

interactions with her child, and her rating of her own ability to parent. A single
dimension or many could account for the signiﬁcant relationship between
parenting stress and mother-child interaction. Maternal depression, for example,
has been found related to less positive mother-child interaction (Jameson,
Gelfand, Kulcsar, & Teti, 1997; McCurdy, 1995; Teti & Gelfand, 1991); however,
this relationship might also be due to the mother and child actually having less
positive interactions and the mother's being aware of such. Examining the
dimensions of parenting stress with regards to mother-child interaction could
potentially yield meaningful information for intervention programs for mothers and
young children.

In addition to parenting stress, family conﬂict was also expected to
negatively predict mother-child interaction. Results from the path analyses did
not support this prediction. Family conﬂict did not directly predict mother-child
interactions at either assessment. A potential explanation for the lack of direct
relationship between family conﬂict and mother-child interaction may lie within
the distribution of the conﬂict variable. Very few mothers report higher amounts
of conﬂict in their families (see Table 2). The mean of the conﬂict items
measured at the child’s 24th month of age and at 36 months indicated that most
mothers disagreed with the statements in the scale.

When the group of mothers whose mean scores suggested that they
agreed (mildly or strongly) with the statements was examined, the conclusions

were a little different from those drawn by examining the manifest variable
89

models (See Appendix O). Examining agreement across time yielded very few
mothers (15.7 percent) who agreed with the conﬂict scale at one time point, and
even fewer (4.2 percent) who agree with the statements at both the 24 and 36
month assessments. Using the across time grouping to examine mean
differences in measures within the study and mother-child interaction at 36
months, signiﬁcant differences arose. Mean mother-child interaction scores at 36
months was signiﬁcantly different across conﬂict groups with mother-child dyads
in which mothers reported agreement with conﬂict also being observed as having
less optimal interactions with one another. No mean differences were identiﬁed
in interactions between mothers and 24 month old children. Like what was
discovered in the path models, parenting stress was also predicted by the across
time conﬂict groupings. Mean parenting stress at 36 months was signiﬁcantly
different across conﬂict groups with mothers who report agreement with conﬂict
also reporting signiﬁcantly higher parenting stress as compared to the mothers
who disagreed with the items at both times and those who reported agreement at
one time point.

The ﬁnding that family conﬂict may effect mother-child interactions has
been supported by other studies with limited-income (Coyl, Roggman, &
Newland, 2002; McKelvey et al. 2002) and middle-income samples (Fisher,
Fagot, 8. Leve, 1998; Garabino, 1976; Gomel et al. 1998; Ostberg & Hagekull,
2000; Peterson & Hawley, 1998; Simons, Whitbeck, Conger, & Melby, 1990).

The ﬁndings of the path analyses were not consistent with the analyses that
90

v_--‘
-0
..—‘

\g": ‘7‘-

compared the families in which mothers reported agreement with the conﬂict
items. Consistently supported by both sets of analyses was the relationship
between parenting stress and family conﬂict. It is important to note that family
conﬂict, taken alone, did not signiﬁcantly predict mother-child interactions, but the
relationships between conﬂict and dyadic interchanges were stronger when
parenting stress was excluded from the model. This might suggest a mediating
effect of parenting stress on the relationship between conﬂict and mother-child
interactions.

Support Seeking — Direct and Moderating Effects

Use of support seeking coping strategies directly impacted mother-child
interaction. The patterns of signiﬁcant path estimates suggest that mother-child
interaction at 24 months was not predicted by support seeking, while the same
measure at 36 months of child’s age was signiﬁcantly predicted by both the 24
and the 36 month support seeking behavior coping scores. What was most
notable in the ﬁndings was that support seeking behaviors at 24 months
positively predicted mother-child interaction at 36 months, and support seeking at
36 months of age negatively predicted interaction at the same age.

This negative relationship between 36 month support seeking and dyadic
interaction appeared to be a suppressor effect (see Appendix I). The zero order
correlation for these variables was small and positive, however, when the shared
variance in support seeking at the two time points was controlled, the partial

correlation between the two concurrently assessed variables, albeit non-
9i

signiﬁcant, became negative. Thus, seeking support when the child was younger
predicted more positive mother-child interaction when the child was older and not
at the concurrent time. This was an interesting ﬁnding which indicated that there
may be some lag time between seeking supports from others to help cope with
problems that face one’s family and impacts on the family itself, or at least
interactions between mothers and toddlers.

This lagged effect may tie into research regarding the ways in which
socially supportive relationships may foster one’s psychological health and sense
of mastery or self-efﬁcacy (see Sarason, Sarason, & Pierce, 1990). Mastery, as
a concept, suggests a belief about one’s own abilities to successfully engage in
activities to cope with stress (Green & Rodgers, 2001). Sarason and colleagues
(1990) purported that the feeling of being supported is crucial to the development
of one’s sense of mastery and research with low-income families supports the
tenet (Green & Rodgers, 2001). Unfortunately, mastery was not considered in
the model, so initial levels of the mothers’ feelings of competency regarding
support seeking were unknown. This could be an interesting concept to include
in forthcoming work with limited-income families. Within the current study, coping
was hypothesized and examined as a result of stress and not as related to one’s
sense of competence regarding their ability to garner support.

Support seeking coping strategies were expected to moderate the
relationships between family stressors and mother-child interactions. Many of

the moderation effects between predictors and support seeking were not related
92

to mother-child interaction. No signiﬁcant moderation effects of support seeking
were demonstrated for 1) dosage predicting mother-child interaction at 24 and 36
months of age, 2) adversity at both collection points, 3) parenting stress at 36
months, or 4) family conﬂict at 36 months. The nearly signiﬁcant effects that
were related to mother-child interaction when the child was at or near 2 years of
age were parenting stress with support seeking at 2 years and family conﬂict with
support seeking at 2 years. These two trends could be determined as signiﬁcant
with a one-sided hypothesis test and were in the direction of prediction based on
previous work and coping theory.

Because much of the theory about coping has been developed based on
work with middle-income families (Prelow, Tein, Roosa, & Wood, 2000) and
speciﬁcally related to illness (Skinner et al., 2003), interpreting the interaction
effect of parenting stress and support seeking should be done with caution. The
direction of the relationship would appear to indicate that for families with lower
amounts of parenting stress, support seeking did little to impact mother-child
interaction scores. For the high stress parents, however, support seeking did
impact mother-child interaction in that within that group of mothers, those who
sought support from others had higher mother-child interaction scores than those
mothers who do not report seeking as much support. The same pattern was
demonstrated with the family conﬂict and support seeking moderation effect. For
families who reported lesser conﬂict, support seeking strategies did not seem to

impact mother-child interaction and in families in which mothers reported higher
93

amounts of conﬂict, those who also sought support from others tended to have
higher scores on the mother-child interaction measure.

Just as the results from these analyses were not consistent regarding the
moderating effects of support, the existing literature was also not consistent. For
example, studies had shown that social support moderated the effects of
parenting stress on discipline behaviors, parental warmth, and sensitivity (Crnic
et al. 1983; Deater-Deckard & Scarr, 1996; Kotch, Browne, Ringwalt, Dufort,
Ruina, Stewart, & Jung, 1997; Rodgers, 1998) and mother-child interaction
(Crnic & Greenberg, 1990; Weinraub & Wolf, 1983). Other studies, such as
those by Quittner, Glueckauf and Jackson (1990) and Pearlin and Schooler
(1978) failed to ﬁnd moderating effects of support.

If one was to interpret the trends in the moderating effects of support
seeking, why would it be that the relationship between stress and mother-child
interactions were different across time? Pearlin and Schooler (1978) failed to
ﬁnd a moderating effect of support and found that self-reliance as a coping
strategy was more efﬁcacious than support from others in reducing the stress
associated with parenting and marriage (Pearlin & Schooler, 1978). Quittner,
Glueckauf and Jackson (1990) concluded that social support functions differently
in chronic versus speciﬁc stress conditions and suggested that increased support
may be helpful for short-term stressors, while support in the context of chronic
conditions may be appraised as “intrusive” or “suggestive of incompetence”

(p. 1276).
94

The support measure in this study was qualitatively different than in the
coping studies mentioned above. The measure of support as utilized in the
current study was the active seeking of support and as such did not entail the
quantity, quality (usefulness), or type of support sought. It may be that the more
support that was sought by a mother coping with what she construed to be a
problem for her family yielded a greater likelihood of encountering support that
was appraised as useful when her child was younger versus older. It might also
have been that mothers sought different types of support when parenting a two
year old versus when parenting a three year old. If the support being sought was
informational or tangible, not gaining ﬁnancial self-reliance over the course of a
year might have brought a negative aspect to support seeking. Parenting advice,
for example, when parenting a younger child might be more welcome than as the
child ages as the likelihood for the support to be suggestive of incompetence or
intrusive is greater. It could also have been that the quality of the support being
sought diminished across time. Further still, earlier supportive relationships may
have increased the mothers’ sense of mastery at dealing with stressful situations
for her family, such that seeking support when her child was older was less
efﬁcacious for the family or even appraised as negative.

This concept of mastery may play a larger part in understanding coping
process. A base tenet for all of the models of parenting reviewed is that stress
weakens the family system, making positive family interactions more and more

difﬁcult to achieve. Omitted from these models, this included, is the role that
95

stress may play in strengthening the system. Successfully dealing with stress or
surviving in a situation that one has appraised as stressful may work to build
one’s sense of competency, but even more, might work to alter the initial
appraisals of circumstances that would once have been considered stressful.
The mothers in this sample have been demonstrated to score very highly on their
perceived ability to meet their families’ basic needs, despite a median income of
$667 monthly (McKelvey et al. 2002). This could suggest that successfully living
with stressful situations has served to alter the appraisal of the situation to be 1)
less stressful than it might have been in the past or potentially not at all stressful
and 2) a source of esteem or mastery building.

In addition to the relationships between support seeking and mother-child
interaction, there was also a positive relationship found between the coping
constructs. Support seeking and cognitive reframlng were positively and
signiﬁcantly related at both 24 and 36 months; mothers who reported more highly
utilizing cognitive reframing strategies also reporting using seeking more support
from others. While support seeking coping directly impacted mother-child
interaction; this effect was not demonstrated for cognitive reframing strategies.
Cpgnitive Reframing - Direct and Moderating Effects

According to the results of the direct effects of coping model, cognitive
reframlng coping strategies did not directly impact mother-child interaction. As
demonstrated with support seeking, the dosage moderated by cognitive

reframlng interaction terms did not predict mother-child interaction. One
96

cognitive reframlng moderating effects signiﬁcantly and another nearly
signiﬁcantly predicted mother-child interaction when the child was at or near 2
years of age; parenting stress at 2 years and family adversity at 2 years,
respectively. There was also a signiﬁcant prediction of parenting stress at 2
years by the interaction of family conﬂict and cognitive reframing at the same
time.

The relationship between parenting stress and mother-child interaction was
moderated by cognitive reframing. For families with low amounts of parenting
stress, mother-child interaction scores were better for mothers who did more
cognitive reframlng. For the high parenting stress mothers, however, there was a
negative relationship between cognitive reframlng and mother-child interaction in
that within that group of mothers, those who did more cognitive reframing had
lower mother-child interaction scores than those mothers who did less cognitive
reframlng. This relationship was in the opposite direction from what was
predicted, but partially mimics the ﬁndings from another study of stress and
coping using this sample. McKelvey and colleagues (2002) examined the
mediating role of coping on the relationship between stress and mother-child
interaction. Findings from that study found a direct relationship between stress
and coping, but coping did not directly impact a mother’s interaction with her
infant. Speciﬁcally, mothers who perceived lower stress reported seeking more

support from friends and family and engaging in more cognitive reframing

97

techniques, but mothers who were experiencing greater life difﬁculties were not
utilizing these coping mechanisms in the same manner.

One explanation to consider may be that there is a threshold for stress and
those mothers who were experiencing the greatest difﬁculty were under too much
stress to use the same number of coping resources as the lower stress mothers.
Peterson and Hawley (1998) found a threshold of three stressors for stress to
impact family functioning and parenting attitudes. They found families with fewer
than three stressors reported signiﬁcantly healthier attitudes towards parenting,
more family cohesion and less family conﬂict. Families with no reported
stressors scored signiﬁcantly better than higher stress groups on measures of
empathic parenting and negative attitudes toward physical punishment. The
Peterson and Hawley study did not directly examine coping, however it may be
that increased stress makes coping more difﬁcult to accomplish as well.

