o

t ' . ,
fier‘fimfr':
“xiii-9".) ‘1'“;-

.:7.+£1' ° "1
‘ 1‘ ,

. _.J
31...:
~11 -.

.vu’.‘
- _, -
mu...
...

5‘...-.-..‘.

.. - .- way-v

' «32:

.' f“ a“
.V ‘ ‘
x

‘ “11‘4“ tr,

." '5' .
v? 1‘3"-
‘ a A w.‘
.. w

"a

.F.T"‘.DI lan—.I ‘

_.-,

: “‘11???”
m :_ :g-

(‘0
I‘r
. M2-

;. 1
{111% H 31%;?

,t
,-
2
A
.
.

.=:;'.. < 2.,-

 

" o. . -'
M~..u—:

 

THESIS

L

lllllllllllllllllllllllllllllllllllllllllllllllllllllllll||l

31293 01710 380

This is to certify that the

thesis entitled
Developmental Changes in Risk Indicators and the Prediction

of Parental Physical Aggression

presented by

Lucy H. Seabrook

has been accepted towards fulfillment
of the requirements for

Masters degree in Psgchology

(NAM? ¥nZ—<f
vy Major p%essor

Date y’Af’ 7;—

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

LIBRARY

Michigan State
Unlversity

 

 

 

PLACE IN RETURN BOX
to remove this checkout from your record.
To AVOID FINES return on or before date due.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1M mm.mu

 

DEVELOPMENTAL CHANGES IN RISK INDICATORS AND THE PREDICTION OF

PARENTAL PHYSICAL AGGRESSION

Lucy H. Seabrook

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements

for the degree of

MASTER OF ARTS

Department of Psychology
1998

Professor Hiram E. Fitzgerald

ABSTRACT

DEVELOPMENTAL CHANGES IN RISK INDICATORS AND THE PREDICTION OF

PARENTAL PHYSICAL AGGRESSION

Lucy H. Seabrook

Studies of risk indicators for maltreatment have been largely limited by cross-
sectional data, which fail to account for developmental changes in the family context.
Examining the deveIOpmental nature of risk levels may provide evidence that risk factors
operate differentially in families to influence the emergence of child maltreatment. This
study examined the developmental contexts of families to ascertain the nature of risk
indicators most likely to predict parental physical aggression. Risk indicators of child
externalizing behaviors and family conflict emerged as significant predictors of parental
physical aggression, while parental alcohol consumption, depression, and SES did not.
Risk indicators that predicted parental physical aggression were consistently high across
time. Fluctuations in risk level (as measured by risk indicators) over time did not predict

parental physical aggression.

ACKNOWLEDGMENTS

Gratitude to my advisor, Hiram E. Fitzgerald, Ph.D., for his guidance,

encouragement, and always-engaging conversation. Also thanks to C. Raymond

Bingham, Ph.D., for patience, support, and many hours of hands-on help.

iii

TABLE OF CONTENTS

List of Tables
List of Figures

Introduction/Rationale

Theory
Developmental-Ecological Framework
The Family Context: Change or Stability in Risk Status?
Risk Indicators
Depression
Alcohol Consumption
Perception of Child Difficulty
Conflict in the Family Environment
Socioeconomic Status
A Developmental Framework of Risk:
Interrelationships among Risk Indicators
Defining Child Maltreatment
Conclusion

Purpose and Goals
Hypotheses

Methods

Participants

Recruitment

Data Collection

Measures
Parental Depression
Parental Alcohol Consumption
Perceived Child Difficulty
Family Conflict
Socioeconomic Status
Parental Physical Aggression

Missing Data Estimation
Longitudinal Data Estimation
Cross-sectional Data Estimation
Bias in Estimation

Analyses

iv

33
33
34
35
35
35
36
37
38
38
39

41
42

45
45

Results

Correlations among Change Variables
Correlations among Concurrent Variables
Relationship between High- and Low-Aggression Parents
& Risk Indicators
Relationship between Change in Risk Indicators & Degree of
Physical Aggression
Relationship between Concurrent Risk Indicators & Degree
of Physical Aggression
Predicting Physical Aggression from Partner’s Risk Indicators

Discussion

Risk Levels & Aggressive Parenting

Change versus Stability in Predicting Parental Aggression
Perception of Child Difficulty
Conflict

Stability in Risk Indicators

Differences between Fathers and Mothers

Summary

Implications

Risk Markers for Parental Aggression
Efi‘ects of Parental Aggression on Children
Intervention

Future Directions
Strengths & Limitations

Appendices
References

49
58
59

62

69

7O
74

74
76
77
79
82
84
86
89

90
91
91
94
95
96

102
149

LIST OF TABLES

Table 1: Correlations among Fathers’ Risk Indicators & Stable Characteristics 50
Table 2: Correlations among Mothers’ Risk Indicators & Stable Characteristics 52
Table 3: Correlations among Fathers’ & Mothers’ Variables 55

Table 4: Correlations between Fathers’ Concurrent Variables & Physical
Aggression 60

Table 5: Correlations between Mothers’ Concurrent Variables and Physical
Aggression 61

Table 6: Multivariate Analysis of Variance for High- and Low-Aggression
Parents 64

Table 7: Summary of Hierarchical Regression Analyses for Physical Aggression
Regressed onto Stable Background Characteristics and Concurrent
Predictor Variables 71

Table 8: Predicting Physical Aggression from Partner’s Risk Indicators 75

Table 9: Variable Means, Standard Deviations, & Range of Predictor
Variables (Appendix A) 103

Table 10: Means, Standard Deviations, & Range of Mothers’ Predictor
Variables (Appendix B) 104

Table 11: Means, Standard Deviations, & Range of Fathers’ Predictor
Variables (Appendix C) 106

Table 12: Correlations among Fathers’ Change Variables, Background
Characteristics & Physical Aggression (Appendix D) 108

Table 13: Correlations among Mothers’ Change Variables, Background
Characteristics & Physical Aggression (Appendix E) 109

Table 14: Summary of Hierarchical Regression Analysis for Physical
Aggression Regressed onto Stable Background Characteristics and
Predictor Variables (Appendix F) 110

Table 15: Summary of Regression Analyses for Physical Aggression
Regressed onto Change Variables Separately (Appendix G)

Table 16: Summary of Hierarchical Regression Analysis for Physical
Aggression Regressed onto Predictor Variables (Appendix H)

Table 17: Summary of Hierarchical Regression Analysis for Physical
Aggression Regressed onto Predictor Variables and Stable
Background Characteristics

Table 18: Summary of Regression Analysis for Physical Aggression
Regressed onto Concurrent Variables Separately (Appendix J)

Table 19: Summary of Hierarchical Regression Analysis for Physical
Aggression Regressed onto Concurrent Predictor Variables

Table 20: Summary of Hierarchical Regression Analysis for
Physical Aggression Regressed onto Concurrent Predictor Variables
and Stable Background Characteristics (Appendix L)

Table 21: Means, Medians, & Missing Cases for Original Sample
(Appendix M)

Table 22: Original Dataset without Estimation of Missing Values
(Appendix N)

Table A1: Correlations among Fathers’ Risk Indicators and Stable
Background Characteristics (Revised)

Table A2: Correlations among Mothers’ Risk Indicators &
Stable Background Characteristics (Revised)

Table A3: Correlations between Fathers’ & Mothers’ Variables (Revised)

Table A4: Correlations between Fathers’ Concurrent Variables & Physical
Aggression (Revised)

Table A5: Correlations between Mothers’ Concurrent Variables & Physical
Aggression (Revised)

Table A6: Repeated Measures Multivariate Analysis for High- and
Low-Aggression Parents (Revised)

vii

112

113

114

115

116

117

119

120

128

130

132

134

135

136

 

Table A15: Summary of Regression Analyses for Physical Aggression
Regressed onto Change Variables Separately (Revised)

Table A16: Summary of Hierarchical Regression Analysis for Physical
Aggression Regressed onto Change Predictor Variables (Revised)

Table A14: Summary of Hierarchical Regression Analyses for Physical

Aggression Regressed onto Stable Background Characteristics & Change

Predictor Variables (Revised)

Table A17: Summary of Hierarchical Regression Analysis for Physical
Aggression Regressed onto Change Predictor Variables & Stable
Background Characteristics (Revised)

Table A18: Summary of Regression Analyses for Physical Aggression
Regressed onto Concurrent Variables Separately (Revised)

Table A19: Summary of Hierarchical Regression Analyses for Physical
Aggression Regressed onto Concurrent Predictor Variables (Revised)

Table A7: Summary of Hierarchical Regression Analysis for Physical
Aggression Regressed onto Stable Background Characteristics &
Concurrent Predictor Variables (Revised)

Table A20: Means, Standard Deviations, & Range of Values for
Mothers and Fathers (Revised)

viii

138

139

140

142

143

146

148

LIST OF FIGURES

Figure 1: Interaction between Parent Sex & Amount of Alcohol Consumed 65
Figure 2: Interaction between Aggression Level & Child Extemalizing

Behaviors 67
Figure 3: Interaction among Conflict in the Home Environment, Parent Sex,

& Aggression Level 68
Figure 4: Fathers’ Risk Group Distribution 123
Figure 5: Mothers’ Risk Group Distribution 124
Figure 6: Distribution of Sampling Groups into Risk Categories (Original) 125

Figure 7: Distribution of Sampling Groups into Risk Categories (Revised) 126

ix

Introduction/Rationale

Our increasingly sophisticated understanding about parenting and its determinants
highlights that parenting practices are multiply determined (e.g., Sameroff, 1975;
Ammerman, 1990; Belsky, 1980, 1993; Cicchetti & Rizley, 1981; Sameroff & Seifer,
1983). From the earliest years of research on child maltreatment, there has been
considerable effort directed toward discovering the etiology of and identifying reliable
risk factors for abuse. Such elements are critical in (a) screening for high-risk groups in
order to provide preventive interventions, and (b) recognizing abused children and their
families so that appropriate treatments can be implemented to remediate the short- and
long-term deleterious effects associated with maltreatment (Ammerman, 1991).

In the last two decades, studies (see Ayoub, Willett, & Robinson, 1992, for a
review) have demonstrated that families at risk for physical abuse could indeed be
identified in the general population, but the identification of characteristics that
differentiate the high-risk from the low-risk family is an area of both ongoing controversy
and success. Research (e.g., Belsky, 1984, 1993) has allowed the identification of a
constellation of contributing factors associated with risk of physical maltreatment,
including previous maltreatment, parental psychopathology, stress and social support,
socioeconomic status, intelligence, child personality variables, and various other
demographic variables (e.g., single parenthood).

Despite the identification of such risk indicators for physical maltreatment, little is

known about how they manifest over time within the family context. Cross-sectional

data, upon which most maltreatment research is based, does not illuminate the nuances
involved in the changing family environment and how developmental changes in risk
factors may predict the emergence of aggressive parenting. Currently there is little
information regarding the developmental course of risk indicators associated with child
maltreatment, and whether a particular “developmental map” of risk differentiates
between maltreating and nonmaltreating families.

The question left unaddressed in the current child maltreatment literature is
whether a “developmental map to parental physical aggression” exists, and whether it is
characterized by changes in the family context which increase or decrease risk, or by
stability in risk levels of a “sufficient” degree. To fully characterize family contexts in
regard to risk for parental physical aggression, stability or changes in risk must be
examined repeatedly over time in a developmental fi'amework. In addition, the intensity
level of each risk indicator (e.g., amount of alcohol consumption, degree of family
conflict) must be assessed for its contribution to the at—risk nature of the family context.
By inspecting such contextual changes, one can investigate whether there is heterogeneity
across families in terms of risk levels, or whether physically aggressive families share the
same history of changes in risk indicators. Understanding how such factors affect
parenting over time can have a profound impact on our ability to successfully intervene
with at-risk parents and to create interventions that are more sensitive to the “long-term”

developmental view of the parenting context.

Theog

Various theoretical models have been proposed to explain the etiology of child
maltreatment. The psychopathology model (e.g., Steele & Pollock, 1968), which
attributed child abuse and neglect to psychiatric disorders or personality disturbances,
has fallen out of interest due to the failure of many studies to identify a syndromal pattern
specific to maltreating parents (Wolfe, 1985). In the early 1970's, the socio-cultural
model (e.g., Gelles, 1973; Gil, 1970) emerged to emphasize the role of stress engendered
by poverty, unemployment, and educational disadvantage. The related social-
interactional model (e.g., Parke & Collmer, 1975) proposed that child and parental
characteristics combine with situational factors that lead to abuse. Thus the probability of
maltreatment increases when child factors (e. g., severe behavior problems) interact with
parental variables (e.g., low frustration tolerance) during situations of conflict (e.g.,
bedtime). This growing focus on multiple contributing factors, rather than on a single
cause, stimulated interest in increasingly integrative models of causation.

The next generation of maltreatment models are characterized by increased levels
of complexity, which resulted from attempts to simultaneously consider multiple
causative factors. Belsky (1980) derived a child abuse model from Bronfenbrenner’s
general ecological model, which delineates four levels of causative influence in etiology:
ontogenetic, microsystem, exosystem, and macrosystem. Ontogenetic variables are those
characteristics of the individual that contribute to (or prevent) maltreatment, such as
parental IQ and poor parenting skills. The microsystem includes those aspects of the

family that increase (or decrease) the probability of maltreatment, such as marital conflict

or children with behavioral problems. The exosystem involves community or social
forces, such as unemployment. Finally, the macrosystem consists of cultural
determinants of maltreatment, such as societal acceptance of corporal punishment or
economic prosperity.

Cicchetti and Rizley (1981) went beyond the ecological model in proposing the
transactional model, in which potentiating and compensatory factors for maltreatment are
examined. Temporal distinctions are made for both categories of risk factors. That is,
there are transient risk factors that fluctuate and may indicate a temporary state, and there
are enduring factors, which represent more permanent conditions or characteristics. In
addition, biological, psychological, and social elements are hypothesized to combine in
influencing the etiology of maltreatment. Cicchetti and Toth (1987; 1995) and others
maintain that at the level of the microsystem, the balance of potentiating and
compensatory factors, or stressors and supports (Belsky, 1980, 1993; Garbarino &
Sherman, 1980), determines the presence or absence of violence within the family
environment. In addition, the effects of multiple, non-specific risk factors may be
cumulative in the sense that the presence of more risk factors is related to a higher
certainty of negative outcome (Seifer, Samerofi‘, Baldwin, & Baldwin, 1992).

Wolfe (1987) built on these integrative models with his transitional model. As
with the ecological and transactional models of maltreatment, the transitional model
focuses on the importance of multiple causes and risk/protective factors. However, the
transitional model views child abuse as the extreme end of a continuum of adequate to

deviant parenting, in which three stages of parent-child conflict are described that

progressively heighten the probability of abuse: Stage One, reduced tolerance for stress
and disinhibition of aggression; Stage Two, poor management of abuse crises and
provocation; and Stage Three, habitual patterns of arousal and aggression with family
members. In each stage, destabilizing factors facilitate the escalation of conflict, whereas
compensatory factors prevent passage into other stages. Thus abuse is viewed as a
heterogeneous form of physically punitive parenting, rather than a dichotomous
phenomenon. According to this view, the parent-child relationship was either never well
established from the beginning, or it began to disintegrate during periods of
developmental change or family stress (Wolfe, 1993).

These models of maltreatment integrated the various factors of early
unidimensional models into more complex, multidimensional models. In addition, they
have accounted for the dynamic nature of the family context. Although there are
important differences in the models described in the section above, all of these theoretical
approaches focus on the importance of interactions and changes over time as pertaining to
the emergence of child maltreatment. Yet most studies utilizing these complex theories
continue to produce af‘snapshot” of family risk factors and functioning, based on cross-
sectional data or a retrospective research design (Azar, Fantuzzo, & Twentyman, 1984;
Hillson & Kuiper, 1994; Pianta, 1984), which delimits their explanatory depth. Azar (as
cited by Hillson & Kuiper, 1994) described these maltreatment studies as “... merely lists
of the components of single factor theories with little attempt to specify contingent

relationships between components or prioritize their contribution to causality.”

Developmental-Ecological Framework

For the present study, a developmental-ecological framework is applied. From a
developmental perspective, etiology is viewed as an evolving process with the possibility
of continuities and discontinuities in predictor variables as children and families change
over time (Fitzgerald et al., 1993). An ecological approach dictates that etiology is
assessed around multiple “contexts of maltreatment” (Belsky, 1993), including the
developmental context (the role of parent and child factors) and the immediate
interactional context (such as conflict in the family environment). Combining the
strengths of developmental theory with an ecological framework, risk is assessed fi‘om a
related set of variables across time. Researchers using this perspective seek to document
the relationship of maturation to environmental trigger events, to establish the extent to
which the basic structure of developmental processes can be modified by environmental
experiences, and to describe the person-environment interchanges that are necessary to
bring about change or to resist it (Zucker et al., 1995). Thus, this framework is concerned
with constancy and change, and the desire to map the multiple, interacting factors that
produce behavior (Zucker et al., 1995).

As such, the developmental-ecological framework is particularly suitable for
examining the flow of events in the family context and how certain patterns of interaction
of risk factors may produce child maltreatment. Both parent and child develop through a
series of stages, at any of which the potential for conflict and abuse arises (Cicchetti &
Toth, 1987; Hoekstra, 1984) via parental, child, and environmental characteristics

(Belsky, 1980; 1993). Capturing the developmental dynamics of risk factors in family

contexts will provide insight into the types of challenges that are most likely to
predispose a family for eventual maltreatment, and answer the need to move beyond
purely descriptive cases of maltreatment and risk factors to more complete process-
oriented explanations of etiology (Hillson & Kuiper, 1994; Hoekstra, 1984).

The Family Context: Change or Stabilng in Risk Status?

When assessing the conditions that give rise to maltreatment, scientific evidence
is needed to ascertain whether there is a stable vulnerability component to families that
eventually maltreat their children, or if families move in and out of risk status. That is, if
risk is conceptualized as the additive effect of multiple factors in the family system, do
those factors change within families over time? And is it possible to distinguish
physically aggressive parents from nonaggressive parents by differences in families’ “risk
history”?

Experts express concern about whether parental, situational, family, and
individual change in functioning occurs over time, and whether it affects the presence or
return of maltreatment (Ayoub et al., 1992). Some evaluation-based studies have
reported stability in family functioning across time. For example, Kowal et a1. (1989)
found that family functioning at entry of an intervention is significantly correlated with
family ftmctioning at exit, indicating that family functioning tends to be moderately
stable. Pianta and Egeland (1990) found that relationship patterns between mothers and
infants at six months were consistent at 42 months. Pianta et a1. (1989) observed that
76% of the mothers identified as maltreating when their offspring were two years of age

were still maltreating four years later, suggesting that the stability of maltreatment may be

 

quite high (Belsky, 1993). Others (Roberts, Block, & Block, 1984) have also reported a
considerable amount of consistency in parenting styles from child’s age three to twelve.

Such reports indicate that family contexts are relatively stable, suggesting that
level of risk for maltreatment does not change within families. However, other studies
(Cicchetti & Howes, 1991; Justice & Duncan, 1976) have found variations in family
ftmctioning over time related to increased “movement” in the abusive family context, in
that parents must cope with multiple aversive events in their environment (Spicer &
Franklin, 1994). Specifically, it has been shown that abuse and neglect correspond with
periods of stressful role transition for parents, such as the postnatal period of attachment,
the early childhood period of increasing socialization pressures, the times of family
instability and disruption, and the times following detachment from social supports and
services (Belsky, 1980; Wolfe et al., 1988). Behavioral patterns that differentiate abusive
from nonabusive mothers may develop during infancy (Gelardo & Sanford, 1987), but
Egeland and Farber (1984) provide evidence that such behavioral patterns do change over
time, related in part to either increases or decreases in life stress. Willett, Ayoub, and
Robinson (1991) found that the number of parenting problems and the number of
violence/maltreatment problems experienced by a family were predictive of changes in
functioning over time, reporting that there is considerable variation in how much change
in family functioning is observed (Ayoub & J acewitz, 1982).

Such studies indicate that global family functioning is subject to change over
time, although they do not provide information about the respective changes in family

risk indicators. However, several studies have provided indirect evidence that risk factors

change over time. For example, Crnic, Greenberg, Robinson, and Ragozin (1984) found
that maternal risk factors were only moderately stable across an eighteen-month time
span. Focusing specifically on risk factors in families who participated in a parent
education and support program, Whipple and Wilson (1994) found that maternal
depression decreased significantly over time, as did negative life changes. Such findings
suggest that maltreating families experience changes in the intensity of particular risk
indicators. However, it is important to note that such findings are tied to studies devised
around interventions.

Despite these findings, the literature lacks a clear picture of the pathways that lead
to parental aggression. Several options are feasible, including the possibility that family
risk levels are stable across time: highly physically aggressive parents experience
consistently high levels of risk relative to non-physically aggressive families, who
experience consistently low levels of risk. Evidence for such a pathway would lead to a
dichotomous view of families, with risk status that is unlikely to change over time.
However, an alternate route to parental physical aggression may be a family context that
is characterized by change in the levels of risk factors. Such a pathway indicates a high-
risk status that is related to turmoil and change, while a low-risk family history may be
associated with stable risk levels.

Risk Indicators

The importance of different types of risk factors—social, emotional, and

economic-in the etiology of child maltreatment is well documented (Egan, 1983;

Garbarino & Crouter, 1978; Gelfand, Teti, & Fox, 1992; Justice & Duncan, 1976;

Koverola, Manion, & Wolfe, 1985; Muller, Fitzgerald, Sullivan, & Zucker, 1994; Pianta
& Egeland, 1990; Schinke et al., 1986; Whipple & Webster-Stratton, 1991). Despite
knowledge about the types of risk factors that are associated with child maltreatment,
greater attention is needed to address the developmental processes by which some parents
gradually acquire the preconditions that lead to maltreatment. Wolfe (1993) maintains
that rather than focusing on observable factors that are often present once a family has
been reported, researchers must look at the process by which more subtle, preexisting
factors associated with the individual parent, child, or family situation become
transformed over time into a high-risk or maltreating situation.

It is the relationships among these factors, not the individual factors per se, that
generate patterns of risk in the family system. However, understanding the individual
components that create risk is imperative. As such, several risk indicators that are
associated with child maltreatment are described in greater detail in the following
sections. In keeping with an ecological philosophy about the etiology of maltreatment,
risk factors are considered from individual, microsystem, and social contexts. Research
on parental depression and alcoholism, perception of child difficulty, conflict in the
family environment, and socioeconomic status has revealed that these risk factors are not
static and constant over time, but rather are variable and subject to fluctuation in people’s
lives. How these possible fluctuations over time relate to eventual parental physical
aggression is presented as the consideration in focus. In addition, differences between
mothers’ and fathers’ experiences of risk factors, and whether such differences affect the

outcome of physical aggression, are examined.

10

mm

During the 1960's parental psychopathology was implicated as the primary cause
of child abuse and neglect. Since then, however, the idea that all abusive parents are
mentally ill has been rejected (Wolfe, 1987), and in fact only 10% to 15% of abusive
parents are diagnosed with a specific psychiatric condition (Ammerman, 1990; Bavolek
& Henderson, 1990). Yet parent variables are highly predictive of inadequate care in
high-risk families (Egeland & Brunnquell, 1979), and they are more powerful
contributors to maltreatment than are child characteristics (Ammerman, 1990). Thus, the
role of parental psychopathology continues to deserve examination.

Despite the fact that no particular “abusive” personality has been discovered
(Ammerman, 1990), parental psychopathology is associated with physical child abuse and
is a major burden for families with dependent children. Depression is the most widely
studied disturbance in parental psychological functioning (Belsky, 1984; Vondra, 1990)
as a well-known risk factor for impaired parent-child interactions, including abuse and
neglect (Bland & 0m, 1986; Egeland, Jacobvitz, & Sroufe, 1988; Gelardo & Sanford,
1987; Gelfand et al., 1992; Pianta & Egeland, 1990; Whipple & Webster-Stratton, 1991;
Whipple & Wilson, 1994). Early identification of maternal depression in particular is
shown to be one promising strategy in child abuse prevention (Scott, 1992).

A predictable association exists between parental depression and child
maltreatment in that depression is closely associated with stress, and that depressed mood
of parents may result in decreased effectiveness in handling discipline situations (Ketch

et al., 1995; Whipple & Webster-Stratton, 1991). Depressed mothers are found to make

11

more negative statements about their children, to use guilt- and anxiety-provoking
methods of controlling their children, and to have unrealistic expectations regarding their
children’s ability to self-regulate (Ammerman, 1990; Jameson, Kulcsar, Gelfand, & Teti,
1995). Depressed mothers create a disruptive, hostile, rejecting home environment for
their children, which in turn undermines child functioning (Campbell, Cohn, & Meyers,
in press; Colletta, 1983). Moreover, an examination of mothers who were able to break
the cycle of intergenerational abuse revealed that depression significantly differentiated
the abusers from the nonabusing mothers (Egeland et al., 1988). The depression scores
obtained by the mothers who continued the intergenerational cycle of abuse highlighted
both (a) depression that was chronic and intrapsychic in nature, and (b) depression that
was situational and reactive due to fi'ustrations and helpless feelings.

Researchers (see Dore, 1993; Whipple & Wilson, 1994) have concluded that
parents who are psychologically distressed have a much lower threshold of tolerance for
child misbehavior and resort more quickly to high authoritarian, overcontrolling
parenting. Gelfand et al. (1992) affirmed that the greater the severity of maternal
depression, the more likely were depressed mothers to experience parent-related stress.
Most important, depressed mothers tend to be more hostile and punitive in their responses
to children at times (Davies et al., 1991; Zucker et al., 1994), and more withdrawn and
psychologically absent at other times (Dore, 1993). Evidence suggests that negative
efl‘ects of parental depression on children are related more to such patterns of inconsistent
parenting of depressed mothers than to the depression itself (Dore, 1993).

Little is known about how changes in depressive episodes affect parenting.

12

Examining such changes relative to parental physical aggression is important because the
vast majority of episodes of depression in adulthood are recurrences (Kessler & Magee,
1994)

9mm in 12an depression over time. Due to the episodic nature of
depression, it is critical to consider the timing of assessment. Much of the work on
maternal depression has focused on women who have suffered an episode of depression
during the child’s lifetime, but the depression was not necessarily current at the time of
the assessment. The data have been equivocal: although one study found that a history of
depression had an impact on later mother-child interactions, even when the depression
was no longer current, others suggest that current symptom levels are more important
than a history of depression (Campbell et al., in press; Susman, Trickett, Iannotti,
Hollenbeck, & Zahn-Waxler, 1985).

Other studies that have examined changes in maternal depression and parenting
have found that mothers with past depression are more oriented toward letting their
children take chances and try new things as they grow up (Susman et al., 1985). Davies
and colleagues (Davies et al., 1989) found tentative evidence that when a parent has been
depressed earlier in life, it is probably not a risk factor for disturbance in the later parent-
child relationship. However, higher levels of self-reported current depression were
associated with more negative affective parenting (Egeland et al., 1988; Gelardo &
Sanford, 1987; Gelfand et al., 1992; Pianta & Egeland, 1990; Whipple & Webster-
Stratton, 1991; Whipple & Wilson, 1994).

Additional evidence for the importance of assessing impact of changes in maternal

l3

depression on parenting comes from the multitude of studies on attachment of the
mother-infant dyad (see Egeland & Farber, 1984). Egeland and Farber (1984) revealed
that maternal affective behaviors are particularly important for maintaining a secure
attachment once it is formed, and that mothers’ personality variables associated with

depression (e.g., aggression and hostility) have a negative impact on the mother-child

 

relationship over time.

WW. Studies have found that for both mothers and
fathers, current depression is associated with more negatively affective parenting (Davies
et al.,1989). Further, it has been found that both fathers and mothers who report greater
ratings of current and worst-ever depression have higher rates of aggression toward their
children (Reider et al., 1989; Zucker et al., 1994). Interestingly, Davies and colleagues
(1989) found that maternal and paternal depression operate in different ways to affect
parenting. Specifically, this study found evidence that there is independence between the
father’s parenting and the mother’s depression (i.e., maternal depression did not afi‘ect the
father’s parenting), but that the mother’s parenting is strongly afl‘ected by the father’s
level of depression.

MW

Alcoholism and problem drinking are positively correlated with violence among
family members (Barber & Crisp, 1994; Bavolek & Henderson, 1990; Famularo, Stone,
Barnum, & Wharton, 1986; Muller et al., 1994; Pianta, 1984; Reider, Zucker, Maguin,
Noll, & Fitzgerald, 1989; Steinberg, Catalano, & Dooley, 1981; Whipple & Wilson,

1994). Studies of the link between parental substance abuse and child maltreatment

14

 

suggest that chemical dependence is present in at least half of the families known to the
public child welfare system (Dore, Doris, & Wright, 1995). Some place this figure as
high as 80% (Barth, 1994).

Researchers have found significant relationships between parental alcohol abuse
and child maltreatment of all types, including physical, sexual, and emotional abuse as
well as physical and emotional neglect (see Dore et al., 1995). In one in-depth
comparison of substance abusing and non-substance abusing parents involved in child
protective cases, Murphy and colleagues (1991) found that parents with documented
substance abuse histories were more likely than other parents to: (a) be repeat offenders
with regard to child abuse and neglect; (b) fail to follow through with court-ordered
services; and ( c) eventually lose care and custody of their children.

Nevertheless, the evidence connecting alcohol and child abuse has weaknesses
(Steinhausen, 1995). Although increased rates of physical abuse have been reported
among alcoholic parents, data also suggest that parental withdrawal is more common than
physical abuse (Zucker, 1979). In a national survey, Wolfrrer and Gelles (1993) found no
significant difl’erences for self-reported rates of maltreatment among abstainers, those
who drink 1 - three times/month, and those who drink at least weekly. However, profiles
of at-risk families frequently include histories of and current drug or alcohol abuse (Bath
& Haapala, 1993; Bavolek & Henderson, 1990; Kowal etal., 1989; Muller et al., 1994),
and some researchers believe that parental alcoholism is one of the most destructive of
many risk factors for physical child abuse (F amularo et al., 1986).

A number of explanations have been offered for the hypothesized connection

15

between physically aggressive parents and alcohol problems, including that aggressive
parents, in attempting to cope with the stresses of their daily lives, resort to behaviors
involving alcohol use or exceedingly harsh punishment of their children (Wolfe, 1987).
Another possibility relates to expectations of interpersonal interactions, such that under
the influence of alcohol, individuals who are already highly stressed may misinterpret
cues and resort to aggression (Widom, 1993).

One critical element that may determine the pairing of alcohol consumption with
parental physical aggression is the presence of comorbidity of alcoholism with antisocial
personality disorder (ASPD). Abusive parents are more likely to receive diagnoses of
alcoholism and ASPD (Bland & Om, 1986; Dinwiddie, 1993). Even without diagnostic
labels of alcoholism and antisocial personality disorder, child abusers in general are found
to have a history of disciplinary problems and property destruction, and as adults to
engage in other violent behaviors. It could be that these maltreating parents are
“contextually embedded” (see Zucker & Fitzgerald, 1991) as individuals who are
predisposed for aggression or violence and demonstrate patterns over the lifespan that
reflect a bias toward attracting reinforcing factors.

Although ASPD occurs in only 4 percent of the US. male noninstitutional
population, it is 12 times more common among those with alcohol dependence (Zucker,
Ellis, Bingham, & Fitzgerald, 1996), and is the psychiatric disorder with the strongest
association with alcohol disorders. This dose of increased aggressivity can play a role for
both males and females with alcohol problems (Department of Health and Human

Services Special Report, 1993), although it is not present for all alcoholics (i.e., ASPD is

16

found in approximately 14% of the alcohol abuse/dependence population—see Zucker &
Fitzgerald, 1991). This heterogeneity in alcoholism may give alcohol consumption
several different pathways to aggression, with antisociality as a distal factor in a causal set
of processes in only some cases.

cages in alcohol cogumption over time. Research shows that alcoholism is

not a static component, but rather demonstrates variable patterns of behavior across time.

 

Longitudinal studies of alcoholic behaviors have shown that there are changes in patterns
of use, both in frequency and intensity (Paolino, McCrady, & Kogan, 1978). A multitude
of studies (see Zucker et al., 1995) report significant shifts into and out of problem
drinking classifications over intervals as short as one year, suggesting that consumption
level and related problems are unstable over longer intervals of time. Studies showing
the greatest stability across time were those where the initial level of risk was greatest
(Bingham, Zucker, & Fitzgerald, 1996). Risk may be more fluid over time in populations
involving low to midrange levels of risk (Zucker et al., 1996).
How such patterns of consumption relate to child maltreatment has been

' investigated: periods of active drinking increase the likelihood of direct child abuse, via
parental violence and severe incapacity (F amularo et al., 1986). In addition, within
aggressive parent populations there is an overrepresentation of parents with a past history
of alcohol problems. Reider et al. (1989) found that both mothers’ and fathers’ degree of
long-term alcohol-related difficulty was positively correlated with severity of aggression
toward their children. Those parents with greater long-term alcohol-related troubles had

higher rates of aggression toward their children, but current rates of drinking were not

17

— t

 

 

significantly related. Other studies have also found that current alcohol consumption was
not related to parenting measures (Davies, Zucker, Noll, & Fitzgerald, 1991). The finding
that a history of alcoholic drinking in the past is a risk factor for current maltreatment
reveals the degree to which alcoholism may have a deep and wide-ranging impact on
parental functioning, even beyond the immediate effects of actual consumption.

Ema; and maternal alcohol consummgn. Fathers’ problem drinking is linked
to family stress and less positive father-child relationships (Durnka & Roosa, 1993;
Zucker, Ojala, Fitzgerald, & Noll, 1994). In addition, fathers’ problem drinking has been
shown to interfere with positive mother-child relationships (Dumka & Roosa, 1993).
Data are particularly scarce concerning the effects of maternal alcohol consumption and
parenting (Steinhausen, 1995), although researchers (Davies, Zucker, Fitzgerald, & Noll,
1992) have predicted more severe effects with maternal alcoholism due to the
traditionally increased responsibility for child-rearing held by the mother. However, this
prediction has received equivocal support: Krauthamer (1979) found that alcoholic
mothers tend to be more ambivalent, confused, and inconsistent about parenting than
nonalcoholic mothers, while another study (Zucker et al., 1994) found that mothers’ level
of alcohol-related difficulties was not predictive of any aspect of their relationships with
their sons. Other studies (see Dumka & Roosa, 1993) have shown that alcoholic families
are most likely to be harmful to children when both parents are caught up in problem

drinking and family routines and rituals are not maintained.

18

Pergeption Qf Child Difficulty

A number of child characteristics are implicated in maltreatment in general, and in
physical abuse in particular. It has been proposed that certain child features, such as
difficult-to-manage and behavior-problem children, increase parental fi'ustration and add
to the level of overall familial stress (Crnic & Acevedo, 1995; Gelfand et al., 1992;
Koverola et al., 1985; Pianta, 1984; Whipple & Webster-Stratton, 1991; Wolfe, 1985).
Deviant child behavior and noncompliance are frequent precipitants of abuse
(Ammerman, 1990; Gelardo & Sanford, 1987; Kadushin & Martin, 1981), as children
who are abused tend to present parents with more discipline situations via disruptive and
aversive behavior (Azar et al., 1984; Whipple & Wilson, 1994). These “abuse-
provoking” characteristics include prolonged and irritating crying, oppositional and
defiant behavior, and conduct problems (see Ammerman, 1990; 1991). These child risk
factors also contribute to the maintenance of parental aggression in those families where
violence is primarily caused by other factors (Ammerman, 1991).

The importance of the child’s behavior in the etiology of abuse has been debated,
as prospective studies of abusive families have failed to demonstrate a significant
causative role for child behavior problems in abuse (Ammerman, 1990; 1991). For
example, Egeland and Brunnquell (1979) found that aspects of infant temperament (e.g.,
infant orientation, irritability, and consolability), in contrast to parental characteristics,
added little information in distinguishing between adequate and inadequate care groups.

However, parental perceptions of child behavior have a dramatically stronger

influence than actual child behavior (Mash & Johnston, 1990; Wolfe, 1985). The parent-

19

child conflict in abusive families often does not appear directly attributable to difficult
child behavior (Mash & Johnston, 1990), but children who are perceived as “difficult”
have been found to be at higher risk of abuse and to have highly stressed parents
(Garbarino & Sherman, 1980; Vietze et al., 1980). Research (Ammerman, 1990; 1991;
Dore, 1993; Houck & King, 1989; Mash et al., 1983) has shown that abusive parents’
perceptions are incongruent with observed child behavior; they overestimate the degree to
which their children show problematic behavior and expect more troublesome behavior
from their children. Whipple and Webster-Stratton (1991) found that abusive mothers are
significantly more critical in their interactions with their children and report more child
behavior problems, despite the fact that independent observations on home visits did not
reveal significant differences in the amount of child deviance in abusive and nonabusive
homes.

Physically abusive parents’ expectations and perceptions of their children differ
from those of nonabusive parents in such a way that they are more likely to choose harsh
disciplinary tactics (Egeland & Brunnquell, 1979; Haskett et al., 1995). Maltreating
parents report less satisfaction with the parenting role, have more negative perceptions of
their children, and attribute their children’s misbehavior to deliberate efforts to challenge
their authority (Azar et al., 1984; Cicchetti & Howes, 1991; Cicchetti & Lynch, 1993;
Dore, 1993; Trickett, Aber, Carlson, & Cicchetti, 1991; Creighton, 1985). Moreover,
families with unrealistic expectations and cognitive distortions regarding children’s
intentions are significantly less likely to improve even with intervention (Kowal et al.,

1989).

20

Changes over time in perceptions of child difficulg. Very few studies have
examined changes in parental perception of child difficulty and their relationship to

parental aggression (Ammerman, 1991). One study (Whipple & Wilson, 1994) that
examined families in a parent education and support program found that there were no
significant decreases in parental reports of child misbehavior and parental discipline over
the nine-month assessment span. In addition to this finding, results revealed that despite
participation in the program, parents persisted in inaccurately perceiving their children’s
abilities, had difficulty being aware of their needs, maintained strong beliefs in the value
of corporal punishment, and expected their children to be responsible for much of their
happiness. A number of other studies have pointed to similar attitudes among risk
populations and the difficulty in changing such beliefs (Barber, 1992). However, such
findings are based on interventions with groups of parents who have been specifically
identified as at-risk for maltreatment. There is little information about changes in
perception of child difficulty over time, and how this relates to parental aggression, in a
non-intervention context.

Assessing parental perceptions of child difficulty over time is important due to the
nature of the changing parent—child relationship. The particular developmental period
through which children are progressing may be a critical determinant of the stressors
associated with parenthood (Crnic & Acevedo, 1995), as the child’s age has been shown
to be a significant contributor to the prediction of maltreatment by mothers (Reider et al.,
1989). It is critical to understand the potential for stage-salient transition periods to affect

stressful experience and how that potential may predict aggressive parenting.

21

temal and maternal rce tions of child ifficul . Although studies have
shown that in general mothers and fathers equivalently rate their child’s behavior
problems (Bingham, Fitzgerald, & Zucker, 1996; Ellis, Bingham, Zucker, & Fitzgerald,
1996, under review), other studies (Fitzgerald, Zucker, Maguin, & Reider, 1994) have
found that mothers’ and fathers’ perceptions of child behavior are both different and
subject to change over time. Specifically, Fitzgerald and colleagues (1994) found that as
fathers increased the amount of time spent with their sons, parental perceptions became
increasingly concordant. Fathers who spent less time with their sons perceived them as
less troubled than mothers perceived them to be. As such, depending on amount of time
directly spent with a child, mothers’ and fathers’ perceptions of child difficulty may
follow different developmental courses. A broader literature base has shown that in
general there are low-order correlations between mothers’ and fathers’ perceptions of
child behavior.

' t ' e 'l viro n
A family-level component that affects the potential for parental physical

aggression is the presence of conflict in the family environment. High levels of family
conflict have been shown to set a foundation for coercive interactions that increase the
likelihood of child maltreatment (Mollerstrom, Patchner, & Milner, 1992). In particular,
conflict in the marital relationship has been hypothesized to play an important role in the
development of physical aggression, as marital discord has been associated with child
abuse (Barton and Baglio, 1993; Garbarino & Crouter, 1978; Pianta, 1984; Wolfe, 1985;

Gelardo & Sanford, 1987; Houck & King, 1989). Abusive mothers describe significantly

22

more distressed marriages and conflict than nonabusive mothers (Cicchetti & Lynch,
1993; Whipple & Webster-Stratton, 1991). The effects of spousal relations, including
violence among parents, general negativity and aversive behavior, on parenting has been
implicated by studies of quite different developmental periods (Wolfe, 1985; Belsky,
Lemer & Spanier, in press; Vondra, 1990). Both fathers and mothers in abusive families
report greater levels of family conflict (Perry et al., 1983), and demonstrate particular
difficulty in resolving disagreements through discussion (Silber, 1990).

In a comprehensive review of the literature, Wolfe (1985) cites studies that clearly
indicate that child maltreatment is significantly associated with observable levels of
conflict in the home, and that the tone of family interactions is less positive than in
nonaggressive homes. The consequences of such negative and conflictual family
environments are likely to be associated with poorer family system health (e.g., cohesion
and adaptability) (Durnka & Roosa, 1993).

Chggps rp' pppflipt in the m1): environment pve; p'mp. Changes in conflict
levels in the family environment over time is a construct that incorporates elements of
crisis, communication, and interaction among family members. Such global
considerations of family conflict are necessary in longitudinal studies, which include
changes in family composition, such as divorces, remarriages, and other parental partners
in the family environment. Changes in family composition may have a profound effect
on the degree of conflict in the home (Bolton & MacEachron, 1986). For example,
Kowal et al. (1989) found that the most notable risk factor accounting for worsening

mother-child relationships was a change in family formation.

23

Disintegration and reorganization of family composition have been shown to
increase the degree of stress within the family, particularly within the first two years
following a separation (Bolton & MacEachron, 1986), and are likely to alter the level of
conflict in the family environment. Moreover, heightened abuse potential is significantly
associated with single-parent or nonparent custody (Kolko et al., 1993), and assessing
changes in such family characteristics over time may reveal fluidity in risk status.
mm

Of the several extrafamilial factors investigated in the etiology of child
maltreatment, none has received greater research attention than economic stress.
Although there has been long-lasting controversy regarding whether differences in SE8
affect the etiology and prevalence of abuse (T rickett et al., 1991), support for the
connection between maltreatment and economic factors has been gleaned at both the
aggregate and individual levels of analysis. Poverty and its associated stressors of low
education and unemployment have repeatedly been found to negatively impact
childrearing (Garbarino & Crouter, 1978; Trickett et al., 1991; Whipple & Wilson, 1994)
by providing a stressful context for abuse (Whipple & Webster-Stratton, 1991; Zuravin,
1989).

Measures of poverty and/or unemployment have been found to clearly distinguish
between physically abusive and nonabusive families (Cantrell, Carrieo, Franklin, &
Grubb, 1990; Cicchetti & Lynch, 1993; Gelardo & Sanford, 1987; Glachan, 1991;
Whipple & Webster-Stratton, 1991) and families who reabuse (F erleger, Glenwick,

Gaines, & Green, 1988). Likewise, unemployment, educational underachievement, and

24

low family income are related to intrafamily violence in general (Ammerman, 1990;
Kolko et al., 1993). Economic change at the community level is reflected in aggregate
abuse records (Steinberg et al., 1981), and overall rates of abuse are higher in regions
characterized by high proportions of low-income families and in areas with high
unemployment rates (Barton & Baglio, 1993; Cantrell et al., 1990; Garbarino & Crouter,
1978b; Steinberg et al., 1981).

Despite the preponderance of evidence, the role of SES in the etiology of
maltreatment is debated in the literature. It is likely that the relationship between low
SES and maltreatment is inflated because lower SES families come to the attention of
investigative authorities, given their more frequent contact with social service agencies
when contrasted with higher SES families (Ammerman, 1990; Trickett et al., 1991). Low
SES is neither a sensitive nor a specific marker for child abuse and neglect: child
maltreatment occurs in all socioeconomic groups (Ammerman, 1990), and the majority of
underprivileged families do not engage in maltreatment (Houck & King, 1989).

However, recent studies have indicated that although there may be a reporting bias
such that lower-SES abusive families are proportionately more likely to be reported to the
authorities than are middle-class abusive families, it is likely that maltreatment in fact
occurs more frequently among lower-SES families (T akeuchi, Williams, & Adair, 1991;
Trickett et al., 1991). In a national survey, poor respondents reported 1.5 times more
abusive violence than more affluent families (W olfirer & Gelles, 1993).

Regardless of status of low socioeconomic standing, job loss and economic

instability—circumstances likely to befall maltreating families (Steinberg et al.,

25

l981)—may be especially detrimental to family frmctioning (Pianta, 1984; Vondra, 1990).
In fact, financial problems consistently emerge as the most prevalent risk factor in high-
risk families (Herrenkohl, Herrenkohl, Rupert, Egolf, & Lutz, 1995; Kowal et al., 1989).

Changes in spcipeconomic stag overtime. The theoretical connection between
SES and child maltreatment is based on the well-established generalization that
psychological distress varies inversely with level of income (Catalano & Dooley, 1983).
However, the notion of SES as a stable variable is not accurate. From cross-sectional
data, one might deduce that maltreating families are either “in” or “out” of financial
troubles; one does not get a picture of how SES may change at the family level. Some
studies (Bingharn et al., 1996) that have examined multiple-risk families have assessed
changes in SES and family income. Stability for family SES measured approximately
three years apart was r = .69, while family annual income was r = .71. Both numbers
indicate some degree of change across time. Indeed, it has been shown (Takeuchi et al.,
1991) that socioeconomic status is a dynamic condition, with families entering and
exiting poverty every year.

Early work (Justice & Duncan, 1976) found that abusive parents often go through
many ups and downs in terms of income and expenditures, and that both
contemporaneous and lagged relationships between financial strain and maltreatment
exist (Zuravin, 1989). Gil (1970) reported that maltreatment is more likely among those
families in which the father experienced unemployment during the preceding year.

Several studies have shown that econorrric and financial changes at any SES level

cause stress upon families, and that both negative and positive changes in SES cause

26

stress. Thoits and Harman (1979) examined the relationship between change in income
and stress levels by studying the impact of an income-maintenance experiment, for which
participants’ income level was increased. Several race, sex, and marital status groups
responded to the income-maintenance treatments with significantly increased distress.
These responses of increased distress occurred early in the experiment, and increased
over time—even when controlling for magnitude of financial change. Thus, the magnitude
of the financial change appeared to be less important than the change itself (Thoits &
Harman, 1979). In addition, it had been previously shown that the experiment
significantly increased the divorce rate among participants, their rate of unemployment,
and their rate of geographic mobility (see Thoits & Harman, 1979). However, no research
to date has focused on the processes that could account for a relationship between
changes in SES and abuse rates, and recent literature (Wolfe, 1993) agrees that the effects
of socioeconomic stress and family disadvantage have not received adequate attention.

v 10 F ew rk f 'sk: In errelatio hi am n 'sk Indicat rs

Developmental theory emphasizes the importance of viewing patterns of

adaptation and change as dynamic systems operating in multiple contexts and through
time (see Zucker & Fitzgerald, 1991). Utilizing such a lens when considering child
maltreatment etiology, it is highly likely that risk indicators interact over time to create
multiply-determined contexts that may foster maltreatment. Theories that attempt to
explain the etiology of maltreatment concur that individual risk factors rarely function in
isolation, and that a single risk factor does not account for a causal path to maltreatment.

Most maltreating family contexts demonstrate the presence of multiple risk indicators

27

simultaneously, and as the number of concurrent risk factors increases, the likelihood for
negative individual outcomes also increases (e.g., Sameroff & Seifer, 1983). This is
related to additive risk cumulation theory (Cicchetti & Toth, 1995) and the concept of risk
load (see Bingham et al., 1996).

It is likely that families in which risk factors are of a particularly intense level are
more prone to experience multiple risk factors for maltreatment. To illustrate, depression
occurs in a context that more often than not includes marital conflict, low social support,
and low socioeconorrric status (Campbell, Cohn, & Meyers, in press; Gelfand et al., 1992;
Jameson et al., 1992). In addition, maternal depressed mood is related to increasingly
negative perceptions of child behavior (Krech & Johnston, 1992). Poverty-stricken
parents are subject to the special environmental distresses and strains associated with
socioeconorrric deprivation (Justice & Duncan, 1976) which incorporates a multitude of
risk factors, including increased psychological distress and increased marital and family
conflict (Cantrell et al., 1990).

Partners in alcohol-complicated marriages in general experience greater marital
distress, marital violence, and marital dissolution than non-alcohol marriages (Dumka &
Roosa, 1993), and there are significant relationships between fathers’ and mothers’
alcoholism and depression (Reider et al., 1989) and between paternal alcoholism and
maternal depression (Davies et al., 1991; Fitzgerald et al., 1993; Zucker et al., 1994).
Alcoholism is also related to increased aggression toward children and spouses, family
stress and arguments, decreased attention to children’s basic needs, decreased job

performance, lowered income, and poor health (Department of Health and Human

28

Services Special Report, 1993; F amularo et al., 1986). Also, maltreatment is more highly
correlated with alcoholism when alcoholism is paired with recurrent depression (Bland &
Cm, 1986) and family conflict (Dumka & Roosa, 1993). At the rrricrosystem level,
changes in degree of family conflict may represent the manifestation of increasing
troubles with alcoholism, depression, change in family composition, and/or stresses with
finances (Bolton & MacEachron, 1986).

Despite such evidence of multiple risk simultaneously impinging upon the family,
theories that focus on an additive nature of risk cumulation or risk load seem to suggest a
sense of stable vulnerability in the family context, without indicating whether risk status
is fluid. Further, due to the limits of cross-sectional designs, many studies cannot map
out patterns of change regarding risk level in the family system. However, Zucker, Ellis,
Bingham, and Fitzgerald (1996) draw attention to a probabilistic framework for viewing
how risk increases and decreases over time (Zucker et al., 1996). These researchers
utilized longitudinal data to describe the downward social mobility associated with
alcoholics who were also antisocial. Families with such challenges demonstrated
movement toward worsening conditions over time.

Thus, risk may be viewed as a fluid characteristic that increases or decreases
depending on the introductions, resolutions, and interactions among causal components
(see Zucker, Fitzgerald, & Moses, 1995), allowing for the assessment of alteration in risk
level related to restricted social environments and opportunity. Within the developmental
perspective, change and instability is as likely as stability, depending on the causal

interplay among systenric interrelationships (Zucker, Fitzgerald, & Moses, 1995).

29

Applying this framework to the assessment of risk for parental physical aggression, one
can view across time changes in the family system that may lead to the emergence of
maltreatment.

Defining Child Maltreatment

Despite impressive progress in the field of child maltreatment research in the past
three decades, there remains continuing controversy regarding the seemingly most basic
issue of maltreatment: definition. The implications of defining maltreatment are
profound and relate to issues of policy, intervention and services, and legal ramifications
for perpetrators and victims. At the current time, the states have been left with the duty of
defining types of child maltreatment for their inhabitants, leaving the United States with a
patchwork of policies and services. This difficulty of defining maltreatment is reflected
in and reinforced by research on the issue.

For the purposes of the current study, child maltreatment is evaluated as parental
physical aggression. Child abuse per se is not defined within this study. Sexual abuse,
which appears to be a phenomenon quite separate from physical abuse (Hillson & Kuiper,
1994), is not addressed here, nor is child neglect and emotional abuse. Parental physical
aggression is operationalized as the parent’s use of physical force against a child as a
means of resolving conflict.
thelnaiea

Despite the large amount of research on risk factors for child maltreatment, very
little research has been conducted on the possibility that risk indicators change over time

to predict the emergence of maltreatment. Understanding how risk factors change over

30

time, and how these changes impact our ability to predict child maltreatment, may reveal
important nuances about the family contexts most likely to explain physically aggressive
parenting. Charting changes in the family context over time may also help explain why it
is so difficult to predict maltreatment based only on number of concurrent risk factors.
Any one parenting outcome may be the result of very different patterns of resources and
stressors (V ondra, 1990), and at this point, the processes by which these factors interact to
lead to abuse remain largely speculative (Ammerman, 1990).

The approach used in this study answers the callings of students of developmental
psychopathology that researchers “analyze the risk and protective factors and mechanisms
operating in the individual and his or her environment across the life span,” and seek to
understand the individual differences in responding to stressors at different times and in
difi‘erent contexts (p.542; Cicchetti & Toth, 1995). In addition, many previous studies
have overemphasized factors associated with abuse for mothers by virtue of using entirely
female samples (Perry et al., 1983). Examining fathers’ role in the etiology of parental
physical aggression and searching for possible differential determinants and outcomes of
maternal versus paternal risk factors is imperative. In light of known difl’erences in the
parenting experiences of fathers and mothers, it seems likely that the dynamics among
risk indicators and outcomes will differ for the two parenting roles (Mash & Johnston,
1990).

We
The purpose of the current study was to examine the role of several risk indicators

for parental physical aggression. These risk indicators, parental depression and alcohol

31

consumption, perception of child externalizing behavior, conflict in the family, and
socioeconomic status, were assessed for levels of intensity both concurrent with the
measure of physical aggression and in the (approximately) nine-years prior to the measure
of parental physical aggression. One goal was to ascertain differences in risk levels
between high- and low-aggression parents. Another goal was to predict parental physical
aggression based on current and past risk levels. This study aimed to track the continuity
and discontinuity of risk factors in order to deterrrrine whether parental physical
aggression occurs in a stable vulnerability context or is deterrrrined by family contexts
characterized by disruption and change.

flatmates

Hypothesis 1: Parents who report high levels of parental physical aggression will
experience higher levels of risk factors relative to parents with low levels of parental
physical aggression.

Hypothesis 2: Risk factors levels experienced by the high-aggression parents will
increase in intensity over time (i.e., increasing levels of parental alcohol consumption,
parental depression, perception of child difficulty, and conflict in the family environment;
decreasing levels of SES) relative to low-aggression parents.

Hypothesis 3: Greater degrees of change in risk indicators over time will predict
more parental aggression.

Hypothesis 4: Higher concurrent level of risk factor will predict higher level of
parental physical aggression.

Hypothesis 5: Both greater degree of change and concurrent higher level of risk

32

factor will jointly predict more parental physical aggression.

Follow-up questions: Families with higher levels of parental physical aggression
will be characterized by lower parental IQ and parental education, and higher parental
antisociality than families with low levels of parental physical aggression. Examination
of these additional risk indicators is based on findings that have tied these enduring parent
factors to both parenting and physical aggression (see Cantrell et al., 1990; Davies et al.,
1989; Davies et al., 1991; Perry et al., 1983).

Methods
Emigrants

Participants for the proposed study are families involved in the Michigan State
University/University of Michigan Longitudinal Study (Zucker & Fitzgerald, 1991a).
This ongoing longitudinal project utilizes population-based recruitment strategies to
assess alcoholic men and their families and a contrast group of families with non-
substance abusing parents. The sample included 218 parents, 115 mothers and 104
fathers.

All target children were males between the ages of nine and twelve. Boys were
selected as the target group for the larger study because in the general population
alcoholism, the domain of primary interest, is more common with males (Zucker et al.,
1996). All families are Caucasian. Census data indicated that in the area where data
collection for the main study took place, non-Caucasian families comprised less than 10%
of the entire sample. Given the extensive literature demonstrating a substantial

relationship between patterns of alcohol involvement and ethnic/racial status, and that

33

effective analysis of such differences were not possible due to sample size, the
ethnic/racial variation was excluded in the larger study, and subsequently, in the proposed
study.
Mm

Based on the focus of the larger project, the target population consisted of
alcoholic men with intact families. Maternal alcoholism was not a criterion for either
inclusion or exclusion from the sample. Alcohol-abusing males, recruited from the
population of all convicted male drunk drivers in the mid-Michigan area, were identified
by local district courts (probation officers), and were asked for permission regarding
potential involvement in a “study on child development and family health” (Muller et al.,
1994). At the time of recruitment, fathers were required to be living with a biological son
between the ages of three and six years old, and the child’s mother. All such candidates
were asked their pernrission to have their names released to project personnel. These
candidates were visited in their homes by project staff, who further screened the family
for suitability and if appropriate, recruited them into the project. To date, 79% agreed to
have their names released; of those potential candidates contacted by project staff, 92%
agreed to participate. After reviewing the confidentiality of the study, the project staff
designed an assessment schedule that included both parents and the target child.

Additional subjects were recruited via door-to-door neighborhood canvassing.
These families, recruited from the same census tract as the court-recruited alcoholic
families, were identified as either non-alcoholic controls or community alcoholics.