Like with the interpretation of the trends found in support seeking, this
potential moderating effect should be interpreted with caution. The interaction of
family adversity and cognitive reframlng nearly signiﬁcantly predicted mother-
child interaction assessed at or near the child’s 24"1 birthday. This interaction
may indicate that for families with both low and high amounts of family adversity,
mother-child interaction scores were better for mothers who did more cognitive
reframing. For mothers in families with greater adversity, there was a stronger
relationship between cognitive reframing and mother-child interaction scores than

for mothers in families with lesser adversity.
98

The interaction of family conﬂict and cognitive reframlng signiﬁcantly
predicted parenting stress at or near the child’s 24‘“ birthday. This interaction
indicates that for families with both low and high amounts of family conﬂict,
parenting stress scores were lower for mothers who did more cognitive
reframing. For mothers in families with greater conﬂict, however, there was a
weaker relationship between cognitive reframlng and parenting stress scores
than for mothers in families with greater conﬂict. There was a direct relationship
between parenting stress and cognitive reframlng, in which mothers who
reported more parenting stress also reported lesser use of cognitive reframing
strategies at 24 months and at 36 months of age. This may be due to the
cognitive effort that this particular coping strategy necessitates. Furthermore, the
measure of parenting stress was not unidimensional (Reitman, Currier, & Stickle,
2002) and included items regarding the mother’s perceptions of the child’s
developmental ability and of having dysfunctional interactions with her child, as
well as the mother’s distress, generally and speciﬁcally as related to the
responsibilities of parenting (Abidin, 1990). Mother’s general distress, which
were items originally drawn from the Parenting Stress Index - Long Forrn’s
depression subscale (Abidin, 1990) may be related to a greater inability to
cognitively reframe sources of stress.

Like with the results of the support seeking coping analyses, there were
inconsistencies in the prediction of mother-child interaction across time.

Interestingly, all of the modeled autoregressive relationships, those that were
99

among the same construct across time, were signiﬁcant except one. Mother-
child interaction observed at or near the child’s 24th month did not signiﬁcantly
predict the mother-child interaction when the child was near 36 months of age.
This ﬁnding was contrary to prediction and had not been reported in other studies
using this measure and might account for the differential impacts across time.
Possible reasons for this lack of prediction could include the developmental
changes that occur within the dyad across time or the measure itself.

The interaction measure, the NCAST Teaching Scale, was an observation
in which mother and child were scored on their behaviors as the mother
attempted to teach the child a task at which competency had not yet been met.
Common activities selected for the 24 month old child included teaching the child
to balance on one foot, or to pull a zipper up and down, or to button a button.
Common teaching tasks chosen for a 36 month old child included teaching the
child to draw a shape with a crayon, sorting blocks by color or arranging by
number. Although the choice of task was not addressed as a signiﬁcant variable
in the scoring of the instrument, a greater number of tasks for the younger child
were motor coordination tasks, while those for the 3 year old child were more
educational or cognitive in nature (Sumner & Spietz, 1994).

Although the test-retest reliabilities as discussed by the instrument’s
creators were relatively high, they were computed based on the scores of 30
children when they were aged 1, 4, 8, and 12 months and were not directly re-

assessed as the child ages into toddlerhood (Sumner & Spietz, 1994). This lack
100

of predictability in interactions from 24 months to 36 months may have made a
signiﬁcant impact on the conclusions drawn from this study. Many of the
predictors of mother-child interaction (apart from parenting stress) were
signiﬁcant when the child was 2 versus 3 or vice versa. The lack of reliability of
the mother-child interaction variable from one time point to another might account
for the lack of uniformity across ﬁndings.

Program Impacts

Program dosage was hypothesized to directly affect mother-child
interaction assessed at or near the child’s 2nd and 3rd birthdays. From the
program’s inception, a primary goal of the Jackson EHS program was to improve
parent-infant interaction. Effects of the program on mother-child interaction had
been demonstrated in previous studies with this sample based upon the intent to
treat model (Van Egeren, McKelvey, Schiffman, & Fitzgerald, 2002). Van Egeren
and colleagues (2002) demonstrated differences in interactions across time
comparing the randomly assigned program and comparison families to one
another without consideration of the amount of program services received.

An unexpected ﬁnding in the current study was that program dosage did
not predict mother-child interaction as observed when the child was near two or
three years old. Dosage also did not relate to other variables within the model.
One possible reason for this unexpected result is, as described in the measures,
the program dosage variable was a continuous variable that represents the total

number of hours that families in the EHS program received services. As such,
IOI

the program dosage variable existed for only half of the families in the study.
The randomized comparison group received no EHS services and was assigned
zeros for this construct. Perhaps, the inclusion of the comparison group in the
analyses diminished the effects of the EHS program as the comparison group
was also free to seek services from other agencies within the community and as
such families within the comparison group may had received similar services
within their community that might account for changes in mother-child interaction
across time.

Another unexpected ﬁnding was that program dosage was not signiﬁcantly
related to the number of adversities reported by mothers at enrollment.
Although, similar to the effects of program dosage on mother-child interaction, it
was possible that the inclusion of the comparison group in the analyses also
masked the relationship between adversity and program dosage, this lack of
relationship had been found in other studies of early interventions (Barton, 2002;
Berlin, O’Neal, Brooks-Gunn, 1998). Families characterized as having high risk
and thus need for intervention may have had differential experiences of the
program. According to Berlin and colleagues (1998), families with greater
amounts of adversity may be harder to engage or more difﬁcult to reach or may
be more demanding of attention from intervention staff. Given the combination of
these two very different responses to intervention services, the effects of amount

of adversity on program dosage may be disguised. Guided by the conclusions

l02

drawn from the adversity analyses, it may also be that the families do not
perceive themselves as in need of services.

Examining the amount of program delivered to the families assigned to the
EHS program group was done to ascertain possible effects of EHS program
services independent of the comparison group. Correlations between the total
number of hours in the program and other variables within the model, while
controlling for family baseline adversity, were computed (see Appendix P).
Results from this set of analyses suggest a relationship between support seeking
behaviors reported by mothers with a 2 year old and amount of hours in the
program. This was the only signiﬁcant ﬁnding demonstrated for the families
assigned to Early Head Start programming. It was a potentially important effect,
however, given the direct impact of support seeking behaviors when parenting a
2 year old on interactions between mothers and 3 year old children.

The lack of effects of EHS program dosage was puzzling, especially given
earlier ﬁndings, based upon this sample, which demonstrated effects of random
assignment on mother-child interaction (Van Egeren et al. 2002). These results
were also demonstrated with the national EHS sample (of which these families
were a part) that supported differences between the program and comparison
groups on measures of mother-child interaction and other indices of parenting
(US. Department of Health and Human Services, 2002). What was not
examined in the current study with regards to program dosage was a family’s

readiness to engage in the program. Researchers (Barton, 2002; Vivian &
103

Wilcox, 2000) had found a relationship between program intensity and
participant’s identiﬁcation of needs and identiﬁcation of the usefulness of
services. Within these early intervention studies, it was found that participants
who reported fewer needs also received fewer services, regardless of their needs
assessment by the program (Barton, 2002) and that families were more likely to
use services when the intervention was thought particularly useful for them
(Vivian & Wilcox, 2000). These seem to be signiﬁcant options to explore in trying
to account for the lack of program dosage effects.

Another potential complexity may lie within the program itself. No program
is exactly alike for each participant and, as such what was not examined by using
the total number of hours dedicated to the family during EHS programming was
what actually occurred within the context of the program for each family.
Collected from the program charts was information regarding the number and
types of referrals made on the family’s behalf and the number of other supports
offered by the program. A more careful examination of the differences in
program offered to each family might yield a better understanding of the program
and its potential impacts.

Strengths and Limitations

The current study sought to underline the nature of coping with stressful
life circumstances for low-income families. Particularly important were the
potentially buffering effects of coping for the maintenance of positive interactions

between mothers and young children. As coping has been studied in the context
104

of illness and for middle-class majority families (Prelow et al. 2000; Skinner at al.
2003), studying coping for low-income families helps to inform the literature about
how these processes may differ across socioeconomic groups.

One strength of the current study was the longitudinal nature of the
design. Understanding how processes in the family may change across time is
crucial for the development of successful interventions and scientiﬁc studies.
Without having data available from both the child’s second and third years,
conclusions could have been drawn that would not have been supported when
the child aged. The longitudinal nature of the study yielded information regarding
a lagged effect of support seeking on mother-child interaction, which would not
have been known in a cross sectional study. This lag time may tie into the
importance of including mastery in future models of determining the usefulness of
coping for buffering stress. Another important ﬁnding in this study that was
determined based on the longitudinal nature of the data was the ﬁnding that
mother-child interaction at the child’s 24th month did not predict mother-child
interaction at 36 months of age.

Unfortunately, the size of the sample limited the number of constructs that
could be simultaneously examined. Examining support seeking and reframlng
behaviors separately did not allow for the modeling of shared variance in the
constructs, for example. The smaller sample also precluded the addition of
constructs to the model that might have helped yield explanatory power to the

analyses. With such a limited number of individuals, examining all facets of the
105

parenting stress construct could have been done, but at the omission of other
constructs within the models.

A number of variables should be examined in future studies of stress and
coping. With relative ease, research scientists and practitioners have determined
eligibility for programming based on a plethora of environmental or contextual
characteristics of the family. In this sample, those family characteristics did not
appear to be related to problematic mother-child interactions. Perhaps,
consideration of these risks independently might yield different conclusions, but
taken as a whole, adversity did not relate to parenting stress, family conﬂict,
coping or mother-child interactions. Personal appraisals of these family
contextual variables were not included in the current study. The appraisals of
stress conditions could be important for determining not only mother-child
interaction, but use and acceptance of early intervention services. Previous work
with early interventions (Barton, 2002; McKelvey et al. 2002) has described a
schism between the needs of low-income families as expected by interventionists
and researchers and as perceived by the families themselves. This lack of
concordance between the appraisal of the stressor and the assumption about
expected stress is a key concept which was not considered in the current study.
Future studies should consider the event about which one would evoke
strategies for coping and to be certain that it was indeed stressful for the

respondent.

106

One’s feelings of competency about dealing with problems as they arise
might yield insights into the longer term effects of coping and support. Within the
current study, coping was considered a reaction to stress, but examining the
ways in which successfully or unsuccessfully coping with stress in the past
impacts coping strategies for present and future stressors was left unexamined.
These coping constructs could be thought of as preceding the stressful situation
as well. Green and Rodgers (2001) suggest social support is important for
building one’s sense of control over one’s life, which could have impacts on the
types of coping strategies used in parenting context.

Summam and Closing Comments

The current study sought to understand the relationships between stress,
coping, and mother-child interactions when parenting a child at age two and age
three. Changes in relationships among target variables differed across time,
perhaps because the children and their families had developed, because of
experiencing differing stressors across time, or because coping strategies for
dealing with stress are differentially effective when parenting a two versus a three
year old. The single consistent predictor of mother-child interaction was
parenting stress. This construct negatively impacted mother-child interaction at
both 24 and 36 months. Findings from the path analyses suggest that reports of
family conﬂict did not directly impact dyadic interactions, but did inﬂuence
mothers’ reports of parenting stress, in that mothers who reported more conﬂict

experienced greater amounts of parenting stress. A post hoc analysis of the
107

.I_ “W‘—

conﬂict construct demonstrated negative impacts on mother-child interaction
directly when examining the mothers who reported agreeing that there was
conﬂict in her family. Interestingly, family adversity did not predict parenting
stress, family conﬂict or mother-child interactions as assessed at either age of
the child.

The effects of coping demonstrated in this study appeared to support their
direct effects more consistently than their moderating inﬂuences. An interesting
result from the study was in ﬁnding a greater percentage of variance explained
for mother-child interaction at 24 months when considering the role of cognitive
reframing, while support seeking explained a greater percentage of variance in
the same measure at 36 months. Support seeking coping reported at 24 and 36
months demonstrated direct effects on mother-child interaction at 36 months,
unexpectedly in different directions. Although the negative relationship between
interactions and support seeking at 36 months was deemed spurious, it was
quite unpredicted to ﬁnd a lagged effect of support seeking at 24 months on the
same outcome. This ﬁnding may suggest some other process at work for
building a sense of mastery or self-efﬁcacy that should be examined in future
studies. Cognitive reframing, on the other hand did not directly predict mother-
child interaction. What was demonstrated was a relationship between parenting
stress and cognitive reframlng in that heightened stress related to lesser
cognitive reframlng. lntriguingly, parenting stress and cognitive reframing

interacted to signiﬁcantly predict mother-child interactions (all assessed at 24
108

months). What was most notable was that parents who were more stressed and
who utilized more cognitive reframing were observed to have less positive dyadic
interactions. This may indeed be a rather signiﬁcant ﬁnding to be further
explored, if one also considers the ways in which coping ineffectively could have
potentially damaging effects on relationships within the family.