Families were matched on a number of variables, including child age (i .5 years), sex,

34

and sibling structure.
Data Collection

Data for the proposed study were collected via questionnaires given and
interviews conducted in the participants’ homes. Data were collected by trained project
staff who were blind to family risk status. All families in the study were paid for their
participation. Data used for the proposed study will be taken from the first, second, and
third waves of measurement. Each wave was collected approximately three years apart.
MM

Emmi}. Parental depression was assessed with the Beck Depression
Inventory (BDI) and the Hamilton Rating Scale for Depression (HRSD). The BDI was
used to produce an indication of current depressive symptoms, while the HRSD was used
to assess the worst-ever depression episode in between measurement periods (i.e., a span
of approximately three years).'

The Beck Depression Inventory (BDI) (Beck, Ward, Mendelson, Mock, &
Erbaugh, 1961) is a widely used phenomenological measure of depressive state that has
received extensive validational study (Carroll et al., 1973). The short form of the BDI is
a 13-item self-report instrument that assesses cognitive, emotional, motivational, and
physical manifestations of depression. It is a better measure of psychological depression

than of the biological aspects, and is more sensitive to moderate levels of depression than

 

1

Although the Harrrilton Rating Scale for Depression also assesses current depressive
status, the use of the BDI for current and the HRSD for worst-ever has been used before
(Ellis et al., 1996, under review; Bingham et al., 1996).

35

more profound states. It is used here, along with the Hamilton Rating Scale as one of
several indicators of this area of functioning. Beck reports a split-half reliability of .93.
Scores on the long and short forms of the BDI correlate between .89 and .97 (Beck, Sterr,
& Garbin, 1988). A meta-analysis of 25 years of data on the BDI yielded an internal
consistency mean coefficient alpha of .86 for psychiatric patients and .81 for non-
psychiatric patients. Current parental depression was assessed by 13 statements
concerning different areas of functioning known to be affected by depression.

The Hamilton Rating Scale for Depression (HRSD) (Hamilton, 1960, 1967), an
instrument for the clinical rating of depression, is coded via interview. This rating covers
a variety of behavioral, affective, somatic, and psychological dimensions associated with
depression, and the score is based in the subject’s responses as well as the clinician’s
judgments. The clinician makes both a current depression rating and a rating of the level
of the subject’s depression at the point in their life when they were most depressed. For
the study proposed here, only the scale that assesses worst-ever depression will be used.
Interrater reliabilities have ranged from .89 to .90 (Hanrilton, 1969). HRSD interrater
reliabilities obtained from previous studies of the population examined in the proposed
study, based on a sample of 16 individuals, were .78 for current depression and .80 for
worst-ever depression.

W. Information on the parents’ level of alcohol
consumption in the last six months is used to compute a score for current drinking. The
Quantity-Frequency- Variability Index-Revised (QFV-R) (Zucker & Davies, 1989) was

used. This instrument is a revision of the Quantity-Frequency Variability (QFV) Index

36

created by Cahalan, Cisin, and Crossley (1969). Scores about drinking are obtained by
multiplying the QV class times the approximate number of drinking episodes per year
(based on the reported average fiequency). This yields a zero to 21,000 score which is
then subjected to a logarithmic transformation (base 10). This revision of the scoring
system increases the sensitivity of the measure over the original QFV, which is necessary
given that almost all of the men in the current sample would be classified as heavy
drinkers or abstainers.

Perceived child difficulty. The Child Behavior Checklist was used to assess the
parents’ perception of the child’s externalizing behaviors. The Child Behavior Checklist
(CBCL) (Achenbach & Edelbrock, 1983) is completed by each parent independently.
This provides an objective assessment of behavior problems via the child’s social and
emotional functioning. The instrument has been normed on children 4 to 16 years of age
and yields standardized scores on eight narrow band subscales and two broad band
subscales concerning externalizing and internalizing psychopathology and social
competence. The scale also yields an overall index of child behavior problems known as
the Total Behavior Problem Score. For this study, the externalizing subscale will be
used, which includes iterrrs regarding the behavior problems.

The CBCL has also been administered to parents of children less than four years
of age. Although the subscale scores of these children must be interpreted with caution,
Fitzgerald, Sullivan, Ham, Zucker, Bruckel, Schneider, and N011 (1993) found no
substantive differences in three-year-old children’s scores and those obtained by four- and

five-year-olds. When applied to the same population assessed in the proposed study, in

37

each analysis from the Time 1 data set, approximately 30% of the high risk children
scored in the clinical range on one or more of the CBCL scales.

Family Qnflict. The degree of conflict present in the home environment was
assessed using the conflict subscale of the Family Environment Scale. The FES (Moos &
Moos, 1976, 1981) is a 90-item true-false measure of various aspects of the family
psychosocial environment that yields 10 subscale scores. As conceptualized by Moos,
these subscales fall into one of three overarching mesosystem domains of the family’s
psychosocial environment: interpersonal relationships, personal growth, and system
maintenance. The FES is an empirically based taxonomy of family social environments
as perceived by the family members themselves. It requires fifth- or sixth-grade level
reading skills. With slight modifications, it may be rated by a home observer or
interviewer. Form R of the FES consists of a number of scales that describe dimensions
if the family climate with which each individual member must cope. The scores on the
subscales yield a profile with the family as a central focus; or they may be used to
compare the extent of agreement between family members; or they may be used to
compare and contrast family groups. The conflict subscale of the F ES will be used in the
current study. The instrument has been subjected to extensive reliability and validity
studies.

WM. Information regarding family socioeconomic status and
household income was assessed as part of a demographic questionnaire. Family SES was
measured using the Revised Duncan Socioeconomic Index (SE1) (Duncan, 1977) of

occupational prestige (T SE12). Items ask about recent occupation based on measure of

38

socioeconomic status, job achievement level, and job stability. Mothers’ and fathers’
prestige scores were averaged to provide an indicator of family SES, with higher TSE12
scores reflecting higher SES.

Pmntal Physical Agggession. Parental physical aggression was assessed via the
parent-to-child subscale of the Conflict Tactics Scale (CTS) (Straus, Gelles, & Steinmetz,
1980). The CTS measures three factorially separate tactics used in interpersonal conflict:
(a) Reasoning—the use of rational discussion and agreement, (b) Verbal Aggression—the
use of verbal and nonverbal expressions of anger, and ( c) Violence-the use of physical
force or violence. Data from the Violence Scale (items 11 through 15) were used for the
current study.

Parents were asked independently to indicate the number of times they used each
tactic during the past year on the target child. The items become gradually more violent
toward the end of the list. The CTS can be interviewer-administered or self-administered;
for the proposed study a self-administered questionnaire format was used. The 15 items
of the CTS were revised in the MSU-UM Longitudinal Study with minor changes in
some of the wording. The respondent is asked to rate how often he or she used each of
the listed tactics during the past 12 months (a = never, B = once, C = 2 - 3 times, D = 4 -
6 times, E = 7 -11 times, F = monthly, G = about twice a month, H = weekly, I = about
twice a week, J = more than twice a week but less than daily, and K = daily).

Straus (1979) reported high reliability with item-total correlations ranging from
.70 to .88 using the first version, Form A (14 items). Cronbach alphas for the Reasoning,

Verbal Aggression, and Violence scales of the Form N (19 items) ranged from .77 to .88,

39

.62 to .88, and .50 to .76, respectively. Correlations between the CTS and other measures
relevant to family violence are also shown to be high. Past studies have collapsed all
items from the verbal aggression, minor violence, and severe violence subscales into a
global dimension of maltreatment, for which parallel-forms reliability was estimated at
.77 (Muller et al., 1994).

Based on the most recent theories that view maltreatment as a continuum, parents’
physical aggression toward children will be assessed according to ‘degree’. To determine
degree of parental physical aggression, responses to the Violence Subscale of the Conflict
Tactics Scale (items 11 through 15) were be weighted according to fiequencies indicated
by response categories and summed together to produce a weighted summed score.
Straus’ (1979) form N had response category ranges flour 0 to 6; for scoring purposes, he
suggests substituting for the 0 to 6 scale, 0, 1, 2, 4, 8, 15, and 25.

The items of the CTS in the MSU-UM Longitudinal Study were revised to include
a greater range of responses for number of times the action occurred during the past 12
months. The respondent was asked to rate how often he or she used each of the listed
tactics, ranging from never to daily. Extending Straus’ suggested weighting method to
the Longitudinal Study version of the CTS, scores were weighted according to frequency,
along the following dimensions: (0) never, (1) once, (3) 2 - 3 times, (5) 4 - 6 times, (9) 7 -
11 times, (12) monthly, (24) about twice a month, (52) weekly, (104) about twice a week,
(156) more than twice a week but less than daily, and (365) daily. Each parent received a
summed weighted frequency score, indicating the overall level of physical violence

reported during past 12 months.

40

Missing Data Estimation

The original datasetconsisted of 224 parents, derived of all parents with complete
Time 3 data for the Conflict Tactics Scale. Missing data was checked across the
remaining six instruments at all three waves of data collection (see Appendix M for
descriptive statistics). Ifan individual was missing more than 50% of data across all
three waves, this case was eliminated. Eliminating these five cases resulted in a data set
than included 219 parents.

There were 3942 data points of repeated measures (i.e., six instruments,
measuring current depression, worst-ever depression, alcohol consumption, perception of
child difficulty, conflict in the family environment, and socioeconomic status X three
waves of data collection X number of subjects) and 657 data points for stable background
characteristics (i.e., three instruments for antisociality, education, and IQ X number of
subjects). The sample’s missing values are presented in Appendix N.

An additional case was deleted from the sample due to the large percentage of
missing data, leaving a sample size of 21 8. Bias analyses, conducted on the data
collected fi'om the remaining 218 individuals, indicated that participants missing one or
more data points were not significantly different from those with complete data.
However, the analyses revealed differences between mothers and fathers, and for parents
based on risk status for the larger MSU-UM Longitudinal Study (i.e., high risk, moderate

risk, and low risk, based on scores for lifetime alcohol problems and antisocial

41

behavior)2, suggesting that data elimination should be conducted separately for these six
groups. The sample sizes for these six groups are as follows: low risk females (N=80),
low risk males (N=34), moderate risk females (N=19), moderate risk males (N =44), high
risk females (N=10), and high risk males (N=3 1). As unequal sample sizes threaten
external validity and the power of a test to detect real differences (Cohen, 1988), the
missing data were estimated prior to substantive hypothesis testing.

Missing data imputation was completed separately for mothers and fathers in each
of the risk groups. For the purposes of this study, only missing scale (instrument) scores,
rather than individual item values, were imputed.

Scale scores were estimated using two different procedures. A longitudinal data
estimation procedure developed by Peterson (1987; see Bingham, 1993; Bingham,
Steinmuler, Petersen, & Graber, under review) was used to estimate the missing scale
scores for instruments that were administered at Times 1, 2, and 3 (i.e., BDI, HRSD,
QFVR, CBCL, FES, DSEI). For those instruments whose scores were used only at one
data point (i.e., Antisociality, Education, and IQ), and for individuals missing scores for

two of three waves of the same instrument, cross-sectional within-group mean

substitution was used.
W. Approximately 230 data points needed to be

estimated (5% of the entire data set). The longitudinal data estimation procedure

 

Risk groups were also formed based on diagnoses of lifetime alcohol problems
and antisocial personality disorder (ASPD) (see Appendix 0); this alteration did not
change the results from initial risk group results.

42

(Petersen, 1987) was used to estimate 205 of these data points (89.1% of all missing
values and 4.5% of the total data set). This estimation procedure was used only in cases
where subjects were missing data at only one of the three waves.

This estimation procedure utilized two components; the nomothetic, or across-
subjects, component and the ideographic, or within-subj ects, component. The nomothetic
component consisted of the scale score means of each group (i.e., low-, moderate-, and
high-risk mothers and fathers) at each wave of data collection. The ideographic
component is the average distance, in units of standard deviation, between the subjects’
data points at the waves where data are not missing and the nomothetic component at the
wave with missing data (Bingham et al., 1996).

First, the nomothetic component of each variable (calculated using non-missing
data)was computed for each of the six subgroups at each wave. Next, an SPSS program
was written to compute the missing data. A deviation score was computed for each case
at each of the three waves by subtracting each individual’s score fi'om the subgroup mean

for each of the three waves. For example, the deviations from the mean for Variable X

were calculated as follows:
DXl = MXl - X1.
DX2 = MX2 - X2.
DX3 = MX3 - X3,

in which DXl represents the deviation score for the Wave 1 variable, whereas MXl
represents the subgroup mean for Variable X at Wave 1, and X1 represents the Wave 1
score for Variable X. An average was computed using DXl , DX2, and DX3. An average

score was computed across all three deviation scores (i.e., DXl + DX2 + DX3/3).

43

If one of the data points was missing, then two deviation scores were computed
(i.e., either DXl & DX2, DXl & DX3, or DX2 & DX3), and the average of these two
deviation scores was used to compute the deviation (across Waves 1, 2, and 3) from the
mean of Variable X. For example, DX = DXl or DX = DX2 or DX = DX3. This
deviation score was then used to compute the estimated data as follows: EDI = MXl -
DX. The missing data point was estimated by subtracting the subgroup mean score of X
from the deviation score. Ifthe scale score was not missing, the original score was
retained for that instrument (Bingham, 1993; Bingham et al., 1996).

Cross-sectional Dsta Estimation. Following the longitudinal data estimation, the
entire sample of data, including estimated and original data, was examined for scale
scores that remained missing. Twenty-five data points out of a possible 4599 data points
(.5%) were still missing. These data points were missing because the instrument was
only collected at Wave 1, and therefore could not be estimated with the longitudinal
strategy.

These missing data points were estimated by using selected items from the sample
which were highly correlated with the non-missing variable. These selected variables
were then regressed onto the variable to be estimated with ordinary least squares
regression. For example, education and antisociality were correlated with IQ scores.
Means were computed for education and antisociality for each of the risk (low-,
moderate-, and high-risk) X parent (mother, father) subgroups and substituted for the

missing values so that missing value estimates could be obtained.

Bias in Estimation. Upon completion of data estimation, analyses were conducted
in order to ensure that data estimation had not biased the measures. Bivariate
correlational analyses were conducted between all observed measures and the outcome
(physical aggression) and between all estimated measures and the outcome. The
correlations for corresponding observed and estimated values were compared.
Discrepancies of no more than .05 were accepted as evidence that the rank-order
associations among the variables had not been unduly disrupted by the missing data
estimation.

Am. Following missing data estimation, each participant’s responses to
items in the Violence subscale of the Conflict Tactics Scale were recoded based on
frequency into a Physical Aggression score. The sample’s Physical Aggression scores
ranged from 0 to 208 (x = 5.02, s = 17.87, md = 1.00). For the first analyses the sample
was divided into high- and low-aggression groups based on a median split of the physical
aggression score. Physical Aggression scores from 0 to 1 were recoded as Low
Aggression; Physical Aggression scores fiom 1.1 to 208 were recoded as High
Aggression.

Hypothesis 1 was that the degree of risk indicator of high-aggression parents
would be greater than low-aggression parents, and hypothesis 2 asserted that risk factors
levels experienced by the high-aggression parents would increase in intensity over time
(i.e., increasing levels of parental alcohol consumption, parental depression, perception of
child difficulty, and conflict in the family environment; decreasing levels of SES) relative

to low-aggression parents.

45

The association between parental physical aggression and degree of risk factor and
the association between parental physical aggression and movement in risk factors across
time, was tested with a 2(mothers, fathers) X 2(high aggression, low aggression) X
3(times of measurement) repeated-measures AN OVA design. This (2 x 2 x 3) repeated-
measures AN OVA was run for each risk factor, with degree of risk as the dependent
measure. Hypothesis 1, that high- versus low-aggression parents differ in degree of stress
indicators, was tested with the between-subj ects segment of the analysis [2(mothers,
fathers) X 2(high aggression, low aggression)]. Hypothesis 2, that risk indicator level for
highly-aggressive parents worsen over time relative to low-aggression parents, was tested
with the within-subjects segment of the analysis [(2(high aggression, low aggression) X
3(times of measurement)]. In addition, differences between risk indicator level for
mothers and fathers was examined. Lastly, interactions among group, time, and parent sex
were assessed.

For the remaining hypotheses, physical aggression scores were rank-transformed,
such that percentile rank scores were substituted for observed scores for each case. For
tied values, the average value for that rank was calculated and assigned to all ties. The
percentile ranked physical aggression score was calculated separately for fathers and
mothers. This method was used to lirrrit the effects of outliers in regression analyses.
Regression analyses were conducted with the percentile ranked values.

Hierarchical regression was used to assess the association between parental
physical aggression and cross-temporal change in risk factors (Hypothesis 3). For these

analyses, the change scores were calculated as follows:

46

A = |(Xi2 - Xil)| + |(Xi3 - Xi2)|. The change score is the Stun of the absolute values of
differences between Times 1 and 2 and between Times 2 and 3, an indicator of overall
change per risk factor for each parent. A change score was calculated for each risk factor,
with a higher score indicating a greater degree of change.

The regression models were constructed using a proximal-to-distal approach in
which the indicators were ordered according to their relative proximity to current parental
experiences. The order of entry for the change measures was individual-level variables
first [(1) parental depression, (2) parental alcohol consumption, and (3) perception of
child difficulty], familial-level variables second [(4) conflict in the family environment],
and community-level variables third [(5) represented by SES]. For indicators assessed
with multiple measures, the change scores representing multiple measures of the same
risk factor were entered simultaneously. To assess their place within the model, the
background characteristics were added both first and last, given their high cross-temporal
stability and their temporal separation from current parental experiences.

The order of proximity was internal states (i.e., depression), behaviors (i.e.,
alcohol consumption), ratings or perceptions of others (i.e., perception of child difficulty),
family-level variables (i.e., conflict within the home), and contextual elements of the
developmental framework (i.e., socioeconomic status). This method, which allows for
the examination of differential effects of variables within and between distal-proximal
levels, has been previously used with data from the larger study (Bingham, Fitzgerald,
W., Fitzgerald, H., & Zucker, 1996; Bingham, Fitzgerald, Townsend, & Zucker, 1995;

Jansen, Fitzgerald, Ham, & Zucker, 1995), as well as other studies (Bingham, Miller, &

47

Adams, 1990; Miller & Bingham, 1989).

From these analyses, a model predicting parental physical aggression by degree of
change in risk factor was generated for mothers and fathers. The model was trimmed of
nonsignificant predictors. Hierarchical regression was also used to test patterns of
association between parental physical aggression and concurrent degree of risk factor
(Hypothesis 4 and follow-up questions). For these analyses, parental physical aggression
was regressed onto degree of risk indicator fi'om Time 3. In addition, background
characteristics of parental antisociality, IQ, and education were added to the model.
Order of entry of each concurrent degree of indicator into the formula was the same as the
previous models, but with concurrent predictors rather than change scores.

To compare the relative importance of degree of change and concurrent degree of
risk factor in predicting level of parental physical aggression (Hypothesis 5), hierarchical
regression was used to generate a model for which significant change indicators (fiom
results of analyses to Hypothesis 3) and significant concurrent level indicators (from
results of analyses to Hypothesis 4) were combined. Essentially, the trimmed models
produced by the previous analyses were combined in a hierarchical order so that both
concurrent degree and change in risk factors were used to predict parental physical
aggression. Based on this analysis, it was ascertained whether: (a) a history of change in
risk indicators is predictive of parental physical aggression, (b) concurrent degree of risk
factor is predictive of parental physical aggression, or ( c) both degree of change and
concurrent degree of risk factor difl’erentiate between types of parent, and whether degree

of change or concurrent level is a stronger predictor.

48

Results

The first set of analyses addressed the question of differences in levels of parental
risk indicators. Means, standard deviations, and range of values for all variables were
calculated for the entire sample and then separately for mothers and fathers. These
statistics are presented in Appendix A (Entire sample, N = 218), Appendix B (Mothers, N
= 115), and Appendix C (Fathers, N = 103). Child externalizing behaviors and conflict in
the home were higher for mothers than fathers on average. Fathers’ mean levels of
alcohol consumption, antisociality, education, IQ, and socioeconomic status were higher
than those of mothers.

Bivariate correlations were conducted to ascertain interrelationships among
predictor variables. These correlations were conducted separately for fathers (Table 1)
and mothers (Table 2). Then relationships between fathers’ and mothers’ risk factors
were assessed (Table 3).

Correlations among fathers’ variables revealed that higher current depression was
related to higher depression in the past, and that reports of high child externalizing
behaviors were associated with greater conflict in the home environment. Several Time 1
variables were positively associated with many of the risk factors: in particular, paternal
antisociality, depression, and ratings of child externalizing behaviors at Time 1 were
positively correlated with virtually all other paternal risk factors. Lower paternal IQ and
education corresponded to lower SES.

Mothers’ variables were more highly related than fathers’, suggesting that

maternal risk occurred in a context of highly integrated indicators (for fathers, the matrix

49

 

O

O.

 

 

 

 

 

 

 

 

 

8. _ 8. _N. cu. m _ . vm. .N 2. MN. am. 5
o. o
8._ _ _. Nm. 3. o _ . S. no. 8. no. n0
3.. e
84 on. .5... 8 no. 2. a. 8. NO
e
8; 8. on. 0.. 2. 3. B. .0
.9. 9 IO. .0!
c9— nn. MN. NV. 3. 3. n:
o: o: .5 :5
8. _ Nn. av. 5N. NV. NI
0: o. :5
8.. _n. ”N. am. :—
09. CC.
.5.— no. 8. mm
.9.
84 an. «m
.m

 

OI

m

u 033—.

 

50

 

 

 

 

 

 

 

 

 

 

 

 

.8.vn_ooo .c.vn: mQV...
o: oo o oo o

8.. n9. 3.- cu. mN. 5N. N... 2. .N. m..- 8.- o..- No. ...- a..- .c. X..- .c. v..- 8.- N... O.
a: o: e: a: a oo o
8.. on. - on. on. cm. 3. m... h... 3.. n..- n..- .6. «.9.- c..- 2.. vN. - hN. - e. .- .e.- N..- 2N...

oo o a: o oo o o: co a. o: e
8.. nN. . vN. - 9n. - 0.. mm. ha. ... NN. 0n. 8. 0.. .n. 69. ”N. 5n. 2. ac. NN. am<

:9 a: o o. o. o o
8.. m». ah. ha . 8.. S . no . o... 9N. . n... v... N..- c..- nN. - 0N. - NN. - n ..- QN. - mm

coo 99 o

8.. ms. 3.. 8.- S.- 8. 8.- a..- .o.- 8.. N..- N..- 5N. - vN. - a..- N..- c..- Nm

.9 e o
8 . .c No.- s - o. - o..- 6N. . N... o... 2... No.- v..- cN. - h..- m . .- ON; .m

o: e: o: o: e: o
8 . 3.. .h. NV. he. NM. 3.- n..- 3. no. N.. h... s... cc. cN. MD

a: o e: o

8.. no. N. .v. .N. 3.- 8.. X..- h.... 3. 6.. n.. m .. 3. NU

so a: o: o. o. co
8.. cu. hm. an. R..- . ..- ac. 9. . h . . nN. 0N. . .. mm. .0

coo co. co

8.. «0. av. N. . Z. N.. am. 3. 8. n . . no: 2. Mm

o: e o
8.. 6c. 3. 0.. N.. N.. 2. 2. NN. 0.. mN. Nm

1 . O O
a w e m m e m m m m m m m M w m m m . mm

9.50 .. 03“....

51

 

 

 

 

 

 

 

 

 

 

O O GOO GOD I O 590
8. .N a. n. 8 8. «N. 2. 8. 2. .m
GO Oi O! 9.. co.
8.. 8. 8. 8. 2. em 8. 9.. 8. no
.99 .
8.. n... 8.- 2. z. z. 8. 8. no
8.. 8. 2. 8. S. 8.- 8.- 5
GOO CO. 0.. GO. GO.
8.. 8. 8. e... 8. 8. n:
009 .0 GOO CO!
8.. 9. a. .m. 8. N:
O. GO GO.
8.. 8. an. e... E
900 .9.
8.. 8. 8. B
9.0
8.. 8. 8
.m

 

N o3:-

 

52

 

 

 

 

 

 

 

 

 

 

 

 

8.2.... 8.1.: 27...
099 5 GOO 9.. Oil 9
8.. 8. e..- Ne. .... N... z. 8. 2. e..- 8.- 8.. .e. 8.- 8.- 8.- 2.- ....- 2. .N... 8.- o.
GOO GOO GOO CO. .0 O
8.. 8.- 8. 8. .m. 8... 8... 2... 2.- N..- 8.- 8. 8.- 8.- 2.- .- 8.. 8.- 8.- 2.. 2.8.
9.. O. O. O! O 9 996 0 GO. .0! GOO 0.0 O. O.
8.. 8.- NN.- 8.- 2. 2. “N. ..N. ..N. 8. 8.. 2. 8. N... .... 8. 8. ... .N. as...
GOO .9. O r O. O 909 GO
8.. 8. 2.. 2.- N..- 2.- ..N.- 2.- NN; 8.- 2.- 2.- N.. N.... NN.. 2.. e... e..- 8
GO. l. O
8.. 8. 8.- 8. N...- N..- o..- N..- 2.. e..- 2.- 2.- .N.- NN. . 8.- N... 2..- 8
O O O. O
8.. 2.- 3.- .- 8.- 2.- ..N.- 8. 2.- 2.- .7 8.. .N.. 8.- ....- 2..- .m
Oi. Oi. GOO 00. O O. O 0. 9 GOO IO. .0
8.. 8. .e. 8. 8. N... z. 8. .N. N. .N. 2. 8. 8. 8. 8
GOO GOO .00 O. O. O 0 {GO 6 O. I. GO!
8.. 8. 8. 8. 8. 2. 8. .N. .N. 8. 8. N. «N. 8. No
O O O. 0 GOO GOD I. GO. GO. GO.
8.. .N. «N. 8. 2. .N. 2. 8. 8. N. 8. 8. .e. .o
GOO 099 5. OOO CO 0 GO GO.
8.. N... 2.. 8. 8. 2. 8. 8. 8. 8. NN. 2. 2.
Ci. .0 0 CI. 00. O 0.0
8.. 8. 2. 8. 2. 8. 8. 2. NN. 8. 2. Nm
.mwmmammmmmm .N.....

.28 .N 2%...

GIL-IO

53

does not create this impression because the correlations were lower). In particular,
increased maternal depression, pronounced child externalizing behaviors, and high levels
of family conflict were associated with “worse” levels of essentially all other maternal
risk factors. Higher maternal antisociality also emerged as a key variable, related to
higher levels of depression, increased alcohol consumption, more child behavior
problems, and lower socioeconomic status. Correlations for both mothers’ and fathers’
alcohol consumption at Times 1, 2, and 3 demonstrated a lack of stability, while
correlations among conflict scores were stable for both mothers and fathers, regardless of
possible changes in the family structure over time. Mothers’ stable background
characteristics (i.e., antisociality, education, and IQ) were highly interrelated, in that
higher antisociality was associated with lower IQ and lower educational status.

Parental physical aggression was also examined regarding its correlations with
risk indicators and stable background characteristics. Mothers’ physical aggression was
significantly correlated at all three data points with both child behavior problems ( r = .21 ,
p < .05; r = .25, p < .01; r = .33, p < .001) and family conflict (r = .29, p < .001; r = .38, p
< .001; r = .22, p < .05). For fathers, level of physical aggression was significantly related
to child externalizing behaviors, but at Time 3 only (r = .23, p< .05), whereas family
conflict was significantly related at all three time points ( r = .22, p < .05; r = .27, p < .01;
and r = .34, p < .001, respectively). For both mothers and fathers, physical aggression
and child behavior problems were more strongly correlated at Time 3 than during earlier
time periods. However, family conflict operated difiemnfly for mothers and fathers: the

relationship between paternal physical aggression and conflict in the home grew stronger

54

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2.. ...- .- 8.. 2..- N... N.. .... 2..- 2..- N.. N.. .... 8.- 8.- 8.- 8. 2..- 2..- 8.- 2..- w
90 O O. W
N... 2.. ..N. ...- ....- 2.- 2.. 8. 2.. 8. 2.. 2. 2.. ..N. 8. ....- 8.- 2.. 2. 2..- 8. m
m
o o o m.
8.- 2..- 2. N..- 2.- ....- 2. 8. .N. 2.. ... NN. 2.. 8. ..N. 8. 2..- 8. 2. 2..- N..
O O O 0 0.. i
8. .....- 2.. 8.. N... 8.- 8 2. 8. NN. 2. N.. 2.. 2... 2..- .... 2. .... 2. 2.. ..N. m
9 GO. O. Q I. O GO GO 0.9 O. GO .00 m
8.- 2.- ..N. 8.- NN.- 8.- 8. .N. .N. .... 2. 2. 2.. 2..- 8. .8. 2.. ..N. NN. .... 8.
o a o o W
8. N..- ... 8.- 8.- 2.- 2. N.. 2. .....- ....- 8. 8.. 2..- 2..- 8. 2.. 8. 2. 2. 2. u
. o. ace 2-
8.- ....- ... 2.- ...- 2..- 2. ... ... .... 2.. 8.- 8.. 2... 2..- 8. N... 8. 8. 8. 8.
e a so a. 2. D
... N..- 2. 2.- .... N..- 2. .... ... 2. N. .. 2.- .N. 2.. 2. 2.. 2.. 8. 8. 8. m
o 2. ..i a. o: m
2.. 2.- 2. ....- 2..- ...- 2. ..N. 2. 8. N.. .... .....- 8.- ...- 8. N... 2.. 8. 8. e... m
n N . n N . n N . n N . n N . n N .
o. 82. e2 28 .85.... was... 2.8.2.: ......o 88.550 8.8.82 .25 8..-.882
62.80 .85.. .e was... .85.... .28.... .85... .53 .85.. .850 .85..
, > . e . 8

 

n 03....