There were no demonstrated impacts of the total number of EHS program
hours devoted to the families on mother-child interaction. There are multiple
reasons for this lack of effect. They stem from logistic and measurement issues,
to potential differences in willingness to accept support. Not only should the
individual differences in families assigned to EHS be further explored, but the
programmatic differences between families should also be considered. A more
thorough examination of data concerning families’ perceptions of their own
needs, family engagement in the program and perhaps variables such as staff
turnover within family are all potential predictors of differential outcomes for
mothers and children.

In sum, the family processes surrounding stress and coping and how they
may interact to predict mother—child relationships are rich and complex and
deserve further exploration. Investigating constructs such as stress appraisals,
self-efﬁcacy or mastery, parental perceptions of need and engagement in early
intervention will help develop a better understanding of family stress, coping, and
program experiences and how they work to predict parenting and mother-child

relationships. Continuing to explore these processes and how they may change
109

for a family when parenting children of differing ages could serve to inform
programs much like that examined within the pages of this study as to provide

the most beneﬁcial services for those in need.

110

REFERENCES

Abidin, R. (1990). Parenting Stress Index/Short Form. Charlottesville, VA:
Pediatric Psychology Press.

Abidin, R. (1992). The determinants of parenting behavior. Journal of Clinical
Child Psychology, 21, 407—412.

Acock, A., (1997). Working with missing data. Family Science Review, 10, 76-
102.

Almeida, D., Wethington, E., & Chandler, A. (1999). Daily transmission of
tensions between marital dyads and mother-child dyads. Journal of Marriage
and the Family, 61, 49-57.

Barnard, K. E. (1997). Inﬂuencing mother-child interactions for children at risk.
In M. J. Guralnick (Ed.), The effectiveness of earty intervention (pp. 249-268).
Baltimore, MD: Paul H. Brookes.

Barnard, K., & Kelly, J. (1990). Assessment of mother-child interaction. In S.
Meisels and J. Shonkoff (Eds.), Handbook of early childhood intervention.
(pp. 278-302). New York: Cambridge University Press.

Baron, R.M., & Kenny, DA. (1986). The moderator-mediator variable distinction
in social psychological research: Conceptual, strategic and statistical
considerations. Journal of Personality and Social PsychologY, 51 , 1173-

1 182.

Barratt, M. S., & Roach, M. A. (1995). Early interactive processes: Parenting by
adolescent and adult single mothers. Infant Behavior and Development, 18,
97-109.

Bar-Tal, Y., Cohen-Mansﬁeld, J., & Golander, H. (1998). Which stress matters?
The examination of the temporal aspects of stress. The Journal of
Psychology, 132, 569-576.

Barth, R., Hacking, S., & Ash, J. (1988). Preventing child abuse: An

experimental evaluation of the Child Parent Enrichment Project. Journal of
Primary Prevention, 8, 201 -21 7.

III

Barton, L. M. R. (2002). Reported needs of low-income mothers: Impact on
service utilization and intervention program outcomes. Dissertation Abstract
International, 63, 6118.

Beck, A. T. (1987). Cognitive models of depression. Journal of Cognitive
Psychotherapy, 1, 5-38.

Belsky, J. (1984). The determinants of parenting: A process model. Child
Development, 55, 83-96.

Belsky, J. (1993). Etiology of child maltreatment: A developmental-ecological
analysis. Psychological Bulletin, 1 14, 413 - 434.

Belsky, J., Youngblade, L., Rovine, M., & Volling, B. (1991). Paterns of marital
change and mother-child interaction. Joumal of Marriage and the Family, 53,
487-498.

Benzies, K. M., Harrison, M. J., & Magill-Evans, J. (1998). Impact of marital
quality and parent-infant interaction on preschool behavior problems. Public
Health Nursing, 15, 35-43.

Biederrnan, J., Milberger, S., Faraone, S. V., Kiely, K., Guite, J., Mick, E.,
Ablon, S., Warburton, R., & Reed, E. (1995). Family—environment risk factors
for attention-deficit hyperactivity disorder: A test of Rutter’s indicators of
adversity. Archives of General Psychiatry, 52, 464-70.

Berlin, L.J., O’Neal, 0., & Brooks-Gunn, J. (1998). What makes early
intervention programs work? The program, its participants ,and their
interaction. Zero to Three, 18, 4-15.

Billings, A., & Moos, R. (1982). Stressful life events and symptoms: A
longitudinal model. Health Psychology, 1, 99-118.

Black, M., Nair, P., Kight, C., Wachtel, R., Roby, P., & Schuler, M. (1994).
Parenting and early development among children of drug-abusing women:
Effects of home intervention. Pediatrics, 94, 440-448.

Blair, C., Ramey, C., & Hardin, J. (1995). Early intervention for low birthweight,
premature infants: Participation and intellectual development. American
Journal on Mental Retardation, 99, 542-554.

Bollen, K.A., & Long, J.S. (Eds), (1993). Testing structural equation models.
Thousand Oaks, CA; Sage.

112

Boomsma, A. (2000). Reporting analyses of covariance structures. Structural
Equation Modeling, 7, 74-81.

Brayden, R., Altemeier, W., Tucker, D., Dietrich, M., & Vietz, P. (1992).
Antecedents of child neglect in the ﬁrst two years of life. Journal of
Pediatrics, 120, 426-429.

Brito, C., & Pearl, J. (2002). A new identiﬁcation condition for recursive models
with correlated errors. Structural Equation Modeling, 9, 459.474.

Bronfenbrenner, U. (1986). The ecology of the family as a context for human
development. Developmental Psychology, 22, 723-742.

Brooks-Gunn, J., & Duncan, G. (1997). The effects of poverty on children. The
Future of Children (Vol. 7), Children and Poverty, 55-71.

Brooks-Gunn, J., Leventhal, T., & Duncan, G. J. (2000). Why poverty matters
for young children: Implications for policy. In J. D. Osofsky & H. E. Fitzgerald
(Eds), WAIMH handbook of infant mental health: Vol. 3. Parenting and child
care (p. 89-131). New York: Wiley.

Brophy-Herb, H., Schiffman, R., McKelvey, L., & Cunningham-DeLuca, M.
(2001). Quality improvement: Lessons learned from an Infant Mental Health
based Early Head Start program. Infants and Young Children, 14, 77-85.

Brown, M., & Cudeck, R. (1993). Alternative ways of assessing model ﬁt. In K.
Bollen & J. Long (Eds), Testing Stmctural Equation Models (pp. 136-162).
London: Sage.

Browne, K., & Saqi, S. (1988). Approaches to screening for child abuse and
neglect. In K. Browne 8 C. Davies (Eds), Early prediction and prevention of
child abuse (pp. 57 - 85). Chichester, England: Wiley.

Buehler, C., Anthony, C., Krishnakumar, A., Stone, G., Gerard, J., & Pemberton,
S. (1997). lnterparental conﬂict and youth problem behaviors: A meta-
analysis. Journal of Child and Family Studies, 6, 233-247.

Buehler, C., & Gerard, J. M. (2002). Marital conﬂict, ineffective parenting, and
children’s and adolescents’ maladjustment. Joumal of Marriage and Family,
64, 78-92.

Campbell, T. (1994). Impact of prenatal/early infancy home visitation on family
health. Family Systems Medicine, 12, 81-85.

113

Censullo, M. (1994). Strategy for promoting greater responsiveness in
adolescent parent/infant relationships: Report of a pilot study. Journal of
Pediatric Nursing, 9, 326-332.

Cicchetti, D., & Lynch, M. (1993). Toward and ecological/transitional model of
community violence and child maltreatment: Consequences for children’s
development. Psychiatry, 56, 96 - 118.

Cicchetti, D., & Rizley, R. (1981). Developmental perspectives on the etiology,
intergenerational transmission and sequelae of child maltreatment. New
Directions for Child Development, 11, 31 - 55.

Conger, R. D., Conger, K. J., Elder, G. H., Jr., Lorenz, F. O., Simeons, R. L., &
Whitbeck, L. B. (1992). A family process model of economic hardship and
adjustment of early adolescent boys. Child Development, 63, 526-41.

Conger, R. D., & Elder, G. (1994). Families in troubled times: Adapting to change
in rural America. Hillsdale, NJ: Aldine de Gruyter.

Conger, R., Ge, X., & Lorenz, F. (1994). Economic stress and marital relations.
In R. Conger, & G. Elder (Eds), Families in troubled times: Adapting to
change in rural america (pp. 187-203). Hillsdale, NJ: Aldine De Gruyter.

Conger, R., Lorenz, F., Elder, G., & Simons, R. (1993). Husband and wife
differences in response to undesirable life events. Joumal of Health and
Social Behavior, 34, 71-88.

Conger, R., McCarty, J., Yang, R., Lahey, B, & Kropp, J. (1984). Perception of
child, child-rearing values, and emotional distress as mediating links between
environmental stressors and observed maternal behavior. Child
Development, 55, 2234-2247.

Cowen, P. (2001). Effectiveness of a parent education intervention for at risk
families. Joumal of the Society of Pediatric Nurses, 6, 73-88.

Coyl, D., Roggman, L., & Newland, L. (2002). Stress, maternal depression, and
negative mother-infant interactions in relation to infant attachment. Infant
Mental Health Journal, 23, 145-163.

Crnic, K., Greenberg, M., Ragozin, A., Robinson, N., & Basham, R. (1983).

Effects of stress and social support on mothers and premature and full-term
infants. Child Development, 54, 209-217.

114

Crnic, K., 8 Acevedo, M. (1995). Everyday stresses and parenting. In Marc
Bornstein (Eds), Handbook of Parenting: Volume 4, Applied and Practical
Parenting (pp.277-297). Mahwah, NJ; Lawrence Earlbaum Associates.

Crnic, K. A., & Greenberg, M. T. (1990). Minor parenting stresses with young
children. Child Development, 61 , 1628-1637.

Cummings, E. M., & Davies, P. T. (1994). Maternal depression and child
development. Joumal of Child Psychology and Psychiatry, 35, 73 - 112.

Curran, P. J., West, S. G, & Finch, J. F. (1996). The robustness of test statistics
to nonnorrnality and speciﬁcation error in conﬁrmatory factor analysis.
Psychological Methods, 1, 16-29.

Dawson, P., van Doornick, W., & Robinson, J. (1989). Effects of home-based,
informal social support on child health. Developmental and Behavioral
Pediatrics, 10, 63-67.

de Chateau, P. (2000). Longitudinal aspects of early parent-infant interactions
and contacts with mental health agencies. In J. D. Osofsky & H. E. Fitzgerald
(Eds), WAIMH handbook of infant mental health: Vol. 1. Perspectives on
infant mental health (pp. 63-90). New York: Wiley.

Deater-Deckard, K., & Scarr, S. (1996). Parenting stress among dual-eamer
mothers and fathers: were there gender differences? Journal of Family
Psychology, 10, 45-59.

Dix, T. (1991). The affective organization of parenting: Adaptive and maladaptive
processes. Psychological Bulletin, 110, 3-25.

Dorn, L., Burgess, E., Susman, E., von Eye, A., De Bellis, M., Gold, P., &
Chrousos, G. (1996). Response to oCRH in depressed and non-depressed
adolescents: did gender make a difference? Joumal of the American
Academy of Child & Adolescent Psychiatry, 35, 764-773.

Downey, G., & Coyne, J. (1990). Children of depressed parents: An integrative
review. Psychological Bulletin, 108, 50-76.

Duggan, A., McFarland, E., Windham, A., Rohde, C., Salkever, D., Fuddy, L.,

Rosenberg, L., Buchbinder, S., & Sia, C. (1999). Evaluation of Hawaii’s
Healthy Start Program. The Future of Children, 9, 66-90.

115

Egeland, B., Jacobvitz, D., & Papakola, K. (1987). Intergenerational continuity of
abuse. In Gelles, R., & Lancaster, J. (Eds), Child abuse and neglect:
Biosocial dimensions (pp. 255-276). Chicago: Aldine de Gruyter.

Erel, O., & Burman, B. (1995). Interrelatedness of marital relations and mother-
child relations: A meta-analytic review. Psychological Bulletin, 118, 106-132.

Feldman, M., Leger, M., & Walton-Allen, N. (1997). Stress in mothers with
intellectual disabilities. Journal of Child and Family Studies, 6, 471 -485.