 

 

 

 

55

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o o: .9 so 8. a.
8. 8. 8.- .... 8. .... 2.. 2.. 2..- 2..- 2..- 2.- 8.. 2.. 2.- 2..- 2.- .3. 2.. .....- .....- a
... a: e: o: a: e o E.
N. .n. 8.. mm. .m. 3.. 8.. 8.- o..- S.- 8: m... .N S.- 8: 2.- 8.- v... mm- o... X..- m
a: .9 on a. e o o .9
N..- «.2- uN. N.- «N... N: NN. S. .... 2. .N. .N. 8.- 8.- S. .... 2. 8. ... 8.. .... w
e. e: e e: a: a: o .5 e e
.N. an. ".N.- 8.. 3. N2 8.- no... ...- 8.- .... .N.- 2.. v..- n... .... «.N.- ...N: «N; n..- .N.-
.... o: e a: e: e: ea 0
SUN. .8. NN.- v». .... n... 8.- 8.- 8.- no. 8.- »..- N9- o..- v..- ...- oNr NN.- .N. .- ...- ....-
.. e: .3 o: a: a: a. o S
..N. mm. .2- N... ...... 8.. 8.. Not 8.- .... ....- ..N: N... .N.- 8: z..- N..- o..- v..- v..- NN.. a.
e e: e: o: o
8. 8.- .N. .N.. m ..- a..- 3.. an. an. 8... ... 8. 3.- 8.- .... .... o. .- X..- v.. 2.. 8..
o a: e: o: e
2.. 2.. .N. n..- o..- 8.- 89. 3.. 3. ... oN. .... 2..- v... 2.. no. 8.- N...- N.. ... ....
3
c a: a: o: .3 e W
No.- .... N. n ..- 2..- 8: an. an. 3. So. 3.... n2. 8.- .nNr no. .... 8. c..- .N. .... .N. m.
a: o a: a: o o
2..- n..- 8. o... 8.. ....- .n. ..N. .N. am. an. 8.. 8.- 2. ..N. «N. 2..- 8: .... no. N..
. oe coo co. 99 o o D
S.- c..- 2. N..- 8.- v..- N. .... 2. NM. ..n. uN. N...- N. .N. ... 3.- 2..- No. 8. 2.. m
o o. e. a a a: a: a: W
2..- ..N.- ..N. N.- a..- NN.- QN. ... 2.. 3“. on. a... .... c2. 2. 02.. 2.. not 8. 3. 2. .u

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.....8 .m 2%.:

S6

over time, while the relationship between maternal physical aggression and conflict in the
home was strongest early in the study.

Several patterns emerged among correlations for fathers’ and mothers’ risk
indicators and stable characteristics, providing interesting information about the family
context of the sample (Table 3). Depression (both current and past levels) was positively
correlated for fathers and mothers. Mothers’ higher past levels of depression were
associated with fathers’ reports of increased family conflict, especially during Times 2
and 3. In addition, mothers’ rates of past depression were related to lower socioeconomic
status. Parents’ ratings of both child difficulty and conflict in the home were highly
correlated. Stable background characteristics (i.e., antisociality, education, & IQ) were
highly related between fathers and mothers, as well as both groups of parents’
background characteristics with socioeconomic status.

Higher paternal antisociality was related to several maternal variables, including
higher levels of worst-ever depression (Time 2), increased alcohol consumption (Time 2),
higher rates of perceived child difficulty (Time 1), and elevated conflict in the home
(Times 1, 2, and 3). Mothers’ higher antisociality was also correlated with fathers’
variables, most notably his perception of more child externalizing behaviors (Times 1 and
2). In addition, higher maternal education and IQ were associated with lower current
paternal depression, although mothers’ current depression was not significantly correlated

with fathers’ education or IQ.

57

Qogelatigns among Change Variables

Since historical changes in risk factor level were an integral part of this study,
correlations between mothers’ and fathers’ change scores and their respective degree of
physical aggression were also calculated. For fathers, larger degrees of change in three
risk indicators was correlated with other variables: more change in reports of family
conflict was associated with variability in worst-ever depression; greater changes in
alcohol consumption were related to both higher antisociality and lower IQ; and more
change in socioeconomic status was related to higher paternal education. (Appendix D).
These relationships indicate that absolute amount of change, rather than direction of
change, is related among risk indicators. In no case did risk factors change in directions
opposite of the other; that is, greater change in one risk factor was not related to less
change in another.

For mothers, there were five significant correlations among her change variables
and background characteristics. Greater changes in current depression were related to
greater changes in depression levels in the past. In turn, greater changes in past
depression were associated with higher maternal antisociality. Variation in amount of
alcohol consumed was related to less change in family conflict and higher antisociality.
In addition, changes in maternal perception of child difficulty were associated with
variation in socioeconomic status, which was related to higher maternal education and IQ
(Appendix E). For neither parent was physical aggression related to changes in these risk

indicators.

58

 

Correlations among Concurrent Variables

Next, bivariate correlations were run among current (T3) levels of risk factors and
parental physical aggression (i.e., risk indicators and physical aggression measured at
Time 3). For fathers, physical aggression was related to elevated perceptions of both
child difficulties (r = .229, p < .05) and conflict in the home ( r = .340, p < .001). In turn,
higher levels of current child difficulties was associated with both higher worst-ever
depression and conflict in the home (r = .420, p < .001). Higher current socioeconomic
status was related to lower current rates of depression ( r = .222, p < .05) and antisociality
( r = -.249, p < .01), and higher education and IQ ( r = .590, p < .001; and r = .258, p <
.01, respectively) (T able 4).

For mothers, concurrent risk indicators were highly related. Like paternal
physical aggression, maternal physical aggression was associated with perceptions of
more child behavior problems ( r = .332, p < .001) and conflict in the home (r = .217, p <
.05). Higher levels of maternal current depression were related to multiple factors,
including higher levels of past depression (r = .491 , p < .001), higher alcohol
consumption (r = .274, p < .01), the perception of more child externalizing behaviors (r
= .260, p < .01), more conflict in the home
( r = .361, p < .001), and higher antisociality (r = .422, p < .001). Higher levels of worst-
ever depression were related to the same variables and to lower current socioeconomic
status (r = -.222, p < .01) as well.

Maternal perception of current high child externalizing, besides being

significantly correlated with physical aggression and current and worst-ever depression,

59

Table 4

 

 

 

 

 

 

 

 

 

 

 

 

o la 'ons tween Fa ’ Concurrent V 'a les i A si n
a? g.. Egg. g, “~53 Ea g... ‘2’ 23
05 u a o< gfifi 08 0%
u a. 0
Phys 1.00
Asst
Current -.02 1.00
Level
of Dep
Level -.02 .42 1 .00
Worst Dep ""
Current .13 .03 .05 1.00
Ale Consn
Current Perc .16 15 .30 .12 1.00
of Child "
Diff
Current .34 .17 .08 -.Ol .42 1.00
confliCt it! it.
Current .01 -.22 -.10 .05 -.02 -.06 1.00
SES *
A83 -.07 .19 .07 .094 .l 1 .16 -.25 1.00
it
Educ .05 -.16 -.11 -.014 -.04 .04 .59 -.32 1.00
it. It.
IQ -.13 -.14 .014 .02 -.13 .17 .26 -.04 .43 1.00
it ...
*p<.05 "p<.01 “*p<.001

60

 

Table 5

 

 

 

 

 

 

 

 

 

 

 

 

 

 

C rre tions be u Mothers’ Coneu at V lee and Ph ice] A n
.2 2 g 1’3 o 2 = a: a
a 3 .3 .>';’ 2 8 6 5 g 3 5°. 8 9
i a. Q m < 23
a“: < E 5
o “5
Phys 1.00
A28!
Current .14 1.00
Level
of Dep
Level .09 .49 1.00
worst it.
Dep
Alc .05 .27 .29 1.00
Consn I. it.
Child .33 .26 .40 .04 1.00
Difi‘ til I. ‘0‘
Conflict .22 .36 .27 .14 .38 1.00
t tfit it it.
SES .01 -. 14 -.22 -.06 -.20 -. 19 1.00
it t 0
A88 .12 .26 .42 .17 .20 .18 -.32 1.00
it 0“ t it.
Educ -.06 -.02 -. 16 .04 -. l 7 .07 .55 -.30 1.00
it. it.
IQ -.03 .10 -.01 .01 -.10 .14 .42 -.19 .56 1.00
tit t ...
'p<.05 "p<.01 "‘p<.001

61

 

was also related to higher current family conflict ( r = .378, p < .001), lower current SES
( r = -.197, p < .05), and higher antisociality ( r = .200, p < .05). Higher socioeconomic
status was related to lower recent levels of depression ( r = -.222, p < .05), fewer child
externalizing behaviors ( r = -.197, p < .05), and less conflict in the home (r = -.187, p <
.05) (Table 5).

These matrices suggest that both fathers’ and mothers’ current risk factors were
highly related, but that risk factors were more interrelated for mothers and that the nature
of her current risk “status” was dependent on multiple and interacting components.

elti hi betweenHi - Low- sinP ents 'kIn'

Parents were divided into two groups based on the degree of their reported use of
physical aggression (i.e., low or high levels of physical aggression). The relationships
between group status (i.e., low- or high-aggression) and risk indicators (and across time)
were examined with repeated measures multivariate analysis of variance. Significance
for all multivariate tests was determined using Wilks Lambda. Within-subjects and
between-subjects effects for the two groups of parents are presented in Table 6.

mm. Significant main effects of Time were found for Current [F(1, 214) =
6.48; p < .01] and Worst-ever Depression [F(1, 214) = 6.51; p < .01]. As a whole,
parents experienced increases in current depression over time, but decreases in worst-ever
depression. In addition, mothers and fathers across groups difl‘ered, with mothers
consistently reporting greater worst-ever depression [F(1, 214) = 12.42; p < .01].
However, levels of depression did not significantly differ between low- and high-

aggression parents.

62

Alcohol Consumption. There was a significant Parent Sex X Time interaction for
Alcohol Consumption [F(1, 214) = 7.68; p < .001]. For the entire sample, there was a
significant decrease in amount of alcohol consumed across time. In addition, mothers’
and fathers’ levels of drinking changed in different ways over time, with fathers’
consumption decreasing more than mothers’ (Figure 1). At Time 1, fathers’ consumption
was significantly greater than mothers’, although this difference dissipated across time:
differences between mothers’ and fathers’ alcohol consumption that were significant at
Time 1 of data collection were not significant at Times 2 or 3. Moreover, levels of
alcohol consumption did not differ significantly between low- and high-aggression
parents.

Child Emmalizing Behau'or. A significant multivariate interaction of Group X
Time was found for child externalizing behavior [F ( 1, 214) = 3.26; p < .05]. High-
aggression parents reported high levels of child externalizing behaviors that were
maintained over time. In contrast, low-aggression parents reported moderate levels of
child externalizing behaviors when children were ages three to five (Time 1), then
reported diminished child behavior problems over the next six to seven years (Times 2 &
3). At the time when parental physical aggression was measured (Time 3), high- and
low-aggression parents reported significant differences in their children’s acting-out
behaviors (Figure 2).

Coof_lict in the Home finm’ omen}. A significant three-way multivariate
interaction of Parent Sex X Group X Time was found for conflict in the home

environment [F(1, 214) = 4.70; p < .01]. This interaction arises from: (1) differences

63

Table 6

Regated Measures Multivariate Analysis of Variance for High- & Low-Aggression

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Parents
Within-Subjects
Multivariate F
Variable Parent Sex x Group x Time Parent Sex x Time
Group x Time Time

Current Depression 0.05 0.64 0.05 6.48"
Worst-ever .66 .54 .07 6.51“
Depression
Alcohol 1.29 1.07 7.68"" 990’"
Consumption
Child Externalizing .50 3.26“ .61 4.59"
Behaviors
Family Conflict 4.70" .44 .57 1.29
SES .57 .26 .01 93"”
Between-Subjects

Variable Sex of Parent Group (High/Low Sex of Parent X Within + Residual

Aggression) Group Error
df F df F df F df F

Current 1 3.13 l 2.92 1 2.32 214 (22.95)
Depression
Worst-ever l 12.42”“ 1 .44 1 2.10 214 (109.18)
Depression
Alcohol 1 7.23“ l .28 l .19 214 (14869614)
Consumption
Child Ext 1 3.33 1 15.10“” 1 .12 214 (88.96)
Behaviors
Conflict l 3.26 l 19.18‘" 1 .06 214 (9.83)
SES 1 .01 l .47 l .l 1 214 (37023.14)
Antisocial 1 27.72""'"'I l .00 1 1.54 214 (76.27)
Behavior
Education 1 4.33“ 1 .01 1 .1 1 214 (3.55)
1Q 1 15.66"” 1 .48 1 .16 214 (181.45)

 

 

 

 

 

 

 

 

 

64

 

 

3539.00 .0582 he EsoE< w xow Ewan. consume c2323.:

 

_. 0.59“.
$653.... 2265......

m mEP N 35. _. 95...
0
com
83
com?
ooow
oomN

m.>n_0 :0 28m

between high- and low-aggression families and mothers and fathers, (2) high aggression
fathers reporting lower aggression at Times 1 and 2 than high-aggression mothers, but
equal levels at Time 3, and (3) low-aggression mothers and fathers differ at Time 3.

High-aggression mothers consistently reported the highest conflict scores, which
remained high across time. High-aggression fathers also reported relatively high degrees
of conflict in the home, and mean conflict scores increased over time. For low-
aggression mothers, ratings of conflict changed over time, with a decrease in conflict at
Time 2, when the child was approximately 6 to 8 years of age, and a return to higher
ratings of conflict at Time 3, when the child was approximately 9 to 11 years of age.
Low-aggression fathers rated conflict the lowest of all four groups in this interaction, with
low levels of conflict that were consistent and stable across time (Figure 3).

High- and low-aggression parents were significantly different regarding this risk
factor. Difference between high- and low-aggression groups was maintained consistently
over time. In addition, ratings of conflict for high aggression mothers and fathers
converged at Time 3, while conflict scores for low aggression mothers and fathers
diverged at Time 3.

W. A significant main effect of Time was found for
Socioeconomic Status [F(1, 214) = 9.3; p < .001]. All families in the sample experienced

improved occupational prestige over time, which did not differentiate low- and high-

aggression parents.

66

mao_>m:mm mEszEme EEO w _o>o.._ :o_mmo._mm< :mozson :ozoESE

 

 

 

 

N 959".
TEEmn. 56363.33... .35qu 223989591...—
m 95 N 95 - P 85¢
_ b m
_ .
I: m
.... or
la
.... N_.
r T ._

 

 

 

H.

homo :0 28m

.65.. 22mmo2mm< a x90. «2222 .u26E2o.._>2m 220: 2_ 2:200 «3025 2289.32.

m 239“.

 

 

2625“. 223953.33..- Eocuoi 223953-264
265m". 223235-291... mam—=05. co_mmo._mmm-2m_:¢

 

 

m as:

N 022:.

 

mm". 20 Boom

 

 

 

 

Sgble Background Characteristics. Regarding stable background characteristics,
fathers had significantly higher antisocial behavior scores [F(1, 214) = 27.72; p < .001],
educational status [F(1, 214) = 4.33; p < .05], and IQs [F(1, 214) = 15.66; p < .001] than
did mothers. However, these features did not differentiate between high- and low-
aggression parents.

For the next set of analyses (i.e., hierarchical regressions testing hypotheses 3 &
4), physical aggression was viewed as a continuous variable, rather than the discreet high-
and low-aggression groups. In this continuum, scores for physical aggression ranged
from 0, indicating no threat or use of physical aggression, to 208, indicating frequent
threat or use of physical aggression.

Reltio in; be cc Lil'eln " kInc'cato | - u -.- ocfP -2 2 ° at-“

The relationship between change in risk indicators and parental physical
aggression was examined within a proximal-to-distal hierarchical regression model’. The
role of parental background characteristics (i.e., antisociality, education, and IQ) in
helping to predict physical aggression was examined by entering simultaneously as the
first step in the proximal-to-distal model based on the theoretical logic that variables
inherent to the individual parent would have more predictive power than variables farther

removed fiom the individual. Thus the order of entry into the regression equation was as

 

Parental physical aggression was also regressed onto change scores independently
(Appendix G), within a hierarchical regression model of change scores without the
inclusion of stable background characteristics (Appendix H), and within a hierarchical
regression model of change scores for which stable background characteristics were
added as a last step in the equation (Appendix I). In each of these cases, parental physical
aggression was not significantly predicted.

69

follows: 1. antisociality, education, IQ; 2. change in depression (Current, Worse); 3.
change in alcohol consumption; 4. change in child externalizing behaviors; 5. change in
conflict; 6. change in SES. For both mothers and fathers, the change model did not
significantly predict physical aggression‘ (Appendix F).

Rolatiooshio between Concorlent Risk Qdigtors Q pegge of Physiofl Agggssion

The next analyses examined concurrent, rather than historical, levels of risk
indicators as predictors of parental physical aggression. In this case, risk indicators were
measured at the same time that parental physical aggression was assessed (i.e., Time 3).

Stable background characteristics were added simultaneously as the first step of
the equation. Thus, the variables were entered in the following order: 1. antisociality,
education, IQ; 2. concurrent depression (Current, Worse); 3. concurrent alcohol
consumption; 4. concurrent child externalizing behaviors; 5. concurrent conflict; 6.
concurrent SES.

This model significantly predicted parental physical aggression. Fifteen percent
of mothers’ physical aggression was accounted for, with high levels of child externalizing
behaviors emerging as the only significant predictor of higher physical aggression in this
sample (Table 7). When concurrent child externalizing behavior was considered alone

(i.e, parental physical aggression regressed onto concurrent Time 3 child externalizing

 

The inclusion of antisociality, education, and IQ added little predictive power to
the model, accounting for an additional 3% of fathers’ physical aggression (up from 6%
in the model that included change variables only), and adding no predictive power to
mothers’ physical aggression model (see Appendices H & I). In addition, whether these
stable background characteristics were considered first or last within the model did not
alter their predictive importance.

70

Table 7

Summag of Hierarchical Roggession Analysis for Physical Agggession Regressed
onto Stable Bacflound Characteristics and Concurrent Predictor Variables

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fathers Mothers
Step Variable [3 R2 AR’ [3 R’ AR1
1 IQ - .18 .03 n/a .01 .01 n/a
Antisociality - .04 .1 1
Education .11 - .03
2 IQ - .18 .03 .00 - .01 .03 .02
Antisociality - .03 .08
Education .11 - .02
Concurrent Depression:
Current - .02 .12
Worst .01 - .01
3 1Q - .18 .05 .02 - .01 .03 .00
Antisociality - .05 .08
Education .11 - .03
Concurrent Depression:
Current - .02 .12
Worst .00 - .01
Concurrent Alcohol Consumption .14 .01
4 1Q -.14 .10 .05 .01 .13" .10
Antisociality - .07 .08
Education .08 .00
Concurrent Depression:
Current - .02 .08
Worst - .07 - .14
Concurrent Alcohol Consumption .12 .04
Concurrent Child Externalizing Behavior .23‘ .35‘"
5 1Q - .24" .21“ .11 .00 .13" .01
Antisociality - .1 1 .07
Education .08 - .01
Concurrent Depression:
Current - .10 .06
Worst - .01 - .13
Concurrent Alcohol Consumption .14 .03
Concurrent Child Externalizing Behavior .05 .32”
Concurrent Conflict .40"" .09

 

71

 

 

 

(Table 7, con’t)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6 IQ - .25' .21" .00 - .03 .15‘ .02

Antisociality - .l 1 .08

Education .09 - .08

Concurrent Depression:

Current - .09 .06

Worst - .01 - .13

Concurrent Alcohol Consumption .15 .04

Concurrent Child Extemalizing Behavior .04 .32"

Concurrent Conflict .39'” .13

Concurrent SES -.01 .16

Trimmed Models
Fathers Mothers
B R’ l3 R’

Concurrent Conflict .38‘" .15'"“l Concurrent Child 33”" .11"“
1Q - .19“ Externalizing Behavior

(" p s .05 " p s .01 ‘“ p s .001)

72

 

 

behaviors), Rz remained significant but dropped from .15 to .11 (p s .001).

This model also significantly predicted paternal physical aggression. Twenty-one
percent of fathers’ physical aggression was accounted for. Two factors emerged as
significant predictors of his physical aggression: lower IQ scores and higher concurrent
conflict in the home.’

Follow-up regression analyses were conducted to ascertain the importance of
these two significant predictors in a trimmed model. When concurrent conflict and IQ
were considered together outside of the model, R2 remained significant but dropped from
.21 to .15. Concurrent conflict emerged as the stronger predictor of the two (conflict [3 =
.38; p s .001), but both remained significant (IQ B = -.19; p s .05). This finding indicates
that in circumstances where the level of family conflict was high, fathers with lower IQs
were more likely to be physically aggressive toward their sons.

A partial correlation was run on these two significant predictors to clarify their
relation to physical aggression. The correlation between current conflict and physical
aggression controlling for IQ was .27 (p s .01); the correlation between IQ and physical
aggression controlling for current level of conflict was -.14 (nonsignficant). Thus,

fathers’ perceptions of current conflict in the home was the more essential predictor of

 

Concurrent risk indicators of child externalizing and family conflict also
significantly predicted parental physical aggression when assessed separately (Appendix
I). These Time 3 risk indicators also significantly predicted parental physical aggression
without consideration of stable background characteristics, and when stable background
characteristics were entered last into the regression equation (Appendix L). It did not
matter if stable background characteristics were considered first or last within the model;
the amount of variance accounted for did not change.

73

paternal physical aggression.
meting Physical Aggression from Partner’s Risk Indicators

In an effort to more fully understand the pathways that lead to parental physical
aggression, additional regression analyses were conducted. Of particular interest here
was the prediction of a parent’s degree of physical aggression toward his or her children
based on the partner ’s risk factors (Table 8).

Fathers’ physical aggression was not predicted by any maternal risk factors but
was predicted by higher levels of maternal physical aggression (R2 = .12, p < .001).
Maternal physical aggression was predicted by fathers’ current experience of conflict in
the home (R2 = .19, p < .001) and fathers’ reports of concurrent child externalizing
behavior (R2 = .06, p < .01). In addition, mothers’ higher physical aggression was
predicted by higher physical aggression by fathers (R2 = .12, p < .001).

Regressing maternal physical aggression onto these three significant paternal
variables (Fathers’ concurrent conflict, Fathers’ concurrent child externalizing behaviors,
and Fathers’ physical aggression), accounted for 11% of the variance in maternal
aggression.

Discussion

Much research on risk for parental physical aggression is based on cross-sectional
data. Such studies fail to address whether risk associated with a family history of
disruption and change is more likely to predict physical aggression than levels of risk
factors impinging upon a parent at a given point in time. The main finding from the

current study was that parents’ current risk levels were predictive of physical aggression,

74

Table 8

ngicting Physical Aggggoion from Partner’s Risk Indicators

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

75

Predicting Fathers ’ Physical R2 Predicting Mothers ’ R2

Aggression fi-om: Physical Aggression fiom:
Mothers’ Antisociality, .01 Fathers’ Antisociality, .00
Education, & IQ Education, & IQ

Change in mothers’ Depression .02 Change in fathers’ Depression .00
Change in mothers’ Alcohol .00 Change in fathers’ Alcohol .00
Consumption Consumption

Change in mothers’ rating of .01 Change in fathers’ rating of .00
Child Extemalizing Behavior Child Extemalizing Behavior

Change in mothers’ reports of .00 Change in fathers’ reports of .01
Conflict Conflict

Change in mothers’ SES .00 Change in fathers’ SES .01
Mothers’ Concurrent .00 Fathers’ Concurrent .03
Depression Depression

Mothers’ Concurrent Alcohol .01 Fathers’ Concurrent Alcohol .00
Consumption Consumption

Mothers’ Concurrent rating of .03 Fathers’ Concurrent rating of .06”
Child Extemalizing Behavior Child Extemalizing Behavior
Mothers’ Concurrent reports of .03 Fathers’ Concurrent reports of .19*”
Conflict Conflict

Mothers’ Concurrent SES .00 Fathers’ Concurrent SES .00
Mothers’ .12 Fathers’ .12
Physical Aggression "* Physical Aggression *"

I'p s .05 "p s .01 *"p s .001

 

while amount of change in the parental history or risk was not. In particular, current
levels of perceived child difficulty and family conflict predicted parental physical
aggression. Moreover, these two risk indicators were elevated consistently across time
for high-aggression parents. The results of each hypothesis tested are discussed in the
following sections.

Risk Levels & Aggressive Parenting

Hypothesis 1 suggested low- and high-aggression parents would differ in their
levels of risk factors. That is, parents who demonstrated higher levels of aggression
toward their children would experience greater levels of depression, alcohol consumption,
child behavior problems, conflicted family environments, and lower socioeconomic
status.

Depression, alcoholism, and lower socioeconomic status did not differ for high
versus low aggression parents. However, Hypothesis 1 was supported by the finding that
physically aggressive parents were more likely to have children who were high
extemalizers and to live in home environments that were more conflicted than low-
aggression parents.

Hypothesis 2, that risk factor levels experienced by the high-aggression parents
would increase in intensity over time relative to low-aggression parents, was not
supported. The difl'erences between high- and low-aggression parents that emerged with
Hypothesis l-high-aggression parents experience greater child behavior problems and
more conflict at home—were stable across time, suggesting that risk for physical abuse is

present early on and is maintained over time. Thus, parents did not appear to move in and

76

out of risk status so much as remain at a risk level characterized by consistently elevated
levels of conflict and child behavior problems.
vers ta ii inPredic in P en A essi n

The notion of risk for physical aggression as stable versus fluid was addressed by
comparing the predictive importance of history of changes in risk indicators to concurrent
levels of risk indicators. Hypothesis 3, that a parent’s history of changes in risk indicators
predicted aggression toward children, was not supported, while Hypothesis 4, that the
levels of risk factors currently experienced by a parent predicted physical aggression, was
supported by the data.

Overall, a family’s map of developmental changes in risk factors was not
predictive of aggressive parenting, while a “snapshot” of current levels of certain risk
components was: the prediction of physical aggression was strongest when risk levels
were measured concurrently. In particular, the current amount of conflict within a family
and the perception of children’s current externalizing behaviors were significant
predictors of parental physical aggression. In addition, conflicted families with fathers
possessing fewer intellectual resources were at greater risk for physical aggression. Thus,
in the case of parental aggression, it appears that the “here and now”of the parenting
context is more influential of parents’ use of aggression than are historical risk levels.

An important trend should be noted that adds depth to the notion of risk for
parental aggression. The risk indicators that predicted parental physical aggression—high
current levels of child externalizing behavior and family conflict-were stable and

elevated across time. This element of stability is supported by earlier studies that have

77

 

also shown that enduring risk indicators place families at greater risk than change, and
that prolonged exposure to stress (rather than isolated stressful events) distinguishes
between abusive and nonabusive parents (Ammerman, 1991; Spicer & Franklin, 1994).
The cumulative impact over time of experiencing “typical” chronic parenting
hassles-being nagged or whined to, being interrupted, dealing with difficult behavior,
continuously cleaning up afier children—and the parental response to them, may
eventually have an adverse influence on the quality of parenting (Crnic & Acevedo,
1995).