Felitti, V., Anda, R., Nordenberg, D., Williamson, 0., Spitz, A., Edwards, V., Koss,
M., & Marks, J. (1998). Relationship of childhood abuse and household
dysfunction to many of the leading causes of death in adults: The Adverse
Childhood Experiences study. American Joumal of Preventive Medicine, 14,
245-258.

Fergusson, D. M., Lynskey, M. T., & Horwood, L. J. (1993). The effects of
maternal depression on maternal ratings of child behavior. Journal of
Abnormal Child PsychologY, 21, 245—69.

Field, T. (1984). Early interactions between infants and their post-partum
depressed mothers. Infant Behavior and Development, 7, 527-532.

Field, T., Widmayer, S., Adler, S., & de Cubas, M. (1990). Teenage parenting in
different cultures, family constellations, and caregiving environments: Effects
on infant development. Infant Mental Health Journal, 11, 158-174.

Fisher, P., Fagot, B., & Leve, C (1998). Assessment of family stress across low-,
medium-, and high-risk samples using the Family Events Checklist. Family
Relations, 47, 215-219.

Fitzgerald, H. E. & Barton, L. R. (2000). Infant Mental Health: Origins and
emergence of an interdisciplinary ﬁeld. In J. D. Osofsky & H. E. Fitzgerald
(Eds), World Association for Infant Mental Health (WAIMH) Handbook ____of

Infant Mental Health Volume1: Perspectives on Infant Mental Health (pp. 2-
36). New York: John Wiley&Sons.

Fitzgerald, H. E., Puttler, L.l., Mun, E. Y., & Zucker, R. A. (2000). Prenatal and
postnatal exposure to parental alcohol use and abuse. In J. D. Osofsky & H.
E. Fitzgerald (eds). WAIMH handbook of infant mental health. Vol. 4: Infant
mental health in groups at high risk. (pp. 123-160). New York: Wiley.

116

Fitzgerald, H. E., Zucker, R. A., & Yang, E. Y. (1995). Developmental systems
theory and alcoholism: Analyzing patterns of variation in high risk families.
Psychology of Addictive Behaviors, 9, 8-22.

Flanagan, P. (1998). Teen mothers: Countering the myths of dysfunction and
developmental disruption. In C. Garcia-Coll, J. Surrey, and K. Weingarten
(Eds), Mothering against the Odds: Diverse Voices of Contemporary
Mothers. The Guilford Press: New York.

Ford, D., & Lerner, R. (1992). Developmental Systems Theory. Newbury Park,
CA: Sage.

Garabino, J. (1976). A preliminary study of some ecological correlates of child
abuse: The impact of socioeconomic stress on mothers. Child Development,
47, 1 78-1 85.

Garcia Coll, C., & Pachter, L. (2002). Ethnic and minority parenting. In M.H.
Bornstein (Ed), Handbook of parenting Vol. 4. Applied parenting (2"d ed., pp.
1-20). Mahwah: NJ: Lawrence Earlbaum Associates.

Garrnezy, N. (1983). Stressors of childhood. In N. Garrnezy & M. Rutter (Eds),
Stress, coping and development in children (pp.43-84). New York: McG raw-
Hill.

Gelfand, D. M., Teti, D. M., & Fox, C. E. R. (1992). Sources of parenting stress
for depressed and nondepressed mothers of infants. Joumal of Clinical Child
Psychology, 21, 262 - 272.

Gomel, J., Tinsley, B., Parke, R., 8. Clark, K. (1998). The effects of economic
hardship on family relations among African American, Latino, and Euro-
American families. Journal of Family Issues, 19, 436-467.

Gorzka, P. (1999). Homeless parents’ perceptions of parenting stress. Journal
of Child and Adolescent Psychiatric Nursing, 12, 7-16.

Gray, 8., & Ruttle, K. (1980). The family-oriented home visiting program: A
longitudinal study. Genetic Psychology Monographs, 102, 299-316.

Green, 8., & Rodgers, A. (2001). Determinants of social support among low-

income mothers: A longitudinal analysis. American Journal of Community
Psychology, 29, 419-427.

117

Grych, J. H., & Fincham, F. (1990). Marital conﬂict and children’s adjustment: A
cognitive-contextual framework. Psychological Bulletin, 108, 267-290.

Hadadian, A., & Merbler, J. (1996). Mother’s stress: Implications for attachment
relationships. Eariy Child Development and Care, 125, 59-66.

Hanson, M. (1996). Early interactions: The family context. In M. Hanson (Ed),
Atypical infant development (2"d ed., p. 235-272). Austin, TX: Pro-Ed.

Harris, T., Brown, G. W., & Bifulco, A. (1986). Loss of parent in childhood and
adult psychiatric disorder: The role of lack of adequate parental care.
Psychological Medicine, 16, 641-59.

Hillson, J. M. C., & Kuiper, N. A. (1994). A stress and coping model of child
maltreatment. Clinical Psychology Review, 14, 261 - 285.

Holden, E. W., & Banez, G. A. (1996). Child abuse potential and parenting stress
within maltreating families. Journal of Family Violence, 11, 1 - 12.

Holden, E. W., Willis, D. J., & Foltz, L. (1989). Child abuse potential and
parenting stress: Relationships in maltreating parents. Psychological
Assessment: A Journal of Consulting and Clinical Psychology, 1 , 64 - 67.

Holden, G. W., & Ritchie, K. A. (1989). Linking extreme marital discord, child
rearing and child behavior problems: Evidence from battered women. Austin:
University of Texas, Dept. of Psychology.

Holmbeck, G. N. (1997). Toward terminological, conceptual, and statistical
clarity in the study of mediators and moderators: Examples from the child-
clinical and pediatric psychology Iiteratures. Journal of Counseling and
Clinical Psychology, 65, 399-410.

Hoyle, R. H., & Panter, A. (1995). Writing about structural equation models. In
R.H. Hoyle (Ed), Structural equation modeling: Concepts, issues and
applications (pp. 158-176). Thousand Oaks, CA: Sage.

Hu, L., & Bentler, PM. (1995). Evaluating model ﬁt. In R.H. Hoyle (Ed),
Stmctural equation modeling: Concepts, issues and applications (pp. 76-99).
Thousand Oaks, CA: Sage.

Jameson, P., Gelfand, D., Kulcsar, E., & Teti, D. (1997). Mother-toddler
interaction patterns associated with maternal depression. Development and
Psychopathology, 9, 537-550.

118

Jarvis, P., & Creasey, G. (1991). Parental stress, coping, and attachment on
families with an 18-month-old infant. Infant Behavior and Development, 14,
383-395.

Jekielek, S. (1998). Parental conﬂict, marital disruption and children’s emotional
well-being. Social Forces, 76, 905-936.

Joreskog, K., & Sorbom, D. (2003). LISREL 8. 54. Chicago: Scientiﬁc Software
International.

Jouriles, E., Pﬁffner, L., & O’Leary, S. (1987). Marital conﬂict, parenting, and
toddler conduct problems. Journal of Abnormal Child Psychology. 16, 197-
206.

Judge, S. (1998). Parental coping strategies and strengths in families of young
children with disabilities. Family Relations, 47, 263-268.

Kaufman, J., & Zigler, E. (1987). Do abused children become abusive parents?
American Joumal of Orthopsychiatry, 57, 186-192.

Kelley, S. J. (1998). Stress and coping behaviors of substance abusing mothers.
Journal of the Society of Pediatric Nurses, 3, 103-108.

Kolvin, l., Miller, F. J. W., Fleeting, M., & Kolvin, P. A. (1988a). Social and
parenting factors affecting criminal-offense rates: Findings from the
Newcastle Thousand-Family Study (1947-1980). British Journal of Psychiatry,
152, 80-90. -

Kolvin, 1., Miller, F. J. W., Fleeting, M., & Kolvin, P. A. (1988b). Risk/protective
factors for offending with particular reference to deprivation. In M. Rutter
(Ed), Studies of psychosocial risk: The power of longitudinal data (pp. 77-95).
Cambridge, England: Cambridge University Press.

Kotch, J., Browne, D., Ringwalt, C., Dufort, V., Ruina, E., Stewart, P., & Jung, J.
(1997). Stress, social support, and substantiated maltreatment in the second
and third years of life. Child Abuse and Neglect, 21, 1025-1037.

Kraemer, H.C., Stice, E., Kazdin, A., Offord, D., & Kupfer, D (2001). How do risk
factors work together? Mediators, moderators, and independent, overlapping,
and proxy risk factors. The American Joumal of Psychiatry, 158, 848-856.

119

Kraemer, H.C., Wilson, G.T., Fairburn, C.G., & Agras, W.S (2002). Mediators
and moderators of treatment effects in randomized clinical trials. Archives of
General Psychiatry, 59, 877-883.

Larson, R., & Almeida, D. (1990). Emotional transmission in the daily lives of
families. Joumal of Marriage and the Family, 61, 5-20.

Lavee, Y., & Olson., D. (1991). Family types and response to stress. Joumal of
Maniage and the Family, 53, 786-798.

Lazarus, R., & Folkman, S. (1984). Stress, appraisal and coping. New York:
Springer.

Leonard, K., Eiden, R., Wong, M., Zucker, R., Puttler, L., Fitzgerald, H., Hussong,
A., Chassin, L., & Mudar, P. (2000). Developmental perspectives on risk and
vulnerability in alcoholic families. Alcoholism: Clinical and Experimental
Research, 24, 238-240.

Lester, B. M., Boukydis, C. F. 2., & Twomey, J. E. (2000). Maternal substance
abuse and child outcome. In C. H. Zeanah, Jr. (Ed), Handbook of infant
mental health (2"d ed., p. 161-175). New York: The Guilford Press.

Liaw, F., & Brooks-Gunn, J. (1994). Cumulative familial risks and low-birthweight
children’s cognitive and behavioral development. Joumal of Clinical Child
Psychology, 23, 360-372.

Longfellow, C. Zelkowitz, P., & Saunders, E. (1982). The quality of mother-child
relationships. In D. Belle (Ed), Lives in stress: Women and depression.
Beverly Hills, CA: Sage.

Loukas, A., Twitchel, G., Piejak, L., Fitzgerald, H., & Zucker, R. (1998). The
family as a unity of interacting personalities. In L. L’Abete (Ed), Family
psychopathology: The relational roots of dysfunctional behavior (pp. 35-59).
New York: The Guilford Press.

Lupien, S. J., King, 8., Meaney, M. J., & McEwen, B. S. (2001). Can poverty get
under your skin? Basal cortisol levels and cognitive function in children from
low and high socioeconomic status. Development and Psychopathology, 13,
653-676.

120

Luster, T., & Brophy-Herb, H. (2000). Adolescent mothers and their children.
In J.D. Osofsky & H.E. Fitzgerald (Eds), The WAIMH handbook of infant
mental health (Vol, IV): Infant mental health in high risk groups (pp.369-414).
New York: Wiley.

Luthar, S., Cushing, G., Marikangas, K., & Rounsville, B. (1998). Multiple
jeopardy: Risk and protective factors among addicted mothers' offspring.
Development and Psychopathology, 10, 117-136.

Maccoby, E. E., & Martin, J. A. (1983). Socialization in the context of the family:
Mother-child interaction. In P. Mussen & E. Hetherington (Eds), Handbook of
child psychology: Socialization, personality, and social development (Vol. 4)
(pp. 1-101). Wiley: New York.

MacCullum, R. C., & Austin, J. T., (2000). Applications of structural equation
modeling in psychological research. Annual Review of Psychology, 51, 201-
226.

MacCuIlum, R. C., Browne, M. W., & Sugawara, H. M. (1996). Power analysis
and determination of sample size for covariance structure modeling.
Psychological Methods, 1, 130-149.

Magill-Evans, J., & Harrison, M. J. (1999). Mother-child interactions and
development of toddlers born preterm. Western Joumal of Nursing Research,
21, 292-312.

Marcenko, M., & Spence, M. (1994). Home visitation services for at-risk
pregnant and postpartum women: A randomized trial. American Journal of
Orthopsychiatry, 64, 468-478.

Mash, E. J., & Johnston, C. (1990). Determinants of parenting stress:
Illustrations from families of hyperactive children and families of physically
abused children. Journal of Clinical Child Psychology, 19, 313-328.

McCubbin, H., Joy, 0., Cauble, A., Comeau, J., Patterson, J., & Needles R.
(1980). Family stress and coping: A decade of review. Joumal of Marriage
and the Family, 42, 855-871.