Researchers have suggested that the nature of daily parenting hassles is mediated
by both the family’s and the child’s developmental stage, such that different factors exert
their influence at different periods in the life course (Crnic & Acevedo, 1995). However,
during the time frame of the current study, which covered multiple developmental
changes (including the child’s entry into school and early adolescence), parental
aggression was oot influenced by different risk factors at different periods: the two risk
indicators predictive of parental aggression, perceived child externalizing behavior and
family conflict, were consistently influential over time. Although the results of the
current study cannot be compared to past parenting practices (i.e., one cannot know if this
sample’s parental physical aggression emerged over time or existed all along), they
provide tentative evidence that cumulative daily hassles associated with parenting may,
over time, change the nature of parent-child relationships to become more problematic,
conflictual, and aggressive. Almost a decade of parental exposure to perceived child

behavior problems and a conflicted family environment appears to be the risk factor that

78

 

leads some parents to physically aggress against their sons. These two indicators are
examined in greater detail in the following sections.

Perception of Child Difficulty. This study provides strong evidence that the
perception of child behavioral problems is indeed a marker of risk for physical
aggression. Children who are perceived as “difficult” (e.g., overactive, aggressive,
noncompliant) are more likely to engage in conflict-ridden and coercive interactions with
their parents, are at higher risk of abuse, and tend to have highly stressed parents
(Campbell, 1997; Garbarino & Sherman, 1980; Milner, 1993; Vietze et al., 1980).
Parents with externalizing children report more negative impact on their social life, more
negative and less positive feelings about parenting, and higher child-related stress;
indeed, parents of high-extemalizing children report levels of stress as high as those
reported by parents of children with autism (Donenberg & Baker, 1993). These parents
also typically use aversive, yet ineffective discipline and high rates of aggression in
interactions with their children (Patterson, 1982).

Parent-child interactions are often a focal point for the manifestation of the effects
of a variety of other proximal and distal sources of parenting stress (Mash & Johnston,
1990; Shaw, Owens, Vondra, Keenan, & Winslow; 1996). Holden and Banez (1996)
found that levels of abuse potential did not differ significantly across low, medium, and
high levels of child-associated stress when parent stress was law. However, at high
levels of parent stress, abuse potential was significantly higher and increased substantially
as child-associated (i.e., externalizing behaviors) stress increased. Both child behavior

and negative parental control contribute unique and independent variance to maternal

79

 

ratings of externalizing behavior problems (Campbell, 1997).

In the current study, high-aggression parents reported higher levels of child
behavior problems than did low-aggression parents consistently across time, suggesting
that parents in this group were consistently more stressed by their children's behavior.
Low-aggression parents reported fewer troublesome behaviors over time. Campbell’s
recent (1997) summary of behavior problems in young children aptly provides an
explanation for this finding. She observes that when initial symptoms are more severe,
mothers are negative and controlling, and family stress is elevated and chronic, child
behavior problems are likely to persist. Other children with lower levels of initial
problems, living in less stressful family circumstances, will continue on a normal
developmental trajectory. Parental perceptions of their child's behavior may or may not
reflect the reality of child behavior, but are (regardless of the accuracy of these
perceptions) more important in predicting parenting practices than are observers’ ratings
of child behavior (Mash, Johnston, & Kovitz, 1983). The finding that high-aggression
parents view their sons as high extemalizers can be interpreted in several ways. While
the accuracy of parental perceptions in this sample cannot be addressed within the
confines of this study, multiple explanations of the results should be considered.

The high-aggression parents may be biased toward viewing their sons' behavior in
a negative light. In particular, mothers prone to physically abuse their children hold
inaccurate interpretations and expectations of their children, and these inaccurate
interpretations and expectations become more distorted and biased as distress levels

increase (Ammerman, 1991; Milner, 1993). Many other studies have found persistently

80

 

 

- n

inaccurate parental perceptions of their children’s behaviors, despite participation in a
program designed to address such incorrect perceptions (Barber, 1992; Perry, Wells, &
Doran, 1983; Whipple & Wilson, 1994).

Parents who view their children as high extemalizers may be less able to alter the
negative patterns established early on between themselves and their children, while low-
aggression parents experience improving perceptions of their children over time. Mash
and Johnston (1990) describe some parenting behaviors as “automatic” in the sense that a
history of repeated experiences with a particular child (e.g., acting out behaviors) may
result in certain parenting responses (e.g., physical aggression), preceded by little parental
appraisal of the event. They report that “parents of problem children form a cognitive
map for anticipating deviant behavior that functions not only as a guide for parent
behavior but also as a filter through which child behavior is interpreted” (p. 316). In fact,
the best predictors of stability in early child behavior problems are a family climate of
disruption and child-rearing practices (Campbell, Breaux, Ewing, & Szumowski, 1986).

Another interpretation is that certain families genuinely experience difficult child
management problems as children progress through developmental stages, and thus resort
to physical aggression, while children in other families do not present such challenges
(and thus are not victims of parental physical aggression). Donenberg and Baker (1993)
speculate that in a reasonably well-adjusted family, there may be a cumulative process,
with stress focused in the early years on difficulties posed by the child and progressively
generalizing to the family system if the child’s problems have not been reduced. With

this interpretation, child behavior is an initial stimulus and ongoing “cause” of parental

81

aggression. However, there is little evidence to support the notion that difficult child
behavior in abused children precedes harsh parental treatment (Mash & Johnston, 1990).
Thus, when a mother reports that she uses considerable aggression when disciplining her
child, she may also exaggerate her child’s behavioral problems, justifying her use of
aggression. This possibility has been suggested elsewhere (O’Keefe, 1994).

Children’s behavioral difficulties may be the cause of parental stress, the
consequence of parental stress, or, more likely, a combination of both, in an escalating
cycle of child misbehavior and parental reactions (Baker & Heller, 1996). Children with
externalizing problems are often reciprocating family aggression and parental
punitiveness with high levels of aggression (Dadds, et al., 1992; O’Keefe, 1994;
Patterson, 1982). Such child behavioral difficulties are closely related to widespread
negative family interaction patterns (Dadds, Sanders, Morrison, & Rebgetz, 1992), and
enhance existing levels of conflict in the home environment.

Conflict Results of the current study strongly support many others (Belsky,
Lerner & Spanier, in press; O’Keefe, 1994; Perry, Wells, & Doran, 1983; Straus &
Gelles, 1990; Vondra,, 1990; Wolfe, 1985) that have found greater parental aggression in
highly conflicted homes. Physically aggressive families in this study demonstrated no
healthy resolve to family conflict. Consistently high levels of conflict may suggest that
changes in the structures of these families’ relationships did not occur, and that family
members were “stuck” in persistent conflict with each other. Rutter (1994) found that
family discord was a substantial risk factor in the absence of family changes (e.g.,

divorce, separation, loss of parent through death), whereas family changes constituted a

82

risk only when they reflected discord. When family relationships in general are strained,
parents tend to express their negative feelings on one child (Rutter, 1994). Thus,
behavior disturbances in children are part of a system that models and reinforces the
particular behavioral “symptoms” (Patterson, 1982). More frequent family conflict
increases children’s negative responding to such conflict (Cummings, 1994).

The importance of family conflict in predicting physical aggression by fathers is
affected by his IQ. That is, fathers with general deficits in cognitive functioning may be
less able to tolerate family conflict before “losing their cool,” as discordant home
environments may overtax the coping resources of lower-IQ fathers. Alternatively, the
role of paternal IQ here may be representative of characteristics that affect the father-child
interaction without being specific to the parenting role (Mash & Johnston, 1990). That is,
fathers with fewer intellectual resources may demonstrate less tolerance for conflict in
other dimensions of their lives.

As a global indicator of the degree of fiiction present in the family environment,
conflict may include marital troubles, difficulty between parents and children, and
contention among siblings, and as such, has been identified as a key factor in
understanding the etiology and maintenance of aggressive behavior (Doumas, Margolin,
& John, 1994). Indeed, different forms of family violence co-occur and are highly
correlated, including verbal, physical, and escalated aggression (McCloskey, Figueredo,
& Koss, 1955). Thus, a home environment that is characterized by conflict is a
particularly strong marker for high levels of aggression—suggesting not only parent-to-

child aggression, but parent-to-parent, child-to-parent, and sibling aggression as well

83

(Jouriles, Barling, & O’Leary, 1987).

The effects of exposure to various forms of family conflict are profound on
children by elevating children’s stress, fear, and feelings of helplessness, often resulting
in behavioral and emotional disturbances (Abidin, Jenkins, & McGaughey, 1992;
O’Keefe, 1994). Children are sensitive to conflict and discord in the family even when it
is nonviolent, and children’s perceptions of level of conflict in the home are more
predictive of children’s psychological adjustment than are parents’ reports of conflict
(McCloskey, Figueredo, & Koss, 1995). Certainly this area deserves continued study.
Stobilig io Risk Indicators

The results of the current study indicate that chronicity of risk is associated with
physical aggression by parents. Risk fluidity mattered little in predicting parental
aggression. Does this mean that one can predict parental physical aggression by
identifying a young child as problematic or a family conflicted at one point in time? Not
necessarily. Roughly half of the children identified with problems at preschool age will
continue on a path toward externalizing problems, but the other half will improve with
development (Campbell, 1997). Similarly, many families typically experience periods of
conflicted interactions based on transitions and change (Emery, 1992), but are able to
firnction without resorting to physical aggression. Such conflict and externalizing
behaviors may be age-appropriate and short-lived manifestations of stress. According to
the data presented here, these temporary periods of increased family stress are not likely
to predict physically aggressive parenting.

The factors that account for such different pathways to aggression or

84

nonaggression may be parenting and the family environment. Some parents may
skillfully help some hard-to-manage children, while others may exacerbate conflict and
fuel noncompliance. For example, fathers with fewer intellectual resources may
exacerbate conflict instead of helping to resolve it productively. Variations in parental
warmth, support, and appropriate control may also be major factors in determining the
outcomes of early externalizing behavior (see Campbell, 1997). In addition, aspects of
the family environment, such as marital distress and other challenges, may impact
outcome. These environmental factors may directly affect the child, because they create a
climate of tension and conflict in the home. They may also have indirect effects on the
child due to their disruptive effect on parenting (Campbell, 1997).

The impact of the chronicity of risk factors may be further intensified by the
interconnected nature of such components within the family system. That is, conflict
within the family, child externalizing problems, and parental aggression can all be
considered both a cause and a product within the larger family system. Many studies
have highlighted the systemic nature of family conflict and aggression, in that the
behavior of all participants effects the balance in multiple levels of the family ecosystem

(Emery, 1992).6 As parents and children are two separate but connected elements of the

 

Although reciprocal influences in the parent-child interaction are acknowledged,
there are presumed to be asymmetries in the amount of influence that child, parent, or
environmental characteristics can exert (Mash & Johnston, 1990): other longitudinal
studies have determined that family discord preceded children’s externalizing behaviors,
and that reductions in family discord are associated with reductions in children’s acting-
out behaviors (Rutter, 1994). Increased conflict in the home results in the perception of
increased child difficulties, as mothers and fathers from conflicted homes display a
general tendency to rate their children more negatively than parents from low-conflict

85

larger family system, so too are individual parents independent contributors to the family
environment.
Differences between Fathers and Mothers

One unanticipated but significant finding of the current study was gender
differences in the role of risk indicators and physical aggression. Other studies have also
discovered differences between parents’ views of their children and the parenting role.
Baker and Heller (1996) found that mothers and fathers did not differ in actual perceived
level of child behavior problems, although both believed that mothers saw more
problems. Moreover, mothers experienced increased stress and a need for help with both
moderate and high child externalizing behaviors, whereas fathers were not elevated on
these measures unless the child’s externalizing behaviors were very high. This difference
between mothers’ and fathers’ perceptions of child difficulty may be explained in part by
the fact that fathers have been found to spend more time with their male children, and act
toward children in more gender stereotyped ways than do mothers, and thus might be
more accepting of externalizing behaviors (Baker & Heller, 1996). 7

The current study builds upon these findings by indicating the role of child
externalizing behavior in fathers’ parenting stress in two dimensions: [1] levels of child

externalizing behavior are indeed important in predicting fathers’ physical aggression, but

 

homes (Smith, Berthelsen, & O’Connor, 1997).

7

Perry and colleagues (1983) also found that fathers from abusive households differed
from their non-abusive controls in their perceptions of greater family conflict and
inaccuracy in their expectations for their children, while maternal abusiveness was related
to her lower-self esteem and greater anxiety.

86

that [2] particularly high levels of child externalizing behavior may be subsumed into
fathers’ rating of the broader family context. Thus, fathers are affected by sons’
externalizing behaviors, especially if they are particularly high. If behavior problems are
suficiently high, fathers experience greater stress levels and include such problematic
behavior in a global rating of family system negativity and coercion.

A different picture emerges for mothers: while fathers’ aggression was mainly
predicted by overall conflict in the home, mothers’ physical aggression was dependent on
her ratings of child difficulty. This finding is not surprising given mothers’ greater
sensitivity to sons’ externalizing behaviors. It could be that the level of externalizing
behaviors reported in this study are sufficient to increase the likelihood of maternal
aggression, while insufficient to serve as the primary cause of paternal aggression.

The importance of mothers’ perceptions of child difficulty has been established in
the literature. Abidin, Jenkins, and McGaughey (1992) report that mothers are more
likely than fathers to rate sons higher on behaviors associated with conduct disorder. In
turn, other studies (Dadds, et al., 1992) have found that mothers-but not fathers-of high
extemalizers display more aversive behavior to their children than non-behaviorally
troubled children, and that mothers of high extemalizers experience higher levels of stress
and feelings of depression than do fathers (Baker & Heller, 1996). This differential
response to children is especially important given that mothers’ characteristics, rather
than fathers’, are more closely associated with boys’ future problem behaviors (Abidin,
Jenkins, & McGaughey, 1992).

Another interesting finding is that mothers’ physical aggression is predicted from

87

fathers’ risk factors, while paternal physical aggression is not predicted from mothers’
risk factors. Several explanations regarding gender and social role differences may
account for this finding. Firstly, it may indicate that fathers in this sample are unaffected
by or insensitive to the level of stress experienced by their female partners. Mothers may
be more “tuned in” to and affected by the risk indicators in her husband’s life, indicating
a spillover effect whereby paternal factors affect maternal parenting styles. This may be
especially notable in families where mothers hold greater responsibility for discipline, so
that high levels of paternal and maternal stress are channeled through her to the child.
Because women carry the major burden of child care, and are socially expected to do so,
mothers as a group may be more psychologically involved in the role of parent (Crnic &
Acevedo, 1995).

Other studies (Baker & Heller, 1996) have found that although mothers and
fathers both acknowledged their children’s behavioral difficulties, fathers were much less
likely to assume personal responsibility for those problems. Thus, greater maternal
reactivity to child externalizing behavior may be due to “traditio ” social forces
influencing parental roles.

Relatedly, there may exist a discrepancy in power between men and women in this
sample, where mother “follows the lead” set by father and acts aggressively against her
child based on fathers’ perception of their circumstances, rather than her own. Mothers in
high-conflict homes have reported that they were undermined in their parenting by their
partner, and that they would change their parenting practices according to the presence or

absence of their partner (Smith, Berthelsen, & O’Connor, 1997). Thus mothers may be

88

physically aggressive based on fathers" view of the need for such discipline.
Alternatively, mothers may try to protect their children from being the target of their
partner’s temper or violence by excessively controlling or suppressing those aspects of
their child which might irritate or provoke their partner (e.g., externalizing behaviors).
The fact that maternal physical aggression is better predicted by fathers’ risk indicators
than her own indicates that the relationship between mothers and parental physical
aggression may be mediated by fathers’ stress levels.

In general, mothers’ behaviors may indicate a better and more sensitive marker for
family problems than fathers’ behaviors. Fathers’ reports of conflict in the home, as well
as his perceptions of difficulty with children’s behavior, should be examined not only in
their own right, but also for the impact that such factors may have on mothers’ parenting.
Regardless of which parent may be targeted for intervention due to concern of physical
maltreatment, his or her partner’s individual risk indicators should be addressed and
included both in research and in treatment.
5mm

Based on the findings presented above, it is affirmed that parental physical
aggression is an outcome associated with the parent’s attempt to cope with multiple
aversive factors in his or her family environment (i.e., high levels of family conflict and
child externalizing behaviors) that are constant across developmental periods. Thus,
results from this study in general support the notion that consistent, daily life stressors
have a greater impact on parenting behaviors than do major life events.

Moreover, it is especially important to note that variables specifically related to

89

the parenting and family roles, rather than individual or community level factors, are the
greater predictors of parental physical aggression. Although it is unclear to what extent
child characteristics contribute to the etiology of parental physical aggression, the results
of this study add to the growing body of data that implicates certain child behaviors in the
escalation of parent-child conflict. In addition, these results also focus our attention on
family-level interchanges, rather than individual parent characteristics, as factors
influencing parental aggression. This may aid in removing stigrnas and negative
assumptions that are associated with certain individual parental risk factors, such as
substance use and depression.

The findings of the current study reveal that aggressive parents feel that child
noncompliance and family conflict are stable features of their family context. In sum,
high aggression families consistently experience more opportunities to discipline their
children, are more likely to respond with punitive techniques, and are characterized by
mutually aversive interchanges among family members.

Implications

Examination of the relationships among parental risk indicators across time and
between parental risk factors and physical aggression ofl‘ers valuable evidence about risk
markers for physically aggressive parenting. The results of this examination also indicate
the need to address [1] the effects of aggressive parenting on children, [2] critical points

of intervention, and [3] future directions for parenting research.

90

Risk Markers for Parental Agggession

Current knowledge of what constitutes “risk” for parental physical aggression is
updated with the findings presented in this study. Despite the many variables that have
been associated with child maltreatment, including parental psychopathology,
socioeconomic status, and substance abuse, only conflict within the family and child
externalizing behaviors were implicated in predicting parental physical aggression.
While these risk indicators have been previously considered in relation to child
maltreatment, a study that examines family conflict and child externalizing behavior over
long periods of time has not existed in the literature. In addition, the predictive power of
perceived conflict and child problem behavior, relative to other known risk indicators, has
not been previously explored.

Cross-sectional studies that have examined risk indicators for aggressive
parenting may present a misinformed view of what is relevant. For example,
maltreatment studies that point to parental depression or substance use as causal agents
may in fact be missing the larger picture of what predicts parental physical aggression. It
seems that factors closest to the parenting context (i.e., conflict at home and child
behavior), not to the individual parent (i.e., depression, alcohol consumption) per se, nor
to broader community elements (i.e., socioeconomic status), are most relevant when
considering physical aggression fi'om parent to child.

ff ts f P s ' on '1 n
Aggressive parents may be teaching their children that aggressive behavior is

appropriate. Parental aggressiveness has been associated with a range of negative

91

 

developmental outcomes including delayed impulse control and lowered social
competence, impaired psychological adjustment, aggression, delinquency, and physical
child abuse, reduced responsiveness to the pain of others, and aggression in offspring
several generations removed (Campbell, 1997; Doumas, Margolin, & John, 1994;
Holden, Coleman, & Schmidt, 1995; McCord, 1988).

McCord (1988) found that children reared by aggressive parents were more
expressive in general, including being fiequently annoyed and showing warmth toward
intimate partners, suggesting an absence of inhibition in children raised in aggressive
homes. As adults, these children were more likely to feel that expressive (including
injurious) behavior is normal and often justified. Aggression and externalizing behaviors
are especially stable in males (Crockenberg & Lourie, 1996; Doumas, Margolin, & John,
1994).

Studies suggest that there is both a direct effect of harsh discipline on aggression
in the home and an indirect effect on aggression with peers through a hostile attributional
bias: parents’ power assertion is a strong predictor of children’s power assertion with
peers and peer rejection (Campbell, 1997; Patterson, Dishion, & Bank, 1984; Rutter,
1994; Strassberg, Dodge, Bates, &. Pettit, 1992). Mothers’ and fathers.’ use of coercion
with their children correlated both concurrently at age 2 and longitudinally with
children’s self-reported conflict strategies with peers and parent-rated adjustment and
social competence at age 6 (Crockenberg & Lourie, 1996). Specifically, parents’ use of
aggression is related to more negative behavior with and social rejection by peers,

indicating the long-term deleterious effects of aggressive parenting on both peer relations

92

and adult criminality (Strassberg, et al., 1992).

Family conflict itself is a strong predictor of child externalizing problems. Living
in a home characterized by high levels of conflict is stressful for children and increases
their aggression: between 9 and 25% of the variance in children’s externalizing disorders
is typically accounted for by conflict within the home (see Cummings, 1994). Children
report feelings of anger, sadness, or fear in response to family “background anger,” and in
some contexts also report feelings of guilt, shame, and worry (Cummings, 1994).

Similar to child externalizing behaviors, a conflicted home environment has
negative effects on child-parent and peer relationships (Cummings, 1994). Cummings’
(1994) review of the literature reveals that general levels of conflict are better predictors
of disturbances in the mother-child relationship than certain maternal characteristics (e.g.,
depression). Moreover, there is evidence that sensitization to destructive conflict occurs,
resulting in the child’s greater reactivity over time. The evidence of children’s m
vulnerability to family conflict as children get older draws attention to the need for early
intervention.

Physical punishment from parents-most notably in the form of spanking—has been
and continues to be widely practiced among parents (Holden, Coleman, & Schmidt,
1995). Yet spanking, especially frequent spanking, is a humiliating technique that may
diminish problem-solving capacities among children (Hyman, 1995). Society can
function without resorting to the use of corporal punishment on their children. Currently,
Sweden, Finland, Denmark, Poland, Norway, and Austria ban the use of corporal

punishment on children. Government sanctions against corporal punishment in the home

93

 

have not led to governmental micromanagement of home discipline or arrests for
spanking, as opponents allege would happen in America (Hyman, 1995). Hyman (1995)
points out that these sanctions do set the moral tone for a country and result in far fewer
spankings and perhaps lower levels of child abuse.

Intervontion. Efforts to reduce parental reliance of physical aggression are crucial,
given its association to multiple negative outcomes in children. Parent-training programs
show evidence of improvement on measures of family well-being following intervention
(Donenberg & Baker, 1993). An important caveat is the need to focus on more effective,
rather than simply more, parental discipline. Studies consistently show that parents of
high extemalizers tend to be more punitive than are parents of low-extemalizers, and
utilizing more extreme punishment, such that many researchers feel that such excessive
punishment causes the externalizing behavior in children (see Patterson, Dishion, &
Bank, 1984). In cases of high extemalizers, the reaction to aggressive punishment is very
likely to be that of accelerating his/her ongoing coercive behavior. Parent use of “time-
out” or withdrawal of privileges has been shown both to suppress the ongoing aggression
and also weaken the stimulus-response; parents trained in the use of these family
management skills have produced significant reductions in observed externalizing
behaviors, and the effects have been shown to persist for at least 12 months (Patterson,
Dishion, & Bank, 1984).

In cases where reports of conflict and child behavior problems come to attention,
interveners should not assume that the difficulties experienced by these families are

transitional. Attention should be focused on early identification of and intervention for

94

 

children with high levels of externalizing behaviors, preferably before elementary school
begins and these behaviors-become more ingrained in the child’s repertoire. In addition,
when planning family-based early intervention programs, it is important to understand
how mothers and fathers may see and react to their children’s behavior differently. Early
intervention is critical in that sequelae of adverse experiences in early childhood may be
less than those associated with the same experiences in middle childhood (Rutter, 1994).

In addition, parent-training programs could be broadened to include a greater
emphasis on parents’ stress and coping. Lowering family stress may enhance and
maintain changes in child behavior following interventions, as high parental stress can
interfere with carrying out a behavior management program (Donenberg & Baker, 1993).
Program components should focus on long-term support for child- and family-related
stress.

Future Directions

This study leaves room for future examination of the role of parental competence
in predicting aggression toward children. Competence has been shown to be a powerful
moderator between child-related stressors and child abuse potential, and its relationships
to parental depression and alcohol consumption, as well as child behavior and family
conflict, deserves attention. Other variables that play a role in the etiology of parental
aggression, such as parental and child personal resources and the parent-child
relationship, must be included in future research in order to begin to understand the
etiology and maintenance of aggressive behavior in the family system.

Another important area for continued study is the role of cognitive appraisal in

95

 

parents’ perceptions of child behavior. Researchers believe that parents prone to physical
abuse suffer from a lack of information integration, such that they maintain explanations
of their child’s behavior that are consistent with their own rigidly held cognitive
distortions and biases, which are associated with the use of power-assertive behaviors.
These physically aggressive parents engage in more automatic processing of child-related
data in ambiguous and stressful situations (Ammerman, 1991; Milner, 1993).

In addition, the study of risk factors with the inclusion of sibling measures may
reveal essential nuances in the role of parental physical aggression. Rutter (1994) notes
that, for most outcomes, family influences that impinge differentially on children within
the same family tend to make a greater impact than family influences that impinge on the
family as a whole to a roughly equal degree. It would be helpful to determine the extent
to which each child in the family is subject to parental physical aggression.

Strengths & Limitations

There are several points that must be considered when placing the results of this
study in a larger research context. Strengths of the research design are presented in the
following paragraphs.

First, parental aggression is examined in a sample of initially intact community
families who have not been identified by treatment or clinical status. The participants in
this study were not originally selected on the basis of parenting practices. This unselected
sample provides an opportunity to investigate these relations as they occur in a sample
more representative of the general population than is afforded by the use of samples

selected based on child protective services or child abuse prevention programs. This

96

community sample allows for a more realistic assessment of parenting practices and the
occurrence of parental physical aggression, and greater generalizability of the findings.
However, the trade off for such increased generalizability is the low base rate of parental
physical aggression among parents, which makes the discovery of theoretical connections
more difficult.

Second, information on aggression is gathered from both husband and wife
whenever possible to broaden the perspective and increase the reliability of data.
Previous studies of parenting practices have often neglected to assess the role of the
father in families. Here, the parents are considered as individuals, rather than as a couple,
so that there is no mix of perpetrator/nonperpeu'ator risk indicators. The results of this
study emphasize that a lack of information about fathers’ risk indicators may leave out
critical details of both his parenting practices as well as what is affecting mothers’
parenting.

Third, an expanded definition of parental aggression, rather than physical abuse, is
used, which includes spanking. It is the author’s view that attention to these “less severe”
forms of physical aggression fi'om parents to children offers a greater band of knowledge
about negative parenting practices that are more common than physical abuse. Recent
research that examines spanking has focused public attention on the long-term negative
consequences associated with this common parenting discipline practice, including low
self esteem, greater disobedience, and antisocial behavior (“Spanking,” 1997).

Fourth, data are analyzed separately for mothers and fathers to determine the

differential effects of aggression by gender. Such a strategy revealed important

97

differences in the roles that fathers’ versus mothers’ risk factors and aggression play in
parenting.

Several limitations of the reported study are related to restrictions based on the
sample and/or the dataset. First, the distribution of physical aggression scores were
highly J-shaped. Such a distribution required the use of percentile ranking before
conducting regression analyses to reduce potential bias produced by several outlying
cases. In this case, the strategy is more conservative and less likely to result in Type 1
errors. However, the impact of such outliers may be valuable information lost, and
further research is suggested to test the significance of such a skewed distribution. In
increasing the understanding of dimensions of family stress, it is imperative to consider
the nature of the samples on which studies are based, and the degree to which findings are
sensitive to these samples. The following limitations are based on this consideration.

A second limitation is that several of the highest risk families that are part of the
larger (MSU-UM Family) Study were not included in this sample due to the selection
process. That is, these families completed a shortened version of the Conflict Tactics
Scale, the measure used to create the outcome variable of physical aggression, and
information provided by them could not be used. The benefits and costs of the exclusion
of such cases are that the distribution of physical aggression may more realistically
resemble the population making the results generalizable, but that key relationships
between indicators of higher-risk families and physical aggression may remain
undiscovered.

Third, the parenting practices used to create the outcome measure of physical

98

 

aggression were self-reported, and may not realistically represent the amount of physical
aggression occurring within homes. It has been posited that people have either a negative
or positive reporting bias that is used for both reports of the self and of others, thereby
increasing the association between parent and child variables (Doumas, Margolin, &
John, 1994). However, self-report data about parenting practices can provide information
about the history of relationships beyond that provided by observations of interactions
(Strassberg, Dodge, Bates, & Pettit, 1992). Moreover, studies have shown that self-
reports of parental aggression and physical punishment are coded reliably and have high
internal consistency (Holden, Coleman, & Schmidt, 1995). .