McCubbin, H. I., McCubbin, M. A., Thompson, A. I., Han, Sae-Young, Allen, C. T.

(1997). Families under stress: What makes them resilient? Journal of Family
and Consumer Sciences, 89, 2-11.

121

McCubbin, H., Olson, D. H., & Larsen, A. (1987). F-COPES: Family Crisis
Oriented Personal Evaluation Scales. In H. McCubbin & A. Thompson (Eds),
Family assessment inventories for research and practice (pp. 195-205).
Madison, WI: University of Wisconsin - Madison.

McCubbin, H., & Patterson, J. (1983). The family stress process: The double
ABCX model of family adjustment and adaptation. Maniage and Family
Review, 6, 7-37.

McCurdy, K. (1995). Risk assessment in child abuse prevention programs.
Social Work Research, 19, 77 - 87.

McDonald, R. P. (1997). Haldane’s lungs: A case study in path analysis.
Multivariate Behavioral Research, 32, 1-38.

McDonald, R. P., & Ho, M. H. R., (2002). Principles and practice in reporting
structural equation analyses. Psychological Methods, 7, 64-82.

McEwen, B. S. (1998). Protective and damaging effects of stress mediators. New
England Journal of Medicine, 338, 171-179.

McGrath, G., & Beehr, T. (1990). Time and the stress process: Some temporal
issues in the conceptualization and measurement of stress. Stress Medicine,
6, 93-1 04.

McKelvey, L., Fitzgerald, H., Schiffman, R., & von Eye, A. (2002). Family stress
and coping: The mediating role of coping. Infant Mental Health Joumal, 23,
164-181.

Melson, G., Windecker—Nelson, E., & Schwarz, R. (1998). Support and stress in
mothers and fathers of young children. Early Education and Development, 9,
261-281.

Meyer, 8., Chrousos, G., & Gold, P. (2001). Major depression and the stress
system: A life span perspective. Development and Psychopathology, 13,
565-580.

Milner, J. S. (1993). Social information processing and physical child abuse.
Clinical Psychology Review, 13, 275—294.

122

Minde, K. (2000). Prematurity and serious medical conditions in infancy:
Implications for development, behavior, and intervention. In C. H. Zeanah, Jr.
(Ed), Handbook of infant mental health (2Ind ed., p. 176-194). New York: The
Guilford Press.

Moos, R. (1974). Preliminary manual for Family Environment Scale, Work
Environment Scale and Group Environment Scale. Palo Alto, CA: Consulting
Psychologists Press.

Morano, CL. (2003). Appraisal and coping: Moderators or mediators of stress in
Alzheimer’s disease caregivers? Social Work Research, 27, 116-128.

Muthén, L. K., & Muthén, B. O. (2002). How to use a monte carlo study to decide
on sample size and determine power. Structural Equation Modeling, 9, 599-
620.

Myrtveit, I., Stensrud, E., & Olsson, U., (2001). Analyzing data sets with missing
data: An empirical evaluation of imputation methods and likelihood-based
models. IEEE Transactions on Software Engineering, 27, 999-1013.

Nitz, K., Ketterlinus, R., & Brandt, L. (1995). The role of stress, social support,
and family environment in adolescent mothers’ parenting. Journal of
Adolescent Research, 10, 358-382.

O’Brien, M., & Bahadur, M. (1998). Marital aggression, mother’s problem-
solving behavior with children, and children’s emotional and behavioral
problems. Journal of Social and Clinical Psychology, 17, 249-272.

Olds, D., Henderson, 0., Chamberlin, R., & Tatlebaum, R. (1986). Preventing
child abuse and neglect: A randomized trial of nurse home visitation.
Pediatrics, 78, 65-78.

Olds, 0., Henderson, C., & Kitzman, H. (1994). Did prenatal and infancy nurse
home visitation had enduring effects on qualities of parental caregiving and
child health at 25 to 50 months of life? Pediatrics, 93, 89-98.

Olds, 0., Henderson, C., Tatelbaum, R., & Chamberlin, R. (1988). Improving the
life-course development of socially disadvantaged mothers: A randomized
trial of nurse home visitation. American Journal of Public Health, 78, 1436-
1445.

123

Ostberg, M., & Hagekull, B. (2000). A structural modeling approach to the
understanding of parenting stress. Journal of Clinical Child Psychology, 29,
61 5-625.

Parker, G. (1993). Parental rearing style: Examining for links with personality
vulnerability factors for depression. Social Psychiatry and Psychiatric
Epidemiology, 28, 97-100.

Patterson, J. (2002). Integrating family resilience and family stress theory.
JoumaI of Marriage and Family, 64, 349-360.

Pearlin, L., & Schooler, L. (1978). The structure of coping. Joumal of Health and
Social Behavior, 19, 2-21.

Peterson, J., & Hawley, D. (1998). Effects of stressors on parenting attitudes
and family functioning in a primary prevention program. Family Relations, 47,
221-227.

Podolski, C., & Nigg, J., (2001). Parent stress and coping in relation to child
ADHD severity and associated child disruptive behavior problems. Joumal of
Clinical Child Psychology, 30, 503-513.

Prelow, H.M., Tein, J.Y., Roosa, M.W., & Wood, J. (2000). Do coping styles
differ across sociocultural groups? The role of measurement equivalence in

making this judgment. American Journal of Community Psychology, 28, 225-
244.

Quittner, A. L, Glueckauf, R. L., & Jackson, D. N. (1990). Chronic parenting
stress: Moderating versus mediating effects of social support. Journal of
Personality and Social Psychology, 59, 1266 - 1278.

Raikes, H., & Love, J. (2002). Early Head Start: A dynamic new program for
infants and toddlers and their families. Infant Mental Health Journal, 23, 1-13.

Reiss, D., Hetherington, M., Plomin, R., Howe, G. W., Simmens, S. J.,
Henderson, S. H., O’Connor, T.J., Bussell, D. A., Anderson, E. R., & Law, T.
(1995). Genetic questions for environmental studies: Differential parenting
and psychopathology in adolescence. Archives of General Psychiatry, 52,
925-36.

Reitman, D., Currier, R.O., & Stickle, TR. (2002). A critical evaluation of the

Parenting Stress Index - Short Form (PSI-SF) in a Head Start Population.
Journal of Clinical Child and Adolescent Psychology, 31, 384-392.

124

Reppucci, N., Britner, P., & Woolard, J. (1997). Preventing child abuse and
neglect through parent education. Baltimore: Paul H. Brooks.

Ritchie, K., & Holden, G. (1998). Parenting stress in low income battered and
community women: Effects on parenting behavior. Early Education and
Development, 9, 97-112.

Rodgers, A. (1998). Multiple sources of stress and parenting behavior. Children
and Youth Services Review, 20, 525-546.

Rodriguez, 0., & Green, A. (1997). Parenting stress and anger expression as
predictors of child abuse potential. Child Abuse and Neglect, 21, 367-377.

Rodriguez, C., & Murphy, L. (1997). Parenting stress and abuse potential in
mothers of children with developmental disabilities. Child Maltreatment, 2,
245-251.

Roosa, M., Fitzgerald, H., Carlson, N. (1982). A comparison of teenage and
older mothers: A systems analysis. Joumal of Marriage and the Family, 367-
377.

Rovine, M.J., & Delaney, M. (1990). Missing data estimation in developmental
research. In A. von Eye (Ed), Statistical methods in longitudinal research:
Principles and structuring change (pp. 35-80). San Diego, CA: Academic
Press, Inc.

Rudolph, K., Hammen, C., Burge, D., Lindberg, N., Herzberg, D., & Daley, S.
(2000). Toward an interpersonal life stress model of depression: The
developmental context of stress generation. Development and
Psychopathology, 12, 215-234.

Rutter, M. (1979). Protective factors in children’s responses to stress and
disadvantage. In M. W. Kent & J. E. Rolf (Eds), Primary prevention of
psychopathology: Vol. 3., Social competence in children (pp. 49-74).
Hanover, NH: University Press of New England.

Rutter, M. (1981). Stress, coping and development: Some issues and some
questions. Journal of Child Psychology and Psychiatry, 22, 323-56.

Rutter, M. (1985). Resilience in the face of adversity: Protective factors and
resistance to psychiatric disorder. British Joumal of Psychiatry, 147, 598-611.

125

Rutter, M. (1987). Psychosocial resilience and protective mechanisms. American
Journal of Orthopsychiatry, 57, 316 - 331.

Rutter, M. (1990). Psychosocial resilience and protective mechanisms. In J. Rolf,
A. Masten, D. Cicchetti, K. Nuechterlein, & S. Weintraub (Eds), Risk and
protective factors in the development of psychopathology (pp. 181-214). New
York: Cambridge University Press.

Rutter, M. (1996). Transitions and turning points in developmental
psychopathology: As applied to the age span between childhood and mid-
adulthood. International Joumal of Behavioral Development, 19, 603-26.

Rutter, M. (2000). Resilience reconsidered: conceptual considerations. In J. P.
Shonkoff & S. J. Meisels (Eds), Handbook of eariy childhood intervention (pp.
651-75). New York: Cambridge University Press.

Rutter, M., Champion, L., Quinton, D., Maughan, B., & Pickles, A. (1995).
Understanding individual differences in environmental risk exposure. In P.
Moen, G. H. Elder Jr., & K. Luscher (Eds), Examining lives in context:
Perspectives on the ecology of human development (pp. 61 -93). Washington,
DC: American Psychological Association.

Rutter, M., Dunn, J., Plomin, P., Simonoff, E., Pickles, A., Maughan, B., Orrnel,
J., Meyer, J., & Eaves, L. (1997). Integrating nature and nurture: Implications
of person-environment correlations and interactions for developmental
psychopathology. Development and Psychopathology, 9, 335-64.

Rutter, M., & Quinton, D. (1977). Psychiatric disorder - Ecological factors and
concepts of causation. In H. McGurk (Ed), Ecological factors in human
development (pp. 173-87). Amsterdam: North-Holland.

Rutter, M., & Quinton, D. (1984). Parental psychiatric disorder: Effects on
children. Psychological Medicine, 14, 853-80.

Rutter, M., Quinton, D., & Hill, J. (1990). Adult outcomes of institution-reared
children: Males and females compared. In L. N. Robins & M. Rutter (Eds),
Straight and devious pathways from childhood to adulthood (pp. 135-57).
Cambridge, UK: Cambridge University Press.

Rutter, M., & Sandberg, S. (1992). Psychosocial stressors: Concepts, causes
and effects. European Child & Adolescent Psychiatry, 1, 3-13.

126

Sandler, I. (2001). Quality and ecology of adversity as common mechanisms of
risk and resilience. American Joumal of Community Psychology, 29, 19-61.

Sameroff, A. (1975). Transactional models in early social relations. Human
Development, 18, 65 - 79.

Sameroff, A. J., Bartko, W. T., Baldwin, A., Baldwin, C., & Seifer, R. (1998).
Family and social inﬂuences on the development of child competence. In M.
Lewis & C. Feiring (Eds), Families, risk and competence (pp. 161 - 185).
Mahwah, NJ: Erlbaum.

Sameroff, A. J., & Feil, L. A. (1985). Parental concepts of development. In I. E.
Sigel (Ed), Parental belief systems (pp. 83 - 105). Hillsdale, NJ: Erlbaum.

Sameroff, A. J., & Seifer, R. (1983). Familial risk and child competence. Child
Development, 54, 1254 - 1268.

Sameroff, A. J., & Seifer, R. (1995). Accumulation of environmental risk and
child mental health. In H. E. Fitzgerald, B. M. Lester, & B. Zuckerrnan (Eds),
Children of poverty (pp. 233 - 254). New York: Garland Publishing.

Sameroff, A. J., Seifer, R., & Bartko, W. T. (1997). Environmental perspectives
on adaptation during childhood and adolescence. In S. S. Luthar, J. A.
Burack, D. Chicchetti, & J. R. Weisz (Eds), Developmentalpsychopathology:
Perspective on adjustment, risk, and disorder (pp. 507-526). New York:
Cambridge University Press.

Sameroff, A. J., Seifer, R., Zax, M., & Barocas, R. (1987). Early indicators of
developmental risk: The Rochester Longitudinal Study. Schizophrenia
Bulletin, 13, 383 - 393.

Sarason, B.R., Sarason, R.G., & Pierce, GR. (1990). Traditional views of social
support and their impact on assessment. In B. Sarason, l. Sarason, & G.
Pierce (Eds), Social support: An interactional view (pp.1-26). New York:
Wiley.