Fourth, due to the nature of the sample for the larger (MSU-UM Family) Study,
physical aggression could be examined at Wave 3 only. There is evidence that rates of
parental physical aggression peak with children who are from three to five years old and
then decreases (Ammerman, 1991; Holden, Coleman, & Schmidt, 1995). Therefore, the
design of this study may have “missed” families who were highly aggressive in early
years of parenting. However, predicting physical aggression when children are older, as
was done here, may indicate cases with more serious parental aggression whose past
physical aggression remains unknown. In addition, research has shown that abusive
dimensions in parent-child relationships can begin at any time (Milner, 1993); thus,
longitudinal examination of levels of parental physical aggression, and how such levels
related to risk levels over time, should reveal fruitful detail regarding the nature of
parental aggression.

Relatedly, younger children tend to score significantly higher on externalizing

99

behavior scales than older children (O’Keefe, 1994), and there is evidence that parents
judge aggressive acts performed by 12-year-olds far more severely than they judge
aggressive acts performed by 5-year-olds (Wenger, Berg-Cross, L., & Berg-Cross, G.,
1980). Thus parental perception of child aggressiveness may be a key determinant of
their choice of “appropriate” punishment, and the age of the child may be the most
important factor in determining their response to child aggression. In this broader view, it
may be that parents increase their socialization efforts over time, in which case the
advanced age of the children examined here is apropos for understanding aggression
within families.

Fifth, due to the nature of the sample for the larger (MSU-UM Family) Study, this
sample includes only male children. The path of parental physical aggression for boys
and for girls may be different, leaving room for future studies to examine the models
tested here with female children. Many studies have pointed to boys’ and girls’ different
developmental trajectories for aggression and externalizing behaviors. Specifically,
mothers become more negative about aggression in girls and less negative about it in boys
between the ages of 4 and 6, and children understand that aggression by girls is more
likely to be punished than aggression by boys (Crockenberg & Lourie, 1996; Doumas,
Margolin, & John, 1994).

Yet the association between exposure to familial aggression and poor child
adjustment is exhibited more in boys than girls (Abidin, Jenkins, & McGaughey, 1992;
Doumas, Margolin, & John, 1994). Hetherington and colleagues (1982) and Abidin et al.

(1992) ofier considerable evidence that boys are much more susceptible than girls to

100

 

negative outcomes in response to stressful events in the family; these authors suggest that
parents are less restrained in the presence of their sons, tending to quarrel and fight more
often in their presence than in daughters’ presence. J ouriles and LeCompte (1991) also
found that marital aggression covaries with higher levels of both mothers’ and fathers’
aggression toward boys but not toward girls, consistent with the hypothesis that female
victims of marital aggression displace some of the anger they feel toward their husbands
onto their male children.

However, past literature (Holden, Coleman, & Schmidt, 1995; Wenger, Berg-
Cross, L., & Berg-Cross, G., 1980) also indicates that parents do not respond differently
to male and female children’s aggression, and that there are no significant differences in
responses of different sex parents to children’s aggression. In addition, O’Keefe (1994)
found no gender differences in children’s externalizing behaviors based on severity of

mother-child aggression. Future research is needed to clarify such discrepant findings.

101

 

APPENDICES

Appendix A

Table 9

Variable Means, Standard Deviations, & Range of Predictor Variables
(Total Sample, N=218)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variable Mean (>7) Standard Minimum Maximum
Deviation (o)
ASB 13.3 1 9.28 l 56
BECKI 2.48 2.91 00 14
BECK2 3.06 3.39 00 19
BECIG 3.11 3.51 00 20
CHEXTI 1 1.82 6.23 00 28
CHEXTZ 1 1.14 6.35 00 30
CHEXT3 10.25 7.27 00 38
CNFLCTI 3.51 2.14 00 9
CNFLCT2 3.50 2.04 00 9
CNFLCT3 3.64 2.20 00 9
EDUC 13.42 1.89 9 20
SESl 317.75 103.78 171 640
SES2 343.93 130.62 162.5 770
SES3 339.46 127.29 177 866
HAMl 17.56 7.80 6.73 48
HAM2 16.42 7.39 8 38
HAM3 15.44 8.01 00 47
QFVRI 1484 3998.83 1 21000
QFVR2 1013.65 3127.02 1 21000
QFVR3 333.93 1528.33 1 18900
IQ 99.47 13.90 62 146
PHYSAGG 5.02 17.87 0 208

 

103

 

Appendix B

 

 

 

 

 

 

 

Means, Standard Deviations 2113:111ng 2f Mothers’ Predictor Variables
(N =1 1 5)

Variable Mean (>7) Standard Deviation (0) Minimum Maximum
ASB 10.27 6.79 1 35
BECKI 2.78 3.29 00 14
BECKZ 3.40 3.58 00 19
BECK3 3.39 3.86 00 20
CI-IEXTI 12.33 6.20 00 28
CHEXTZ 12.01 6.47 00 30
CHEXT3 1 1.01 7.45 00 38
CNFLCTl 3.81 1.94 00 8
CNFLCT2 3.72 1.94 00 9
CNFLCT3 3.85 2.12 00 9
EDUC 13.17 1.82 9 l9
SESl 317.13 103.11 171 640
SES2 344.30 133.37 162.5 770
SES3 339.10 127.29 177 866
HAMl 19.03 8.34 8 48
HAM2 17.65 7.68 8 38
HAM3 16.76 8.69 00 47
QFVRl 580.08 2067.73 1 15750
QFVR2 785.98 3150.5 1 21000
QFVR3 300.24 1837.25 1 18900
IQ 95.96 13.32 62 135
A BECK 4.27 3.82 00 19
A HAM 12.85 8.85 00 39
A QFVR 1461.42 4890.80 0 31475.0

104

 

 

(Table 10, con’t)

 

 

 

 

 

 

 

Variable Mean (2) Standard Deviation (a) Minimum Maximum
A Cl-I EXT 8.65 5.15 00 22.06

A CONFLC 2.34 1.69 00 8

A SES 116.53 126.54 00 798
PHYSAGG 7.29 24.19 0 208

 

105

 

AppendixC

 

 

 

 

 

 

l 1
Means, Standard Deviations &Tl::nege 1of F amefi’ Predictor Variables
(N=103)
Variable I Mean (2) Standard Deviation (a) Minimum Maximum
ASB 16.16 10.36 2 56
BECKI 2.1 1 2.34 00 1 1
BECKZ 2.61 3.16 00 13
BECK3 2.63 2.86 00 13
CHEXTI 10.98 6.03 1 27
CHEXTZ 10.09 6.08 00 30
CHEXT3 9.33 7.05 00 37
CNFLCTI 3.15 2.30 00 9
CNFLCT2 3.25 2.13 00 8
CNFLCT3 3.42 2.28 00 9
EDUC 13.80 1.92 1 1 20
SESl 322.44 105.62 171 640
SES2 347.30 129.86 162.5 770
SES3 343.65 129.55 177 866
HAMl 15.43 6.35 6.73 41
HAM2 14.73 6.66 8 38
HAM3 13.80 6.77 8 36
QFVRl 2540.84 5331.35 1 21000
QFVR2 1271.42 3160.25 1 21000
QFVR3 365.83 1 102.03 1 9220.1
1Q 103.45 13.73 76 146
A BECK 3.34 3.65 00 23
A HAM 11.10 8.31 1 41
A QFVR 3068.80 5694.99 00 24370

106

 

(Table 11, con’t)

 

 

 

 

 

 

 

Variable Mean ( >7) Standard Deviation (a) Minimum Maximum
A CH EXT 7.83 5.71 00 30

A CONFLC 2.48 2.01 00 11

A sas 109.17 103.80 00 496
PHYSAGG 2.39 3.46 ' o 18

 

107

 

SQVQ 1:... .c.Vn .... mQVQ ...

 

 

 

 

 

 

 

 

 

 

8.. 36mm... N8.- aNc. no... No. .3..- 3...- an... 5N7 O.
8.. .....MNmr $8. 3..- aNcr N... 8..- >8. .3. 26m
8.. .N..- 28. v3. :nNN. or... 3.. .8.- mm<
8.. No. .- o8. m8.- eco. 8o. 58. mum <
8.. mm .. 8.. .cmN. oN.. .N... 88220 4
.Exm
8.. N3. 5.. 3.. 8.. 2..-.0 <
22260
8.. n8. 3... m8. 2< <
non—.26
8.. an... on..- 58>» 4
.g
8.. a8. .2250 <
a?
8.. S...
mmm 2:260 m2.~..an§m 22.92328 2238an ..o>o - 2232qu Amw<
o. 2.8 2?. a < ......o < .282 a 3263. < < .2...
22mm. «34236 .3 -.. a .2883... 95.3.: ...“. ... ...-.“> .....B . .-u “:11 -. .. .. .1 .
N. 0.22..-

Q x3209?

108

 

 

 

 

SQVQE: .o.va: mc.vn._
8.. ......Nen. o3..- . 8m. 3.. 2...- n8.- nncr :3. «co.- 0.
8.. 3.8m.- .. 8.. 28. 2....- ..8: no..- :3.- N8.- 26m
8.. one: pm... 3... 2.3m. .33. 3.. N2..- mm<
8.. 2...- 3.... 3...- So. «we. 08.- mmm 4
65:00
8.. m8. .. .2. - 98.- 3.... 2.... <
.Sxm
8.. «3. m3. «3. 35. ......0 <
5.200
8.. 8... m2. m2. o.< <
con. .85
8.. 3cm. 3... 3.83 d
.g
8.. 85. 82.30 4.
a}.
8.. 2:
mum 8.....80 .5xm. 5.980 ion. .85 ion. 5m< -
o. 2.8 mm< < < 25 < 2< < .353 < .856 < 2:

 

 

 

 

 

 

no.8- m “(Such ... #8.. .3889. ..

o W.

m. 03a...

m 522.?

4mm fi—...-$.>Ou 3.190 .30. .2 “=03-.. -..-

109

Appendix F

Table 14

Summm of Hierarchical Regression Analysis for Physical Aggression Regressed onto
Stable Backgmund Characteristics and Predictor Variables

 

 

 

 

 

 

 

 

 

 

Fathers Mothers
Step Variable B R’ AR’ [3 R’ AR2
1 IQ - .18 .03 n/a .01 .01 n/a
Antisociality - .04 .11
Education .11 - .03
2 IQ - .19 .07 .04 .00 .03 .02
Antisociality .00 .08
Education .11 - .01
Change in Depression:
Current - .03 .03
Worst - .19 .12
3 IQ - .20 .07 .00 .00 .04 .01
Antisociality -.Ol .06
Education .1 l - .01
Change in Depression:
Current - .05 .02
Worst - .22“ .12
Change in Alcohol Consumption .01 ..ll
4 IQ - 20 .09 .02 .00 .04 .00
Antisociality -.01 .05
Education .12 - .01
Change in Depression:
Current -.05 .02
Worst - .22‘ .11
Change in Alcohol Consumption .01 .11
Change in Child Extemalizing .16 .05
5 IQ - .19 .09 .00 .00 .04 .00
Antisociality -.01 .05
Education .12 - .01
Change in Depression:
Current -.05 .02
Worst - .23" .12
Change in Alcohol Consumption .01 .11
Change in Child Extemalizing .15 .04
Change in Conflict .04 .Ol

 

110

(Table 14, con’t)

 

 

Step Variable B R1 AR’ [3 R’ AR1
6 IQ - .19 .09 .00 .00 .04 .00
Antisociality -.Ol .05
Education .12 - .01
Change in Depression: .
Current - .05 .02
Worst - .23‘ .11
Change in Alcohol Consumption .01 .11
Change in Child Extemalizing .15 .04
Change in Conflict .04 .02
Change in SES .00 .04
'ps.05 "p501 ”‘psflOl

Ill

Appendix G

Table 15

Summm of Regression Analyses for Physical Aggression Regressed onto Change

ngables Separately

 

 

 

 

 

 

 

 

 

 

 

 

ModeWariable(s) Fathers Mothers

B R2 [3 R2
Change in Depression: .04 .02
Current - .03 .04
Worst - .19 14
Change in Alcohol Consumption .02 .00 .13 .02
Change in Child Extemalizing Behaviors .10 .01 .06 .00
Change in Conflict in Family Environment .02 .00 - .01 .00
Change in SES .03 .00 - .04 .00

 

* p s .05 “ p s .01 *** p s .001

112

 

Appendix H

 

 

 

 

 

 

 

 

 

 

Table 16
Sum of Hierarchical Regression Analysis for Physical Aggression Regressed onto
Predictor Variables
Fathers Mothers
St Variable [3 R2 AR2 B R2 AR2
CP
1 Change in Depression: n/a n/a
Current - .03 .04 .04 .02
Worst - .19 .14
2 Change in Depression:
Current - .03 .04 .00 .02 .04 .02
Worst - .19 .13
Change in Alcohol .04 .12
Consumption
3 Change in Depression:
Current - .05 .06 .02 .02 .04 .00
Worst - .29“ .13
Change in Alcohol .04 .12
Consumption .15 .05
Change in Child Extemalizing
4 Change in Depression:
Current - .06 .06 .00 .02 .04 .00
Worst - .23“ .13
Change in Alcohol .03 .12
Consumption .14 .05
Change in Child Extemalizing .06 .02
Change in Conflict
5 Change in Depression:
Current - .06 .06 .00 .02 .04 .00
Worst - .23“ .12
Change in Alcohol .04 .12
Consumption .14 .04
Change in Child Extemalizing .06 .02
Change in Conflict .03 .03
Change in SES

 

*ps.05 **p$.o1

m p s .001

113

 

 

Appendix I
Table 17

SW of Hierarchical Regression Analysis for Physical Aggression Regressed onto
Predictor Variables and Stable Background Characteristics

 

 

 

 

 

 

 

 

 

Fathers Mothers
Step Variable B R2 AR2 B R2 AR2
1 Change in Depression: .04 n/a .02 n/a
Current -.03 .04
Worst - .19 .14
2 Change in Depression: .04 .00 .04 .02
Current - .04 .02
Worst - .19 .13
Change in Alcohol Consumption .04 .12
3 Change in Depression: .06 .02 .04 .00
Current -.05 .02
Worst - .22* .13
Change in Alcohol Consumption .04 .12
Change in Child Extemalizing Behaviors .15 .05
4 Change in Depression: .06 .00 .04 .00
Current - .06 .02
Worst - .23“ .13
Change in Alcohol Consumption .03 .12
Change in Child Extemalizing Behaviors .14 .05
Change in Conflict .06 .02
5 Change in Depression: .06 .00 .04 .00
Current - .06 .02
Worst - .23"' .12
Change in Alcohol Consumption .04 .12
Change in Child Extemalizing Behaviors .14 .04
Change in Conflict .06 .02
Change in SES .03 .03
6 Change in Depression: .09 .03 .04 .00
Current - .06 .02
Worst - .23" .11
Change in Alcohol Consumption .01 .11
Change in Child Extemalizing Behaviors .15 .04
Change in Conflict .04 .02
Change in SES .00 .04
IQ - .19 .00
Antisociality -.01 .05
Education .12 - .01

 

 

 

 

 

 

 

 

 

 

"‘ p s .05 " p s .01 "”' p s .001
114

Appendix J

Table 18

Summary of ngsign Analyses for Physical Aggression Rggressed onto

Concurrenr Variables Separately

 

 

 

 

 

 

 

 

 

 

 

Fathers Mothers
ModeWariable(s) B R2 B R2
Concurrent Depression: .00 .02
Current - .02 .12
Worst - .01 .03
Concurrent Alcohol Consumption .13 .02 .05 .00
Concurrent Child Extemalizing Behaviors .23* .05* 33*" .11""‘
Concurrent Conflict in Family Environment 34*“ .12"* .22‘ .05"
Concurrent SES .0] .00 .01 .00

 

"' p s. .05 " p s .01 ""' p s .001

115

 

Appendix K

Table 19

Summary of Hierarchical Regression Analysis for Physical Aggression Regressed
onto Concurrent Predictor Variables

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fathers Mothers
Step Variable B R2 ARz B R2 AR2
1 Concurrent Depression:
Current - .02 .00 n/a . 12 .02 n/a
Worst - .01 .03
2 Concurrent Depression:
Current - .02 .02 .02 .12 .02 .00
Worst - .02 .02
Concurrent Alcohol Consumption .14 .01
3 Concurrent Depression:
Current - .02 .07 .05 .09 .12” .10
Worst - .09 - .ll
Concurrent Alcohol Consumption .11 .04
Concurrent Child Extemalizing Behavior .25" 35"”
4 Concurrent Depression:
Current - .07 .15" .08 .06 .13“ .01
Worst - .06 - .lO
Concurrent Alcohol Consumption .13 .04
Concurrent Child Extemalizing Behavior .11 .32"
Concurrent Conflict .31" .IO
5 Concurrent Depression:
Current - .07 .15" .00 .06 .14” .01
Worst - .05 -.09
Concurrent Alcohol Consumption .13 .04
Concurrent Child Extemalizing Behavior .11 .33"
Concurrent Conflict .3 I " .I l
Concurrent SES .01 .08
Trimmed Models
Fathers Mothers
B R’ Step B R2
Concurrent Conflict 34"" .12‘" l Concurrent Conflict .11 .12""’
Concurrent Child Extemalizing
Behavior .29”
2 Concurrent Child Extemalizing .I 1‘"
Behavior .33‘"
" p s .05 " p s .01 ”'" p s .001

116

 

 

Appendix L

 

 

 

 

 

 

 

Table 20 Summary of Hierarchicg Rggression Analysis (Series E) for Phygigl
Aggression Regressed onto Concurrent Predictor Variables and Stable
Bacl_cgrgund Characteristics * p s .05 “ p s .01 ‘" p s .001
Fathers Mothers
Step Variable B R2 AR2 B R2 AR”
1 Concurrent Depression: .00 n/a .02 n/a
Current - .02 .12
Worst - .01 .03
2 Concurrent Depression: .02 .02 .02 .00
Current - .02 .12
Worst - .02 .02
Concurrent Alcohol Consumption .14 .01
3 Concurrent Depression: .07 .05 .12" .14
Current - .02 .09
Worst - .09 - .11
Concurrent Alcohol Consumption .11 .04
Concurrent Child Extemaling Behavior .25" .35‘"
4 Concurrent Depression: .15“ .08 .12" .00
Current - .07 .06
Worst - .06 - .ll
Concurrent Alcohol Consumption .13 .04
Concurrent Child Extemalizing Behavior .11 .32”
Concurrent Conflict .31“ .10
5 Concurrent Depression: .15" .00 .14" .02
Current - .07 .06
Worst - .05 - .09
Concurrent Alcohol Consumption .13 .04
Concurrent Child Extemalizing Behavior .11 .33”
Concurrent Conflict .31 " .1 l
Concurrent SES .01 .08
6 Concurrent Depression: .21" .05 .15" .01
Current - .09 .06
Worst - .Ol - .13
Concurrent Alcohol Consumption .15 .04
Concurrent Child Extemalizing Behavior .05 .32”
Concurrent Conflict .39‘" .13
Concurrent SES -.Ol .16
IQ - .25‘ - .03
Antisociality - .l l .08
Education .09 - .08

 

 

 

 

 

 

 

 

 

117

 

Table 20 (con’t)

 

 

 

 

 

 

 

 

Trimmed Models
Fathers Mothers
B R2 9 R2
Concurrent Conflict 38*" .15*** Concurrent Child 33*” .11***
IQ - .19“ Ext Beh

 

118

 

 

Table 21

Means, Medians, Ranges & Missing Cases for Original Sample

Appendix M

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variable Mean Median Range Valid Missing
N N
Physical Aggression 5.00 1.00 208.00 219 0
Antisociality 13.24 11.00 55.00 217 ~ 2
Education 13.42 13.00 11.00 219 0
IQ 99.60 99.00 84.00 194 25
Worst-ever Depression 1 17.614 16.00 40.00 215 4
Current Depression 1 2.47 2.00 14.00 213 6
Alcohol Consumption 1 1444.41 90.00 20999.00 215 4
Child Extemalizing 1 11.91 12.00 28.00 210 9
Conflict l 3.51 3.00 9.00 213 6
SES 1 317.47 292.00 469.00 219 0
Worst-ever Depression 2 16.12 14.00 30.00 195 24
Current Depression 2 3.06 2.00 19.00 195 24
Alcohol Consumption 2 1042.05 60.00 20999.00 192 27
Child Extemalizing 2 11.05 10.00 30.00 198 21
Conflict 2 3.48 3.00 9.00 191 28
SES 2 346.01 310.50 607.50 197 22
Worst-ever Depression 3 15.43 12.50 47.00 218 1
Current Depression 3 3.10 2.00 20.00 219 0
Alcohol Consumption 3 288.20 3.00 18899.00 21 1 8
Child Extemalizing 3 10.16 8.50 38.00 206 13
Conflict 3 3.64 3.00 9.00 214 5
SES 3 344.53 316.50 689.00 201 18

 

119

 

Table 22

Appendix N

Original Dataset without Estimation of Missing Values (N=219)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Variable Time 1 Time 2 Time 3
No. of % No. of % N0. of %
Cases Missing Cases Missing Cases Missing
Missing Missing Missing
Antisociality l .5 n/a n/a
IQ 24 11.0 n/a n/a
Education 0 0 n/a n/a
Current Depression 5 2.3 23 10.6 0 0
Worst-ever Depression 4 1.8 23 10.6 1 .5
Alcohol Consumption 3 1.4 26 1 1.9 8 3.7
Child Difficulty 8 3.7 20 9.2 13 6.0
Conflict 5 2.3 27 12.4 5 2.3
Socioeconomic Status 0 0 21 9.6 18 8.3
Missing Values by Case(s)

Variable No. Of % No. Of % No. Of % of

Cases of Cases of Cases Total

Missing Total Missing Total Missing

67% 33% 0%

of of data data

data
Current Depression 1 .5 28 12.8 190 86.8
Worst-ever Depression 1 .5 27 12.3 191 87.2
Alcohol Consumption 3 1.4 33 15.1 183 83.6
Child Extemalizing Behavior 4 1.8 35 16.0 180 82.2
Conflict 3 1.4 33 15.1 183 83 .6
SES 4 1.8 32 14.6 183 83.6

 

 

 

 

 

 

 

120

 

Appendix 0
Revised Data Analyses

Analyses were re-run to compare effects of changes in the distribution of subjects
into risk groups. In the initial dataset, all subjects were divided into risk groups based on
both their (DSM-III-R) [a] Lifetime Alcohol Diagnosis score and [b] score on the
Antisocial Behavior Scale (ASB). Lifetime Alcohol Diagnoses were given to families
based on the fathers’ diagnosis at Time 1. The ASB was administered to both fathers and
mothers at Time 1, with different cut-off scores for each sex indicating high versus low
antisociality (Fathers’ ASB s 24, mothers’ ASB s 18). Based on these two measures,
subjects were assigned into High Risk (high ASB alcoholic), Medium Risk (low ASB
alcoholic), and Low Risk (non-alcoholic) groups (separately for male subjects and female
subjects). From these risk groups, estimation for missing data was completed as
described in the Methods sections, and analyses were run.

For the sake of comparison, the process described above was repeated with one
difference: risk groups were reformulated based on subjects’ (DSM-III-R) [a] Lifetime
Alcohol Diagnosis score and [b] diagnosis of Antisocial Personality Disorder (ASPD).
Both Lifetime Alcohol Diagnoses and ASPD Diagnoses were given to families based on
the fathers’ data at Time 1. Data re-estimation based on these new groups followed, and
the same hypotheses were tested with this revised dataset. The revised distribution of
subjects into risk groups is presented in the following table, and displayed graphically in

Figures 4 and 5.

121

 

Distribution of Sample into Risk Groups

 

 

 

 

 

 

Original Dataset Revised Dataset
(groups based on ASB) (groups based on ASPD)
Fathers Mothers Fathers Mothers
N=103 N=115 N=103 N=115
Freq % Freq % Freq % Freq %

High Risk 25 24.3 16 13.9 17 16.5 19 16.5
Medium Risk 44 42.7 19 16.5 47 45.6 50 43.5
Low Risk 34 33.0 80 69.6 39 37.9 46 40.0

 

 

 

 

 

 

 

 

 

 

 

The distribution of subjects into risk groups changed due to the revised strategy
(based on ASPD diagnoses). For fathers, subjects’ assignment into risk groups became
more conservative: High Risk membership dropped by 7.8% (N=8), Medium Risk
membership grew by 2.9% (N=3), and Low Risk membership grew by 4.9% (N=5). For
mothers, however, the revised risk groups resulted in more subjects’ placement into
higher categories: High Risk mothers increased by 2.6% (N=3), Medium Risk mothers
increased by 27.0% (N=31), and Low Risk mothers decreased by 30.4% (N=35).

Distribution of Sampling Groups into Risk Groups

Subjects for the larger study, the MSU-UM Family Study, were recruited in
several ways: alcoholics were enlisted through the court system and within the
community, and matched non-alcoholic control subjects were recruited through door-to-
door canvassing. The distribution of these three types of sample groups into risk groups

changed from the original to the revised dataset (Figures 6 & 7).

122

 

5:39.35 985 me .2056“. ”v 059;
mm 9 en xm: 26.. E. o. 3 xm: 32 E 3 mm xm: :9: Hz 5 3:20
Emir mm<a_

 

 

88 Em 33 ...an Em 5282 88 Em :9:

 

 

 

 

 

 

 

 

 

28.9.5. motomofio .21 25 3:20 mEEEmm .6 5:39:55 5 059m

.3 Fuz. Em 338

82. Ex 528.2 :82. Ex :9...

C. O. : O.

CU .. .I Q. U. .I .0 .O .I .U
8 C. i C I 3 I. .C .0
...... .0 O. .0 .C I.

I I .- O. O.

o

 

°::: :

........................... om

 

 

£9.50! .. ..
82 558m ”mm” .............................................. 2.
82 58a

 

 

 

.om

 

125

 

 

' ECourt Alcs

EComm Alcs

_ nControls

 

 

 

 

 

oo‘oooaoeoooooaeoaTeao ..............
oooooooooooooooooooooooooooooooooo
oooooooooooooooooooooooooooooooooo
ooooooooooooooooooooooooooooooooooo
.......................................................
ooooooooooooooooooooooooooooooooooo
38338333338883838383833!!! 8888888888
ooooooooooooooooooooooooooooooooooo
oooooooooooooooooooooooooooooooo
oooooooooooooooooooooooooooooooooooo
ooooooooooooooooooooooooooooooooo

:3!!:3::8::3:=3::3323328333382888888

...................................
IIIIIIIIIIIIIIIIIIIIIIIIIIIIII

 

 

 

100

 

k (N=97) Low Risk (N=85)

##l’i'Ififili m

High Risk (N=36)Medium
Figure 7: Distribution of Sampling Grou

(Revised)

ps into Risk Categories

Correlations

Correlations among stress components are presented in Tables A1 (Revised
correlations among fathers’ stress components), A2 (Revised correlations among
mothers’ stress components), and A3 (Revised correlations between fathers’ and mothers’
variables)‘. Patterns in the revised correlation matrices strongly resemble the original
correlational matrices.

Similarly, revised correlations among Concurrent and Change variables changed
little from original correlations. The exceptions: for fathers, Concurrent Perception of
Child Difficulty no longer correlated with his physical aggression; for mothers,
Concurrent level of SES is more weakly related to Concurrent level of Worst-ever
Depression and Concurrent Conflict (the last correlation became nonsignificant).

Correlations among Concurrent stress components and parental physical aggression are

presented in Tables A4 and A5 (fathers’ and mothers’ variables, respectively.

RM-MANOVA
Hypotheses I and 2 predicted differences between high- and low-aggression
parents and their levels of stress components. In both the original and the revised
analyses, differences were not found between high- and low-aggression groups for

depression, alcohol consumption, and socioeconomic status, while there were differences

 

' Please note that revised tables (based on ASPD-formed risk groups) are numbered
to correspond to the original tables. An “A” in front of the table number indicates results based
on “revised group” data.

127

 

...

 

 

 

 

 

 

 

 

 

 

 

 

we. 8. 2. ~.. ~.. 2. 2. Na 2. mm. mm
D O O .0 O {O 9 GO
. 3. mm. «N. .... on. «N. an. mm. mm. .m.
9*
. .... .m. 8. .N. .... 3. 3. .... no
9.. I
. an. 8...- o... 8. n.. on. 8. NO
. m... 2. 2. .... .... 3. .0
{.9 l {O O.
. 3. .N. N... v... n... m...
.9. {I .9 Q!
. mm. .... ..N. .v. N:
O. .9 O.
. .m. an. hm. .1
9.. Oil
. no. 3. mm
0..
. an. «m.
. .m
m m aw. m m. 3 3 m 0 O O m H M m
. .. - m m m m m .. m M w m m. m

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.< 0.2....

 

128

 

 

 

 

 

 

 

 

 

 

 

 

.82.... 8.2.: 27...
9.! O. O O. O
. 2.. x..- 8. a. R. 2. 2. .m 2.- 2..- 2.- .... 2.- 2.. .... 3.- .... ....- 8- 2.. ,o.
900 {Oi CO. 99. O O.
. 8.- an. on. x. 3. 2.. 8. x..- 2.- z.- 2... 8.. 2.. ...- 3.- E.- v..- 8.- 2.- 38
i O. 999 O Q} Q 99 O O. O! O. l
. 2.- a.- ..m- 2. a. R. .... a. R. 2. .q ...... 8. 8. ...... 2. 8. 8. ..2
*5. if. O O. G O
. 2.. 2. 8.- S.- 8- 2..- o..- 8.- .... ....- N..- 2.. 3- 8.- 8.. 2.- 4n- mm
GO. DO 9
. 2. 3.. 8- S.- 8. 8.- 2.- 3.- S.- N..- 2.- a.- ..N- 2.- 2.- ....- S
O} O O
. .... 8.- 8.- 2.. e..- R..- 3.. 2- 8.- 2..- ....- ..~.- 2.- 2.- 8.- ..m
it. 9.. {GO .9! 99. G
. a. E. .4. 2.. a. 8.. 2.- 3. 8. 2. 8. 2. 8. R. 8
GO. 6! .9. Q
. 2.. a. .... .N. 8.. 8.- x..- S.- 2. .... 2. 2. 2. 8
09 to. GOO 99 O} O.
. em on. 2.. 8.- ...- 8. 2. 2. R. 3. .... 2. .o
2..... e: as
. 8. 2.. 2. 2. 2. .... 2.. .... 2. 2..- 2. 2.
m n w s m m m o o H H

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.....8 .2 2......

129

 

 

 

 

 

 

 

 

 

 

 

 

 

9.. G. G ii. .90 9 DO!
8. 2. a. 2. mm. 8. 2. .N. 2. 2. 8
O 9 it. GO. 9 0 9.}!
_ 8. 2. 2. an. 3”. Na 8. 3.. 2. .m
.9 I! 9. .09 DO
2 3. 8. R. 2. mm. a. 2.. a. no
999
_ 8. 8.- 2. 2. 8. 2. 8. 8
_ 8. 2. 8. s. 8.- 3.- 5
:9 O: :9 .9. {GO
_ a. a. 2.. 2.. an. n:
Oi} DO .90 .99
_ 2‘. R. a. 2.. N:
GO GOO GOO
_ 8. am. 8. E
GOO GOO
_ 8. a. 8
:9
_ 2.. mm
2 .m
m m w s w 3 3 b O O H M m
...-m. .. mm. mmmwmmm m

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

130

 

 

 

 

 

 

 

 

 

 

 

 

_oc.vno: S-Vmoo mc.vn§
o: o ...: 2.5 :5
_ em. if NV. NV. NV. 3. ma. 2. 2: Sr 86 8. 2r cor Sr ”-6 ¢_.- 2. ”Na. a8: 0—
. 2.3 .3... ...... 2..... .9 .
_ cm.- on. Vn. 3. he. no. NO. :6 V_.- mo: na- o_.- .- Er «Ni nN.- Nc.- war :r 0:8
o: 9.. ..i 3 o ... .9... :3 o: .9... :9 o.
_ Z...- th «N6 3. £- nN. _N. ON. on. 2. SR NM. NV. —V. 3. 0N. :. ..MN. 32
...: 2..... o o... o g u: 2.
_ 3.. Vb. 36 :.- V_.- Er 26 ON: nor 3. 2: NN.- ~V.- 6N6 V_.- m—.- 26 mm
... CG G
_ 2.. car 3. NO.- N_.- .2.- N—.- no. V_.- 2: o—.- VN.- NN.- 86 VO.- no: Nm
5 a .3.
_ :r 8.. .- ch 2v _N.- 8. 2: 3r 2.. eNr _N.- «av AZ.- hcc .m
2..... ...: ...: :3 o ¢ .5 o o: o: .3
_ a0. 3. an. 2.”. 2 . m- . S .- ~N. 0N. NN. m— . on. Vn. 0N. m0
2:. o: o: .3 ... o ... :9 o 5.. co :5
_ aw. Vm. mm. 5N. _ _. 0N. NN. ON. QN. cN. VN. mN. 3. NO
5 ... ..i .i a: o: 2.. o: ...: o:
_ ~N. MN. mN. 2. NN. o— . an. hm. VN. an. NM. —V. —U
...: o: .3 2: ...... ¢ ...... ...:
_ Nb. nV. S. aN. 2. 6V. aNr MN. 0N. Nm. 2. mm
m m w m. m. m. m m m 0 O O m H H
...-...... :4: :2:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3.8 .9. 2%.:

131

 

 

 

 

 

 

 

 

 

 

 

 

 

 

132

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8. ...- 2.. 8.- 3.- 3. 2. 3.- 3.- 8.- 2. 2. 8. 8.- 8.- 3.- 8. 3.- 3.- 8. 8.- w
... O. m.-
3. ...- 3. ...- 2.- 2.- 3. 8. 8. 8. 2. 2. 8. 2. 3. 3.- 8.- 8. 2. 8.- 3. W.
9 9 O O .m-
8.- 8.- 3. 2.- 2.- 2.- 2.. 3. .N. 8. 2. 8. 8. 8. 8 3. 8.- 8. 2. 3.- 2. m
... O Q a. O.
8. 8.- 8. 3.- 2.- 8.- 3. 2. 3. 3. 2. 2. 8. 3.- 3.- 3. 2. 3. 2. 8. 2. m
... .99 Gs. {Q Q Q... 99 G. O .29 O. W
3.- 2.- 3. 8.- 3.- 3.- 3. 3. 3. 2. 2. 2. 8.- 8.- 8. 3. 2.. 3. 3. 2. 8.
9 5 . 9 m
3. 2.- 2. 3.- 3.- 2.- 2. 2. 2. 8.- 8.- 3. 3.- 8.- 3.- 3. 3. 3. 2. 2. 2. m
5 9 to
3.- 2.- ... 2.- ...- 3.- 3. 2. 3. 2. 3. 8.- 3.- 8.- 8.- 3. 8. 8. 2. 8. 2.
O Q 99 2 Q 99 .9. m
... 2.- 2. 2.- ...- ...- 3. 3. 3. 2. 3. 2. 2.. 3. 8. 2. 8. 8. 3. 2. 8. .-
.. 2.. o... .2. :3 m
8. 2.- 2. 2.- 3.- 3.- 2. 3. 2. 3. 2. 3. 8.- 8.- ...- 3. 8. 8. 3. 3. 2.. w
352323 2.53 m N . m N . n 3 . n 3 . n N . n 3 . >
, .5532
o. 2..... a... 23 .205: 8.5. 3.325 2.5 839585 8..-.8.qu .25 8.325
83.5 .2053... ..o 8888.. .28.... .92.... .233 22.3. .850 .85....
.85....

 

 

 

 

 

 

 

 

 

2 0.5.

 

83222, .80—.82 u 2.8 > 8323.» ...-.05....— u 2.3 x

 

 

 

 

 

 

 

 

 

 

 

 

 

33.1.... 3.1.: 8v...
9 {99 O 9.0 {0.9 99.9 9
3. 2.. 3.- .... 8. .... 8. 2.- 8.- 8.- 8.- 2.. 8.- 2.- 2.- 8.. 2.. 3.. 8. 8.- 8.- a
QC 9.9 CO. 0*} {GO 9 O
3. ...- 8- 8. 3. 8. 3.- 8.- 2.- 8. 8.- 2.. 8. ...- 8.- 2.- 8.- 2.- 3.- 2.- 3.- m
its. 09 i. {O O Q Q.
2.- 3.- 8. 3.- 8.- 3.- 3. 8. 2. 2. 2. 3. 8.- 8.- 8. 2. 2. 8. ... 3.. 8. w
G *2. Q .52 ii. 99* O O O O O
3. 8. 3.- 3.. 3. 8. 8.- 8.- ...- 8.- 2.- 3.- 3. 2.- 2.- .... 3.- mm- 3.- 2.- 3.- n
Q .92. G {ii 999 {ii .6 G
3. 8. 3.- 8. 8. 2.. 8.- 8.- 8.- 8. 8.- 2.- 8.- 2.- 2.- 2.. 3.- 3.- 2.- .... 2.- 3
9* {9* a... 0.0 {OI Oi! O. O I S
3. 8. 8.- 8. 8. 8.. 3.. 8.- 8.- 2.- 2.- 3.- 3.- 3.- 8.- 3.- 2.- 2.- 2.- 2.- 3.- . a.
2 2!! 5.! 9.. G
3. 3.- 3. 3.- 2.- 2.- 8. 8. 8. 3. 2. 8. 2.- 8.- 2. 2. 8.- 3.- 2. 8. 2. m
G .2. GO! 999 O
8. 8. 3. 2.- 8.- 8. 8. 8. 8. 2. 3. 2. 8.- 2.- 8. 8. 3.- 8.- 2. ... 2. 3
Q .9. *9. *6! O 99 O M
8.- 3. 3. 2.- 8.- 8.- 8. 8. 2.. 8. 8. 2. 3.- 3- 8. 2. 3. 8.- 3. 2. 3. . n.
9! i {in {ii 0 G G
3.- 2.- 2. 2.- 9 2.- 8. 2. 3. 8. 8. 2. 3. 2. 3. 3. 8.- 8.- 2. 8. 2. ...
1 {IO 9!} DO I «flu.
8.- 2.- 2. 2.- 8.- 2.- 3. 2. 2. 8. 8. 3. 8.- 3. 3. ... 3.- 8.- 8. 8. 8. 3 m
3.. I. Q O *9! {GO 299 M.
8.- 3.- 3. 3.- 2.- 3.- 3. ... 8. 8. 3. 8. 8. 2. 2. 8. 3. 8.- 8. 3. 2. . m

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2:8 .8.- 0.8.:

133

Correlations between Fathers’ Concurrent Variables & Physical Agmsion

Table A4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a) a u u o-o t u 6 a m 9 O
>‘ 3 e .. '- -- = m = -—
he; Eg- Egg- E2 E3, g? E‘u’q’ < E
U U U ‘3 E U U U m
U U
Phys 1.00
Aggr
Current -.10 1.00
Level
of Dep
Level -.06 .42 1.00
Worst Dep "“
Current .08 .06 .08 1.00
Ale Consn
Current .16 .15 .31 .15 1.00
Child "
Diff
Current .26 .17 .08 -.02 .41 1.00
Conflict " """
Current .07 -.22 -.10 .01 —.02 -.60 1.00
SES *
ASB -.O7 .19 .07 .13 .10 .16 «.25 1.00
t.
Educ .05 -.16 -.ll -.05 -.04 .04 .59 -.36 1.00
it! It.
IQ -.l3 -.14 .01 .01 -.12 .17 .26 -.04 .43 1.00
it fit.
*p<.05 """p<.01 "*p<.001

134

 

Table A5

Correlatigfi between Mothers] Concurrent Variables & Physigg Aggssion

 

 

 

 

 

 

 

 

 

 

 

 

>5 : _ 0 E C H u- G .2 C m 3 l-I
538-33; E2 @9- E8 Ea“ B
02 0< ,E 90 Um
U 0
Phys 1.00
Aggr
Current .14 1.00
Dep
Level .09 .49 1.00
Worst Dep ***
Current .05 .27 .27 1.00
Alc Consn " *“'
Current .33 .26 .40 .01 1 .00
Child Diff ** "”" ”*
Current .22 .36 .26 .13 .38 1.00
Conflict * 13*. it ##t
Current .02 -.14 -.22 -.05 -.19 -.18 1.00
SES * *
ASB .12 .26 .42 .16 .21 .18 -.31 1.00
ii iii * it
Educ -.06 -.02 -.16 .03 -.17 .07 .56 -.30 1.00
it! it
IQ -.03 .10 -.01 .01 -.10 .14 .42 -.19 .56 1.00
ti! 1 it!

 

 

 

*p<.05 I""'p<.01 "*p<.001

135

Table A6

Remated Measures Multivariate Analysis of Variance for High- and Low-Aggression

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Parents
Within-Subjects
Multivariate F
Variable Parent Sex x Group x Time Parent Sex x Time
Group x Time Time

Current .02 .60 .05 6.56"
Depression
Worst-ever .63 .40 .09 7.16"
Depression
Alcohol .96 .76 8.44"” 13.33""
Consumption
Child .33 4.02“ .53 653"”
Extemalizing
Behaviors
Family Conflict 4.79" .53 .68 1.34
SES .51 .20 .01 1 1.47‘"
Between-Subjects

Variable Sex of Parent Group (High/Low Sex of Parent X Within + Residual

Aggression) Group Error
df F df F df F df F

Current 1 3.23 1 2.35 l 2.32 214 (22.82)
Depression
Worst-ever 1 . 12.15”" 1 .49 1 2.04 214 (l 10.02)
Depression
Alcohol 1 6.78" 1 .33 1 .14 214 (1485501 1)
Consumption
Child Ext 1 3.47 l 1508"" 1 .07 214 (89.53)
Behaviors
Conflict l 3.26 l 19.24"" 1 .05 214 (9.84)
SES 1 .01 l .50 1 .10 214 (37115.59)
Antisocial 1 27.72"" 1 .00 1 1.54 214 (76.27)
Behavior
Education 1 4.33‘ l .01 l .11 214 (3.55)
IQ 1 15.66“" 1 .48 l .16 214 (181.45)

 

 

 

 

 

 

 

 

 

 

136

for child externalizing behaviors and conflict in the family (i.e., high-aggression parents
reported consistently higher levels of child difficulty and conflict relative to low-
aggression parent). The results of the revised Repeated Measures Multivariate Analysis
of Variance are presented in Table A6.

Hierarchical Regression Analyses
Prom -to-Distal Model with Change Predictog.

The first set of regressions examined the prediction of parental physical
aggression from each change stress component independently (Table A15) and within the
proximal-to-distal model for change stress components (Table A16). These results of
these analyses did not change from the original: neither series significantly predicted
parental physical aggression

Next, regression was used to test the proximal-to-distal model of change stress
components with the inclusion of stable background characteristics first (Table A14) and
last (Table A17). Neither of these new analyses changed from the previous results: the
proximal-to-distal change model did not predict parental physical aggression. One minor
alteration: for fathers, change in worst-ever depression became a significant variable after
changes in child externalizing behaviors was added to the model, rather than after change
in alcohol consumption was added to the model. This alteration did not affect the
substantive results of the analyses.

Pro ' - i l wi 11
Regressions were rerun on Concurrent Variables to predict parental physical

aggression. The first set of regressions examined the prediction of parental physical

137

 

 

Table A15

Summg of Regression Analyses for Physical Aggression Regressed onto Change
Variables Separately

 

ModeWariable(s) Fathers Mothers

 

13 R2 13 R2

 

 

Change in Depression:

Current -.O37 .04 .040 .02
Worst -.188 .134

Change in Alcohol Consumption .010 .000 .141 .020

 

Change in Child Extemalizing Behaviors .111 .012 .065 .004

 

Change in Conflict in Family Environment .021 .000 -.016 .000

 

 

 

 

 

 

 

Change in SES .026 .001 .033 .001

 

* p s .05 ** p s .01 *** p s .001

138

Table A16

Summary of Hierarchical Regression Analysis for Physical Aggression Regressed onto
Change Predictor Variables

 

 

 

 

 

 

 

 

I Fathers I Mothers
Step Variable B R2 AR2 [3 R2 AR2
1 Change in Depression: n/a n/a
Current - .037 .04 .040 .02
Worst - .188 .134
2 Change in Depression:
Current - .040 .04 .00 .025 .04 .02
Worst - .191 .125
Change in Alcohol Consumption .030 .126
3 Change in Depression:
Current - .057 .06 .02 .023 .04 .00
Worst - .218" .122
Change in Alcohol Consumption .030 .124
Change in Child Extemalizing .156 .051
4 Change in Depression:
Current - .063 .06 .00 .024 .04 .00
Worst - .229" .123
Change in Alcohol Consumption .026 .128
Change in Child Extemalizing .150 .049
Change in Conflict .056 .018
5 Change in Depression:
Current - .063 .06 .00 .023 .04 .00
Worst - .229“ .121
Change in Alcohol Consumption .029 .131
Change in Child Extemalizing .148 .044
Change in Conflict .058 .019
Change in SES .025 .029

 

"' p s .05 ** p s .01 *" p s .001

139

Table A14

Summgy of Hierarchical Regression Analysis for Physical Aggression Regessed onto
Stable Background Characteristics and Change Predictor Variables

 

 

 

 

 

 

 

 

 

 

Fathers Mothers
Step Variable B R2 AR2 B R2 AR1
1 1Q - .178 .03 n/a .008 .01 n/a
Antisociality ~ .036 .109
Education .115 - .026
2 IQ - .193 .07 .04 -.003 .03 .02
Antisociality -.001 .081
Education .115 - .008
Change in Depression:
Current - .033 .020
Worst - .193 .114
3 IQ - .192 .07 .00 .004 .04 .01
Antisociality -.002 .055
Education .115 - .009
Change in Depression:
Current - .034 .020
Worst - .194 .113
Change in Alcohol Consumption .005 .113
4 IQ - .199 .09 .02 .001 .04 .00
Antisociality -.003 .052
Education .122 - .005
Change in Depression:
Current -.052 .018
Worst - .222‘ .112
Change in Alcohol Consumption .004 .112
Change in Child Extemalizing .164 .046
5 IQ -.l94 .10 .01 .000 .04 .00
Antisociality -.004 .050
Education .122 - .006
Change in Depression:
Current -.056 .019
Worst - .229‘ .113
Change in Alcohol Consumption .003 .115
Change in Child Extemalizing .160 .045
Change in Conflict .036 .013

 

140

(Table A14, con’t)

 

Step Variable B R3 AR2 B R1 AR’
6 IQ -.194 .10 .01 -.004 .04 .00

Antisociality -.004 .050

Education .122 - .011

Change in Depression:

Current - .056 .019

Worst - .228“ .109

Change in Alcohol Consumption .003 .119

Change in Child Extemalizing .160 .038

Change in Conflict .036 .016

Change in SE8 -.001 .035

 

*ps.05 "ps.01"‘"ps.001

141

Table A17

Summary of Hierarchical Regression Analysis for Physical Aggession Regressed onto
Change Predictor Variables & Stable Background Characteristics

 

 

Fathers l Mothers

 

 

 

 

 

 

 

 

 

Step Variable B R2 AR2 B R2 AR2
1 Change in Depression: .04 n/a .02 n/a
Current -.037 .040
Worst - .188 .134
2 Change in Depression: .04 .00 .04 .02
Current - .040 .025
Worst - .191 .125
Change in Alcohol Consumption .030 .126
3 Change in Depression: .06 .02 .04 .00
Current -.057 .023
Worst - .218" .122
Change in Alcohol Consumption .030 .124
Change in Child Extemalizing Behaviors .156 .051
4 Change in Depression: .06 .00 .04 .00
Current - .062 .024
Worst - .229" .123
Change in Alcohol Consumption .026 .128
Change in Child Extemalizing Behaviors .150 .049
Change in Conflict .056 .018
5 Change in Depression: .06 .00 .04 .00
Current - .063 .023
Worst - .229“ .121
Change in Alcohol Consumption .029 .131
Change in Child Extemalizing Behaviors .148 .044
Change in Conflict .058 .019
Change in SES .025 .029
6 Change in Depression: .10 .04 .04 .00
Current - .056 .019
Worst - .228‘ .109
Change in Alcohol Consumption .003 .119
Change in Child Extemalizing Behaviors .160 .038
Change in Conflict .038 .016
Change in SE8 -.001 .035
IQ - .194 -.004
Antisociality -.004 .050
Education .122 - .01 l
"‘ p s .05 " p s .01 "" p s .001

142

Table A18

Summary of Regression Analyses for Physical Aggression Regressed onto Concurrent
Variables Separately

 

 

 

 

 

 

 

 

 

 

 

Fathers Mothers
ModeWariable(s) B R2 B R2
Concurrent Depression:
Current -.015 .00 .125 .02
Worst -.015 .024
Concurrent Alcohol Consumption .170 .03 .048 .00
Concurrent Child Extemalizing Behaviors .228“ .05“ 330"" .11"*
Concurrent Conflict in Family Environment .341 *** .12"* .217“ .05"
Concurrent SES .014 .00 .015 .00

 

* p s .05 " p 3 .01 *" p s .001

143

 

Table A19

Summgy 9f Hierarchical Regression Analysis for Physical Aggression Reggssed Qnto

Concurrent Predictor Variables

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fathers Mothers
Step Variable B R2 AR2 B R2 AR2
1 Concurrent Depression:
Current - .015 .00 n/a .124 .02 n/a
Worst - .015 .024
2 Concurrent Depression:
Current - .022 .03 .02 .123 .02 .00
Worst - .026 .022
Concurrent Alcohol Consumption .174 .009
3 Concurrent Depression:
Current - .026 .08 .05 .087 .12" .10
Worst - .095 - .113
Concurrent Alcohol Consumption .144 .051
Concurrent Child Extemalizing Behavior .239“ 353""
4 Concurrent Depression:
Current - .076 .16" .08 .061 .13" .01
Worst - .059 - .112
Concurrent Alcohol Consumption .172 .045
Concurrent Child Extemalizing Behavior .098 .323"
Concurrent Conflict .321” .095
5 Concurrent Depression:
Current - .074 .16" .00 .063 .14" .01
Worst - .058 -.098
Concurrent Alcohol Consumption .172 .044
Concurrent Child Extemalizing Behavior .098 .330"
Concurrent Conflict .322" .104
Concurrent SES .012 .085
Trimmed Models
Fathers Mothers
B R2 Step ~ B R2
Concurrent Conflict .341‘" .116"” 1 Concurrent Child .331‘" .109‘”
Extemalizing Behavior
’ps.05 "p501 *“ps.001

144

 

 

 

 

aggression from each Concurrent Stress Component independently (Table A18) and

within the proximal-to-distal model for Concurrent Stress Components (Table A19). The

results of these analyses did not change from the original:

0 For both mothers and fathers, physical aggression was significantly predicted
independently by concurrent levels of child externalizing behaviors and conflict in
the home.

- The concurrent stress component model significantly predicted physical
aggression for both parents. For mothers, higher levels of concurrent child
externalizing behaviors predicted higher levels of parental physical aggression.
For fathers, higher levels of concurrent conflict predicted higher levels of paternal
physical aggression.

Next, regression was used to test the proximal-to-distal model of Concurrent

Stress Components with the inclusion of stable background characteristics first (Table

A7) and last (Table A20). For both fathers and mothers, these models significantly

predicted parental physical aggression in the original and the revised analyses. For

mothers, the results did not change.
Thus, changes in risk groups (High, moderate, and low) based on ASP did not
change results fi'om the original findings, based on risk groups formulated from the ASB.

Revised tables follow.

145

 

 

Table A7

Summary of Hierarchical Regression Analysis for Physical Aggression Regressed onto
S ble Back ound Characteri tics & Con nt Predictor V 'a les

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fathers Mothers
Step Variable B R’ AR’ B R1 AR’
1 IQ - .178 .031 n/a .008 .01 n/a
Antisociality - .036 .109
Education .115 - .026
2 IQ - .181 .032 .001 - .010 .03 .02
Antisociality - .032 .081
Education .1 14 - .025
Concurrent Depression:
Current - .024 .123
Worst .007 - .014
3 [Q - .181 .052 .02 - .010 .03 .00
Antisociality - .049 .080
Education .105 - .026
Concurrent Depression:
Current - .025 .122
Worst -.001 - .015
Concurrent Alcohol Consumption .143 .007
4 IQ - .141 .098 .046 .015 .13“ .10
Antisociality - .074 .076
Education .080 -.003
Concurrent Depression:
Current - .021 .081
Worst - .073 - .140
Concurrent Alcohol Consumption .120 .048
Concurrent Child Extemalizing Behavior .232‘ 351""
5 IQ - .247" .212” .114 .004 .13“ .02
Antisociality - .1 15 .067
Education .084 - .012
Concurrent Depression:
Current - .091 .059
Worst - .016 - .138
Concurrent Alcohol Consumption .148 .043
Concurrent Child Extemalizing Behavior .048 319""I
Concurrent Conflict 396"" .093

 

146

 

 

 

(Table A7, con’t)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6 IQ - .247“ .212" .00 - .023 .15" .02

Antisociality - .115 .080

Education .088 - .082

Concurrent Depression:

Current - .092 .066

Worst - .016 - .130

Concurrent Alcohol Consumption .148 .044

Concurrent Child Extemalizing .048 .315"l

Behavior 396"“I .126

Concurrent Conflict -.006 .152

Concurrent SES

Trimmed Models
Fathers Mother:
9 R, p R:

Q «193‘ .152‘" Concurrent Child .331'" .109'”
Concurrent Conflict .375": Extemalizing Behavior

(" p s .05 ” p s .01 W" p s .001)

147

 

Table A20

 

Means, Standard Deviations, & Range of Values for Mothers and Fathers

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Mothers (Valid N = 115) Fathers (Valid N = 103)
Variable Mean Std. Dev. Min. Max. Mean Std. Dev. Min. Max.
Physical 7.29 24.19 0 208 2.50 3.62 0 18
Aggression
ASB 10.27 6.79 1 35 16.71 10.47 2 56
Education 13.17 1.82 9 19 13.7 1.93 10 20
1Q 95.96 13.32 62 135 103.4 13.54 76 146
Beckl 2.77 3.29 0 14 2.13 2.36 0 11
Beck2 3.41 3.56 0 19 2.68 3.13 0 13
Beck3 3.39 3.86 0 20 2.8 3.06 0 13
Haml 19.04 8.34 8 48 15.98 6.84 6.8 41
Harn2 17.64 7.69 8 38 15.11 6.92 8 38
Ham3 16.78 8.71 0 47 13.97 6.92 8 36
QFV-Rl 584.58 2068.64 1 15750 2468.9 5209.94 1 21000
QFV-R2 79.011 3149.16 1 21000 1234.12 3100.27 1 21000
QFV-R3 262.16 1843.80 1 18900 3 16.38 1 122.97 1 8706.29
CBCLl 12.37 6.20 0 28 l 1.29 6.26 1 27
CBCL2 12.02 6.53 0 30 10.14 6.08 0 30
CBCL3 1 1.06 7.48 0 38 9.35 7.01 0 37
FESl 3.81 1.94 0 8 3.17 2.3 0 9
FESZ 3.71 1.95 0 9 3.25 2.13 0 8
FES3 3.85 2.12 0 9 3.39 2.28 0 9
SES] 317.13 103.11 171 640 318.43 105.03 171 640
SES2 344.97 133.60 162.5 770 343.78 128.09 162.5 770
SES3 339.27 127.61 177 866 339.91 127.92 177 866

 

148

 

 

REFERENCES

REFERENCES

Abidin, R.R., Jenkins, C.L., & McGaughey, MC. (1992). The relationship of
early family variables to children’s subsequent behavioral adjustment. Journal of
Clinical Child Psychology, 21 (l), 60 - 69.

Ammerman, R.T. (1990). Etiological models of child maltreatment: A
behavioral perspective. Behavior Modification, 14 (3), 230 - 254.

Ammerman, R.T. (1991). The role of the child in physical abuse: A reappraisal.
Violence and Victims, 6(2). 87 - 101.

Ayoub, C., & Jacewitz, M.M. (1982). Families at risk of poor parenting: A
descriptive study of sixty at risk families in a model prevention program. Child Abuse &
Neglect, 6, 413 - 422.

Ayoub, C.C., Willett, J .B., & Robinson, D.S. (1992). Families at risk of child
maltreatment: Entry-level characteristics and growth in family functioning during
treatment. ChildAbuse & Neglect, 16, 495 - 511.