Schiffman, R. F ., & McKelvey, L. M. (2002, October). Early Head Start: Results of
National and Local Studies. Community Action Agency Board of Directors
Meeting, Hillsdale, Ml.

Schiffman, R. F., Omar, M. A., & McKelvey, L. M. (2003). Mother-infant
interaction in low-income families. MCN The American Joumal of Maternal
Child Nursing, 28 (4), 246-251.

127

Schiffman, R. (1996). Pathways project: Research into directions for family
health and service use. Administration for Children, Youth and Families
(90YF0010).

Schumacker, R., & Lomax, R. (1996). A beginner’s guide to structural equation
modeling. Mahwah, NJ: Earlbaum.

Seabrook, L. H. (2001). Examining physical child abuse potential as a function of
risk load, maternal affective risk, and self-rated parenting skill. Dissertation
Abstract International, 62, 564.

Seccombe, K. (2002). “Beating the odds" versus “changing the odds”: Poverty,
resilience, and family policy. Journal of Maniage and Family, 64, 384-394.

Seifer, R., & Dickstein, S. (2000). Parental mental illness and infant
development. In C.H. Zeanah, Jr. (Ed), Handbook of infant mental health
(2"d ed., p. 145-160). New York: The Guilford Press.

Seifer, R., Sameroff, A. J., Baldwin, C. P., & Baldwin, A. (1992). Child and family
factors that ameliorate risk between 4 and 13 years of age. Journal of the
American Academy of Child and Adolescent Psychiatry, 31, 893 - 903.

Seitz, V., Rosenbaum, L., & Apfel, N. (1985). Effects of family support
intervention: A ten-year follow-up. Child Development, 56, 376-391.

Simons, R., & Johnson, C. (1996). Mother’s parenting. In B. Adams and D.
Klein (Eds), Understanding the differences between divorced and intact
families: Stress, interaction, and child outcome. Thousand Oaks, CA: Sage.

Simons, R., Whitbeck, L., Conger, R., & Melby, J. (1990). Husband and wife
differences in determinants of parenting: A social learning model of parental
behavior. Journal of Marriage and the Family, 52, 375-392.

Skinner, E., Edge, K., Altman, J., & Sherwood, H. (2003). Searching for a
structure of coping: A review and critique of category systems for classifying
ways of coping. Psychological Bulletin, 129, 216-269.

Smith, J., Berthelsen, D., & O’Connor, 1. (1997). Child adjustment in high conﬂict
families. Child Care, Health and Development, 23, 113-133.

Steiger, J. H. (1990). Structural model evaluation and modiﬁcation: An interval
estimation process. Multivariate Behavioral Research, 25, 173-180.

128

Sumner, G., & Speitz, A. (1994). Caregiver-mother-child interaction teaching
manual. Seattle, WA: NCAST Publications.

Tanaka, J.S. (1993). Multifaceted conceptions of ﬁt in structural equation
models. In K. A. Bollen & J. S. Long (Eds), Testing structural equation
models (pp.10-39). Thousand Oaks, CA; Sage.

Teti, D. M., & Gelfand, D. M. (1991). Behavioral competence among mothers of
infants in the ﬁrst year: The mediational role of self-efﬁcacy. Child
Development, 65, 918 - 929.

Tronick, E. Z. (1986). Interactive mismatch and repair: Challenges to the coping
infant. Zero to Three, 6, 1-6.

US. Department of Health and Human Services, Administration for Children,
Youth, and Families (2002). Making a difference in the lives of infants and
toddlers and their families: The impacts of Early Head Start. Vol. I: Final
Report. Washington, DC.

Van Egeren, L. V., McKelvey, L. M., Fitzgerald, H. E., 8*. Schiffman, R. F. (2002,
July). Change in mother-child interaction in low-income families: Links to
father status. Paper presented at World Association for Infant Mental
Health’s 8th World Congress, Amsterdam.

Wallace, I. F., Roberts, J. E., & Lodder, D. E. (1998). Interactions with African-
American infants and their mothers: Relations with development at 1 year of
age. Joumal of Speech, Language, and Hearing Research, 41, 900-912.

Webster-Stratton, C. (1990). Stress: A potential disruptor of parent perceptions
and family interactions. Journal of Clinical Child Psychology, 19, 302-312.

Weinraub, M, & Wolf, B. (1983). Effects of stress and social supports on
mother-child interactions in single- and two-parent families. Child
Development, 54, 1297-131 1.

Whipple, E. E., & Webster-Stratton, C. (1991). The role of parental stress in
physically abusive families. Child Abuse 8. Neglect, 15, 279 - 291.

Wills, T., Belchman, E., & McNamara, G. (1996). Family support, coping and
competence. In M. Hetherington & E. Blechman (Eds), Resiliency in children
and families (pp. 107-133). Mahwah, NJ: Earlbaum.

129

Woodward, L., & Fergusson, D., (2002). Parent, child, and contextual predictors
of childhood physical punishment. Infant and Child Development, 11, 213-
235.

Younger, J., Kendell, M., & Pickler, R. (1997). Mastery of stress in mothers of
preterm infants. Joumal for the Society of Pediatric Nurses, 2, 29-36.

Zeanah, C. H., Jr., Larrieu, J. A., Heller, S. S., & Valliere, J. (2000). Infant-

parent relationship assessment. In C. H. Zeanah, Jr. (Ed), Handbook of
infant mental health (2"d ed., p. 222-235). New York: The Guilford Press.

130

Appendix A: Missing Data Analysis

Missingness on the Parenting Stress Index
At the 24-month data collection point, there were 154 families (92.7%)

from whom PSI data were collected and 12 families from whom 24-month PSI
responses were not attained. Of the 154 families, one case was missing a single
item (Item I, which reads “when you go to a party you usually expect to had a bad
time”). Nationally, this item had caused some problems as many data collectors
left it blank as opposed to marking it “not sure” (3) which was not to be read
aloud as a response code (many respondents appear to had stated “I don’t go to
parties”). The PSI Parental Distress scale, from which this item comes, was
computed by Mathematica Policy Research, Inc. Recently at the national level, it
was decided that cases with missing values for this item should automatically
revert to the “not sure”. There were no other cases with single missing items at
24 months and all of the values of the PSI Parent-Child Dysfunctional Interaction
scale were present.

At the 36-month child’s birthday related data point, there were 151 families
(91%) from whom PSI data were collected, yielding 15 families missing PSI data.
There was one case missing several items, nearly all of the items on the PSI
Parent-Child Dysfunctional Interaction Scale. It did not appear that these items
were skipped for this case based on the layout of the instrument. This yields 151
PSI Parental Distress scale scores and 150 PSI Parent-Child Dysfunctional

Interaction scale scores at 36 months.
131

Missingness on the Familijnvironment Scale. Conﬂict Scala

There were 144 cases (86.7%) that had all Family Environment Scale
(FES) Conﬂict items at the 24-month birthday related interview and 143 cases
(86.1%) that had all FES conﬂict items at 36 months. There were no cases with
single missing items on the FES conﬂict scale.
Missingness on the Family Crisis Oriented Personal Scale

At the 24-month data collection point 152 cases (91.6%) had Family Crisis
Oriented Personal Scale (F-COPES) items. There was one case in that 152 that
refused to answer or skipped item 5 (“When faced with problems in your family,
do you seek advice from relatives?”), which belongs to the support seeking scale.

At the 36-month data collection point 144 families (86.7%) had FCOPES
items, with one case that refused to answer or skipped item 6 (“When faced with
problems in your family, do you seek assistance from community agencies or

programs?”), which belongs to the support seeking scale.

Missingness on theidrsing Child Assessment Satellite Training (NCAS I ) —

Teaching Scale

 

At 24 months, there were 145 mother-child dyads (87.3%) for whom the
NCAST observation was attained. There were 136 cases with complete data,
which represents 81.9% of the sample. There were nine dyads that were
missing individual items. Seven of those nine cases were missing one item, and
the remaining two dyads were missing two items. There was no overlap of

behaviors that were not observed or scored.
132

At 36 months, 141 (84.9%) dyads had NCAST observations and 136
(81.9%) had complete items. Therefore, there were ﬁve families missing at least
one item. Of those ﬁve cases, one was missing two items and the remaining four
were missing only one item. Like as seen with the 24-month data, there appears
to be no pattern of missing items.

Mﬂmess on Program Dosage

As approximately half of the cases (49%, N = 82) in the study were
assigned to the comparison sample, the EHS program dosage variable will be nil
for these families. Of the 84 program families in the study, there were two
missing program chart data (in one case the chart could not be located at the
Community Action Agency and in the other case the ﬁle was located, but the
progress notes from which dosage information was extracted were not contained
within).

Missingness on Adversig

There were no families missing adversity variables. These variables were
treated as summations of risk as and such, missing values were treated as a
non-instance of adversity. For a breakdown of the individual scale items, see

Table 2.

133

Appendix B: Family Characteristics by Data Presence

Demographic Characteristics

There were no differences between families from whom we collected data at 24

or 36 months (or both) and families from whom we did not collect data at either

point on demographic characteristics:

Race of the respondent

Age of mother at enrollment

Education level of mother at enrollment

Mother’s employment status at enrollment (a greater percentage of the
families from whom we were able to collect data at 24 or 36 months were
families in which the mother was employed at enrollment, 28% versus
4.2%, p < .05).

Primary language (a greater percentage of the families from whom we
were not able to collect data at 24 or 36 months spoke a language other
than English in the home at enrollment, 16.7% versus 4.7%, this
difference was statistically signiﬁcant, p < .05)

Living arrangements at enrollment, whether there was a male present in
the home

Gender of the focus child

Birth order of the focus child

Child’s age at enrollment

134

Appendix C: Outlier Analysis

Upon examination of the distributions of the data, outliers were discovered
that were out of range and were determined likely to change the results of the
analyses (by changing the slopes of individual regressions). These values were
truncated to the next nearest in range value, or windsorized (Foster, 1986).
There were no outliers, high or low, on several variables in the study. Mother-
child interaction at 24 months, parenting stress at 36 months, or in support
seeking coping behaviors assessed when the child was 24 and 36 months of
age. Outliers were truncated for mother-child interaction at 36 months of age,
parenting stress at 24 months, family conﬂict at both assessments, cognitive
reframlng at both assessments, program dosage and family adversity. The
changes that were made were as follows (means and standard deviations for
comparison can be found in Table 1):

1. Mother-Child Interaction — 36 month assessment: One value was an outlier
on the low end of the distribution (value 40 replaced 34).

2. Parenting Stress Index — 24 month assessment: Four high out of range
values were discovered and were trimmed to the next highest value (values of
110, 95, 89, and 81 were changed to 78).

3. Family Conﬂict — 24 month assessment: Four high end values were replaced
(3.2 and 3.0 were reported by two persons each, they were replaced with 2.8,

the next highest value in the distribution).

135

. Family Conﬂict — 36 month assessment: Two high values were replaced (3.6
and 3.8 were changed to the next highest value in the distribution, which was
3.4).

. Cognitive Reframing — 24 month assessment: Two low end values were
changed (2.0 and 2.5 were replaced with 3.125).

. Cognitive Reframing — 36 month assessment: One low end values were
changed (value 2.5 was replaced with 3.25).

. Program Dosage: There were two families in the EHS group who received
extensive service from program staff. One family received 350 hours of
programming and the other received 266 hours. These two families dosage
was truncated to 200 hours, which was the next highest level of programming
in the distribution.

. Family Adversity: There were two low end and two high end counts for the
number of family adversities (0 and 8), which were changed to the next
closest values in the sample.

Many of the outliers were the only value changed for a particular family;

however, there were 5 families for which two values were changed. One family

scored on both the low extreme of mother-child interaction and the high extreme

of family conﬂict assessed at the child’s 36th month, two families scored at the

high extreme of parenting stress and low extreme of cognitive reframlng

assessed at the child’s 24th month, one family reported extremely high conﬂict at

136

both assessments, and one of the families for whom program dosage was

truncated also reported extremely high family stress at 24 months.