Azar, S.T., Fantuzzo, J .W., & Twentyman, CT. (1984). An applied behavioral
approach to child maltreatment: Back to basics. Advances in Behavioral Research
Therapy, 6, 3 - 11.

Baker, B.L., & Heller, TL. (1996). Preschool children with externalizing
behaviors: Experience of fathers and mothers. Journal of Abnormal Child Psychology,
24(4), 513 - 532.

Barber, J .G., & Crisp, BR (1994). The effects of alcohol abuse on children and
the partner’s capacity to initiate change. Drug and Alcohol Review, 13, 409 - 416.

Barton, K., & Baglio, C. (1993). The nature of stress in child-abusing families: A
factor analytic study. Psychological Reports, 73, 1047 - 1055.

Bath, H.I., & Haapala, DA. (1993). Intensive family preservation services with
abused and neglected children: an examination of group differences. Child Abuse &
Neglect, 17, 213 - 225.

Bavolek, S.J., & Henderson, H.L. (1990). Child maltreatment and alcohol abuse:

Comparisons and perspectives for treatment. Journal of Chemical Dependency
Treatment, 3 (l), 165 - 184.

150

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, 114(3), 413 - 434.

Bertolli, J ., Morgenstem, H., & Sorenson, SB. (1995). Estimating the
occurrence of child maltreatment and risk-factor effects: Benefits of a mixed—design
strategy in epidemiologic research. Child Abuse & Neglect, 19 (8), 1007 - 1016.

Bingham, C.R., Fitzgerald, W.W., Fitzgerald, H.E., & Zucker, RA. (1996, June).
Prenatal hist arental alcoho i m and the develo me t of resc -a f
alcoholics. Poster presented at the Joint Scientific Meeting of the Research Society on
Alcoholism and the International Society for Biomedical Research on Alcoholism,
Washington, DC.

Bingham, C.R., Fitzgerald, H.E., Townsend, M.Z., & Zucker, RA. (1995, June).

hi1 ood tem rarnent and r 1 vi r: etiolo h i . Poster

presented at the 25th Anniversary Celebration of the National Institute on Alcohol Abuse
and Alcoholism, Steamboat Springs, CO.

Bingham, C.R., Fitzgerald, H., & Zucker, R. (1996). The two worlds of

ontogenetic risk in childhood: How fathers and mothers from alcoholio homos groeive

and contribute to their sons’ grommment gd behavior problems. Unpublished
manuscript, Michigan State University, East Lansing, MI.

Bingham, C.R, Miller, B.C., & Adams, GR. (1990). Correlates of age at first
sexual intercourse in a national sample of young women. Journal of Adolescent
Research, 5 (1), 18 - 33.

Bingham, C.R., Zucker, RA., & Fitzgerald, HE. (1996, in press). Risk load and
problem behavior development across the preschool years among sons of alcoholics:
Variation across subtypes of alcoholic families. Journal of Abnormal Psychology.

Bland, RC, & Om, H. (1986). Psychiatric disorders, spouse abuse and child
abuse. Acta Psychiatrica Belgica, 86 (4), 444 - 449.

Bolton, F.G., & MacEachron, A. (1986). Assessing child maltreatment risk in the
recently divorced parent-child relationship. Journal of Family Violence, 1 (3), 259 - 275.

Bronfcnbrenner. U- (1979). WWW. Cambridge,
MA: Harvard University Press.

151

Campbell, SB. (1997). Behavior problems in preschool children: Developmental
and family issues. In T.H. Ollendick and R.J. Prinz (Eds), Advances in Clinical Child
Psychology: Volume 19. New York: Plenum Press.

Campbell, S.B., Breaux, A.M., Ewing, L.J., & Szumowski, E.K. (1986).
Correlates and predictors of hyperactivity and aggression: A longitudinal study of parent-
refcrred problem preschoolers. Journal of Abnormal Child Psychology, 14 (2), 217 - 234.

Campbell, S.B., Cohn, J .F., & Meyers, T. (1993, in press). Depression in first-
time mothers: Mother-infant interaction and depression chronicity.

Cantrell, P.J., Carrico, M.F., Franklin, J .N., & Grubb, H.J. (1990). Violent
tactics in family conflict relative to familial and economic factors. Psychological
Reports, 66, 823 - 828.

Catalano, R. & Dooley, D. (1983). Health effects of economic instability: A test
of economic stress hypothesis. Journal of Health and Social Behavior, 24, 46 - 60.

Child Maltreatment (1993). Reports from the States to the National Center on
Child Abuse and Neglect. US. Department of Health and Human Services.

Cicchetti, D., & Howes, P.W. (1991). Developmental psychopathology in the
context of the family: Illustrations fi'om the study of child maltreatment. Canadian
Journal of Behavioural Science, 23 (3), 257 - 281.

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., & Toth, S.L. (1995). A developmental psychopathology
perspective on child abuse and neglect. Journal of the American Academy of Child and
Adolescent Psychiatry, 34(5), 541 - 565.

Cicchetti, D., & Toth, S.L. (1987). The application of a transactional risk model
to intervention with multi-risk maltreating families. Zero to Three, Bulletin of the
National Center for Clinical Infant Programs, 7(5), 1 - 8.

Creighton, TD. (1985). Children of alcoholics: A systems approach. Journal of
Child Care, 2 (5), 67 - 82.

Crnic, K. & Acevedo, M. (1995). Everyday stresses and parenting. In M. H.
Bomstein(Ed.), H - - - - _ - - -- . pp 277-
297. Mahwah, NJ. Lawrence Erlbaum Associates, Inc.

 

152

Crnic, K.A., Greenberg, M.T., Robinson, N.M., & Ragozin, AS. (1984).
Maternal stress and social support: Effects on the mother-infant relationship from birth to
eighteen months. American Journal of Orthopsychiatry, 54 (2), 224 - 23S.

Crockcnberg, S., & Lourie, A. (1996). Parents’ conflict strategies with children
and children’s conflict strategies with peers. Merrill-Palmer Quarterly, 42 (4), 495 - 518.

Cummings, EM. (1994). Marital conflict and children’s functioning. Social
Development, 3 (1), 16 - 36.

Dadds, M.R., Sanders, M.R., Morrison, M., & Rebgetz, M. (1992). Childhood
depression and conduct disorder: 11. An analysis of family interaction patterns in the
home. Journal of Abnormal Psychology, 101 (3), 505 - 513.

Dalgard, O.S., Bjark, S., & Tambs, K. (1995). Social support, negative life
events, and mental health. British Journal of Psychiatry, 166, 29 - 34.

Davies, W.H., Zucker, R.A., Fitzgerald, H.E., & Noll, RB. (1992, August).
Individual and familial characteristics related to number of alcoholic parents. Presented
at the meeting of the American Psychological Association, Washington, DC.

Davies, W.H., Zucker, R.A., Noll, R.B., & Fitzgerald, HE. (1991, under review).
Early socialization practices in alcoholic families: The relationship of child-rearing
patterns to demographics and parental psychopathology.

Davies, W.H., Zucker, R.A., Noll, RB, & Fitzgerald, HE. (1989, June).
Parental psychopathology and child-rearing practices in young alcoholic families.
Presented at the meetings of the Research Society on Alcoholism, Beaver Creek, CO.

Dinwiddie, S.H., & Bucholz, K.K. (1993). Psychiatric diagnoses of self-reported
child abusers. Child Abuse & Neglect, 17 (4), 465 - 476.

Donenberg, G., & Baker, BL. (1993). The impact of young children with
externalizing behaviors on their families. Journal of Abnormal Child Psychology, 21 (2),
179 - 198.

Dore, M.M. (1993). Family preservation and poor families: When
“Homebuilding” is not enough. Families in Society: The Journal of Contemporary
Human Services, November.

Dore, M.M., Doris, J .M., & Wright, P. (1995). Identifying substance abuse in
maltreating families: A child welfare challenge. Child Abuse & Neglect, 19 (5), 531 -
543.

153

Doumas, D., Margolin, G., & John, RS. (1994). The intergenerational
transmission of aggression across three generations. Journal of Family Violence, 9 (2),
157 - 175.

Dumka, L.E., & Roosa, M.W. (1993). Factors mediating fathers’ problem
drinking and mothers’ marital and personal adjustment. Journal of Family Psychology, 7
(3), 333 - 343.

Egan, K.J. (1983). Stress management and child management with abusive
parents. Journal of Clinical Child Psychology, 12 (3), 292 - 299.

Egeland, B., Breitcnbuchcr, M., & Rosenberg, D. (1980). Prospective study of
the significance of life stress in the etiology of child abuse. Journal of Consulting and
Clinical Psychology, 48 (2), 195 - 205.

Egeland, B., & Farber, EA. (1984). Infant-mother attachment: Factors related to
its development and changes over time. Child Development, 55, 753 - 771 .

Egeland, B., Jacobvitz, D., & Sroufe, LA. (1988). Breaking the cycle of abuse.
Child Development, 59, 1080 - 1088.

Ellis, D.A., Bingham, C.R., Zucker, R.A., & Fitzgerald, HE. (1996, under
review). Typological differences in patterns of risk load and child adaptation among
young alcoholic families.

Emery, RE. (1992). Family conflicts and their developmental implications: A
conceptual analysis of meanings for the structure of relationships. In C. Shantz & W.

Hartup’s Confliot in Child and Aoolescoot fivelooment, pp 270 - 298. London:

Cambridge University Press.

Famularo, R., Stone, K., Barnum, R, & Wharton, R. (1986). Alcoholism and
severe child maltreatment. American Journal of Orthopsychiatry, 56 (3), 481 - 485.

Ferleger, N., Glenwick, D.S., Gaines, R.R., & Green, AH. (1988). Identifying
correlates of reabuse in maltreating parents. Child Abuse & Neglect, 12, 41 - 49.

Fitzgerald, H.E., Sullivan, L.A., Ham, H.P., Zucker, R.A., Bruckel, S., Schneider,

A.M., & Noll, RB. (1993). Predictors of behavior problems in three-year-old sons of
alcoholics: Early evidence for the onset of risk. Child Development, fl (1), 110 - 123.

154

Fitzgerald, H.E., & Zucker, RA. (1995). Socioeconomic status and alcoholism:
The contextual structure of developmental pathways to addiction. In H.E. Fitzgerald,
B.M. Lester, and B. Zuckerman (Eds), Chilgen of Povegty, pp. 125 - 148. New York:
Garland Publishing.

Fitzgerald, H.E., Zucker, R.A., Maguin, E.T., & Reider, EB. (1994). Time spent
with child and parental agreement about preschool children’s behavior. Perceptual and
Motor Skills, .7_9 (1), 336 - 338.

Gaines, R., Sandgrund, A., Green, A.H., & Power, E. (1978). Etiological factors
in child maltreatment: A multivariate study of abusing, neglecting and normal mothers.
Journal of Abnormal Psychology, 87(5), 531 - 540.

Garbarino, J ., & Crouter, A. (1978). Defining the community context for parent-
child relations: The correlates of child maltreatment. Child Development, 49, 604 - 616.

Garbarino, J ., & Sherman, D. (1980). High-risk neighborhoods and high-risk
families: The human ecology of child maltreatment. Child Development, 51, 188 - 198.

Gaudirr, J .M., & Pollane, L. (1983). Social networks, stress, and child abuse.
Children and Youth Services Review, 5, 91 - 102.

Gelardo, M.S., & Sanford, BE. (1987). Child abuse and neglect: A review of the
literature. School Psychology Review, 16 (2), 137 - 155.

Gelfand, D.M., Teti, D.M., & Fox, C.E.R. (1992). Sources of parenting stress for
depressed and nondepressed mothers of infants. Journal of Clinical Child Psychology,
21 (3), 262 - 272.

Gelles, RJ. (1973). Child abuse as psychopathology: A sociological critique and
reformulation. American Journal of Orthopsychiatry, 43, 611 - 621 .

Gil, D. (1970). Vi 1 n
Sgtes. Cambridge, MA: Harvard University Press.

 

Glacharr, M. (1991). Child abuse: A social and cultural phenomenon. Early
Childhood Development and Care, 74, 95 - 102.

Gordis, E.B., Margolin, G., & John, RS. (1997). Marital aggression, observed

parental hostility, and child behavior during triadic family interaction. Journal of Family
Psychology, 11 (1), 76 - 89.

155

 

Haskett, M.E., Scott, S.S., & Farm, K.D. (1995). Child Abuse Potential
Inventory and parenting behavior: Relationships with high-risk correlates. Child Abuse &
Neglect, 19 (12), 1483 - 1495.

Herrenkohl, E.C., Herrenkohl, R.C., Rupert, L.J., Egolf, B.P., & Lutz, 1.6. (1995).
Risk factors for behavioral dysfunction: The relative impact of maltreatment, SES,
physical health problems, cognitive ability, and quality of parent-child interactions. Child
Abuse & Neglect, 19(2), 191 - 203.

Hillson, J .M.C., & Kuiper, NA. (1994). A stress and coping model of child
maltreatment. Clinical Psychology Review, 14 (4), 261 - 285.

Hoekstra, K.O’C. (1984). Ecologically defining the mistreatment of adolescents.
Children and Youth Services Review, 6, 285 - 298.

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

Holden, G.W., Coleman, S.M., & Schmidt, KL. (1995). Why 3-year-old
children get spanked: Parent and child determinants as reported by college-educated
mothers. Merrill-Palmer Quarterly, 41 (4), 431 - 452.

Houck, G.M., & King, MC. (1989). Child maltreatment: Family characteristics
and developmental consequences. Issues in Mental Health Nursing, 10, 193 - 208.

Hyman, LA. (1995). Corporal punishment, psychological maltreatment, violence,
and punitiveness in America: Research, advocacy, and public policy. Applied &
Preventive Psychologi, 4, 113 - 130.

Jameson, P., Kulcsar, E., Gelfand, D., & Teti, D. (May, 1992). Context, quality
of interaction, and developmental outcome in children with depressed mothers.
Presented at the meeting of the Western Psychological Association.

Jansen, R.E., Fitzgerald, H.E., Ham, H.P., & Zucker, RA. (1995). Pathways to
risk: Temperament and behavior problems in three- to five-year old sons of alcoholics.
Alcoholism Clinical and Experimental Research, 19(2), 510 - 509.

Jouriles, E.N., Barling, J ., & O’Leary, K.D. (1987). Predicting child behavior

problems in maritally violent families. Journal of Abnormal Child Psychology, 15 (2),
165 — 173.

156

Jouriles, E.N., & LeCompte, SH. (1991). Husbands’ aggression toward wives
and mothers’ and fathers’ aggression toward children: Moderating effects of child gender.
Journal of Counseling and Clinical Psychology, 59 (l), 190 - 192.

Justice, B., & Duncan, DR (1976). Life crisis as a precursor to child abuse.
Public Health Reports, 91 (2), 110 - 115.

Kessler, R.C., & Magee, WJ. (1994). Childhood family violence and adult
recurrent depression. Journal of Health and Social Behavior, 35, 13 - 27.

Kolko, D.J., Kazdin, A.E., Thomas, A.M., & Day, B. (1993). Heightened child
physical abuse potential: Child, parent, and family dysfunction. Journal of Interpersonal
Violence, 8 (2), 169 - 192.

Kotch, J .B., Browne, D.C., Ringwalt, C.L., Stewart, P.W., Ruina, E., Holt, K.,
Lowman, B., & Jung, J -W. (1995). Risk of child abuse or neglect in a cohort of low-
income children. Child Abuse & Neglect, 19(9), 1115 - 1130.

Koverola, C., Manion, I., & Wolfe, D. (1985). A microarralysis of factors
associated with child-abusive families: Identifying individual treatment priorities.
Behavioral Research Therapy, 23 (5), 499 - 506.

Kowal, L.W., Kottrneier, C.P., Ayoub, C.C., Komives, J .A., Robinson, D.S., &
Allen, J .P. (1989). Characteristics of families at risk of problems in parenting: Findings
from a home-based secondary prevention program. Child Welfare, 68 (5), 529 - 538.

Krauthamer, C. (1979). Maternal attitudes of alcoholic and nonalcoholic upper
middle class women. International Journal of the Addictions, 14, 639 - 644.

Krech, K.H., & Johnston, C. (1992). The relationship of depressed mood and life
stress to maternal perceptions of child behavior. Journal of Clinical Child Psychology,
21 (2), 115 -122.

Linsky, A.S., Straus, M.A., & Colby, J .P. (1984). Stressful events, stressful
conditions and alcohol problems in the United States: A partial test of Bales’ theory.
Journal of Studies on Alcohol, 46 (l), 72 - 80.

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 (4), 313 - 328.

157

Mash, E.J., Johnston, C., & Kovitz, K. (1983). A comparison of the mother-child
interactions of physically abused and non-abused children during play and task situations.
Journal of Clinical Child Psychology, 12 (3), 337 - 346.

McCloskey, L.A., Figueredo, A.J., & Koss, MP. (1995). The effects of systemic
family violence on children’s mental health. Child Development, 66, 1239 - 1261.

McCord, J. (1993). Considerations of causes in alcohol-related violence. In

Alcohol and integmrsonal violence: Fostering multidisciolmm' firsmo'ves (NIAAA
Monograph 24; Publication No. 93-3496). Rockville, MD: NIH.

McCord, J. (1988). Parental behavior in the cycle of aggression. Psychiatry, 51,
14 - 23.

Miller, B.C. & Bingham, CR (1989). Family configuration in relation to the
sexual behavior of female adolescents. Journal of Marriage and the Family, 51, 499 -
506.

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

Mollerstrom, W.W., Patchner, M.A., & Milner, J .S. (1992). Family functioning
and child abuse potential. Journal of Clinical Psychology, 48 (4), 445 - 454.

Muller, R.T., Fitzgerald, H.E., Sullivan, L.A., & Zucker, R.A. (1994). Social
support and stress factors in child maltreatment among alcoholic families. Canadian
Journal of Behavioural Science, _2_6 (3), 438 - 461.

O’Keefe, M. (1994). Linking marital violence, mother-child/father-child
aggression, and child behavior problems. Journal of Family Violence, 9(1), 63 - 78.

Paolino, T.J., McCrady, B.S., & Kogan, KB. (1978). Alcoholic marriages: A
longitudinal empirical assessment of alternative theories. British Journal of Addiction,
73, 129 - 138.

Patterson, GR. (1982). A social learning approach to family intervention: Vol. 3.
Coercive Family Process. Eugene, OR: Castalia

Patterson, G.R., Dishion, T.J., & Bank, L. (1984). Family interaction: A process
model of deviancy training. Aggressive Behavior, 10, 235 - 267.

158

 

Perry, M.A., Wells, E.A., & Doran, L.D. (1983). Parent characteristics in
abusing and nonabusing families. Journal of Clinical Child Psychology, 12 (3), 329 -
336.

Pianta, B. (1984). Antecedents of child abuse: Single and multiple factor models.
School Psychology International, 5, 151 - 160.

Pianta, R.C., & Egeland, B. (1990). Life stress and parenting outcomes in a
disadvantaged sample: Results of the mother-child interaction project. Journal of
Clinical Child Psychology, 19 (4), 329 - 336.

Pogge, D.L. (1992). Risk factors in child abuse and neglect. Journal of Social
Distress and the Homeless, 1 (3/4), 237 - 248.

Reider, E.E., Zucker, R.A., Maguin, E.T., Noll, R.B., & Fitzgerald, HE. (1989,
August). Alcohol involvement and violence towards children among high risk families.
Presented at the meeting of the American Psychological Association, New Orleans.

Reider, E.E., Zucker, R.A., Maguin, E.T., & Fitzgerald, HE. (1996). Alcohol
involvement and violence towards children among alcoholic families in the community.

Roberts, G.C., Block, J .H., & Block, J. (1984). Continuity and change in parents’
child-rearing practices. Child Development, 55, 586 - 597.

Rutter, M. (1994). Family discord and conduct disorder: Cause, consequence, or
correlate? Journal of Family Psychology, 8 (2), 170 - 186.

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

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

Schaufeli, W.B., & Vaaneren, NW. (1993). Success and failure in the labor
market. Journal of Organizational Behavior, 14, 559 - 572.

Schinke, S.P., Schilling, R.F., Kirkham, M.A., Gilchrist, L.D., Barth, R.P., Blythe,
BJ. (1986). Stress management skills for parents. Journal of Child and Adolescent
Psychotherapy, 3 (4), 293 - 298.

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 (5), 893 - 903.

159

 

Shaw, D.S., Owens, E.B., Vondra, J .I., Keenan, K., & Winslow, EB. (1996).
Early risk factors and pathways in the development of early disruptive behavior problems.
Development and Psychopathology, 8 (4), 679 - 699.

Silber, S. (1990). Conflict negotiation in child abusing and nonabusing families.
Journal of Family Psychology, 3 (4), 368 - 384.

Smith, J ., Berthelsen, D., & O’Connor, 1. (1997). Child adjustment in high
conflict families. Child: Care, Health, and Development, 23 (2), 113 - 133.

Smith, P.D., Rivers, P.C., & Stahl, K.J. (1992). Family cohesion and conflict as
predictors of drinking patterns: Beyond demographics and alcohol expectancies. Family
Dynamics of Addiction Quarterly, 2 (2), 61 - 69.

Spanking kids has negative effects. (1997, August 17). W
Journal p. 10A.

 

Spicer, S.J., & Franklin, C. (1994). Exploratory effects of social support, stress,
and locus of control on the conflict tactics of parents at risk for child maltreatment.
Journal of Social Service Research, 19 (3 -4), 1 - 22.

Steinberg, L.D., Catalano, R., & Dooley, D. ( 1981). Economic antecedents of
child abuse and neglect. Child Development, 52, 975 - 985.

Steinglass, P. (1987). The alcoholic family. New York: Basic Books.

Steinhausen, H-C. (1995). Children of alcoholic parents: A review. European
Child and Adolescent Psychiatry, 4 (3), 143 - 152.

Strassbcrg, 2., Dodge, K.A., Bates, J .E., & Pettit, G.S. The longitudinal relation
between parental conflict strategies and children’s sociometric standing in kindergarten.
Merrill-Palmer Quarterly, 38 (4), 477 - 493.

Straus, MA. (1979). Measuring intrafamily conflict and violence: The Conflict
Tactics (CT) scales. Journal of Marriage and the Family, 35, 75 - 88.

Straus, M.A., Gelles, R.J., & Steinmetz, SK. (1980). mm
Wiggly Garden City, NJ: Anchor Books.

Susman, E.J., Trickett, P.K., Iannotti, R.J., Hollenbeck, B.E., & Zahn-Waxler, C.
(1985). Child-rearing patterns in depressed, abusive, and normal mothers. American
Journal of Orthopsychiatry, 55, 237 - 251.

160

Takeuchi, D.T., Williams, D.R., & Adair, R.K. (1991). Economic stress in the
family and children’s emotional and behavioral problems. Journal of Marriage and the
Family, 53, 1031 - 1041.

Thoits, P., & Hannan, M. (1979). Income and psychological distress: The impact
of an income-maintenance experiment. Journal of Health and Social Behavior, 20, 120 -
138.

Trickett, P.K., Aber, J .L., Carlson, V., & Cicchetti, D. (1991). Relationship of
socioeconomic status to the etiology and developmental sequelae of physical child abuse.
Developmental Psychology, 27(1), 148 - 158.

United States Department of Health and Human Services. (1993). Alcohol a_rld
Health (DI-IHS Special Report; Publication No. ADM-28l-91-0003). Virginia: EEI.

United States Department of Health and Human Services (1993). M

intemrsofl violonce: Fogoring mflo'digiolmary' gromtivos (NIAAA Monograph 24;
Publication No. 93-3496). Rockville, MD: NIH.

van den Boom, D.C., & Hoeksma, J .B. (1994). The effect of infant irritability on
mother-child interaction: A growth-curve analysis. Developmental Psychology, 30(4),
581 - 590.

Vietze, P., Falsey, S., Sandler, H., O’Conner, S., Altemier, W.A. (1980).
Transactional approach to prediction of child maltreatment. Infant Mental Health
Journal, 1 (4), 248 - 261.

Vondra, 1.1. (1990). Sociological and ecological factors. In RT. Ammerman &
M. Hasen (Eds), hildre ' '
and neglgt. New York: Plenum Press.

Weissman Wind, T., & Silvern, L. (1994). Parenting and family stress as
mediators of the long-term effects of child abuse. Child Abuse & Neglect, 18 (5), 439 -
453.

Wekerle, C., & Wolfe, DA. (1993). Prevention of child physical abuse and
neglect: Promising new directions. Clinical Psychology Review, 13, 501 - 540.

Wenger, S., Berg-Cross, L., & Berg-Cross, G. (1980). Parents’ judgements of
children’s aggressive behavior. Merrill-Palmer Quarterly, 26(2), 161 - 169.

161

 

Whipple, E.E., Fitzgerald, H.E., & Zucker, R.A. (1995). Parent-child interactions
in alcoholic and nonalcoholic families. American Journal of Orthopsychiatry, 65 (1), 153
- 159.

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

Whipple, E.E., & Wilson, SR (1994). Evaluation of a parent education and
support program for families at risk of physical child abuse. Families in Society, 77 (4),
227 - 239.

Widom, CS. (1993). Child abuse and alcohol use and abuse. In Alghol ood

interpersonal violence: Fogering multidisciplim oersgtives (NIAAA Monograph 24;
Publication No. 93-3496). Rockville, MD: NIH.

Willett, J .B., Ayoub, C.C., & Robinson, D. (1991). Using growth modeling to
examine systematic differences in growth: An example of change in the functioning of
families at risk of maladaptive parenting, child abuse, or neglect. Journal of Consulting
and Clinical Psychology, 59(1), 38 - 47.

Wolfe, DA. (1993). Prevention of child neglect: Emerging issues. Criminal
Justice and Behavior, 29 (1), 90 - 111.

Wolfe, DA. (1987). Child b :Im 1ic ti re '1 develo
osychooatholog. Newbury Park, California: Sage Publications.

Wolfe, DA. (1985). Child-abusive parents: An empirical review and analysis.
Psychological Bulletin, 97 (3), 462 - 482.

Wolfe, D.A., Edwards, B., Manion, I., & Koverola, C. (1988). Early intervention
for parents at risk of child abuse and neglect: A preliminary investigation. Journal of
Consulting and Clinical Psychology, 56 (1), 40 - 47.

Wolfirer, G.D., & Gelles, R.J. (1993). A profile of violence toward children: A
national study. Child Abuse & Neglect, 17, 197 - 212.

Zucker, RA., Bingham, C.R, Fitzgerald, H.E., Schulenberg, J .E., Wadsworth,
K.N., Burchart, A.T., Maggs, J .L., Donovan, J .E., Leonard, K.E., Eiden, R.D., Schuckit,
M.A., & Sher, K.J. (1996). Longitudinal research on alcohol problems: The flow of risk,
problems, and disorder over time. Alcohol Clinical Experimental Research, 20 (8 suppl),
93A - 95A.

162

Zucker, R.A., Ellis, D.A., Bingham, C.R., & Fitzgerald, HE. (1996). The
development of alcoholic subtypes: Risk variation among alcoholic families during the
early childhood years. Alcohol Health & Research World, E (l), 46 - 54.

Zucker, R.A., & Fitzgerald, HE. (1991). Early developmental factors and risk
for alcohol problems. Alcohol Health & Research World, 15 (1), 18 - 24.

Zucker, R.A., Fitzgerald, H.E., & Moses, H.D. (1995). Emergence of alcohol
problems and the several alcoholisms: A developmental perspective on etiologic theory
and life course trajectory. In D. Cicchetti and D. J. Cohen (Eds. ) ,Womal

. . 2.- _‘ an“. 677- 711. J.Wiley&

 

Sons: New York.

Zucker, RA., Ojala, C.J., Fitzgerald, H.E., & Noll, RB. (1994). Evidence for the
early onset of risk in alcoholic families, II: The interplay between parent psychopathology
and parent-child interactional process. To appear in: R. Vrasti (Ed.), Algoholismnlflew
W. Timis: Romania.

Zucker, RA., Reider, E. E., Ellis, D. A., & Fitzgerald, H. E. (1995). Alcohol
involvement over the life span: A developmental perspective on etiology and course. In
PC RrversandERShore(Eds), =

W Westpert, CT: Greenwood ’—

Zuravin, SJ. (1989). The ecology of child abuse and neglect: Review of the
literature and presentation of data. Violence and Victims, 4 (2), 101 - 120.

 

163

 

l’lICHIGRN STATE UNIV. LIBRnRIES
‘lWWIIWIIIWIWW“WWIWIIHHIUIHIIWWI
31293017103809