137

Appendix D: Fit lndices for Base Model - Independently predicting:

Degrees of Freedom = 16

Minimum Fit Function Chi-Square = 30.92 (P = 0.014)
Normal Theory Weighted Least Squares Chi-Square = 30.21 (P = 0.017)
Estimated Non-centrality Parameter (NCP) = 14.21

90 Percent Conﬁdence Interval for NCP = (2.45 ; 33.75)
Minimum Fit Function Value = 0.19

Population Discrepancy Function Value (F0) = 0.087

90 Percent Conﬁdence Interval for F0 = (0.015 ; 0.21)

Root Mean Square Error of Approximation (RMSEA) = 0.074
90 Percent Conﬁdence Interval for RMSEA = (0.031 ; 0.11)
P-Value for Test of Close Fit (RMSEA < 0.05) = 0.15
Expected Cross-Validation Index (ECVI) = 0.53

90 Percent Conﬁdence Interval for ECVI = (0.41 ; 0.60)
ECVI for Saturated Model = 0.44

ECVI for Independence Model = 1.34

Chi-Square for Independence Model with 28 Degrees of Freedom = 202.70
Independence AIC = 218.70

Model AIC = 86.21

Saturated AIC = 72.00

Independence CAIC = 251.60

Model CAIC = 201.34

Saturated CAIC = 220.03

Normed Fit Index (NFI) = 0.85

Non-Normed Fit Index (NNFI) = 0.85

Parsimony Normed Fit Index (PNF I) = 0.48

Comparative Fit Index (CFI) = 0.91

Incremental Fit Index (IF I) = 0.92

Relative Fit Index (RFI) = 0.73

Critical N (CN) = 171.79

Root Mean Square Residual (RMR) = 16.14

Standardized RMR = 0.099

Goodness of Fit Index (GFI) = 0.96

Adjusted Goodness of Fit Index (AGF I) = 0.90

Parsimony Goodness of Fit Index (PGF I) = 0.42

138

Appendix E: Fit lndices for Base Model - Relationships between
Predictors:

Degrees of Freedom = 10

Minimum Fit Function Chi-Square = 7.04 (P = 0.72)
Normal Theory Weighted Least Squares Chi-Square = 6.84 (P = 0.74)
Estimated Non-centrality Parameter (NCP) = 0.0

90 Percent Conﬁdence Interval for NCP = (0.0 ; 6.22)
Minimum Fit Function Value = 0.043

Population Discrepancy Function Value (F0) = 0.0

90 Percent Conﬁdence Interval for F0 = (0.0 ; 0.038)

Root Mean Square Error of Approximation (RMSEA) = 0.0
90 Percent Conﬁdence Interval for RMSEA = (0.0 ; 0.062)
P-Value for Test of Close Fit (RMSEA < 0.05) = 0.91
Expected Cross-Validation Index (ECVI) = 0.43

90 Percent Conﬁdence Interval for ECVI = (0.43 ; 0.47)
ECVI for Saturated Model = 0.44

ECVI for Independence Model = 1.34

Chi-Square for Independence Model with 28 Degrees of Freedom = 202.70
Independence AIC = 218.70

Model AIC = 74.84

Saturated AIC = 72.00

Independence CAIC = 251.60

Model CAIC = 214.65

Saturated CAIC = 220.03

Normed Fit Index (NFI) = 0.97

Non-Normed Fit Index (NNFI) = 1.05

Parsimony Normed Fit Index (PNFI) = 0.34

Comparative Fit Index (CF I) = 1.00

Incremental Fit Index (IFI) = 1.02

Relative Fit Index (RF 1) = 0.90

Critical N (CN) = 544.94

Root Mean Square Residual (RMR) = 16.56
Standardized RMR = 0.035

Goodness of Fit Index (GFI) = 0.99

Adjusted Goodness of Fit Index (AGFI) = 0.96

Parsimony Goodness of Fit Index (PGF I) = 0.27

139

Apmndix F: Fit lndices for Base Model — Relationships between PredictorsI
Parsimonious Model:

Degrees of Freedom = 16

Minimum Fit Function Chi-Square = 7.26 (P = 0.97)
Normal Theory Weighted Least Squares Chi-Square = 7.05 (P = 0.97)
Chi-Square Difference with 1 Degree of Freedom = 0.00 (P = 1.00)
Estimated Non-centrality Parameter (NCP) = 0.0

90 Percent Conﬁdence Interval for NCP = (0.0 ; 0.0)
Minimum Fit Function Value = 0.044

Population Discrepancy Function Value (F0) = 0.0

90 Percent Conﬁdence Interval for F0 = (0.0 ; 0.0)

Root Mean Square Error of Approximation (RMSEA) = 0.0
90 Percent Conﬁdence Interval for RMSEA = (0.0 ; 0.0)
P-Value for Test of Close Fit (RMSEA < 0.05) = 1.00
Expected Cross-Validation Index (ECVI) = 0.39

90 Percent Conﬁdence Interval for ECVI = (0.39 ; 0.39)
ECVI for Saturated Model = 0.44

ECVI for Independence Model = 1.34

Chi-Square for Independence Model with 28 Degrees of Freedom = 202.70
Independence AIC = 218.70

Model AIC = 63.05

Saturated AIC = 72.00

Independence CAIC = 251.60

Model CAIC = 178.19

Saturated CAIC = 220.03

Normed Fit Index (NFI) = 0.96

Non-Normed Fit Index (NNFI) = 1.09

Parsimony Normed Fit Index (PNFI) = 0.55

Comparative Fit Index (CFI) = 1.00

Incremental Fit Index (IFI) = 1.05

Relative Fit Index (RFI) = 0.94

Critical N (CN) = 727.78

Root Mean Square Residual (RMR) = 16.63

Standardized RMR = 0.036

Goodness of Fit Index (GFI) = 0.99

Adjusted Goodness of F it Index (AGFI) = 0.98

Parsimony Goodness of Fit Index (PGFI) = 0.44

140

Apgndix G: Fit lndices for Modeling Coping - Independent Predictors:

Degrees of Freedom = 46

Minimum Fit Function Chi-Square = 98.11 (P = 0.00)

Normal Theory Weighted Least Squares Chi-Square = 96.04 (P = 0.00)
Chi-Square Difference with 1 Degree of Freedom = 0.00 (P = 1.00)
Estimated Non-centrality Parameter (NCP) = 50.04

90 Percent Conﬁdence Interval for NCP = (25.73 ; 82.12)
Minimum Fit Function Value = 0.59

Population Discrepancy Function Value (F0) = 0.31

90 Percent Conﬁdence Interval for F0 = (0.16 ; 0.50)

Root Mean Square Error of Approximation (RMSEA) = 0.082
90 Percent Conﬁdence Interval for RMSEA = (0.059 ; 0.10)
P-Value for Test of Close Fit (RMSEA < 0.05) = 0.014
Expected Cross-Validation Index (ECVI) = 1.13

90 Percent Conﬁdence Interval for ECVI = (0.91 ; 1.25)
ECVI for Saturated Model = 0.96

ECVI for Independence Model = 2.44

Chi-Square for Independence Model with 66 Degrees of Freedom = 373.11
Independence AIC = 397.11

Model AIC = 184.04

Saturated AIC = 156.00

Independence CAIC = 446.46

Model CAIC = 364.97

Saturated CAIC = 476.74

Normed Fit Index (NFI) = 0.74

Non-Normed Fit Index (NNFI) = 0.76

Parsimony Normed Fit Index (PNFI) = 0.51

Comparative Fit Index (CF I) = 0.83

Incremental Fit Index (IFI) = 0.84

Relative Fit Index (RF 1) = 0.62

Critical N (CN) = 120.75

Root Mean Square Residual (RMR) = 11.36

Standardized RMR = 0.10

Goodness of Fit Index (GFI) = 0.91

Adjusted Goodness of Fit Index (AGFI) = 0.85

Parsimony Goodness of Fit Index (PGFI) = 0.54

141

Apgndix H: Fit lndices for Modeling Coping: Step 2 - Modeling
Relationships Between Predictors:

Degrees of Freedom = 46

Minimum Fit Function Chi-Square = 25.50 (P = 0.99)
Normal Theory Weighted Least Squares Chi-Square = 24.60 (P = 1.00)
Chi-Square Difference with 5 Degrees of Freedom = 0.62 (P = 0.99)
Estimated Non-centrality Parameter (NCP) = 0.0

90 Percent Conﬁdence Interval for NCP = (0.0 ; 0.0)
Minimum Fit Function Value = 0.15

Population Discrepancy Function Value (F0) = 0.0

90 Percent Conﬁdence Interval for F0 = (0.0 ; 0.0)

Root Mean Square Error of Approximation (RMSEA) = 0.0
90 Percent Conﬁdence Interval for RMSEA = (0.0 ; 0.0)
P-Value for Test of Close Fit (RMSEA < 0.05) = 1.00
Expected Cross-Validation Index (ECVI) = 0.75

90 Percent Conﬁdence Interval for ECVI = (0.75 ; 0.75)
ECVI for Saturated Model = 0.96

ECVI for Independence Model = 2.44

Chi-Square for Independence Model with 66 Degrees of Freedom = 373.11
Independence AIC = 397.11

Model AIC = 112.60

Saturated AIC = 156.00

Independence CAIC = 446.46

Model CAIC = 293.53

Saturated CAIC = 476.74

Normed Fit Index (NFI) = 0.93

Non-Normed Fit Index (NNFI) = 1.10

Parsimony Normed Fit Index (PNFI) = 0.65

Comparative Fit Index (CFI) = 1.00

Incremental Fit Index (IFI) = 1.06

Relative Fit Index (RFI) = 0.90

Critical N (CN) = 461.81

Root Mean Square Residual (RMR) = 11.56

Standardized RMR = 0.044

Goodness of Fit Index (GFI) = 0.98

Adjusted Goodness of Fit Index (AGFI) = 0.96

Parsimony Goodness of Fit Index (PGFI) = 0.58

142

Apgndix I: Partial Correlation Coefﬁcients Support Seeking

Variable Mean Standard Dev Cases
SUPPRT36 3.1174 .7632 166
MTOTAL4 54.6801 5.7466 166
SUPPRT24 2.8906 .7726 166

P A R T I A L C O R
T S

R E L A T I O N C O E F F I C I E N

Zero Order Correlations

SUPPRT36
SUPPRT36 1.0000
( O)
P:
MTOTAL4 .0253
( 164)
P = .747
SUPPRT24 .4746
( 164)
P = .000

P A R T I A L C O R
T S

Controlling for SUPPRT24 (support seeking behaviors at 24

months)
SUPPRT36
SUPPRT36 1.0000
( O)
p =
MTOTAL4 -.1210
I 163)
P = .122

(Coefficient / (D.F.)

MTOTAL4 SUPPRT24

.0253 .4746
( 164) ( 164)
P = .747 P = .000
1.0000 .2694
( O) ( 164)
P = . P = .000
.2694 1.0000
( 164) 1 0)
P = .000 P =

R E L A T I O N C O E F F I C I E N

MTOTAL4
-.1210

( 163)

P = .122
1.0000

( O)

p =

/ 2-tailed Significance)

143

Appendix J: Fit lndices for Modeling Moderating Effects of Support Seeking

Degrees of Freedom = 122

Minimum Fit Function Chi-Square = 129.80 (P = 0.30)
Normal Theory Weighted Least Squares Chi-Square = 120.58 (P = 0.52)
Estimated Non-centrality Parameter (NCP) = 0.0

90 Percent Conﬁdence Interval for NCP = (0.0 ; 28.29)
Minimum Fit Function Value = 0.79

Population Discrepancy Function Value (F0) = 0.0

90 Percent Conﬁdence Interval for F0 = (0.0 ; 0.17)

Root Mean Square Error of Approximation (RMSEA) = 0.0
90 Percent Conﬁdence Interval for RMSEA = (0.0 ; 0.038)
P-Value for Test of Close Fit (RMSEA < 0.05) = 1.00
Expected Cross-Validation Index (ECVI) = 1.46

90 Percent Conﬁdence Interval for ECVI = (1.46 ; 1.63)
ECVI for Saturated Model = 2.10

ECVI for Independence Model = 3.99

Chi-Square for Independence Model with 153 Degrees of Freedom = 614.31
Independence AIC = 650.31

Model AIC = 254.58

Saturated AIC = 342.00

Independence CAIC = 724.32

Model CAIC = 530.08

Saturated CAIC = 1045.15

Normed Fit Index (NFI) = 0.79

Non-Normed Fit Index (NNFI) = 0.98

Parsimony Normed Fit Index (PNFI) = 0.63

Comparative F it Index (CFI) = 0.98

Incremental Fit Index (IFI) = 0.98

Relative Fit Index (RFI) = 0.74

Critical N (CN) = 205.97

Root Mean Square Residual (RMR) = 52.78
Standardized RMR = 0.073

Goodness of Fit Index (GFI) = 0.92

Adjusted Goodness of Fit Index (AGFI) = 0.89

Parsimony Goodness of Fit Index (PGFI) = 0.66

Apgndix K: Fit lndices for Modeling Moderating Effects of Support
Seeking - Parsimonious Model

Degrees of Freedom = 69

Minimum Fit Function Chi-Square = 85.11 (P = 0.091)
Normal Theory Weighted Least Squares Chi-Square = 79.47 (P = 0.18)
Chi-Square Difference with 3 Degrees of Freedom = 1.99 (P = 0.57)
Estimated Non-centrality Parameter (NCP) = 10.47

90 Percent Conﬁdence Interval for NCP = (0.0 ; 36.73)
Minimum Fit Function Value = 0.52

Population Discrepancy Function Value (F0) = 0.064

90 Percent Conﬁdence Interval for F0 = (0.0 ; 0.23)

Root Mean Square Error of Approximation (RMSEA) = 0.031
90 Percent Conﬁdence Interval for RMSEA = (0.0 ; 0.057)
P-Value for Test of Close Fit (RMSEA < 0.05) = 0.87
Expected Cross-Validation Index (ECVI) = 1.10

90 Percent Conﬁdence Interval for ECVI = (0.95 ; 1.18)
ECVI for Saturated Model = 1.29

ECVI for Independence Model = 2.29

Chi-Square for Independence Model with 91 Degrees of Freedom = 344.93
Independence AIC = 372.93

Model AIC = 179.47

Saturated AIC = 210.00

Independence CAIC = 430.50

Model CAIC = 385.07

Saturated CAIC = 641.76

Normed Fit Index (NFI) = 0.75

Non-Normed Fit Index (NNFI) = 0.92

Parsimony Normed Fit Index (PNFI) = 0.57

Comparative Fit Index (CFI) = 0.94

Incremental Fit Index (IFI) = 0.94

Relative Fit Index (RFI) = 0.67

Critical N (CN) = 193.38

Root Mean Square Residual (RMR) = 15.14

Standardized RMR = 0.071

Goodness of Fit Index (GFI) = 0.94

Adjusted Goodness of Fit Index (AGF I) = 0.90

Parsimony Goodness of Fit Index (PGFI) = 0.61

145

Appendix L: Fit lndices for Modeling Moderating Effects of Cognitive
Reframing

Degrees of Freedom = 120

Minimum Fit Function Chi-Square = 172.17 (P = 0.0013)
Normal Theory Weighted Least Squares Chi-Square = 158.72 (P = 0.010)
Estimated Non-centrality Parameter (NCP) = 38.72

90 Percent Conﬁdence Interval for NCP = (10.03 ; 75.51)
Minimum Fit Function Value = 1.04

Population Discrepancy Function Value (F0) = 0.24

90 Percent Conﬁdence Interval for F0 = (0.062 ; 0.46)

Root Mean Square Error of Approximation (RMSEA) = 0.044
90 Percent Conﬁdence Interval for RMSEA = (0.023 ; 0.062)
P-Value for Test of Close Fit (RMSEA < 0.05) = 0.68
Expected Cross-Validation Index (ECVI) = 1.82

90 Percent Conﬁdence Interval for ECVI = (1.53 ; 1.94)
ECVI for Saturated Model = 2.10

ECVI for Independence Model = 3.30

Chi-Square for Independence Model with 153 Degrees of Freedom = 501.64
Independence AIC = 537.64

Model AIC = 296.72

Saturated AIC = 342.00

Independence CAIC = 611.65

Model CAIC = 580.45

Saturated CAIC = 1045.15

Normed Fit Index (NFI) = 0.66

Non-Normed Fit Index (NNFI) = 0.81

Parsimony Normed Fit Index (PNFI) = 0.52

Comparative Fit Index (CF I) = 0.85

Incremental Fit Index (IFI) = 0.86

Relative Fit Index (RFI) = 0.56

Critical N (CN) = 153.33

Root Mean Square Residual (RMR) = 12.72

Standardized RMR = 0.078

Goodness of Fit Index (GFI) = 0.90

Adjusted Goodness of Fit Index (AGFI) = 0.86

Parsimony Goodness of F it Index (PGFI) = 0.63

146

Appendix M: Fit lndices for Modeling Moderating Effects of Cognitive
Reframing — Parsimonious Model

Degrees of Freedom = 93

Minimum Fit Function Chi-Square = 118.39 (P = 0.039)
Normal Theory Weighted Least Squares Chi-Square = 105.71 (P = 0.17)
Chi-Square Difference with 1 Degree of Freedom = 2.55 (P = 0.11)
Estimated Non-centrality Parameter (NCP) = 12.71

90 Percent Conﬁdence Interval for NCP = (0.0 ; 42.23)
Minimum Fit Function Value = 0.72

Population Discrepancy Function Value (F0) = 0.078

90 Percent Conﬁdence Interval for F0 = (0.0 ; 0.26)

Root Mean Square Error of Approximation (RMSEA) = 0.029
90 Percent Conﬁdence Interval for RMSEA = (0.0 ; 0.053)
P-Value for Test of Close Fit (RMSEA < 0.05) = 0.92
Expected Cross-Validation Index (ECVI) = 1.37

90 Percent Conﬁdence Interval for ECVI = (1.20 ; 1.46)
ECVI for Saturated Model = 1.67

ECVI for Independence Model = 2.87

Chi-Square for Independence Model with 120 Degrees of Freedom = 435.75
Independence AIC = 467.75

Model AIC = 223.71

Saturated AIC = 272.00

Independence CAIC = 533.55

Model CAIC = 466.32

Saturated CAIC = 831.23

Normed Fit Index (NFI) = 0.73

Non-Normed Fit Index (NNFI) = 0.90

Parsimony Normed Fit Index (PNF I) = 0.56

Comparative Fit Index (CFI) = 0.92

Incremental Fit Index (IF I) = 0.93

Relative Fit Index (RFI) = 0.65

Critical N (CN) = 178.89

Root Mean Square Residual (RMR) = 9.19

Standardized RMR = 0.068

Goodness of Fit Index (GFI) = 0.93

Adjusted Goodness of Fit Index (AGFI) = 0.89

Parsimony Goodness of Fit Index (PGFI) = 0.63

147

Appendix N: Risk Thresholds

 

 

Adversity

Categorized Mean SD

Program Dosage 0 to 2 risks (N = 40) 45.73 66.01
3 or more (N = 126) 38.73 55.14
PSI (24 months) 0 to 2 risks (N = 40) 45.97 12.25
3 or more (N = 126) 46.31 13.31

PSI (36 months) 0 to 2 risks (N = 40) 46.36 11.51
3 or more (N = 126) 45.99 13.89

Conﬂict (24 months) 0 to 2 risks (N = 40) 1.76 0.42
3 or more (N = 126) 1.769 0.51

Conﬂict (36 months) 0 to 2 risks (N = 40) 1.89 0.56
3 or more (N = 126) 1.82 0.59

CogRef (24 months) 0 to 2 risks (N = 40) 4.31 0.45
3 or more (N = 126) 4.32 0.45

CogRef (36 months) 0 to 2 risks (N = 40) 4.38 0.47
3 or more (N = 126) 4.42 0.44

Support (24 months) 0 to 2 risks (N = 40) 2.92 0.71
3 or more (N = 126) 2.88 0.80

Support (36 months) 0 to 2 risks (N = 40) 3.19 0.68
3 or more (N = 126) 3.09 0.79

Interaction (24 months) 0 to 2 risks (N = 40) 53.71 6.55
3 or more (N = 126) 55.57 6.37

Interaction (36 months)t 0 to 2 risks (N = 40) 56.14 4.78

3 or more (N = 126) 54.22 5.96

tdenotes signiﬁcance at p < .10

148

Appendix 0: Examining Conﬂict across Time

Conﬂict Scores at Dichotomized at mean Agreement: 24 Months

 

E" a a 3 a
o '9 o E o
‘8 8 s 8 s 8
9 O. 0. 3 CL
LL 0
Family Conﬂict
Conﬂict at scale
24 Months indicates 15° 90'4 90'4 90'4
Non-
Agreement
Conﬂict
scale 16 9.6 9.6 100.0
indicates
Agreement
Total 166 100.0 100.0

 

 

149

Conﬂict Scores at Dichotomized at mean Agreement: 36 Months

 

8 ‘E a 3 ‘E
o ‘9 o 111' o
‘8 8 8° 8 8 8
9 D. D. a 0.
LL 0
Family Conﬂict
Conﬂict at scale
36 Months indicates 142 85's 85'5 85'5
Non-
Agreement
Conﬂict
scale 24 14.5 14.5 100.0
indicates
Agreement
Total 166 100.0 100.0

150

 

 

Conﬂict Scores across Time

 

 

 

>. E e
0 *u 0) a ‘0‘
5 5 8 «:5 8
a 8 e 88
IE 0. % 3 CL
> 0
Family
Conﬂict -
D'sagiee- 133 80.1 80.1 80.1
Across ment at
me Both 24 and
36 Months
Agreement 25 15.7 15.7 95.8
at Either 24
or36
Months
Agreement
at Both 24 7 4.2 4.2 100.0
and 36
Months
Total 166 100.0 100.0

151

DESCRIPTIVE STA TIS TICS

N

Mean

Std. Deviation

Std. Error

 

Parenting
Stress at 36
Months*

Disagreement
at Both 24 and
36 Months
Agreement at
Either 24 or 36
Months
Agreement at
Both 24 and 36
Months

Total

133

26

166

44.84

48.67

59.86

46.08

13.12

12.89

10.71

13.32

1.14

2.53

4.05

1.03

 

Cognitive
Reframing at
36 Monthst

Disagreement
at Both 24 and
36 Months
Agreement at
Either 24 or 36
Months
Agreement at
Both 24 and 36
Months

Total

133

26

166

4.44

4.34

4.09

4.41

.44

.44

.43

.45

.038

.09

.16

.035

 

Suppon
Seeking at 36
Months

Disagreement
at Both 24 and
36 Months
Agreement at
Either 24 or 36
Months
Agreement at
Both 24 and 36
Months

Total

152

133

26

166

3.11

3.15

3.16

3.12

.74790

.89

.65

.76

.06485

.17

.24

.06

DESCRIPTIVE STA Trs rigs (Cont’d)

 

 

N Mean Std. Deviation Std. Enor

Mother-Child Disagreement
Interaction at at Both 24 and 133 54.73 5.71 .50
36 Months* 36 Months

Agreement at

Either 24 or 36 26 56.14 4.79 .94

Months

Agreement at

Both 24 and 36 7 48.35 6.26 2.37

Months

Total 166 54.68 5.75 .45

153

ANOVA TABLE

 

 

1 Sum of Mean
Squares df Square F %_
Parenting
Stress at 36 Between
Months Groups 1707.684 2 853.842 5.048 .007
Within Groups 27573.222 163 169.161
Total 29280.906 165
Cognitive
Reframing at Between
36 Months Groups .969 2 .484 2.486 .086
Within Groups 31.761 163 .195
Total 32.730 165
Suppon
Seeking at Between
36 Months Groups .045 2 .023 .038 .962
Within Groups 96.075 163 .589
Total 96.120 165 '
Mother-Child
Interaction at Between
36 Months Groups 335.967 2 167.983 5.355 .006
Within Groups 5112.960 163 31.368
Total 5448.927 165

 

 

 

 

 

 

 

 

154

Appendix P: Exploring EHS Program Dosage

Variable Mean Standard Dev Cases
P8124 45.0132 12.7988 84
PSI36 45.1891 13.5103 84
BZP_CONF 1 .7655 .4956 84
B3P_CONF 1.8044 .5921 84
REFRAM3 4.3297 .4504 84
REFRAMF 4.4437 .4276 84
SUPPRT24 2.8869 .8204 84
SUPPRT36 3.1280 .7562 84
MTOTAL3 55.3871 5.8903 84
MTOTAL4 53.8564 6.1203 84
DOSAGE 79.8709 58.7824 84
ADVRSITY 3.5833 1 .4987 84

PARTIAL CORRELATION COEFFICIENTS
Controlling for.. ADVRSITY
(Coefﬁcient / (D.F.) / 2-tailed Signiﬁcance)

DOSAGE

P8124 -.0208 REFRAMF .0142
( 81) ( 81)
P= .852 P= .899

P8136 -.1058 SUPPRT24 .2596
( 81) ( 81)

P= .341 P= .018
BZP_CONF -.1292 SUPPRT36 .1608
( 81) ( 81)

p: .244 P= .146
B3P_CONF .0691 MTOTAL3 .0530
( 81) ( 81)

P= .535 P= .634
REFRAM3 .0105 MTOTAL4 .0844
( 81) ( 81)

P= .925 P= .448

155

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

11111111111111111111111